Pay Ratio Disclosure

 
CONTENT

Federal Register, Volume 80 Issue 159 (Tuesday, August 18, 2015)

Federal Register Volume 80, Number 159 (Tuesday, August 18, 2015)

Rules and Regulations

Pages 50103-50187

From the Federal Register Online via the Government Publishing Office www.gpo.gov

FR Doc No: 2015-19600

Page 50103

Vol. 80

Tuesday,

No. 159

August 18, 2015

Part III

Securities and Exchange Commission

-----------------------------------------------------------------------

17 CFR Parts 229, 240, and 249

Pay Ratio Disclosure; Final Rule

Page 50104

-----------------------------------------------------------------------

SECURITIES AND EXCHANGE COMMISSION

17 CFR Parts 229, 240, and 249

Release Nos. 33-9877; 34-75610; File No. S7-07-13

RIN 3235-AL47

Pay Ratio Disclosure

AGENCY: Securities and Exchange Commission.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: We are adopting amendments to Item 402 of Regulation S-K to implement Section 953(b) of the Dodd-Frank Wall Street Reform and Consumer Protection Act. Section 953(b) directs the Commission to amend Item 402 of Regulation S-K to require disclosure of the median of the annual total compensation of all employees of a registrant (excluding the chief executive officer), the annual total compensation of that registrant's chief executive officer, and the ratio of the median of the annual total compensation of all employees to the annual total compensation of the chief executive officer. The disclosure is required in any annual report, proxy or information statement, or registration statement that requires executive compensation disclosure pursuant to Item 402 of Regulation S-K. The disclosure requirement does not apply to emerging growth companies, smaller reporting companies, or foreign private issuers.

DATES: Effective Date: October 19, 2015.

Compliance Date: Registrants must comply with the final rule for the first fiscal year beginning on or after January 1, 2017.

FOR FURTHER INFORMATION CONTACT: John Fieldsend, Special Counsel in the Office of Rulemaking, at (202) 551-3430, in the Division of Corporation Finance; 100 F Street NE., Washington, DC 20549.

SUPPLEMENTARY INFORMATION: We are adopting amendments to Item 402 \1\ of Regulation S-K \2\ and a conforming amendment to Form 8-K \3\ under the Securities Exchange Act of 1934 (the ``Exchange Act'').\4\

---------------------------------------------------------------------------

\1\ 17 CFR 229.402.

\2\ 17 CFR 229.10 et seq.

\3\ 17 CFR 249.308.

\4\ 15 U.S.C. 78a et seq.

---------------------------------------------------------------------------

Table of Contents

  1. Background

    1. Section 953(b) of the Dodd-Frank Act

    2. Summary of the Proposed Rule

    3. Summary of the General Comments on the Proposed Rule

    4. Summary of Changes in the Final Rule

    1. Non-U.S. Employee Exemptions and Additional Permitted Disclosure

    2. Employees of Consolidated Subsidiaries

    3. Employed on Any Date Within Three Months of the Last Completed Fiscal Year

    4. Identifying the Median Employee Once Every Three Years

    5. Initial Compliance Date

    6. Transition Period for New Registrants

    7. Additional Transition Periods

  2. Discussion

    1. Scope of Final Rule

      1. Pay Ratio Disclosure Requirements Under New Paragraph (u) to

        Item 402 of Regulation S-K

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

      2. Pay Ratio Disclosure in Filings That Require Item 402 of

        Regulation S-K Information

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

      3. Excluded Registrants--Smaller Reporting Companies, Foreign Private Issuers, MJDS Filers, and Emerging Growth Companies

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

    2. Requirements of Final Rule

      1. ``All Employees'' Covered Under the Rule

        a. Types of Employees

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        b. Employed on Any Date Within Three Months of the Last Completed Fiscal Year

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        c. Employees Located Outside the United States

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        (a) Non-U.S. Employees Generally

        (b) Foreign Data Privacy Law Exemption

        (c) De Minimis Exemption

        (d) Cost-of-Living Adjustment

        d. Employees of Consolidated Subsidiaries

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        e. Any PEO Compensation in the Last Full Fiscal Year

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        f. Additional Information is Permissible

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        g. Annualizing Permanent Employees is Permissible, but Other Compensation Adjustments are Prohibited

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

      2. Identifying the Median Employee and Calculating Annual Total Compensation

        a. Identifying the Median Employee

        i. Once Every Three Years

        (a) Proposed Rule

        (b) Comments on the Proposed Rule

        (c) Final Rule

        ii. Using Annual Total Compensation, Another Consistently Applied Compensation Measure, Statistical Sampling, Reasonable Estimates, or Other Reasonable Methods

        (a) Proposed Rule

        (b) Comments on the Proposed Rule

        (i) Flexibility

        (ii) Statistical Sampling

        (iii) Consistently Applied Compensation Measures

        (iv) The ``Median'' Employee

        (c) Final Rule

        (i) Flexibility

        (ii) Statistical Sampling

        (iii) Consistently Applied Compensation Measures

        (iv) The ``Median'' Employee

        b. Calculating Annual Total Compensation

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

      3. Disclosure of Methodology, Assumptions, and Estimates

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

      4. Meaning of ``Annual''

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

      5. ``Filed'' Not ``Furnished''

        a. Proposed Rule

        b. Comments on the Proposed Rule

        c. Final Rule

      6. Timing of Disclosure

        a. Updating Pay Ratio Disclosure for the Last Completed Fiscal Year

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        b. Omitting Salary or Bonus Information for the PEO in Reliance on

        Instruction 1 to Item 402(c)(2)(iii) and (iv), and Technical Amendment to Item 5.02(f) of Form 8-K

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        c. Initial Compliance Date

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        d. Transition Period for New Registrants

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

        e. Additional Transition Periods

        i. Proposed Rule

        ii. Comments on the Proposed Rule

        iii. Final Rule

  3. Economic Analysis

    1. Background

    2. Baseline

    3. Economic Effects From Mandated Disclosure Requirements

      1. Benefits

      2. Costs

        a. General

        b. Compliance Cost Estimates in Comment Letters

        i. Center on Executive Compensation Survey

        ii. Chamber of Commerce Survey

        Page 50105

        iii. Other Specific Comments

        c. Quantification of Compliance Costs

        d. Indirect Costs

      3. Other Economic Effects

    4. Economic Effects From Exercise of Discretion

      1. General

      2. Implementation Choices and Alternatives

      a. Filings Subject to Pay Ratio Disclosure Requirements

      b. Registrants Subject to the Pay Ratio Disclosure Requirements

      c. Employees Included in the Determination of the Median

      i. Types of Employees

      ii. Workers Not Employed by the Registrant

      (i.e., ``Leased'' Workers'')

      iii. Employees of Consolidated Subsidiaries

      iv. Employees Located Outside of the United States

      v. Foreign Data Privacy Law Exemption

      vi. De Minimis Exemption

      vii. Calculation Date

      d. Adjustments to the Compensation of Employees

      e. Frequency of Identifying the Median Employee

      f. Method of Identifying the Median Employee

      i. Consistently Applied Compensation Measure

      ii. Statistical Sampling

      g. Disclosure of Methodology, Assumptions, and Estimates

      h. Determination of Total Compensation

      i. Defining ``Annual''

      j. Updating the Pay Ratio Disclosure for the Last Completed

      Fiscal Year

      k. Status of Disclosure as ``Filed''

      l. Compliance Date

      m. Transition Periods

  4. Paperwork Reduction Act

    1. Background

    2. Summary of Information Collections

    3. Summary of Comment Letters and Revisions to Proposal

    4. Revisions to PRA Reporting and Cost Burden Estimates

      1. Estimated Internal Burden Hours

      2. Estimated Cost Burdens

      3. Estimated Cost and Hour Burdens for Each Collection of Information

      a. Regulation S-K

      b. Form 10-K

      c. Form 8-K

      d. Proxy Statements on Schedule 14A

      e. Information Statements on Schedule 14C

      f. Form S-1

      g. Form S-4

      h. Form S-11

      i. Form N-2

      j. Form 10

    5. Summary to Changes to Annual Compliance Burden in Collection of Information

  5. Final Regulatory Flexibility Act Certification

  6. Statutory Authority

  7. Background

    1. Section 953(b) of the Dodd-Frank Act

      Section 953(b)(1) of the Dodd-Frank Wall Street Reform and Consumer Protection Act (the ``Dodd-Frank Act'') \5\ directs us to amend Item 402 of Regulation S-K (``Item 402'') \6\ to require each registrant, other than an emerging growth company, as that term is defined in Section 3(a) of the Exchange Act, to disclose in any filing of the registrant described in Item 10(a) of Regulation S-K (or any successor thereto): \7\ (A) The median of the annual total compensation of all employees of the registrant, except the chief executive officer (``CEO'') (or any equivalent position) of the registrant; (B) the annual total compensation of the CEO (or any equivalent position) of the registrant; and (C) the ratio of the median of the total compensation of all employees of the registrant to the annual total compensation of the CEO of the registrant. Section 953(b)(2) specifies that, for purposes of Section 953(b), ``total compensation'' of an employee of a registrant shall be determined in accordance with Item 402(c)(2)(x) of Regulation S-K as in effect on the day before the date of enactment of the Dodd-Frank Act. As discussed in detail below, we are adopting amendments to Item 402 to implement Section 953(b).\8\ We refer to this disclosure of the median of the annual total compensation of all employees of the registrant, the annual total compensation of the principal executive officer (``PEO'') of the registrant,\9\ and the ratio of the two amounts as ``pay ratio'' disclosure.

      ---------------------------------------------------------------------------

      \5\ Public Law 111-203, sec. 953(b), 124 Stat. 1376, 1904 (2010), as amended by Public Law 112-106, sec. 102(a)(3), 126 Stat. 306, 309 (2012). Section 102(a)(3) of the JOBS Act amended Section 953(b) of the Dodd-Frank Act to provide an exemption for registrants that are emerging growth companies as that term is defined in Section 3(a) of the Exchange Act.

      \6\ 17 CFR 229.402. As discussed in greater detail below, consistent with Section 953(b), the final rule requires a registrant to provide the pay ratio disclosure in any filing described in Item 10(a) of Regulation S-K that calls for executive compensation disclosure under Item 402, including annual reports on Form 10-K, registration statements under the Securities Act and Exchange Act, and proxy and information statements, to the same extent that these forms require compliance with Item 402. Therefore, any company that provides such a filing is subject to the final rule. Section 953(b) refers to any such company as an ``issuer.'' In this release, to be consistent with other releases, we generally refer to such a company as a ``registrant.'' For the purposes of this release, unless otherwise expressly specified, these terms are used interchangeably.

      \7\ 17 CFR 229.10(a).

      \8\ On September 18, 2013, we proposed amendments to implement Section 953(b). See Pay Ratio Disclosure, Release No. 33-9452 (Sept. 18, 2013) 78 FR 60560 (``Proposing Release'').

      \9\ The term ``CEO'' in the executive compensation rules was replaced by the term ``PEO'' as part of the 2006 amendments to Item 402 in order to maintain consistency with the nomenclature used in Item 5.02 of Form 8-K. See Executive Compensation and Related Person Disclosure, Release No. 33-8732A, n. 326 (Aug. 29, 2006) 71 FR 53158 (``2006 Adopting Release''). Consistent with the language of current Item 402, both the proposed rule and the final rule use the term ``PEO'' in lieu of ``CEO.''

      ---------------------------------------------------------------------------

      Congress did not expressly state the specific objectives or intended benefits of Section 953(b), and the legislative history of the Dodd-Frank Act also does not expressly state the Congressional purpose underlying Section 953(b).\10\ As discussed below, based on our analysis of the statute and comments received, we believe Section 953(b) was intended to provide shareholders with a company-specific metric that can assist in their evaluation of a registrant's executive compensation practices. Accordingly, we have sought to tailor the final rule to meet that purpose while avoiding unnecessary costs.

      ---------------------------------------------------------------------------

      \10\ See letters from National Association of Manufacturers (Jul. 6, 2015) (``NAM II'') (stating that it concurs with our conclusion in the Proposing Release that neither the statute nor the related legislative history directly states the objectives or intended benefits of the provision) and WorldatWork (Jul. 6, 2015) (``WorldatWork II'').

      ---------------------------------------------------------------------------

      In informing our understanding of the Congressional purpose of Section 953(b), we have considered the surrounding provisions of the Dodd-Frank Act \11\ as well as the comments that we received during this rulemaking. Subtitle E of Title IX of the Dodd-Frank Act, headed --``Accountability and Executive Compensation'' is, as explained in the Conference Report for the legislation, ``designed to address shareholder rights and executive compensation practices.'' \12\ Its provisions, including Section 953(b), address various aspects of executive compensation with a focus on encouraging shareholder engagement in executive compensation matters by, among other things, increasing the transparency of compensation. In Section 951, for example, Congress required companies to provide for periodic shareholder votes on executive compensation. In implementing Congress's directive, we noted that a key function of the disclosures required incident to the new voting requirement was to ``provide shareholders and investors with timely information'' that was potentially useful to them ``as they

      Page 50106

      consider voting and investment decisions.'' \13\ Section 952 requires, in turn, that both compensation committee members of registrants and their advisors be independent. We noted that the rules implementing Section 952 could serve an informational purpose that benefits ``investors to the extent they enable compensation committees to make better informed decisions regarding the amount or form of executive compensation.'' \14\ Further, as we noted in the release proposing implementation of Section 953(a), that section is intended to provide shareholders with metrics that will help them assess executive compensation relative to the registrant's performance.\15\ The Section 953(a) information is intended, among other things, to assist shareholders when exercising their say-on-pay voting rights under Section 951.

      ---------------------------------------------------------------------------

      \11\ See, e.g., Medtronic, Inc. v. Lohr, 518 U.S. 470, 496 (1996) (``Congress' intent, of course, primarily is discerned from the language of the . . . statute and the `statutory framework' surrounding it. Also relevant, however, is the `structure and purpose of the statute as a whole,' as revealed not only in the text, but through . . . reasoned understanding of the way in which congress intended the statute and its surrounding regulatory scheme to affect business, consumers, and law.'') and Maine Public Utilities Comm'n v. FERC, 545 F.3d 278, 282 (D.C. Cir. 2006) (``This court applies the traditional tools of statutory interpretation in determining congressional intent, looking to the text, structure, purpose, and legislative history of a statute.'').

      \12\ Dodd-Frank Act, H.R. Rep. 111-517, at 872 (2010) (Conf. Rep.).

      \13\ Shareholder Approval of Executive Compensation and Golden Parachute Compensation, Release No. 33-9178 (Jan. 25, 2011) 76 FR 6010, 6037 (Feb. 2, 2011).

      \14\ Listing Standards for Compensation Committees, Release No. 33-9330 (June 20, 2012) 77 FR 38422, 38447 (Jun. 27, 2012).

      \15\ Pay Versus Performance, Release No. 34-74835 (Apr. 29, 2015) 80 FR 26329 (May 7, 2015).

      ---------------------------------------------------------------------------

      We believe that Section 953(b) should be interpreted consonant with Subtitle E's general purpose of further facilitating shareholder engagement with executive compensation. Thus, we believe that Congress intended Section 953(b) to supplement the executive compensation information available to shareholders. Particularly, Section 953(b) provides new data points that shareholders may find relevant and useful when exercising their voting rights under Section 951.\16\ Several commenters stated affirmatively that they would find the new data points, including pay ratio disclosure, relevant and useful when making voting decisions.\17\ Some commenters in the pre-proposing period \18\ suggested specifically that shareholders of public companies could use the pay ratio information, together with pay-versus-performance disclosure, to help inform their say-on-pay votes, which could also be a tool for shareholders to hold companies accountable for their CEO compensation.\19\ A significant consideration for us in fashioning a final rule implementing Section 953(b), then, is the extent to which elements of the final rule further Congress's apparent goal of giving shareholders additional executive compensation information to enhance the shareholder engagement envisioned by Section 951.\20\

      ---------------------------------------------------------------------------

      \16\ We note that the say-on-pay votes extend to certain other senior officers at the registrant beyond the PEO while the pay ratio disclosure is solely focused on a comparison of the PEO's compensation to the median employee's compensation. However, we do not think that this diminishes the overall utility of the pay ratio disclosures to say-on-pay votes. The PEO will typically be the highest compensated officer at a registrant and, to the extent shareholders rely on the pay ratio disclosure to determine whether the PEO's compensation is appropriate or not, it also may inform shareholders' relative assessment of the compensation of the other senior officers whose compensation is subject to say-on-pay votes.

      \17\ See, e.g., letters from AFL-CIO (Dec. 2, 2013) (``AFL-CIO I''); American Federation of State County and Municipal Employees (Nov. 27, 2013) (``AFSCME''); Amalgamated Bank (Nov. 25, 2013) (``Amalgamated''); Bricklayers & Trowel Trades International Pension Fund (Nov. 27, 2013) (``Bricklayers International''); California State Teachers' Retirement System (Dec. 2, 2013) (``CalSTRS''); Calvert Investments (Dec. 5, 2013) (``Calvert''); Chevy Chase Trust (Nov. 25, 2013) (``Chevy Chase Trust''); Corporate Governance (Nov. 25, 2013) (``CorpGov.net''); Form Letter C; Form Letter D; Form Letter E; Form Letter F; Laborers' International Union of North America (Oct. 10, 2013) (``LIUNA''); Local Authority Pension Fund Forum (Nov. 10, 2013) (``LAPFF''); New York State Comptroller (Nov. 27, 2013) (``NY State Comptroller''); Pax World Management LLC (Oct. 10, 2013) (``Pax World Funds''); Public Citizen (Nov. 26, 2013) (``Public Citizen I''); The Forum for Sustainable and Responsible Investment (Dec. 2, 2013) (``US SIF''); Trillium Asset Management, LLC (Dec. 2, 2013) (``Trillium I''); Trillium Asset Management, LLC (Jul. 31, 2014) (``Trillium II''); and UAW Retirement Medical Benefits Trust (Nov. 21, 2013) (``UAW Trust'').

      \18\ To facilitate public input on rulemaking required by the Dodd-Frank Act, we provided a series of email links, organized by topic, on our Web site at http://www.sec.gov/spotlight/regreformcomments.shtml, so that the public could provide comments before we proposed a rule. The comments relating to Section 953(b) are located at http://www.sec.gov/comments/df-title-ix/executive-compensation/executive-compensation.shtml (``Pre-Proposing Release Web site''). Comments that we received after we published the Proposing Release are located at http://www.sec.gov/comments/s7-07-13/s70713.shtml. Our references to comment letters refer to the comments on the proposal unless otherwise specified.

      \19\ See, e.g., letters from CtW Investment Group (Jun. 20, 2011) (``CtW Investment Group pre-proposal letter) and Steven Towns (Feb. 6, 2012) (``S. Towns pre-proposal letter).

      \20\ We note that some commenters contended that the pay ratio disclosure is intended to publicly ``shame'' registrants concerning the size of the disparity between their CEO's compensation and their typical worker's compensation. See, e.g., letters from The Honorable Carolos Cardozo Campbell, Former Assistant Secretary of Commerce for Economic Development (Nov. 27, 2013) (``Former Assistant Secretary Campbell''); Center on Executive Compensation (Jul. 6, 2015) (``COEC III''); Hyster-Yale Materials Handling, Inc. (Dec. 2, 2013) (``Hyster-Yale''); NACCO Industries, Inc. (Dec. 2, 2013) (``NACCO''); and U.S. Chamber of Commerce (Dec. 2, 2013) (``Chamber I''). As discussed above, we have reached a different conclusion based on principles of statutory construction and have taken no such objective into account in framing the rule. In crafting the final rule, we have sought to carefully tailor the pay ratio disclosure requirement so that it provides shareholders with a company-specific metric that is relevant and useful to their say-on-pay voting.

      ---------------------------------------------------------------------------

      Consistent with this understanding of the Congressional purpose of Section 953(b), we believe the final pay ratio rule should be designed to allow shareholders to better understand and assess a particular registrant's compensation practices and pay ratio disclosures rather than to facilitate a comparison of this information from one registrant to another.\21\ As we noted in the Proposing Release, we do not believe that precise conformity or comparability of the pay ratio across companies is necessarily achievable given the variety of factors that could cause the ratio to differ. Consequently, we believe the primary benefit of the pay ratio disclosure is to provide shareholders with a company-specific metric that they can use to evaluate the PEO's compensation within the context of their company.

      ---------------------------------------------------------------------------

      \21\ See, e.g., letters from CalSTRS, LAPFF, RPMI Railpen Investments (Dec. 2, 2013) (``RPMI''), Rep. Keith Ellison et al. (Dec. 2, 2013) (``Rep. Ellison et al. I''), Rep. Keith Ellison et al. (Mar. 17, 2015) (``Rep. Ellison et al. II''), and Sen. Robert Menendez et al. (Dec. 16, 2014) (``Sen. Menendez et al. II'').

      ---------------------------------------------------------------------------

      On the other hand, some commenters asserted that the pay ratio disclosure would not provide meaningful or material information to shareholders in making voting or investment decisions.\22\ In support of this contention, some of these commenters cited studies demonstrating that shareholders are not interested in this information,\23\ some commenters cited shareholder votes indicating a high level of support for executive pay and little support for shareholder proposals advocating for pay ratio disclosure,\24\

      Page 50107

      and some commenters contended that pay ratio disclosure would confuse shareholders because they would rely on it without fully considering a company's detailed narrative disclosures.\25\ Notwithstanding the disagreement among commenters on the value of the pay ratio disclosure, in adopting the final rule we have sought to implement Congress's apparent determination that the pay ratio disclosure would be useful to shareholders.

      ---------------------------------------------------------------------------

      \22\ See, e.g., letters from James J. Angel, Ph.D., CFA, Visiting Associate Professor, The Wharton School, University of Pennsylvania (Nov. 22, 2013) (``Prof. Angel''); Avery Dennison Corporation (Nov. 26, 2013) (``Avery Dennison''); Bill Barrett Corporation (Dec. 2, 2013) (``Bill Barrett Corp.''); Business Roundtable (Dec. 2, 2013) (``Business Roundtable I''); Center on Executive Compensation (Dec. 2, 2013) (``COEC I''); Center on Executive Compensation (Sep. 26, 2014) (``COEC II''); COEC III; Hyster-Yale; Mercer, Inc. (Dec. 2, 2013) (``Mercer I''); NACCO; and Pearl Meyer and Partners (Dec. 2, 2013) (``PM&P'').

      \23\ See, e.g., letters from Chamber I (citing the Center for Audit Quality's 7th annual Main Street Investor Survey, which ranked CEO compensation last on the list of factors used by shareholders to make investment decisions, with only 16% saying it was essential to their decision); COEC I (acknowledging that, while some literature focuses on pay disparities among employees with comparable jobs, the study frequently cited for the impact of disparities on collaboration, ``Pay Disparities Within Top Management Groups: Evidence of Harmful Effects on Performance of High-Technology Firms'' by Phyllis Siegel and Donald C. Hambrick, concerns executive pay and pay disparities among top executives, it does not discuss pay disparities between the CEO and median employee), and International Bancshares Corporation (Nov. 25, 2013) (``IBC'') (citing a Wall Street Journal article that says only 10% of individuals polled believed pay ratio would have value to shareholders) .

      \24\ See, e.g., letters from Chamber I (indicating that the average support for management compensation in public companies was 90.1%, and 97.6% of companies received majority shareholder support for executive compensation) and COEC I (noting that, since 2010, there have been 14 shareholder proposals advocating that companies provide pay ratio disclosure, and those proposals averaged 93% opposition from shareholders, with none receiving at least 10% support from shareholders).

      \25\ See, e.g., letters from Bill Barrett Corp., COEC I, National Investor Relations Institute (Oct. 17, 2013) (``NIRI''), and Semtech Corporation (Nov. 27, 2013) (``Semtech''). Two of these commenters suggested specifically that we should undertake an education effort to help shareholders understand the limits of pay ratio disclosure and remind them that they can find other information, such as an executive summary of the Compensation Discussion and Analysis section of a company's proxy statement, which can provide a more complete understanding of corporate pay practices. See letters from NIRI and Semtech.

      ---------------------------------------------------------------------------

      We also recognize that many commenters raised significant concerns about the costs of providing the required pay ratio disclosure. In implementing the statutory requirements, we have exercised our exemptive authority and provided flexibility in a manner that we expect will reduce costs and burdens for registrants, while preserving what we perceive to be the purpose and intended benefits of the disclosure required by Section 953(b).\26\ The significant cost estimates of the pay ratio disclosure submitted by some commenters support our view that some accommodations are appropriate.\27\ The final rule, therefore, both maintains the flexibility and accommodations from the proposal (such as permitting the use of statistical sampling and a consistent compensation measure to identify the median employee and reasonable estimates to calculate total compensation) and provides additional flexibility as follows: The final rule takes a flexible approach to the methodology a registrant can use to identify its median employee and calculate the median employee's annual total compensation; provides a de minimis exemption for non-U.S. employees and an exemption for registrants where, despite reasonable efforts to obtain or process the information necessary for compliance with the final rule, they are unable to do so without violating a foreign jurisdiction's laws or regulations governing data privacy; permits cost-of-living adjustments for the compensation of employees in jurisdictions other than the jurisdiction in which the PEO resides so that the compensation is adjusted to the cost of living in the jurisdiction in which the PEO resides; gives registrants the ability to make the median employee determination only once every three years and to choose as a determination date any date within the last three months of a registrant's fiscal year; and provides transition periods for new registrants, registrants engaging in business combinations or acquisitions, and registrants that cease to be smaller reporting companies or emerging growth companies.

      ---------------------------------------------------------------------------

      \26\ We are mindful of the principle that ``no legislation pursues its purposes at all costs,'' Rodriguez v. United States, 480 U.S. 522, 525-26 (1987), and we believe that the accommodations that we have included within the final rule reflect an appropriate balance.

      \27\ The potential costs arising from the requirements of Section 953(b), as well as the potential costs relating to the final rule, are discussed in detail below in Section III of this release under the heading ``Economic Analysis''.

      ---------------------------------------------------------------------------

      Overall, we think the final rule will provide investors with information Congress intended them to have to assess the compensation and accountability of a company's PEO while seeking to limit the costs and practical difficulties of providing the disclosure.

      Finally, we recognize the possibility that, based on the specific facts and circumstances of a registrant's work force and corporate operations, the pay ratio disclosure may warrant additional disclosures from a registrant to ensure that, in the registrant's view, the pay ratio disclosure is a meaningful data point for investors when making their say-on-pay votes. While Congress appears to have believed that the pay ratio disclosure would be a useful data point, we recognize that its relative usefulness--taken alone without accompanying disclosures to provide potentially important context--may vary considerably. Rather than prescribe a one-size-fits-all catalogue of additional disclosures that registrants should provide to put the pay ratio disclosure in context, we believe it is the better course to provide registrants the flexibility to provide additional disclosures that they believe will assist investors' understanding of the meaning of pay ratio disclosure when making say-on-pay votes. In this way, we believe we can best fulfill Congress's directive in Section 953(b) while avoiding unnecessary costs and complexities that might result from mandating additional disclosures.

    2. Summary of the Proposed Rule

      In September 2013, we proposed a new rule to implement Section 953(b) of the Dodd-Frank Act.\28\ The proposal's goal was to implement the statutory directive, while minimizing costs. In response to public comments we received prior to the proposal about the significant potential costs of complying with this requirement, the proposed rule would allow registrants flexibility in developing the disclosure required by the statute. We recognized that a one-size-fits-all approach would not be appropriate, given the wide range of affected registrants and the disparate burdens that would be imposed on them based on such factors as their business types and the complexity of their payroll systems. We therefore proposed to implement Section 953(b) in a manner that we believed would lower the cost of compliance while remaining consistent with the requirements of Section 953(b).

      ---------------------------------------------------------------------------

      \28\ See Proposing Release.

      ---------------------------------------------------------------------------

      The proposed rule would require companies to disclose the median of the annual total compensation of all its employees except the PEO, the annual total compensation of its PEO, and the ratio of the two amounts. The proposed rule would not have specified a single calculation methodology for identifying the median employee. Instead, it would permit registrants to select a methodology for identifying the median employee that was appropriate to the size and structure of their business and the way they compensate employees. Under the proposal, registrants could have chosen to identify the median employee by analyzing their full employee population or by using statistical sampling or another reasonable method. Also, to identify the median, registrants could have used ``total compensation,'' as defined in our existing rules, namely Item 402(c)(2)(x), or any consistently applied compensation measure, such as information derived from tax and/or payroll records. The proposed rule would not prescribe a particular methodology or specific computation parameters.

      Once the median employee was identified, the proposed rule would require the registrant to calculate the annual total compensation for that median employee in accordance with the definition of ``total compensation'' set forth in Item 402(c)(2)(x), which requires companies to provide extensive compensation information for the PEO and other named executive officers. ``Total compensation'' under Item 402(c)(2)(x) is not ordinarily calculated for all employees. The proposed rule, therefore, would permit registrants to use reasonable estimates in calculating

      Page 50108

      any element of total compensation and in calculating the annual total compensation of the median employee. Also, the proposed rule would define ``annual total compensation'' to mean total compensation for the last completed fiscal year, which would be consistent with our existing executive compensation disclosure requirements.

      Under the proposal, if a registrant used a compensation measure other than annual total compensation to identify the median employee, it would be required to disclose the compensation measure it used. Also, the registrant would be required to briefly describe and consistently apply any methodology it used to identify the median and any material assumptions, adjustments, or estimates used to identify the median employee or determine total compensation or any elements of total compensation for that employee or the PEO, and the registrant would need to clearly identify any amounts it estimated. Finally, registrants would be permitted, but not required, to supplement their disclosure with a narrative discussion or additional ratios if they chose to do so.

      Section 953(b) does not define the term ``employee.'' The proposed rule would define that term, for purposes of pay ratio disclosure, to include any individual employed by the registrant or any of its subsidiaries as of the last day of the registrant's last completed fiscal year. The proposed definition would encompass any full-time, part-time, seasonal, or temporary employees of the registrant or any of its subsidiaries, including any non-U.S. employee. Also, a registrant would be permitted, but not required, to annualize the total compensation for a permanent employee who was employed at year-end but did not work for the entire year. In contrast, full-time equivalent adjustments for part-time employees, annualizing adjustments for temporary and seasonal employees, and cost-of-living adjustments for non-U.S. employees would not be permitted.

      Also, under the proposal, registrants would be required to provide the proposed pay ratio disclosure in registration statements, proxy and information statements, and annual reports required to include executive compensation information as set forth under Item 402. Registrants, however, would not be required to provide their pay ratio information in reports that did not include Item 402 executive compensation information, such as current and quarterly reports. Additionally, registrants would not be required to update their annual pay ratio disclosure until they filed their annual report on Form 10-K for their last completed fiscal year or, if later, their definitive proxy or information statement for their next annual meeting of shareholders (or written consents in lieu of such a meeting). Registrants, however, would still be required to file their pay ratio information no later than 120 days after the end of the last fiscal year as provided in General Instruction G(3) of Form 10-K.\29\

      ---------------------------------------------------------------------------

      \29\ 17 CFR 249.310.

      ---------------------------------------------------------------------------

      The proposal would provide a transition period for newly public companies. For these companies, initial compliance would be required with respect to compensation for the first fiscal year commencing on or after the date the company became subject to the reporting requirements. Also, as provided by the Jumpstart Our Business Startups Act (``JOBS Act''),\30\ the proposed rule would not apply to emerging growth companies. Finally, the proposed rule would not apply to smaller reporting companies or foreign private issuers.

      ---------------------------------------------------------------------------

      \30\ 15 U.S.C. 78c(a).

      ---------------------------------------------------------------------------

    3. Summary of the General Comments on the Proposed Rule

      In the Proposing Release, we requested comment on many aspects of the proposed rule, including whether the proposed rule would address sufficiently the practical difficulties of data collection, whether other alternative approaches consistent with Section 953(b) could provide the potential benefits of pay ratio information at a lower cost, and whether the proposed flexible approach would appropriately implement Section 953(b). We received a large volume of comment letters from a variety of stakeholders. We received more than 287,400 comment letters, including over 1,540 individual letters that reflected a wide range of views concerning the proposed rule and the potential costs and benefits associated with its requirements. We received comments that addressed the proposed rule as a whole (including commenters that supported or opposed the rule in its entirety) as well as comments directed toward particular requirements of the rule, such as its application to foreign, part-time, temporary, and seasonal employees. In this section, we summarize the general comments on the proposal as a whole. Comments on particular provisions of the proposed rule are addressed as part of the discussion of each specific provision of the rule in Section II below.

      Of the over 287,400 total comment letters we received, over 285,900 were form letters regarding the proposed rule. There were 12 types of form letters,\31\ and these letters either supported our proposed rule or supported the idea of adopting a rule based on Section 953(b) without specifically referencing the proposal.

      ---------------------------------------------------------------------------

      \31\ We have received 285,936 form letters. Form Letter A (19,965 letters) supporting a strong rule generally; Form Letter B (12,942 letters) supporting a strong rule generally; Form Letter C (20 letters) supporting the Proposed Rule; Form Letter D (5,428 letters) supporting the Proposed Rule; Form Letter E (1,688 letters) supporting the Proposed Rule; Form Letter F (1,167 letters) supporting the Proposed Rule; Form Letter G (15,304 letters) supporting the Proposed Rule; Form Letter H (70,338 letters) supporting a rule generally; Form Letter I (36,299 letters) supporting the adoption of a rule; Form Letter J (75,333 letters) supporting a strong rule; Form Letter K (15,247 letters) supporting a strong rule generally; and Form Letter L (32,275 letters) supporting a rule.

      ---------------------------------------------------------------------------

      For example, one form letter asserted that the pay ratio disclosure is material to investors because high pay disparities can impair employee morale and productivity and have negative consequences on a company's overall performance and because investors will have a ``valuable additional'' measure for evaluating executive compensation, including when making say-on-pay voting decisions.\32\ Also, the letter supported the proposed rule's inclusion of all employees, including non-U.S. and part-time employees, and its flexibility in identifying the median employee. Other form letters also indicated that the pay ratio disclosure is material to investors,\33\ including some that noted that the disclosure would aid them in making voting decisions.\34\

      ---------------------------------------------------------------------------

      \32\ See Form Letter C.

      \33\ See letters from Form Letter A, Form Letter B, Form Letter D, Form Letter E, Form Letter F, and Form Letter G.

      \34\ See letters from Form Letter D (``Pay ratio disclosure will help investors evaluate CEO pay levels when voting on executive compensation matters. The ratio of the CEO-to-worker pay is a valuable metric for investors, because it places CEO pay levels into a broader perspective.''), Form Letter E (``Pay ratio disclosure will help investors evaluate CEO pay levels when voting on executive compensation matters. The ratio of the CEO-to-worker pay is a valuable tool for investors in evaluating and voting on CEO pay; scrutinizing the performance of Boards of Directors; and, identifying possible investment risks.''), Form Letter F (``A pay ratio disclosure will help investors better evaluate CEO pay levels when voting on executive compensation matters. Compensation experts have found that there is a correlation between high CEO pay and poor performance. By mandating disclosure of the ratio of CEO to worker pay, inequities will be become more transparent.'').

      ---------------------------------------------------------------------------

      Additionally, the vast majority of the over 1,500 unique comment letters were from individual commenters who, like those submitting the form letters, supported the proposed rule or supported the idea of adopting a rule based on Section 953(b) without specifically referencing the proposal. Most of these individuals supported the

      Page 50109

      proposed rule or the pay ratio disclosure because they believed it would:

      Inform shareholders about executive compensation matters, especially with regard to say-on-pay voting; \35\

      ---------------------------------------------------------------------------

      \35\ See, e.g., letters from D.A. Alexander (Nov. 22, 2013) (``Alexander''), Jean M. Blair (Sep. 25, 2013) (``J. Blair''), Cathy Clemens (Sep. 25, 2013) (``Clemens''), Beth Finchler (Sep. 27, 2013) (``Finchler''), Amy Hevron (``Sep. 28, 2013) (``Hevron''), Emanuel Jacobowitz (Sep. 24, 2013) (``Jacobowitz''), Rachel LaBruyere (Oct. 17, 2013) (``LaBruyere''), Gabrielle Loperfido (Sep. 25, 2013) (``Loperfido''), Carol Nix (Sep. 24, 2013) (``Nix''), Bonnie Overcott (Jan. 4, 2014) (``Overcott''), Lynn Reilly (Sep. 22, 2013) (``Reilly''), Kendall Simmons (Sep. 24, 2013) (``Simmons''), Dory Storms (Sep. 24, 2013) (``Storms''), Amy Sullivan-Greiner (Sep. 24, 2013) (``Sullivan-Greiner''), Jackie Tortora (Oct. 21, 2013) (``Tortora''), and Bryan Taylor (Nov. 21, 2013) (``Taylor'').

      ---------------------------------------------------------------------------

      demonstrate a company's focus on its long-term health as opposed to short-term gains that benefit its executives at the expense of its shareholders; \36\

      ---------------------------------------------------------------------------

      \36\ See, e.g., letters from Anonymous (Dec. 1, 2013) (``Anonymous''), Eric C. Gade (Sep. 18, 2013) (``Gade''), Linda Kranen (Oct. 20, 2013) (``Kranen''), Alyce Lomax (Dec. 2, 2013) (``Lomax''), Holly Schroeder (Nov. 5, 2013) (``Schroeder''), Erika Skornia-Olsen (Nov. 5, 2013) (``Skornia-Olsen''), and Calvin Vu (Oct. 26, 2013) (``Vu'').

      ---------------------------------------------------------------------------

      discourage the pay practices that led to the 2008 financial crisis; \37\

      ---------------------------------------------------------------------------

      \37\ See, e.g., letters from Lisbeth Caccese (Sep. 25, 2013) (``Caccese''), Hope Carr (Sep. 24, 2013) (``Carr''), Sarah McKee (Sep. 24, 2013) (``McKee''), Thomas Motes (Sep. 24, 2013) (``Motes''), Karl David Reinhardt (Sep. 24, 2013) (``Reinhardt''), Cynthia Sommer (Sep. 24, 2013) (``Sommer''), Cody Spann (Sep. 24, 2013) (``Spann''), Kathy Van Dame (Sep. 24, 2013) (``Van Dame''), Marietta Whittlesey (Sep. 24, 2013) (``Whittlesey''), Susan Williams (Sep. 24, 2013) (``S. Williams''), Mary M. Williamson (Sep. 24, 2013) (``M. Williamson''), and Robin Wittrock (Sep. 24, 2013) (``Wittrock'').

      ---------------------------------------------------------------------------

      reduce the inequitable wealth distribution in the U.S.; \38\ and

      ---------------------------------------------------------------------------

      \38\ See, e.g., letters from Paul Cohen (Oct. 3, 2013) (``Cohen''), Alan Harris (Sep. 19, 2013) (``Harris''), Liz A. King (Sep. 23, 2013) (``L. King''), Ben Leet (Oct. 27, 2013) (``Leet''), Laurie H. Norton (Sep. 18, 2013) (``Norton''), Debbie Notkin (Sep. 24, 2013) (``Notkin''), and Desmonde Printz (Sep. 19, 2013) (``Printz'').

      ---------------------------------------------------------------------------

      highlight potential problems in a company due to the negative impact of a high pay ratio on employee morale and productivity.\39\

      ---------------------------------------------------------------------------

      \39\ See, e.g., letters from Daniel Grossman (Nov. 8, 2013) (``Grossman''), Peter Linton (Nov. 17, 2013) (``Linton''), Michael R.K. Mudd (Nov. 3, 2013) (``Mudd''), Vivian Rosati (Nov. 6, 2013) (``Rosati''), and Walden Asset Management (Nov. 15, 2013) (``Walden'').

      ---------------------------------------------------------------------------

      As discussed in greater detail below, many commenters supported the proposed rule's overall flexibility.\40\ Some commenters asserted that the permitted flexibility would lessen the costs and burdens of the proposed rule without reducing the rule's benefits,\41\ be consistent with the directives of Section 953(b),\42\ and have minimal effect on the pay ratio disclosure.\43\ Other commenters, however, opposed the proposed rule because they believed it provided too much flexibility, which they asserted would allow registrants to manipulate the ratio in their favor,\44\ decrease the ratio's utility (especially for comparing the ratios of different companies),\45\ and still lead to high costs.\46\ One commenter suggested that the final rule consist of little more than ``a simple restatement'' of Section 953(b), doing ``little more than rearranging a few commas and adding a word or two to the statutory language.'' \47\

      ---------------------------------------------------------------------------

      \40\ See, e.g., letters from Prof. Angel; Aon Hewitt (Dec. 3, 2013) (``Aon Hewitt''); Bacirctirente, Canada et al. (Nov. 28, 2013) (``Bacirctirente et al.''); Best Buy Co., Inc., et al. (Dec. 17, 2013) (``Best Buy et al.''), CalSTRS; Marcia K. Campbell, Trustee, Illinois Teachers Retirement System (Nov. 21, 2013) (``Trustee Campbell''); Capital Strategies Consulting, Inc. (Dec. 2, 2013) (``Capital Strategies''); Center for Effective Government (Nov. 26, 2013) (``CEG''); Cummins Inc. (Dec. 2, 2013) (``Cummins Inc.''); CUPE Employees' Pension Plan (Nov. 21, 2013) (``CUPE''); Davis Polk & Wardwell LLP (Dec. 4, 2013) (``Davis Polk,''); Ernst & Young LLP (Dec. 2, 2013) (``E&Y''); First Affirmative Financial Network (Nov. 12, 2013) (``First Affirmative''); Fonds de solidariteacute FTQ (Nov. 27, 2013) (``FS FTQ''); Freeport-McMoRan Copper & Gold Inc. (Dec. 2, 2013) (``Freeport-McMoRan''); Interfaith Center on Corporate Responsibility (Oct. 17, 2013) (``ICCR''); International Brotherhood of Teamsters (Oct. 9, 2013) (``Teamsters''); International Union of Bricklayers and Allied Craftworkers, Administrative District Council 1 of Illinois (Nov. 15, 2013) (``IL Bricklayers and Craftworkers Union''); International Union of Bricklayers and Allied Craftworkers, Local 5 New York (Nov. 15, 2013) (``NY Bricklayers and Craftworkers Union''); Intel Corporation (Nov. 27, 2013) (``Intel''); Johnson and Johnson (Dec. 4, 2013) (``Johnson & Johnson''); Marco Consulting Group (Nov. 15, 2013) (``Marco Consulting''); McMorgan & Company LLC (Dec. 2, 2013) (``McMorgan & Co.''); Mercer, Inc. (Dec. 2, 2013) (``Mercer I''); Meridian Compensation Partners, LLC (Dec. 2, 2013) (``Meridian''); Microsoft Corporation (Dec. 4, 2013) (``Microsoft''); Nathan Cummings Foundation (Nov. 21, 2013) (``Cummings Foundation''); Network for Sustainable Financial Markets (Dec. 2, 2013) (``NSFM''); New York City Bar Association (Dec. 5, 2013) (``NYC Bar''); Novara Tesija, PLLC (Nov. 6, 2013) (``Novara Tesija''); Organizational Capital Partners (Nov. 24, 2013) (``OCP''); Oxfam America (Oct. 16, 2013) (``Oxfam''); PGGM (Dec. 2, 2013) (``PGGM''); Public Citizen I; Public School Teachers' Pension and Retirement Fund of Chicago (Nov. 25, 2013) (``Chicago Teachers Fund''); Quintave (Dec. 1, 2013) (``Quintave''); Rep. Ellison et al. II, SEIU Master Trust (Dec. 2, 2013) (``SEIU''); Sen. Menendez et al. II, Socially Responsive Financial Advisors/First Affirmative Financial Network (Oct. 11, 2013) (``Socially Responsive Financial Advisors''); Society of Corporate Secretaries and Governance Professionals (Dec. 16, 2013) (``Corporate Secretaries''); Trillium I, UAW Trust; Vectren Corporation (Dec. 2, 2013) (``Vectren Corp.''); Washington State Investment Board (Nov. 26, 2013) (``WA State Investment Board''); and WorldatWork (Dec. 2, 2013) (``WorldatWork I'').

      \41\ See, e.g., letters from American Bar Association (Mar. 7, 2014 (``ABA''), AFSCME, Chris Barnard (Nov. 6, 2013) (``Barnard''), Bricklayers International, Council of Institutional Investors (Nov. 6, 2013) (``CII''), Kenneth Fowler (Nov. 3, 2013) (``Fowler''), LIUNA, Walter Mirczak (Oct. 21, 2013) (``Mirczak''), PNC Financial Services Group, Inc. (Dec. 2, 2013) (``PNC Financial Services''), Sen. Menendez et al. II, US SIF, Vivient Consulting LLC (Dec. 2, 2013) (``Vivient''), and Walden.

      \42\ See, e.g., letters from Domini Social Investments LLC (Nov. 27, 2013) (``Domini'') and PM&P.

      \43\ See, e.g., letters from Capital Strategies and WorldatWork I.

      \44\ See, e.g., letters from Sherry Bupp (Oct. 4, 2013) (``Bupp''), Chris Corayer (Oct. 8, 2013) (``Corayer''), Russell J. Fedewa (Oct. 4, 2013) (``Fedewa''), Eleanor J. Fox (Oct. 3, 2013) (``Fox''), Gary G. Friend II (Oct. 3, 2013) (``Friend''), Charles Grotzke (Oct. 3, 2013) (``Grotzke''), Bruce Hlodnicki (Sep. 19, 2013) (``Hlodnicki''), Karla Kizzort (Oct. 4, 2013) (``Kizzort''), Christine Maly (Sep. 25, 2013) (``Maly''), B. A. Petricoin (Oct. 3, 2013) (``Petricoin''), and Jasmine Van Pelt (Oct. 4, 2013) (``Van Pelt'').

      \45\ See, e.g., letters from Amundi Asset Management (Nov. 28, 2013) (``Amundi''); British Columbia Investment Management Corporation (Dec. 2, 2013) (``BCIMC''); Paul Ciatto (Sep. 26, 2013) (``Ciatto''); Paul Glenn (Oct. 3, 2013) (``Glenn''); IBC; Karl T. Muth, Lecturer in Economics and Public Policy, Northwestern University (Sep. 24, 2013) (``Prof. Muth''); and NIRI.

      \46\ See, e.g., letters from ABA (stating that registrants will still incur significant costs even with the ability to select a methodology) and Financial Services Roundtable (Dec. 2, 2013) (``FSR'').

      \47\ See letter from Public Citizen (Jul. 6, 2015) (``Public Citizen II'').

      ---------------------------------------------------------------------------

      A number of commenters were critical of the proposed rule or particular aspects of it, as discussed in greater detail below. Some commenters stated specifically that they opposed the proposed rule or Section 953(b)'s requirement that we adopt any pay ratio rule. Some of these commenters asserted that the rule would not provide shareholders with material information.\48\ Other commenters noted that the pay ratio disclosure would not allow for meaningful comparisons among registrants.\49\ One commenter asserted that Section 953(b) and the proposed rule violate the First Amendment.\50\

      ---------------------------------------------------------------------------

      \48\ See, e.g., letters from American Benefits Council (Jan. 9, 2014) (``American Benefits Council''); Mark Appleby (Oct. 10, 2013) (``Appleby''); Avery Dennison; Sean Bearly (Nov. 7, 2013) (``Bearly''); Joe Beltran (Nov. 21, 2013) (``Beltran''); Renato Berzolla (Nov. 21, 2013) (``Berzolla''); Former Assistant Secretary Campbell; Jonnie Dodge (Nov. 7, 2013) (``Dodge''); FSR; Hyster-Yale; IBC (supporting Congressional efforts to repeal of Section 953(b)); Jim Meyer (Oct. 18, 2013) (``Meyer''); NACCO; National Association of Manufacturers (Dec. 2, 2013) (``NAM I'') (supporting Congressional efforts to repeal of Section 953(b)); NAM II (same); National Retail Federation (Nov. 26, 2013) (``NRF'') (expressing ``concern'' with the Proposed Rule); Elaine St. Miller (Nov. 21, 2013) (``St. Miller''); Towers Watson (Dec. 2, 2013) (``Towers Watson'') (discussing ``reservations'' about a pay ratio rule); and WorldatWork I.

      \49\ See, e.g., letters from ABA (stating that ``the required disclosure will have little utility to investors other than to enable them to see the ratio of principal executive to employee compensation for a specific registrant from year to year'') and COEC I.

      \50\ See letter from COEC I. We do not believe that the pay ratio disclosure that Congress has mandated is inconsistent with the First Amendment. We believe that, in passing Section 953(b), Congress determined that the disclosure advances an important government interest, and we have carefully tailored the disclosure through this rulemaking to further that interest. Moreover, consistent with Congress's apparent purpose, commenters have stated that the pay ratio disclosure would be useful to shareholders in making say-on-pay votes. Accordingly, we believe the disclosure fits comfortably within the class of securities law disclosures that have been deemed to be consistent with the First Amendment. See Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 758, n.5 (1985) (citing Ohralik v. Ohio State Bar Ass'n, 436 U.S. 447, 456 (1978)). See also SEC v. Wall Street Publ'g Inst., 851 F.2d 365, 373 (D.C. Cir. 1988).

      ---------------------------------------------------------------------------

      Page 50110

      A few commenters stated that we should not adopt a final rule until we demonstrate that the rule is consistent with our mission and fully explain the benefits and costs of the rule.\51\ In this regard, one of these commenters criticized us for not making a statement about our precise goals or objectives for the rule, especially when Congress failed to hold hearings on Section 953(b).\52\ The commenter also stated that, without this statement and further explanations as to why we rejected less costly options, commenters cannot be fully informed and provide constructive comments. Several commenters argued that the proposed rule would be very costly to implement though many of these did not provide specific cost estimates.\53\ A majority of these commenters indicated that navigating their payroll systems and creating a single database of all their employees' compensation would be the most costly aspect of the proposed rule--especially with respect to non-U.S. employees.\54\ Commenters also mentioned other activities that would contribute to the costs, including data privacy compliance, foreign exchange calculations, data testing, establishing corporate guidelines, obtaining legal services, auditing results, public relations tasks, and litigation risk.\55\ As discussed below, some commenters provided specific cost and burden estimates about the proposed rule to demonstrate generally that it would impose high costs and burdens on registrants.\56\ In addition, some commenters argued that the pay ratio disclosure would impose a burden on competition or would cause competitive disadvantages for particular types of companies.\57\ Our analysis of the costs of the final rule, as well as an assessment of its impact on competition, is contained in the Economic Analysis section below.

      ---------------------------------------------------------------------------

      \51\ See, e.g., letters from Chamber I and Technical Compensation Advisors (Dec. 2, 2013) (``TCA.'').

      \52\ See letter from Chamber I.

      \53\ See, e.g., letters from Aon Hewitt; Business Roundtable I; Chesapeake Utilities Corporation (Dec. 17, 2013) (``Chesapeake Utilities''); Garmin, Ltd. (Nov. 11, 2013) (``Garmin''); IBC; KBR, Inc. (Nov. 26, 2013) (``KBR''); McGuireWoods LLP (Nov. 7, 2013) (``McGuireWoods''); National Association of Corporate Dealers (Dec. 1, 2013) (``NACD''); NAM I; NAM II; NIRI; PNC Financial Services; and Semtech.

      \54\ See, e.g., letters from Aon Hewitt: Avery Dennison; Business Roundtable I; Chamber I; COEC I; Corporate Secretaries; Eaton (Dec. 2, 2013) (``Eaton''); FEI Company (Oct. 16, 2013) (``FEI''); FuelCell Energy, Inc. (Nov. 8, 2013) (``FuelCell Energy''); IBC; KBR; NACCO; NAM I; NAM II; and NIRI.

      \55\ See, e.g., letters from Aon Hewitt, Corporate Secretaries, and McGuireWoods.

      \56\ See, e.g., letters from ASA; Avery Dennison; COEC I; COEC II; Corporate Secretaries; Dover Corporation (Nov. 26, 2013) (``Dover Corp.''); Eaton; Exxon Mobil Corporation (Dec. 2, 2013) (``ExxonMobil''); FEI; FuelCell Energy; General Mills, Inc. (Dec. 2, 2013) (``General Mills''); Hyster-Yale; Intel; NACCO; NRF; and U.S. Chamber of Commerce (May 22, 2014) (``Chamber II'').

      \57\ See, e.g., letters from ABA; Prof. Angel; Former Assistant Secretary Campbell; Chamber I; COEC I; Corporate Secretaries; IBC; NAM I; NAM II; and Korok Ray, Assistant Professor of Accounting, The George Washington University (Dec. 2, 2013) (``Prof. Ray'').

      ---------------------------------------------------------------------------

      A number of commenters recommended that we take additional preliminary steps before adopting a final rule. Some commenters requested that we extend the proposed rule's comment period.\58\ Another commenter suggested that we re-solicit comments after publishing the concerns it expressed.\59\ A few commenters advocated that we involve stakeholders in the rulemaking process by holding a roundtable, engaging in negotiated rulemaking, and/or conducting pilot programs.\60\ One of these commenters also recommended submitting the proposed rule to the Office of Information and Regulatory Affairs for an enhanced regulatory review.\61\ A few commenters suggested that we defer adopting a final rule under Section 953(b) until we adopt other rules required under the Dodd-Frank Act, particularly the pay-versus-

      performance rule mandated by Section 953(a).\62\

      ---------------------------------------------------------------------------

      \58\ See, e.g., letters from American Insurance Association, et al. (Oct. 9, 2013) (``AIA et al.''); Bill Barrett Corp.; COEC I; IBC; NIRI; and WorldatWork I.

      \59\ See letter from Chamber I.

      \60\ See, e.g., letters from Chamber I, Chamber II, IBC, and NIRI.

      \61\ See letter from Chamber II.

      \62\ See, e.g., letters from Chamber I and WorldatWork I.

      ---------------------------------------------------------------------------

      As discussed above, members of the public interested in making their views known were invited to submit comment letters in advance of the official comment period for the proposed rule. In addition, we have continued to review and consider all comment letters submitted during and after the end of the comment period. Also, as discussed further in the Economic Analysis section below, we have considered and analyzed the numerous comments received regarding the costs and complexities of the mandated disclosure and have taken them into account in the final rule. Finally, we added to this rulemaking's public comment file additional analyses by the Commission's Division of Economic and Risk Analysis staff on the potential effects of excluding different percentages of employees from the pay ratio calculation, and commenters were expressly invited to comment on the analyses.

      This robust and public debate has informed us in developing our final rule. Overall, we believe interested parties have had sufficient opportunity to review the proposed rule, as well as the comment letters submitted, and to provide views on the proposal and on the other comment letters and data to inform our consideration of the final rule. Accordingly, we do not believe it is necessary to solicit additional public input before adopting the final rule.

    4. Summary of Changes in the Final Rule

      The final rule we are adopting generally is consistent with the proposed rule. After considering all of the comments received on the proposal, however, and in particular, after considering specific suggestions from commenters on alternatives that could help to mitigate compliance costs and practical difficulties associated with the proposed rule, we are adopting a number of revisions in the final rule. We believe these revisions generally will preserve Congress's intent to require the disclosure of information that reflects the ratio of the PEO's compensation to the median employee's compensation while helping to minimize the expected costs and unintended consequences of the required disclosure. We summarize some of these changes here and discuss them in greater detail in Section II, below.

      1. Non-U.S. Employee Exemptions and Additional Permitted Disclosure

        We proposed that an ``employee'' would include any U.S. and non-

        U.S. employee of a registrant. We acknowledged in the Proposing Release that the inclusion of non-U.S. employees would raise compliance costs for multinational companies, would introduce cross-border compliance issues, could raise additional comparability concerns, and could have an adverse impact on competition. We indicated, however, that the inclusion of non-U.S. employees in the calculation of the median is consistent with the ``all employees'' language of the statute.

        Page 50111

        The final rule defines the term ``employee'' to include U.S. employees and employees located in a jurisdiction outside the United States (``non-U.S. employees'') of a registrant, as proposed. We continue to believe that this is most consistent with the statutory language of Section 953(b) and with the purpose of providing a company-

        specific metric that shareholders can use to evaluate a registrant's executive compensation. Including both U.S. and non-U.S. employees will result in pay ratio disclosure that reflects the actual composition of the registrant's workforce. Even assuming the statutory language could be viewed as ambiguous on this issue, we also believe that this approach is most consistent with the general nature of our disclosure regime, which does not limit registrants' disclosure obligations only to factors, events, or circumstances that exist in or take place within the United States. For example, a registrant must disclose the PEO's compensation whether or not the PEO actually works within the United States.

        To help address concerns about compliance costs, and consistent with the commenters' suggestions, the final rule provides two tailored exemptions from the definition of ``employee,'' which otherwise includes all of a registrant's U.S. and non-U.S. employees in the median employee determination. First, the final rule provides an exemption for circumstances in which foreign data privacy laws or regulations make registrants unable to comply with the final rule. Second, the final rule permits registrants to exempt non-U.S. employees where these employees account for 5% or less of the registrant's total U.S. and non-U.S. employees, with certain limitations.

        The Proposing Release acknowledged that data privacy laws or regulations in various foreign jurisdictions could affect a registrant's ability to gather the necessary data to identify its median employee. We did not propose any accommodation to address this concern, however, because we believed the flexibility of the proposed rule would permit registrants to manage any potential costs arising from these laws. In response to significant concerns expressed by a number of commenters over cross-border compliance issues that may arise from the pay ratio disclosure requirement, and consistent with commenters' suggestions, the final rule permits registrants to exclude from their determination of the median employee an employee who is employed in a foreign jurisdiction in which the laws or regulations governing data privacy are such that, despite its reasonable efforts to obtain or process the information necessary for compliance with the final rule, the registrant is unable to do so without violating such data privacy laws or regulations.

        The registrant's reasonable efforts must include using or seeking an exemption or other relief under any governing data privacy laws or regulations. If a registrant excludes any non-U.S. employees in a particular jurisdiction under this exemption, it must exclude all non-

        U.S. employees in that jurisdiction, list the excluded jurisdictions, identify the specific data privacy law or regulation, explain how complying with the final rule violates such data privacy law or regulation (including the efforts made by the registrant to use or seek an exemption or other relief under such law or regulation), and provide the approximate number of employees exempted from each jurisdiction based on this exemption. In addition, the registrant must obtain a legal opinion from counsel that opines on the inability of the registrant to obtain or process the information necessary for compliance with the final rule without violating that jurisdiction's data privacy laws or regulations, including the registrant's inability to obtain an exemption or other relief under any governing laws or provisions.

        In addition to the data privacy exemption for non-U.S. employees, the final rule includes a de minimis exemption for non-U.S. employees. Under the final rule, if a registrant's non-U.S. employees account for 5% or less of its total employees, it may exclude all of those employees when making its pay ratio calculations. In this circumstance, however, if the registrant chooses to exclude any non-U.S. employees, it must exclude all of them. If a registrant's non-U.S. employees exceed 5% of the registrant's total U.S. and non-U.S. employees, it may exclude up to 5% of its total employees who are non-U.S. employees. If a registrant excludes any non-U.S. employees in a particular jurisdiction, it must exclude all non-U.S. employees in that jurisdiction. The registrant must also disclose the jurisdictions from which its non-U.S. employees are being excluded, the approximate number of employees excluded from each jurisdiction under the de minimis exemption, the total number of its U.S. and non-U.S. employees irrespective of any exemption (de minimis or data privacy), and the total number of its U.S. and non-U.S. employees used for its de minimis calculation.

        In calculating the number of non-U.S. employees that may be excluded under the de minimis exemption, a registrant must count any non-U.S. employee exempted under the data privacy exemption against the availability. A registrant may exclude any non-U.S. employee that meets the data privacy exemption, even if the number of excluded employees exceeds 5% of the registrant's total employees. If, however, the number of employees excluded under the data privacy exemption equals or exceeds 5% of the registrant's total employees, the de minimis exemption will not be available. Additionally, if the number of employees excluded under the data privacy exemption is less than 5% of the registrant's total employees, the registrant may use the de minimis exemption to exclude no more than the number of non-U.S. employees that, combined with the data privacy exemption, equals 5% of the registrant's total employees.

        Finally, the final rule permits registrants to make cost-of-living adjustments for the compensation of employees in jurisdictions other than the jurisdiction in which the PEO resides to identify the median and calculate annual total compensation. In identifying the median employee, whether using annual total compensation or any other compensation measure that is consistently applied to all employees included in the calculation, the registrant may, but is not required to, make cost-of-living adjustments for the compensation of employees in jurisdictions other than the jurisdiction in which the PEO resides so that the compensation is adjusted to the cost of living in the jurisdiction in which the PEO resides. If the registrant uses a cost-

        of-living adjustment to identify the median employee, and the median employee identified is an employee who does not reside in the same jurisdiction as the PEO, the registrant must use the same cost-of-

        living adjustment in calculating the median employee's annual total compensation and disclose the country in which the median employee is located. The registrant is also required to briefly describe the cost-

        of-living adjustments it used to identify the median employee and briefly describe the cost-of-living adjustments it used to calculate the median employee's annual total compensation, including the measure used as the basis for the cost-of-living

        Page 50112

        adjustment.\63\ To provide context for the Item 402(u)(1)(iii) disclosure, a registrant electing to present the pay ratio in this manner must also disclose the median employee's annual total compensation and pay ratio without the cost-of-living adjustments. To calculate this pay ratio, the registrant will need to identify the median employee without using any cost-of-living adjustments.

        ---------------------------------------------------------------------------

        \63\ For example, registrants may use cost-of-living adjustments based on purchasing power parity (``PPP'') conversion factors. A PPP conversion factor is the ratio of PPP exchange rate to the nominal exchange rate. For example, conversion factors for the US dollar are available at http://data.worldbank.org/indicator/PA.NUS.PPPC.RF. This ratio provides the number of units of a country's currency required to buy the same amount of goods and services in the domestic market as a U.S. dollar would buy in the United States.

        ---------------------------------------------------------------------------

        In the Proposing Release, we noted that some comments received prior to the proposal requested that the rule allow registrants to present separate pay ratios covering U.S. and non-U.S. employees to mitigate concerns that the comparison of the PEO to non-U.S. employees could substantially affect the pay ratio disclosure. The proposal did not prohibit such disclosure but did not expressly state it was permitted. For clarification, therefore, the final rule states that registrants are permitted, but not required, to provide additional pay ratios as long as any additional pay ratios are not misleading and are not presented with greater prominence than the required ratio.

      2. Employees of Consolidated Subsidiaries

        We proposed requiring a registrant's pay ratio disclosure to include the employees of any of its subsidiaries (including officers other than the PEO), in addition to its direct employees, in its pay ratio disclosure. Unlike the proposed rule, however, the final rule defines ``employee'' to include only the employees of the registrant's consolidated subsidiaries. As discussed in greater detail below, defining a ``subsidiary'' based on whether a registrant consolidates a company in its financial statements will likely decrease the costs and burdens on a registrant without significantly affecting the pay ratio because most registrants consolidate based on their ownership of over 50% of the outstanding voting shares of their subsidiaries and guidance is readily available on when consolidation is appropriate.

      3. Employed on Any Date Within Three Months of the Last Completed Fiscal Year

        The proposed rule would define ``employee'' as an individual employed as of the last day of the registrant's last completed fiscal year because this calculation date would be consistent with the one used for the determination of the three most highly compensated executive officers under existing Item 402(a)(3)(iii). In the Proposing Release, we also noted our preliminary view that a bright line calculation date for determining who is an employee would ease compliance for registrants by eliminating the need to monitor changes in workforce composition during the year. Further, we assumed the potential benefits of the pay ratio disclosure would not be significantly diminished by covering only individuals employed at year-

        end, although we acknowledged that this approach could be costlier to registrants with seasonal or temporary employees who are employed at year end as opposed to other times during the year.

        Taking into consideration concerns raised by commenters about the desire for flexibility in choosing the calculation date, the final rule permits registrants to use any date within three months prior to the last day of their last completed fiscal year to identify the median employee. If in subsequent years the registrant changes the date it uses to identify the median employee, it must disclose this change and provide a brief explanation about the reason or reasons for the change. This provision provides consistency for individual registrants from year to year while also providing registrants with flexibility to choose the determination date. To provide additional transparency about how the pay ratio disclosure has been calculated, the final rule requires registrants to disclose the date used to identify the median employee.

      4. Identifying the Median Employee Once Every Three Years

        The proposed rule would require registrants to identify the median employee every year. To help minimize compliance costs, we are revising the rule, as suggested by commenters, to allow registrants to identify the median employee every three years unless there has been a change in its employee population or employee compensation arrangements that the registrant reasonably believes would result in a significant change in the pay ratio disclosure. However, the registrant must still calculate the identified median employee's annual total compensation and use that figure in calculating its pay ratio every year. If there have been no changes that the registrant reasonably believes would significantly affect its pay ratio disclosure, the registrant must disclose that it is using the same median employee in its pay ratio calculation and describe briefly the basis for its reasonable belief. For example, the registrant could disclose that there has been no change in its employee population or employee compensation arrangements that it believes would significantly impact the pay ratio disclosure. If there has been such a change, the registrant must re-identify the median employee for that fiscal year.

        Under the final rule's approach, the registrant will identify its median employee for year one and then be permitted to use that employee or one who is similarly compensated (if, for example, the median employee is no longer in the same position or is no longer employed by the registrant) in the following two years for calculating the median employee's annual total compensation and the registrant's pay ratio. The registrant must calculate the median employee's annual total compensation in year one and then re-calculate the annual total compensation for that employee in year two and again in year three. If the median employee identified in year one is no longer in the same position or no longer employed by the registrant on the median employee determination date in year two or three, the final rule permits the registrant to replace its median employee with an employee in a similarly compensated position.

      5. Initial Compliance Date

        We proposed that a registrant's first reporting period would begin in its first fiscal year commencing on or after the effective date of the final rule. Therefore, under the proposed rule, the registrant's initial pay ratio disclosure would be included in its first annual report on Form 10-K or proxy or information statement for its annual meeting of shareholders following the end of such year. Unlike the proposal, the compliance date set forth in this adopting release provides that the registrant's first reporting period for the pay ratio disclosure is its first full fiscal year beginning on or after January 1, 2017 (instead of on or after the effective date of the final rule).

      6. Transition Period for New Registrants

        The proposed rule would not have required pay ratio disclosure by new registrants subject to the rule in a registration statement on Form S-1 or

        Page 50113

        Form S-11 for an initial public offering or registration statement on Form 10. Consistent with the revised transition period for existing registrants, the final rule provides that the first pay ratio reporting period begins for new registrants with their first fiscal year commencing on or after January 1, 2017 that is after the date they first become subject to the requirements of Section 13(a) or 15(d) of the Exchange Act. In this way, new registrants will not become subject to the final rule sooner than existing registrants. Such registrants are also permitted to omit their initial pay ratio disclosure from their registration statements (or any other filing) made before their first annual report or proxy or information statement following the end of that reporting period, but not later than 120 days after the end of the fiscal year.

      7. Additional Transition Periods

        We did not propose a transition period for registrants that cease to be smaller reporting companies or emerging growth companies, nor did we provide any special rules for registrants that engage in business combinations and/or acquisitions. We did, however, request comment on whether there should be transition periods in these situations and, if so, the appropriate length of time for any such transition period. One commenter requested that we include such a transition period.\64\

        ---------------------------------------------------------------------------

        \64\ See letter from ABA (recommending a transition period for issuers that cease to be smaller reporting companies in which such registrants would not be required to provide their pay ratio disclosure until the first full fiscal year commencing on or after the first anniversary of the end of the fiscal year in which the issuer is no longer a smaller reporting company).

        ---------------------------------------------------------------------------

        The final rule provides that registrants that cease to be smaller reporting companies or emerging growth companies are not required to provide pay ratio disclosure until they file a report for the first fiscal year commencing on or after they cease to be a smaller reporting company or emerging growth company. The final rule also permits registrants that engage in business combinations and/or acquisitions to omit the employees of a newly-acquired entity from their pay ratio calculation for the fiscal year in which the business combination or acquisition occurs. In these cases, a registrant does not have to include these individual employees in its median employee calculation until the first full fiscal year following the acquisition. Registrants that exclude employees as a result of a business combination must disclose the relevant acquired business and the approximate number of employees that are excluded from the pay ratio calculation.

  8. Discussion

    1. Scope of Final Rule

      1. Pay Ratio Disclosure Requirements Under New Paragraph (u) to Item 402 of Regulation S-K

        a. Proposed Rule

        We proposed new paragraph (u) of Item 402 to require disclosure of: (A) the median of the annual total compensation of all employees of the registrant, (except the registrant's PEO); (B) the annual total compensation of the registrant's PEO; and (C) the ratio of the amount in (A) to the amount in (B), presented as a ratio in which the amount in (A) equaled one, or, alternatively, expressed narratively in terms of the multiple that the amount in (B) bears to the amount in (A).

        Although Section 953(b) calls for a ratio showing the median of the annual total compensation of all employees to the PEO's annual total compensation, it does not specify how the ratio should be expressed. To promote consistent presentation and address the potential for confusion, therefore, the proposed rule specified that registrants must express the ratio as one in which the median of the annual total compensation of all employees is equal to one.

        b. Comments on the Proposed Rule

        No commenters objected to use of the term PEO or including the pay ratio disclosure requirements in new paragraph (u) to Item 402, and commenters discussing other aspects of the proposal did so on the assumption that we would include the pay ratio disclosure requirements in Item 402(u). Several commenters agreed that the pay ratio should show the PEO's compensation divided by the median employee's compensation because it would be easier to understand.\65\

        ---------------------------------------------------------------------------

        \65\ See, e.g., letters from FEI and Dr. Sue Ravenscroft, Professor of Accounting, Iowa State University (Sep. 20, 2013) (``Prof. Ravenscroft I'').

        ---------------------------------------------------------------------------

        c. Final Rule

        After considering the comments, we are adopting the final rule as proposed. The final rule adds new paragraph (u) to Item 402 and requires disclosure of: (A) The median of the annual total compensation of all employees of the registrant (except the registrant's PEO); (B) the annual total compensation of the registrant's PEO; and (C) the ratio of the amount in (B) to the amount in (A), presented as a ratio in which the amount in (A) equals one, or, alternatively, expressed narratively in terms of the multiple that the amount in (B) bears to the amount in (A).\66\

        ---------------------------------------------------------------------------

        \66\ We did not propose to require that the pay ratio disclosure be provided in interactive data format, and are not adopting such a requirement for this disclosure. To the extent that we consider more generally the tagging of disclosures in XBRL format in our rules, we may consider revisiting the format in which the pay ratio disclosure is provided.

        ---------------------------------------------------------------------------

        Consistent with the proposal, the final rule also requires registrants to disclose the ratio such that the PEO's annual total compensation is always compared to the median employee's annual total compensation. Registrants may not present the median employee's annual total compensation as a percentage of the PEO's compensation. We believe expressing the ratio as ``a factor rather than a fraction'' makes the ratio easier to understand because allowing the inverse may be confusing.\67\ In other words, the ratio must always show how much larger or smaller the PEO's annual total compensation is as compared to the median employee's annual total compensation. We believe that requiring registrants to present the ratio in this manner will make it easier for shareholders to comprehend and allow them to use it in making voting decisions on executive compensation.\68\

        ---------------------------------------------------------------------------

        \67\ See letter from FEI. See also letter from Prof. Ravenscroft I (stating that ``the ratio of CEO compensation to median worker would be easier for most people to grasp than the ratio of median worker compensation to CEO compensation, a switch that I would see as not changing the intent of the law'').

        \68\ In the rare cases in which the PEO's yearly compensation is nominal (or is otherwise less than the median employee's compensation), the resulting ratio will be a number smaller than one. Despite this anomalous result, we believe that in the vast majority of cases setting the median compensation equal to one will result in a ratio that is easier to understand than the inverse. See generally Illinois Commerce Comm'n v. I.C.C., 776 F.2d 355, 359 (D.C. Cir. 1985) (``General rules need not work perfectly in all their applications''). We also note that registrants are permitted to provide additional narrative discussion in cases where they feel the disclosed pay ratio may be confusing or incomplete without further explanation.

        ---------------------------------------------------------------------------

        The final rule permits registrants to choose one of two options to express the ratio. Registrants may disclose the pay ratio with the median of the annual total compensation of all employees equal to one and the PEO's compensation as the number compared to one.\69\ For

        Page 50114

        example, if a registrant's median annual total compensation for employees is $50,000 and the annual total compensation of the PEO is $2,500,000, the PEO's compensation is 50 times larger than the median employee's compensation. The registrant may describe the pay ratio as 50 to 1 or 50:1. Alternatively, registrants may disclose the pay ratio narratively by stating how many times higher (or lower) the PEO's annual total compensation is than that of the median employee. For example, the registrant may state that ``the PEO's annual total compensation is 50 times that of the median of the annual total compensation of all employees.''

        ---------------------------------------------------------------------------

        \69\ We note that some commenters recommended that the final rule include a safe harbor or simplified reporting method such that a registrant may stipulate that its pay ratio exceeds 300-to-1. See letters from Hyster-Yale and NACCO. Under this suggestion, registrants would be permitted to forgo calculating and disclosing their company-specific pay ratio and would instead be permitted to simply disclose that the ratio exceeds the stipulated 300-to-1 statistic. We have not adopted the suggestion to allow a registrant to disclose that its pay ratio exceeds a stipulated statistic, such as 300-to-1, because we do not believe that it would be consistent with Congress's apparent intent to provide a useful, relevant, company-specific pay ratio disclosure for investors to utilize when undertaking their say-on-pay votes.

        ---------------------------------------------------------------------------

      2. Pay Ratio Disclosure in Filings That Require Item 402 of Regulation S-K Information

        a. Proposed Rule

        The proposed rule required registrants to include their pay ratio disclosure in any filing described in Item 10(a) that requires executive compensation disclosure under Item 402, including annual reports on Form 10-K,\70\ Securities Act and Exchange Act registration statements, and proxy and information statements, to the same extent that these forms require compliance with Item 402. Section 953(b) does not direct us to amend any of our forms to add the pay ratio disclosure requirements to filings that do not already require disclosure of Item 402 information, and we did not propose to do so. Additionally, we proposed not to require registrants to update their pay ratio disclosure for the most recently completed fiscal year until they file their annual reports on Form 10-K, or, if later, their proxy or information statements for their next annual meeting of shareholders (or written consents in lieu of such a meeting) but, in any event, not later than 120 days after the end of such fiscal year.

        ---------------------------------------------------------------------------

        \70\ 17 CFR 249.310.

        ---------------------------------------------------------------------------

        b. Comments on the Proposed Rule

        All of the commenters discussing the issue agreed that we should limit pay ratio disclosure to the filings described in Item 10(a) that require executive compensation disclosure under Item 402, as proposed.\71\ One commenter suggested, however, that the final rule should allow registrants to include their pay ratio disclosure in other filings if they choose to do so.\72\

        ---------------------------------------------------------------------------

        \71\ See, e.g., letters from ABA, AFL-CIO I, California Public Employees' Retirement System (Dec. 2, 2013) (``CalPERS''), Calvert, Capital Strategies, CII, Connecticut State Treasurer (Dec. 2, 2013) (``CT State Treasurer''), Davis Polk, Intel, Johnson & Johnson, McGuireWoods, NIRI, NRF, Pax World Funds, PM&P, Prof. Ray, and WorldatWork I.

        \72\ See letter from Capital Strategies.

        ---------------------------------------------------------------------------

        c. Final Rule

        We are adopting the final rule as proposed. It requires registrants to include their pay ratio disclosure in any filing described in Item 10(a) that calls for executive compensation disclosure under Item 402, including annual reports on Form 10-K, registration statements under the Securities Act and Exchange Act, and proxy and information statements to the same extent that these forms require compliance with Item 402, consistent with the statutory directive. Registrants must follow the instructions in each form to determine whether Item 402 information is required, including any instructions that allow for the omission of Item 402 information.\73\ The final rule does not require registrants to include the pay ratio disclosure in any filings that are not required to include Item 402 information, but registrants can voluntarily include non-mandated information in any of their filings if they choose to do so and the information is not misleading in the context of that filing. Further, registrants do not need to update their pay ratio disclosure for their most recently completed fiscal year until they file their annual report on Form 10-K, or, if later, their proxy or information statement for their next annual meeting of shareholders (or written consents in lieu of such a meeting) but, in any event, not later than 120 days after the end of such fiscal year.

        ---------------------------------------------------------------------------

        \73\ See, e.g., General Instructions I(2)(c) and J(1)(m) to Form 10-K containing special provisions for the omission of Item 402 information by wholly-owned subsidiaries and asset-backed registrants.

        ---------------------------------------------------------------------------

        We do not read Section 953(b) to require pay ratio disclosure in filings that do not contain other executive compensation information. In our view, the most meaningful way to present pay ratio disclosure is in context with other executive compensation disclosure, such as the Summary Compensation Table required by Item 402(c) and the Compensation Discussion and Analysis required by Item 402(b), rather than provided on a stand-alone basis. In this manner, the pay ratio information will be presented in the same context as other information that shareholders can use in making their voting decisions on executive compensation. Finally, although we understand the primary purpose of the pay ratio disclosure to be to inform shareholder's say-on-pay votes under Section 951, we acknowledge that some commenters indicated the disclosure could be useful to investors in making investment decisions. For that reason, and in light of the statutory language of Section 953(b), the final rule retains the requirement to include this disclosure in registration statements under the Securities Act.\74\

        ---------------------------------------------------------------------------

        \74\ Although the final rule requires registrants to include the pay ratio disclosure in registration statements under the Securities Act, as discussed below, the final rule permits new registrants to delay compliance so that the pay ratio requirement is not required in a registration statement on Form S-1 or Form S-11 for an initial public offering or registration statement on Form 10.

        ---------------------------------------------------------------------------

      3. Excluded Registrants--Smaller Reporting Companies, Foreign Private Issuers, MJDS Filers, and Emerging Growth Companies

        a. Proposed Rule

        In the Proposing Release, we noted that the reference to ``each issuer'' in Section 953(b) could be read to apply to all registrants, including smaller reporting companies,\75\ foreign private issuers,\76\ U.S.-Canadian Multijurisdictional Disclosure System filers,\77\ and emerging growth companies. Because Section 953(b) refers specifically to the definition of total compensation in Item 402(c)(2)(x), and is silent on whether pay ratio disclosure should be required for registrants not previously subject to Item 402(c) requirements, however, we proposed to limit the pay ratio disclosure requirement to registrants required to provide Item 402(c) disclosure. As a result, the proposed rule stated that smaller reporting companies, foreign private issuers, and MJDS filers did not have to provide pay ratio disclosure in any of their filings.

        ---------------------------------------------------------------------------

        \75\ A ``smaller reporting company'' is an issuer that had a public float of less than $75 million as of the last business day of its most recently completed second fiscal quarter or had annual revenues of less than $50 million during the most recently completed fiscal year for which audited financial statements are available. 17 CFR 229.10(f)(1).

        \76\ A ``foreign private issuer'' is any foreign issuer other than a foreign government, except for a registrant that, as of the last business day of its most recent fiscal year, has more than 50% of its outstanding voting securities held of record by United States residents and any of the following: a majority of its officers and directors are citizens or residents of the United States, more than 50% of its assets are located in the United States, or its business is principally administered in the United States. 17 CFR 240.3b-

        4(c).

        \77\ A U.S.-Canadian Multijurisdictional Disclosure System (``MJDS'') filer is a registrant that files reports and registration statements with us in accordance with the requirements of the MJDS.

        ---------------------------------------------------------------------------

        Also, the JOBS Act,\78\ which was passed by Congress subsequent to the

        Page 50115

        Dodd-Frank Act but prior to publication of the Proposing Release, specifically excluded registrants that qualify as emerging growth companies, as that term is defined in Section 3(a) of the Exchange Act,\79\ from the requirements of Section 953(b). To give effect to the statutory exemption, we proposed an instruction to Item 402(u) providing that a registrant that is an emerging growth company is not required to comply with Item 402(u).

        ---------------------------------------------------------------------------

        \78\ See Section 102(a)(3) of the JOBS Act.

        \79\ An ``emerging growth company'' is an issuer that had total annual gross revenues of less than $1 billion during its most recently completed fiscal year, has not reached the fifth anniversary of the date of the first sale of its common equity securities pursuant to an effective registration statement under the Securities Act, had not issued $1 billion in non-convertible debt during the previous 3-year period, or is deemed to be a ``large accelerated filer.'' 15 U.S.C. 78c(a).

        ---------------------------------------------------------------------------

        b. Comments on the Proposed Rule

        Most commenters concurred with the proposed rule's exclusion of emerging growth companies, consistent with the JOBS Act.\80\ One commenter noted specifically that, while it ``believes all registrants accessing U.S. capital markets should be subject to comparable financial regulation,'' including smaller reporting companies, foreign private issuers, and MJDS filers, the pay ratio information ``is best viewed in the context of other compensation disclosures and the pay ratio disclosure should be limited to those registrants required to provide a summary compensation disclosure.'' \81\

        ---------------------------------------------------------------------------

        \80\ See, e.g., letters from CalPERS; Davis Polk; Hay Group, Inc. (Dec. 2, 2013) (``Hay Group''); NY State Comptroller; PM&P; and US SIF. Some commenters, however, disagreed or were uncomfortable with this exclusion. See letters from CII and Andrew Kushner (Nov. 12, 2013) (``Kushner'').

        \81\ See letter from CalPERS.

        ---------------------------------------------------------------------------

        Additionally, most of the commenters who addressed the issue agreed that we should exclude smaller reporting companies from the pay ratio requirements.\82\ One commenter reasoned that, by excluding emerging growth companies, Congress demonstrated its intent to relieve this category of registrants from the costs and burdens of compliance, and because both emerging growth companies and smaller reporting companies are subject to scaled executive compensation disclosure, it would be consistent with Congressional intent to exclude smaller reporting companies.\83\ Another commenter recommended that the final rule exempt any business with revenues of less than $500 million or a market capitalization of less than $1 billion.\84\

        ---------------------------------------------------------------------------

        \82\ See, e.g., letters from ABA, Prof. Angel, CalPERS, Capital Strategies, Davis Polk, Hay Group, NIRI, NY State Comptroller, PM&P, Vivient, and WorldatWork I.

        \83\ See letter from ABA.

        \84\ See letter from Hay Group.

        ---------------------------------------------------------------------------

        By contrast, a few commenters asserted that the final rule should include smaller reporting companies.\85\ One of these commenters asserted that Congress did not expressly exclude smaller reporting companies because the phrase ``each issuer'' in Section 953(b)(1) signals its intent that there should be no exemption for any particular registrant and that excluding certain registrants from the disclosure would defeat the purpose and policy of Section 953(b).\86\

        ---------------------------------------------------------------------------

        \85\ See, e.g., letters from CII; Ashley Ray, University of Idaho College of Law (Nov. 13, 2013) (``Ray''); and US SIF.

        \86\ See letter from Ray.

        ---------------------------------------------------------------------------

        Finally, some commenters agreed that the proposed rule should exclude foreign private issuers and MJDS filers,\87\ while a few other commenters disagreed with excluding them.\88\ One commenter noted that, in ``view of the Commission's long-standing rules allowing foreign private issuers and MJDS filers to provide information about their executive compensation programs based on the applicable disclosure requirements of their home jurisdiction, we would find it anomalous to single out this specific Item 402-based disclosure requirement for mandatory application to these registrants without regard to important policy considerations that have led the Commission for decades to permit disclosure in this area based on home-country law.'' \89\

        ---------------------------------------------------------------------------

        \87\ See, e.g., letters from ABA, CalPERS, Capital Strategies, Davis Polk, Hay Group, and PM&P.

        \88\ See, e.g., letters from CII, Ray, and US SIF.

        \89\ See letter from ABA.

        ---------------------------------------------------------------------------

        c. Final Rule

        After considering the comments, we are adopting the final rule as proposed. The final rule, therefore, does not require pay ratio disclosure by smaller reporting companies, foreign private issuers, MJDS filers, and emerging growth companies.\90\

        ---------------------------------------------------------------------------

        \90\ Registered investment companies will also not be required to provide Item 402(u) disclosure. Business development companies are a category of closed-end investment company that are not registered under the Investment Company Act 15 U.S.C. 80a-2(a)(48) and 80a-53-64. Business development companies will be treated in the same manner as issuers other than registered investment companies and therefore will be subject to the pay ratio disclosure requirement.

        ---------------------------------------------------------------------------

        As stated above, Congress explicitly excluded emerging growth companies from the pay ratio disclosure requirement. Regarding smaller reporting companies, Section 953(b)(2) requires total compensation to be calculated in accordance with Item 402(c)(2)(x). Smaller reporting companies, however, are permitted to follow the scaled disclosure requirements set forth in Items 402(m)-(r),\91\ and therefore are not required to calculate compensation in accordance with Item 402(c)(2)(x). Also, the requirement set forth in Item 402(n) for disclosure of Summary Compensation Table information, which includes disclosure of ``total compensation,'' does not require smaller reporting companies to include the same types of compensation required to be included in total compensation for other registrants under Item 402(c)(2).\92\ Congress's express reference to Item 402(c)(2)(x) to calculate total compensation (without mentioning Item 402(n)(2)(x)) is consistent with the exclusion of smaller reporting companies from the pay ratio disclosure requirement.

        ---------------------------------------------------------------------------

        \91\ See Item 402(l).

        \92\ See Item 402(n)(2)(viii) (indicating that smaller reporting companies are not required to include the aggregate change in the actuarial present value of pension benefits that is required for companies subject to Item 402(c)(2)(viii)).

        ---------------------------------------------------------------------------

        Requiring smaller reporting companies to provide the pay ratio disclosure would compel them to collect data and calculate compensation for the PEO in ways that they otherwise are not required to do. Nothing in the statute indicates that was Congress's intent, and no commenters indicated that they believed there was such an intent. To clarify further that smaller reporting companies are excluded from the final rule, we are making a technical amendment to paragraph (l) of Item 402 to add Item 402(u), as proposed, to the list of items that are not required for smaller reporting companies.\93\

        ---------------------------------------------------------------------------

        \93\ Smaller reporting companies are permitted to choose whether they want to comply with either the scaled disclosure requirements or the larger company disclosure requirements on an ``a la carte'' basis. As we discussed in the scaled disclosure adopting release, the staff evaluates compliance by smaller reporting companies with only the Regulation S-K requirements applicable to smaller reporting companies, even if the company chooses to comply with the larger company requirements. See Smaller Reporting Company Regulatory Relief and Simplification, Release No. 33-8876 (Dec. 19, 2007) 73 FR 934, at 941.

        ---------------------------------------------------------------------------

        The final rule similarly does not apply to foreign private issuers and MJDS filers, which we believe is consistent with excluding registrants that are not currently required to provide Summary Compensation Table disclosure pursuant to Item 402(c). Foreign private issuers file annual reports and registration statements on Form 20-F \94\ and MJDS filers file annual reports and registration

        Page 50116

        statements on Form 40-F.\95\ Neither of these forms requires Item 402 disclosure. As with smaller reporting companies, requiring foreign private issuers and MJDS filers to provide the pay ratio disclosure would require these registrants to collect data and calculate compensation for the PEO in ways they otherwise would not be required to do. The final rule, therefore, does not apply to foreign private issuers or MJDS filers.

        ---------------------------------------------------------------------------

        \94\ 17 CFR 249.220f.

        \95\ 17 CFR 249.240f.

        ---------------------------------------------------------------------------

        Finally, for the same reasons, the final rule, consistent with the proposal, does not change existing Item 402(a)(1) with respect to foreign private issuers. Item 402(a)(1) states that a ``foreign private issuer will be deemed to comply with Item 402 if it provides the information required by Items 6.B and 6.E.2 of Form 20-F, with more detailed information provided if otherwise made publicly available or required to be disclosed by the registrant's home jurisdiction or a market in which its securities are listed or traded.'' Foreign private issuers that file annual reports on Form 10-K, therefore, are still able to satisfy Item 402 requirements by following Items 6.B and 6.E.2 of Form 20-F and are not required to provide the pay ratio disclosure mandated by Section 953(b).

    2. Requirements of Final Rule

      1. ``All Employees'' Covered Under the Rule

        The final rule defines ``employee'' to include a registrant's U.S. and non-U.S. employees, as well as its part-time, seasonal, and temporary employees, as proposed. We believe that the ``all employees of the issuer'' language in Section 953(b) is best implemented by including rather than excluding broad categories of employees. Further, even assuming there was any ambiguity in the statutory language, we believe that a more inclusive approach better serves Section 953(b)'s purpose of providing shareholders with additional information about a registrant's compensation practices that can be used in making voting decisions on executive compensation because it results in a pay ratio that is more reflective of the actual composition of the registrant's workforce.\96\ As discussed in greater detail below, however, in response to particular issues and concerns raised by comments, we have provided two tailored exemptions from the general requirement to include all employees. In particular, for the reasons discussed below, we have provided an exemption for employees in foreign jurisdictions in which it is not possible for a registrant to obtain or process information necessary to comply with the rule without violating the data privacy laws or regulations of that jurisdiction and a de minimis exemption for non-U.S. employees.

        ---------------------------------------------------------------------------

        \96\ As discussed below, commenters have made competing arguments that the pay ratio disclosure would be ``distorted'' in one way or another if the final rule included (or excluded) various categories of employees. See, e.g., letter from Prof. Ray (asserting that the pay ratio will be distorted depending on whether the final rule includes non-U.S. employees). We appreciate that a particular registrant's pay ratio information may vary--perhaps, in some cases, significantly--depending on whether the final rule includes or excludes these employee categories (e.g., non-U.S. employees; part-

        time, temporary and seasonal employees; and leased workers). We do not, however, view the choices made on these issues as involving a ``distortion'' of the pay ratio disclosure that Congress directed should be provided to investors. As noted elsewhere in this release, we have been guided by our general view that Section 953(b) reflects Congress's intention that the pay ratio disclosure should be broadly inclusive of all types of a registrant's employees; thus, absent a reason that takes into account the statutory objective, we have declined to make choices in the final rule that would exclude broad categories of employees from the process of identifying the median employee. At the same time, however, in an effort to mitigate the potential costs and burdens of the final rule, we have built in some flexibility and provided several other accommodations to elements of the final rule, where we have concluded that these measures would not result in any undue impact on the required pay ratio disclosure.

        ---------------------------------------------------------------------------

        a. Types of Employees

        i. Proposed Rule

        The proposed rule included in the definition of ``employee'' all of a registrant's full-time, part-time, seasonal, and temporary workers, including officers other than the PEO. In the Proposing Release, we reasoned that these individuals should be included in the rule because Section 953(b)(1)(A) expressly requires disclosure of the median of the annual total compensation of ``all employees,'' which would encompass full-time, part-time, seasonal, and temporary workers. Also, we proposed to include all of a registrant's officers other than the PEO in the definition of ``employee.''

        Workers not employed by a registrant (or its subsidiaries), however, such as independent contractors, ``leased'' workers, or other workers who are employed by a third party, were not covered by our proposed definition of ``employee.'' As an example, we noted that, if a registrant pays a fee to another company (such as a management company or an employee leasing agency) that supplies workers to the registrant, and those workers receive compensation from that other company, these workers should not be considered employees of the registrant for purposes of the disclosures required by Section 953(b) of the Dodd-

        Frank Act.

        ii. Comments on the Proposed Rule

        A number of commenters supported the proposed rule's requirement that registrants include their part-time, temporary, and seasonal employees in addition to their full-time employees in their median employee determination.\97\ Some commenters asserted that the reference in Section 953(b) to ``all employees'' demonstrates Congress's intent was not to limit the pay ratio to only full-time employees.\98\ Some commenters contended that including temporary employees would cost registrants very little because they routinely develop that information for their own internal use.\99\ Commenters supporting the proposed rule also contended that, if part-time, temporary, and seasonal employees were excluded from the pay ratio, the disclosure would be incomplete, inaccurate, and/or misleading.\100\ One commenter suggested that the exclusion of part-time, seasonal, and temporary employees would not reduce the regulatory burdens on registrants because registrants with such employees would have substantial flexibility in identifying the median employee.\101\

        ---------------------------------------------------------------------------

        \97\ See, e.g., letters from AFSCME, CalPERS, Calvert, Chicago Teachers Fund, Cummings Foundation, CUPE, Domini, Susan A. Estep (Nov. 15, 2013) (``Estep''), Frank Gould (Oct. 4, 2013) (``Gould''), ICCR, IL Bricklayers and Craftworkers Union, Marco Consulting, McMorgan & Co., Novara Tesija, NY State Comptroller, Oxfam, Pax World Funds, Sen. Robert Menendez et al. (Nov. 26, 2013) (``Sen. Menendez et al. I''), Sen. Menendez et al. II, Socially Responsive Financial Advisors, Teamsters, Trillium I, Trustee Campbell, US SIF, and Walden.

        \98\ See letters from AFL-CIO (Jul. 6, 2015) (``AFL-CIO II''), Institute for Policy Studies (Oct. 30, 2013) (``IPS''), and Sen. Menendez et al. II.

        \99\ See, e.g., letters from Virginia Fischer (Oct. 3, 2013) (``Fischer'') and Public Citizen II.

        \100\ See, e.g., letters from AFL-CIO I, AFL-CIO II, Americans for Financial Reform (Dec. 2, 2013) (``AFR''), Bacirctirente et al., Bricklayers International, CII, CT State Treasurer, FS FTQ, Public Citizen I, Public Citizen II, and John Theodore (Nov. 5, 2013) (``Theodore'').

        \101\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        Other commenters contended that the final rule should include only full-time employees.\102\ Many of these commenters claimed that applying the rule to part-time, temporary, and seasonal employees would make the pay ratio disclosure less meaningful because the compensation of these different types of employees are not comparable to each other or to the PEO's compensation.\103\ Some commenters

        Page 50117

        also asserted that including employees other than full-time employees would be burdensome and increase costs.\104\ One commenter noted that, according to a survey it conducted of companies with more than 10,000 employees, ``86 percent of the average employer's employees are full-

        time, with the median employer having 95 percent of its workforce as full-time employees.'' This commenter asserted that the incremental information that would be obtained from including part-time or seasonal employees does not justify the effort to collect it and could provide a distorted picture of the employee's annual income.\105\

        ---------------------------------------------------------------------------

        \102\ See, e.g., letters from Best Buy et al., Brian Foley & Company, Inc. (Dec. 2, 2013) (``Brian Foley & Co.''), Chamber I, FuelCell Energy, Garmin, Hyster-Yale, Mercer I, NACCO, PM&P, Semtech, Steven Hall and Partners (Dec. 2, 2013) (``SH&P''), Vectren Corp., WorldatWork I, and WorldatWork II.

        \103\ See, e.g., letters from American Apparel & Footwear Association (Dec. 2, 2013) (``AAFA I''), American Apparel & Footwear Association (Dec. 23, 2013) (``AAFA II''), American Benefits Council, COEC I, Corporate Secretaries, Davis Polk, Dover Corp., ExxonMobil, FSR, General Mills, Hay Group, IBC, KBR, Meridian, NACD, NIRI, NRF, NYC Bar, Retail Industry Leaders Association (Dec. 2, 2013) (``RILA''), and WorldatWork I.

        \104\ See, e.g., letters from American Benefits Council, COEC I, Corporate Secretaries, Davis Polk, and General Mills.

        \105\ See letter from COEC I.

        ---------------------------------------------------------------------------

        Another commenter asserted that the final rule should exclude part-

        time, seasonal, and temporary employees unless a majority of a registrant's employees work on a part-time, temporary, and/or seasonal basis.\106\ A few commenters recommended that the rule should allow registrants to exclude any employee who was not employed for at least four months during the calendar year.\107\ Other commenters indicated that, if the rule is not limited to full-time employees, registrants should be able to annualize or make full-time equivalent adjustments to the compensation of part-time, seasonal, and/or temporary employees.\108\ One commenter suggested, in the alternative, that if the rule includes part-time, seasonal, and temporary employees, it should only include such employees who have been employed during the 90-day period ending on the last day of the registrant's fiscal year.\109\ Finally, another commenter suggested, in the alternative, that the rule should require a ``primary disclosure'' that compares only full-time employees to the PEO's compensation.\110\

        ---------------------------------------------------------------------------

        \106\ See letter from ABA.

        \107\ See, e.g., letters from Hyster-Yale and NACCO (stating that such a provision could contain a consistency requirement to prevent companies from selectively choosing which employees to include or exclude).

        \108\ See, e.g., letters from AAFA II, American Benefits Council, Brian Foley & Co., Corporate Secretaries, Hay Group, KBR, NIRI, and NYC Bar.

        \109\ See letter from Meridian.

        \110\ See letter from NRF.

        ---------------------------------------------------------------------------

        Some commenters noted that neither Section 953(b) nor its legislative history states explicitly that Congress intended for the ``all employees'' term to include part-time, seasonal, and temporary employees. These commenters contended we could thus interpret the ``all employees'' language to exclude such employees from the final rule.\111\ Some commenters believed that the minimal effect these employees would have on the pay ratio would not justify the high costs required to include those employees in determining the pay ratio.\112\ Conversely, commenters in support of including part-time, seasonal, and temporary employees in the pay ratio contended that failing to do so would distort registrants' pay ratios because many of their employees would not be included in the median calculation.\113\

        ---------------------------------------------------------------------------

        \111\ See, e.g., letter from ABA.

        \112\ See, e.g., letters from COEC I (stating that two-thirds of respondents to a survey it conducted indicated that limiting the application of the rule to full-time employees would reduce their costs, that the ``average savings for these respondents would be approximately 20 percent,'' and that the burden imposed by including ``global, full-time, part-time and seasonal employees'' is not offset by other benefits) and General Mills (``We would expect moderate cost savings from limiting the analysis to full-time employees, versus covering our entire workforce, but the savings could be significant for registrants in other industries. . . Conversely, there has been little or no evidence to suggest that the benefits of the Proposed Rule would be diminished as a result of limiting its scope to full-time employees.'').

        \113\ See, e.g., letters from AFL-CIO I, AFR, Bacirctirente et al., Bricklayers International, CII, CT State Treasurer, FS FTQ, Public Citizen I, and Theodore.

        ---------------------------------------------------------------------------

        Several commenters agreed that a registrant's pay ratio should exclude ``leased'' workers.\114\ One of these commenters noted that including ``leased'' workers would add significant costs and distort the pay ratio.\115\ Also, according to this commenter, such workers are not ``statutory'' employees and the third parties employing these workers may be unwilling to provide the information. A few commenters recommended that the final rule require a registrant to include such workers in its pay ratio.\116\ One commenter asserted that a registrant should be required to clearly describe its reliance on ``leased'' workers if they comprise more than 40% of its workforce.\117\

        ---------------------------------------------------------------------------

        \114\ See, e.g., letters from ABA; American Staffing Association (Nov. 21, 2013) (``ASA''); CalPERS; CII; Corporate Secretaries; Emergent BioSolutions, Inc. (Nov. 27, 2013) (``Emergent''); ExxonMobil; Hyster-Yale; Intel; McGuireWoods, NACCO; and WorldatWork I.

        \115\ See letter from ABA.

        \116\ See, e.g., letters from Demos (Nov. 22, 2013) (``Demos I''), Fischer (referring to ``independent contractors''), and Vectren Corp. (arguing that the rule should include full-time, U.S.-

        based contractors if they make up a significant portion of the registrant's workforce).

        \117\ See letter from LAPFF.

        ---------------------------------------------------------------------------

        iii. Final Rule

        After considering the public comments, we have concluded that the final rule's definition of ``employee'' should include the full-time, part-time, seasonal, and temporary employees employed by the registrant or any of its consolidated subsidiaries. Because this definition refers to workers ``employed by the registrant,'' workers who provide services to the registrant or its consolidated subsidiaries as independent contractors or ``leased'' workers are excluded from the definition as long as they are employed, and their compensation is determined, by an unaffiliated third party. The final rule includes in the definition of ``employee'' all of a registrant's officers other than the PEO, as proposed. Section 953(b)(1)(A) expressly directs disclosure of the median of the annual total compensation of ``all employees of the issuer, except the chief executive officer (or any other equivalent position) of the issuer.''

        We believe this statutory language indicates that Congress intended the final rule to include all types of a registrant's employees, including part-time, seasonal, and temporary workers, and we do not think it is appropriate to provide a wholesale exemption for those broad categories of employees that are not employed full-time.\118\ Any such exemption would risk producing pay ratio disclosure that is significantly different than the pay ratio disclosure that Congress expressly directed us to require when it said ``all employees.'' Further, as noted above, we have generally limited our use of discretionary or exemptive authority to those items that would not have an appreciable effect on the information that Congress intended that shareholders have when they make their say-on-pay votes. To the extent there is any statutory ambiguity, we would still elect this inclusive approach because we believe that it is more reflective of the actual composition of the registrant's workforce and thus furthers the purpose of providing shareholders with useful information about a registrant's overall compensation practices. While we are sensitive to

        Page 50118

        concerns raised by some commenters that inclusion of these broad categories of employees means that compliance with the final rule will be more costly than if we adopted a broad exemption, we note that the final rule provides other types of flexibility and accommodations designed to reduce compliance costs while remaining faithful to our understanding of the statutory directive and purpose of Section 953(b).

        ---------------------------------------------------------------------------

        \118\ See, e.g., County of Oakland v. Federal Housing Finance Agency, 716 F.3d 935, 940 (6th Cir. 2013) (explaining that ``a straightforward reading of the statute leads to the unremarkable conclusion that when Congress said `all taxation,' it meant all taxation'') (emphasis in original); Marie O. v. Edgar, 131 F.3d 610, 620 (7th Cir. 1997) (interpreting the phrase ``all infants and toddlers'' and explaining that ``all is unambiguous; it means every eligible child''); cf. GEICO v. Fetisoff, 958 F.2d 1137, 1142 (D.C. Cir. 1992) (describing ``the word `all''' as ``one of the least ambiguous words in the English language'').

        ---------------------------------------------------------------------------

        A registrant can supplement its pay ratio disclosure or provide additional pay ratios for its shareholders to consider if it wants to explain the effect of including part-time, seasonal and temporary employees on its pay ratio disclosure. While we do not believe a purpose of the rule is to facilitate comparisons among registrants, the opportunity to supplement the pay ratio disclosure and to provide additional pay ratios should help mitigate some concerns that shareholders may draw unwarranted conclusions from comparing one registrant's disclosed ratio to the ratio of others. In addition, our change to the proposed rule to allow a registrant some flexibility in selecting the date for identifying the median employee may enable registrants that employ temporary or seasonal employees only during a very limited period at the end of their fiscal year to choose a date that allows them to exclude these employees.

        One commenter pointed out that Item 402 does not contain a reference to hourly or overtime compensation and contended, therefore, that the rule should not apply to non-salaried employees who receive ``wages plus overtime,'' rather than salary.\119\ We believe that the ``all employees'' language is not limited to salaried employees. Moreover, we are concerned that a contrary reading would be arbitrary and would eliminate an entire category of employees from the pay ratio disclosure, potentially depriving shareholders of a more complete understanding of the median employee's compensation when making their say-on-pay votes. Thus, we believe that it is appropriate to include these employees as part of a registrant's pay ratio disclosure to reflect the manner in which the registrant establishes its workforce.

        ---------------------------------------------------------------------------

        \119\ See letter from Chamber I.

        ---------------------------------------------------------------------------

        The final rule excludes from the definition of ``employee'' those workers who are employed, and whose compensation is determined, by an unaffiliated third party but who provide services to the registrant or its consolidated subsidiaries as independent contractors or ``leased'' workers. Although it is unclear whether Congress intended to include these workers as ``employees of the issuer,'' or even considered the issue, we believe, as a matter of policy, these workers should not be included as ``employees of the issuer.''

        While one commenter stated that ``leased employees and other workers employed by a third party are not ``statutory'' employees of a registrant,'' \120\ some definitions of ``employee'' may include workers who are not employed directly by the registrant or its consolidated subsidiaries, such as independent contractors or ``leased'' or ``borrowed'' workers, if they are employed by a third party.\121\ We note that the statute specifies employees ``of the issuer,'' and in light of this, to the extent there is any ambiguity on this point, we believe that the better reading of the statute is to exclude from the final rule's definition of ``employee'' workers who are not employed by the registrant or its consolidated subsidiaries, such as independent contractors or ``leased'' workers, if they are employed, and their compensation is determined, by an unaffiliated third party.

        ---------------------------------------------------------------------------

        \120\ See, e.g., letter from ABA.

        \121\ See, e.g., Black's Law Dictionary 602 (9th ed. 2009) (defining ``employee'' as a ``person who works in the service of another person (the employer) under an express or implied contract of hire, under which the employer has the right to control the details of work performance,'' and defines a ``borrowed employee'' as an ``employee whose services are, with the employee's consent, lent to another employer who temporarily assumes control over the employee's work'').

        ---------------------------------------------------------------------------

        We believe excluding such workers is appropriate because registrants generally do not control the level of compensation that these workers are paid. Instead, the registrant provides a payment for their services to an unaffiliated third party, which determines the compensation for the employees. As one commenter noted, there can be no assurance, therefore, that the registrant even has access to the workers' compensation information, which could make it difficult or impossible to obtain the information.\122\

        ---------------------------------------------------------------------------

        \122\ See letter from ABA.

        ---------------------------------------------------------------------------

        We do not believe it is appropriate for registrants to voluntarily include workers employed by third parties in their required pay ratio disclosure ``if such persons make up a significant portion of the workforce,'' as one commenter suggested, even if doing so may add to the ``flexibility'' of the final rule.\123\ For the reasons described above, we have not included these workers within the definition of employee, and we are concerned that allowing registrants the option to elect to include them in their required disclosures would introduce the potential for registrants to manipulate the pay ratio disclosure. Registrants, however, may discuss their reliance on ``leased'' workers, as suggested by another commenter, in their narrative disclosure.\124\ Also, they may provide additional ratios that factor in those workers, as long as any additional ratios are not misleading and are not more prominently displayed than the required ratio.

        ---------------------------------------------------------------------------

        \123\ See letter from Vectren Corp.

        \124\ See letter from LAPFF.

        ---------------------------------------------------------------------------

        b. Employed on Any Date Within Three Months of the Last Completed Fiscal Year

        i. Proposed Rule

        The proposed rule would define ``employee'' as an individual employed as of the last day of the registrant's last completed fiscal year. We proposed this calculation date for determining who is an employee because it is consistent with the one used for the determination of PEO and principal financial officer and the other three most highly compensated executive officers under Item 402(a)(3)(iii). In the Proposing Release, we noted that the composition of a company's workforce typically changes throughout the fiscal year, and in some industries and businesses, it can change constantly. Although Section 953(b) requires the median calculation to cover ``all employees,'' it does not prescribe a particular calculation date for the determination of who should be treated as an employee for that purpose.

        We reasoned in the Proposing Release that a single date for determining who is an employee would ease compliance for registrants by eliminating the need to monitor changing workforce composition during the year, while providing a recent snapshot of the registrant's entire workforce. Also, we indicated that a requirement to track which employees have been continuously employed for the entire annual period could increase costs for registrants, and suggested that the most appropriate calculation date would be one that is consistent with the calculation date for determining the named executive officers under current Item 402 requirements.

        In proposing this approach, we assumed that the potential benefits of the disclosure mandated by Section 953(b) would not be significantly diminished by covering only individuals employed on a specific date

        Page 50119

        at calendar year-end, rather than covering every individual who was employed at any time during the year. Although this approach could help limit compliance costs for registrants, we acknowledged that it could have other costs. For example, this approach would not capture seasonal or temporary employees who are not employed at year end, with the result that a registrant with a significant number of such workers might identify a median employee from a pool that does not fully reflect the workforce that it requires to run its business. This approach might also cause the disclosure to be costlier for, and thereby have an anti-competitive impact on, registrants whose temporary or seasonal workers are employed at calendar year-end as opposed to other times during the year because registrants with temporary or seasonal employees at calendar year-end would have to include them in their median calculations but other registrants with temporary or seasonal employees at other times of the year would not have to do so. Finally, we noted that it would be possible, but unlikely, that registrants could try to structure their employment arrangements to reduce the number of lower paid employees employed on the determination date.

        ii. Comments on the Proposed Rule

        Most commenters that discussed the issue agreed that registrants should be permitted to identify the median employee based on the composition of their workforce on a particular day of the year as opposed to the workforce employed throughout the year.\125\ Only a few commenters, however, supported using the last day of the fiscal year calculation date.\126\ It seems that these commenters supported this provision more to limit the calculation to a particular day of the year, thereby limiting the need to monitor a changing workforce during the year, than because they believed the appropriate date should be the last day of the registrant's last fiscal year.\127\ A number of commenters contended that the final rule should allow registrants the flexibility to choose a calculation date within the registrant's last fiscal year.\128\ As some of these commenters noted, requiring registrants to use the last day of the fiscal year could adversely affect retailers,\129\ may not allow enough time for registrants to collect and report on their pay ratio information,\130\ and could make it difficult for registrants that are not calendar year-end companies to use information derived from its tax and/or payroll records to calculate the ratio.\131\ Some commenters suggested that, if the final rule permits registrants to choose a determination date other than a registrant's fiscal year end, it should also require registrants to be consistent from year to year and/or briefly explain the reasons for not using the fiscal year-end date.\132\

        ---------------------------------------------------------------------------

        \125\ See letters from ABA, Capital Strategies, FEI, Hyster-

        Yale, Johnson & Johnson, and NACCO.

        \126\ See, e.g., letters from CII and Vectren.

        \127\ Id.

        \128\ See, e.g., letters from ABA, Best Buy et al., Capital Strategies, COEC I, COEC II, Corporate Secretaries, Davis Polk, FEI, Hyster-Yale, Johnson & Johnson, Microsoft, NACCO, NAM I, PM&P, RILA, SH&P, and WorldatWork I. See also letter from Brianne H. McCoy, University of Idaho College of Law (Nov. 28, 2013) (``McCoy'') (stating that a registrant should be required to choose a date that ``corresponds with the month in which the registrant had its highest gross operating revenues from the previous year'').

        \129\ See letter from PM&P.

        \130\ See, e.g., letters from COEC I, Corporate Secretaries, Davis Polk, Microsoft, and NAM I.

        \131\ See, e.g., letters COEC I, Microsoft, NAM I, and NAM II.

        \132\ See, e.g., letters from Capital Strategies, Microsoft, and SH&P.

        ---------------------------------------------------------------------------

        One commenter suggested that the rule could allow registrants to choose a calculation date within a designated time window, ``such as a date occurring during the 90-day period preceding the fiscal year end.'' \133\ Regarding the concern that allowing registrants to select the calculation date would create the potential for manipulation of the pay ratio, this commenter stated that the concern ``is unwarranted, particularly if the choice is restricted to a limited time period (such as the last fiscal quarter), since in general the employee population of a registrant would not vary significantly over such a period.'' \134\

        ---------------------------------------------------------------------------

        \133\ See letter from Davis Polk.

        \134\ Id.

        ---------------------------------------------------------------------------

        Another commenter proposed that registrants should be required to calculate the median annual compensation of all employees employed at any time over the preceding 365 days to ensure accurate disclosure for registrants that employ a high number of seasonal employees.\135\

        ---------------------------------------------------------------------------

        \135\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        One commenter recommended that the final rule permit registrants to use different determination dates for different segments of their workforce based on tax, payroll, and/or other established recordkeeping systems, accompanied by a brief statement of the basis for the different disclosure dates because a number of companies maintain their human resource/payroll systems for U.S. employees on a calendar-year basis, but do so for their foreign employees on a fiscal-year basis.\136\ The commenter also noted that using the end of the second or third fiscal quarter as a determination date would not be feasible because most payroll systems are set up to collect information on fiscal year-end or calendar year-end bases.

        ---------------------------------------------------------------------------

        \136\ See letter from ABA.

        ---------------------------------------------------------------------------

        Some commenters responded to the Proposing Release's request for comment on whether the rule's definition of ``employee'' would cause a registrant to change its corporate structure. Most of the commenters that responded said that the definition would not cause registrants to alter their corporate structure or employment arrangements,\137\ although one commenter disagreed.\138\

        ---------------------------------------------------------------------------

        \137\ See, e.g., letters from ABA, Capital Strategies, and WorldatWork I.

        \138\ See letter from Prof. Ray.

        ---------------------------------------------------------------------------

        iii. Final Rule

        After considering commenters' desire for flexibility in choosing the median employee determination date, we are revising the final rule from the proposal. Unlike the proposed rule, which would define ``employee'' as an individual employed as of the last day of the registrant's last completed fiscal year, the final rule defines ``employee'' as an individual employed on any date of the registrant's choosing within the last three months of the registrant's last completed fiscal year. The final rule also requires registrants to disclose the date used to identify the median employee.

        Some commenters recommended that the final rule require registrants to explain the reason they selected a determination date other than the last day of the fiscal year.\139\ We do not believe such an explanation would lead to useful disclosure, as registrants would likely state that they chose any date other than the end of the fiscal year to provide more time to take the steps necessary to identify the median employee. If, however, a registrant changes the determination date from the prior year, we believe it should disclose the reason for the change. Under the final rule, therefore, if a registrant changes the date it uses to identify the median employee, the registrant must disclose the change and provide a brief explanation about the reason or reasons for the change.

        ---------------------------------------------------------------------------

        \139\ See, e.g., letters from Microsoft and SH&P.

        ---------------------------------------------------------------------------

        We note that allowing registrants to choose a determination date within a defined window, rather than be required to use the last day of the fiscal year, is a change from the proposal and differs from the approach in determining the

        Page 50120

        three most highly compensated executive officers under existing Item 402(a)(3)(iii).\140\ We believe permitting registrants to choose a date within the last three months of their last completed fiscal year, as suggested by a commenter, is appropriate because it will provide registrants with some flexibility and could permit them additional time to identify their median employee in advance of their fiscal year end.\141\ At the same time, establishing a particular date certain will provide some consistency from year to year. We also note that this change may help to avoid some of the unintended consequences identified by commenters, such as anti-competitive effects on retailers with a significant number of employees at year end or inefficient changes in corporate structure made simply to avoid employing workers on the last day of the fiscal year. Finally, as discussed in the Proposing Release, we continue to believe that requiring the determination of the employee to be made as of a specific date, rather than over the course of the year, will ease compliance for registrants by eliminating the need to monitor changes in their workforce composition throughout the year.

        ---------------------------------------------------------------------------

        \140\ 17 CFR 230.402(a)(3)(iii).

        \141\ See letter from Davis Polk (noting that a ``90-day period preceding the fiscal year end'' would permit ``a registrant to begin the task of identifying its median employee in advance of its fiscal year end, which is the most costly and time-consuming part of the pay ratio calculation'').

        ---------------------------------------------------------------------------

        c. Employees Located Outside the United States

        i. Proposed Rule

        We proposed a definition of ``employee'' that would include any U.S. and non-U.S. employee of a registrant. In the Proposing Release, we acknowledged that the inclusion of non-U.S. employees raises compliance costs for multinational companies, introduces cross-border compliance issues, and could raise concerns about the impact of non-

        U.S. pay structures on the comparability of the data to companies without off-shore operations. We also recognized that differences in relative compliance costs could have an adverse impact on competition. We weighed these considerations and proposed that the disclosure requirements would nonetheless cover all employees without exemptions for specific categories of employees, including non-U.S. employees.

        Additionally, we were cognizant that data privacy laws in various jurisdictions could have an impact on gathering and verifying the data needed to identify the median of the annual total compensation of all employees. Commenters in the pre-proposal period expressed concern that, in some cases, data privacy laws of foreign countries could prohibit a registrant's collection and transfer of personally identifiable compensation data that would be needed to identify the median employee. We also noted that some data privacy laws may make the collection or transfer of the underlying data more burdensome, but do not actually prohibit transfer of compensation data.

        For example, we indicated that multinational companies based in the United States might need to ensure compliance with data privacy regulations when taking certain actions to comply with the proposal, such as transmitting personally identifiable human resources data (``personal data'') of European Union (``E.U.'') employees onto global human resource information system networks in the United States; sending personal data in hard copy from the E.U. to the United States; or making personal data ``onward transfers'' to third-party payroll, pension, and benefits processors outside of the E.U.\142\ In some E.U. countries, employee consent is required, while in other countries consent may not be sufficient. We noted that other jurisdictions, such as Peru, Argentina, Canada, and Japan also have data privacy laws that could be implicated by the gathering of data for purposes of the proposed pay ratio disclosure.

        ---------------------------------------------------------------------------

        \142\ The E.U. Directive 95/46/EC, 1995 O.J. L 281 (European Union Directive on the Protection of Individuals with Regard to the Processing of Personal Data and on the Free Movement of Such Data) sets forth the regulatory framework governing the transfer of personal data from an E.U. Member State to a non-E.U. country.

        ---------------------------------------------------------------------------

        Although we did not propose any specific accommodation to address this concern, we stated our belief that the flexibility afforded to all registrants under the proposed rule could permit registrants to manage any potential costs arising from applicable data privacy laws. For example, the proposed rule would permit registrants in this situation to estimate the compensation of affected employees. We requested comment on whether the proposed flexibility afforded to registrants in selecting a method to identify the median, such as the use of statistical sampling or other reasonable estimation techniques and the use of consistently applied compensation measures to identify the median employee, could enable registrants to better manage any potential costs and burdens arising from local data privacy regulations or if there are other alternatives that would be consistent with Section 953(b).

        ii. Comments on the Proposed Rule

        Many commenters agreed that non-U.S. employees should be included in a registrant's pay ratio disclosure.\143\ Some commenters contended that failing to include non-U.S. employees would cause the pay ratio disclosure to be incomplete, less informative, and misleading.\144\ Some commenters stressed that Congress intended to include non-U.S. employees because Section 953(b) refers specifically to ``all employees.'' \145\ One commenter suggested that the exclusion of non-

        U.S. employees would not reduce the regulatory burdens on registrants because registrants with such employees would have substantial flexibility in identifying the median employee.\146\

        ---------------------------------------------------------------------------

        \143\ See, e.g., letters from AFL-CIO I, AFL-CIO II, AFSCME, Bricklayers International, CalPERS, Calvert, Chicago Teachers Fund, Corayer, CT State Treasurer, Cummings Foundation, CUPE, Estep, Fedewa, First Affirmative, Form Letter C, Gould, Douglas J. Matteson (Oct. 5, 2013) (``Matteson''), ICCR, IL Bricklayers and Craftworkers Union, Marco Consulting, McMorgan & Co., Novara Tesija, NY Bricklayers and Craftworkers Union, NY State Comptroller, Oxfam, Pax World Funds, Charmaine M. Phillips (Oct. 16, 2013) (``C. Phillips''), Public Citizen I, Public Citizen II, Douglas C. Rand (Sep. 26, 2013) (``Rand''), Sen. Menendez et al. I, Sen. Menendez et al. II, Socially Responsive Financial Advisors, Stephen Spofford (Sep. 24, 2013) (``S. Spofford''), Teamsters, Trillium I, Trustee Campbell, US SIF, and Walden.

        \144\ See, e.g., letters from AFR, Bacirctirente et al., CII, Domini, and FS FTQ.

        \145\ See, e.g., letters from AFL-CIO II, IPS, and Sen. Menendez et al. II.

        \146\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        Many commenters, however, disagreed with the proposed rule and contended that the final rule should include only the registrant's U.S. employees because including non-U.S. employees would be very costly and/or distort the pay ratio.\147\ A number of commenters asserted that the costs to registrants of including non-U.S. employees would outweigh any benefits of the disclosure to shareholders,\148\ and offered a variety of different estimates of how greatly the inclusion of non-U.S. employees would affect costs.\149\

        ---------------------------------------------------------------------------

        \147\ See, e.g., letters from Appleby, Avery Dennison, Chamber I, Chesapeake Utilities, Dover Corp., FuelCell Energy, Garmin, Hay Group, Meridian, NACD, PM&P, SH&P, Vectren Corp., Vivient, WorldatWork I, and WorldatWork II.

        \148\ See, e.g., letters from ABA, American Benefits Council, Business Roundtable I, Chamber I, Corporate Secretaries, ExxonMobil, Frederic W. Cook & Co., Inc. (Nov. 29, 2013) (``Frederic W. Cook & Co.''), RILA, Semtech, and Society for Human Resource Management (Dec. 30, 2013) (``SHRM'').

        \149\ See, e.g., letters from American Benefits Council (stating that costs could be 20 to 30 times higher if non-U.S. employees are included and also that some of its ``member companies have estimated that the annual cost to make the median employee determination on a worldwide basis could be hundreds of thousands of dollars''), Business Roundtable I (stating that costs could decrease by over 50% in some cases if non-U.S. employees are excluded), ExxonMobil, and FEI (stating that costs would decrease by 90% if the final rule excluded non-U.S. employees).

        ---------------------------------------------------------------------------

        Page 50121

        Additionally, commenters noted that companies with international operations almost always have multiple payroll systems and databases for their employees' compensation that are difficult, if not impossible, to reconcile,\150\ with some of the commenters providing the following examples:

        ---------------------------------------------------------------------------

        \150\ See, e.g., letters from American Benefits Council, Aon Hewitt, BCIMC, Business Roundtable I, COEC I, COEC II, COEC III, Cummins Inc., Eaton, ExxonMobil, Freeport-McMoRan, MVC Associates International (Nov. 28, 2013) (``MVC Associates''), NIRI, Semtech, SHRM, and Tesoro Corporation (Nov. 21, 2013) (``Tesoro Corp'').

        ---------------------------------------------------------------------------

        One commenter indicated that it has 15 payroll systems that are not integrated, and those systems would have to be manually reconciled with ``substantial costs'' and ``extensive staff hours;'' \151\

        ---------------------------------------------------------------------------

        \151\ See letter from Freeport-McMoRan.

        ---------------------------------------------------------------------------

        one commenter stated that it has 30 payroll systems that do not interface; \152\

        ---------------------------------------------------------------------------

        \152\ See letter from Cummins Inc.

        ---------------------------------------------------------------------------

        one commenter cited a Human Resource Policy Association survey concluding that 84% of respondents could not easily calculate worldwide enterprise cash compensation for all their employees; \153\

        ---------------------------------------------------------------------------

        \153\ See letter from MVC Associates.

        ---------------------------------------------------------------------------

        one commenter cited its own survey finding that a registrant on average maintains 46 different payroll systems in 34 different countries; \154\ and

        ---------------------------------------------------------------------------

        \154\ See letter from COEC I.

        ---------------------------------------------------------------------------

        one commenter stated that it does not have a single payroll system.\155\

        ---------------------------------------------------------------------------

        \155\ See letter from Tesoro Corp.

        ---------------------------------------------------------------------------

        Several commenters stated that many countries have data privacy and other laws that prevent registrants from transferring payroll data outside that country's borders (even if the transfer would be within the same company), which would make compiling the information necessary for the pay ratio problematic or even illegal.\156\ One of these commenters recommended that, if the final rule did not exclude all non-

        U.S. employees, it should permit registrants to exclude from their methodology for identifying the median employee the data for any employees in a jurisdiction where such collection, analysis, and transmission would violate a registrant's existing data privacy obligations.\157\ A few commenters stated that the proposed rule's flexibility afforded to all registrants could permit registrants to manage any potential costs arising from applicable data privacy laws.\158\

        ---------------------------------------------------------------------------

        \156\ See, e.g., letters from AAFA II, ABA, American Benefits Council, BCIMC, Business Roundtable I, Chamber I, COEC I, COEC II, Corporate Secretaries, Cummins Inc., Eaton, ExxonMobil, Freeport-

        McMoRan, Hyster-Yale, IBC, NACCO, NAM I, NAM II, NIRI, RILA, Semtech, SHRM, and WorldatWork I.

        \157\ See letter from ABA.

        \158\ See, e.g., letters from Capital Strategies (stating that data privacy laws would not affect registrants because U.S. firms must know their employee expenses and can use sampling techniques to negate any data privacy effects) and WorldatWork I (stating that, while data privacy laws will have a negative effect on some registrants, others may be able to gather the required information under existing waivers granted to them by the EU, and some registrants can estimate total compensation of their employees in countries with data privacy laws by placing employees in ``bands'' of similar compensation and benefits levels, and estimating total compensation using those bands).

        ---------------------------------------------------------------------------

        One commenter contended that any data privacy concerns can be addressed easily by anonymizing payroll data sets or conducting statistical sampling.\159\ If, however, these privacy safeguards proved insufficient, the commenter recommended that registrants be permitted to exclude employees in such countries only if they disclose the number of employees in the excluded countries and obtain and file as an exhibit to the periodic report in which the pay ratio disclosure appears a legal opinion by qualified outside counsel demonstrating that the privacy safeguards are inadequate under local law.

        ---------------------------------------------------------------------------

        \159\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        Additionally, a number of commenters contended that including non-

        U.S. employees in the final rule would distort the pay ratio because of (1) differences in local pay practices,\160\ with some of these comment letters stating that, unlike many U.S. companies, companies in other countries include as ``compensation'' transportation, food, housing, wedding, birth, education, and phone expenses, as well as profit-

        sharing arrangements and government provided benefits; \161\ (2) the exchange rates of foreign currencies; \162\ and (3) cost-of-living differences among countries.\163\

        ---------------------------------------------------------------------------

        \160\ See, e.g., letters from AAFA I, ABA, Aon Hewitt, American Benefits Council, Business Roundtable I, COEC I, Cummins Inc., Eaton, ExxonMobil, FEI, Freeport-McMoRan, Hyster-Yale, IBC, KBR, NACCO, NYC Bar, Prof. Ray, RILA, Semtech, and SHRM.

        \161\ See, e.g., letters from Eaton, Freeport-McMoRan, and SHRM.

        \162\ See, e.g., letters from American Benefits Council, BCIMC, Business Roundtable I, Cummins Inc., ExxonMobil, FEI, Freeport-

        McMoRan, Prof. Ray, IBC, NAM I, NAM II, NYC Bar, Semtech, and SHRM.

        \163\ See, e.g., letters from AAFA I, American Benefits Council, BCIMC, COEC I, Corporate Secretaries, Cummins Inc., ExxonMobil FEI, Freeport-McMoRan, IBC, NAM I, NAM II, NYC Bar, and RILA.

        ---------------------------------------------------------------------------

        A few commenters argued that excluding non-U.S. employees was supported by statutory construction. They contended that, despite Section 953(b)'s reference to ``all employees,'' this statutory reference does not require the final rule to include non-U.S. employees.\164\ Some of these commenters asserted that excluding non-

        U.S. employees would be consistent with Section 953(b) because the statute is silent as to whether ``all employees'' means non-U.S. employees.\165\ Others insisted that interpreting ``all employees'' to exclude non-U.S. employees would be appropriate because of a presumption against the extraterritoriality of U.S. laws.\166\

        ---------------------------------------------------------------------------

        \164\ See, e.g., letters from COEC I, ExxonMobil, Frederic W. Cook & Co., Freeport-McMoRan, NIRI, and NYC Bar.

        \165\ See, e.g., letters from Freeport-McMoRan, and NIRI.

        \166\ See, e.g., letters from COEC I and ExxonMobil.

        ---------------------------------------------------------------------------

        Some commenters advocated for a de minimis exemption for non-U.S. employees \167\ because, as one of these commenters stated, excluding a small number of employees is unlikely to affect ``in a material way'' the pay ratio and the nominal differences in ratios would be outweighed by the cost savings to registrants.\168\ Commenters provided a number of suggestions for a de minimis exemption. One commenter suggested that, if non-U.S. employees make up less than 20% of all a registrant's employees, the registrant should be permitted to exclude all non-U.S. employees.\169\ Another commenter stated that registrants should be permitted to exclude non-U.S. employees in any foreign country that comprises less than 5% of the registrant's aggregate global workforce.\170\

        ---------------------------------------------------------------------------

        \167\ See, e.g., letters from American Benefits Council (arguing in the alternative to excluding non-U.S. employees), Corporate Secretaries, ExxonMobil, Financial Services Institute (Dec. 2, 2013) (``FSI''), FSR, NACCO (arguing in the alternative to excluding non-

        U.S. employees), NYC Bar (arguing in the alternative to excluding non-U.S. employees), and PNC Financial Services.

        \168\ See letter from PNC Financial Services.

        \169\ See letter from NACCO.

        \170\ See letter from American Benefits Council.

        ---------------------------------------------------------------------------

        A different commenter recommended that a registrant be permitted to exclude non-U.S. employees if they account for less than 5% of the registrant's total workforce because they would represent a de minimis number of employees.\171\ Also, according to that commenter, if the registrant's foreign employees account for more than 5% of all employees, this commenter recommended that the registrant should be permitted to exclude employees in any single foreign jurisdiction if they comprise less than 2% of total

        Page 50122

        employees, with an aggregate cap of 5%, and that the registrant be allowed to choose which country's employees to exclude.\172\ Another commenter noted that a registrant should be permitted to exclude non-

        U.S. employees in a foreign country if the number of employees in that country is less than 1% of the registrant's total workforce.\173\

        ---------------------------------------------------------------------------

        \171\ See letter from FSR.

        \172\ See id.

        \173\ See letter from ExxonMobil.

        ---------------------------------------------------------------------------

        One commenter recommended that a registrant be able to exclude non-

        U.S. employees if its CEO is based in the United States and more than 50% of the registrant's employees also are based in the United States.\174\ Conversely, a different commenter contended that we should use our discretion to limit the scope of the term ``all employees'' to only U.S. employees if ``most of the registrant's employees'' and its principal executive officer ``work primarily outside the United States.'' \175\ Alternatively, the commenter suggested that, if the final rule includes non-U.S. employees, the rule should not include as ``compensation'' the value of pension accruals, employer matching contributions for 401(k)s, and non-cash benefits. The commenter also advocated that, if non-U.S. employees are included, the final rule should exclude from the median identification employees who work in any E.U. jurisdiction that has strict data privacy rules and employs fewer than 50 people.

        ---------------------------------------------------------------------------

        \174\ See letter from NYC Bar.

        \175\ See letter from ABA. See also letter from NACCO (suggesting that, if non-U.S. employees make up more than 80% of a registrant's employees, the registrant should be exempt from the rule entirely).

        ---------------------------------------------------------------------------

        Another commenter contended that the final rule should include a principles-based exclusion that would permit companies the flexibility to exclude substantial percentages of employees if their compensation data is difficult to obtain and the impact would not be significant.\176\ The commenter cited the two memoranda analyzing the potential effects of excluding different percentages of employees on the pay ratio calculation that the staff from the Division of Economic and Risk Analysis (``DERA'').\177\ According to the commenter, these memoranda concluded that excluding a ``large share'' of employees from the pay ratio calculation would ``not have a significant impact'' on the ratio. As an example, the commenter noted that the June 30 Memorandum states that excluding 40% of a registrant's employee population could reduce the pay ratio by 10.77% or increase it by 12.08%. The commenter asserted these amounts would be ``negligible,'' and therefore registrants should be permitted to exclude these employees from their pay ratio calculations under the principles-based exclusionary approach advocated by the commenter. Additionally, the commenter noted that the average respondent to a survey it conducted indicated that 40% of its employees are located overseas and excluding these employees would decrease the compliance costs of the rule by 47%.

        ---------------------------------------------------------------------------

        \176\ See letter from COEC III.

        \177\ See Memorandum of the Division of Economic and Risk Analysis regarding the potential effect on pay ratio disclosure of exclusion of different percentages of employees at a range of thresholds (Jun. 4, 2015) (``June 4 Memorandum'') available at http://www.sec.gov/comments/s7-07-13/s70713-1556.pdf and Memorandum from the Division of Economic and Risk Analysis regarding an extension of the analysis of the potential effect on pay ratio disclosure of exclusion of different percentages of employees at a range of thresholds (Jun. 30, 2015) (``June 30 Memorandum'') available at http://www.sec.gov/comments/s7-07-13/s70713-1559.pdf. When the June 4 Memorandum was placed in the public comment file, we provided a press release announcing it was issued and an electronic alert through our RSS feed. The press release expressly advised that additional staff analyses might be placed in the comment file.

        ---------------------------------------------------------------------------

        iii. Final Rule

        (a) Non-U.S. Employees Generally

        After considering the comments, we have determined to include in the final rule's definition of ``employee'' a registrant's U.S. and non-U.S. employees, as proposed. We believe that the inclusion of non-

        U.S. employees is the policy choice that most closely captures what Congress directed us to do in stating that the ratio should reflect ``all employees.'' As noted above, we believe that the use of the word ``all'' provides a general direction in favor of inclusion rather than exclusion of broad categories of employees. Although we recognize that our reading may impose more costs on registrants than if we excluded non-U.S. employees, given that Congress is undoubtedly aware that many U.S. registrants operate globally, we think Congress's use of ``all employees'' without any territorial limitation is a strong indication that Congress did not want us to categorically exclude non-U.S. employees. Further supporting our conclusion is the fact that, historically, the disclosures that are required of registrants under the securities laws apply to events, assets, conduct, or persons irrespective of whether those are located in the United States or abroad, and there is no indication that Congress intended to depart from this historical approach. However, as discussed below, we are adopting several tailored exemptions to address specific concerns raised by commenters.

        With respect to the assertion by some commenters that the rule should exclude non-U.S. employees because there is a presumption against the extraterritoriality of U.S. laws,\178\ we do not believe that including foreign employees within the ``employee'' definition constitutes an extraterritorial application of the statute. Generally, whether a particular application of a statute is ``extraterritorial'' turns on an analysis of whether the conduct that is the Congressional focus of the statute occurs here or abroad.\179\ The Congressional focus of Section 953(b) is disclosure by registrants that avail themselves of the U.S. public markets and thereby submit to U.S. law.\180\ As discussed above, Section 953(b)(1) directs us to amend Item 402, ``to require each issuer to disclose'' the pay ratio information ``in any filing'' described in Item 10(a). Companies are only required to provide a ``filing'' with us if they offer and/or sell securities in the United States and become subject to the our registration and filing requirements. Therefore, the final rule affects only registrants that come under the umbrella of United States laws.

        ---------------------------------------------------------------------------

        \178\ See, e.g., letters from COEC I and ExxonMobil.

        \179\ See Morrison v. Nat'l Australia Bank Ltd., 561 U.S. 247 (2010).

        \180\ In this regard, we note that it is important that shareholders have access to information about both the domestic and foreign operations of a registrant given that a shareholder's investment in a U.S. registrant is typically exposed to the risks of, and the returns generated by, the global operations of the registrant.

        ---------------------------------------------------------------------------

        Our exemptive authority under Section 36 of the Exchange Act and Section 28 of the Securities Act would allow us to exempt registrants from including non-U.S. employees in the median employee determination required by Section 953(b).\181\ However,

        Page 50123

        after careful consideration, we decline to exercise our discretion to grant a wholesale exemption for non-U.S. employees. As we understand Section 953(b), Congress thought that it was important for shareholders to have pay ratio disclosure that reflects ``all employees'' of a registrant when making their say-on-pay votes. We do not believe it would be appropriate to second-guess Congress by granting a wholesale exemption for non-U.S. employees in a manner that we believe could fundamentally change the pay ratio information that Congress directed be provided to shareholders.\182\

        ---------------------------------------------------------------------------

        \181\ Section 36(a) of the Exchange Act and Section 28 of the Securities Act afford us general exemptive authority to conditionally or unconditionally exempt any person, security, or transaction, or any class or classes of persons, securities, or transactions, from any provision or provisions of this title (i.e., the Exchange Act and the Securities Act, respectively) or of any rule or regulation thereunder, to the extent that such exemption is necessary or appropriate in the public interest, and is consistent with the protection of investors. Although Section 953(b) was not expressly incorporated into either Act, our view is that, to the extent that the statutory criteria for invoking exemptive authority under these sections are met, the exemptive authorities afforded by Section 36(a) and Section 28 are available here. We construe Section 953(b) as a Congressional directive to us to rely on our Exchange Act and Securities Act rulemaking authorities to amend Sec. 229.402 of title 17, Code of Federal Regulations to require the pay ratio disclosure. The pay ratio amendments that we are adopting, therefore, are rules or regulations under the Exchange Act and the Securities Act and, thus, fall within the express terms of Section 36(a) and Section 28.

        \182\ Our use of exemptive authority where foreign data privacy laws are involved (as discussed below) may result in different pay ratio disclosure in certain instances, but we believe the use of exemptive authority is appropriate because it reflects a carefully tailored accommodation necessary to address a situation (i.e., foreign data privacy laws) that Congress may not have contemplated.

        ---------------------------------------------------------------------------

        While the final rule does not exclude non-U.S. employees, in response to concerns that the inclusion of non-U.S. employees could raise compliance costs for multinational companies, introduce cross-

        border compliance issues, and have an adverse impact on competition, we are exercising our exemptive authority to provide two tailored exemptions that we believe will alleviate some of these concerns: (1) an exemption that applies when a foreign jurisdiction's data privacy laws or regulations are such that, despite its reasonable efforts to obtain or process information necessary to comply with the rule, a registrant is unable to do so without violating those laws or regulations, and (2) a de minimis exemption. These exemptions are discussed in detail below.

        (b) Foreign Data Privacy Law Exemption

        The first instance in which we believe it is appropriate to provide an exemption to the general requirement that non-U.S. employees be included in the pay ratio disclosure is when a jurisdiction's data privacy laws or regulations are such that, despite a registrant's reasonable efforts to obtain or process information necessary to comply with the rule, it is unable to do so without violating those laws or regulations. A number of commenters noted that many countries have data privacy and other laws that prevent registrants from transferring payroll data outside that country's borders (even if the transfer would be within the same company), which would make compiling the information necessary for the pay ratio disclosure illegal.\183\ For example, commenters noted that the E.U. prohibits the transfer of personal data to a third country that does not ensure an adequate level of privacy protections (the United States is considered not to ensure adequate privacy protections) and that China, Japan, Mexico, Canada, Peru, Australia, Russia, Switzerland, Argentina, and Singapore have adopted or are considering similar rules.\184\

        ---------------------------------------------------------------------------

        \183\ See, e.g., letters from AAFA II, ABA, American Benefits Council, BCIMC, Business Roundtable I, Chamber I, COEC I, Corporate Secretaries, Cummins Inc., Eaton, ExxonMobil, Freeport-McMoRan, Hyster-Yale, IBC, NACCO, NAM I, NAM II, NIRI, RILA, Semtech, SHRM, and WorldatWork I.

        \184\ See, e.g., letters from ABA, American Benefits Council, Business Roundtable I, COEC I, Cummins Inc., NAM I, NAM II, RILA, and WorldatWork I.

        ---------------------------------------------------------------------------

        One of these commenters acknowledged, however, that ``it would be reasonable to expect that registrants which employ workers abroad already have an understanding of their obligations under the data privacy laws of each jurisdiction in which they operate, and have undertaken to comply with those laws,'' but the commenter was concerned that existing actions taken by registrants to comply with those laws may not be sufficiently flexible to facilitate compliance with the rule.\185\ In this regard, the commenter noted that, under the E.U. Directive, data may be exempt from the dictates of the Directive if it is truly anonymous such that the data cannot be attributed to any identifiable person. It would seem, therefore, that a registrant employing hundreds or thousands of employees in an E.U. jurisdiction could collect compensation data for purposes of complying with Section 953(b) in a way that would preserve employee anonymity while a registrant that employs only a handful of employees in an E.U. jurisdiction may not be able to collect the data in such a manner.\186\ The commenter therefore recommended that, if the final rule did not exclude all non-U.S. employees, it should permit registrants to exclude from the definition of ``employee'' any employee in a jurisdiction where such collection, analysis, and transmission would violate a registrant's existing data privacy obligations.\187\

        ---------------------------------------------------------------------------

        \185\ See letter from ABA.

        \186\ In such situations, we recognize that the de minimis exemption may be available.

        \187\ See letter from ABA (recommending that we consider permitting registrants to exclude from the identification of the median employee those employees who work in an E.U. jurisdiction that maintains strict data privacy laws and in which the registrant employs fewer than 50 employees).

        ---------------------------------------------------------------------------

        After considering the comments received, we are persuaded that a tailored exemption from the definition of ``employee'' is appropriate where a foreign country's data privacy laws or regulations are such that a registrant is not able to comply with the rule without violating those laws or regulations in spite of its reasonable efforts to obtain or process the necessary information.\188\

        ---------------------------------------------------------------------------

        \188\ As required by Section 28 of the Securities Act and Section 36 of the Exchange Act, we find that the exemption here is consistent with investor protection and is necessary or appropriate in the public interest. We make these findings based on the fact that, without the exemption, registrants operating in countries with applicable privacy laws could be forced into the difficult situation of violating either that country's laws or U.S. law, and we believe that because of the limited and tailored nature of the exemption, it will not materially impact the pay ratio disclosure.

        ---------------------------------------------------------------------------

        Although, as noted above, we believe the inclusion of non-U.S. employees is consistent with the Congressional directive and is important for providing pay ratio information that reflects a registrant's overall employment practices, we do not have any indication that Congress intended that a registrant should have to choose between complying with our disclosure rules and violating the laws of a foreign jurisdiction. We believe that, on balance, providing an accommodation in such situations would not substantially affect the utility of the Section 953(b) disclosures for shareholder say-on-pay votes.

        To prevent any potential manipulation, the rule requires the registrant to exercise reasonable efforts to obtain or process the information necessary for compliance with the final rule. As part of its reasonable efforts, the registrant must seek an exemption or other relief under the applicable jurisdiction's governing data privacy laws or regulations and use the exemption if granted.

        If a registrant excludes any non-U.S. employees in a particular jurisdiction under the data privacy exemption, it must exclude all non-

        U.S. employees in that jurisdiction. Additionally, the registrant must list the excluded jurisdictions, identify the specific data privacy law or regulation, explain how complying with the final rule violates the law or regulation (including the efforts made by the registrant to use or seek an exemption or other relief under such law or regulation), and provide the approximate number of employees exempted from each jurisdiction based on this exemption.

        Also, the registrant must obtain a legal opinion that opines on the inability of the registrant to obtain or process the information necessary for compliance with the final rule without violating that jurisdiction's laws or regulations governing data privacy, including the registrant's inability to obtain an exemption or other relief under any

        Page 50124

        governing laws or regulations. The legal opinion must be filed as an exhibit with the filing in which the pay ratio disclosure is included. For filings other than proxy or information statements, the legal opinion must be filed as an exhibit under Exhibit 99.\189\

        ---------------------------------------------------------------------------

        \189\ See Item 601(b)(99) of Regulation S-K 17 CFR 229.601(b)(99). The exhibit must be filed with proxy or information statements pursuant to Item 25 of Schedule 14A.

        ---------------------------------------------------------------------------

        (c) De Minimis Exemption

        The second instance in which we believe it is appropriate to provide an exemption from the general requirement to include non-U.S. employees in identifying the median employee is when a de minimis number of a registrant's employees work outside the United States. The de minimis exemption is a change from the proposed rule. Under this exemption, registrants whose non-U.S. employees make up 5% or less of their total U.S. and non-U.S. employees may exclude all of them when identifying their median employee. If such a registrant chooses to exclude any non-U.S. employees under this exemption, it must exclude all of them. A registrant with more than 5% non-U.S. employees may also exclude non-U.S. employees up to the 5% threshold; provided that, if such a registrant excludes any non-U.S. employees in a particular foreign jurisdiction, it must exclude all the employees in that jurisdiction. The registrant may not pick and choose which employees to exclude in any one jurisdiction.

        We believe a de minimis exemption provides flexibility in a manner that will not meaningfully alter the pay ratio disclosure. We are persuaded that a de minimis exemption is appropriate after considering the potential cost savings to registrants and the small effect it would have on the pay ratio, as discussed below. The final rule establishes the de minimis threshold at 5%. The commenters that suggested specific de minimis thresholds did not provide reasons why the particular thresholds they suggested were suitable, but several of these commenters suggested a threshold of 5%.\190\ We believe the 5% threshold will both limit the exemption to an amount that is, in fact, de minimis and help address the payroll or other data challenges that may arise for registrants with a small percentage of non-U.S. employees.

        ---------------------------------------------------------------------------

        \190\ See letters from PNC Financial Services (suggesting 5% as the de minimis threshold) and FSR (recommending a registrant be permitted to exclude non-U.S. employees if they account for less than 5% of the registrant's employees, but, if the registrant's foreign employees account for more than 5% of all employees, the registrant may exclude employees in any single foreign jurisdiction if they comprise less than 2% of total employees, with an aggregate cap of 5%).

        ---------------------------------------------------------------------------

        Although commenters did not provide data about the effect on the pay ratio of potential de minimis thresholds, staff in DERA performed an analysis of the potential effect on pay ratio disclosure of excluding different percentages of employees at a range of thresholds and posted to the comment file two memoranda containing the analysis.\191\ As discussed in further detail both in the memoranda and in the Economic Analysis section below, under such analysis and based on the assumptions set forth in the analysis, the exclusion of 5% of employees may cause the pay ratio calculation to decrease by up to 3.4% or to increase by up to 3.5%. We believe such analysis confirms that the effect of the 5% threshold on the pay ratio disclosure will be de minimis. We note that one commenter observed that under one of the scenarios analyzed in the June 30 Memorandum, excluding 40% of a registrant's employees may cause the pay ratio calculation to decrease by up to 10.77% or to increase by up to 12.08%.\192\ The commenter called this impact ``not significant'' and ``negligible'' and, based on this impact, contended that the final rule should permit registrants to exclude up to 40% of their non-U.S. employees. We are not adopting this suggestion. We believe that the exclusion of 40% of employees would be a fundamentally different type of exclusion than the one we adopting here--that is, a de minimis exclusion designed to allow companies to exclude employees in jurisdictions where there are only a limited number of employees and where the costs of including such employees may be disproportionately greater than the incremental information they would add to the disclosure. The exclusion of 40% of employees is not a de minimis amount of employees, and, in contrast to the assertions of the commenter, we believe that a decrease in the pay ratio calculation of up to 10.77% or increase by up to 12.08% is significant and more than negligible, and we do not believe that it is de minimis. Additionally, as the commenter acknowledges, under other scenarios in the memorandum, exclusion of up to 40% of a registrant's employees could have an even greater effect on the pay ratio (e.g., by causing it to decrease by up to 25.06% or to increase by up to 33.43%). In contrast, the 5% de minimis threshold both results in the exclusion of a de minimis number of employees and has a de minimis effect on the pay ratio, regardless of which scenario is considered. Accordingly, the final rule's de minimis threshold for non-U.S. employees does not exceed 5%.

        ---------------------------------------------------------------------------

        \191\ See the June 4 Memorandum and the June 30 Memorandum. See also, Section III.D.2.c.vi below.

        \192\ See letter from COEC III.

        ---------------------------------------------------------------------------

        As one commenter warned, there is a possibility for intentional manipulation in identifying the median employee when a de minimis exemption is provided and a registrant is permitted to choose which jurisdictions to exclude.\193\ To provide safeguards against any potential manipulation, the final rule requires that, if a registrant with 5% or fewer non-U.S. employees chooses to exclude those employees from the calculation of its median employee, it may not pick and choose which of the 5% to exclude and must exclude all of its non-U.S. employees. Similarly, if a registrant with more than 5% non-U.S. employees excludes any employees in any jurisdiction, it must exclude all the employees in that jurisdiction. In this regard, we recognize that this requirement could prevent some registrants with more than 5% non-U.S. employees from excluding any of its foreign employees. A purpose of the de minimis exemption is to provide relief from the need to determine how to integrate payroll systems and compensation arrangements in jurisdictions where the number of employees may not justify the effort, and we believe that setting the threshold at 5% establishes an appropriate measure of relief.

        ---------------------------------------------------------------------------

        \193\ See letter from FSR.

        ---------------------------------------------------------------------------

        The final rule also requires a registrant using the de minimis exemption to provide certain disclosures. If the registrant excludes any non-U.S. employees under the de minimis exemption, it must disclose the jurisdiction or jurisdictions from which employees are being excluded, the approximate number of employees excluded from each jurisdiction under the de minimis exemption, the total number of its U.S. and non-U.S. employees irrespective of any exemption (data privacy or de minimis) and the total number of its U.S. and non-U.S. employees used for its de minimis calculation.

        In calculating the number of non-U.S. employees that may be excluded under the de minimis exemption, a registrant must count any non-U.S. employee exempted under the data privacy exemption against the availability. A registrant may exclude any non-U.S. employee that meets the data privacy exemption, even if the number of excluded employees exceeds 5% of the registrant's total employees. If, however,

        Page 50125

        the number of employees excluded under the data privacy exemption equals or exceeds 5% of the registrant's total employees, the registrant may not use the de minimis exemption to exclude additional non-U.S. employees.

        For example, a registrant has non-U.S. employees located in two foreign jurisdictions. One of the jurisdictions has 10% of the registrant's total employees who are non-U.S. employees and has data privacy laws that, despite its reasonable efforts to obtain or process the information necessary for compliance with the final rule, the registrant is unable to do so without violating those data privacy laws or regulations. The other jurisdiction has an additional 5% of the registrant's total employees who are non-U.S. employees and has no such data privacy laws or regulations. The registrant may exclude all the non-U.S. employees in the first jurisdiction, which has 10% of the registrant's total employees. In that situation, however, the registrant may not exclude the non-U.S. employees in the second jurisdiction, which has the additional 5% of the total employees, even though the 5% would otherwise constitute a de minimis amount of non-

        U.S. employees, because the registrant is already excluding over 5% of its employees under the data privacy exemption.\194\

        ---------------------------------------------------------------------------

        \194\ We do not envision a scenario in which a registrant can forgo the data privacy exemption in favor of the de minimis exemption in the above or similar situations. The data privacy exemption is permitted only for circumstances in which a foreign jurisdiction's laws or regulations governing data privacy are such that a registrant is unable to comply with the final rule without violating the that jurisdiction's laws or regulations. If a registrant is in a position to forgo the data privacy exemption, it would not be considered eligible for the exemption.

        ---------------------------------------------------------------------------

        Moreover, if the number of non-U.S. employees excluded under the data privacy exemption is less than 5% of the registrant's total employees, the registrant may use the de minimis exemption to exclude no more than the number of non-U.S. employees that, combined with the data privacy exemption, equals 5% of the registrant's total employees.

        For example, a registrant has non-U.S. employees located in two foreign jurisdictions. One of the jurisdictions has 2.5% of the registrant's total employees who are non-U.S. employees and has data privacy laws that, despite its reasonable efforts to obtain or process the information necessary for compliance with the final rule, the registrant is unable to do so without violating those data privacy laws or regulations. The other jurisdiction has an additional 2.5% of the registrant's total employees who are non-U.S. employees and has no such data privacy laws or regulations. The registrant may exclude the 2.5% of total employees who are non-U.S. employees in the first jurisdiction under the data privacy exemption. The registrant may also exclude the additional 2.5% of the registrant's total employees who are non-U.S. employees from the second jurisdiction because the total number of exempted non-U.S. employees under both the data privacy and the de minimis exemptions equal only 5% of the registrant's total employees.

        Alternatively, in the above example, if the number of non-U.S. employees in the second jurisdiction was 3% of the registrant's total employees, the registrant could not exclude the non-U.S. employees in that jurisdiction because the registrant's number of excluded non-U.S. employees in both jurisdictions would be over 5% of its total employees.

        (d) Cost-of-Living Adjustment

        In the Proposing Release, we requested comment on whether we should permit cost-of-living adjustments for employees in different countries. A number of commenters who addressed this issue contended that unadjusted cost of living differences between countries would cause the inclusion of non-U.S. employees to render the pay ratio disclosure misleading.\195\ Accordingly, some of these commenters suggested that the pay ratio disclosure could be more meaningful for some registrants if the final rule permitted cost-of-living adjustments.\196\ However, other commenters objected to permitting cost-of living-

        adjustments,\197\ asserting that the compensation of a non-U.S. employee ``is more directly comparable to the total compensation of a registrant's PEO without a cost-of-living adjustment,'' \198\ and that permitting cost-of-living adjustments could add a level of subjectivity to the pay ratio disclosure.\199\

        ---------------------------------------------------------------------------

        \195\ See, e.g., letters from AAFA I, American Benefits Council, BCIMC, Corporate Secretaries, Cummins Inc., ExxonMobil, NAM I, NAM II, and SH&P.

        \196\ See letter from ExxonMobil (``Example 2: Differences in cost of living. Two employees hold similar jobs in two different countries, C and D. The employee in Country C receives total annual compensation of $100,000 and the employee in Country D receives total annual compensation of $75,000. The cost of living in Country D is approximately 50% of the cost of living in Country C. In real economic terms, the employee in Country D enjoys significantly higher pay than the employee in Country C. However, under the rules as proposed the appearance is reversed.'') (Emphasis in original.). See also, letters from NAM I, NAM II, and SH&P.

        \197\ See letters from ABA, Prof. Ray, and WorldatWork I.

        \198\ See letter from ABA (``We believe that the Commission should not allow registrants to make cost-of-living adjustments for non-U.S.-based employees (should the agency determine to include them), other than the annualization and full-time equivalent adjustments discussed above in our responses to Questions 23 and 24. We believe the total compensation of such full-time employees is more directly comparable to the total compensation of a registrant's principal executive officer without a cost-of-living adjustment than with it.'').

        \199\ See letter from Prof. Ray.

        ---------------------------------------------------------------------------

        We acknowledge that differences in the underlying economic conditions of the countries in which registrants operate likely have an effect on the compensation paid to employees in those jurisdictions. As a result, requiring registrants to determine their median employee and calculate the pay ratio without permitting them to adjust for these different underlying economic conditions could result in what some would consider a statistic that does not appropriately reflect the value of the compensation paid to individuals in those countries. The final rule, therefore, allows registrants the option to make cost-of-

        living adjustments to the compensation of their employees in jurisdictions other than the jurisdiction in which the PEO resides when identifying the median employee (whether using annual total compensation or any other consistently applied compensation measure), provided that the adjustment is applied to all such employees included in the calculation.\200\ If the registrant chooses this option, the compensation of such employees will have to be adjusted to the cost of living in the jurisdiction in which the PEO resides. Further, if the registrant uses a cost-of-living adjustment to identify the median employee, and the median employee identified is an employee in a jurisdiction other than the one in which the PEO resides, the registrant must use the same cost-of-living adjustment in calculating the median employee's annual total compensation and disclose the median employee's jurisdiction. If a registrant does not make cost-of-living-

        adjustments to its employees when identifying the median employee, the registrant is not permitted to make cost-of-living adjustments to the median employee's annual total compensation if the median employee is an employee in

        Page 50126

        a jurisdiction other than the jurisdiction in which the PEO resides.

        ---------------------------------------------------------------------------

        \200\ We are utilizing our authority under Section 28 of the Securities Act and Section 36 of the Exchange Act to permit registrants to elect to identify the median employee by first adjusting the compensation of employees in jurisdictions other than the jurisdiction in which the PEO resides to the cost of living in the CEO's jurisdiction of residence. Moreover, for the reasons described above, we believe that this conditional exemption is consistent with the public interest and investor protection.

        ---------------------------------------------------------------------------

        In the Proposing Release, we said that we preliminarily believed that certain adjustments, including cost-of-living adjustments, ``could distort an understanding of the registrant's compensation practices.'' Based on our fuller understanding of the Congressional purpose underlying the pay ratio disclosure and the comments received on the proposal, however, we are persuaded that allowing registrants the option of a cost-of-living adjustment for employees in jurisdictions other than the jurisdiction in which the PEO resides could be useful to investors as they make their say-on-pay votes. Put simply, a cost-of-

        living adjustment could provide a more meaningful comparison of the PEO's compensation to the actual value of the median employee's compensation by effectively filtering out that part of the difference in compensation that results from differences in the cost of living between the PEO's place of residence (typically, the United States) and the median employee's jurisdiction. For some shareholders making their say-on-pay votes, we believe that what may matter is the value of compensation received by the median employee, rather than the dollar amount of the compensation paid. Although we are not mandating that registrants adjust for these cost-of-living considerations, we believe that it is appropriate to give them the option to make such adjustments where they determine that doing so would provide more useful information to their shareholders as they vote on executive compensation.\201\

        ---------------------------------------------------------------------------

        \201\ Although the final rule gives registrants the option to make cost-of-living adjustments for employees in jurisdictions other than the jurisdiction in which the PEO resides, we are not giving registrants the option of adjusting part-time or seasonal employees' compensation as though they were full time employees. For U.S. part-

        time or seasonal employees, the unadjusted compensation reflects the actual relative value of the compensation received by that employee, unlike non-U.S. employees who may be working in countries with a significantly lower cost of living. Moreover, adjusting the compensation of part-time or seasonal employees to what they would have received if they had been full time employees would cause the median to not be reasonably representative of the registrant's actual employment arrangements for its workforce during the period.

        ---------------------------------------------------------------------------

        We recognize that providing registrants the flexibility to make cost-of-living adjustments could add a level of subjectivity to the pay ratio disclosure, make compliance with the rule more burdensome, or permit registrants to alter the reported ratio to achieve a particular objective with the ratio disclosure.\202\ Registrants with a significant number of employees in countries with higher cost-of-living than the jurisdiction in which the PEO resides may be unlikely to adjust those compensation figures downward, while registrants with a sizable work force in countries with a lower cost-of-living may be likely to adjust the compensation figures upward.

        ---------------------------------------------------------------------------

        \202\ See letters from Prof. Ray (stating that permitting cost-

        of-living adjustments ``will only make the pay ratio more subjective,'' because they ``add subjectivity'' to the disclosure) and WorldatWork I (``Cost-of-living adjustments and full-time compensation adjustments would make compliance more burdensome by requiring more context in the explanation of how the ratio was calculated.'').

        ---------------------------------------------------------------------------

        We believe, however, that the final rule mitigates these concerns by requiring registrants to briefly describe any cost-of-living adjustments they used to identify the median employee or to calculate annual total compensation, including the measure used as the basis for the cost-of-living adjustment, and disclose the country in which the median employee is located. Additionally, the final rule requires that any registrant electing to present the pay ratio using a cost-of-living adjustment must also disclose the median employee's annual total compensation and pay ratio without the cost-of-living adjustments. To calculate this pay ratio, the registrant will need to identify the median employee without using any cost-of-living adjustments. In this way, shareholders would have pay ratio information both in terms of the value of compensation received by the employee and in terms of the compensation paid by the registrant.\203\

        ---------------------------------------------------------------------------

        \203\ We believe that requiring this disclosure of the unadjusted pay ratio for those registrants who choose to include a cost-of-living adjustment will help to mitigate the concerns noted in the Proposing Release about the impact that a cost-of-living adjustment could have on an understanding of a registrant's compensation practices.

        ---------------------------------------------------------------------------

        For registrants who choose to present the pay ratio using a cost-

        of-living adjustment, the pay ratio required by Item 402(u)(1)(iii) will be the cost-of-living adjusted pay ratio. Disclosure of the unadjusted pay ratio will be available to provide context for the registrant's required pay ratio. Because the cost-of-living adjustment will be optional for registrants, we assume they will choose to avail themselves of this option only to the extent they believe the benefits of doing so will justify any additional costs to make the adjustment.

        d. Employees of Consolidated Subsidiaries

        i. Proposed Rule

        The proposed rule would cover employees of both a registrant and its subsidiaries, which is similar to the approach taken for other Item 402 information.\204\ In the context of Item 402, a subsidiary of a registrant is an affiliate controlled by the registrant directly or indirectly through one or more intermediaries, as set forth in the definition of ``subsidiary'' under both Securities Act Rule 405 and Exchange Act Rule 12b-2. Therefore, the proposal would cover an employee if he or she was employed by the registrant or a subsidiary of the registrant as defined in Rule 405 and Rule 12b-2.

        ---------------------------------------------------------------------------

        \204\ See Item 402(a)(2) and Instruction 2 to Item 402(a)(3).

        ---------------------------------------------------------------------------

        ii. Comments on the Proposed Rule

        The majority of commenters that discussed this issue recommended that the rule only require parent registrants to incorporate into their pay ratio disclosure the employees of their consolidated subsidiaries.\205\ One of the commenters claimed that there would be a 91% increase in compliance costs if the final rule included minority-

        owned subsidiaries and joint ventures because registrants would otherwise be required to ``engage in an extensive information gathering process'' without ``access to the payroll and human resources information needed for the pay ratio from subsidiaries or other entities with a more tenuous connection, such as joint ventures.'' \206\ Another commenter claimed that registrants do not exercise much influence on the compensation policies and practices of entities in which they have only a minority or nominal interest.\207\ Only one commenter asserted that the final rule should exclude compensation information from all subsidiaries and be limited to only the compensation of employees directly employed by the registrant.\208\

        ---------------------------------------------------------------------------

        \205\ See, e.g., letters from ABA (limiting the requirement only to employees of the registrant's wholly-owned or majority-owned subsidiaries with consolidated financial statements, but not subsidiaries that are portfolio companies of business development companies), Best Buy et al., Business Roundtable I, COEC I, COEC II, Corporate Secretaries, CT State Treasurer, Davis Polk, Eaton, ExxonMobil, Mercer I, Meridian, NACCO, NAM I, and NAM II.

        \206\ See letter from COEC I.

        \207\ See letter from ABA.

        \208\ See letter from WorldatWork I.

        ---------------------------------------------------------------------------

        Some commenters suggested that the final rule require registrants to incorporate only employees of their wholly-owned subsidiaries and not employees of their joint ventures.\209\ Other commenters stated that the final rule should allow a subsidiary to exclude its pay ratio disclosure in its filings if the subsidiary's employees are incorporated into its parent registrant's

        Page 50127

        pay ratio disclosure.\210\ One commenter asserted that the final rule should require a registrant to include the compensation information of its subsidiary only if the issuer has ``control'' over the subsidiary (as ``control'' is defined in our rules).\211\ Another commenter maintained that the final rule should require a registrant to include its subsidiary only if the registrant has ``actual control'' over the compensation decisions made at the subsidiary level.\212\ Finally, a few commenters contended that the final rule should require registrants to include the employees of their subsidiaries in their pay ratio generally without specifying the types of subsidiaries.\213\

        ---------------------------------------------------------------------------

        \209\ See, e.g., letters from AAFA I, Business Roundtable I, Corporate Secretaries, NACCO, NAM I, NAM II, and PM&P.

        \210\ See, e.g., letters from FSI and FSR.

        \211\ See letter from Capital Strategies.

        \212\ See letter from Cummins Inc.

        \213\ See, e.g., letters from AFL-CIO I, Domini, and US SIF.

        ---------------------------------------------------------------------------

        iii. Final Rule

        After considering these comments, we are revising the final rule from the proposal. Unlike the proposed rule, the final rule defines ``employee'' to include only the employees of the registrant and its consolidated subsidiaries rather than employees of subsidiaries that were affiliates it controlled directly or indirectly through one or more intermediaries, as set forth in the definition of ``subsidiary'' under both Securities Act Rule 405 and Exchange Act Rule 12b-2. This change should reduce costs and burdens for registrants, while maintaining the benefits of the pay ratio rule, as discussed below.

        Rule 12b-2 and Rule 405 define a ``subsidiary'' as ``an affiliate controlled by an entity directly, or indirectly through one or more intermediaries,'' while an ``affiliate'' is defined as ``a person that directly, or indirectly through one or more intermediaries, controls or is controlled by, or is under common control with, the person specified.'' \214\ The term ``control'' (including the terms ``controlling,'' ``controlled by'' and ``under common control with'') means the possession, direct or indirect, of the power to direct or cause the direction of the management and policies of a person, whether through the ownership of voting securities, by contract, or otherwise.\215\ One commenter described this definition of ``subsidiary'' as ``very expansive'' because it includes not just affiliates controlled directly by the registrant, but also those controlled indirectly by the registrant through one or more intermediaries or that are under common control with the registrant.\216\ In the Section 16 context, we have noted that many practitioners believe that individuals or entities holding as little as 10% or more of the voting equity securities of a registrant may likely be considered an affiliate or control person.\217\ Further, whether an affiliate is controlled by an entity is based on the facts and circumstances of each situation, so whether a company should be considered a ``subsidiary'' of a registrant is not always clear.\218\ Therefore, depending on the specific facts and circumstances of the situation, if the rule used the term ``subsidiary'' as defined under Rules 12b-2 and 405, a registrant could potentially be required to include the employees of a company in which it holds as little as a 10% ownership stake. As commenters noted, obtaining compensation and payroll data from unconsolidated entities could be costly, burdensome, or potentially impossible.\219\ Additionally, one commenter suggested that, because compensation disclosure is designed to facilitate the comparison of the PEO's pay to the performance of the company based on its consolidated financial statements, the pay ratio should relate to the same consolidated financial performance of the company and not to non-consolidated entities and other factors if the purpose of the rule is to enhance compensation disclosure.\220\

        ---------------------------------------------------------------------------

        \214\ 17 CFR 240.12b-2 and 17 CFR 230.405.

        \215\ Id.

        \216\ See letter from ABA.

        \217\ See Revision of Rule 144, Rule 145 and Form 144, Release No. 33-7391, Section III.B (Jun. 27, 1995) (``Under the proposal to make revisions to Rule 144, 145, and Form 144, the same criteria used to determine those persons that are not `insiders' under Exchange Act Section 16 would be used for Rule 144. Many practitioners already used Section 16 criteria as a guide. The Commission believes it is likely that most persons who are not officers, directors, or 10% holders are not in a `control' position.'').

        \218\ See, e.g., letters from ABA, Best Buy et al., Business Roundtable I, COEC I, and Davis Polk.

        \219\ See, e.g., letters from ABA, Best Buy et al., Business Roundtable I, COEC I, Corporate Secretaries, Davis Polk, Eaton, ExxonMobil, NACCO, and NAM I.

        \220\ See letter from NAM II.

        ---------------------------------------------------------------------------

        In contrast, defining a ``subsidiary'' based on whether a registrant consolidates a company in its financial statements likely will decrease the costs and burdens on a registrant compared with the proposal because most registrants consolidate based on their ownership of over 50% of the outstanding voting shares of their subsidiaries and more guidance is readily available on when consolidating subsidiaries is appropriate than when an entity should be considered a ``subsidiary'' based on the concept of control. For example, the United States generally accepted accounting principles (``U.S. GAAP'') traditionally has required a company holding the ``controlling financial interest'' of another company to consolidate that company.\221\ The usual condition for consolidation is a controlling financial interest through majority ownership of over 50% of the outstanding voting shares.\222\ Determining whether a company is a ``subsidiary'' under the consolidated financial statement method, as opposed to the using the definition of ``subsidiary'' under Rules 405 and 12b-2, generally will provide a higher quantitative threshold and thus a smaller pool of employees to include in the median employee determination, which should help to reduce costs associated with making such a determination. Overall, the standard for consolidation and the definition of ``control'' under Rules 405 and 12b-2 are both driven by very similar concepts of control. Use of a consolidated subsidiary standard will typically exclude employees of entities where a company holds between a 10% to 50% voting interest in such entity.

        ---------------------------------------------------------------------------

        \221\ See Volume I, Section 1160.1.12 of PWC 2010 Accounting and Reporting Manual.

        \222\ See ASC 810-10-15-8 (``The usual condition for a controlling financial interest is ownership of a majority voting interest, and, therefore, as a general rule ownership by one reporting entity, directly or indirectly, of more than 50 percent of the outstanding voting shares of another entity is a condition pointing toward consolidation. The power to control may also exist with a lesser percentage of ownership, for example, by contract, lease, agreement with other stockholders, or by court decree.'').

        ---------------------------------------------------------------------------

        Although this change from the proposal generally will result in a smaller pool of employees being used for the median employee determination, we do not believe it will undermine the usefulness of the required disclosures or conflict with the purposes of Section 953(b). As one commenter indicated, ``registrants do not exercise much, if any, influence on the compensation policies and practices of entities in which they have only a minority or nominal interest (unless the employees of such entities provide services directly to the registrant).'' \223\ According to the commenter, limiting the final rule to employees of a registrant's consolidated subsidiaries, therefore, ``would not deprive investors of useful information or important insights into a registrant's compensation structure.'' We believe that requiring registrants to consider only their employees and the employees of their consolidated subsidiaries in identifying their median employee should not limit the usefulness of the pay ratio disclosure as a data point for shareholders to use in making their

        Page 50128

        voting decisions on executive compensation under Section 951 of the Dodd-Frank Act in any significant way.

        ---------------------------------------------------------------------------

        \223\ See letter from ABA.

        ---------------------------------------------------------------------------

        e. Any PEO Compensation in the Last Full Fiscal Year

        i. Proposed Rule

        The Proposing Release did not discuss the compensation information that would be required if one or more of a registrant's PEOs served only part of a fiscal year.

        ii. Comments on the Proposed Rule

        Some commenters suggested that a registrant should be required to disclose the compensation information only for the PEO holding the position at the end of the last completed fiscal year.\224\ One commenter suggested an exception where the PEO has served for most of a fiscal year but departs before the last day, in which case the commenter recommended that we require compensation information disclosure for the person who served as PEO for the majority of the fiscal year.\225\ Another commenter suggested the use of the PEO as of the last day of the most recently completed fiscal year and noted that the issue of that person's annual total compensation representing less than a full year's compensation ``could be addressed by requiring registrants to annualize the annual total compensation for the chief executive officer serving at fiscal year end.'' \226\

        ---------------------------------------------------------------------------

        \224\ See, e.g., letters from ABA, Corporate Secretaries, Davis Polk, McGuireWoods, and PM&P.

        \225\ See letter from ABA.

        \226\ See letter from Davis Polk.

        ---------------------------------------------------------------------------

        iii. Final Rule

        The final rule allows a registrant a choice of two options in calculating the annual total compensation for its PEO in situations in which the registrant replaces its PEO with another PEO during its fiscal year. In these situations, the registrant must disclose which option it chose and how it calculated its PEO's annual total compensation. First, a registrant may take the total compensation calculated pursuant to Item 402(c)(2)(x), and reflected in the Summary Compensation Table, provided to each person who served as PEO during the year and combine those figures. This figure would constitute the registrant's annual total PEO compensation.

        Alternatively, a registrant may look to the PEO serving in that position on the date it selects to identify the median employee and annualize that PEO's compensation. For example, if the registrant chooses October 15 as the date to determine its median employee, the registrant would calculate the compensation of the person serving as PEO on that date and annualize that PEO's compensation. If the person was PEO for six months and received $100,000 of total compensation, the registrant would use $200,000 as the annual total compensation of its PEO. This approach is consistent with annualizing the total compensation of permanent employees, discussed below, which is permitted under the final rule. It is also similar to the approach suggested by one commenter.\227\

        ---------------------------------------------------------------------------

        \227\ See id.

        ---------------------------------------------------------------------------

        We are not adopting the approach advocated by some commenters of disclosing compensation information for the PEO holding that position at the end of the fiscal year. Section 953(b) requires the disclosure of the ``annual total compensation of the chief executive officer (or any equivalent position) of the issuer,'' and we think the better interpretation of that language is that it is intended to capture the annual compensation paid for that position (regardless of the individual who holds the position). Also, we believe that allowing the disclosure of only partial year compensation would fundamentally alter the pay ratio disclosure and would not capture the ratio of PEO compensation to median employee compensation.

        f. Additional Information Is Permissible

        i. Proposed Rule

        In the Proposing Release, we noted that we received some comments prior to the proposal suggesting that the rule should allow registrants to present separate pay ratios covering U.S. and non-U.S. employees to mitigate concerns that the comparison of the PEO to non-U.S. employees could distort the disclosure.\228\

        ---------------------------------------------------------------------------

        \228\ See, e.g., letters from AFL-CIO (Dec. 13, 2010) (``AFL-CIO pre-proposal letter''), Americans for Financial Reform (Mar. 23, 2011) (``AFR pre-proposal letter''), Walden Asset Management (Apr. 29, 2011) (``Walden pre-proposal letter''), and Social Investment Forum (Apr. 21, 2011) (``SIF pre-proposal letter'').

        ---------------------------------------------------------------------------

        The Proposing Release stated that we did not believe that it was necessary to include instructions in the new rule expressly permitting registrants to add disclosure to accompany the pay ratio. We indicated, that, as with other mandated disclosure under our rules, registrants would be permitted to supplement their required disclosure with a narrative discussion or additional ratios if they chose to do so. We indicated also that, as with other disclosure under our rules,\229\ any additional ratios should not be presented with greater prominence than the required pay ratio. We requested comment on whether the final rule should permit or require registrants to include two separate pay ratios covering U.S. employees and non-U.S. employees.

        ---------------------------------------------------------------------------

        \229\ See, e.g., Exchange Act Rule 12b-20 and Commission Guidance Regarding Management's Discussion and Analysis or Financial Condition and Results of Operations, Release No. 33-8350 (Dec. 19, 2003) 68 FR 75056, at 75060.

        ---------------------------------------------------------------------------

        ii. Comments on the Proposed Rule

        Some commenters opposed any requirement for two separate ratios.\230\ Other commenters also opposed a requirement for two separate ratios, but indicated that the final rule should permit registrants to provide separate ratios if they chose to do so.\231\ One commenter acknowledged that the Proposing Release permitted additional ratios but suggested that we should indicate that any additional ratios should not be misleading and not presented with greater prominence than the required ratio.\232\

        ---------------------------------------------------------------------------

        \230\ See, e.g., letters from American Benefits Council, Aon Hewitt, Frederic W. Cook & Co., and WorldatWork I.

        \231\ See, e.g., letters from ABA, Capital Strategies, CII, COEC I, CT State Treasurer, Cummings Foundation, Emergent, Hyster-Yale, Johnson & Johnson, NACCO, Comptroller of the City of New York (Nov. 27, 2013) (``NYC Comptroller''), US SIF, and Vivient.

        \232\ See letter from E&Y.

        ---------------------------------------------------------------------------

        iii. Final Rule

        After considering the comments, we have added an instruction to the final rule stating expressly that registrants may present additional ratios or other information to supplement the required ratio, but are not required to do so. The instruction states also that, if a registrant includes any additional ratios, the ratios must be clearly identified, not misleading, and not presented with greater prominence than the required ratio. Additional pay ratios are not limited to any particular information, such as pay ratios covering U.S. and non-U.S. employees.

        g. Annualizing Permanent Employees is Permissible, but Other Compensation Adjustments are Prohibited

        i. Proposed Rule

        The proposed rule included an instruction permitting, but not requiring, registrants to annualize the total compensation for permanent ``employees'' who did not work for the entire year, such as new hires, employees on leave under the Family and Medical Leave Act of 1993,\233\ employees called for active military duty, or employees who took an unpaid

        Page 50129

        leave of absence during the period for another reason. We did not propose to require registrants to perform this type of adjustment, however, because we did not believe that the costs of requiring companies to make this extra calculation would be justified. The proposed rule applied to individuals who were employed on the last day of the fiscal year because it referred specifically to an ``employee.''

        ---------------------------------------------------------------------------

        \233\ 29 U.S.C. 2601 et seq.

        ---------------------------------------------------------------------------

        In addition, the proposed instruction would prohibit a registrant from annualizing some eligible employees and not others, and the instruction prohibited adjustments that would cause the ratio to not reflect the actual composition of the workforce, such as annualizing the compensation of seasonal or temporary employees. A registrant could annualize the compensation for a permanent part-time employee who had only worked a portion of the year (such as an employee who is permanently employed for three days a week and who took an unpaid leave of absence under the Family and Medical Leave Act for a portion of the year). In such a case, we explained that the adjustment should reflect compensation for the employee's part-time schedule over the entire year, but should not adjust the part-time schedule to a full-time equivalent schedule.

        Although we proposed to permit the annualizing adjustments described above, the proposed rule would not have permitted certain other adjustments or assumptions, such as full-time equivalent adjustments for part-time employees or annualizing adjustments for temporary or seasonal employees. We believed such adjustments would cause the median to not be reasonably representative of the registrant's actual employment and compensation arrangements for its workforce during the period and could diminish the potential usefulness of the disclosure.

        ii. Comments on the Proposed Rule

        Some commenters concurred that registrants should be permitted to annualize the total compensation for all permanent employees (which would exclude temporary and seasonal positions) that were employed by the registrant for less than the full fiscal year.\234\ Some commenters also agreed that registrants should not be permitted to adjust the salaries of part-time, temporary, and seasonal employees to the equivalent of full-time status contending that it would be misleading to do so.\235\

        ---------------------------------------------------------------------------

        \234\ See, e.g., letters from ABA, Intel, NACCO, PM&P, SH&P, and WorldatWork I.

        \235\ See, e.g., letters from Domini, Capital Strategies, and CT State Treasurer (stating that substantial reliance on non-full-time employees is important for shareholders' understanding of a company's workforce and pay structure).

        ---------------------------------------------------------------------------

        Many commenters, however, contended that the final rule should permit registrants to provide full-time equivalent adjustments for the salaries of part-time, temporary, and seasonal employees.\236\ Some of these commenters asserted that not doing so would distort the pay ratio.\237\ One of these commenters asserted that permitting full-time equivalent adjustments would not undermine disclosure or make it less accurate because any potential concern about shareholders' understanding of the pay ratio would be mitigated by requiring registrants to disclose their full-time equivalency and the approximate number of part-time, seasonal, and temporary employees for which it made the calculation.\238\ Another commenter indicated that, if the final rule requires disclosure of any adjustment calculations, the disclosure should be limited to brief statements.\239\

        ---------------------------------------------------------------------------

        \236\ See, e.g., letters from ABA, ASA, Business Roundtable I, Chesapeake Utilities, COEC I, COEC II, Dover Corp., ExxonMobil, Fedewa, FEI, Frederic W. Cook & Co., Johnson & Johnson, LAPFF, McCoy, PM&P, SH&P, SHRM, TCA, and Vectren Corp.

        \237\ See, e.g., letters from ABA, ASA, Chesapeake Utilities, COEC I, and COEC II.

        \238\ See letter from ABA.

        \239\ See letter from Corporate Secretaries.

        ---------------------------------------------------------------------------

        iii. Final Rule

        After considering the comments, we are adopting the final rule as proposed. Annualization and full-time equivalent adjustments are separate concepts. Annualization involves taking the compensation of an employee who worked for only part of the registrant's fiscal year and projecting that compensation as if the employee worked the full fiscal year at the schedule that the employee worked for the portion of the year the employee worked. Annualization is allowed under the rule for full-time and part-time employees who did not work for the registrant's full fiscal year for some reason, such as they were employees who were newly hired, on leave under the Family and Medical Leave Act of 1993, called for active military duty, or took an unpaid leave of absence during the period. Annualization is only allowed for permanent employees; it is not allowed under the final rule for seasonal or temporary employees. A full-time equivalent adjustment involves taking the compensation of a part-time employee and projecting what the employee would have made if the employee were employed on a full-time basis. Full-time equivalent adjustments are prohibited under the final rule under all circumstances.

        We are taking this approach in the final rule because we believe it most accurately captures the workforce and compensation practices that the registrant has chosen to employ. The limited ability to annualize a permanent full-time or part-time employee reflects the fact that these employees are a permanent part of the registrant's workforce despite having only worked for part of that particular year. In contrast, a temporary or seasonal employee is not a permanent part of the registrant's workforce. Full-time equivalent adjustments of these employees' compensation would reflect a different workforce composition and compensation structure than used by the registrant. To the extent a registrant believes that not making full-time equivalent adjustments for temporary or seasonal employees might not provide shareholders a complete understanding of the registrant's compensation practices as they exercise their say-on-pay votes, the registrant is permitted under the final rule to provide additional disclosure.

      2. Identifying the Median Employee and Calculating Annual Total Compensation

        a. Identifying the Median Employee

        i. Once Every Three Years

        (a) Proposed Rule

        The proposed rule required registrants to disclose the ``median of the annual total compensation of all employees of the registrant.'' \240\ The proposed rule defined ``annual total compensation'' to mean ``total compensation for the registrant's last completed fiscal year,'' \241\ and ``employee'' to mean ``an individual employed by the registrant or any of its subsidiaries as of the last day of the registrant's last completed fiscal year.'' \242\ Therefore, the proposed rule suggested that registrants would need to undertake the full process of identifying anew the median employee for each completed fiscal year.

        ---------------------------------------------------------------------------

        \240\ See proposed Item 402(u)(1)(i) of Regulation S-K.

        \241\ See proposed Item 402(u)(2)(ii) of Regulation S-K.

        \242\ See proposed Item 402(u)(3) of Regulation S-K.

        ---------------------------------------------------------------------------

        (b) Comments on the Proposed Rule

        A few commenters suggested that the final rule should allow registrants to undertake the full process of identifying the median employee periodically (such as once every three years) and use the

        Page 50130

        compensation of that same or a similarly situated employee during the intervening two years as long as no material or substantial changes to the workforce had occurred.\243\ These commenters indicated that, to remain consistent with Section 953(b), a registrant should be required to undertake the process to re-identify the median employee for any year in which it has experienced a change in its employee population which the registrant reasonably believes would result in a significant change in the pay ratio disclosure. One of these commenters suggested that the registrant should exercise ``requisite diligence'' annually to determine whether such a material change had occurred.\244\

        ---------------------------------------------------------------------------

        \243\ See, e.g., letters from ABA, COEC I, and COEC II.

        \244\ See letter from ABA.

        ---------------------------------------------------------------------------

        (c) Final Rule

        The final rule allows a registrant to identify the median employee whose compensation will be used for the annual total compensation calculation once every three years unless there has been a change in its employee population or employee compensation arrangements that it reasonably believes would result in a significant change in the pay ratio disclosure. If there have been no changes that the registrant reasonably believes would significantly affect its pay ratio disclosure, the registrant must disclose that it is using the same median employee in its pay ratio calculation and describe briefly the basis for its reasonable belief. For example, the registrant could disclose that there has been no change in its employee population or employee compensation arrangements that it believes would significantly affect the pay ratio disclosure. If the registrant is using the same median employee, it must calculate that median employee's annual total compensation each year and use that figure to update its pay ratio disclosure each year.

        For example, the registrant is required to identify the median employee and calculate that median employee's annual total compensation in year one. In years two and three, however, the registrant may use that same median employee (or an employee whose compensation is substantially similar to the original median employee based on the compensation measure used to select that median employee, as discussed below) to re-calculate the annual total compensation for that employee without re-identifying the median employee as would otherwise be required under the final rule if it satisfies the above conditions.

        We believe this approach is appropriate because, as commenters noted, it is consistent with the requirements of Section 953(b) to provide annual pay ratio disclosure, while at the same time reducing registrants' costs and burdens of re-calculating the median employee more than once every three years unless there has been a change in the registrant's employee population or employee compensation arrangements that it reasonably believes would result in a significant change in the pay ratio disclosure.\245\ Also, we note that the final rule permits reasonable estimates in identifying the median employee. Permitting registrants to identify the median employee once every three years, absent a change in the registrant's employee population or employee compensation arrangements that it reasonably believes would result in a significant change in its pay ratio disclosure, is another way of allowing an estimate of the median employee in a situation where it is unlikely to result in a significant change to the pay ratio disclosure. Therefore, this provision will help to minimize burdens and costs while not significantly affecting registrants' pay ratios. We do not believe that allowing this flexibility will limit the usefulness of the pay ratio for shareholders as a data point in making their voting decisions on executive compensation under Section 951 of the Dodd-Frank Act.

        ---------------------------------------------------------------------------

        \245\ See, e.g., letters from ABA and COEC I.

        ---------------------------------------------------------------------------

        Further, we note that allowing a registrant in appropriate circumstances to go up to three years without engaging in the full process to re-identify the median employee is consistent with the outer limit established by Section 951 for a registrant's say-on-pay vote. In our view, just as registrants must have a new say-on-pay vote three years after the previous vote, we think it is reasonable and appropriate that they engage in the process of re-identifying the median employee no less frequently than every three years.

        Also, there may be situations in which there has been no change in a registrant's employee population or employee compensation arrangements but it is no longer appropriate for the registrant to use the median employee identified in year one as the median employee in years two or three because of a change in that employee's circumstances. In such a situation, the registrant may use another employee whose compensation is substantially similar to the original median employee based on the compensation measure used to select the median employee in year one. For example, if the median employee identified in year one is no longer employed by the registrant in years two or three or that employee's compensation significantly changed in years two and three (for example, a promotion that significantly increased his or her compensation), the registrant is permitted to identify its median employee in each of the following two years from among employees that had similar compensation to the median employee in year one. If no other employee has similar compensation, however, the registrant must re-identify the median employee as required under the final rule.

        ii. Using Annual Total Compensation, Another Consistently Applied Compensation Measure, Statistical Sampling, Reasonable Estimates, or Other Reasonable Methods

        (a) Proposed Rule

        In the Proposing Release, we noted that Congress specifically chose the ``median'' as the point of comparison for Section 953(b), rather than the average or some other measure. Therefore, the proposed rule required the median. Section 953(b) does not prescribe a methodology that must be used to identify the median. To allow the greatest degree of flexibility while remaining consistent with the statutory provision, the proposed rule did not specify any required calculation methodologies for identifying the median. Instead, we provided instructions and guidance designed to allow registrants to choose from several alternative methods to identify the median, so that they would be able to use the method that worked best for their own facts and circumstances.

        For instance, registrants would be able to provide the proposed disclosure using total compensation for each employee under Item 402(c)(2)(x), statistical sampling, reasonable estimates, or the use of any consistently applied compensation measures to identify the median. Once the registrant identified the median employee based on the selected compensation measure applied to each employee in the sample, the registrant would calculate that employee's annual total compensation in accordance with Item 402(c)(2)(x) and disclose that amount as part of the pay ratio disclosure. The proposal did not prescribe what a reasonable estimate would entail because that would necessarily depend on the registrant's particular facts and circumstances. In addition, the proposed rule did not

        Page 50131

        prescribe specific estimation techniques or confidence levels for identifying the median employee because we believed that companies would be in the best position to determine what is reasonable in light of their own employee population and access to compensation data. We proposed to require registrants to briefly describe and consistently use the methodology and any material assumptions, adjustments, or estimates applied to identify the median employee, and for any estimated amounts to be clearly identified as such.

        We proposed this flexible approach because we believed that the most appropriate and cost effective methodology would necessarily depend on a registrant's particular facts and circumstances, including, among others, such variables as: The size and nature of the registrant's workforce; the complexity of its organization; the stratification of pay levels across the workforce; the types of compensation paid to employees; the extent that different currencies are involved; the number of tax and accounting regimes involved; and the number of payroll systems the registrant has and the degree of difficulty involved in integrating payroll systems to readily compile total compensation information for all employees. We believed that these likely are the same factors that would cause substantial variation in the costs of compliance.

        In the Proposing Release, we stated our belief that the proposed rule's flexibility would enable registrants to manage compliance costs more effectively. We also stated that, by allowing registrants to better manage costs, a flexible approach could mitigate, to some extent, any potential negative effects on competition arising from the mandated requirements. We recognized, however, that a flexible approach could increase uncertainty for registrants that would prefer more specificity on how to comply with the proposed rule, particularly for those registrants that do not use statistical analysis in the ordinary course of managing their businesses.

        In the Proposing Release, we offered guidance on two permissible methodologies under the proposal: (1) Statistical sampling and (2) use of a consistently applied compensation measure. The variance of underlying compensation distributions (that is, how widely employee compensation is spread out or distributed around the mean) could appreciably affect the sample size needed for reasonable statistical sampling. We conducted an analysis about sample size that we described in the Proposing Release. Our analysis used mean and median wage estimates from the U.S. Department of Labor's Bureau of Labor Statistics (``BLS'') at the 4-digit North American Industry Classification System (``NAICS'') industry level (290 industries) and assumed a lognormal wage distribution, a 95% confidence interval with 0.5% margin of error. The analysis focused on the registrants that have a single business or geographical unit. The analysis also assumed that the sampling method would be a true random sampling because it would not be biased by region, occupation, rank, or other factor. In our analysis, the appropriate sample size for the registrants with a single business or geographical unit varied between 81 and 1,065 across industries, with the average estimated sample size close to 560.

        We acknowledged, however, that variation in the types of employees at a registrant across business units and geographical regions would add complexity to the sampling procedure. While we generally agreed that a relatively small sample size would be appropriate in some situations, a reasonable determination of sample size would ultimately depend on the underlying distribution of compensation data. We noted that reasonable estimates of the median for registrants with multiple business lines or geographical units could be arrived at through more than one statistical sampling approach. All approaches, however, would require drawing observations from each business or geographical unit with a reasonable assumption on each unit's compensation distribution and inferring the registrant's overall median based on the observations drawn. Certain cases may not easily generate confidence intervals around the estimates or prescribe the appropriate minimum sample size. As a result, compliance costs would vary across registrants according to the characteristics of their compensation distributions. Nevertheless, we concluded that permitting registrants to use statistical sampling could lead to a reduction in compliance costs as compared with other methods of identifying the median.

        Additionally, we noted that the identification of a median employee would not necessarily require a determination of exact compensation amounts for every employee included in the sample. A registrant could, rather than calculating exact compensation, identify the employees in the sample that have extremely low or extremely high pay that would fall completely on either end of the pay spectrum. Since identifying the median involves finding the employee in the middle, it might not be necessary to determine the exact compensation amounts for every employee paid more or less than the employee in the middle.

        In addition to statistical sampling, the Proposing Release also highlighted the use of a consistently applied compensation measure. We recognized concerns about expected compliance costs arising from the complexity of the ``total compensation'' calculation under Item 402(c)(2)(x) and, in particular, the determination of total compensation in accordance with Item 402(c)(2)(x) for employees when identifying the median. To address these concerns, the proposed rule would allow companies to use total direct compensation (such as annual salary, hourly wages, and any other performance-based pay) or cash compensation to first identify a median employee and then calculate that median employee's annual total compensation in accordance with Item 402(c)(2)(x). We noted that this approach would provide a workable identification of the median for many registrants, and we expected that the costs of compliance would be reduced if registrants were permitted to identify the median using a less complex, more readily available figure, such as salary and wages, rather than total compensation as determined in accordance with Item 402(c)(2)(x).

        This approach could also reduce costs for registrants that are unable to use statistical sampling techniques, as registrants would be permitted to use a consistently-applied compensation measure to identify the median employee regardless of whether they use statistical sampling. Further, because of concerns that using cash compensation could be just as burdensome to calculate for registrants with multiple payroll systems in various countries, we did not propose to require companies to use a specific compensation measure like cash compensation or total direct compensation when they were identifying the median employee. Instead, we believed that registrants would be in the best position to select a compensation measure that was appropriate to their own facts and circumstances and that a consistently applied compensation measure would result in a reasonable estimate of a median employee at a substantially reduced cost. Therefore, the proposed rule permitted a registrant to identify a median employee based on any consistently applied compensation measure, such as information derived from its tax and/or payroll records, as long as the registrant briefly disclosed the measure that it used.

        Page 50132

        We also recognized that the annual period used for payroll and/or tax recordkeeping could sometimes differ from the registrant's fiscal year. For purposes of calculating the annual total compensation amounts when using a consistently applied compensation measure, the proposed rule permitted registrants to use the same annual period that was used in the payroll and/or tax records from which the compensation amounts were derived. We did not propose to define or limit what would qualify as payroll and/or tax records. We noted, however, that the proposed accommodation was intended to be construed broadly enough to allow registrants to use information that they already tracked and compiled for payroll and/or tax purposes. We thought that permitting registrants to use compensation information in the form that it was maintained in their own books and records would reduce compliance costs without appreciably affecting the quality of the disclosure.

        Additionally, although our proposed flexible approach could reduce comparability of the pay ratio disclosure across registrants, we stated our belief that precise conformity or comparability of the ratio across companies was not necessary and indicated that a possible benefit of the pay ratio disclosure would be providing a company-specific metric that shareholders could use to evaluate the PEO's compensation within the context of his or her own company. Accordingly, we did not believe that improving the comparability of the disclosure across companies by mandating a specific method for identifying the median would be justified in light of the costs that would be imposed on registrants by a more prescriptive rule.

        Also, even assuming the benefits of comparability across registrants as a desirable goal, we did not believe that mandating a particular methodology would necessarily improve comparability because of the numerous other factors that could also cause the ratios to be less meaningful for company-to-company comparison, such as differences in industry and business type; variations in the way companies organize their workforces to accomplish similar tasks; differences in the geographical distribution of employees (domestic or international, as well as in high- or low-cost areas); degree of vertical integration; reliance on contract and outsourced workers; and ownership structure. We also note that some commenters asserted that disclosing the pay ratio could potentially increase the likelihood that a registrant's competitors could infer proprietary or sensitive information about the registrant's business, which could cause a competitive disadvantage for registrants.\246\

        ---------------------------------------------------------------------------

        \246\ See, e.g., letters from Huan Lou (Dec. 2, 2013) (``Lou'') and Prof. Ray.

        ---------------------------------------------------------------------------

        Finally, we recognized that allowing registrants to select a methodology for identifying the median, including identifying the median employee based on any consistently applied compensation measure and allowing the use of reasonable estimates, rather than prescribing a methodology or set of methodologies, could potentially permit a registrant to alter the reported ratio to achieve a particular objective with the pay ratio disclosure, thereby potentially reducing the usefulness of the information. But, as we explained, we believed that requiring the use of a consistently applied compensation measure should lessen this concern.

        (b) Comments on the Proposed Rule

        (i) Flexibility

        A large number of commenters indicated that they supported the flexibility permitted in the proposed rule generally, or more specifically supported the flexibility of the proposed rule in permitting registrants to choose a methodology for calculating the median.\247\ Some commenters contended that the flexibility would lessen the costs and burdens of the proposed rule without reducing the rule's benefits.\248\ A few commenters suggested that the proposed rule's flexibility would not undermine the rule and would be consistent with the directives of Section 953(b).\249\ Other commenters indicated that, although the proposed rule's flexibility might alter pay ratio disclosures, any distortion would be minimal.\250\ One commenter claimed that the proposed rule was not flexible enough because Instruction 2(iii) to Item 402(u) would permit a registrant to identify the median employee only using ``compensation'' measures, whereas the commenter advocated for use of a ``job-level'' measure.\251\ Despite the permitted flexibility, another commenter remained concerned about the complexity of finding the median employee.\252\

        ---------------------------------------------------------------------------

        \247\ See, e.g., letters from Prof. Angel, Aon Hewitt, Bacirctirente et al., Best Buy et al., CalSTRS, Capital Strategies, CEG, Chicago Teachers Fund, Corporate Secretaries, Cummings Foundation, Cummins Inc., CUPE, Davis Polk, E&Y, First Affirmative, Freeport-McMoRan, FS FTQ, ICCR, IL Bricklayers and Craftworkers Union, Intel, Johnson & Johnson, Marco Consulting, McMorgan & Co., Mercer I, Meridian, Microsoft, Novara Tesija, NSFM, NY Bricklayers and Craftworkers Union, NYC Bar, OCP, Oxfam, PGGM, Public Citizen I, Quintave, Rep. Ellison et al. II, SEIU, Sen. Menendez et al. II, Socially Responsive Financial Advisors, Teamsters, Trillium I, Trustee Campbell, UAW Trust, Vectren Corp., WA State Investment Board, and WorldatWork I.

        \248\ See, e.g., letters from ABA, AFSCME, Barnard, Bricklayers International, CII, LIUNA, Fowler, Mirczak, PNC Financial Services, Sen. Menendez et al. II, US SIF, Vivient, and Walden.

        \249\ See, e.g., letters from Domini and PM&P.

        \250\ See, e.g., letters from Capital Strategies and WorldatWork I.

        \251\ See letter from Towers Watson. According to the commenter, in a job-level approach, a global company assigns all jobs in its organization to one of X number of job levels (determined by how the particular job role contributes to the organization based on tasks, skills, expertise, leadership, functional strategy and business strategy). The company would readily be able to determine the job or job family that would fall at the median of skill levels within the company based on a ratio that compares the total number of employees at each job level to the total employee population. The job or job family in the median of skill levels would be where the median-level employee resides and, once it is determined that the median employee resides at a particular level, the company would then be able to apply a statistical sampling approach to that job level, taking into account compensation earned in different locations or countries, to further reduce the number of payroll files that will need to be examined. The commenter indicated that more details of this approach can be found at: http://www.towerswatson.com/en/Services/Tools/job-leveling-global-grading-and-career-map.

        \252\ See letter from Vectren Corp.

        ---------------------------------------------------------------------------

        One commenter contended that a registrant should be permitted to develop its own methodology for identifying their median employee to mitigate costs \253\ if the rule required the registrant to accurately describe its methodology, perform consistent calculations each year, disclose when and how it chose to deviate from the prior year's methodology, select the methodology in good faith, and make reasonable assumptions, adjustments, and estimates. Some commenters recommended that, once a registrant chooses its methodology for identifying the median employee, the registrant should be required to use that methodology going forward.\254\ Other commenters supported the proposed rule's flexibility in allowing registrants to choose a methodology for identifying the median employee, provided that registrants are required to disclose the methodologies they used.\255\ One commenter suggested that the final rule provide some type of check generally on a registrant's methodology.\256\

        ---------------------------------------------------------------------------

        \253\ See letter from ABA.

        \254\ See, e.g., letters from PGGM and RPMI.

        \255\ See, e.g., letters from Australian Council of Superannuation Investors (Dec. 2, 2013) (``ACSI'') and IPS.

        \256\ See letter from Mirczak.

        ---------------------------------------------------------------------------

        Some commenters were opposed to the proposed rule's flexibility. Most of these commenters were individuals who claimed that the proposed rule's flexibility would allow registrants to reduce the ratio in inappropriate

        Page 50133

        ways.\257\ Other commenters contended that the permitted flexibility would decrease the ratio's utility, especially for comparing the ratios of different companies.\258\ A few commenters maintained that the proposed rule would still lead to high costs, even taking into account its flexibility.\259\

        ---------------------------------------------------------------------------

        \257\ See, e.g., letters from Bupp, Corayer, Fedewa, Fox, Friend, Grotzke, Hlodnicki, Kizzort, Maly, Petricoin, and Van Pelt.

        \258\ See, e.g., letters from Amundi, BCIMC, Ciatto, Glenn, IBC, Prof. Muth, and NIRI.

        \259\ See, e.g., letters from ABA (stating that registrants will still incur significant costs even with the ability to select a methodology) and FSR.

        ---------------------------------------------------------------------------

        Some commenters recommended that the final rule include a safe harbor for identifying the median employee to minimize burdens, provide greater comparability, and limit liability.\260\ One commenter encouraged us to retain the proposed rule's flexibility in allowing a registrant to remove some percentage of the distribution from both the high and low ends of an employee sample because this would not distort the median employee determination while reducing costs.\261\

        ---------------------------------------------------------------------------

        \260\ See, e.g., letters from NSFM, OCP, PM&P, Quintave, and SHRM.

        \261\ See letter from E&Y.

        ---------------------------------------------------------------------------

        Only a few commenters commented specifically on using reasonable estimates for identifying the median employee. One commenter declared that the final rule should not permit any estimates at all.\262\ Other commenters contended that the final rule should permit reasonable estimates.\263\ One of these commenters noted that permitting reasonable estimates would mitigate the rule's costs while not ``materially'' impacting its usefulness.\264\ Also, some commenters recommended that the final rule not specify any requirements, guidance, or safe harbors regarding the estimates.\265\ Several commenters asserted that the final rule should not provide guidance regarding assumptions about error rates or confidence levels.\266\ One of these commenters expressed concern that, if the final rule provided such guidance, the assumptions would become de facto requirements.\267\

        ---------------------------------------------------------------------------

        \262\ See letter from Dennis T. (Nov. 19, 2013) (``Dennis T'').

        \263\ See, e.g., letters from ABA, Capital Strategies, Davis Polk, Hyster-Yale, Johnson & Johnson, NACCO, and WorldatWork I.

        \264\ See letter from ABA.

        \265\ See, e.g., letters from ABA (noting that further guidance is not needed because there is already sufficient deterrence for unreasonable estimates with the principles-based disclosure framework and anti-fraud provisions), Capital Strategies, Hyster-

        Yale, Johnson & Johnson, and NACCO.

        \266\ See, e.g., letters from Prof. Angel, Hyster-Yale, and NACCO.

        \267\ See letter from Prof. Angel.

        ---------------------------------------------------------------------------

        A number of commenters provided suggestions on the methodologies a registrant should be permitted to use in identifying the median employee. A few commenters stated that the final rule should provide more explicit guidance on what ``other reasonable methods'' for identifying the median employee are available.\268\ One of these commenters suggested allowing the following methods: (1) Specific safe harbor assumptions about the statistical distribution of compensation within the company and its business units; (2) formulaic, numerical, and other computational approaches to estimate the median compensation; and (3) disclosure of a reasonable range of outcomes rather than requiring the ``right'' outcome.\269\

        ---------------------------------------------------------------------------

        \268\ See, e.g., letters from E&Y and TCA.

        \269\ See letter from TCA.

        ---------------------------------------------------------------------------

        Additionally, other commenters provided specific suggestions for methods that we should consider ``reasonable'' in the final rule. One commenter requested that the final rule allow registrants to identify the median employee based on rates of pay or compensation schedules applicable to classes of employees instead of pay actually earned over the course of the year.\270\ Another commenter suggested that, where a registrant has an even number of employees and, therefore, is unable to select one median employee, the registrant should be permitted to select and disclose an average of the compensation of the two employees nearest the median.\271\ Finally, one commenter indicated that a registrant's methodology should be based on a ``good faith compliance'' standard that is akin to that type of standard in our proposed crowdfunding rules.\272\

        ---------------------------------------------------------------------------

        \270\ See letter from RILA (``In order to reduce the cost of using payroll or other data and annualizing information for employees such as new hires who are employed for less than an entire fiscal or other year, we would propose that registrants be permitted as an alternative to determine the median employee based upon employee rate of pay on the measurement date. For some issuers, identifying the median employee based upon rates of pay, rather than pay earned over the course of a year, will reduce the burden and minimize the skewing effects on the ratio of a large number of part-

        time, temporary and seasonal employees. Provided that the method is disclosed, the final rule should give issuers flexibility in this regard.'').

        \271\ See letter from SH&P.

        \272\ See letter from FSR. See also Crowdfunding, Release No. 33-9470 (Oct. 23, 2013) 78 FR 66427.

        ---------------------------------------------------------------------------

        (ii) Statistical Sampling

        We received many comments on using statistical sampling for identifying the median employee, with a majority of these commenters supporting the use of statistical sampling, as permitted in the proposed rule. Many of these commenters suggested that allowing the use of statistical sampling would reduce the costs for registrants, without specifically quantifying a reduction.\273\ Some commenters that supported statistical sampling recommended that the final rule not specify requirements for statistical sampling (such as appropriate sample size, confidence levels, or other requirements).\274\ One of these commenters contended that specifying requirements for statistical sampling would unduly constrain registrants from developing the most appropriate methodology and would be inconsistent with flexibility.\275\

        ---------------------------------------------------------------------------

        \273\ See, e.g., letters from AFL-CIO I, AFSCME, Prof. Angel, Bricklayers International, Calvert, Chesapeake Utilities, Chicago Teachers Fund, CUPE, Davis Polk, First Affirmative, FS FTQ, FSI, ICCR, IL Bricklayers and Craftworkers Union, LIUNA, Marco Consulting, McMorgan & Co., Meridian, NACD, Novara Tesija, NRF, NY Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, Rep. Ellison et al. I; Socially Responsive Financial Advisors, TCA, Teamsters, Trillium I, Trustee Campbell, US SIF, and WA State Investment Board.

        \274\ See, e.g., letters from ABA, IBC, Johnson & Johnson, and WorldatWork I.

        \275\ See letter from ABA.

        ---------------------------------------------------------------------------

        Another commenter asserted that the statute permits statistical sampling because Section 953(b) does not prescribe a particular way to identify the median employee.\276\ Some commenters stated that registrants would likely use statistical sampling in identifying their median employee,\277\ and that statistical sampling is feasible.\278\ A few commenters agreed with the proposal that registrants should be allowed to identify the median employee by using statistical sampling based on a definition of compensation other than ``annual total compensation'' under Item 402.\279\ Some commenters indicated that the final rule should include a safe harbor for statistical sampling.\280\

        ---------------------------------------------------------------------------

        \276\ See letter from Michael Ohlrogge, Stanford Law School and Stanford Department of Management Science and Engineering (Sep. 25, 2013) (``Ohlrogge I'').

        \277\ See, e.g., letters from Capital Strategies and Mercer I.

        \278\ See, e.g., letters from Ohlrogge I, Michael Ohlrogge, Stanford Law School and Stanford Department of Management Science and Engineering (Jul. 5, 2015) (``Ohlrogge II''), and Mike Petty (Oct. 21, 2013) (``M. Petty'').

        \279\ See, e.g., letters from Emergent and McGuireWoods.

        \280\ See, e.g., letters from American Benefits Council; E&Y; FSR; Dr. Sandy J. Miles, Professor of Human Resource Management, Murray State University (Nov. 29, 2013) (``Prof. Miles''); Strus and Associates Inc. (Nov. 30, 2013) (``Strus and Assoc.''); and TCA.

        ---------------------------------------------------------------------------

        Although the majority of commenters that discussed statistical sampling supported its use in identifying the median employee, one commenter stated specifically that the final rule should discourage the use of statistical sampling in favor of information

        Page 50134

        derived from tax and/or payroll records to determine actual employee pay rather than an estimated amount.\281\ Other commenters, while not necessarily opposed to statistical sampling, contended that it would not mitigate costs of collecting and assembling employee compensation data, which, in their view, is the most expensive part of the rule.\282\

        ---------------------------------------------------------------------------

        \281\ See letter from LAPFF.

        \282\ See, e.g., letters from Business Roundtable I, Chamber I, COEC I, COEC II, ExxonMobil, Freeport-McMoRan, FuelCell Energy, Garmin, Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and WorldatWork I.

        ---------------------------------------------------------------------------

        (iii) Consistently Applied Compensation Measures

        Most commenters that discussed using consistently applied compensation measures, such as information derived from tax and/or payroll records, to identify the median employee agreed with the proposed rule's approach.\283\ Generally, these commenters contended that permitting the use of such measures would reduce costs while not impairing the pay ratio's usefulness. For example, one commenter noted that, while using a consistently applied compensation measure may exclude benefits, perquisites, and other allowances, it would still capture salary, incentive cash earned, and stock awards.\284\ Therefore, the measure would include ``the substantial majority of compensation and would not lead to distortion of the median.'' Another commenter asserted that the final rule should be flexible enough to encompass different approaches across jurisdictions.\285\

        ---------------------------------------------------------------------------

        \283\ See, e.g., letters from ABA, AFL-CIO I, Calvert, Chesapeake Utilities, Chicago Teachers Fund, CII, First Affirmative, FS FTQ, ICCR, IL Bricklayers and Craftworkers Union, Johnson & Johnson, Marco Consulting, Meridian, Microsoft, Novara Tesija, NY Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, SH&P, WA State Investment Board, and WorldatWork I.

        \284\ See letter from Microsoft.

        \285\ See letter from ABA.

        ---------------------------------------------------------------------------

        One commenter disagreed with permitting the use of consistently applied compensation measures on the basis that it would alter the pay ratio because not all compensation would be included.\286\ Some commenters noted that using consistently applied compensation measures would not reduce the costs for registrants with non-U.S. employees.\287\ Therefore, several of these commenters recommended that the final rule should allow registrants to use one or more comparable consistently applied compensation measures and not be limited to a single consistently applied compensation measure for employees in different international jurisdictions to reduce registrants' costs and burdens.\288\

        ---------------------------------------------------------------------------

        \286\ See letter from Tumeh.

        \287\ See, e.g., letters from Corporate Secretaries, Davis Polk, and FSR.

        \288\ See, e.g., letters from Corporate Secretaries (``For example, a company may be able to use W-2 or payroll information to relatively efficiently and accurately quantify annual cash compensation for US employees, but this data could be impracticable to replicate for certain non-US employees due to privacy concerns, comparability of compensation schemes, tax systems, or otherwise. . . . In light of this, we propose that the Commission consider including in the final rules that for employees based in non-US jurisdictions, registrants be permitted to use (i) the same compensation measure used in the US or (ii) another reasonably comparable compensation measure for any non-US jurisdiction where the same compensation measure is not used, unless a different application is required or compensation cannot be estimated in a particular jurisdiction for a particular reason (i.e., data privacy laws).'') and Davis Polk (``For example, a registrant using cash compensation as its consistently applied compensation measure for purposes of determining its median employee may intend to include only base salary and bonus amounts for its U.S. employees, but may find it appropriate to also include other benefits commonly considered to be part of base compensation for its non-U.S. employees, such as meal stipends or automobile allowances. We believe that it would be appropriate for registrants to use these reasonably comparable measures, even if not the exact same measure, across different employee populations in order to provide for more meaningful comparisons across the varying compensation structures in international jurisdictions and to reduce unnecessary burdens on registrants.'').

        ---------------------------------------------------------------------------

        (iv) The ``Median'' Employee

        A number of commenters recommended that the rule use compensation of an employee, or employees, other than the median employee as is required in Section 953(b). A few of these commenters suggested that, instead of using their median employee in their pay ratio, registrants should be permitted to use their ``average'' employee.\289\ The commenters contended that using ``average'' instead of ``median'' would reduce costs, would be better understood by the public, and could be calculated easily using tax records.\290\ One of these commenters acknowledged that the statutory language of Section 953(b) uses the word ``median'' rather than ``average,'' but pointed out that the Proposing Release quotes from two letters submitted by members of Congress, including two co-sponsors of Section 953(b), in which the members refer to the ``average'' and ``typical'' employee instead of the ``median'' employee when discussing the statute.\291\ Also, the commenter noted that we stated in the Proposing Release that ``Section 953(b) does not expressly set forth a methodology that must be used to identify the median, nor does it mandate that we must do so in our rules.'' \292\

        ---------------------------------------------------------------------------

        \289\ See, e.g., letters from Hyster-Yale, Mercer I, NACCO, and Powers.

        \290\ See, e.g., letters from Hyster-Yale and NACCO. But see letter from Suzanne Laatsch (Nov. 29, 2013) (``Laatsch'') (asserting that using the median employee, as opposed to using the average employee, would actually reduce costs).

        \291\ See letter from NACCO.

        \292\ See letter from NACCO.

        ---------------------------------------------------------------------------

        Other commenters recommended that the final rule allow registrants to use existing BLS data to calculate their pay ratios, which is based on average employee compensation figures.\293\ One of these commenters contended that allowing registrants to use existing data sources, including BLS data, would be faithful to the intent of Section 953(b) and not be ``materially'' different than using the median, even though it would be less precise.\294\

        ---------------------------------------------------------------------------

        \293\ See, e.g., letters from Brian Foley & Co., NACD, and NIRI.

        \294\ See letter from NIRI. But see letter from IPS (disagreeing specifically with the NIRI letter and asserting that the final rule should not permit registrants to use BLS data).

        ---------------------------------------------------------------------------

        A few commenters contended that registrants should be permitted to use the compensation of a range of employees as the median compensation instead of the compensation of the median employee.\295\ One of these commenters noted that it would be unlikely for a registrant, using a convenient and cost-effective measure, to determine a single employee that is the median.\296\ According to this commenter, it would be more likely that the registrant's calculations would yield a group of employees, any of whom who could serve as the median employee. Therefore, once the range of employees is determined, registrants should be permitted to use any reasonable method to determine which employee to use as the median employee.

        ---------------------------------------------------------------------------

        \295\ See, e.g., letters from Capital Strategies, Hyster-Yale, Mercer I, and NACCO.

        \296\ See letter from NYC Bar.

        ---------------------------------------------------------------------------

        Similarly, one commenter recommended that the final rule permit registrants, after identifying the median employee using whatever methodology they select, to use another employee as the median employee if that employee is within a 1% variance of the median and the original employee has anomalous compensation characteristics that would create the risk of a distorted pay ratio.\297\ The commenter recognized that Section 953(b) refers to the ratio of the ``median'' employee's annual compensation to the compensation of the PEO but contended that some deviation from that precise statutory language should be acceptable if it furthers the statute's intent to show the

        Page 50135

        ratio of the compensation of the typical or representative employee to that of the PEO.

        ---------------------------------------------------------------------------

        \297\ See letter from PNC Financial Services.

        ---------------------------------------------------------------------------

        Another commenter advocated using one of two alternative approaches based on whether the registrant has international employees that would segregate employees with similar positions into different groups. Registrants could identify in which group the median resides based on a ratio that compares the total number of employees at each job level to the total employee population and use sampling or another technique to identify the median of that group, which the registrant would use as its median for pay ratio purposes.\298\

        ---------------------------------------------------------------------------

        \298\ See letter from Tower Watson. See also letter from Mercer I (advocating a somewhat similar approach using multiple statistical samples).

        ---------------------------------------------------------------------------

        (c) Final Rule

        We are adopting the final rule as proposed. Consistent with the proposal, the final rule does not specify any required methodology for registrants to use in identifying the median employee. Instead, the final rule permits registrants the flexibility to choose a method to identify the median employee based on their own facts and circumstances. To identify the median employee, registrants may use a methodology that uses reasonable estimates. The median employee may be identified using annual total compensation or any other compensation measure that is consistently applied to all employees included in the calculation, such as information derived from tax and/or payroll records. Also, in determining the employees from which the median is identified, a registrant is permitted to use its employee population or statistical sampling and/or other reasonable methods. In any event, the final rule requires a registrant to briefly describe the methodology it used to identify the median employee and any material assumptions, adjustments (including any cost-of-living adjustments), or estimates it used to identify the median employee or to determine total compensation or any elements of total compensation, which shall be consistently applied. The registrant also must clearly identify any estimates used.

        (i) Flexibility

        As we noted in the Proposing Release, we believe that allowing registrants the flexibility to choose a method that works best for their particular facts and circumstances will help them comply with the rule in a relatively cost-efficient manner while still fulfilling the purpose of Section 953(b). We recognize that a flexible approach could increase uncertainty for registrants that prefer more specificity on how to comply with the final rule, particularly for those registrants that do not use statistical analysis in the ordinary course of managing their businesses. We believe that any negative effects caused by any uncertainty would be offset by the positive effects of permitting flexibility. Also, the final rule establishes certain methodologies and permissible uses of estimates, such that registrants may use reasonable estimates both in the methodology used to identify the median employee and in calculating annual total compensation or any elements of total compensation for employees other than the PEO; use their employee population, statistical sampling, or another reasonable methods in determining the employees from which the median employee is identified, and identify the median employee using annual total compensation or any other compensation measure that is consistently applied to all employees included in the calculation.

        We believe also that any uncertainty provided by the final rule's flexibility is offset by other benefits. Particularly, as we noted in the Proposing Release, the final rule's flexibility allows registrants to provide the required disclosure in a relatively cost-efficient manner based on the registrant's own facts and circumstances using total compensation for each employee under Item 402(c)(2)(x), statistical sampling, reasonable estimates, or the use of any consistently applied compensation measures to identify the median. A large number of commenters supported this flexibility because it would reduce the rule's costs without significantly diminishing its benefits.\299\ In this regard, we do not believe permitting this flexibility will limit the usefulness of the pay ratio for shareholders as a data point in making their voting decisions on executive compensation under Section 951 of the Dodd-Frank Act.

        ---------------------------------------------------------------------------

        \299\ See, e.g., letters from ABA, AFSCME, Barnard, Bricklayers International, Capital Strategies, CII, Domini, LIUNA, Fowler, Mirczak, PM&P, PNC Financial Services, US SIF, Vivient, Walden, and WorldatWork I.

        ---------------------------------------------------------------------------

        In determining their methodology, registrants may consider, among other factors, such variables as: The size and nature of the workforce; the complexity of the organization; the stratification of pay levels across the workforce; the types of compensation the employees receive; the extent that different currencies are involved; the number of tax and accounting regimes involved; and the number of payroll systems the registrant has and the degree of difficulty involved in integrating payroll systems to readily compile total compensation information for all employees. These likely are the same factors that could cause substantial variation in the costs of compliance, but the final rule's flexibility should help registrants reduce these costs.

        As part of the flexibility permitted by the final rule, registrants may use a methodology that uses reasonable estimates in identifying the median employee and calculating the annual total compensation or any elements of compensation for employees other than the PEO, as proposed. Most commenters on this issue agreed that the final rule should permit reasonable estimates to mitigate the rule's costs.\300\ Some of these commenters suggested that the final rule should not include further requirements or guidance for making reasonable estimates, including safe harbors, assumptions for error rates, or confidence levels.\301\ We are not persuaded that further guidance is necessary.

        ---------------------------------------------------------------------------

        \300\ See, e.g., letters from ABA, Capital Strategies, Davis Polk, Hyster-Yale, Johnson & Johnson, NACCO, and WorldatWork I.

        \301\ See, e.g., letters from ABA, Capital Strategies, Hyster-

        Yale, Johnson & Johnson, and NACCO. But see letters from NSFM, OCP, PM&P, Quintave, and SHRM (recommending that the final rule include a safe harbor for identifying the median employee).

        ---------------------------------------------------------------------------

        Further, as proposed, in determining the employees from which the median is identified, the final rule permits registrants to use ``other reasonable methods'' in addition to using its employee population or statistical sampling. Some commenters provided suggestions on the ``other reasonable methods'' a registrant should be permitted to use in identifying the median employee.\302\ We are not specifying the ``other reasonable methods'' that may be appropriate because we seek to allow each registrant

        Page 50136

        the flexibility to determine the method that best suits its own facts and circumstances, which may include some of the suggestions made by these commenters.

        ---------------------------------------------------------------------------

        \302\ See, e.g., letters from E&Y (``In the adopting release, we believe the Commission should identify additional methods that, if used, would satisfy the requirements for determining which employees should be included in the analysis.''), FSR (indicating that an issuer's methodology should be based on a ``good faith compliance'' standard that is akin to our proposed crowdfunding rules), RILA (requesting that the final rule allow issuers to identify the median employee based on rates of pay instead of pay earned over the course of the year), TCA (suggesting that final rule allow the following methods: (1) Specific safe harbor assumptions about the statistical distribution of compensation within the company and its business units; (2) formulaic, numerical, and other computational approaches to estimate the median compensation; and (3) disclosure of a reasonable range of outcomes rather than requiring the ``right'' outcome), and Towers Watson (recommending that the final rule permit registrants to ``employ a methodology using salary grades or job levels, as appropriate, to reasonable identify the median'' (emphasis in original)).

        ---------------------------------------------------------------------------

        (ii) Statistical Sampling

        Consistent with the proposal, the final rule permits registrants to use statistical sampling in determining the employees from which the median is identified. We believe permitting statistical sampling is appropriate under Section 953(b). Statistical sampling is based on statistical theory to make sampling more efficient. For example, with large populations, results accurate enough to be useful can be obtained from samples that represent only a small fraction of the population.\303\ We note that one commenter recommended that we discourage statistical sampling in favor of using the information derived from tax and/or payroll records to determine actual employee pay rather than allowing registrants to use an estimation.\304\ We believe, however, that statistical sampling can be a reliable means of identifying the median employee.

        ---------------------------------------------------------------------------

        \303\ See generally Cochran, W. G. (1977), Sampling Techniques, 3rd Edition, New York: Wiley. Different statistical sampling methods have been developed and have been well established in literature and practice since 1786. See Laplace, P.S. (1786). ``Sur Les Naissances, Les Mariages Et Les Morts,'' In Histoire de L'Academie Royale des Sciences, 1783, Paris, 693-702; Stephan, F. F. (1948). History of the uses of modern sampling procedures. Journal of the American Statistical Association 43:12-37; Godambe, V. P. (1954). A unified theory of sampling from finite populations. Journal of the Royal Statistical Society 17(2):269-278; Sukhatme, P. V. (1966). Major developments in sampling theory and practice. Research papers in statistics. F.N. David Ed. John Wiley & Sons. 367-409pp; Lohr, S.L. (2009), Sampling: Design and Analysis, 2nd Edition, Cengage Learning; and Rao, J.N.K. (2011). Impact of frequentist and Bayesian methods on survey sampling practice: a selective appraisal. Statistical Science 26(2): 240-256pp.

        \304\ See letter from LAPFF.

        ---------------------------------------------------------------------------

        Some commenters indicated that permitting statistical sampling in the final rule would not mitigate costs because the final rule would still require registrants to collect and assemble employee compensation data, which the commenters view as the most expensive part of the rule.\305\ We believe, however, that permitting registrants to use statistical sampling could lead to a reduction in compliance costs as compared with other methods of identifying the median employee without significantly affecting the pay ratio because registrants are not required to calculate the total compensation for each of their employees. Therefore, although the final rule still requires registrants to collect and assemble employee compensation data, the availability of statistical sampling may allow them to assemble far less employee compensation data than if the final rule prohibited such sampling. We note that a number of other commenters indicated that permitting statistical sampling would reduce costs.\306\ Also, because of the reliability of the result achieved through appropriately conducted statistical sampling, we do not believe the use of sampling will limit the usefulness of the pay ratio for shareholders as a data point in making their voting decisions on executive compensation under Section 951 of the Dodd-Frank Act.

        ---------------------------------------------------------------------------

        \305\ See, e.g., letters from Business Roundtable I, Chamber I, COEC I, ExxonMobil, Freeport-McMoRan, FuelCell Energy, Garmin, Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and WorldatWork I.

        \306\ See, e.g., letters from AFL-CIO I, AFSCME, Bricklayers International, Calvert, Chesapeake Utilities, Chicago Teachers Fund, CUPE, Davis Polk, Dennis T, First Affirmative, FS FTQ, FSI, ICCR, IL Bricklayers and Craftworkers Union, LIUNA, Marco Consulting, McMorgan & Co., Meridian, NACD, Novara Tesija, NRF, NY Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, Rep. Ellison et al. I, Socially Responsive Financial Advisors, TCA, Teamsters, Trillium I, Trustee Campbell, Prof. Angel, US SIF, and WA State Investment Board.

        ---------------------------------------------------------------------------

        The final rule does not provide specific parameters for statistical sampling, including the appropriate sample size. We agree with commenters that specifying requirements for statistical sampling, including appropriate sample sizes, confidence levels, or other requirements would unduly constrain registrants from developing the most appropriate methodology.\307\ Instead, we believe registrants must make their own determinations on what is appropriate based on their own facts and circumstances.

        ---------------------------------------------------------------------------

        \307\ See, e.g., letters from ABA, IBC, Johnson & Johnson, and WorldatWork I.

        ---------------------------------------------------------------------------

        We are, however, providing some guidance for registrants when using statistical sampling. In this regard, based on the analysis we described in the Proposing Release, we believe that a relatively small sample size may be appropriate in certain situations.\308\ A reasonable determination of sample size ultimately depends on the underlying distribution of compensation data. Further, we believe that reasonable estimates of the median for registrants with multiple business lines or geographical units may be determined using more than one statistical sampling approach. Additionally, all statistical sampling approaches should draw observations from each business or geographical unit with a reasonable assumption on each unit's compensation distribution and infer the registrant's overall median based on the observations drawn.

        ---------------------------------------------------------------------------

        \308\ In that analysis, we determined that the appropriate sample size for the registrants with a single business or geographical unit varied between 81 and 1,065 employees across industries, with the average estimated sample size close to 506.

        ---------------------------------------------------------------------------

        Moreover, as we noted in the Proposing Release, the identification of a median employee does not necessarily require a determination of exact compensation amounts for every employee included in the sample. For example, rather than calculating exact compensation, a registrant could identify the employees in its sample that have extremely low or extremely high pay that would, therefore, fall on either end of the compensation spectrum. Since identifying the median involves finding the employee in the middle, it may not be necessary to determine the exact compensation amounts for every employee paid more or less than that employee in the middle. Instead, just noting that the employees are above or below the median may be sufficient for finding the employee in the middle of the compensation spectrum.

        (iii) Consistently Applied Compensation Measures

        As proposed, the final rule permits registrants to use a consistently applied compensation measure, such as information derived from tax and/or payroll records, in determining the employees from which the median is identified as long as the registrant discloses the compensation measure used. Due to concerns about expected compliance costs arising from the complexity of using the ``total compensation'' calculation under Item 402(c)(2)(x) when identifying the median, we sought a reasonable alternative to identifying the median employee that is easier to calculate.

        As we noted in the Proposing Release, this approach provides a workable identification of the median employee for many registrants, and we expect it will reduce the costs of compliance. Most commenters discussing this issue agreed with this position and supported permitting registrants to use consistently applied compensation measures, such as information derived from tax and/or payroll records, to identify the median employee because it would reduce costs while not significantly affecting a registrant's pay ratio.\309\ As one commenter noted, while

        Page 50137

        a consistently applied compensation measure may exclude benefits, perquisites, and other allowances, it will still capture salary, incentive cash earned, and stock awards, which will encompass ``the substantial majority of compensation and should not lead to distortion of the median.'' \310\ In light of these comments, we do not believe this provision will hinder shareholders in using the pay ratio as a potentially useful data point in making their voting decisions on executive compensation under Section 951 of the Dodd-Frank Act.

        ---------------------------------------------------------------------------

        \309\ See, e.g., letters from ABA, AFL-CIO I, Calvert, Chesapeake Utilities, Chicago Teachers Fund, CII, First Affirmative, FS FTQ, ICCR, IL Bricklayers and Craftworkers Union, Johnson & Johnson, Marco Consulting, Meridian, Microsoft, Novara Tesija, NY Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, SH&P, WA State Investment Board, and WorldatWork I.

        \310\ See letter from Microsoft.

        ---------------------------------------------------------------------------

        One commenter asserted that the final rule should be flexible enough to permit a registrant to use an internally consistent measure that is not necessarily internally identical.\311\ According to the commenter, the consistently applied compensation measure could be a measure based on the registrant's individual organizational structure that is reasonably designed to identify the median employee. As an example, the commenter stated that if a registrant selects to use ``taxable wages'' as the consistently applied compensation measure, it is possible that the measure may be defined differently across multiple jurisdictions and be calculated over differing time periods. In such a situation, the commenter suggested that, although the registrant should attempt to use the non-United States equivalent to a Form W-2 for purposes of conducting its analysis, the final rule should provide sufficient flexibility to permit the use of other reasonable data sources for collecting the comparable information relating to non-U.S. employees as long as such measures are consistently applied within each subject jurisdiction.

        ---------------------------------------------------------------------------

        \311\ See letter from ABA.

        ---------------------------------------------------------------------------

        We agree. We note that a consistently applied compensation measure, such as ``taxable wages'' or ``cash compensation,'' \312\ may be defined differently across jurisdictions and may include different annual periods. For purposes of calculating the annual total compensation amounts when using a consistently applied compensation measure, the final rule permits registrants to use a measure that is defined differently across jurisdictions and may include different annual periods as long as within each jurisdiction, the measure is consistently applied. A registrant, however, would not be permitted to use an entirely different type of measure across jurisdictions that would not be consistently applied. The final rule does not require registrants to use any specific compensation measure when identifying the median employee. We continue to believe that registrants are in the best position to select a compensation measure that is appropriate to their own facts and circumstances. Therefore, consistent with the proposal, the final rule permits registrants to identify a median employee based on any consistently applied compensation measure, such as information derived from tax and/or payroll records, as long as the registrant briefly discloses the measure that it used. After the median employee is identified, registrants must calculate that median employee's annual total compensation in accordance with Item 402(c)(2)(x).

        ---------------------------------------------------------------------------

        \312\ See, e.g., letters from Corporate Secretaries and Davis Polk.

        ---------------------------------------------------------------------------

        (iv) The ``Median'' Employee

        Some commenters recommended that the final rule permit registrants to calculate the ratio using a figure other than the median, such as the employee earnings estimates available through the BLS,\313\ which may reduce costs for registrants and promote comparability across companies. As we stated in the Proposing Release, although such an approach would greatly reduce the compliance burden for registrants, we do not believe it is consistent with Section 953(b).

        ---------------------------------------------------------------------------

        \313\ See, e.g., letters from Brian Foley & Co., NACD, and NIRI. Some commenters recommended that the final rule permit registrants to use the average salary of its employees instead of the median. See, e.g., letters from Hyster-Yale, and NACCO, which they suggest also may reduce costs for registrants and promote comparability across companies. As with using the BLS data, however, we do not believe this approach is consistent with Section 953(b).

        ---------------------------------------------------------------------------

        Consistent with the proposal, the final rule requires that registrants provide their pay ratio disclosure using the compensation of their median employee. Section 953(b)(1)(A) states specifically that registrants must disclose ``the median of the annual total compensation of all employees of the issuer, except the chief executive officer.'' We believe, therefore, that Congress intended that the final rule should require registrants to use their median employee in their pay ratio determination.

        Although we considered commenters' arguments that ``median,'' as used in Section 953(b), can be interpreted to mean a measure other than ``median,'' such as ``average'' or ``mean,'' we believe the better reading of the statutory language is that it means the statistical median of all employee compensation. ``Median'' has a specific meaning in statistics and probability theory,\314\ and there is no reason to believe that, when Congress chose to use this term in Section 953(b), it intended that it not have its established meaning. In the context of the disclosure required by Section 953(b), the alternative measures suggested by commenters could produce a significantly different number than the ``median,'' and we do not believe that would be appropriate in light of the statutory language and our understanding of its purpose. Further, we believe the median may provide a more useful or relevant data point for shareholders making their say-on-pay votes than would a mathematical average because the use of median can decrease the significance of outliers.

        ---------------------------------------------------------------------------

        \314\ The dictionary defines ``median'' as ``the middle number in a sequence, or the average of the two middle numbers when the sequence has an even number of numbers.'' Random House Webster's Dictionary, 411 (2d ed. 1996).

        ---------------------------------------------------------------------------

        Similarly, we note that some commenters suggested that a registrant should be permitted to use the compensation of a range of employees as the median compensation instead of the compensation of the exact median employee.\315\ We disagree. No matter what method a registrant chooses to identify its median employee, it must identify an actual employee and determine that employee's annual total compensation to use in the pay ratio disclosure. We note, however, that the final rule does not require a registrant to disclose any personally identifiable information about that employee other than his or her compensation. A registrant may choose to generally identify the employee's position to put the employee's compensation in context, but the registrant is not required to provide this information and should not do so if providing the information could identify any specific individual.

        ---------------------------------------------------------------------------

        \315\ See, e.g., letters from Capital Strategies, Hyster-Yale, Mercer I, NACCO, and NYC Bar.

        ---------------------------------------------------------------------------

        Another commenter recommended that a registrant, after identifying the median employee, should be able to select another employee as the median if that employee is within a 1% variance of the median and the original employee has anomalous compensation characteristics that would result in a pay ratio that did not accurately reflect the relationship between the compensation practices for a typical employee and the compensation of the

        Page 50138

        CEO.\316\ Given the significant flexibility that the rule provides registrants in identifying the median employee, we believe it would be appropriate to substitute another substantially similarly situated employee in these circumstances. Thus, when calculating the total compensation for that median employee in accordance with Item 402(c)(2)(x), if the registrant reasonably determines that there are anomalous compensation characteristics of that employee's compensation that would have a significant higher or lower impact on the pay ratio, we will not object if the registrant substitutes another employee with substantially similar compensation to the original median employee based on the compensation measure used to select the median employee. The registrant must, however, disclose this fact as part of its brief description of the methodology it used to identify the median employee.

        ---------------------------------------------------------------------------

        \316\ See letter from PNC Financial Services (noting that if the median employee is chosen by a metric other than total Item 402(c) compensation, when that median employee's compensation is then calculated using Item 402(c), it is possible that such employee may have pay elements or be missing pay elements that would make his or her compensation anomalous when compared with others at the same overall compensation level, and providing some examples such as if the employee does not participate in certain benefit programs that are reflected in Item 402(c) compensation but not in W-2 compensation).

        ---------------------------------------------------------------------------

        b. Calculating Annual Total Compensation

        i. Proposed Rule

        The proposed rule would define ``total compensation'' by reference to Item 402(c)(2)(x). In the Proposing Release, we stated that, because of the complexity of the requirements of Item 402(c)(2)(x), registrants typically compile information required by Item 402(c) manually for the named executive officers, which takes significant time and resources. Given the specificity of the definition used in Section 953(b), the proposed rule incorporated the Item 402(c)(2)(x) definition of ``total compensation'' for purposes of disclosing the median of the annual total compensation of employees and the pay ratio. Because the total compensation calculation using Item 402(c)(2)(x) would only be required for one additional employee (the median employee), we did not propose to simplify the total compensation definition that is required to be used to disclose the median employee compensation and the ratio.

        Additionally, the proposed rule permitted the use of reasonable estimates in determining any elements of total compensation of employees other than the PEO under Item 402(c)(2)(x). For registrants using estimates, an instruction to the proposed rule would require them to disclose and consistently apply any material estimate used to identify the median employee or to determine that employee's total compensation or any elements of total compensation, and to clearly identify any estimates used. In using an estimate for annual total compensation (or for a particular element of total compensation), a registrant would be required to have a reasonable basis to conclude that the estimate approximates the actual amount of compensation under Item 402(c)(2)(x) (or for a particular element of compensation under Item 402(c)(2)(iv)-(ix)) awarded to, earned by, or paid to the employee. We did not specify what a reasonable basis would entail because we believed that would necessarily depend on the registrant's particular facts and circumstances.

        Because the requirements of Item 402(c)(2)(x) were promulgated to address executive officer compensation, rather than compensation for all employees, we considered some interpretive questions that registrants could face in applying the requirements of Item 402(c)(2)(x) to employees who are not executive officers and proposed ways to address those questions. Those included questions concerning: applying the definition of ``total compensation'' to an employee who is not an executive officer and valuation issues for certain elements of total compensation, including for non-U.S. employees.

        ii. Comments on the Proposed Rule

        One commenter agreed with the Proposed Rule's requirement that ``total compensation'' be calculated using the requirements of Item 402(c)(2)(x).\317\ Also, this commenter and a few others stated specifically that they supported the proposed rule's flexibility in permitting the use of reasonable estimates to calculate the annual total compensation or any elements of total compensation for employees other than the PEO.\318\ One commenter indicated that it expected its clients to use reasonable estimates for calculating total compensation.\319\ Some commenters requested additional guidance as to what estimates would be considered reasonable,\320\ but other commenters said that no additional guidance is required.\321\

        ---------------------------------------------------------------------------

        \317\ See letter from COEC II.

        \318\ See, e.g., letters from ABA, Prof. Angel, COEC I, COEC II, Davis Polk, and Vectren.

        \319\ See letter from Mercer I.

        \320\ See, e.g., letters from Aon Hewitt (requesting also a safe harbor in the final rule) and PM&P.

        \321\ See, e.g., letters from ABA and COEC II.

        ---------------------------------------------------------------------------

        One commenter stated that it was ``deeply troubled'' that the proposed rule would require registrants to calculate the median employee's total compensation using Item 402(c)(2)(x) because no registrant uses this measure to calculate a non-executive officer's compensation.\322\ According to the commenter, a significant difficulty with using Item 402(c)(2)(x) is the inclusion of pension benefits because the actuarial value of pensions vacillate dramatically from year to year, which would significantly impact total compensation. The commenter recommended excluding pension value from the total compensation calculation. If the final rule does not exclude pensions, the commenter suggested that government-related pensions for non-U.S. employees should be excluded as they are for named executive officers under Item 402(c)(2)(x).

        ---------------------------------------------------------------------------

        \322\ See letter from PM&P.

        ---------------------------------------------------------------------------

        Other commenters also contended that the final rule should not require calculating total compensation using Item 402(c)(2)(x) because certain compensation measures are excluded from that item requirement, such as benefits, non-discriminatory plans, perquisites, and personal benefits that aggregate less than $10,000, which would cause the median employee's total compensation to be understated.\323\ Another commenter, while not necessarily advocating against using Item 402(c)(2)(x) to calculate total compensation, noted that it would be difficult for registrants to use that item requirement because it would require them to include as compensation bonuses, stock compensation, pensions, and other benefits for the median employee that are usually not factored into annual salary amounts in company records.\324\

        ---------------------------------------------------------------------------

        \323\ See, e.g., letters from Hyster-Yale, NACCO, and NACD.

        \324\ See letter from IBC.

        ---------------------------------------------------------------------------

        Some commenters suggested that the final rule prescribe a methodology for calculating total compensation.\325\ One commenter requested specifically that we include guidance about excluding government-mandated pension plans.\326\ Another commenter suggested that registrants should be permitted (but not required) to include benefits, non-discriminatory plans, perquisites, and personal benefits that aggregate less than $10,000 in total compensation because Item 402(c)(2)(x) merely

        Page 50139

        permits and does not require exclusion of such items for executive officers.\327\ Other commenters contended that the calculation of total compensation in the final rule should include all compensation of a registrant's PEO and median employee, such as benefits, perks, bonuses, stock options, and other forms of compensation; \328\ be limited to cash and stock-based compensation; \329\ or include only compensation on the W-2 Form of the PEO and median employee.\330\

        ---------------------------------------------------------------------------

        \325\ See, e.g., letters from Friend and William Preston (Oct. 5, 2013) (``Preston'').

        \326\ See letter from ABA.

        \327\ Id.

        \328\ See, e.g., letters from Corayer, Fedewa, Gould, LAPFF, Matteson, and Anne C. Somers (Oct. 24, 2013) (``Somers'').

        \329\ See, e.g., letters from Brian Foley & Co., Dover Corp., and Semtech.

        \330\ See, e.g., letters from Capital Strategies, FEI, Hyster-

        Yale, and NACCO.

        ---------------------------------------------------------------------------

        One commenter recommended that total compensation either include the average change in defined benefit pension values or exclude defined benefit pension values entirely.\331\ Another commenter indicated that total compensation should include welfare and retirement benefits included to union members.\332\ One commenter indicated that total compensation should be calculated based upon the same method used when identifying the three most highly compensated executive officers in accordance with Instruction 1 of Item 402(a)(3), which excludes the value of the aggregate change in actuarial present value of defined pension benefits under Item 402(c)(2)(viii).\333\

        ---------------------------------------------------------------------------

        \331\ See letter from Mercer I.

        \332\ See letter from Vectren Corp.

        \333\ See letter from SH&P.

        ---------------------------------------------------------------------------

        Finally, one commenter suggested that, if the final rule allows registrants to identify the median employee only every three years, registrants should similarly be allowed to calculate total compensation either every year or only when a new median is determined unless there has been a material change in annual total compensation.\334\

        ---------------------------------------------------------------------------

        \334\ See letter from ABA.

        ---------------------------------------------------------------------------

        iii. Final Rule

        The definition of ``total compensation'' in the final rule is unchanged from the proposal. The final rule requires that ``total compensation'' for both the median employee and PEO be calculated using the requirements of Item 402(c)(2)(x). As with the proposed rule, the final rule provides registrants with flexibility in identifying the median employee, and does not require registrants to identify the median employee by calculating the total compensation for each employee. Because the total compensation calculation using Item 402(c)(2)(x) is only required for one additional employee (the median employee), we do not believe it necessary in the final rule to simplify the total compensation definition that is required to be used in that calculation.

        THE FINAL RULE PERMITS REGISTRANTS TO USE REASONABLE ESTIMATES IN CALCULATING THE ANNUAL TOTAL COMPENSATION OF THEIR MEDIAN EMPLOYEE, INCLUDING ANY ELEMENTS OF THE TOTAL COMPENSATION, UNDER ITEM 402(C)(2)(X). A FEW COMMENTERS SUPPORTED SUCH FLEXIBILITY.\335\ WE BELIEVE, AS WE NOTED IN THE PROPOSING RELEASE, THAT THE USE OF REASONABLE ESTIMATES DOES NOT DIMINISH THE POTENTIAL USEFULNESS OF THE PAY RATIO DISCLOSURE, IS CONSISTENT WITH SECTION 953(B), AND WILL RESULT IN LOWER COMPLIANCE COSTS ON REGISTRANTS.

        ---------------------------------------------------------------------------

        \335\ See, e.g., letters from ABA, Capital Strategies, COEC I, Davis Polk, Johnson & Johnson, Mercer I, Vectren, Corp., and WorldatWork I.

        ---------------------------------------------------------------------------

        Under the final rule, registrants must clearly identify any estimates used. Additionally, registrants must have a reasonable basis to conclude that their estimates approximate the actual amounts of Item 402(c)(2)(x) compensation, or a particular element of compensation under Item 402(c)(2)(iv)-(ix), that are awarded to, earned by, or paid to the median employee. Some commenters requested that we provide additional guidance as to what estimates would be considered reasonable.\336\ Consistent with the Proposing Release, we are not prescribing what a reasonable basis would entail in the final rule because we believe that will necessarily depend on the registrant's particular facts and circumstances.

        ---------------------------------------------------------------------------

        \336\ See, e.g., letters from Aon Hewitt (requesting also a safe harbor in the final rule) and PM&P.

        ---------------------------------------------------------------------------

        As we discussed in the Proposing Release, our rules for compensation disclosure focus on the compensation of executive officers and directors rather than compensation for all employees. Some commenters urged us not to require registrants to calculate total compensation using Item 402(c)(2)(x).\337\ We believe, however, that it is appropriate to require ``total compensation'' to be calculated using Item 402(c)(2)(x) for both the median employee and PEO. Using different measures of total compensation for the median employee and the PEO would be inconsistent with the statutory language and alter the pay ratio. Due to the concerns about calculating median employee compensation using requirements meant only for executive offices, however, we are reiterating the discussion that we included in the Proposing Release to help registrants understand our views on how the final rule should be applied.

        ---------------------------------------------------------------------------

        \337\ See, e.g., letters from Hyster-Yale, IBC, NACCO, NACD, and PM&P.

        ---------------------------------------------------------------------------

        In calculating the annual total compensation of employees in accordance with Item 402(c)(2)(x), applicable references to ``named executive officer'' in Item 402 and the related instructions are deemed in the final rule to refer instead to ``employee,'' as proposed. Also, the final rule clarifies that, for non-salaried employees, references to ``base salary'' and ``salary'' in Item 402 are deemed to refer instead, as applicable, to ``wages plus overtime.'' We are adopting this provision to help registrants calculate the total compensation for a median employee that happens to be non-salaried.

        Additionally, registrants may use reasonable estimates, as described above, in determining an amount that reasonably approximates the aggregate change in actuarial present value of an employee's defined pension benefit for purposes of Item 402(c)(2)(viii), as we stated in the Proposing Release. For example, in the case of pension benefits provided to union members in connection with a multi-employer defined benefit pension plan, the participating employers typically do not have access to information (or may not have access in the timeframe needed to compile pay ratio disclosure) from the plan administrator that would be needed to calculate the aggregate change in actuarial present value of the accumulated benefit of a particular individual under the plan.\338\ In such circumstances, we believe it would be appropriate for a registrant to use reasonable estimates in determining an amount that reasonably approximates the aggregate change in actuarial present value of an employee's defined pension benefit for purposes of Item 402(c)(2)(viii).

        ---------------------------------------------------------------------------

        \338\ Section 101(k) and related regulations under the Employee Retirement Income Security Act of 1974, as amended 21 U.S.C. 1021(k), govern the requirements for plan administrators to provide actuarial reports relating to the plan. Under the rules, a plan administrator has thirty days to respond to a request for an actuarial report, and it is not required to provide access to any reports that have not been its possession for more than thirty days. In addition, the rules prohibit the disclosure of reports that include information that the plan administrator reasonably determines to be personally identifiable information regarding a plan participant, beneficiary or contributing employer. See 29 CFR 2520.101-6.

        ---------------------------------------------------------------------------

        The instructions to Item 402(c)(2)(ix) permit the exclusion of personal benefits as long as the total value for the

        Page 50140

        employee is less than $10,000, based on the basis of the aggregate incremental cost to the registrant.\339\ In calculating any such amounts for purposes of calculating the annual total compensation of employees other than the PEO, a registrant may use reasonable estimates in the manner described above, as proposed. In light of concerns about the difficulty and complexity in the valuation of government-mandated pension plans, we acknowledged in the proposing release that some registrants might need clarity as to how to treat government-mandated pension plans for purposes of calculating an employee's total compensation and, specifically, for purposes of determining the aggregate change in actuarial present value of defined pension benefits under Item 402(c)(2)(viii). Item 402(c)(2)(viii) applies to a defined benefit plan, which, as explained in the 2006 Adopting Release, is a retirement plan in which the company pays the executive specified amounts at retirement that are not tied to the investment performance of the contributions that fund the plan.\340\ In contrast, under many government-mandated pension plans, the employee ultimately receives the pension benefit payment from the government, not the employer, and the purpose of the mandated pension benefit is not to provide compensation to the employee for services performed for the employer.\341\ Notwithstanding any amounts that an employer may be obligated to pay (typically as a tax) to the government in respect of an employee or amounts the employee may be obligated to have withheld from wages and paid to the government, where a pension benefit is being provided to the employee from the government and not by the registrant, a government-mandated defined benefit pension plan should not be considered a ``defined benefit plan'' for purposes of Item 402(c)(2)(viii) and any accrued pension benefit under such a plan should not be considered compensation for purposes of Item 402(c)(2)(x).

        ---------------------------------------------------------------------------

        \339\ See Instruction 4 to Item 402(c)(2)(ix). This instruction would apply to perquisites and personal benefits. Accordingly, perquisites provided to executive officers who are included in the identification of the median would be treated as set forth in Instruction 4. For this purpose, however, benefits that were provided to all employees or all salaried employees would not have been considered ``perquisites.''

        \340\ See 2006 Adopting Release at 53175. This definition serves to distinguish defined benefit pension plans from defined contribution plans, in which the amount payable at retirement is tied to the performance of the contributions that fund the plan.

        \341\ Although Item 402(a)(2) includes compensation transactions between a registrant and a third party where the purpose of the transaction is to furnish compensation to the employee, we generally would not consider a government-mandated pension plan to be such a transaction.

        ---------------------------------------------------------------------------

        Some commenters expressed concern over including pension benefits in calculating total compensation.\342\ One commenter indicated that it would be difficult to use Item 402(c)(2)(x) in calculating ``total compensation'' due to the inclusion of pension benefits because the actuarial value of pensions vacillate dramatically from year to year, which could significantly impact total compensation.\343\ Another commenter requested specifically that we include guidance about excluding government-mandated pension plans.\344\ In response to these comments, we are clarifying that, in calculating ``total compensation,'' registrants may exclude government-related pension benefits for non-U.S. employees, just as Social Security benefits are excluded under ``total compensation'' for U.S. employees. Ultimately, as to both concerns, we believe that the ability to use reasonable estimates should help with calculating total compensation for the median employee who has pension benefits.

        ---------------------------------------------------------------------------

        \342\ See, e.g., letters from ABA and PM&P.

        \343\ See letter from PM&P.

        \344\ See letter from ABA.

        ---------------------------------------------------------------------------

        We acknowledge that the application of the definition of total compensation under Item 402(c)(2)(x) to employees who are not executive officers could understate the overall compensation paid to such employees. Item 402 captures all of the various compensation components received by a named executive officer, excluding certain limited items like benefits under non-discriminatory plans and perquisites and personal benefits that aggregate less than $10,000. By excluding certain benefit plans and perquisites that do not exceed the $10,000 threshold, however, the rules may understate the median employee's actual total compensation. To address this, the final rule permits registrants, at their discretion, to include personal benefits that aggregate less than $10,000 and compensation under non-discriminatory benefit plans in calculating the annual total compensation of the median employee. To be consistent, however, the PEO's total compensation used in the related pay ratio disclosure must also reflect the same approach to these items used for the median employee. The registrant must also explain any difference between the PEO total compensation used in the pay ratio disclosure and the total compensation amounts reflected in the Summary Compensation Table, if material.

        One commenter suggested that, if the final rule allows registrants to identify the median employee only every three years, registrants should similarly be required to calculate total compensation only when a new median is determined or when there is a material change to the annual total compensation figure.\345\ The final rule allows registrants to identify the median employee every three years, but requires total compensation for that employee to be calculated each year. The primary reason for our decision to permit registrants to calculate the median employee every three years is to reduce the costs and burdens to registrants. Based on the comments we received, we understand that much of the cost associated with the proposed rule arises from the task of identifying the median employee.\346\ Several commenters stated that navigating a registrant's payroll systems and creating a single database of all of its employees' compensation, especially non-U.S. employees' compensation, would be the most costly aspect of the proposed rule.\347\ Other activities mentioned by commenters that would contribute to the costs of the proposed rule included data privacy compliance, foreign exchange calculations, data testing, establishing corporate guidelines, obtaining legal services, auditing results, public relations tasks, and litigation risk.\348\ Many of these activities also must be undertaken in identifying the median employee. Once the median employee has been identified, however, it does not appear that calculating the annual total compensation of that one additional employee is a source of significant additional cost. Additionally, since the PEO's compensation will be updated annually, we believe that it is appropriate to have a consistent reflection of that year's compensation both for the PEO and the median employee. Therefore, the final rule requires registrants to calculate the median employee's annual total compensation every year.

        ---------------------------------------------------------------------------

        \345\ See letter from ABA.

        \346\ One commenter stated explicitly that the primary cost associated with the proposed rule would be in identifying the median employee. See letter from McGuireWoods.

        \347\ See, e.g., letters from Aon Hewitt, Avery Dennison, Business Roundtable I, Chamber I, COEC I, Corporate Secretaries, Eaton, FEI, FuelCell Energy, IBC, KBR, NACCO, NAM I, NAM II, and NIRI.

        \348\ See, e.g., letters from Aon Hewitt, Corporate Secretaries, and McGuireWoods.

        ---------------------------------------------------------------------------

        Finally, Section 953(b)(2) states that ``total compensation'' shall be determined in accordance with Item 402(c)(2)(x) ``as in effect on the day before the date of enactment of this

        Page 50141

        Act.'' One commenter suggested that this statement does not preclude any amendment of Regulation S-K subsequent to the passage of the Dodd-

        Frank Act that would alter the definition of ``total compensation'' in Item 402 in effect on the day before the date of enactment of the Act.\349\ In the Proposing Release, we noted that Section 953(b) refers to Item 402(c)(2)(x) in effect on the day before enactment of the Dodd-

        Frank Act, or July 20, 2010. We also indicated that, because no substantive amendments have been made to Item 402(c) since that date, the proposed rule would refer to Item 402(c)(2)(x) without reference to the rules in effect on July 20, 2010. We further stated that we expect to address the impact on the proposed rule of any future amendments to Item 402(c)(2)(x) if and when such future amendments are considered. No substantive amendments have been made to Item 402(c) since July 20, 2010. We continue, therefore, to take the approach articulated in the Proposing Release on this issue.

        ---------------------------------------------------------------------------

        \349\ See letter from Chamber I.

        ---------------------------------------------------------------------------

      3. Disclosure of Methodology, Assumptions, and Estimates

        a. Proposed Rule

        The proposed rule required registrants to briefly describe and consistently apply any methodology used to identify the median and any material assumptions, adjustments, or estimates used to identify the median or to determine total compensation or any elements of total compensation. The proposed rule also provided that, if a registrant changes methodology, material assumptions, adjustments, or estimates from those used in its pay ratio disclosure for the prior fiscal year, and if the effects of any such change are material, the registrant must briefly describe the change and the reasons for the change, and provide an estimate of the impact of the change on the median and the ratio. The proposed rule would not require registrants to provide technical analyses or formulas (such as statistical formulas, confidence levels or the steps used in data analysis).

        b. Comments on the Proposed Rule

        Many commenters indicated that the rule should require registrants to provide narrative information about the methodology and material assumptions, adjustments, or estimates they used in identifying the median or calculating annual total compensation for employees.\350\ Only a few commenters asserted that the rule should not require registrants to provide narrative information.\351\ One of these commenters recommended also that we clarify the nature of the information that we expect registrants to disclose without imposing restrictions on methodologies; state expressly that disclosure is only required if a registrant used material assumptions, adjustments, or estimates; and state that no negative statement is required if a registrant did not use material assumptions, adjustments, or estimates.\352\

        ---------------------------------------------------------------------------

        \350\ See, e.g., letters from AFL-CIO I, Barnard, Business Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT State Treasurer, Domini, Hermes Equity Ownership Services (Nov. 18, 2013) (``Hermes''), Alex Kasner (Nov. 12, 2013) (``Kasner''), LAPFF, Meridian, Microsoft, Somers, Wesley Sze (Nov. 13, 2013) (``Sze''), UAW Trust, US SIF, and WorldatWork I.

        \351\ See, e.g., letters from Prof. Angel, Capital Strategies, and Prof. Ray.

        \352\ See letter from ABA.

        ---------------------------------------------------------------------------

        One commenter expressed concern that the disclosure would not be brief because of the registrant's need to use many estimates and assumptions for the median, especially for non-U.S. employees.\353\ Another commenter cited survey data in which 66% of respondents anticipated that they would feel compelled to provide more than a brief narrative to explain how they determined the pay ratio.\354\ Some commenters did not support requiring any additional narrative disclosure beyond what was already in the proposed rule.\355\ Some of these commenters, as well as a large number of other commenters, asserted that the final rule should permit registrants to provide additional narrative disclosures if they chose to do so.\356\

        ---------------------------------------------------------------------------

        \353\ See letter from Business Roundtable I.

        \354\ See letter from COEC I.

        \355\ See, e.g., letters from ABA, COEC I, Hyster-Yale, Intel, Meridian, NACCO, NYC Comptroller, Vivient, and WorldatWork I.

        \356\ See, e.g., letters from AFL-CIO I, AFSCME, Bacirctirente et al., CBIS, CT State Treasurer, E&Y, First Affirmative, CUPE, FS FTQ, Hyster-Yale, ICCR, IL Bricklayers and Craftworkers Union, Intel, Marco Consulting, McGuireWoods, McMorgan & Co., NYC Comptroller, NACCO, Novara Tesija, NY Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, Rep. Ellison et al. I, Socially Responsive Financial Advisors, Teamsters, Trillium I, Trustee Campbell, Vivient, and Walden.

        ---------------------------------------------------------------------------

        Some commenters noted that the final rule should require registrants to disclose material changes to their assumptions, adjustments, or estimates from previous years, as proposed.\357\ One commenter suggested that the final rule allow a good-faith compliance period of two years in which a registrant can change its initial methodology without having to specifically explain and quantify the change.\358\

        ---------------------------------------------------------------------------

        \357\ See, e.g., letters from ABA, CEG, CT Treasuer, Domini, Kasner, McGuireWoods, and WorldatWork I.

        \358\ See letter from Frederic W. Cook & Co.

        ---------------------------------------------------------------------------

        Finally, a few commenters requested that the final rule require some additional metrics, such as upper and lower quartiles, mean, and standard deviation,\359\ and one commenter suggested that companies voluntarily disclose both the ratio between average employee pay and average executive pay and the ratio of pay between the top and bottom 10% of earners within the company.\360\ Other commenters, however, stated specifically that the final rule should not require the disclosure of any additional metrics.\361\

        ---------------------------------------------------------------------------

        \359\ See, e.g., letters from Prof. Ray and Rebecca Vogel (Nov. 13, 2013) (``R. Vogel'').

        \360\ See letter from LAPFF.

        \361\ See, e.g., letters from Capital Strategies, Johnson & Johnson, and WorldatWork I.

        ---------------------------------------------------------------------------

        c. Final Rule

        The final rule, consistent with the proposal, requires registrants to briefly describe and consistently apply any methodology used to identify the median and any material assumptions, adjustments (including any cost-of-living adjustments), or estimates used to identify the median or to determine total compensation or any elements of total compensation. The final rule also requires a registrant to clearly identify any estimates used. For example, when statistical sampling is used, registrants must describe the size of both the sample and the estimated whole population, any material assumptions used in determining the sample size and the sampling method (or methods) is used. Additionally, although the required descriptions must provide sufficient information for readers to evaluate the appropriateness of the methodologies used, registrants are not required to include any technical analyses, formulas, confidence levels, or the steps used in data analysis.\362\ Although one commenter suggested that the final rule allow a good-faith compliance period of two years in which an issuer may change its initial methodology without having to specifically explain and quantify the change,\363\ the final rule requires registrants to disclose any change in methodology, significant assumption, adjustment, or estimate from the prior year if the effects of any such change are significant. Registrants must also disclose if they changed from using the cost-of-living adjustment to not using that adjustment and if they changed from not using the cost-of-living

        Page 50142

        adjustment to using it. We believe that it is important for shareholders to understand changes to a registrant's methodology so that they may make informed voting decisions on executive compensation under Section 951 of the Dodd-Frank Act.

        ---------------------------------------------------------------------------

        \362\ A registrant, however, must include the measure used as the basis for any cost-of-living adjustments when briefly describing the cost-of-living adjustments it used to identify the median employee and calculate the median employee's annual total compensation.

        \363\ See letter from Frederic W. Cook & Co.

        ---------------------------------------------------------------------------

        We believe that requiring registrants to include the brief overview will make it easier for shareholders to understand the pay ratio disclosure for that company and better evaluate its utility in assessing the compensation and accountability of a registrant's executives, including in making their voting decisions on executive compensation under Section 951. We do not believe that requiring registrants to provide additional metrics, such as a more detailed or technical analysis will help shareholders in this manner. We note that other of our rules require similar disclosures, particularly where registrants are given the flexibility to choose a methodology, such as the valuation method for determining the present value of accrued pension benefits in Item 402(h)(2) or the description of models, assumptions, and parameters in Item 305 of Regulation S-K (quantitative and qualitative disclosures about market risk). Several commenters agreed that the rule should require registrants to provide this narrative description, as proposed, and no additional information.\364\ Further, a number of these commenters indicated that the final rule should clarify that the narrative should be brief.\365\ Consistent with these comments, the final rule specifically states that registrants must ``briefly'' describe this information.

        ---------------------------------------------------------------------------

        \364\ See, e.g., letters from ABA, AFL-CIO I, Barnard, Business Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT State Treasurer, Domini, Hermes, Hyster-Yale, Kasner, Intel, LAPFF, Meridian, Microsoft, NACCO, Comptroller of the City of New York (Nov. 27, 2013) (``New York City Comptroller''), Somers, UAW Trust, US SIF, Vivient, and WorldatWork I.

        \365\ See, e.g., letters from ABA, AFL-CIO I, Business Roundtable I, COEC I, COEC II, Domini, Meridian, Microsoft, and WorldatWork I.

        ---------------------------------------------------------------------------

        We note that some commenters contended that the final rule should require additional metrics,\366\ whereas other commenters stated specifically that the final rule should not require the disclosure of any additional metrics.\367\ As we discussed in the Proposing Release, we are sensitive to the costs of the mandated disclosure, and we believe that narrative disclosure in addition to what is already required about the ratio would not, for many registrants, provide useful information for shareholders. For example, disclosures about employment policies, use of part-time employees, use of seasonal employee workers, and outsourcing and off-shoring strategies are not required under the final rule. The final rule does, however, allow registrants the flexibility to provide those additional disclosures that they believe will assist shareholders' understanding of the meaning of the pay ratio disclosure for their particular circumstances. We believe this approach is preferable to imposing a requirement on all registrants to provide additional metrics that may not be useful in many cases.

        ---------------------------------------------------------------------------

        \366\ See, e.g., letters from Prof. Ray, LAPFF, and R. Vogel.

        \367\ See, e.g., letters from Capital Strategies, Johnson & Johnson, and WorldatWork I.

        ---------------------------------------------------------------------------

      4. Meaning of ``Annual''

        a. Proposed Rule

        The proposed rule would define ``annual total compensation'' to mean total compensation for the last completed fiscal year, consistent with the time period used for the other Item 402 disclosure requirements. This provision was intended to address concerns about the need to update the pay ratio disclosure throughout the year and to make clear that the disclosure does not need to be updated more than once a year. Although we considered other ``annual'' periods that may have reduced compliance costs for registrants by giving them the ability to use information in the form that it is currently compiled for other purposes, we believed it was appropriate for the time period for the pay ratio disclosure to be the same as the time period used for the PEO's compensation. Registrants, therefore, would be required to calculate the total compensation for the median employee for their last completed fiscal year.

        For purposes of identifying the median employee, however, we proposed allowing registrants to use compensation amounts derived from the information derived from tax and/or payroll records for the annual period used in those records. We believed that permitting companies to identify the median employee using compensation information in the form that it is maintained in their own books and records would reduce compliance costs. Registrants using the information derived from tax and/or payroll records to identify the median employee would still be required to calculate the Item 402(c)(2)(x) total compensation for that median employee for the last completed fiscal year, rather than the annual period used in the payroll and/or tax records.

        b. Comments on the Proposed Rule

        Some commenters agreed that the final rule should require the pay ratio to be calculated for the last completed fiscal year, as proposed, rather than some other annual period.\368\ Other commenters, however, contended that the final rule should provide another annual period. Some of these commenters recommended that the final rule permit registrants to use the year prior to the registrant's last completed fiscal year for identifying the median employee, annual total compensation, or both to give the registrants more time to identify the median employee and calculate his or her total compensation.\369\ Using the employee population from the year before, one commenter stated, would not have a material impact on the ratio and the added value of more contemporaneous information would likely be negligible.\370\ Another commenter suggested using this time period for registrants with non-U.S. employees.\371\

        ---------------------------------------------------------------------------

        \368\ See, e.g., letters from ABA, CalPERs and UAW Trust.

        \369\ See, e.g., letters from Aon Hewitt, Business Roundtable I, Corporate Secretaries, and Eaton.

        \370\ See letter from Aon Hewitt.

        \371\ See letter from FSR.

        ---------------------------------------------------------------------------

        Some commenters urged us to adopt a final rule that permitted use of the time periods used for payroll and/or tax records when calculating compensation to identify the median employee and the pay ratio for that employee.\372\ These commenters indicated that the periods could be different across jurisdictions. In this regard, another commenter noted that there is no need to have exact overlap of the time periods because the pay ratio ``won't change all that much.''\373\

        ---------------------------------------------------------------------------

        \372\ See, e.g., letters from Hyster-Yale and NACCO.

        \373\ See letter from Prof. Angel.

        ---------------------------------------------------------------------------

        Other commenters asserted that registrants should be given flexibility to choose any annual period in identifying the median employee and/or that employee's total compensation.\374\ One of these commenters stated, however, that the annual period must substantially relate to the fiscal year for which the pay ratio disclosure is being provided, regardless of whether the last day of such annual period falls before or after the end of the registrant's fiscal year for the purposes of identifying the median employee.\375\ As an example, this commenter stated that, if the registrant has a fiscal year ending on November 30, the registrant should be permitted to identify the median employee based on a compensation

        Page 50143

        measure calculated from January 1 through December 31 of that year, as long as such records substantially relate to the fiscal year for which pay ratio disclosure is being provided.

        ---------------------------------------------------------------------------

        \374\ See, e.g., letters from Davis Polk and WorldatWork I.

        \375\ See letter from Davis Polk.

        ---------------------------------------------------------------------------

        c. Final Rule

        The final rule defines ``annual total compensation'' to mean ``total compensation'' for the registrant's last completed fiscal year, as proposed. Although there were other ``annual'' periods suggested by commenters, such as the year prior to the registrant's last completed fiscal year \376\ or the time periods used for the information derived from tax and/or payroll records,\377\ we believe the registrant's last completed fiscal year is more appropriate. Using the registrant's last completed fiscal year is consistent with the time period used for the other Item 402 disclosure requirements. Registrants are required, therefore, to disclose the ``total compensation,'' using Item 402(c)(2)(x), for their median employee and PEO based on the compensation they provided for these individuals in the last completed fiscal year. We believe that making the time period for the pay ratio disclosure consistent with other Item 402 disclosures will better enable shareholders to use it in conjunction with the other Item 402 disclosures to assess the compensation and accountability of a registrant's executives. For this same reason, we are not permitting registrants to select any annual period or the year prior to the last completed fiscal year to calculate total compensation.

        ---------------------------------------------------------------------------

        \376\ See, e.g., letters from Aon Hewitt, Business Roundtable I, Corporate Secretaries, Eaton, and FSR.

        \377\ See, e.g., letters from Hyster-Yale and NACCO.

        ---------------------------------------------------------------------------

        As discussed above, registrants may use compensation amounts derived from the information derived from their tax and/or payroll records for the same annual period used in those records to identify their median employee because we believe this reduces compliance costs. Registrants using the information derived from tax and/or payroll records to identify the median employee are still required to calculate the Item 402(c)(2)(x) total compensation for that median employee for the registrant's last completed fiscal year, rather than the annual period used in the payroll and/or tax records because identifying the median is a separate process from calculating total compensation.

      5. ``Filed'' Not ``Furnished''

        a. Proposed Rule

        Under the proposal, the pay ratio disclosure would be considered ``filed'' for purposes of the Securities Act and Exchange Act, which is the same as for other Item 402 information.

        b. Comments on the Proposed Rule

        Only one commenter stated explicitly that the pay ratio disclosure should be ``filed,'' as proposed.\378\ This commenter agreed that the information should be filed because Section 953(b) refers to ``filings.'' Further, the commenter stressed that any concerns registrants may have about the pay ratio information being ``filed'' are mitigated by the proposed rule's flexibility.

        ---------------------------------------------------------------------------

        \378\ See letter from US SIF.

        ---------------------------------------------------------------------------

        Commenters that opposed the proposed rule generally indicated that the pay ratio disclosure should be ``furnished'' rather than ``filed.'' \379\ The commenters contending that the pay ratio information should be ``furnished'' argued that, in making the calculations for identifying the median employee and total compensation, registrants will have to review a large amount of data and make a significant number of estimates, assumptions, and judgment calls, which will necessarily lead to imprecision.\380\ Some noted that this imprecision will subject a registrant to potential liability and litigation,\381\ make it difficult to validate the information sufficiently for Sarbanes-Oxley Act \382\ certification purposes,\383\ and/or not permit the information to be audited (or greatly increase the costs of the audits).\384\

        ---------------------------------------------------------------------------

        \379\ See, e.g., letters from AAFA II, ABA, American Benefits Council, Aon Hewitt, Best Buy et al., Bill Barrett Corp., Business Roundtable I, Chamber I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, Eaton, Freeport-McMoRan, General Mills, Intel, Mercer I, NAM I, NIRI, NRF, PM&P, RILA, SHRM, and Vectren Corp.

        \380\ See, e.g., letters from AAFA II, ABA, Business Roundtable I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, Eaton, General Mills, NRF, RILA, and Vectren Corp.

        \381\ See, e.g., letters from ABA, American Benefits Council, Aon Hewitt, Bill Barrett Corp., Chamber I, General Mills, Mercer I, and PM&P.

        \382\ See Pub. L. 107-204, 116 Stat. 745 (2002).

        \383\ See, e.g., letters from ABA, Best Buy et al., Corporate Secretaries, Freeport-McMoRan, Intel, NAM I, NAM II, and SHRM.

        \384\ See, e.g., letters from COEC I, COEC II, Corporate Secretaries, and NIRI.

        ---------------------------------------------------------------------------

        Some commenters asserted that use of the word ``filing'' in Section 953(b) does not demonstrate a Congressional desire that the disclosure be ``filed'' rather than ``furnished.'' Some of these commenters pointed out that the statutory language refers to information to be included in ``filings'' rather than requiring the information to be ``filed.'' \385\ Also, the commenters noted that there is some information in our ``filings'' that is ``furnished,'' such as Items 2.02 and 7.01 of Form 8-K, the glossy annual reports to shareholders, the audit committee reports (Item 407(d)), the stock performance graphs (Item 2.01(e)), the compensation committee reports (Item 407(e)(5)), and that executive compensation information in ``filings'' was ``furnished'' until 2006.\386\

        ---------------------------------------------------------------------------

        \385\ See, e.g., letters from ABA, COEC I, COEC II, Corporate Secretaries, PNC Financial Services, and RILA.

        \386\ See, e.g., letters from ABA, Business Roundtable I, COEC I, General Mills, and Mercer I.

        ---------------------------------------------------------------------------

        One commenter recommended that, in the event that the pay ratio disclosure must be ``filed,'' we consider providing a ``safe harbor'' excluding the disclosure from the portion of a registrant's filings that must be certified pursuant to Exchange Act Rules 13a-14 and 15d-14 and are also subject to Section 906 of the Sarbanes-Oxley Act of 2002.\387\ Additionally, another commenter recommended that, at least initially, we make the pay ratio disclosure an addendum to documents required under Regulation S-K and have that addendum deemed ``furnished.'' \388\ This commenter indicated that this approach could minimize some of the rule's costs and burdens.

        ---------------------------------------------------------------------------

        \387\ See letter from ABA.

        \388\ See letter from Chamber I.

        ---------------------------------------------------------------------------

        c. Final Rule

        The final rule treats the pay ratio disclosure, as with other Item 402 information, as ``filed'' for purposes of the Securities Act and Exchange Act, and, therefore, subject to potential liabilities under those statutes, including Exchange Act Section 18 liability.\389\ Information required to be disclosed by registrants pursuant to the federal securities laws generally is filed with us and subject to the liabilities thereunder, unless a specific exception applies. Although we recognize that identifying the median employee and calculating total compensation may require registrants to review a large amount of data and make a significant number of estimates, assumptions, and judgment calls, we do not believe this fact alone justifies exempting this information from being ``filed.'' Many of the disclosures required by our rules require complex calculations and estimates. Moreover, the fact that registrants will be required to provide disclosure about how they have arrived at their pay ratio calculations, and in particular the required disclosure about the assumptions and methodologies underlying the calculations, will permit registrants to clearly explain to shareholders where potential imprecisions may be introduced into the reported statistic. We also note that all

        Page 50144

        other Item 402 information is considered ``filed'' rather than ``furnished'' and that the disclosure called for by Item 402(u)--

        information pertaining to the registrant's operations and workforce composition--differs from the types of information we typically permit to be ``furnished'' rather than ``filed.'' \390\ For similar reasons, we do not believe these disclosures should be exempted from the certification requirements of Exchange Act Rules 13a-14 and 15d-14 or Section 906 of the Sarbanes-Oxley Act of 2002 or be provided as an addendum to filings referenced in Regulation S-K.

        ---------------------------------------------------------------------------

        \389\ 15 U.S.C. 78r.

        \390\ Some examples of information that are ``furnished'' include: the Results of Operations and Financial Condition information (Item 2.02 of Form 8-K); Regulation FD disclosures (Item 7.01 of Form 8-K); the Stock Performance Graph (Item 201(e) of Regulation S-K); the Audit Committee Report (Item 407(d) of Regulation S-K); the Compensation Committee Report (Item 407(e)(5) of Regulation S-K); and the annual reports to shareholders (Rules 14a-3(b) and 14c-3(a) under the Exchange Act and General Instruction G(2) to Form 10-K).

        ---------------------------------------------------------------------------

        In addition, we note that Section 18 of the Exchange Act does not create strict liability for ``filed'' information. Rather, it states that a person shall not be liable for misleading statements in a filed document if it can establish that it acted in good faith and had no knowledge that the statement was false or misleading.\391\ A plaintiff asserting a claim under Section 18 would need to meet the elements of the statute to establish a claim, including purchasing or selling a security at a price that was affected by the false or misleading statement in reliance on the misstatement, and damages caused by that reliance. Finally, regardless of whether the information is ``filed'' or ``furnished,'' registrants that fail to comply with the final rule could also be violating Exchange Act Sections 13(a) and 15(d), as applicable, and would also be subject to potential liability under Exchange Act Section 10(b) \392\ and Rule 10b-5,\393\ promulgated thereunder, for any false or misleading material statements in the information disclosed pursuant to the rule.

        ---------------------------------------------------------------------------

        \391\ Exchange Act Section 18(a) provides that any person who shall make or cause to be made any statement in any application, report, or document filed pursuant to this title or any rule or regulation thereunder or any undertaking contained in a registration statement as provided in subsection (d) of section 15 of this title, which statement was at the time and in the light of the circumstances under which it was made false or misleading with respect to any material fact, shall be liable to any person (not knowing that such statement was false or misleading) who, in reliance upon such statement shall have purchased or sold a security at a price which was affected by such statement, for damages caused by such reliance, unless the person sued shall prove that he acted in good faith and had no knowledge that such statement was false or misleading. A person seeking to enforce such liability may sue at law or in equity in any court of competent jurisdiction. In any such suit the court may, in its discretion, require an undertaking for the payment of the costs of such suit, and assess reasonable costs, including reasonable attorneys' fees, against either party litigant.

        \392\ 15 U.S.C. 78j.

        \393\ 17 CFR 240.10b-5

        ---------------------------------------------------------------------------

      6. Timing of Disclosure

        a. Updating Pay Ratio Disclosure for the Last Completed Fiscal Year

        i. Proposed Rule

        The proposed rule would not have required the pay ratio for the registrant's last completed fiscal year to be disclosed until the filing of its annual report on Form 10-K for that fiscal year or, if later, the filing of a definitive proxy or information statement relating to its next annual meeting of shareholders (or written consents in lieu of such a meeting) following the end of such fiscal year. The proposed rule would require pay ratio information to be filed, in any event, not later than 120 days after the end of such fiscal year as provided in General Instruction G(3) of Form 10-K. Also, in any filing a registrant made after the end of its last completed fiscal year and before the filing of such Form 10-K or proxy or information statement, as applicable, a registrant that was subject to the proposed rule for the fiscal year prior to the last completed fiscal year would be permitted to include or incorporate by reference the pay ratio disclosure information for that prior fiscal year. We proposed this provision because, as discussed above, the proposed rule would require annual total compensation amounts used in the ratio to be calculated for the registrant's last completed fiscal year. In addition, pay ratio disclosure would be required in any filing by the registrant that required Item 402 disclosure.

        Although the annual update of the pay ratio was not required to be disclosed until the filing of an annual report for the last completed fiscal year, or if later, the filing of a definitive proxy statement or information statement relating to the registrant's annual meeting of shareholders, this provision would not have altered the requirements for Item 402 disclosure under Item 8 of Schedule 14A in other proxy or information statement filings.

        ii. Comments on the Proposed Rule

        Some commenters generally agreed with the proposed rule's requirement that the pay ratio disclosure be updated no earlier than the filing of a registrant's annual report on Form 10-K or, if later, the filing of a proxy or information statement for the registrant's annual meeting of shareholders (or written consents in lieu of such a meeting), and in any event no later than 120 days after the end of its fiscal year.\394\ One of these commenters agreed with the proposal that all registrants, should be allowed to file their pay ratio disclosure no later than 120 days after fiscal year end by either an amended Form 10-K or Form 8-K because permitting the later filing would allow registrants not subject to the proxy rules to have the same amount of time to file their pay ratio disclosure as filers that are subject to the proxy rules.\395\ One commenter indicated that, although timing of the disclosure is ``not important,'' the information should be required in the registrant's annual report on Form 10-K and permitted in other filings.\396\ Another commenter contended that the information should at least be available in a registrant's proxy statement for the annual meeting so that shareholders may use the information for voting.\397\

        ---------------------------------------------------------------------------

        \394\ See, e.g., letters from ABA, CalPERS, Calvert, CII, E&Y, Trillium II, UAW Trust, US SIF, and Vectren Corp.

        \395\ See letter from ABA.

        \396\ See letter from Capital Strategies.

        \397\ See letter from UAW Trust.

        ---------------------------------------------------------------------------

        One commenter stated that it would not object to the proposed delay because it would not diminish the usefulness of the disclosure to investors.\398\ The commenter, however, noted that the proposed delay still might not provide registrants enough time after the end of the fiscal year for all registrants to calculate and disclose the pay ratio in their annual proxy statement. Therefore, the commenter stated that it ``generally would not object to the rules providing for some additional accommodation to the extent that it does not significantly diminish the usefulness of the disclosure to investors.''\399\

        ---------------------------------------------------------------------------

        \398\ See letter from CII.

        \399\ Id.

        ---------------------------------------------------------------------------

        Other commenters disagreed with the proposed rule's requirement that registrants disclose their pay ratio information on Form 10-K, the proxy or information statement, or 120 days after the end of its fiscal year. Mainly, these commenters believed the requirement would not provide sufficient time for registrants to identify the median employee, calculate total compensation and the pay ratio, and file their information.\400\ Most of these commenters recommended that the final rule permit disclosure on Form 8-K at

        Page 50145

        some other time of the year, including when the information is able to be calculated,\401\ ``within some extended period (such as 180 days after fiscal year end, as is the case for Form 11-Ks and other reports),''\402\ any time during the first five months after fiscal year-end,\403\ before the end of the registrant's second quarter,\404\ and 14 days before the annual meeting of shareholders.\405\

        ---------------------------------------------------------------------------

        \400\ See, e.g., letters from American Benefits Council, Brian Foley & Co., Chesapeake Utilities, Frederic W. Cook & Co., Hyster-

        Yale, Mercer I, NACCO, and PM&P.

        \401\ See, e.g., letters from Chesapeake Utilities, Mercer I, and PM&P.

        \402\ See letter from American Benefits Council.

        \403\ See letter from Brian Foley & Co.

        \404\ See letter from Frederic W. Cook & Co.

        \405\ See, e.g., letters from Hyster-Yale and NACCO.

        ---------------------------------------------------------------------------

        iii. Final Rule

        The final rule does not require registrants to provide the pay ratio disclosure information for the registrant's last completed fiscal year until it files its annual report on Form 10-K for that year or, if later, it files the definitive proxy or information statement relating to its next annual meeting of shareholders (or written consents in lieu of such a meeting). In any event, the final rule requires registrants to file their pay ratio information not later than 120 days after the end of such fiscal year in a manner similar to General Instruction G(3) of Form 10-K.\406\ This requirement is consistent with the proposal. Also, consistent with the proposed rule, a registration statement that incorporates by reference a Form 10-K (or amended Form 10-K) containing all Part III information other than updated pay ratio information could be declared effective before the registrant's definitive proxy or information statement containing updated pay ratio information is filed in accordance with General Instruction G(3).

        ---------------------------------------------------------------------------

        \406\ General Instruction G(3) of Form 10-K permits registrants to incorporate by reference Part III of its Form 10-K, which includes the Item 402 information, from their definitive proxy or information statements filed in connection with the registrant's annual meeting if such definitive proxy or information statements are filed within 120 days after the end of the fiscal year covered by the Form 10-K. If a definitive proxy or information statement is not filed within this 120-day period, Items comprising the Part III information must be filed as part of the Form 10-K, or as an amendment to the Form l0-K, not later than the end of the 120-day period.

        ---------------------------------------------------------------------------

        Additionally, although the annual update is not required to be disclosed until the filing of an annual report for the last completed fiscal year, or if later, the filing of a definitive proxy statement or information statement relating to the registrant's annual meeting of shareholders, as discussed in the Proposing Release, this provision does not alter the requirements for Item 402 disclosure under Item 8 of Schedule 14A in other proxy or information statement filings. For example, if a registrant filed a proxy statement (other than the definitive proxy statement for its annual meeting) that required Item 402 information pursuant to Item 8 of Schedule 14A, the registrant would be required to include or incorporate by reference pay ratio disclosure for the most recent period that had been filed in its Form 10-K or definitive proxy statement for its annual meeting.

        We continue to believe this provision is appropriate for the reasons discussed in the Proposing Release. Without it, a registrant would be required to include its pay ratio disclosure in a filing (such as a registration statement) filed after the end of the prior fiscal year, but before it was able to compile its executive compensation information for that fiscal year, which is usually included in a registrant's proxy statement relating to its annual meeting of shareholders following the end of the fiscal year, which could raise additional incremental costs for registrants that elect to provide executive compensation disclosure in their annual proxy statement rather than their annual report and for registrants that are conducting registered offerings at the beginning of their fiscal year.

        We note that a number of commenters agreed with our approach. In response to other comments stating that our approach will not provide registrants sufficient time to identify the median employee, calculate total compensation and the pay ratio, and file their information, we note that the final rule retains the significant flexibility afforded to registrants in the proposal and includes several additional accommodations intended to reduce the burdens of producing the required disclosure. We believe these provisions will make it feasible for registrants to file their pay ratio disclosure within the timeframes set forth in the final rule. We do not believe it would be appropriate to extend the deadline by which the pay ratio disclosure should be updated in light of its relevance to shareholders in making their voting decisions under Section 951 of the Dodd-Frank Act. If registrants were not required to provide the pay ratio disclosure when they file their annual report on Form 10-K or, if later, the definitive proxy or information statement for their next annual meeting of shareholders (or written consent in lieu of such a meeting), this could result in the disclosure not being presented together with other relevant executive compensation information to which it relates and not being available to inform shareholders as they exercise their say-on-

        pay voting rights, which we understand to be the disclosure's primary purpose. For all of these reasons, we believe the timing requirements in the final rule are reasonable and appropriate.

        b. Omitting Salary or Bonus Information for the PEO in Reliance on Instruction 1 to Item 402(c)(2)(iii) and (iv), and Technical Amendment to Item 5.02(f) of Form 8-K

        i. Proposed Rule

        In cases where a registrant is relying on Instruction 1 to Items 402(c)(2)(iii) and (iv) of Regulation S-K to omit salary or bonus of the PEO that is not calculable until a later date, the proposed rule would permit registrants to omit pay ratio disclosure until those elements of the PEO's total compensation are determined. The proposed rule would also have required registrants relying on that instruction to provide their pay ratio disclosure in the same Form 8-K filing in which the PEO's salary or bonus is disclosed. We proposed a conforming amendment to Item 5.02(f) of Form 8-K to reflect the addition of this pay ratio disclosure requirement. Although a filing is triggered under Item 5.02(f) when the omitted salary or bonus becomes calculable in whole or in part, under the proposed amendment to Form 8-K, the pay ratio information would be required only when the salary or bonus became calculable in whole, which would avoid the need for multiple updates to the pay ratio disclosure until the final total compensation amount for the PEO is known.

        ii. Comments on the Proposed Rule

        Only a few commenters discussed this proposed instruction and they generally agreed with the proposed rule.\407\ One commenter contended that, if the registrant is relying on Instruction 1 to Item 402(c)(2)(iii) and (iv), the final rule should require neither an estimate of compensation nor additional supplemental disclosure prior to the Item 5.02 8-K because it would further dilute the utility of the pay ratio information for shareholders.\408\ One commenter suggested that delaying the pay ratio information under Instruction 1 to Item 402(c)(2)(iii) and (iv) would diminish the information's usefulness, but did not object to the proposed instruction because it would only affect a small number of registrants.\409\

        ---------------------------------------------------------------------------

        \407\ See, e.g., letters from ABA, Capital Strategies, and CII.

        \408\ See letter from ABA.

        \409\ See letter from CII.

        ---------------------------------------------------------------------------

        Page 50146

        iii. Final Rule

        As proposed, the final rule permits registrants to omit pay ratio disclosure until the salary or bonus of their PEO's total compensation is determined in cases in which the registrant is relying on Instruction 1 to Items 402(c)(2)(iii) and (iv) of Regulation S-K \410\ to omit the salary or bonus of the PEO that is not calculable until a later date. Commenters on this provision generally agreed with our approach. The final rule also includes a conforming amendment to Item 5.02(f) of Form 8-K to reflect the addition of this pay ratio disclosure requirement. However, although a filing is triggered under Item 5.02(f) when the PEO's omitted salary or bonus becomes calculable in whole or in part, under the conforming amendment to Form 8-K, the pay ratio information is required only when the salary or bonus become calculable in whole, which avoids the need for multiple updates to the pay ratio disclosure until the final total compensation amount for the PEO is known.

        ---------------------------------------------------------------------------

        \410\ Instruction 1 to Items 402(c)(2)(iii) and (iv) of Regulation S-K, under our existing executive compensation disclosure rules, permits registrants to omit disclosure in the Summary Compensation Table of the salary or bonus of a named executive officer if it is not calculable as of the latest practicable date. In that circumstance, the registrant must include a footnote disclosing that fact and providing the date that the amount is expected to be determined, and the amount must be disclosed at that time by filing a Form 8-K. Item 5.02(f) of Form 8-K sets forth the requirements for the filing of information that was omitted from Item 402 disclosure in accordance with Instruction 1 to Items 402(c)(2)(iii) and (iv), including the requirement to include a new total compensation figure for the named executive officer.

        ---------------------------------------------------------------------------

        The final rule includes an instruction that provides that a registrant relying on Instruction 1 to Items 402(c)(2)(iii) and (iv) with respect to the salary or bonus of the PEO would be required to disclose that the pay ratio disclosure is not calculable until the PEO salary or bonus, as applicable, is determined and disclose the date that the PEO's actual total compensation is expected to be determined. The instruction also requires the registrant to include its pay ratio disclosure in the filing on Form 8-K that includes the omitted salary or bonus information as contemplated by Instruction 1 to Items 402(c)(2)(iii) and (iv).

        We believe, as stated in the Proposing Release, that the potential benefits of the complete and up-to-date pay ratio disclosure could be diminished if the pay ratio were to be calculated using less than the entire amount of the PEO's total compensation for the period and that these potential benefits could justify the potential costs to shareholders of a delay in the timing of the disclosure. For example, in some cases, the amount of compensation that is omitted under Instruction 1 to Items 402(c)(2)(iii) and (iv) could be significant, and, therefore, the pay ratio would be lower if presented using that incomplete compensation amount. Similarly, we believe that the potential benefits of the complete and up-to-date pay ratio disclosure could be diminished if the registrant used the prior year's pay ratio information to calculate an approximate pay ratio for the current year, especially if there is a significant change to the PEO's compensation from the prior year. Also, based on the number of registrants that have historically relied on Instruction 1 to Items 402(c)(2)(iii) and (iv),\411\ we do not expect that the instruction will impact a significant number of registrants each year.

        ---------------------------------------------------------------------------

        \411\ For example, based on a review of EDGAR filings for calendar years 2012 and 2013, we estimate that approximately 11 Forms 8-K are filed pursuant to Item 5.02(f) annually and approximately 90% of these relate to disclosure of PEO compensation.

        ---------------------------------------------------------------------------

        c. Initial Compliance Date

        i. Proposed Rule

        We proposed to require a registrant to comply with proposed Item 402(u) with respect to compensation for the registrant's first fiscal year commencing on or after the effective date of the rule. We also proposed to permit a registrant to omit this initial pay ratio disclosure until the filing of its annual report on Form 10-K for that fiscal year or, if later, the filing of a proxy or information statement for its next annual meeting of shareholders (or written consents in lieu of a meeting) following the end of such year. In any event, the information would be required to be filed not later than 120 days after the end of such fiscal year. We recognized in the Proposing Release that a transition period would likely be needed by large, multinational registrants and any registrants that did not have a centralized, consolidated payroll, benefits, and pension system that captures the information necessary to identify the median. We expected that it would take registrants one full reporting cycle to implement and test any necessary systems.

        ii. Comments on the Proposed Rule

        One commenter disagreed with the initial transition period in the proposed rule on the grounds that further delays in having access to the pay ratio disclosure are not in the best interests of shareholders.\412\ Other commenters contended that the transition period in the Proposing Release would disadvantage registrants with fiscal years that end on or close to the effective date of the final rule and suggested that the transition period be extended until:

        ---------------------------------------------------------------------------

        \412\ See letter from IPS.

        ---------------------------------------------------------------------------

        a registrant's first fiscal year commencing on or after six months following the effective date of the final rule; \413\

        ---------------------------------------------------------------------------

        \413\ See, e.g., letters from American Benefits Council, COEC I, Frederic W. Cook & Co., and Microsoft.

        ---------------------------------------------------------------------------

        a registrant's first fiscal year commencing one year after the effective date of the final rule; \414\

        ---------------------------------------------------------------------------

        \414\ See, e.g., letters from AAFA II and NRF.

        ---------------------------------------------------------------------------

        a registrant's first fiscal year commencing after the second anniversary of the effective date of the final rule (or, alternatively, a registrant's first fiscal year commencing on or after December 15 of the year in which the rule becomes effective); \415\

        ---------------------------------------------------------------------------

        \415\ See letter from ABA.

        ---------------------------------------------------------------------------

        a registrant's first fiscal year commencing on or after the first January 1 after the effective date of the final rule; \416\

        ---------------------------------------------------------------------------

        \416\ See, e.g., letters from Best Buy et al., Corporate Secretaries, General Mills, Meridian, PM&P, and SH&P.

        ---------------------------------------------------------------------------

        a registrant's 2016 fiscal year, if the final rule is adopted in 2014; \417\

        ---------------------------------------------------------------------------

        \417\ See, e.g., letters from Chesapeake Utilities, Intel, and Mercer I.

        ---------------------------------------------------------------------------

        a registrant's 2017 fiscal year; \418\

        ---------------------------------------------------------------------------

        \418\ See, e.g., letters from Hay Group, Hyster-Yale, and NACCO.

        ---------------------------------------------------------------------------

        one year after the Proposing Release's compliance date (i.e., one year after a registrant's first fiscal year commencing on or after the effective date of the final rule); \419\

        ---------------------------------------------------------------------------

        \419\ See letter from RILA.

        ---------------------------------------------------------------------------

        two full years after the effective date of the final rule; \420\ and

        ---------------------------------------------------------------------------

        \420\ See, e.g., letters from Business Roundtable I, Eaton, and SHRM.

        ---------------------------------------------------------------------------

        three years after the effective date of the final rule.\421\

        ---------------------------------------------------------------------------

        \421\ See letter from FSI.

        ---------------------------------------------------------------------------

        One commenter recommended delaying compliance with ``the most onerous parts of this rule,'' \422\ and a further commenter requested that we phase in various requirements of the rule.\423\ Neither of these commenters, however, was more specific as to which parts of the rule to delay or phase-in. One commenter suggested that we include a three-year sunset provision in the final rule.\424\ Other commenters suggested various transition periods for companies with non-U.S. employees. Some commenters requested that, if the final rule includes non-U.S. employees,

        Page 50147

        we permit registrants to exclude non-U.S. employees from the pay ratio for an additional two years.\425\ A few commenters recommended that we extend the transition period for multinational registrants to permit them to begin to comply with the final rules with respect to compensation for their first full fiscal year commencing on or after the second anniversary of the effective date of the final rules (assuming foreign employees are not excluded from the ``median employee'' determination).\426\ One commenter urging a transition period for non-U.S. employees stated that ``a staged implementation would allow companies to design methodologies for pay ratio compliance during the first year and test them on an employee population where data collection is more manageable.'' \427\

        ---------------------------------------------------------------------------

        \422\ See letter from Semtech.

        \423\ See letter from Chamber I.

        \424\ See letter from COEC I.

        \425\ See, e.g., letters from FSR (suggesting that the final rule permit registrants with non-U.S. employees at least two full fiscal years to comply, and providing two suggestions for doing so: (1) providing a ``transition period during which the registrant may report its pay ratio disclosure solely on the basis of the data available in respect of its employees based in the United States;'' or (2) providing a transition period for any registrant with more than a de minimis non-U.S. workforce), FuelCell Energy (requesting that we ``provide an additional two years before companies must include overseas workers in their pay ratio calculations''), Garmin (same), NIRI (same), and Semtech (same).

        \426\ See letter from ABA and American Benefits Council.

        \427\ See letter from COEC I.

        ---------------------------------------------------------------------------

        iii. Final Rule

        The final rule provides that registrants' first reporting period is their first full fiscal year beginning on or after January 1, 2017, instead of on or after the effective date of the rule, as proposed. For example, the reporting period for a company with a fiscal year that ends on December 31 will begin on January 1, 2017.\428\ We believe a transition period is appropriate because, as we noted in the Proposing Release, certain registrants may need additional time to implement systems to compile the disclosure and verify its accuracy.

        ---------------------------------------------------------------------------

        \428\ Approximately 70% of registrants have fiscal years that begin on January 1. We determined this figure based on the number of current reporting companies. There are 8,529 total registrants, and 5,799 of these registrants have a fiscal year end of December 31, which is approximately 68% (5,799/8,529=.67991).

        ---------------------------------------------------------------------------

        We are changing our approach from the proposal because a number of commenters contended that the proposed transition period would be burdensome to registrants, and would particularly disadvantage registrants with fiscal years that end on or close to the effective date of the final rule.\429\ Additionally, a number of these commenters indicated at least another additional year would be required for registrants to establish systems to comply with the final rule.\430\ One commenter claimed, in particular, that registrants would need ``an initial year to establish and test the systems that may be necessary to collect and analyze the data required to identify their median employee and develop the necessary disclosure controls and procedures, and then a second year involving a full reporting cycle to actually put their selected system into operation.'' \431\

        ---------------------------------------------------------------------------

        \429\ See, e.g., letters from AAFA II, ABA, American Benefits Council, Best Buy et al., Business Roundtable I, Chamber I, Chesapeake Utilities, COEC I, Corporate Secretaries, Eaton, Frederic W. Cook & Co., FSI, General Mills, Hay Group, Hyster-Yale, Intel, Mercer I, Meridian, Microsoft, NACCO, NRF, PM&P, RILA, Semtech, SH&P, and SHRM.

        \430\ See, e.g., letters from ABA, Business Roundtable I, Chesapeake Utilities, Eaton, FSI, Hay Group, Hyster-Yale, Intel, Mercer I, NACCO, and SHRM.

        \431\ See letter from ABA.

        ---------------------------------------------------------------------------

        We are not providing an additional transition period or staggered compliance for registrants with non-U.S. employees, as requested by some commenters. We believe that the final rule provides sufficient time for all registrants, including multinationals and those with non-

        U.S. employees, to identify the median employee and calculate annual total compensation for that employee and the PEO. Additionally, we note that the de minimis and foreign privacy law exemptions to the definition of ``employee'' in the final rule may help reduce the burden on such registrants in preparing the necessary disclosure.

        d. Transition Period for New Registrants

        i. Proposed Rule

        The proposed rule would permit new registrants to delay compliance so that the pay ratio disclosure would not be required in a registration statement on Form S-1 \432\ or Form S-11 \433\ for an initial public offering or registration statement on Form 10.\434\ Such registrants would be required to comply with proposed Item 402(u) with respect to compensation for the first fiscal year commencing on or after the date the registrant became subject to the requirements of Section 13(a) or Section 15(d) of the Exchange Act, and the registrant could omit this initial pay ratio disclosure from its filings until the filing of its Form 10-K for such fiscal year or, if later, the filing of a proxy or information statement for its next annual meeting of shareholders (or written consents in lieu of a meeting) following the end of such fiscal year. Similar to the proposed instructions for updating pay ratio disclosure, these proposed instructions also would require that this initial pay ratio disclosure be filed, in any event, as provided in connection with General Instruction G(3) of Form 10-K not later than 120 days after the end of such fiscal year.

        ---------------------------------------------------------------------------

        \432\ 17 CFR 239.11.

        \433\ 17 CFR 239.18.

        \434\ 17 CFR 249.210.

        ---------------------------------------------------------------------------

        ii. Comments on the Proposed Rule

        All the commenters that discussed the topic agreed that new registrants should not be required to provide pay ratio disclosure in their initial registration statements on Form S-1, Form S-11, or Form 10.\435\ A few commenters also agreed that a new public company should not have to provide any pay ratio disclosure until it completes its first full fiscal year as a public company.\436\

        ---------------------------------------------------------------------------

        \435\ See, e.g., letters from ABA, CalPERS, CII, Corporate Secretaries, Hyster-Yale, NACCO and PM&P.

        \436\ See, e.g., letters from ABA and Hyster-Yale.

        ---------------------------------------------------------------------------

        iii. Final Rule

        Similar to the transition period for existing registrants, the final rule provides that a new registrant's first pay ratio disclosure must follow its first full fiscal year beginning after the registrant has (i) been subject to the requirements of Sections 13(a) or 15(d) of the Exchange Act for a period of at least twelve calendar months beginning on or after January 1, 2017 and (ii) filed at least one annual report pursuant to Sections 13(a) or 15(d) of the Exchange Act that does not contain the pay ratio disclosure.\437\

        ---------------------------------------------------------------------------

        \437\ For example, company with a fiscal year ending on December 31 that completes its initial public offering on March 1, 2017 will not be required to include any pay ratio information in its registration statement on Form S-1. The registrant will be first required to include pay ratio disclosure in its Form 10-K for its 2018 fiscal year or its definitive proxy or information statement for its 2019 annual meeting of shareholders, but no later than 120 days following the end of its 2018 fiscal year. The registrant's pay ratio disclosure will be required for its 2018 fiscal year because it filed its registration statement after January 1, 2017 (March 1, 2017), it will have been subject to the requirements of Sections 13(a) or 15(d) of the Exchange Act for a period of at least twelve calendar months (March 1, 2017 to March 1, 2018), and it will have filed at least one annual report pursuant to Sections 13(a) or 15(d) of the Exchange Act (for fiscal year 2017).

        ---------------------------------------------------------------------------

        This change aligns the transition for new registrants with the change we made to the initial transition period for existing registrants. As discussed above, the final rule provides that a registrant's first reporting period is its first full fiscal year beginning on or after January 1, 2017, instead of on or after the effective date of the rule, as proposed. Also, this change is consistent with

        Page 50148

        some commenters' recommendation that new registrants not be required to provide any pay ratio disclosure until they complete their first full fiscal year as a public company.\438\

        ---------------------------------------------------------------------------

        \438\ See, e.g., letters from ABA and Hyster-Yale.

        ---------------------------------------------------------------------------

        Additionally, as proposed, the final rule does not require the pay ratio to be disclosed in a registration statement on Form S-1 or Form S-11 for an initial public offering or an initial registration statement on Form 10. Also, new registrants are permitted to omit their pay ratio disclosure from their filings until after the later of (i) their first full fiscal year beginning on the date they first become subject to the requirements of Section 13(a) or 15(d) of the Exchange Act and (ii) January 1, 2017. All commenters that discussed the topic agreed that new registrants should not be required to provide pay ratio disclosure in their initial registration statements on Form S-1, Form S-11, or Form 10.\439\

        ---------------------------------------------------------------------------

        \439\ See, e.g., letters from ABA, CalPERS, CII, Corporate Secretaries, Hyster-Yale, NACCO and PM&P.

        ---------------------------------------------------------------------------

        We noted in the Proposing Release that shareholders might benefit from pay ratio disclosure in connection with an initial public offering or Exchange Act registration. Even so, we continue to believe that it is appropriate to give companies time to develop any needed systems to compile the disclosure and verify its accuracy. This is particularly so since we believe the primary purpose of the pay ratio disclosure is to provide a useful data point for shareholders in making their voting decisions on executive compensation, including their say-on-pay votes, which is unlikely to occur for those registrants until at least a year after the initial public offering has occurred. The transition period for new registrants is similar to the time frame provided for other registrants to comply with pay ratio disclosure requirements following the effective date of the final rule.

        As we stated in the Proposing Release, we are sensitive to the impact that a rule could have on capital formation. Section 953(b), as amended by the JOBS Act, distinguished between certain newly public companies and all other registrants by providing an exemption for emerging growth companies. We note further that, without a transition period, the incremental time needed to compile pay ratio disclosure could cause companies that are not emerging growth companies to delay an initial public offering, which could have a negative impact on capital formation. In this regard, we assume that companies that are no longer eligible for emerging growth company status are likely to be businesses with more extensive operations or a greater number of employees, which could increase the initial efforts needed to comply with the proposed requirements. We continue to believe that providing a transition period for these newly public companies could mitigate this potential impact on capital formation and will not significantly affect shareholders' ability to assess the compensation and accountability of a registrant's executives.

        e. Additional Transition Periods

        i. Proposed Rule

        We did not propose a transition period for registrants that cease to be smaller reporting companies or emerging growth companies or engage in business combinations and/or acquisitions. We did, however, request comment on whether there should be such transition periods and the appropriate length of time for any such transition period.

        ii. Comments on the Proposed Rule

        One commenter recommended a transition period for registrants that cease to be smaller reporting companies.\440\ The commenter recommended that those registrants not be required to provide their pay ratio disclosure until the first full fiscal year commencing on or after the first anniversary of the end of the fiscal year in which the registrant is no longer a smaller reporting company.

        ---------------------------------------------------------------------------

        \440\ See letter from ABA.

        ---------------------------------------------------------------------------

        Several commenters supported generally a transition period for registrants that engage in a business combination and/or an acquisition to delay including any new employees acquired in the transaction in the acquirer's pay ratio.\441\ Other commenters suggested specific transition periods for such registrants, including: six months after the end of the fiscal year in which the transaction closes; \442\ six or more months after the transaction; \443\ one full fiscal year following the transaction; \444\ the fiscal year beginning 18 months after closing of the transaction; \445\ and three years after the transaction.\446\

        ---------------------------------------------------------------------------

        \441\ See, e.g., letters from Eaton, Hyster-Yale, NAM I, NAM II, and PM&P.

        \442\ See letter from Brian Foley & Co.

        \443\ See id. (providing an alternative to initial recommendation of six months after the end of the fiscal year in which the transaction closes).

        \444\ See, e.g., letters from ABA, Business Roundtable I, Chesapeake Utilities, Frederic W. Cook & Co., NACCO, and NYC Bar.

        \445\ See letter from Microsoft.

        \446\ See letter from Cummins Inc.

        ---------------------------------------------------------------------------

        iii. Final Rule

        In response to comments, the final rule provides that a registrant that ceases to be a smaller reporting company or an emerging growth company will not be required to provide pay ratio disclosure until after the first full fiscal year after exiting such status and not for any fiscal year commencing before January 1, 2017. For example, if a calendar year-end smaller reporting company registrant's public float exceeds $75 million as of the end of its second fiscal quarter in 2017, the registrant will cease to be a smaller reporting company as of the beginning of its fiscal year starting on January 1, 2018.\447\ The registrant, therefore, must include its pay ratio disclosure in its Form 10-K for 2018 or a proxy or information statement for its 2019 annual meeting of shareholders (or written consents in lieu of a meeting) following the end of the 2018 fiscal year, but not later than 120 days after the end of the 2018 fiscal year. We believe that this approach is appropriate because, as commenters noted, smaller reporting companies ``will encounter the same challenges in preparing to comply with the pay ratio disclosure requirement as registrants generally and, therefore, will need time to determine how they will collect the data necessary to identify their median employee and prepare the necessary disclosure,'' but, ``as relatively small entities, these registrants are not likely to need as much time as `regular' (larger) registrants to transition to compliance with the pay ratio disclosure requirement.'' \448\ This new transition period is consistent with the one comment we received on this issue, and we believe it will provide sufficient time for these registrants to prepare their disclosure.

        ---------------------------------------------------------------------------

        \447\ See 17 CFR 229.10(f)(1).

        \448\ See letter from ABA.

        ---------------------------------------------------------------------------

        Similarly, in 2017, if a calendar year-end emerging growth company had total annual gross revenues of $1 billion or more, exceeded the $1 billion threshold in non-convertible debt for the previous 3-year period, has reached the fifth anniversary of the date of the first sale of its common equity securities pursuant to an effective registration statement under the Securities Act, had not issued $1 billion in non-

        convertible debt during the previous 3-year period, or is deemed to be a ``large accelerated filer,'' the registrant will cease to be an emerging growth company at the beginning of its fiscal year starting on January 1, 2018.\449\ The registrant,

        Page 50149

        therefore, will be required to include pay ratio disclosure in its Form 10-K for 2018 or a proxy or information statement for its 2019 annual meeting of shareholders (or written consents in lieu of a meeting) following the end of the 2018 fiscal year, but not later than 120 days after the end of the 2018 fiscal year.

        ---------------------------------------------------------------------------

        \449\ 17 CFR 229.10(a).

        ---------------------------------------------------------------------------

        We have decided to adopt this provision because it is consistent with a commenter's similar recommendation for a transition period when a registrant ceases to be a smaller reporting company. The reasoning for the approach for both types of registrants is similar in that emerging growth companies will need time to determine how they will collect the data necessary to identify their median employee and prepare the necessary disclosure.

        The final rule also permits a registrant that engages in a business combination and/or an acquisition to omit the employees of a newly-

        acquired entity from their pay ratio calculation for the fiscal year in which the business combination or acquisition becomes effective. For example, for a calendar year-end registrant that engages in a business combination and/or acquisition in 2017, the registrant's first period for which it will have to include the newly-acquired employees in the pay ratio disclosure would be fiscal year 2018, with the disclosure included in its Form 10-K for 2018 or a proxy or information statement for their next annual meeting of shareholders (or written consents in lieu of a meeting) following the end of the 2018 fiscal year, but not later than 120 days after the end of such fiscal year.

        A number of commenters recommended a transition period for such registrants. Suggestions for the length of the transition period ranged from six months \450\ to three years.\451\ The transition period being adopted is generally consistent with commenters' suggestions that the disclosure not be required until one full fiscal year after the transaction. We also believe that the final rule's approach will allow a registrant sufficient time to incorporate the payroll, compensation, and/or recordkeeping structures of the newly-acquired entity into the registrant's pay ratio disclosure framework.

        ---------------------------------------------------------------------------

        \450\ See letter from Brian Foley & Co. (suggesting that the final rule allow registrants conducting a merger/acquisition transaction to delay reporting ``until at least 6 months after the end of the fiscal year in which the M&A transaction closes, or 6 or more months after the closing date'').

        \451\ See letter from Cummins Inc.

        ---------------------------------------------------------------------------

        Finally, under the provision in the final rule for triennial calculations of median employee compensation discussed above, in the year of the acquisition or business combination, the registrant need not evaluate whether the acquisition or business combination would result in a substantial change to its pay ratio disclosure that would necessitate the re-identification of the median employee. Rather, consistent with the one year transition for incorporating the new employees in the pay ratio disclosure, the first time the registrant must evaluate whether the business combination or acquisition would result in a substantial change to its pay ratio disclosure that would necessitate a re-identification of the median employee is in the fiscal year following the acquisition or business combination. We believe this will provide registrants sufficient time to integrate the new business or acquisition. Nevertheless, those registrants must identify the acquired business excluded and the approximate number of employees for the fiscal year in which the business combination or an acquisition becomes effective to provide transparency about what the pay ratio disclosure does and does not include.

  9. Economic Analysis

    1. Background

      We have performed an economic analysis of the main economic effects that may result from the final rule, relative to the baseline discussed below. Section 2(b) of the Securities Act and Section 3(f) of the Exchange Act require us, when engaging in rulemaking that requires us to consider or determine whether an action is necessary or appropriate in the public interest, also to consider whether the action will promote efficiency, competition, and capital formation.\452\ Further, Section 23(a)(2) of the Exchange Act requires us, when adopting rules under the Exchange Act, to consider the impact that any new rule will have on competition and to not adopt any rule that would impose a burden on competition that is not necessary or appropriate in furtherance of the purposes of the Exchange Act.\453\

      ---------------------------------------------------------------------------

      \452\ See 15 U.S.C. 77b(b) and15 U.S.C. 78c(f).

      \453\ See 15 U.S.C. 78w(a)(2).

      ---------------------------------------------------------------------------

      As discussed above, Section 953(b) of the Dodd-Frank Act directs us to amend Item 402 of Regulation S-K to add a pay ratio disclosure requirement. Section 953(b) imposes a new requirement on registrants to disclose the median of the annual total compensation of all employees and the ratio of that median to the annual total compensation of the CEO. In doing so, Section 953(b) requires registrants to determine total compensation in accordance with Item 402(c)(2)(x). Our rules for compensation disclosure generally focus on compensation matters that relate to executive officers and directors. While registrants subject to Item 402 are required to provide extensive information about the compensation of the PEO and other named executive officers identified pursuant to Item 402(a), current disclosure rules do not require registrants to disclose compensation information for other employees in their filings with us.

      As directed by Congress, we proposed amendments to Item 402 to require the disclosure of the annual total compensation of a registrant's PEO, the median annual total compensation of all employees of that registrant (excluding the PEO), and the ratio of the median annual total compensation of all employees to the annual total compensation of the PEO. We considered the statutory language and exercised our discretion to develop a proposal designed to lower compliance costs while remaining consistent with the mandate of Section 953(b). In particular, among other things, we proposed a rule that would permit registrants to use reasonable estimates to identify the median employee, including by using statistical sampling, and a consistently applied compensation measure (such as payroll or tax records). The proposed rule would also allow the use of reasonable estimates in calculating the annual total compensation or any elements of total compensation for employees. The proposed flexible approach was aimed at decreasing compliance costs while taking into consideration a registrant's particular facts and circumstances. We received thousands of comment letters in response to the proposal.

      To satisfy the statutory mandate of Section 953(b), we are adopting amendments to Item 402 substantially as proposed, with modifications intended to address some of the concerns raised by commenters and provide further flexibility in the determination of the pay ratio. We believe the primary benefit that Congress intended with pay ratio disclosure is to provide shareholders with a company-specific metric that they can use to inform their voting decisions regarding executive compensation under Section 951 of the Dodd-Frank Act. Several commenters stated affirmatively that they would find the new data points, including pay ratio disclosure, relevant and useful when

      Page 50150

      making voting decisions.\454\ As discussed above, while neither the statute nor the related legislative history directly states the objectives or intended benefits of the provision, we believe, based on our analysis of the statute and comments received, that Section 953(b) was intended to provide shareholders with a company-specific metric that can assist in their evaluation of a registrant's executive compensation practices.

      ---------------------------------------------------------------------------

      \454\ See, e.g., letters from AFL-CIO I, AFSCME, Amalgamated, Bricklayers International, CalSTRS, Calvert, Chevy Chase Trust, CorpGov.net, Form Letter C, Form Letter D, Form Letter E, Form Letter F, LIUNA, LAPFF, NY State Comptroller, Pax World Funds, Public Citizen I, Rep. Ellison et al. I, Rep. Ellison et al. II, Trillium I, Trillium II, UAW Trust, and US SIF.

      ---------------------------------------------------------------------------

      We are sensitive to the costs and benefits that stem from the final rule. Some of the costs and benefits stem directly from the statutory mandate in Section 953(b), while others are affected by the discretion we exercise in implementing that mandate. Our economic analysis of the final rule addresses both the costs and benefits that stem directly from the mandate of Section 953(b) and those arising from the policy choices made using our discretion, recognizing that it may be difficult to separate the discretionary aspects of the rule from those elements required by statute.

      In the economic analysis that follows, we first examine the current regulatory and economic landscape to form a baseline for our analysis. We then analyze the likely economic effects--including benefits and costs and impact on efficiency, competition, and capital formation--

      arising from the new mandatory disclosure requirement prescribed by the Dodd-Frank Act and from the choices we have made in exercising our discretion, relative to the baseline discussed below.

    2. Baseline

      To assess the economic impact of the final rule, we are using as our baseline the current state of the market without a requirement for registrants to disclose pay ratio information. At present, registrants that are required to comply with Item 402(c) of Regulation S-K provide disclosure of their PEO's compensation as Section 953(b) requires. Other registrants, such as emerging growth companies, smaller reporting companies, foreign private issuers, and MJDS filers, are not required to comply with Item 402(c).\455\ We do not expect that many registrants, if any, currently maintain payroll and information systems that track total compensation as determined pursuant to Item 402 for all their employees, or make that information publicly available. Some registrants have reported ratios of CEO compensation to employee pay.\456\ We note, however, that the voluntarily reported information is different from the elements covered by Item 402(c)(2)(x), and therefore is not identical to what would be required under the final rule.

      ---------------------------------------------------------------------------

      \455\ These registrants are required to provide disclosure of executive compensation, but the disclosure requirements for these registrants do not fall under Item 402(c)(2)(x).

      \456\ For example, Noble Energy Inc. voluntarily disclosed that ``our Chairman and CEO's total annual direct compensation was approximately 85 times that of the median annual total direct compensation of all of our other employees'' in their 2014 proxy filing and disclosed a pay ratio in their 2015 proxy filing. MBIA Inc. provided voluntary disclosure about average and median salary and bonus for all employees other than the NEOs, and compared them to the CEO's salary and bonus in its 2011 and 2012 proxy filings. NorthWestern Corp. disclosed in its 2011 proxy filing that its CEO compensation (salary, annual incentive and long-term incentive) was ``18 times the median pay of all our employees'' and disclosed a pay ratio of 19-24 in its 2012-2015 proxy filings. Whole Foods Market Inc. disclosed in its filings a ``salary cap'' on executive cash compensation based on a multiple of the employee ``average annual wage''. Other examples of registrants that disclosed a pay ratio or median employee pay in its proxy filings include Advanced Environmental Recycling Technologies Inc., First Real Estate Investment Trust of New Jersey, Inter Parfums Inc., Itex Corporation, and Penn Virginia Corp. See also Simpson Thacher survey of pay ratio disclosures, available at http://www.compensationstandards.com/member/memos/firms/Simpson/03_15_ratio.pdf. We note that the pay ratio in these voluntary disclosures may differ from the pay ratio required to be disclosed in the final rule. In addition, registrants that currently disclose pay ratio are not necessarily the same registrants subject to the final rule.

      ---------------------------------------------------------------------------

      Currently, shareholders cannot calculate registrant-specific median employee compensation or the ratio of the PEO compensation to median employee compensation because there are no existing publicly available sources for this data. In the absence of such data, researchers have approximated the ratio using other available data, such as average employee pay.\457\ Statistics on the median earnings of U.S. workers in various ``industries'' are publicly available from the BLS,\458\ enabling shareholders to approximate the ratio using the industry median employee compensation and the information about PEO compensation for those registrants subject to Item 402(c).\459\ The distribution of the ratios of CEO to industry median employee compensation for a sample of large reporting companies is reported by NAICS industry sectors in the figure below for fiscal year 2013.\460\ Using this data, it is possible, for example, to determine that, for the median manufacturing firm with available data, CEO pay was approximately 105 times industry median employee pay. The 25th-75th percentile range for manufacturing firms was approximately 51-195.

      ---------------------------------------------------------------------------

      \457\ In a working paper entitled ``The CEO-Employee Pay Ratio,'' Dr. Steven Crawford finds that during the 1995-2012 period the ratio of CEO compensation to the average employee pay at U.S. commercial banks was on average 16.6 (with a median of 8.4). See ``The CEO-Employee Pay Ratio'' available at http://ssrn.com/abstract=2529112.

      \458\ See BLS Occupational Employment Statistics available at http://www.bls.gov/oes/.

      \459\ See letter from Prof. Angel (``CEO compensation is already disclosed for public, but not private companies. The only new information is the pay of the median employee. However, there is already pretty good information about median compensation in various industries. For example, a few seconds of Googling leads to http://www.bls.gov/oes/current/oessrci.htm.'').

      \460\ The ratios in the figure are calculated for each registrant with executive compensation data from the Standard and Poor's Compustat Executive Compensation database which tracks compensation for the companies currently or previously in the S&P 1500 index and industry median employee wage information at each 3-

      digit NAICS level from the BLS as of May 2014. The data in the Compustat Executive Compensation database is for fiscal year 2013, which is the most recent fiscal year with complete coverage at the time of this analysis. The distribution of the registrant-level ratios within each NAICS industry sector (2-digit) is represented using horizontal box plots that show the minimum and maximum, and 25th, 50th (median), and 75th percentiles.

      ---------------------------------------------------------------------------

      Page 50151

      GRAPHIC TIFF OMITTED TR18AU15.003

      We caution that any pay ratio estimate that can be made with currently available information would be different from the ratio required under the final rule. The above example uses the BLS median wage information of U.S. workers within the same 3-digit NAICS industries, while the final rule mandates registrants to use registrant-specific information about median employee compensation for ``all employees,'' including employees in workplaces outside the U.S., subject to certain exemptions. Also, the example is based on only wages and does not consider other forms of compensation for employees other than PEOs because the BLS does not report those components for detailed industry definitions. In contrast, the final rule requires registrants to present the ratio using ``total compensation,'' which includes all forms of compensation in Item 402(c)(2)(x). Thus, while existing public data may permit shareholders to estimate median pay ratios across industry sectors, it does not allow for the particularized, registrant-

      specific assessment that, in our view, Section 953(b) was intended to facilitate.

      To assess the economic effects of the final rule, we consider its impact on shareholders, registrants subject to the pay ratio disclosure, and all registrants' employees, including executive officers. We estimate that the final rule applies to approximately 3,571 registrants.\461\

      ---------------------------------------------------------------------------

      \461\ Based on a review of EDGAR filings in calendar year 2014, we estimate that there were approximately 7,619 annual reports on Form 10-K filed in that year with available Xtensible Business Reporting Language (XBRL) data tags (available at http://www.sec.gov/dera/data/financial-statement-data-sets.html). From this number we subtracted the annual reports filed by approximately 678 emerging growth companies (EGCs), 2,958 smaller reporting companies (SRCs), and 412 ABS issuers. These ABS issuers typically omit executive compensation disclosures in accordance with General Instruction J to Form 10-K. To the extent that the number of EGCs is growing each year, we might be underestimating the number of registered EGCs because we look only at registrants that file an annual report on Form 10-K. Registrants can fall into multiple categories among emerging growth companies, smaller reporting companies, and foreign private issuers. For example, 371 smaller reporting companies self-identified also as emerging growth companies. Therefore, we did not include these 371 registrants in the 2,958 smaller reporting companies that we subtracted from the 7,619 registrants that file annual reports on Form 10-K because they were already included as emerging growth companies. Foreign private issuers and MJDS filers that file annual reports on Form 20-F and Form 40-F, respectively, are not required to provide Item 402 information. They are therefore not included in the 3,571 affected registrants estimated above.

      ---------------------------------------------------------------------------

      Important potential determinants of the economic effects of the pay ratio disclosure requirements on the affected registrants are the differences in size, nature, and location of the workforce; complexity of the organization; and the degree of integration of payroll systems that are likely to exist among these registrants. In particular, the number of business and/or geographic segments within a particular registrant can significantly affect the compliance costs associated with the final rule. The registrants that operate in different geographic and business segments will likely have a less homogeneous workforce and are also less likely to maintain a single centralized payroll system.\462\ The average number of geographic and business segments and employees per each segment disclosed by some of the potentially affected

      Page 50152

      registrants in the calendar year 2014 are reported in the table below.\463\

      ---------------------------------------------------------------------------

      \462\ See letter from Chamber II (reporting that 39 companies that conduct operations in more than 50 countries with an average of 90 different ``employee data systems'' worldwide would have average estimated labor costs of $311,800 for first year compliance. In contrast, 37 companies that operate in fewer than 10 countries with an average of 4.2 employee data systems would have average estimated labor costs of $67,200 for first year compliance (according to the 25 firms that provided this data)).

      \463\ The corporate segments data used in the table come from the Standard and Poor's Compustat Segments database for companies with a business or geographic segment listed under ``segment type''. Segment information is self-reported by companies. As such, it is not based on standardized definitions of lines-of-business and geographic areas. The database provides some geographic segment information for approximately 63% of the potentially affected registrants and some business segment information for approximately 68% of the potentially affected registrants.

      Table 1--Characteristics of Registrants With Segments Data

      ----------------------------------------------------------------------------------------------------------------

      Number of

      Average Min Median Max registrants

      ----------------------------------------------------------------------------------------------------------------

      Total Assets ($ millions)....... 13,250 1 1,840 3,248,176 2,681

      Number of Employees per 15,453 0 2,978 537,000 1,575

      Registrant.....................

      Number of Geographic Segments... 3.11 1 2 31 2,263

      Geographic Segment Assets ($ 11,892 3.1 1,512 3,248,176 1,037

      millions)......................

      Number of Employees per 8,742 0 1,522 1,100,000 1,210

      Geographic Segment.............

      Number of Business Segments..... 2.45 1 1 11 2,444

      Business Segment Assets ($ 4,614 1 811 812,044 2,171

      millions)......................

      Number of Employees per Business 7,849 0 1,022 420,000 1,429

      Segment........................

      ----------------------------------------------------------------------------------------------------------------

      Table 1 shows that, in 2014, potentially affected registrants had an average of three geographic segments and two business segments. Also, the average number of employees was approximately 8,700 per geographic segment and 7,800 per business segment. We do not have complete information on how the registrants maintain their payroll systems across multiple geographic and business segments, but we believe that, because it is probable that registrants with multiple geographic and business segments will have multiple payroll systems and therefore lack easily accessible employee-level data on compensation, the number of such segments serves as an indication of the complexity and costs of trying to comply with the final rule (whether by sampling at each segment and aggregating the samples across the segments or by aggregating the payroll observations and sampling from the aggregated pool). The estimated costs associated with compliance for registrants with multiple geographic and business segments employing multiple payroll systems are discussed below.

      One commenter asserted that the pay ratio disclosure may affect PEO compensation.\464\ If the pay ratio disclosure were to significantly affect PEO compensation, the rule may have adverse effects on registrants' ability to attract and retain PEOs focused on such compensation. We note that there may be other factors affecting the ability of a registrant to attract and retain executive talent, such as the general structure and conditions of the labor market for executives. However, we do not have enough information to assess the effect of the new rule on PEO compensation or on the level of competition in the labor market for PEOs.

      ---------------------------------------------------------------------------

      \464\ See letter from Lou (``In spite of the fact that the pay ratio does not necessarily lead to CEO pay cuts, some companies may decrease executive pay if the number is too embarrassing. But this productivity-unrelated deduction can artificially depress the U.S. CEO market and IPOs.'').

      ---------------------------------------------------------------------------

      Relative to the baseline discussed above, the economic analysis that follows focuses initially on the likely economic effects--

      including benefits and costs and impact on efficiency, competition, and capital formation--arising from the new mandatory disclosure requirement prescribed by the Dodd-Frank Act, and it then focuses on those that arise from the choices we have made in exercising our discretion.\465\

      ---------------------------------------------------------------------------

      \465\ Although we have divided our discussion of the economic effects of the rule between mandatory and discretionary features, we do not mean to imply that Congress unambiguously compelled us to adopt all of the items discussed under the mandatory requirements discussion. Specifically, we recognize that we retain exemptive authority and interpretive authority over many aspects of the rule, including many of those that we discuss within the mandatory requirements section. Generally speaking, we have chosen to classify items as mandatory because, in our view, these particular items appear to us to be consistent with the Congressional intent underlying Section 953(b).

      ---------------------------------------------------------------------------

    3. Economic Effects From Mandated Disclosure Requirements

      1. Benefits

        The following discussion is mainly intended to address benefits of the mandated disclosure to shareholders and shareholders of the registrants that are subject to the disclosure requirements mandated by Section 953(b).

        Although Congress neither expressly identified in Section 953(b) a specific market failure intended to be addressed by the new disclosure requirement nor expressly stated the specific objectives and intended benefits of Section 953(b), we nonetheless believe that the context in which the provision appears provides useful evidence of Congress' purpose in enacting the provision. As discussed above, we believe that Congress intended Section 953(b) to enhance the executive compensation information available to shareholders. Particularly, Section 953(b) provides new data points that shareholders may find relevant and useful when exercising their voting rights under Section 951. We believe, therefore, that Section 953(b) should be interpreted consonant with Subtitle E's general purpose of further facilitating shareholder engagement in executive compensation decisions. A significant consideration for us in fashioning a final rule implementing Section 953(b), then, is the extent to which elements of the final rule further Congress' apparent goal of giving shareholders additional executive compensation information to use as part of the shareholder engagement envisioned by Section 951.

        Moreover, as discussed earlier, a number of commenters stated that they would find the pay ratio disclosure relevant when making voting decisions.\466\ We acknowledge the views of these commenters and regard the informational benefit of facilitating shareholder engagement in executive compensation decisions as potentially a significant new benefit to shareholders when they exercise their say-on-pay voting rights. We note that registrants have not historically been required to provide shareholders with access to information that would allow them to

        Page 50153

        assess the level of a PEO's compensation as it compares to employees at the same registrant and, as a result, shareholders generally have not been provided such information.

        ---------------------------------------------------------------------------

        \466\ See, e.g., letters from AFL-CIO I, AFSCME, Amalgamated, Bricklayers International, CalSTRS, Calvert, Chevy Chase Trust, CorpGov.net, Form Letter C, Form Letter D, Form Letter E, Form Letter F, LAPFF, LIUNA, NY State Comptroller, Pax World Funds, Public Citizen I, Rep. Ellison et al. I, Rep. Ellison et al. II, Trillium I, Trillium II, UAW Trust, and US SIF.

        ---------------------------------------------------------------------------

        While we believe that the pay ratio disclosure may provide an informational benefit to shareholders in their say-on-pay voting, we are unable to quantify this benefit. This is so for a number of reasons. First, the primary benefit that results from the pay ratio disclosure is not directly tied to an immediate economic transaction, such as the purchase or sale of a security, which makes it difficult for us to quantify in monetary terms the likely benefit to shareholders of this information. Second, the pay ratio disclosure is but one data point among many considerations that shareholders might find relevant when exercising their say-on-pay votes, which also makes it difficult for us to quantify the precise benefit that shareholders may experience. Third, even in situations where the pay ratio may be a significant or dispositive consideration for shareholders, because the say-on-pay vote is advisory and not binding, it is difficult for us to link the disclosure with certainty to a potential change in PEO compensation and even more speculative for us to link the disclosure to an economic outcome at a registrant. Further, we note that no commenter provided us with data that would allow us to quantify the potential benefits nor did any commenter suggest a source of data or a methodology that we could look to in quantifying the rule's potential benefits.

        We also think it is important to observe that, despite our inability to quantify the benefits, Congress has directed us to promulgate this disclosure rule. Thus, we believe it reasonable to rely on Congress's determination that the rule will produce benefits for shareholders and that its costs (which we discuss further below) are necessary and appropriate in furthering shareholders' ability to meaningfully exercise their say-on-pay voting rights. Because Congress expressly directed us to undertake this rulemaking, we do not believe it would be appropriate to second-guess its apparent conclusion that the benefits from this rule justify its adoption. In any event, as noted above, we concur with Congress's judgment that the pay ratio disclosure could be beneficial for shareholders.

        Commenters also suggested that pay ratio disclosure can be a valuable tool in evaluating PEO compensation practices in general.\467\ Among other uses of the pay ratio information, some commenters suggested that comparing the total compensation of the median employee and PEO would assist investors in their ability to evaluate the PEO's compensation in the context of the registrant's overall business,\468\ and could provide insight into the effectiveness of board oversight and sound board governance.\469\ Other commenters noted that they incorporate social and governance issues, like pay equity, as part of their investment decisions.\470\

        ---------------------------------------------------------------------------

        \467\ See, e.g., letters from Alexander, Bricklayers International, CalSTRS, Calvert, Chicago Teachers Fund, First Affirmative, Grossman, Grosvenor Capital, ICCR, IL Bricklayers and Craftworkers Union, Cummings Foundation, LaBruyere, Linton, Marco Consulting, NEI Investments (Dec. 2, 2013) (``NEI Investments I''), NEI Investments (Feb. 28, 2014) (``NEI Investments II''), Novara Tesija, NY Bricklayers and Craftworkers Union, NY State Comptroller, Oxfam, Pax World Funds, Public Citizen I, RPMI, Taylor, Tortora, US SIF, and Walden.

        \468\ See, e.g., letter from UAW Retiree Medical Benefits Trust (Apr. 15, 2011) (``UAW Trust pre-proposal'') (``Disclosure of internal pay equity, whether the ratio between median employee wages and those of the CEO or the ratio between compensation awarded to the CEO and to other top executives, will ultimately help investors evaluate executive pay practices by better contextualizing the information provided to the shareholders through the proxy statement and other corporate filings.'').

        \469\ See, e.g., letters from Allied Value, LLC (``Allied Value''); Kranen; Lynne L. Dallas, Professor of Law, University of San Diego (Dec. 1, 2013) (``Prof. Dallas''); UAW Trust pre-proposal (noting ``we view Section 953(b) as an essential tool that will increase corporate board accountability to investors''); and WA State Investment Board.

        \470\ See, e.g., letters from AFR, CT State Treasurer, Cummings Foundation, and Form Letter B (``I support Dodd-Frank rule 953(b), which strikes me as being all about the intersection of pay equity and investor value.'').

        ---------------------------------------------------------------------------

        As noted above, we recognize that there are significant limitations to using the pay ratio information for comparative purposes in light of the various factors that could affect employee compensation at a particular registrant and the flexibility we are providing. We believe that the informational benefit to shareholders from the final rule is in providing information about a particular registrant's executive compensation.

        In addition to its utility in assessing PEO compensation practices, commenters identified a number of ancillary benefits that may arise from the required pay ratio disclosure. Some commenters suggested that the new disclosure could offset an upward bias in executive compensation resulting from the practice of benchmarking executive pay solely against the compensation of other executives to the extent that current benchmarking practices are inefficient.\471\ Other commenters suggested that a comparison of PEO compensation to employee compensation could be used by shareholders to approximate employee morale and/or productivity \472\ or analyzed as a measure of a particular registrant's approach to managing human capital.\473\ One commenter cited his own research showing that large pay disparities within a corporation contribute to an unethical culture within the corporation.\474\ Finally, some commenters asserted that the registrant-specific information about the median employee compensation may be used to address a broader public policy concern relating to income inequality and income mobility, which they suggest is exacerbated by increasingly high levels of PEO compensation relative to other workers.\475\

        ---------------------------------------------------------------------------

        \471\ See, e.g., letters from AFR, Cummings Foundation, LIUNA, PGGM, and RPMI.

        \472\ See, e.g., letters from AFL-CIO I, AFR, AFSCME, Alexander, Allied Value, Bacirctirente et al., CalSTRS, Calvert, Cummings Foundation, CUPE, Demos I, Domini, Professors Charles M. Elson and Craig K. Ferrere (Dec. 2, 2013) (``Profs. Elson and Ferrere''), Estep, Form Letter A, FS FTQ, Daniel Greenwood (Oct. 26, 2013) (``Greenwood''), Grossman, IPS, LaBruyere, Linton, McMorgan Co., Mudd, NY State Comptroller, Overcott, Pax World Funds, Public Citizen I, Rosati, SEIU, Socially Responsive Financial Advisors, Somers, Taylor, Teamsters, Tortora, Trillium I, Trillium II, Trustee Campbell, US SIF, and Walden.

        \473\ See, e.g., letters from Bricklayers International, Demos (May 30, 2014) (``Demos II''), FS FTQ, MVC Associates, Quintave, OCP, Rep. Ellison et al. I, and UAW Trust.

        \474\ See letter from Prof. Dallas.

        \475\ See letters from Rep. Ellison et al. I; Rep. Ellison et al. II; and Dr. Sue Ravenscroft, Professor of Accounting, Iowa State University (Jun. 18, 2014) (``Prof. Ravenscroft II'').

        ---------------------------------------------------------------------------

        With respect to employee morale and productivity, commenters did not specify what effect a pay ratio disclosure would have on these conditions relative to other environment-specific and registrant-

        specific factors. In particular, the pay ratio disclosure may be significantly dependent on how a registrant structures its business. For example, one registrant might outsource the labor-related (manufacturing) aspects of its business to a third party to focus on product innovation, while another registrant competing in the same industry might choose to retain the labor aspect of its business. To the extent that product innovation requires higher pay than manufacturing, the outsourcing company will have a lower pay ratio for the same PEO pay. Therefore, the potential value of this disclosure for assessing issues related to employee morale, productivity, and investment in human capital may be

        Page 50154

        diminished by the variation in business structures.

        Some commenters suggested that the pay ratio disclosure would promote capital formation.\476\ The main rationale given for this effect is that the new disclosure would help shareholders understand the assets of the firms they invest in or that it will let shareholders choose registrants that invest in their workforce.\477\ On the other hand, some commenters asserted that the rule would discourage capital formation because it would discourage firms from accessing the U.S. capital markets.\478\ We note that the final rule does not apply to emerging growth companies, which conduct the bulk of initial public offerings.\479\ While the pay ratio disclosure could be costly, it is not clear whether it would significantly affect a registrant's ongoing capital raising activity.

        ---------------------------------------------------------------------------

        \476\ See, e.g., letters from Alexander, Capital Strategies, Change to Win (Nov. 25, 2013) (``Change to Win''), CUPE, Greenwood, Grossman, LaBruyere, Mudd, Overcott, Prof. Ravenscroft I, Taylor, and Tortora.

        \477\ See, e.g., letters from Alexander, Greenwood, Grossman, LaBruyere, Mudd, Overcott, Taylor, and Tortora.

        \478\ See, e.g., letters from Prof. Angel, COEC I, Lou, and NIRI.

        \479\ M. Dambra, L. Field and M. Gustafson, The JOBS Act and IPO Volume: Evidence that Disclosure Costs Affect the IPO Decision, J. of Financial Economics. (2014) (documenting that 88% of the U.S. IPOs filed in the 4/1/12-3/31/14 period are eligible for the EGC revenue threshold).

        ---------------------------------------------------------------------------

        Overall, while certain shareholders may use the pay ratio for their investment decisions, it is unclear whether the final rule would impact the capital formation of U.S. capital markets in a significant way. As discussed above, shareholders may be able to approximate the industry level pay ratio using industry employee compensation data from BLS and the information about PEO compensation for registrants subject to Item 402(c). In this regard, adding the pay ratio statistic to the mix of reported financial and operational data may not change the investment decision of investors who access this data. On the other hand, the pay ratio disclosure is company-specific, which adds information not otherwise available to investors.

        In contrast to commenters supporting the required disclosure, some commenters stated that the disclosure mandated by Section 953(b) would not have any benefit, or would not have benefits sufficient to justify the compliance costs, which many of those commenters anticipate would be substantial.\480\ Some of the commenters questioned the materiality of pay ratio information to an investment decision \481\ This view was also asserted by the minority in the Senate report accompanying the legislation.\482\

        ---------------------------------------------------------------------------

        \480\ See, e.g., letters from AAFA I, American Benefits Council, ASA, Brian Foley & Co. Chamber I, Chamber II, NIRI, Tesoro Corp., WorldatWork I, and WorldatWork II.

        \481\ See, e.g., American Benefits Council, ExxonMobil, Prof. Muth, and RILA. Another commenter noted that pay ratio disclosure that includes all employees and is based on a large, global, full- and part-time pool of employees will not be meaningful without substantial explanation. See letter from WorldatWork II.

        \482\ See S. Rep. No. 111-176 (2010) (``Although provisions like this appeal to popular notions that chief executive officer salaries are too high, they do not provide material information to investors who are trying to make a reasoned assessment of how executive compensation levels are set. Existing SEC disclosures already do this.'').

        ---------------------------------------------------------------------------

        While we acknowledge these concerns about the usefulness or materiality of the mandated disclosure, we note that other commenters asserted that certain shareholders incorporate social and governance issues, like pay equity, as part of their decision making.\483\ These shareholders may realize non-economic benefits associated with their decision making based on this type of information. These commenters, however, did not quantify the extent to which shareholders would value pay ratio information or would incorporate the disclosure required by Section 953(b) into their investment or voting decision, if at all. Academic research suggests that the stock market does not fully incorporate employee satisfaction into stock prices.\484\ As mentioned above, because company-specific pay ratio information is not currently reported, it is not possible to assess the usefulness to shareholders of this information as required by Section 953(b) relative to the usefulness of publicly available statistics of median compensation, or the usefulness of any other company-specific metric of employee compensation or satisfaction.

        ---------------------------------------------------------------------------

        \483\ See, e.g., letters from AFR and Form Letter B.

        \484\ Existing research has studied whether there is a correlation between information about employee satisfaction and long-term equity returns in an effort to understand how the market values a public company's intangible assets. This research uses different information than what is provided in the pay ratio disclosure. This research was based on the equity prices of companies that were identified on Fortune Magazine's list of the ``100 Best Companies to Work For in America.'' See A. Edmans, Does the stock market fully value intangibles? Employee satisfaction and equity prices, J. of Financial Economics 101, 621-640 (2011) (finding evidence implying that the market fails to incorporate intangible assets, like employee satisfaction, fully into stock valuations until the intangible subsequently manifests in tangibles, such as earnings, that are valued by the market, and finding evidence suggesting that the non-incorporation of intangibles into stock prices is not simply due to the lack of salient information about them).

        ---------------------------------------------------------------------------

        Some commenters were particularly concerned that the comparisons of pay ratios across registrants may be inappropriate to the extent that registrants employ workers in different countries that have unique compensation practices,\485\ use different methodologies to calculate the median employee,\486\ employ workers with different skill levels,\487\ and have different corporate structures.\488\ As noted above, we believe that the purpose of the pay ratio disclosure is to provide shareholders of a registrant with new data points that they may find relevant and useful when exercising their voting rights under Section 951. As we noted in the Proposing Release, we believe that a variety of factors can potentially limit the comparability of the pay ratio across registrants. We also acknowledge that the final rule we are adopting allows for significant flexibility in determining the pay ratio to address concerns raised by a number of commenters about the potential costs of the pay ratio disclosure. One result of allowing for this flexibility, however, is that the comparability of the pay ratio from registrant to registrant may be further diminished. We recognize this consequence but believe it is justified in light of the cost savings that such flexibility will provide and because we do not regard precise comparability as the primary objective of the final rule.

        ---------------------------------------------------------------------------

        \485\ See, e.g., letters from Avery Dennison, BCIMC, and COEC I.

        \486\ See letter from BCIMC.

        \487\ See letter from COEC I.

        \488\ See letters from Prof. Angel, COEC I, and Tesoro Corp.

        ---------------------------------------------------------------------------

      2. Costs

        a. General

        The following discussion is mainly intended to address costs to registrants that are subject to the pay ratio disclosure. The analysis of costs focuses on direct compliance costs on registrants.

        As discussed above, the final rule permits registrants to choose from several options to identify the median employee. First, registrants can choose to use Item 402(c)(2)(x) to calculate the annual total compensation for each employee and then identify the median employee. Second, registrants can choose a statistical method that is appropriate to the size and structure of their own businesses and the way in which they compensate employees to identify the median employee, and then use Item 402(c)(2)(x) to calculate the median employee's compensation. Third, registrants can use a consistently applied compensation measure, whether with respect to the entire employee population or in conjunction with

        Page 50155

        statistical sampling, to identify the median employee, and then calculate and disclose that median employee's total compensation in accordance with Item 402(c)(2)(x).

        In addition to providing flexibility in identifying the median employee's compensation, the final rule allows flexibility in several other respects. Registrants may:

        Use reasonable estimates when applying Item 402(c)(2)(x) to calculate the annual total compensation for employees other than the PEO, including when disclosing the annual total compensation of the median employee identified using a consistently applied compensation measure;

        identify the median employee every three years to the extent that there is no significant change in the registrant's employee population or employee compensation arrangements;

        consistently choose any date within the last three months of a registrant's fiscal year to identify the median employee.

        Moreover, the final rule allows flexibility for registrants with non-U.S. employees by providing (1) a foreign data privacy law exemption reducing the burden on registrants that operate in certain foreign jurisdictions, and (2) a de minimis exemption that may reduce the number of payroll systems that need to be used to identify the median employee and will allow registrants some flexibility in addressing payroll matters that may result from having employees in multiple jurisdictions. Finally, the final rule also provides that registrants, when determining the compensation of the median employee for purposes of identifying the median employee and making the pay ratio disclosure, may elect to adjust the compensation of their employees in jurisdictions other than the jurisdiction in which the PEO resides to reflect the cost of living in the PEO's country of residence, but they must also provide disclosure of the registrant's pay ratio calculated without the cost-of-living adjustment. Overall, we believe that the flexible approach allowed by the final rule is consistent with Section 953(b) and, in certain circumstances discussed below, may reduce costs compared to other methods of implementing the statute.

        b. Compliance Cost Estimates in Comment Letters

        In the pre-proposing period, we received estimates of the costs of compliance for certain registrants from some commenters.\489\ These estimates varied significantly and were based on the commenters' initial reading and interpretation of the statute and not on the proposed rule, which would allow for flexibility not accounted for in the pre-proposal letters. For example, prior to the proposal, one commenter estimated the cost of compliance with Section 953(b) would be $250,000 to $500,000 annually, and revised its cost estimate downward to $15,000 annually after the proposed rule was released ``primarily due to the ability afforded by the proposed rule for registrants to use a consistently applied compensation measure, such as payroll records or W-2 reportable wages and the equivalents for non-U.S. employees, to identify the median employee.'' \490\

        ---------------------------------------------------------------------------

        \489\ For example, one pre-proposal comment letter from an industry group reported that a member company estimated that it would require approximately $7.6 million and 26 weeks to prepare the pay ratio disclosure and that a separate member company estimated that it would cost approximately $2 million annually to determine the actuarial value of employee pension benefits. See American Benefits Council et al., (Jan. 19, 2012) (``American Benefits Council et al. pre-proposal letter'').

        \490\ See letter from Intel.

        ---------------------------------------------------------------------------

        In the Proposing Release, we did not estimate the costs of the calculation and disclosure of a registrant's pay ratio because we did not have enough data for such estimation. In response to the Proposing Release, a number of commenters evaluated our estimates of the compliance costs represented by the estimated Paperwork Reduction Act (``PRA'') burdens imposed by the proposed rule. Most commenters generally indicated that those PRA burdens underestimated the compliance costs associated with the disclosure requirement, and some provided more specific cost estimates. For example, one commenter noted that our PRA estimate of an average of 340 hours of internal company time in year one to comply with the proposed rule significantly understates the time that many companies would need to comply (especially those with non-U.S. employees).\491\ Below, we discuss the specific comments that we consider to be the most useful to estimate the compliance costs of the pay ratio disclosure. We note that, in providing specific comments, commenters did not typically distinguish between costs derived from the statutory mandate and costs derived from the exercise of our discretion. Furthermore, they typically did not distinguish between internal costs in burden hours versus external professional costs in dollar amounts for PRA purposes.

        ---------------------------------------------------------------------------

        \491\ See letter from NIRI.

        ---------------------------------------------------------------------------

        Two commenters provided survey studies with several relevant estimates of the compliance costs associated with the proposed rule, as well as characteristics of the types of registrants that would be affected. As discussed in greater detail below, the aggregate initial external compliance cost estimates provided by these commenters range between $187 million \492\ and approximately $711 million.\493\ These estimates are based on responses to the surveys discussed below and may not be representative of all registrants affected by the final rule.

        ---------------------------------------------------------------------------

        \492\ See letter from COEC I. We note that this estimate represents only the cost of outside professionals and does not take into account internal company hours as an additional cost of compliance with the rule. This commenter estimated that compliance would require at least 801,000 hours of in-house personnel time (at least 255,000 additional company hours above the 546,000 company hours we estimated) in addition to a cost of $187 million.

        \493\ See letter from Chamber II. This commenter estimated an annual internal compliance burden of 3.6 million hours in addition to an annual cost of $710.9 million.

        ---------------------------------------------------------------------------

        i. Center on Executive Compensation Survey

        One commenter provided the results of a joint survey it conducted among its members.\494\ The results are based on the responses from 128 public companies out of 1,270 surveyed.\495\ Most of the respondents are large registrants, with average revenue of approximately $28 billion; 59% of the registrants had revenues greater than $10 billion. Nearly 80% of respondents had 10,000 or more employees, most of them employed full-time. In addition, nine out of ten respondents had foreign operations with employees located outside the United States. On average, respondents operated in 34 countries, and about two-fifths of their employees worked in foreign countries. The average number of separate employee data systems that respondents had worldwide was 46.

        ---------------------------------------------------------------------------

        \494\ This survey was jointly conducted by the COEC, Human Resource Policy Association, and Corporate Secretaries. See letters from COEC I and Corporate Secretaries. See also letter from Business Roundtable I (referencing the results from the COEC I survey).

        \495\ See letter from COEC I. We note that the letter from Corporate Secretaries refers to results from the same survey, but for the 127 respondents who are also members of Corporate Secretaries.

        ---------------------------------------------------------------------------

        In its letter, the commenter questioned our estimate, for PRA purposes, of $400 per hour for outside professional costs and the estimated PRA hour burden. More than half of the survey respondents indicated that the average hourly fee for their company's

        Page 50156

        external securities compliance counsel is above $700. The respondents also indicated that, on average, 72% of the estimated initial compliance costs are expected to be incurred in subsequent years. Based on these survey results, the commenter asserted that compliance with the rule would require at least 255,000 additional company hours and an additional $114.1 million in costs across all affected registrants for outside professional services above the our PRA burden estimates in the Proposing Release ($72.77 million). Using these updated estimates, the commenter arrived at a total initial compliance cost estimate of at least $186.9 million.\496\ We note that, although labeled ``total compliance costs'' by the commenter, that estimate of compliance costs includes only the cost of outside professionals, and thus is only part of the expected total compliance costs. The estimate does not take into account internal company hours as an additional cost of compliance with the rule. Additionally, the commenter assumed that all affected registrants will bear the same compliance costs, which may bias its total cost estimate because compliance costs are likely to vary between registrants with and without foreign operations, or between small and large registrants.

        ---------------------------------------------------------------------------

        \496\ The commenter reports the additional cost of outside professional services of $114 million in excess of the cost of outside professional services estimated by us in the Proposing Release. The total cost estimate reported as $186.9 million can be obtained by adding $114 million and $72.77 million, with the difference likely due to rounding. See letter from COEC I.

        ---------------------------------------------------------------------------

        The survey provided several estimates of how compliance costs might change if there were certain changes in the rule. For instance, the commenter's letter argued that the final rule should apply only to a registrant's consolidated subsidiaries, noting that its survey indicated that, if the final rule were to include employees of all minority-owned subsidiaries and joint ventures, a registrant's compliance costs would increase by an average of 91%, with a mid-range of 20%. The letter from the other commenter that jointly conducted the survey also presented information about the inclusion of all minority-

        owned subsidiaries and joint ventures, but its letter presented the survey data in a different format. It presented the average and median anticipated increases categorized based on the company's annual revenue.\497\ According to this comment letter, for registrants with annual revenue of over $30 billion, the median increase in cost would be approximately 35% if employees in minority-owned subsidiaries and joint ventures were included. For registrants with annual revenue between $5 billion and $30 billion, that median increase would be approximately 20%, while for registrants with annual revenue below $5 billion, the median increase in compliance cost would be 10%. These numbers, however, appear to reflect an increase in the compliance cost if the coverage of subsidiaries and joint ventures were to be increased from the suggested coverage under the Proposing Release to complete (100%) coverage of subsidiaries and joint ventures. Thus, they do not directly correspond to the changes we made in this release, including the change that we made to include only employees of a registrant's consolidated subsidiaries, as suggested by the commenter.

        ---------------------------------------------------------------------------

        \497\ See letter from Corporate Secretaries.

        ---------------------------------------------------------------------------

        If a registrant were permitted to calculate the pay ratio based on full-time, permanent employees only, then according to the survey responses, the compliance cost would decrease by a mid-range of 10% or an average of approximately 11%.\498\ Another commenter suggested that the median decrease in compliance costs would be approximately 10% or an average of 12% if a registrant were permitted to calculate the pay ratio based on full-time, permanent employees only.\499\ Requiring only U.S. employees to be used when estimating the pay ratio would decrease costs on average by 40%, while the mid-range decrease would be approximately 50%.\500\ In contrast, if the rule did not contain the flexibility allowed under the proposal and instead total compensation as calculated in the Summary Compensation Table was required to be used to identify the median employee, 99% of the respondents said that their cost would increase and 49.1% said that the cost would increase by over 100%. Another commenter, relying on information from the same survey, suggested that the average cost increase would be 4,689% and the median cost increase would be approximately 175% if total compensation as calculated in the Summary Compensation Table was required to be used to identify median employee pay.\501\

        ---------------------------------------------------------------------------

        \498\ See letter from COEC I. While the survey data attached to the comment letter showed that the average anticipated cost reduction would be 11%, the text of the comment letter discussing the survey also stated that for the subset of respondents that anticipated that limiting the rule to full-time employees would lower costs, ``the average savings would be approximately 20 percent.''

        \499\ See letter from Corporate Secretaries.

        \500\ See letter from COEC I (mentioning in the text of the comment letter that the expected decrease is ``47% on average for firms with non-U.S. employees''). See also letter from Corporate Secretaries.

        \501\ See letter from Corporate Secretaries.

        ---------------------------------------------------------------------------

        ii. Chamber of Commerce Survey

        A different commenter also provided estimates of compliance costs of the proposed rule based on survey results.\502\ This commenter's survey is a version of the COEC survey that included only 118 respondents, approximately ``3.1% of all covered businesses.'' The commenter did not elaborate on how its version of the survey is different from the COEC survey, other than including fewer respondents.\503\ The commenter's letter provides no information on the survey's respondent size characteristics to provide context with respect to the respondents' potential organizational complexity and associated challenges in complying with the proposed rule. Based on the survey, the commenter concludes that the average labor cost per company of complying with the proposed rule would be approximately $185,600 for the initial year. That commenter also estimated, but did not monetize, an annual compliance time of 3.6 million hours.\504\ The survey results also show a wide divergence in cost estimates across survey respondents, with 42 respondents estimating the value of the time necessary to comply with the proposal to be at least $100,000, while 13 respondents estimated this value to be less than $10,000. On average, respondents estimated 952 hours needed to comply with the proposed rule. Respondents that conduct operations in foreign countries will have higher compliance costs according to the survey results. Thirty-

        nine respondents that conduct operations in more than 50 countries indicated an average labor cost of $311,800 to comply with the proposed rule. These respondents also reported an average of 90 different employee data systems worldwide. On the other hand, for 37 respondents that operate in fewer than 10 countries, the average compliance cost was estimated to be $67,200. Based on the survey results, the commenter asserted that the total external compliance costs for the private sector could be approximately $711 million and that total cost could increase to $1.1 billion (in addition to the internal compliance time) if every

        Page 50157

        affected registrant has an average cost of $311,800.

        ---------------------------------------------------------------------------

        \502\ See letter from Chamber II.

        \503\ The letter from Chamber II does not specify how many companies were surveyed but the letter indicates that the Chamber represents over ``3 million businesses and organizations of every size, sector and region.''

        \504\ See letter from Chamber II.

        ---------------------------------------------------------------------------

        iii. Other Specific Comments

        In addition to the two surveys, several other commenters provided the following cost estimates based on the proposed rule. In these estimates, the commenters did not distinguish between the costs arising from the mandated disclosure and the costs arising from the exercise of our discretion. The estimates for the proposal were as follows:

        $500,000 to $1 million to automate a large global registrant's processes; \505\

        ---------------------------------------------------------------------------

        \505\ See letter from KBR.

        ---------------------------------------------------------------------------

        between $1 million and $1.5 million for at least 10 to 15 internal staff members and two or three external advisors, with 100 to 150 hours of internal work and 20 to 40 hours in external consulting time; \506\

        ---------------------------------------------------------------------------

        \506\ See letter from Avery Dennison.

        ---------------------------------------------------------------------------

        annual compliance cost of $15,000 for an issuer with global operations; \507\

        ---------------------------------------------------------------------------

        \507\ See letter from Intel.

        ---------------------------------------------------------------------------

        a ``likely to be conservative'' estimate of $100,000 per company, based on what mid-sized retail corporations informed the commenter.\508\

        ---------------------------------------------------------------------------

        \508\ See letter from NRF. We note that this commenter did not indicate the number of employees of such mid-sized retail corporations.

        ---------------------------------------------------------------------------

        approximately $250,000 for 1,000 internal hours initially and $100,000 per year for 500 hours annually thereafter (but, according to the commenter, the workload would perhaps drop by 90% if the final rule includes only employees employed by the U.S. parent organization and all U.S.-based subsidiaries in addition to other changes recommended by the commenter); \509\

        ---------------------------------------------------------------------------

        \509\ See letter from FEI.

        ---------------------------------------------------------------------------

        ``thousands of dollars to hire a dedicated resource and overhaul our payroll and human resource information system in order to prepare our first pay ratio disclosures under this rule;'' \510\

        ---------------------------------------------------------------------------

        \510\ See letter from FuelCell Energy.

        ---------------------------------------------------------------------------

        between $50,000 and $100,000 to test the commenter's current payroll system for a quote on identifying the median employee, with an ``actual cost'' that ``is indeterminable'' but that the commenter believes could cost over $500,000; \511\

        ---------------------------------------------------------------------------

        \511\ See letter from NACCO.

        ---------------------------------------------------------------------------

        $500,000 to $1 million for 50 internal employees and outside advisors; \512\

        ---------------------------------------------------------------------------

        \512\ See letter from General Mills.

        ---------------------------------------------------------------------------

        3,000 work hours in the initial year and 850 work hours annually thereafter (but, according to the commenter, these costs would be reduced by 90% if the final rule excluded non-U.S. employees); \513\

        ---------------------------------------------------------------------------

        \513\ See letter from ExxonMobil.

        ---------------------------------------------------------------------------

        over $1.6 million not including modifications of payroll or accounting systems; \514\

        ---------------------------------------------------------------------------

        \514\ See letter from Eaton.

        ---------------------------------------------------------------------------

        actual cost is indeterminable, but believed to exceed $500,000 due to substantial non-U.S. employee base; \515\

        ---------------------------------------------------------------------------

        \515\ See letter from Hyster-Yale.

        ---------------------------------------------------------------------------

        cost for many registrants would likely to be in the millions of dollars; \516\ and

        ---------------------------------------------------------------------------

        \516\ See letter from ASA.

        ---------------------------------------------------------------------------

        annual cost to collect required data would exceed $2 million.\517\

        ---------------------------------------------------------------------------

        \517\ See letter from Dover Corp.

        ---------------------------------------------------------------------------

        The overall cost range provided by individual commenters for initial compliance by a large registrant was between $15,000 per year to $2 million.\518\ We note that all of these comments concerned the proposed rule, rather than the final rule. As discussed below, the final rule allows for further flexibility, which we believe will reduce the cost of compliance.

        ---------------------------------------------------------------------------

        \518\ See letters from Dover Corp. and Intel.

        ---------------------------------------------------------------------------

        These estimates provide a significant number of data points on the anticipated compliance costs that we use in our quantification of the estimated compliance costs of the final rule below. However, we caution that these estimates do not necessarily represent an accurate indication of the expected costs because they use different methodologies and assumptions in arriving at these numbers, some of which might change with the different requirements under the final rule. Moreover, only a few commenters discussed the complexity of their payroll systems; \519\ the degree to which the estimated costs reflect internal personnel costs \520\ or the costs of outside service providers \521\ and outside professionals; \522\ and the precise assumptions used in deriving the estimates, all of which may be relevant for assessing the estimates provided.\523\ Also, although most of these estimates do not precisely distinguish between initial and ongoing costs, we expect that, for many registrants, the overall compliance burden will diminish after systems are in place to gather and verify the underlying data.\524\ Some commenters noted that the costs are impossible to determine before they start the process.\525\ The provided cost estimates were also given prior to additional cost-

        reducing measures adopted in the final rule in response to comments, including comments about costs. These measures include: The ability to calculate the median employee once every three years, the exemption from the definition of ``employee'' of a de minimis percentage of non-

        U.S. employees, the requirement to consider only registrant's own employees and those of their consolidated subsidiaries when identifying the median employee, and the exemption for employees in foreign jurisdictions in which it is not possible for a registrant to obtain or process information necessary to comply with the rule without violating the data privacy laws or regulations of that jurisdiction. We expect that these changes will further reduce the costs of compliance.

        ---------------------------------------------------------------------------

        \519\ See letter from Hyster-Yale (commenting that ``of the 4,967 people who received Form W-2s (or similar documents) for 2012, 2,375 (48%) resided outside the U.S. and were paid on twenty different payroll systems (none of which are integrated with our U.S. system or any other system)''). This commenter, however, did not comment on the expected cost of integrating these payroll systems.

        \520\ See letter from Avery Dennison (estimating that they will use 10-15 internal staff members with 100-150 hours of internal work), FEI (estimating 1,000 hours of internal time to develop the database and methodology to derive the data and 500 hours to maintain this database in the following years), and ExxonMobil (estimating 3,000 work hours by internal personnel in the first year and 850 work hours per year thereafter).

        \521\ See letter from ExxonMobil (stating ``are not able to provide a specific estimate, but expect that a significant work effort would also be required on the part of each of our 60 third-

        party vendors'').

        \522\ See letters from Avery Dennison (estimating that they will use two or three external advisors (e.g., legal counsel, human resources consultants) and 20-40 hours of external consulting time,) and Hyster-Yale (stating that it will need to hire an outside consulting firm to assist with the process).

        \523\ See letter from General Mills (providing a list of steps it will take). However, the commenter did not provide detailed cost estimates for these steps.

        \524\ A few commenters addressed this point, and the estimates they provided were very different. See, e.g., letters from ExxonMobil (expecting a 72% drop), FEI (estimating a 60% drop in compliance costs from the first year to the next year), and General Mills (indicating that most of the costs for the first year would also apply to subsequent years of compliance). But see letter from Business Roundtable I (asserting that 42% of respondents to a survey it conducted indicated that they would have to update their methodology every year).

        \525\ See, e.g., letter from NACCO (stating that consulting companies were unable to provide quotes about the cost to assist with statistical sampling until they are able to test their various payroll systems). This commenter also stated that the actual cost is indeterminable but could exceed $500,000, as it will depend on (i) the availability and accuracy of employee data, (ii) the scope of the final rule and (iii) whether registrants choose to disclose the minimum required disclosure or if they decide to provide various alternative disclosures that would provide shareholders with more context.

        ---------------------------------------------------------------------------

        In contrast to these estimates, a significant number of the commenters, generally the same commenters that perceived the benefits of the rule, asserted that the rule would not impose high costs and burdens. The majority of these commenters indicated that the

        Page 50158

        permitted flexibility in complying with the proposed rule would reduce costs,\526\ and that the registrants already have the data necessary to make their pay ratio calculations.\527\ While we agree that the permitted flexibility should lower costs for many registrants, we recognize that registrants who operate in various geographic and business segments may need to reconcile their systems to compile and provide the required information at a potentially significant cost.\528\

        ---------------------------------------------------------------------------

        \526\ See e.g., letters from AFR, Bricklayers International, CalPERS, Calvert, Capital Strategies, Chicago Teachers Fund, Cummings Foundation, First Affirmative, ICCR, IL Bricklayers and Craftworkers Union, LIUNA, Marco Consulting, Novara Tesija, NY Bricklayers and Craftworkers Union, Oxfam, Pax World Funds, Prof. Ravenscroft I, Public Citizen I, Rep. Ellison et al. I (referring specifically to sampling), Sen. Menendez et al. I, Sen. Menendez et al. II, and US SIF.

        \527\ See, e.g., letters from AFR, Allied Value, CalPERS, LAPFF, LIUNA, Pax World Funds, and Theodore.

        \528\ See, e.g., letters from American Benefits Council, Aon Hewitt, Avery Dennison, BCIMC, Business Roundtable I, Business Roundtable (Jul. 21, 2015) (``Business Roundtable II''), Chamber I, COEC I, Corporate Secretaries, Cummins Inc., Eaton, ExxonMobil, FEI, Freeport-McMoRan, FuelCell Energy, IBC, KBR, MVC Associates, NACCO, NAM I, NAM II, NIRI, Semtech, SHRM, and Tesoro.

        ---------------------------------------------------------------------------

        c. Quantification of Compliance Costs

        While our overarching consideration of the costs of the rule takes into account the information provided by a broad range of commenters, the most useful frameworks for considering costs were provided by commenters that provided data on company-wide potential costs. Other commenters provided certain valuable insights into how our rule would be implemented, but were either not as transparent in their analytical frameworks or not easily generalizable in terms of aggregating the costs across multiple registrants.

        Some commenters indicated that the most important driver for the compliance costs of the required disclosure is the presence of foreign operations and the complexity of dealing with, or the need to create compatible employee data and payroll systems that track the compensation of employees of such foreign operations.\529\ Underscoring the importance of foreign operations for the costs of compliance with the rule, one commenter's survey results indicate that the compliance costs for registrants would decrease by a median of 50% or on average by 40% if only U.S. employees were included in the calculation of the median compensation.\530\ In addition, most of the survey participants appear to have had at least some international operations. The participants in the commenter's survey on average operate in 34 countries (including the United States) \531\ and have about 38% of all employees and 44% of their full-time employees outside of the United States. The survey reflected predominantly larger registrants and therefore may not reflect the characteristics of a large number of registrants subject to Section 953(b) requirements. Another commenter suggested that the cost of implementing the final rule would be 20-30 times higher if foreign employees are included in the calculation of the median compensation.\532\ Based on these comments and survey results, we acknowledge that there may be significant differences in the potential costs of the rule between registrants with foreign operations and registrants without foreign operations.

        ---------------------------------------------------------------------------

        \529\ See id.

        \530\ See letter from COEC I. The survey was conducted jointly by COEC, Human Resource Policy Association, and Corporate Secretaries. See supra note 494.

        \531\ See id. The Chamber II survey participants were similarly international in operation. Over 68% of the participants in the Chamber II survey operate in more than 10 countries.

        \532\ See letter from American Benefits Council.

        ---------------------------------------------------------------------------

        Commenters have pointed out other potential cost drivers, such as including part-time, seasonal, and temporary employees in the calculation of the median. However, the effect of these other factors seems to be less significant. For example, one commenter's survey results suggest that the compliance costs for registrants would decrease by a median of 10% and on average by 11% if only full-time permanent employees are included in the determination of the median compensation, compared to an expected median reduction of 50% or average reduction of 40% if no foreign employees are included in the determination.\533\

        ---------------------------------------------------------------------------

        \533\ See letters from COEC I.

        ---------------------------------------------------------------------------

        Some of the compliance costs outlined above may be ameliorated by the de minimis exemption in the final rule that allows for some flexibility in identifying the median employee for registrants that have employees in multiple countries. The final rule also provides a foreign data privacy law exemption that should help to reduce the burden on registrants that operate in certain foreign jurisdictions. For some registrants with small foreign operations, the de minimis exemption and the foreign data privacy law exemption might greatly reduce the importance of foreign operations as a driver of compliance costs.

        Several commenters subject to the proposed rule provided compliance costs estimates specific to their particular situation. Other commenters provided cost estimates for what appear to be anonymous but real companies.\534\ For all estimates received, we have used data available from Standard and Poor's Compustat,\535\ to obtain or confirm information about the number of employees at the registrant. Because, as noted above, certain registrants were not specifically identified, we could not confirm or obtain their number of employees or use them to construct a reliable cost-per-employee estimate, which is a key factor in our analysis below. Accordingly, we decided to focus our analysis on those estimates where the registrant was identified, although we have noted below what the impact on our estimates would be if we were to include the estimates from the unidentified registrants whose employment data we cannot verify.

        ---------------------------------------------------------------------------

        \534\ One commenter mentioned that ``Company B,'' a U.S. multinational manufacturer with approximately 130,000 employees in about 275 locations worldwide, including 30,000 employees in the United States and 100,000 overseas expects that the cost to build the global human resources information system needed to comply with the proposed rule would exceed $18 million. See letters from NAM I and NAM II. Another commenter mentioned that one survey respondent with over 50,000 employees across 69 countries described the costs of data gathering as over $10 million. See letter from Corporate Secretaries. The commenter only notes that the survey respondent has over a certain number of employees.

        \535\ The Standard and Poor's Compustat Database is a comprehensive database of company financials routinely used by us, academics, and practitioners in empirical assessments involving public companies.

        ---------------------------------------------------------------------------

        To estimate the potential compliance costs of the final rule, we analyze the detailed information on compliance costs provided in comment letters from 10 registrants \536\ and provide an assessment, below, of how their estimates might relate more broadly to other affected registrants. The reported expected costs vary in nature, but the common element among the commenters is the cost associated with modifying current payroll or accounting systems to compile the information necessary for the identification of the median employee and the calculation of the employee's compensation. In quantifying these and other reported potential costs, when registrants presented a range of cost estimates, we use the mid-range value (i.e., if a registrant indicated that its costs would range from $0.5 million to $1 million, we use $0.75 million in the analysis).\537\

        Page 50159

        When registrants indicated that they expect to incur at least a certain dollar amount of costs, we use that value in our estimation, and, as such, this represents a lower bound because we have no way of estimating the top of their range (i.e., if a registrant indicated that its costs would be at least $0.5 million, we use $0.5 million in the analysis).

        ---------------------------------------------------------------------------

        \536\ See letters from Avery Dennison, Dover Corp., Eaton, ExxonMobil, General Mills, Intel, NACCO, FEI, Hyster-Yale, and KBR.

        \537\ Since commenters provided no information on how likely it is that the costs would be close to the lower or upper bound of the range, we believe that using the midpoint of the range would provide a reasonable cost estimate.

        ---------------------------------------------------------------------------

        One commenter provided costs in terms of the number of hours rather than in dollar value terms.\538\ In this case, we converted the hours into a dollar amount using the hourly rates reported in one commenter's survey,\539\ according to which the median respondent expected to incur the cost of $700 per hour to comply with the proposed rule. This rate is higher than our estimate of $400 per hour for professional costs in the PRA section of the Proposing Release, and we believe it may overestimate the costs of compliance with the final rule. We continue to believe that $400 per hour is the appropriate rate to use for PRA purposes as this reflects the average cost for outside professional services for all registrants, including smaller and mid-sized registrants. Nevertheless, we have used the $700 per hour rate here to be conservative in our estimates and because the commenter in question is a large registrant with globally diversified operations. The estimate of $700 per hour for outside compliance counsel may be justified for a certain type of registrant with certain size and characteristics but may not be representative of all registrants. We also note that we have monetized the commenter's entire hourly estimate using the $700 per hour rate. This likely overstates the actual dollar value of these costs, as we expect that some of the compliance burden of the final rule will be carried internally by the registrant (e.g., using the registrant's existing workforce) and at rates significantly lower than $700 per hour.

        ---------------------------------------------------------------------------

        \538\ See letter from ExxonMobil.

        \539\ See letter from COEC I.

        ---------------------------------------------------------------------------

        The 10 registrants that quantified expected registrant-wide compliance costs tend to be large registrants (in terms of both assets and revenues) and have globally diverse operations, and as such, may not be representative of the costs incurred by smaller registrants or registrants that do not have foreign operations. Thus, we restrict the use of the information provided by them to estimate the expected compliance costs for only registrants subject to the requirements that we identify as having foreign operations. Since we believe that the compliance costs will be generally proportionate to the size of the registrant's work force, we calculate the cost per employee using the number of employees reported for fiscal year 2013.\540\

        ---------------------------------------------------------------------------

        \540\ This assumption implies that all compliance costs are variable. It is possible that some compliance costs have a fixed component (i.e., payroll system configuration costs and training), and thus compliance costs per employee may be higher for small and mid-size firms with globally diversified operations than for globally diversified large firms, which may result in adverse effects on competition to the extent that the former are not SRCs and are not covered by the de minimis or foreign data privacy exemptions. Data on the number of employees and revenues is taken from Standard and Poor's Compustat.

        ---------------------------------------------------------------------------

        For the 10 registrants with compliance cost estimates, we estimate the ratio of ``Compliance cost estimates'' to ``Number of employees.'' We then take the median ratio and use it to estimate the expected initial compliance costs for registrants with U.S.-based operations and registrants with foreign operations. We use the median instead of the average to diminish the influence of outliers. The individual estimates and the average and median are presented in the table below. We estimate that the average cost-per-employee for these registrants would be $50.70 and the median cost-per-employee would be $38.04. The average cost per registrant is approximately $971,500, while the median is $750,000.\541\ We note that the 10 registrants in this analysis are larger than the average and median registrant subject to the final rule. To adjust for differences in registrant size, we make the assumption that the compliance costs of the rule will be proportionate to the size of the registrant's work force, which enables us to use the cost-per-employee ratio to estimate the potential compliance costs of affected registrants.

        ---------------------------------------------------------------------------

        \541\ Our estimate of the cost per registrant aggregates the cost of outside professional services and monetized internal burden hours. One commenter referred to an average cost of $311,800 per registrant to comply with the pay ratio provision. The commenter also estimated that the median registrant would require 1,825 internal burden hours to comply with the rule. These estimates are based on companies with operations in more than 50 countries that averaged 90 different employee data systems. See letter from Chamber II.

        Table 2--Total Initial Compliance Costs per Employee Based on Commenters' Estimates

        ----------------------------------------------------------------------------------------------------------------

        Total

        compliance Revenue ($ Number of Estimated cost

        Commenter CIK cost estimates millions) employees per employee

        (dollars) (thousands) (dollars)

        ----------------------------------------------------------------------------------------------------------------

        Avery Dennison.................. 0000008818 1,250,000 6,140.0 26 48.08

        Dover Corporation............... 0000029905 2,000,000 8,729.8 37 54.05

        Eaton Corporation............... 0000031277 1,600,000 22,046.0 102 15.69

        ExxonMobil Corp................. 0000034088 2,100,000 390,247.0 75 28.00

        General Mills................... 0000040704 750,000 17,909.6 43 17.44

        Intel........................... 0000050863 15,000 52,708.0 107.6 0.14

        NACCO........................... 0000789933 500,000 932.7 4.1 121.95

        FEI Company..................... 0000914329 250,000 927.5 2.61 95.79

        Hyster-Yale Materials Handling.. 0001173514 500,000 2,666.3 5.1 98.04

        KBR............................. 0001357615 750,000 7,214.0 27 27.78

        -------------------------------------------------------------------------------

        Average......................... .............. .............. .............. .............. 50.70

        Median \542\.................... .............. .............. .............. .............. 38.04

        ----------------------------------------------------------------------------------------------------------------

        We estimate the aggregate costs for the 3,571 registrants to whom the rule will apply, including both registrants with U.S.-based operations and registrants with foreign operations, using data from Standard and Poor's Compustat. We

        Page 50160

        classify registrants as having foreign operations if they report having at least one geographical segment outside the United States. If they report only U.S.-based segments, then we classify them as having U.S.-

        based operations only. We note several challenges in calculating these estimates. First, 13 of both the registrants with foreign operations and the registrants with U.S.-based operations only do not report the total number of employees. For these registrants, we impute the total number of employees by using the median number of employees for the registrants with foreign operations and the registrants with U.S.-based operations, respectively. Second, about 37% of affected registrants do not report geographic segment data, so we cannot classify them as registrants with U.S.-based operations or registrants with foreign operations.\543\ We compare the total assets and total revenues of these to the group of registrants with U.S.-based operations or registrants with foreign operations. These registrants tend to more closely resemble the registrants with U.S.-based operations, so we classify them as such. We recognize that this may underestimate the compliance cost for registrants with non-U.S. operations that choose not to report geographic segments.

        ---------------------------------------------------------------------------

        \542\ If we were to include the data for the anonymous registrants in the NAM I/NAM II letters and the Corporate Secretaries letter, the median would be $51.07. We note that the data for the anonymous registrant in the Corporate Secretaries letter did not specify a single number of employees, indicating instead that registrant had over 50,000 employees, thus making it difficult to compute a cost per employee ratio without additional assumptions. We assume that the number of employees of that anonymous registrant was 50,000 and its compliance cost was $10 million.

        \543\ Compustat, like most other databases, has incomplete coverage of small companies, and reporting companies that do not trade on national exchanges.

        ---------------------------------------------------------------------------

        We estimate the initial compliance costs for the registrants with foreign operations by first aggregating the number of employees across all such registrants and then multiplying that number by the median estimated cost per employee, calculated as $38.04 in Table 2 above.

        To estimate the expected costs for registrants with U.S.-based operations only, we rely on one commenter's survey results that indicate that the median decrease in registrants' compliance costs would be 50% if only U.S. employees are included in the determination of the median employee compensation.\544\ Thus, we assume that the expected compliance costs for registrants with U.S.-based operations only will be 50% lower than the expected compliance costs for registrants with foreign operations. Accordingly, we estimate the compliance costs for registrants with only U.S.-based operations by multiplying the total number of employees across such registrants by the (median estimated costs per employee * (1-0.5)), which is $19.02 per employee.\545\

        ---------------------------------------------------------------------------

        \544\ See letters from COEC I.

        \545\ $38.04*(1-0.5) = $19.02

        ---------------------------------------------------------------------------

        Lastly, to estimate the total compliance costs for registrants that do not report geographic segment data, which we reclassify as registrants with U.S.-based operations only, we first determine whether we have information on their number of employees.\546\ For the 973 registrants that we reclassify as registrants with U.S.-based operations and for which we have information as to their number of employees, we aggregate the total number of employees for those registrants and multiply it by $19.02. For the other 335 registrants that we reclassify as registrants with U.S.-based operations, we do not have information on the number of employees. As mentioned above, these registrants are similar to registrants with U.S.-based operations with respect to total assets and total revenues. Thus, to estimate the compliance cost for those registrants, we estimate the median cost per registrant with U.S.-based operations only and multiply it by the number of such registrants (335). To estimate the median cost per registrant with U.S.-based operations only, we first multiplied $19.02 per employee by the number of employees of each of the 793 registrants with U.S.-based operations only. We then found the median of those 793 amounts, which was $27,008.40. Multiplying this number by the total number of registrants with missing employee data (335), we reached a total cost estimate of $9,047,814. Consistent with other estimates in our analysis, we used median costs rather than average costs to reduce the significance of outliers. We believe that our approach to the estimate is more appropriate because, by using median numbers, we reduce the significance of outliers, but we acknowledge that had we instead estimated based on average numbers, a significantly higher cost estimate for this group would result. Our cost estimates are presented in Table 3 below.

        ---------------------------------------------------------------------------

        \546\ For about a third of registrants with missing segment data, Compustat reports no employee data.

        Table 3--Total Initial Compliance Cost Estimates for Affected Registrants

        ----------------------------------------------------------------------------------------------------------------

        Type of registrant Estimates Calculation

        ----------------------------------------------------------------------------------------------------------------

        Registrants with foreign operations:

        Registrants affected...................... 1,470

        Total number of employees................. 27,595,305

        Cost/employee ratio....................... $38.04

        Total cost................................ $1,049,725,402 27,595,305*$38.04

        Average cost per registrant............... $714,099 $1,049,725,402/1,470

        Registrants with U.S.-based operations only:

        Registrants affected...................... 793

        Total number of employees................. 6,522,626

        Cost/employee ratio....................... $19.02 $38.04*(1-0.5)

        Total cost................................ $124,060,347 6,522,626*$19.02

        Average cost per registrant............... $156,444 $124,060,347/793

        Median cost per registrant................ $27,008.40 This number represents the median of

        (number of employees * $19.02) across the

        793 U.S.-based registrants

        Registrants with missing data, reclassified as

        registrants with U.S.-based operations only:

        Registrants affected...................... 1,308

        Registrants with available employee data.. 973

        Total number of employees for the 973 6,932,754

        registrants.

        Cost/employee ratio....................... $19.02 $38.04*(1-0.5)

        Total cost for the 973 registrants........ $131,860,981 6,932,754*$19.02

        Registrants with no employee data......... 335 1,308-973

        Total cost for the 335 registrants........ $9,047,814 335 * $27,008.4

        Page 50161

        Total cost................................ $140,908,795 131,860,981 + 9,047,814

        Total................................. $1,314,694,544\547\ $1,049,725,402 + $124,060,347 +

        $140,908,795

        ----------------------------------------------------------------------------------------------------------------

        ---------------------------------------------------------------------------

        \547\ If we had included the cost estimates of the anonymous registrants provided by two commenters in calculating the median cost per employee ratio, as referenced above, the median cost per employee would be $51.07, and the total compliance cost for all 3,571 registrants would have been $1,765 million. See letters from NAM I, NAM II (providing the same estimate as in the NAM I letter), and Corporate Secretaries.

        ---------------------------------------------------------------------------

        Based on our calculations, the average initial cost of compliance for a registrant with foreign operations is expected to be approximately $714,099 and for a registrant with U.S.-based operations only is expected to be approximately $156,444. The total initial cost of compliance for all 3,571 registrants affected by the Section 953(b) requirements is expected to be approximately $1,315 million, which includes both internal company costs as well as the costs of outside professionals.\548\ Thus, we estimate the initial cost of compliance for the average registrant to be approximately $368,159.\549\

        ---------------------------------------------------------------------------

        \548\ If average cost per registrant with U.S.-based operations ($156,444) were used in lieu of median for the 335 registrants with missing employee data ($27,008.40), the total cost would instead be estimated as $1,049,725,402 + $124,060,347 + $131,860,981 + (335*$156,444) = $1,358 million.

        \549\ $1,314,694,544/3,571 = $368,159.

        ---------------------------------------------------------------------------

        It is important to note that this estimate does not reflect the de minimis and foreign data privacy exemptions, or the change to include only employees of consolidated subsidiaries, which would lead to some cost reductions for some registrants and which we are not able to fully quantify. Our cost estimate is higher than the survey estimate of $187 million for the cost of outside professionals provided by one commenter, although we note that the commenter also estimated, but did not monetize, an internal company burden of 800,870 hours.\550\ Similarly, our estimated cost is higher than the other commenter's survey estimate of over $710 million, although that commenter also estimated, but did not monetize, an annual compliance time of 3.6 million hours.\551\

        ---------------------------------------------------------------------------

        \550\ See letter from COEC I.

        \551\ See letter from Chamber II.

        ---------------------------------------------------------------------------

        Next, we estimate the ongoing compliance costs. Unlike in the case of the initial compliance costs, we received very few specific estimates of the ongoing compliance costs from commenters. One commenter suggested that ongoing annual costs would be approximately 28% of the initial compliance costs.\552\ Another commenter reported expected ongoing compliance costs of 40%.\553\ One commenter's survey results suggest that the ongoing costs are expected to be about 80% (mid-range) or 72% (average) of the initial compliance costs.\554\ We note that some compliance costs of the final rule, such as burden hours and professional costs associated with making the disclosure, may remain consistent from year to year. Other compliance costs, however, will largely be upfront fixed costs, such as those associated with the modification of payroll or accounting systems to allow a registrant to compile the information and costs associated with developing the methodology needed to identify the median employee and calculate his or her pay.\555\ Given these upfront fixed costs, it is likely that that part of the initial compliance costs would decline after the first year.

        ---------------------------------------------------------------------------

        \552\ See letter from ExxonMobil. The commenter stated that it expects 3,000 work hours for initial compliance costs and 850 work hours for ongoing compliance costs. This suggests the ongoing compliance cost is approximately 28.3% (850/3000) of the initial compliance costs.

        \553\ See letter from FEI. According to this commenter, the initial compliance costs would be approximately $250,000, while the ongoing compliance costs would be approximately $100,000, suggesting that the latter is 40% of the former ($100,000/$250,000).

        \554\ See letter from COEC I.

        \555\ But see letter from Business Roundtable I (asserting that 42% of respondents to a survey indicated that they would have to update their methodology every year). We note that ongoing costs of compliance may represent a higher percentage of the initial costs of compliance for these respondents.

        ---------------------------------------------------------------------------

        The specific estimates provided by commenters (28% to 72%) \556\ yield a range of ongoing compliance cost estimates of between $368 million and $947 million per year, with the median of the estimates provided by these commenters (40%) yielding an ongoing compliance cost of approximately $526 million per year.\557\ We note, however, that the Proposing Release did not provide registrants with the flexibility to identify the median employee every three years. We assume these three estimates are based on the commenters' reading of the Proposing Release, and hence include the requirement that the median employee be identified every year.

        ---------------------------------------------------------------------------

        \556\ Other commenters made more general assertions about ongoing compliance costs, but because they did not provide specific cost estimates, we did not include them in this calculation. See, e.g., letters from General Mills (asserting that ``most of the costs would also apply to subsequent years of compliance'') and Business Roundtable I.

        \557\ $1,315 million * 0.28 = $368 million; $1,315 million * 0.72 = $947 million; $1,315 million * 0.4 = $526 million.

        ---------------------------------------------------------------------------

        d. Indirect Costs

        Registrants covered by the final rule also could be affected by indirect costs.\558\ They could be at a competitive disadvantage to registrants (including private companies, foreign private issuers, smaller reporting companies and emerging growth companies) that are outside the scope of the final rule. Some commenters suggested that the proposed rule would cause competitive disadvantages for public companies,\559\ U.S. companies, especially those with overseas employees,\560\ issuers with subsidiaries,\561\ and issuers with low-

        wage workers.\562\ In addition, we understand from commenters that some registrants covered by the final rule would likely incur higher costs of compliance based on size, business type, and level of integration of payroll and benefits systems--such as large, multinational registrants that do not maintain integrated employee compensation information on a global basis.\563\ Therefore, the competitive impact of compliance with the disclosure requirements prescribed by Section 953(b) could disproportionately fall on U.S. registrants with large workforces and global operations, although the de minimis exemption and the foreign data privacy law exemption in the final rule would likely reduce some of these compliance costs and the competitive effects of the final rule. While we expect that the incremental

        Page 50162

        impact of the fixed components of compliance costs will have a lower impact on larger registrants than on smaller registrants, as discussed in the previous section, the overall compliance costs will likely lessen after systems are in place to gather and verify the underlying data, reducing the competitive effects of the final rule over the years.\564\

        ---------------------------------------------------------------------------

        \558\ We cannot precisely quantify the indirect costs because they depend on the registrant's business structure and competitive environment.

        \559\ See, e.g., letters from ABA, Prof. Angel, Former Assistant Secretary Campbell, Chamber I, COEC I, Corporate Secretaries, NAM I, NAM II.

        \560\ See, e.g., letters from ABA, Chamber I, COEC I, and Corporate Secretaries.

        \561\ See letter from COEC I.

        \562\ See, e.g., letters from IBC and Prof. Ray.

        \563\ See supra note 528.

        \564\ But see letters from Business Roundtable I (noting that 42% of respondents to a survey ``indicated that they expect having to update their pay ratio methodology every year because of changes to their business organization or structure'') and COEC I (arguing that the burden hours and professional costs may decrease after the first year but not by the estimated 50% in the Proposing Release).

        ---------------------------------------------------------------------------

        Registrants subject to the final rule could also face a competitive disadvantage if their competitors are able to infer proprietary or sensitive information from the disclosure of the median of the annual total compensation of all employees.\565\ For example, it could be possible in some instances for a competitor to infer sensitive information about a registrant's cost structure based on information about median levels of employee compensation, especially for small registrants operating in a single industry. While the final rule does not apply to smaller reporting companies and emerging growth companies, we still estimate that at least 131 affected registrants operated in a single business segment and had book value of assets less than $100 million in 2014.\566\ As we noted above, a registrant subject to Section 953(b) could potentially be at a competitive disadvantage when hiring or retaining a PEO if there is pressure to limit PEO wages based on the pay ratio disclosure while non-covered registrants are not subject to the same pressure. However, there may be other factors affecting the ability of a registrant to attract and retain executive talent.

        ---------------------------------------------------------------------------

        \565\ See letter from Prof. Ray. But see letter from Capital Strategies (stating, on the contrary, that competitors will not be able to decipher any proprietary or sensitive information from the pay ratio disclosure).

        \566\ This estimate is based on data from the Standard and Poor's Compustat Segments database as of the end of December 2014.

        ---------------------------------------------------------------------------

        One of the commenters indicated that 55% of the respondents in a survey of members it conducted anticipated indirect costs (i.e., adverse impact on sales, brand damage, increased public relations costs etc.)\567\ Although we acknowledge the possibility of these indirect costs, we cannot quantify them and lack sufficient data to analyze them.

        ---------------------------------------------------------------------------

        \567\ See letter from Corporate Secretaries.

        ---------------------------------------------------------------------------

      3. Other Economic Effects

        Several commenters indicated that the pay ratio rule would promote economic efficiency.\568\ In contrast, one commenter argued the rule would inhibit economic efficiency without providing specific details.\569\ As noted above, the pay ratio disclosure is not well suited to compare pay practices across registrants, and thus, it is unclear whether the final rule would affect economic efficiency. Some commenters suggested that registrants may decide to alter their pay structure or workforce structure in ways that are different from their efficient labor market decisions.\570\ For example, one commenter suggested that registrants may decide to shift their labor force to workers employed by a third party or reduce their foreign operations.\571\ Also, the same commenter asserted that registrants may change the relative wages of employees in a way that increases the median employee pay while reducing the pay of employees below the median employee in the pay distribution.\572\ Another commenter asserted that pressure for a registrant to maintain a low pay ratio could also curtail the expansion of business operations into lower cost geographies.\573\ We expect that such changes, if they were to occur, would move the registrant away from efficient business practices and could result in inefficient outcomes. Other commenters suggested, however, that workforce restructuring in response to the pay ratio disclosure was not likely.\574\ While we believe that registrants are unlikely to make critical labor decisions solely to impact the pay ratio disclosure, we cannot assess the prevalence of such effects at this time because these commenters did not quantify or otherwise provide data relevant to the expected changes in business practices.

        ---------------------------------------------------------------------------

        \568\ See, e.g., letters from Capital Strategies (asserting that ``The proposed rules would promote market efficiency as they would steer investors to the best value CEO, that is, the one producing the most returns for the least pay, when compared to peers.''), Change to Win (asserting that CEO pay increases have been driven by rent seeking behavior and ``if in fact the disclosures provided by Section 953(b) do induce more investors to insist on limiting executive pay, this will result in increased, rather than reduced, economic efficiency''), and Form Letter D (``Disclosure of the pay ratios will help the capital markets better allocate capital to those companies that invest in their workforces.'').

        \569\ See letter from NIRI.

        \570\ See, e.g., letters from CEG (stating that the rule can be ``gamed by outsourcing lower wage jobs''), NIRI (stating that the rule will adversely impact ``U.S. states and cities with lower labor costs''), and Prof. Ray (arguing that employers will change their corporate structure or employment arrangements as a response to the pay ratio rule).

        \571\ See letter from Lou (``Some companies will be incentivized to outsource poorly paid jobs to increase the median payment number. The flip side is the increase of administrative cost and losses of profits generated by the to-be-outsourced department. Or companies will reduce workforce in their foreign subsidiaries where the labor cost is relatively low. But obviously this reduction sacrifices the low labor cost advantage. Therefore the proposal does not motivate CEOs to maximize companies' interests.'').

        \572\ Id.

        \573\ See letter from Prof. Ray (providing the example of a firm that might prefer to hire in the U.S. rather than in India because a strong exchange rate of the U.S. dollar against the Indian rupee will make the Indian wages appear low and can lead to high pay ratio if the firm hires employees in India).

        \574\ See letters from ABA (asserting that registrants would not incur the related costs to alter their organizational structure or workforce in order to improve their pay ratio disclosure), Capital Strategies (asserting that the definition of ``all employees'' would not cause registrants to alter their corporate structure or employment arrangements), and WorldatWork I (same).

        ---------------------------------------------------------------------------

    4. Economic Effects From Exercise of Discretion

      1. General

        In this section, we discuss the choices we have made in implementing the statutory requirements and the associated economic effects, including the likely benefits and costs and the likely impact on efficiency, competition, and capital formation. In addition to the statutory benefits and costs described above, we believe that the use of our discretion in implementing the statute could result in benefits and costs to registrants and users of the pay ratio disclosure.

        In general, the final rule implementing Section 953(b) is designed to comply with the statutory mandate. In light of the significant potential costs that commenters attribute to the requirements of Section 953(b),\575\ we believe that it is appropriate for the final rule to permit registrants certain flexibility in calculating the pay ratio, which we believe should help lower the costs of compliance generally while still providing the information directed by Section 953(b). In addition, the final rule generally seeks to implement Section 953(b) without imposing additional requirements that are not mandated by the Dodd-Frank Act. In this respect, the final rule reflects our consideration of the relative costs and benefits of a more flexible approach as opposed to a more prescriptive approach.

        ---------------------------------------------------------------------------

        \575\ See, e.g., letters from Aon Hewitt, Avery Dennison, Business Roundtable I, Chamber II, Chesapeake, COEC I, COEC II, COEC III, Corporate Secretaries, Eaton, FEI, FuelCell Energy, Garmin, General Mills, Hyster-Yale, IBC, KBR, NACCO, NACD, NAM I, NAM II, NIRI, PNC Financial Services, and Semtech.

        ---------------------------------------------------------------------------

        In evaluating alternatives, we considered whether to adopt a rule that would be prescriptive enough that the

        Page 50163

        resulting pay ratio disclosure would be more directly comparable across registrants. As noted above, we believe that comparability of the ratio across registrants has significant limits, due to the variety of factors that could influence the ratio. We believe that providing a flexible approach would not significantly diminish the potential benefits of the mandated disclosure and would achieve the purposes that Congress intended at a significantly reduced burden for registrants. In this respect, we note that some commenters indicated that the expected benefits of pay ratio disclosure derive from its ability to facilitate a company-specific assessment, by providing a metric by which a PEO's compensation can be evaluated within the context of that particular company.\576\ We also acknowledge that some commenters that support the pay ratio disclosure suggested that it could be used to compare compensation practices between registrants and/or for the same registrant over time.\577\ We note, however, that using the ratios to compare compensation practices between registrants, and for a registrant over time (e.g., in the case of business acquisitions or significant structural or business model changes), without taking into account inherent differences in business models between registrants and for a registrant over time, which may not be readily available information, could potentially lead to unwarranted conclusions.

        ---------------------------------------------------------------------------

        \576\ See, e.g., letters from Bacirctirente et al., Bricklayers International, CalSTRS, Calvert, Domini, FS FTQ, Pax World Funds, Walden, and WA State Investment Board.

        \577\ See, e.g., letters from Bacirctirente et al., Bricklayers International, CalPERS, CalSTRS, Calvert, Domini, EnTrust Capital (Nov. 20, 2013) (``EnTrust''), FS FTQ, LIUNA, Pax World Funds, Public Citizen I, RPMI, Walden, and WA State Investment Board.

        ---------------------------------------------------------------------------

      2. Implementation Choices and Alternatives

        a. Filings Subject to the Pay Ratio Disclosure Requirements

        Commenters suggested that the final rule should apply only to those filings for which the applicable form requires Item 402 disclosure.\578\ Some commenters stated that requiring pay ratio disclosure in every filing would be unnecessary and even confusing.\579\ The final rule follows the proposed approach to require pay ratio disclosure in filings described in Item 10(a) of Regulation S-K that require executive compensation disclosure under Item 402 of Regulation S-K. We believe that requiring pay ratio disclosure in filings that do not contain other executive compensation information would not present this information in the most meaningful context. Some commenters asserted that the pay ratio disclosure would provide another metric to evaluate executive compensation disclosure,\580\ and we believe that the intended purpose of the disclosure is to provide new data points that shareholders can use when exercising their new voting rights under Section 951. We believe that the pay ratio disclosure would be presented in a more meaningful context if it were accompanied by other Item 402 information, such as the Summary Compensation Table required by Item 402(c) and the Compensation Discussion and Analysis required by Item 402(b). Therefore, we believe that this choice preserves the intended benefits of Section 953(b) while reducing reporting costs relative to a requirement to include pay ratio disclosure in every filing, including in filings that do not require other Item 402 information.

        ---------------------------------------------------------------------------

        \578\ See letters from ABA, AFL-CIO I, CalPERS, Calvert, Capital Strategies, CII, CT State Treasurer, Davis Polk, Intel, Johnson & Johnson, McGuireWoods, NIRI, NRF, Pax World Funds, PM&P, Prof. Ray, and WorldatWork.

        \579\ See, e.g., letters from ABA (``In our view, it is neither reasonable nor sensible to mandate the inclusion of pay ratio disclosure in filings where no other executive compensation disclosure required by Item 402 will appear to provide meaningful context.'') and PM&P (``Providing such information in multiple filings (e.g., registration statements, annual reports or other filings) throughout the year is unnecessary and would dilute the usefulness, if any, of the disclosure.'').

        \580\ See supra note 34.

        ---------------------------------------------------------------------------

        b. Registrants Subject to the Pay Ratio Disclosure Requirements

        We recognize that the reference to ``each issuer'' in Section 953(b) could be interpreted to apply to all registrants. However, as a result of the specific reference in Section 953(b) to the definition of ``total compensation'' contained in Item 402(c)(2)(x) and the absence of Congressional direction to apply this requirement to registrants not previously subject to Item 402(c) requirements, the final rule does not apply to registrants that are not subject to Item 402(c) requirements. Thus, smaller reporting companies, foreign private issuers, and MJDS filers are excluded. In addition, Congress exempted emerging growth companies from the requirement in the JOBS Act.

        We considered a number of alternative approaches. We considered whether a broader reading of the statute was warranted in the context of SRCs as suggested by some commenters.\581\ However, most commenters agreed with the approach to exclude SRCs from the requirements.\582\ Commenters either argued that these registrants are not currently required to provide a Summary Compensation Table under Item 402(c) and therefore should not be required to comply with the pay ratio rule \583\ or cited high costs of compliance.\584\ Requiring SRCs to provide the pay ratio disclosure consistent with the requirement for other registrants would require them to collect data and calculate compensation for the PEO in a manner they otherwise would not do, and there would be some incremental costs in doing so. However, these incremental costs may be limited to the extent that smaller reporting companies are less likely to have defined benefit and actuarial pension plans.\585\ In contrast, the costs of complying with the other requirements prescribed by Section 953(b)--namely, identifying the median employee and calculating annual total compensation for that employee--are more extensive. We can estimate those costs using the approach and estimates we made in the ``Quantification of Compliance Costs'' section above. We identify 2,958 registrants as SRCs that are not EGCs as of the end of fiscal year 2014. Of these, 494 have data on segments and number of employees in Compustat. Following the approach in the ``Quantification of Compliance Costs'' section above, we use information on international geographic segments reported in Compustat to identify registrants with and without international operations. Of the 2,958 SRCs, 212 have foreign operations and 282 have U.S.-based operations only; the rest does not have

        Page 50164

        segment data in Compustat.\586\ We next obtain the total number of employees for SRCs with foreign operations and SRCs with U.S.-based operations only from Compustat. We apply the respective ``compliance cost per employee'' ratios for registrants with foreign and U.S.-based operations only estimated in section ``Quantification of Compliance Costs'' above to estimate the average and total compliance costs for SRCs with foreign operations and SRCs with U.S.-based operations only. We reclassify the remaining 2,464 registrants as SRCs with U.S.-based operations only because our analysis suggests that registrants that are not covered by Compustat are usually smaller companies without foreign operations. Consistent with our methodology for Table 3, to estimate the total compliance costs for SRCs that do not report segment data, which we reclassify as SRCs with U.S.-based operations only, we estimate their total number of employees and multiply it by $19.02. For those SRCs we reclassify as SRCs with U.S.-based operations only that do not have employee data available, we estimate the median cost per SRC with U.S.-based operations only and multiply it by the number of such registrants. The table below presents our estimates of the compliance costs of SRCs with foreign operations and SRCs with U.S.-

        based operations only, as well as total compliance costs for SRCs.

        ---------------------------------------------------------------------------

        \581\ See, e.g., letters from CII (indicating that two of three CII members that commented on the proposed rule ``were `not comfortable' with the proposed exemption from the pay ratio disclosure requirements for emerging growth companies, smaller reporting companies, foreign private issuers, and MJDS filers''), Ray (``To support Congress's demand for greater pay-related disclosures, I strongly suggest for the Commission to expand the disclosure requirement to ensure that all smaller reporting companies disclose their pay ratio.''), and US SIF (``While we understand the SEC's reasons for several exemptions from the proposed rule for emerging growth companies, smaller companies and foreign private issuers, we are, nonetheless, uncomfortable with these proposed exemptions.'').

        \582\ See, e.g., letters from ABA, Prof. Angel, CalPERS, Capital Strategies, Davis Polk, Hay Group, NIRI, NY State Comptroller, PM&P, Vivient, and WorldatWork.

        \583\ See, e.g., letters from CalPERS and WorldatWork (``Smaller companies would face a double compliance burden if asked to publish summary compensation tables and calculate the pay ratio.'').

        \584\ See, e.g., letters from Prof. Angel and Vivient.

        \585\ See supra note 92.

        \586\ For approximately 60% of SRCs with missing segment data, Compustat reports no employee data.

        Table 4--Total Initial Compliance Cost Estimates for SRCs

        ----------------------------------------------------------------------------------------------------------------

        Type of registrant Estimates Calculation

        ----------------------------------------------------------------------------------------------------------------

        Registrants with foreign operations:

        Registrants affected...................... 212

        Total number of employees................. 60,280

        Cost/employee ratio....................... $38.04

        Total cost................................ $2,293,051 60,280*$38.04

        Average cost per registrant............... $10,816 2,293,051/212

        Registrants with U.S.-based operations only:

        Registrants affected...................... 282

        Total number of employees................. 89,625

        Cost/employee ratio....................... $19.02 $38.04*(1-0.5)

        Total cost................................ $1,704,668 89,625 *$19.02

        Average cost per registrant............... $6,045 1,704,668/282

        Median cost per registrant................ $1,103 This number represents the median of

        (number of employees * $19.02) across the

        282 U.S.-based registrants

        Registrants with missing data, reclassified as

        U.S.-based operations only:

        Registrants affected...................... 2,464

        Registrants with available employee data.. 1,150

        Total number of employees for the 1,150 245,752

        registrants.

        Cost/employee ratio....................... $19.02 $38.04*(1-0.5)

        Total cost for the 1,150 registrants...... $4,674,203 245,752*$19.02

        Registrants with no employee data......... 1,314 2,464-1,150

        Total cost for the 1,314 registrants...... $1,449,342 1,314 * $1,103

        Total cost................................ $6,123,545 4,674,203+1,449,341

        Total................................. $10,121,264 2,293,051 +1,704,668

        +6,123,545

        ----------------------------------------------------------------------------------------------------------------

        Our decision not to require SRCs to comply with the pay ratio disclosure requirements prescribed by Section 953(b) would save the average SRC with foreign operations approximately $10,816, and the average SRC with U.S.-based operations only approximately $6,045. We note that these cost savings are the savings from not having to identify the median employee and calculate the total compensation for that employee. Those cost savings from exercise of our discretion with respect to SRCs total approximately $10 million.\587\ We expect there also would be cost savings from not having to calculate PEO compensation pursuant to Item 402, but we are unable to quantify those savings.

        ---------------------------------------------------------------------------

        \587\ If average cost per SRC with U.S.-based operations were used in lieu of median for registrants with missing data, the total cost savings would instead be estimated as $2,293,051 + $1,704,668 + $4,674,203 + 1,314*$6,045 = $16.62 million.

        ---------------------------------------------------------------------------

        To the extent that these costs have a fixed component that does not depend on the registrant's size of operations, the compliance burden for small registrants may be disproportionately large.\588\ Moreover, small companies are more likely to operate in a single geographic or business segment, making the disclosure of the median employee pay more likely to reveal sensitive or proprietary information that can put these registrants at a competitive disadvantage.

        ---------------------------------------------------------------------------

        \588\ See letter from Vivient (``From a cost standpoint, requiring smaller reporting companies to disclose the pay ratio would create a significant cost and administrative burden. Given the size of their employee population and administrative budget, very few smaller reporting companies employ full-time senior level human resource professionals who could be assigned the responsibility of complying with the proposed ruling.'').

        ---------------------------------------------------------------------------

        We also considered expanding the coverage of the final rule to registrants, such as foreign private issuers and MJDS filers, which are not currently required to provide Item 402 disclosure. Most commenters agreed with the exclusion of foreign private issuers and MJDS filers,\589\ but other commenters expressed concerns about excluding them.\590\ Although quantifying the costs to these registrants of calculating PEO compensation under Item 402(c)(2)(x) or of complying with the requirements

        Page 50165

        prescribed by Section 953(b) is not currently feasible because of lack of data, we assume that these costs (and in particular the costs of computing the median employee) could be significant. In particular, these costs may be higher for foreign private issuers and MJDS than for SRCs because these registrants are not currently required to provide any Item 402 disclosure. Based on a review of EDGAR filings for calendar year 2014, we estimate that there are approximately 677 foreign private issuers filing on Form 20-F and 143 MJDS filers filing on form 40-F that will benefit from the exclusion from the pay ratio disclosure requirements.

        ---------------------------------------------------------------------------

        \589\ See letters from ABA (stating that Section 953(b) does not require expanding the scope of Item 402 of Regulation S-K to apply to registrants not currently required to comply with the Item 402 disclosure, such as foreign private issuers and MJDS filers, so the Commission should not expand its rules to do so), Capital Strategies, Davis Polk, Hay Group, and PM&P.

        \590\ See letters from CalPERS, CII, and US SIF.

        ---------------------------------------------------------------------------

        c. Employees Included in the Determination of the Median

        Section 953(b) requires disclosure of the median of the annual total compensation of ``all employees of the issuer.'' Consistent with that mandate, the final rule includes in the definition of ``employee'' or ``employee of the registrant'' any U.S. and non-U.S. full-time, part-time, seasonal, or temporary worker (including officers other than the PEO) employed by the registrant or any of its subsidiaries as of the last day of the registrant's last completed fiscal year. Additionally, as set forth above, we are excluding from the determination of the median employee any workers who are not employed by the registrant or its consolidated subsidiaries, such as independent contractors and ``leased'' workers who are employed by, and whose compensation is determined by, an unaffiliated third party.

        i. Types of Employees

        Commenters were generally split on whether the rule should include part-time, seasonal, or temporary employees. A number of commenters agreed with the proposed requirements to include part-time, temporary, and seasonal workers because of Section 953(b)'s reference to ``all employees'' and believed that excluding these employees would distort the pay ratio by rendering it incomplete or misleading.\591\ Other commenters suggested that the final rule should exclude part-time, seasonal, or temporary employees. These commenters asserted that compensation for these employees is not comparable to full-time employees, so their inclusion would distort the pay ratio.\592\ Some commenters believed that the final rule should exclude part-time, seasonal, or temporary employees because the potential benefits from including them would not justify the high costs.\593\ Another commenter recommended that the final rule should exclude part-time, seasonal, and temporary employees unless a majority of a registrant's employees work on a part-time, temporary, and/or seasonal basis.\594\

        ---------------------------------------------------------------------------

        \591\ See letters from AFL-CIO I, AFR, Bacirctirente et al., Bricklayers International, CII, CT State Treasurer, FS FTQ, Public Citizen I, and Theodore.

        \592\ See letters from AAFA I, AAFA II, American Benefits Council, COEC I, Corporate Secretaries, and Davis Polk.

        \593\ See letters from COEC I (``Nearly all firms, 122 of 128, have part-time and/or seasonal employees. With this in mind, two-

        thirds of survey respondents to the survey indicated that limiting the application of the proposed pay ratio rules to full-time employees only would reduce their costs. The average savings for these respondents would be approximately 20 percent.''), General Mills (which estimated that approximately 13% of their employees are part-time, seasonal or temporary), and WorldatWork II (recommending exclusion of part-time, seasonal, and temporary employees from the pay ratio).

        According to one of the commenters, the average percentage of full-time employees reported by survey respondents was 86% (mid-

        range was 95%). See letter from COEC I. Another commenter estimates the average (median) percentage of full-time employees to be 87% (95%). See letter from Corporate Secretaries. Based on BLS data, approximately 81% of workers were employed full time as of 2014. See BLS Labor Force Statistics from the Current Population Survey, available at http://www.bls.gov/cps/cpsaat12.htm. We note that BLS data incorporates workers at a broader range of firms, including privately held firms and small firms, which may not be representative of the composition of the workforce at the registrants subject to the final rule.

        \594\ See letter from ABA.

        ---------------------------------------------------------------------------

        The final rule requires registrants to include part-time, temporary, and seasonal employees when identifying the median employee. We could have chosen an alternative approach, namely to allow registrants to base their pay ratio disclosure on full-time employees only. This approach would have led to a lower cost of compliance. According to one survey, such flexibility would have generated median savings of approximately 10% in compliance costs.\595\ Applying this estimate to our compliance cost estimate of $1,315 million, had we chosen this alternative the total compliance costs would have been approximately $1,183.5 million, or savings of approximately $131.5 million. According to another commenter, excluding part-time employees could reduce costs of compliance by 20%, which would raise the estimate of potential cost savings to approximately $263 million.\596\ Another commenter noted that more than 30 percent of respondents to its survey believe that limiting the median employee calculation to full-time employees would yield cost savings of more than 10 percent and an additional 18 percent of respondents believe that limiting the median employee calculation to full-time employees will yield cost savings of more than 20 percent.\597\ Despite these potential cost savings, we are not adopting this alternative. Several commenters argued that excluding part-time, temporary, and seasonal workers from the definition of ``employee'' would make the disclosure incomplete and/or not representative of the registrant's actual workforce.\598\ We agree and have concluded, as discussed more fully in section B.1.a. above, that this approach could significantly undermine the intent of the rule to allow a company-specific assessment of a registrant's compensation practices with respect to ``all employees''.

        ---------------------------------------------------------------------------

        \595\ See letter from Corporate Secretaries.

        \596\ See letter from COEC III.

        \597\ See letter from WorldatWork I.

        \598\ See, e.g., letters from AFL-CIO I, AFR, Bacirctirente et al., Bricklayers International, CII, CT State Treasurer, FS FTQ, and Public Citizen I.

        ---------------------------------------------------------------------------

        ii. Workers Not Employed by the Registrant (i.e., Leased Workers)

        The final rule, as proposed, excludes independent contractors and ``leased'' workers who are employed by, and whose compensation is determined by, an unaffiliated third party. Commenters generally supported this approach.\599\ As discussed, we believe excluding such workers is appropriate because registrants generally do not control the level of compensation that these workers are paid.

        ---------------------------------------------------------------------------

        \599\ See supra note 114.

        ---------------------------------------------------------------------------

        We recognize that it is possible that a registrant could alter its corporate structure or its employment arrangements to reduce the number of employees covered by the final rule and, therefore, reduce its costs of compliance or alter its pay ratio disclosure to achieve a particular objective. For example, a registrant could choose to use only independent contractors or ``leased'' workers instead of hiring employees. A registrant could also choose to outsource some aspects of its business to achieve similar objectives. Although one commenter asserted that registrants would change their corporate structure or employment arrangements based on the definition of ``employee,'' \600\ other commenters questioned the likelihood of this behavior.\601\ We cannot quantify the expected prevalence of this behavior. However, given the inherent complexities involved in altering a registrant's corporate structures or employment arrangements, we do not expect that many registrants would undertake such changes merely for the

        Page 50166

        purposes of lowering compliance costs or achieving a particular pay ratio.

        ---------------------------------------------------------------------------

        \600\ See letter from Prof. Ray.

        \601\ See, e.g., letters from ABA, Capital Strategies, and WorldatWork I.

        ---------------------------------------------------------------------------

        iii. Employees of Consolidated Subsidiaries

        As discussed above, of the commenters that discussed whether to include employees of a subsidiary, the majority recommended that the final rule require registrants to include only employees of certain types of subsidiaries, in particular consolidated or wholly-owned subsidiaries.\602\ One of those commenters claimed that there would be a median increase in compliance costs of approximately 20% if the final rule included employees of all minority-owned subsidiaries and joint ventures.\603\ Another commenter argued that limiting the final rule to consolidated subsidiaries would reduce costs and burdens and is consistent with Rule 405 of the Securities Act and Rule 12b-2 of the Exchange Act.\604\

        ---------------------------------------------------------------------------

        \602\ See letters from ABA, Best Buy et al., Business Roundtable I, COEC I, Corporate Secretaries, CT State Treasurer, Davis Polk, Eaton, ExxonMobil, General Mills, Mercer I, Meridian, NACCO, NAM I, and NAM II.

        \603\ See letter from Corporate Secretaries.

        \604\ See letter from ABA.

        ---------------------------------------------------------------------------

        The final rule defines ``employee'' to include only the employees of the registrant and its consolidated subsidiaries rather than employees of subsidiaries that were affiliates it controlled directly or indirectly through one or more intermediaries, as set forth in the definition of ``subsidiary'' under both Securities Act Rule 405 and Exchange Act Rule 12b-2. This change will affect registrants that have unconsolidated subsidiaries with a significant number of workers.

        We believe that excluding employees of unconsolidated subsidiaries may provide a better representation of the compensation practices of the registrant itself since the compensation provided by unconsolidated subsidiaries may be beyond the control of the registrant covered by Section 953(b).

        Based on our analysis of Compustat firms for calendar year 2014, excluding firms identified as emerging growth companies, smaller reporting companies, foreign private issuers, and MJDS filers, approximately 23% of firms reported positive equity investments in unconsolidated subsidiaries,\605\ with the median investment of approximately 1.7% of the book value of total assets. The majority of the firms were in the financial, regulated utilities, and oil and gas industries. We lack information on the number of employees at unconsolidated subsidiaries to quantify the potential magnitude of their effect on the pay ratio disclosure.

        ---------------------------------------------------------------------------

        \605\ Our analysis excludes investments in unconsolidated subsidiaries accounted for by the cost method as those are not identified separately by filers in the data available to us.

        ---------------------------------------------------------------------------

        We also reviewed Exhibit 21 to the annual reports on Form 10-K, which contains subsidiary information, for a sample of 24 firms that submitted individual unique comment letters pertaining to the proposal. The median number of subsidiaries in that set of firms was approximately 31. Only three registrants explicitly identified the number of consolidated subsidiaries, with the median being approximately 36. The registrants we studied may not be random as firms with more subsidiaries may be more likely to submit a comment letter. The information in the exhibits indicates that some firms have complex organizational structures but it does not allow us to systematically differentiate between consolidated and unconsolidated subsidiaries.

        The final rule allows registrants to exclude employees from unconsolidated subsidiaries when identifying the median employee. This change from the proposed rule could lead to significant cost savings.\606\ First, limiting the definition in this way will result in a smaller pool of employees from which to identify the median employee, thereby helping to reduce compliance costs associated with this step. Second, registrants are more likely to maintain integrated systems with their consolidated subsidiaries because these subsidiaries have to consolidate their financial statements with those of the registrant, which should make it easier to collect and analyze the relevant data.\607\ Finally, as the consolidated subsidiary standard is commonly applied in other disclosures, there may be less cost for registrants to identify subsidiaries relevant for the disclosure. In summary, the final rule could provide a potential competitive advantage to registrants with a significant percentage of the workforce at unconsolidated subsidiaries over registrants with consolidated subsidiaries due to lower compliance costs associated with having fewer workers covered by the rule.

        ---------------------------------------------------------------------------

        \606\ See, e.g., letters from ABA, COEC I (claiming that there would be a 91% increase on average in costs if registrants were required to include all minority-owned subsidiaries and joint ventures in the definition of ``employee''), and Corporate Secretaries (survey reporting a median increase in costs of 20% if the definition of ``employee'' included all minority-owned subsidiaries and joint ventures of the registrants).

        \607\ See letter from COEC I. The letter indicates that in the majority of cases, a registrant's access to the information necessary to calculate the pay ratio will only extend to wholly-

        owned subsidiaries that consolidate their financial statements with those of the registrant.

        ---------------------------------------------------------------------------

        We have attempted to quantify the expected decrease in compliance costs from the revised definition of subsidiaries of the registrant, but did not obtain estimates on what these costs would be. One commenter's survey results suggested that compliance costs would increase by approximately 20% (median) compared to the proposal if the final rule required registrants to include employees of all minority-

        owned subsidiaries and joint ventures.\608\ However, the effect of allowing registrants to exclude employees of unconsolidated subsidiaries on compliance costs relative to the proposed rule is not clear from this estimate. In light of this uncertainty and because we do not have other data available on the effects on the compliance cost estimate of the exclusion of employees of unconsolidated subsidiaries versus the exclusion of ``minority-owned subsidiaries and joint ventures,'' we are not reducing our initial cost estimates to account for the change to only include employees of consolidated subsidiaries. We acknowledge that our estimates may therefore overstate the compliance cost for companies with unconsolidated subsidiaries. Although we are unable to estimate the magnitude of this cost savings, as noted above, and consistent with the views of commenters, we believe that it could be significant for some companies.

        ---------------------------------------------------------------------------

        \608\ See letter from Corporate Secretaries. See also letter from COEC I.

        ---------------------------------------------------------------------------

        iv. Employees Located Outside the United States

        As discussed above, a number of commenters asserted that non-U.S. employees should be included in the final rule.\609\ Some who supported this view argued that the failure to include foreign workers would substantially affect the pay ratio disclosure.\610\ On the other hand, many commenters indicated that including those employees would lead to significantly higher costs and suggested that the final rule allow registrants to use only their U.S. employees when identifying the median

        Page 50167

        employee.\611\ One commenter indicated that costs would be 20 to 30 times higher, or hundreds of thousands of dollars higher if non-U.S. employees are included.\612\ Other commenters asserted that costs would decrease by over 50% \613\ or 90% \614\ if non-U.S. employees are excluded. Another commenter indicated that nearly half of respondents to its survey expected a U.S.-employee -only ratio to reduce compliance costs by more than 20 percent, while 29 percent of respondents expected it would reduce compliance costs by more than 40 percent.\615\

        ---------------------------------------------------------------------------

        \609\ See, e.g., letters from AFL-CIO I, AFSCME, Bricklayers International, CalPERS, Calvert, Chicago Teachers Fund, Corayer, CT State Treasurer, Cummings Foundation, CUPE, Estep, Fedewa, First Affirmative, Gould, ICCR, IL Bricklayers and Craftworkers Union, Marco Consulting, Matteson, McMorgan Co., Sen. Menendez et al. I, Novara Tesija, NY Bricklayers and Craftworkers Union, NY State Comptroller, Oxfam, Pax World Funds, C. Phillips, Public Citizen I, Rand, S. Spofford, Socially Responsive Financial Advisors, Teamsters, Trillium I, Trustee Campbell, US SIF, and Walden.

        \610\ See, e.g., letters from AFR, Bacirctirente et al., CII, Domini, and FS FTQ.

        \611\ See, e.g., letters from ABA, American Benefits Council, Business Roundtable I, Business Roundtable II, Chamber I, Corporate Secretaries, ExxonMobil, Frederick W. Cook & Co., RILA, Semtech, and SHRM.

        \612\ See letter from American Benefits Council.

        \613\ See letter from Business Roundtable I.

        \614\ See, e.g., letters from ExxonMobil and FEI.

        \615\ See letter from WorldatWork I.

        ---------------------------------------------------------------------------

        Comment letters addressing costs associated with including non-U.S. employees often noted that multinational registrants have multiple payroll systems and databases for their employees' compensation that are difficult if not impossible to reconcile.\616\ One commenter indicated that it has 15 payroll systems that are not integrated, and those payroll systems would have to be manually reconciled with ``substantial costs'' and ``extensive staff hours.'' \617\ Another commenter stated that it used 30 payroll systems that are not connected.\618\ Yet another commenter cited a Human Resource Policy Association survey indicating that 84% of respondents could not easily calculate worldwide enterprise cash compensation for all their employees.\619\

        ---------------------------------------------------------------------------

        \616\ See, e.g., letters from American Benefits Council, Aon Hewitt, BCIMC, Business Roundtable I, COEC I, Cummins, Eaton, ExxonMobil, Freeport-McMoRan, MVC Associates, NIRI, Semtech, SHRM, and Tesoro.

        \617\ See letter from Freeport-McMoRan. Similarly, a different commenter stated that it does not have a single payroll system that can easily analyze the type of data required for the calculation. See letter from Tesoro.

        \618\ See letter from Cummins Inc.

        \619\ See letter from MVC Associates (citing to COEC I).

        ---------------------------------------------------------------------------

        The final rule does not allow registrants to exclude their non-U.S. employees when identifying the median employee, other than the limited exceptions described below. One commenter \620\ estimated that the average company in a survey it conducted had 40% of its workforce located outside the United States and recommended a principles-based approach that would permit registrants to exclude up to and exceeding 40% of their employee population. This commenter estimated that permitting registrants to exclude non-U.S. employees would reduce compliance costs by 47%. Another commenter estimated that the median decrease in the compliance costs for registrants with foreign operations would be approximately 50% if the final rule excluded non-

        U.S. employees.\621\ Given our estimate of aggregate initial compliance costs for registrants with foreign operations of approximately $1,050 million, had we instead excluded non-U.S. employees, the survey results suggest such registrants' initial compliance costs would instead be $525 million.

        ---------------------------------------------------------------------------

        \620\ See letter from COEC III.

        \621\ See letter from Corporate Secretaries.

        ---------------------------------------------------------------------------

        v. Foreign Data Privacy Law Exemption

        The final rule also provides a foreign data privacy exemption that gives registrants the ability to exclude from their median employee computation non-U.S. employees in jurisdictions in which data privacy laws or regulations prohibit the use or transfer of the necessary information required to comply with the final rule. According to one commenter's survey, 45.8% of respondents anticipated ``being prohibited or limited by non-U.S. data privacy laws'' in their efforts ``to access information necessary to collect data to identify the median employee or make the pay ratio calculation.'' \622\ The foreign data privacy exemption may lower the costs of calculating the pay ratio for registrants with employees in such jurisdictions, although we do not have data from which to estimate the magnitude of the cost savings. We recognize that it may also affect the median employee compensation determination. For example, based on the latest available Bureau of Economic Analysis (BEA) data for 2012, U.S. multinational companies were estimated to have approximately 11.5% of their employees at foreign affiliates in the EU, 3.2% in Canada, 4.2% in China, 3.7% in Mexico, 1.4% in Japan, and 2.6% combined in Switzerland, Australia, Argentina, Russia, and Singapore. Lower estimates are obtained when only majority-owned foreign affiliates are considered. Each of these jurisdictions was identified by commenters as having laws that may prohibit or restrict the transfer of information necessary to make the pay ratio calculation.\623\ To the extent that data privacy restrictions may be present both in high-income and low-income jurisdictions, the direction of the effect of the exemption on the pay ratio is ambiguous and may vary from registrant to registrant. To the extent that registrants with non-U.S. workers in jurisdictions with data privacy laws would have experienced a significantly higher cost of calculating the pay ratio than registrants with the same percentage of non-U.S. workers but in jurisdictions without such laws, this change from the proposed rule mitigates the potential adverse competitive effects of the pay ratio disclosure requirement on registrants with non-U.S. workers in jurisdictions with data privacy laws. Consistent with a commenter's suggestion,\624\ the final rule requires registrants to obtain a legal opinion from counsel in that jurisdiction on the inability of the registrant to obtain or process the information necessary for compliance with the final rule without violating that jurisdictions' laws or regulations governing data privacy, including the registrant's inability to obtain an exemption or other relief under any governing laws or regulations. The legal opinion must be

        Page 50168

        filed as an exhibit with the filing in which the pay ratio disclosure is included.

        ---------------------------------------------------------------------------

        \622\ See letter from COEC I.

        \623\ See, e.g., letters from ABA (``The best known of these data privacy regimes is the European Union's Directive 95/46/EC of the European Parliament and the Council of 24 October 1995 on the Protection of Individuals with Regard to the Processing of Personal Data and the Free Movement of Such Data, 1995 O.J. L 281 . . . . Other jurisdictions, including Argentina, Canada, Japan, and Switzerland, have also adopted strong data privacy laws.''), American Benefits Council (``Among other things, the EU data privacy regime prohibits `the transfer of personal data to a third country which does not ensure an adequate level of protection.' We also understand that many other countries, including China, Japan, Mexico, Canada, Peru, and Singapore, have or are in the process of implementing similar data privacy rules.''), Business Roundtable I (``Finally, countries outside of the EU, including Japan and Singapore, either already have developed domestic data privacy regimes similar to the EU Directive or are in the process of doing so.''), Business Roundtable II, COEC I (stating that, ``in addition to the 27 jurisdictions which have implemented the EU Directive, our survey respondents noted that there are several other countries which have restrictive data privacy laws including China, Japan and Mexico''), Corporate Secretaries (indicating that ``there are 27 countries in the EU that have implemented the EU Privacy Law. . .other countries such as Japan and Singapore have developed or are in the process of developing domestic data privacy regulations''), NAM I (``More specifically, compliance with the data protection laws of each European Union member country, as well as data protection laws of Australia, will be a significant obstacle to collection of necessary information. . . . Indeed, a Manufacturer operating in Russia found that, according to that nation's data privacy laws, the company will need to get the personal sign-off from every Russian employee to share the data with the corporate headquarters.''), and WorldatWork I (``Aside from the EU's laws, there are other countries that have confidentiality laws which may impact this information gathering, such as Argentina's confidentiality laws concerning equity awards.'').

        \624\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        The exemption could potentially provide a competitive advantage to registrants with a significant overall percentage of the workforce located in jurisdictions with data privacy laws over other registrants due to lower compliance costs associated with having fewer workers covered by the rule. However, the limited and tailored nature of the exemption, as well as the reduction or elimination of the de minimis exemption for registrants that exclude employees under the foreign data privacy exemption, as discussed below, mitigates this possibility.

        vi. De Minimis Exemption

        While we define the term ``employee'' to include any U.S. and non-

        U.S. employee of a registrant, the final rule provides for a de minimis exemption for employees in foreign countries, up to 5% of a registrant's workforce, under certain conditions.\625\ This type of exemption was suggested by several commenters,\626\ and it should provide cost savings to eligible registrants. The suggested de minimis amount varied significantly across commenters.\627\

        ---------------------------------------------------------------------------

        \625\ See Section II.B.1.c.iii.

        \626\ See letters from American Benefits Council, ExxonMobil, FSI, FSR, NYC Bar, and PNC Financial Services.

        \627\ See, e.g., letters from American Benefits Council (suggesting that a registrant be permitted to exclude non-U.S. employees in any foreign jurisdiction that comprises less than 5% of the issuer's aggregate global workforce), ExxonMobil (indicating that a registrant be permitted to exclude non-U.S. employees in a foreign jurisdiction if the number of employees in that jurisdiction is less than 1% of the issuer's total workforce), FSR (recommending that non-U.S. employees be excluded if they account for less than 5% of the registrant's total workforce or employees in any single foreign jurisdiction if they comprise less than 2% of total employees with an aggregate cap of 5% (if the registrant's non-U.S. employees account for more than 5% of all employees)), and NACCO (suggesting that a registrant be permitted to exclude non-U.S. employees if they make up less than 20% of the employee population).

        ---------------------------------------------------------------------------

        In the June 4 Memorandum and June 30 Memorandum, staff attempted to quantify the effects of the exemption. However, because staff lacked more specific information about potentially affected registrants, including comprehensive data on the intra-company distribution of compensation of these categories of employees at companies that may be subject to the rule, the analyses necessarily relied on certain assumptions. Commenters also did not provide this data.

        The projections in the two staff memoranda were based on evidence obtained from other studies, aggregate statistics, and other assumptions that may result in over- or underestimating the magnitude of the effect on the pay ratio calculation. The memoranda made the following assumptions: companies have excluded the percent of employees equal to the specified percentage threshold; the distribution of pay is described by a lognormal distribution \628\ (with various estimates of the standard deviation of the log of pay \629\ that broadly incorporate the ranges of estimates from the studies cited in the June 4 Memorandum, 0.25, 0.35, 0.45, and 0.55 \630\); and the level of PEO pay is independent of the exclusion threshold. The estimates of the effect on the pay ratio calculation of excluding different percentages of employees were sensitive to the above assumptions.\631\

        ---------------------------------------------------------------------------

        \628\ A commenter noted that a lognormal distribution may be inadequate for actual firms. See letter from Public Citizen II. As we noted in the Proposing Release, registrants that have multiple business or geographical segments may not necessarily have a lognormal distribution of pay. However, a distributional assumption is necessary for the analysis because staff could not observe the actual distribution of wages within the affected firms in the data available to them. This assumption is motivated by the positive skewness in dollar wages and the distribution of log of wages approximating normal distribution. See, e.g., Blundell, R., Reed, H., Stoker, T., 2003, Interpreting aggregate wage growth: The role of labor market participation, American Economic Review, Vol. 93(4), pp. 1114-1131; Measuring the distribution of wages in the United States from 1996 through 2010 using the Occupational Employment Survey, BLS Monthly Labor Review, May 2014, http://www.bls.gov/opub/mlr/2014/article/measuring-the-distribution-of-wages-in-the-united-states-from-1996-through-2010-using-the-occupational-employment-survey-1.htm). The assumptions used in this analysis are for illustration purposes only. As we note, while the lognormal assumption may be appropriate for some registrants, it may not be appropriate for all registrants. While one commenter suggested that the final rule permit a registrant to determine the median employee based on an assumption that compensation is lognormally distributed within a company or segment, we believe that registrants can and thus should determine for themselves whether the use of a lognormal assumption is appropriate given their own compensation distributions. See letter from TCA. The final rule does not specify any required methodology for registrants to use in identifying the median employee and permits registrants the flexibility to choose a method to identify the median employee based on their own facts and circumstances so long as a registrant's methodology uses reasonable estimates. Indeed, more generally, we believe that it is appropriate for registrants to make their own determinations about whether a particular methodological assumption constitutes a reasonable estimate for their particular firms.

        \629\ As we noted in the Proposing Release, each registrant would have a company-specific compensation variance, which is impossible to be generally assumed.

        \630\ Two commenters noted that these assumptions may understate the variability of employee pay within actual publicly-traded companies, particularly companies with part-time, seasonal or temporary employees. See letters from AFL-CIO II and Public Citizen II. As discussed in the June 4 memorandum, the above standard deviation assumptions could understate intra-firm wage variation in employee pay, which would in turn potentially understate the effects of the exclusion on the pay ratio. However, the staff lacked data on the actual intra-firm distribution of wages for registrants affected by the final rule to perform additional analysis.

        \631\ The letter from COEC III cites additional research on wage dispersion within and between firms not cited in the June 4 Memorandum. See Jae Song, David J. Price, Fatih Guvenen, Nicholas Bloom, and Till von Wachter, Firming Up Inequality, NBER Working Paper 21199, May 2015, available at http://www.nber.org/papers/w21199 (``Song et al. (2015)''). The estimates of within-firm wage variation in this paper are in a format from which the staff cannot directly infer the standard deviation estimates required for its analysis, and the staff lacks access to the source data to compute these standard deviation estimates. The paper concludes that within-

        firm wage inequality changed little over time, which is broadly similar to the conclusion in Barth et al. (2014) based on a different dataset, cited in the June 4 Memorandum. The staff analysis utilizes a range of standard deviation assumptions to illustrate the effects of various potential levels of within-firm wage variability.

        ---------------------------------------------------------------------------

        The June 4 Memorandum, under the assumptions above evaluated the effects on the pay ratio calculation of excluding different percentages (between 1% and 20%) of pay observations from a lognormal distribution for each set of assumptions about intra-company standard deviation of the log of pay (sigma) and for each of the two scenarios below concerning excluded pay observations: Scenario I (all excluded observations are below the median for the underlying distribution of pay); and Scenario II (all excluded observations are above the median for the underlying distribution of pay). Under these scenarios, for a given standard deviation level, the effect on the pay ratio is larger in magnitude when a larger percentage of employees are excluded. For example, the exclusion of 5% of employees may cause the pay ratio to decrease by up to 3.4% in Scenario I or to increase by up to 3.5% in Scenario II (an aggregate range of 6.9%). Under a 20% threshold, the pay ratio may decrease by up to 13% or increase by up to 15% (an aggregate range of 28%), depending on the scenario considered.

        The June 30 Memorandum extended the analysis contained in the June 4 Memorandum by showing, under the same assumptions, the potential effects of excluding percentages greater than 20% and up to 95%. As expected, under the same assumptions, excluding a broader range of exclusion thresholds (between 20% and 95%) yielded a larger magnitude of the effect on the pay ratio for Scenarios I and II. In addition, the June 30 Memorandum included different intermediate scenarios between Scenarios I and II, with some observations excluded from above the median and some from below the median of the underlying

        Page 50169

        distribution.\632\ As expected, under the same assumptions, including these intermediate scenarios yielded effects within the range delineated by Scenarios I and II.\633\

        ---------------------------------------------------------------------------

        \632\ Specifically, the June 30 Memorandum included the following three scenarios: Scenario I(a) (75% of the excluded observations are below the median, with the remaining 25% of the excluded observations above the median); Scenario I(b) (50% of the excluded observations are below the median, with the remaining 50% of the excluded observations above the median); and Scenario I(c) (75% of the excluded observations are above the median, with the remaining 25% of the excluded observations below the median). See June 30 Memorandum. One commenter noted that non-random exclusion of low-paid employees would cause the pay ratio estimate to decline more than random exclusion. See letter from Public Citizen II. Consistent with the commenter, as the staff analysis demonstrates, exclusion scenarios in which over half of the excluded workers are paid below the true median cause the pay ratio estimate to decline. As noted in the June 4 Memorandum, non-U.S. employees of U.S. multinational firms outside the United States on average receive lower compensation than employees located inside the United States. See June 4 Memorandum. However, for some firms with employees outside the United States in highly skilled occupations or firms with employees in jurisdictions with high labor costs, some employees outside the United States may receive higher compensation than U.S. employees.

        \633\ We note that, if observations are equally excluded from either side of the median, the estimated effect is zero regardless of the percentage of employees excluded from the distribution.

        ---------------------------------------------------------------------------

        As the memoranda indicate, under the assumptions considered, excluding 5% of employees yields an effect on the pay ratio in the range between -3.4% and 3.5%.\634\

        ---------------------------------------------------------------------------

        \634\ As mentioned above, we lack information about the actual intra-firm distribution of pay for affected registrants but, even if the true distribution is lognormal, we do not have information to know where the effect may fall within the range.

        ---------------------------------------------------------------------------

        We further recognize that the estimates of the effects of the de minimis exemption on the pay ratio are sensitive to the assumptions made and may understate or overstate the actual magnitude of the effect if any of the above assumptions, for instance, the assumptions about lognormal distribution or magnitude of intra-firm variation in wages, do not hold. If the affected registrant's true intra-firm distribution of the log of employee pay is not normal, depending on the shape of the true distribution, the actual effects on the median may significantly differ from the estimated effects reported in the memoranda. Importantly, if the true intra-firm distribution of pay at an affected registrant deviates from the lognormal assumption, estimates of the effects under these scenarios may correspondingly decrease in accuracy as the percentage of the excluded observations increases.

        We note that the de minimis exemption may not affect some registrants because not all registrants will be eligible to use it or choose to use it to exclude up to 5% of the total workforce. We also note that, in some instances, this exemption may result in the exclusion of employees from jurisdictions with low pay, which may increase the difficulty of interpreting the pay ratio. The requirements to disclose the jurisdiction(s) and the approximate number of employees from each jurisdiction being excluded should mitigate this concern.

        We have considered several reasonable alternatives to the final rule's 5% de minimis exemption. One alternative would be to apply a different de minimis threshold. A lower de minimis percentage may increase registrants' costs of calculating the ratio for workers in non-U.S. jurisdictions. Lowering the de minimis threshold below 5% would not meaningfully reduce the impact on the pay ratio under the assumptions in our analysis. A higher de minimis threshold could yield potentially larger savings in the costs of calculating the ratio for registrants with workers in non-U.S. jurisdictions. However, as seen in Table 1 in the June 4 Memorandum, such a threshold would have a potentially larger effect on the pay ratio than the 5% threshold. Specifically, as discussed above, under the assumptions made and depending on the scenario considered, the exclusion of 10% of employees may decrease the pay ratio by up to 6.7% or increase it by up to 7.2%; the exclusion of 15% of employees may decrease the pay ratio by up to 9.9% or increase it by up to 11%; the exclusion of 20% of employees may decrease the pay ratio by up to 13% or increase it by up to 15%. Each of these alternatives thus could result in an impact on pay ratio that is greater than the impact under the de minimis threshold.\635\

        ---------------------------------------------------------------------------

        \635\ See Section II.B.1.c.iii(c) for the definition of the de minimis threshold at 5%.

        ---------------------------------------------------------------------------

        Under alternative de minimis exemptions suggested by commenters, registrants would be permitted to exclude non-U.S. employees in every foreign country that comprises less than 1% \636\ or less than 5% \637\ of the registrant's aggregate global workforce. These alternative definitions of the de minimis exemption can reduce calculation costs for U.S. multinational registrants with a high level of international diversification in their workforce. However, they may potentially result in the exclusion of a large percentage of employees at registrants with a large percentage of non-U.S. employees diversified across countries, which may affect the pay ratio considerably, as we indicate in the discussion above. These alternatives may also offer a larger relative competitive advantage to internationally diversified U.S. registrants compared to U.S. registrants with the same total percentage of non-U.S. employees concentrated in fewer countries and thus ineligible for the exemption under these alternatives.

        ---------------------------------------------------------------------------

        \636\ See letter from ExxonMobil.

        \637\ See letter from American Benefits Council.

        ---------------------------------------------------------------------------

        A different alternative exemption proposed by a commenter would permit registrants to exclude all employees in any single foreign jurisdiction if they comprise less than 2% of total employees, with an aggregate cap of 5% (if the registrant's foreign employees account for more than 5% of all employees).\638\ The exemption in our final rule is defined more broadly than this alternative definition and enables savings in calculation costs for a potentially larger fraction of registrants with non-U.S. workers.

        ---------------------------------------------------------------------------

        \638\ See letter from FSR.

        ---------------------------------------------------------------------------

        Registrants that are eligible for the de minimis exemption and choose to use it may have lower compliance costs than registrants that do not use the exemption.\639\ By excluding foreign workers, the exemption makes eligible registrants with foreign operations more similar to registrants with U.S.-based operations only and reduces the number of employees considered in the identification of the median employee, with the effect of reducing compliance costs for the eligible registrants. Relying on some reasonable assumptions, we are able to quantify some of the cost savings from the de minimis exemption. First, we assume that all registrants with foreign operations will use the exemption and eliminate 5% of their workforce. We also assume that the savings in compliance costs are directly proportionate to the number of employees excluded and that the compliance cost per excluded foreign employee is equal to the estimate for firms with some foreign employees, $38.04. Using these assumptions and our estimates of the number of registrants with foreign operations and the total number of employees for these registrants from Table 3, the total savings from the use of the de minimis

        Page 50170

        exemption would be approximately $52.5 million.\640\

        ---------------------------------------------------------------------------

        \639\ Some commenters noted that the staff analysis did not incorporate a cost-benefit analysis. See letters from COEC III and WorldatWork II. Potential cost savings from the de minimis exemption are discussed in this section. See Section III.D.2.c.iv, above, for a discussion of potential cost savings from the exclusion of all non-U.S. employees. Commenters did not provide more detailed estimates of the cost savings from excluding some but not all non-

        U.S. employees.

        \640\ The savings are calculated as 5% * (27,595,305 * $38.04) for firms with some foreign employees. Under similar assumptions, if companies with foreign operations were permitted to exclude 10%, 20%, 30%, and 40% of their employees, respectively, potential cost savings could amount to that percentage multiplied by (27,595,305 * $38.04) or $105 million, $210 million, $315 million, and $420 million, respectively. Actual cost savings may differ if there are fixed costs associated with the inclusion of non-U.S. employees in the pay ratio calculation. Another commenter estimated from a survey that the mid-range percentage of non-U.S. employees was 40% and that the mid-range cost savings from excluding all non-U.S. employees would be 50%. See letter from COEC.

        ---------------------------------------------------------------------------

        We note that actual cost savings incrementally attributed to the de minimis exemption are likely to be lower. Some registrants may have less than 5% of the total workforce outside the United States. Other registrants may have more than 5% of the total workforce outside the United States but less than 5% of the total workforce in aggregate across foreign jurisdictions in which it can exclude employees under the de minimis exemption (with no more than 5% of the total workforce in each such jurisdiction). The incremental cost savings from the de minimis exemption are further reduced or potentially eliminated for registrants that exclude some employees under the foreign data privacy exemption as such registrants may be ineligible for the de minimis exemption or eligible for the exemption but unable to exclude employees at the maximum level under the de minimis exemption due to the concurrent use of the foreign data privacy exemption, concentrated nature of their non-U.S. workforce, or a combination of the two factors.\641\ Other registrants with non-U.S. workers may elect not to use the exemption. We also note that the actual cost savings could vary significantly depending on whether cost savings increase uniformly with the percent of employees excluded and whether registrants can exclude any employees outside the United States under the foreign data privacy law exemption.

        ---------------------------------------------------------------------------

        \641\ The final rule does not limit a registrant's ability to rely on the foreign data privacy exemption, provided the conditions of the exemption are met.

        ---------------------------------------------------------------------------

        To the extent that registrants with non-U.S. workers experience a higher cost of calculating the pay ratio than registrants with U.S. workers only and that the de minimis exemption reduces such costs, it reduces the potential adverse competitive effects of the pay ratio disclosure requirement on registrants with non-U.S. workers eligible for the exemption. Registrants with non-U.S. workers concentrated in fewer foreign jurisdictions are more likely to exceed the 5% threshold for any single foreign jurisdiction and thus be ineligible for the exemption than registrants with the same total percentage of non-U.S. workers diversified across more foreign jurisdictions. If the inability to use the de minimis exemption increases the costs of calculating the pay ratio significantly (for instance, when the overall percentage of non-U.S. workers is higher than but relatively close to 5%), registrants with geographically concentrated non-U.S. workers may be at a relative competitive disadvantage.

        vii. Calculation Date

        The final rule permits registrants to choose any date within three months of the end of registrant's fiscal year to identify the median employee for that year and requires registrants to disclose the date used. Compared with prescribing a given date, such as the last day of the completed fiscal year, as proposed, this approach may reduce compliance costs by providing flexibility to registrants that may not have enough time to collect and report on their pay ratio information at year-end. Commenters suggested that allowing registrants to select the date would allow them to pick a date that does not coincide with other required reporting or that better utilizes the internal resources of the registrants.\642\ This approach also might reduce compliance costs for registrants that use many employees at the end of the calendar year by permitting those registrants to choose a date on which those seasonal or temporary employees are not employed.\643\ Registrants in industries with more fluctuations in employment within the year due to seasonality may realize larger benefits from this approach.\644\ Hence, it is possible that registrants could choose a date and structure their employment arrangements around that date to reduce the number of workers employed on the calculation date or to alter the reported ratio to achieve a particular objective with the pay ratio disclosure. One commenter specifically addressed this issue and noted its belief that this concern ``is unwarranted, particularly if the choice is restricted to a limited time period (such as the last fiscal quarter), since in general the employee population of a registrant would not vary significantly over such a period.'' \645\ Another commenter suggested that not allowing such an adjustment will produce ``artificially low'' median employee pay for registrants that have many temporary and seasonal workers at year-end.\646\ Yet another commenter recommended allowing flexibility with respect to the measurement date but requiring that the calculation include annual compensation of all employees employed at any time over the preceding 365 days.\647\ However, we note that such an alternative would increase the cost for registrants with significant fluctuations in the number of employees within the year relative to the final rule. Based on the comments we received, we believe that the rule as adopted will reduce compliance costs compared to the proposed rule. However, we did not receive data that would allow us to quantify the cost reduction.

        ---------------------------------------------------------------------------

        \642\ See, e.g., letters from Chamber I, COEC I, COEC II, and Microsoft.

        \643\ See, e.g., letters from American Benefits Council, NACCO, and PM&P.

        \644\ For example, analysis by staff in the Division of Economic and Risk Analysis of BLS employment statistics by industry for 2014 revealed that the following sectors appear to have the largest employment fluctuations due to seasonality, as proxied by the average magnitude of the percentage difference between seasonally adjusted and not seasonally adjusted employment: arts, entertainment and recreation, educational services, and construction, as well as administrative and support services, accommodation and food services (when data for all months is used) and retail trade and warehousing and storage (when data for the final three months is used, to account for the fiscal year end being December for the majority of registrants). See http://www.bls.gov/data/#employment. Agricultural employment is excluded from the above calculations. It is also likely to be subject to significant fluctuations within the year due to seasonality. We note that BLS data may not be representative of the employment fluctuations within the year for the registrants subject to the final rule and that the value of the flexibility to select the calculation date will vary across registrants.

        \645\ See letter from Davis Polk.

        \646\ See letter from PM&P.

        \647\ See letter from AFL-CIO II.

        ---------------------------------------------------------------------------

        d. Adjustments to the Compensation of Employees

        The final rule includes an instruction that permits a registrant to annualize the compensation for all permanent employees (full-time or part-time) employed on the calculation date who did not work for the registrant for the full fiscal year. The final rule does not permit annualization for employees in temporary or seasonal positions. The final rule also does not permit the use of full-time-equivalent adjustments for any of a registrant's employees in the required pay ratio disclosure, although such adjustments are permitted to derive an additional ratio if the registrant chooses. We believe that our approach provides appropriate accommodations to registrants to represent the annual composition of

        Page 50171

        their workforce without significantly diminishing the potential usefulness of the disclosure mandated by Section 953(b).

        We believe that by permitting annualization adjustments for permanent employees but not seasonal or temporary ones, our approach more closely captures the composition of a registrant's workforce and compensation practices. For these annualizing adjustments to have any significant impact on the reported pay ratio, both the fraction of permanent new hires to all employees of the registrant and their annualized compensation would have to be relatively large. We also note that some commenters were supportive of allowing annualizing adjustments.\648\ One commenter suggested that this adjustment will make the ratio more representative of the registrant's labor arrangements.\649\ This procedure is purely optional and registrants do not need to annualize compensation, such as if they believe that the additional cost of the adjustment does not warrant the perceived benefit.

        ---------------------------------------------------------------------------

        \648\ See, e.g., letters from ABA, Corporate Secretaries, Intel, NACCO, PM&P, SH&P, and WorldatWork I.

        \649\ See letter from Corporate Secretaries (``We believe that allowing annualizing adjustments effectuates the Dodd-Frank Act requirement that each registrant disclose the median of the annual total compensation of its employees. Such adjustments would result in a calculation that is closer to a fair and reasonable representation of the registrant's actual compensation practices and labor costs. The adjustment also would help eliminate the potential distorting effects of mid-year hires, including those that result from a merger or acquisition'').

        ---------------------------------------------------------------------------

        By permitting registrants to annualize compensation for these employees, the comparability of disclosure across registrants could be reduced compared to an alternative of either requiring or prohibiting such annualization. As noted above, however, we believe that precise comparability of disclosure from registrant to registrant could be difficult to achieve due to the variety of factors that could cause the ratio to differ. Accordingly, we do not believe that the costs associated with promoting precise comparability would be justified.

        Another alternative would have been to permit a registrant to annualize the compensation for all temporary and seasonal employees who were employed for less than the full fiscal year and to use that annualized compensation in the mandated pay ratio disclosure. Some commenters supported this alternative and indicated that not allowing an annualizing adjustment for these employees would distort the pay ratio.\650\ Some commenters urged us to permit the use of full-time equivalent adjustments for part-time employees, temporary, and seasonal employees.\651\ The final rule does not permit the mandated pay ratio disclosure to include either the use of annualized adjustments for seasonal or temporary employees or the use of full-time-equivalent adjustments for part-time employees. We believe that such adjustments would reflect a different workforce composition and compensation structure than that utilized by the registrant. To the extent that a registrant relies primarily on part-time, temporary, or seasonal workers, computing a ratio based on their annualized compensation of temporary and seasonal workers or the full-time-equivalent of a part-

        time worker, unlike annualizing adjustments for permanent employees, could have a significant impact on the ratio.

        ---------------------------------------------------------------------------

        \650\ See, e.g., letters from AAFA II, American Benefits Council, Brian Foley & Co., Corporate Secretaries, Hay Group, KBR, NIRI, and NYC Bar.

        \651\ See supra note 108.

        ---------------------------------------------------------------------------

        Although we are not permitting full-time-equivalent adjustments or annualization adjustments for seasonal and temporary employees to be made for purposes of calculating the annual total compensation in the mandated pay ratio disclosure, the final rule does permit registrants to provide additional disclosure. For example, registrants can report additional ratios, including ratios that reflect one or more of those adjustments, if they choose, provided that any additional ratio is clearly identified, not misleading, and not presented with greater prominence than the required ratio.

        The final rule permits but does not require registrants to adjust compensation to the cost of living in the PEO's jurisdiction of residence. While some commenters stated that international differences in the cost of living can distort the reported pay ratio,\652\ one commenter suggested that compensation is more directly comparable to the total compensation of a registrant's PEO without a cost-of-living adjustment.\653\ Moreover, as some commenters suggested,\654\ the cost-

        of-living adjustment may alleviate concerns about pay comparability between U.S. employees and those non-U.S. employees that are located in foreign countries where the purchasing power of the average foreign pay in dollar terms deviates significantly from the purchasing power of the average domestic pay in dollar terms.

        ---------------------------------------------------------------------------

        \652\ See, e.g., letters from AAFA I, American Benefits Council, Corporate Secretaries, and ExxonMobil.

        \653\ See letter from ABA.

        \654\ See letters from NAM I, NAM II, and SH&P.

        ---------------------------------------------------------------------------

        Providing the option to use a cost-of-living adjustment is not expected to increase the compliance cost for registrants. Country-level cost-of-living data is widely available. The incremental cost of identifying the median employee based on pay without the cost-of-living adjustment and calculating the pay ratio without the cost-of-living adjustment is expected to be small once pay data for all employees or for samples of employees from individual countries have been obtained. Thus, registrants that believe the cost-of-living adjusted ratio to be more meaningful given the structure of their workforce may benefit from the option to present the pay ratio with the cost-of-living adjustment as the ratio required by Item 402(u)(1)(iii).

        The cost-of-living adjustment of the compensation of a registrant's employees may have an effect on the determination of the median employee and on the calculation of the pay ratio for registrants with employees in countries whose cost of living differs from the cost of living in the PEO's country of residence. We are limited in our ability to quantify the impact of this adjustment on the pay ratio calculation by our lack of data on the intra-firm distribution of pay of employees outside the PEO's country of residence for the affected registrants and by limited data available to us on the distribution of employees by country at the individual registrant level. As noted elsewhere, because we lack data regarding intra-firm distributions, we cannot predict the effects of a cost-of-living adjustment on those distributions, as the adjustment may, in some cases, have an effect on the combined employee pay distribution at the individual registrant level by potentially changing the median employee within the same country or by locating the median employee in a different country. We therefore analyze qualitatively the main factors that may contribute to more significant effects of the cost-of-living adjustment on the determination of the median employee and on the calculation of the pay ratio.

        The cost-of-living adjustment option could affect the pay ratio calculation for registrants with some employees located outside the PEO's country of residence that elect to use this option. The effect of the cost-of-living adjustment could be potentially larger for registrants with a larger percentage of employees outside the PEO's country of residence and for registrants with employees in countries with a cost of living that differs significantly from the PEO's country of residence.

        Page 50172

        According to the results of one commenter's survey, the average (median) respondent had 62% (60%) of its employees located in the United States.\655\ According to another commenter's survey, approximately 55% of respondents reported having the majority of their employees located in the United States.\656\ As set forth in Table 3 above, out of the registrants potentially subject to the final rule for which we have data on the presence of geographic segments,\657\ approximately a third are estimated to have U.S.-based operations only and two-thirds are estimated to have some non-U.S. operations. Based on aggregate BEA data on U.S. multinational companies for 2012,\658\ employees at foreign affiliates and at majority-owned foreign affiliates of U.S. multinational companies comprised approximately 38% and 33-34%, respectively, of the total employment of U.S. multinational companies. We note that the respondents in the surveys cited above, companies that report geographic segments, and companies in the BEA sample may not be representative of the full set of registrants subject to the final rule, and the PEO's country of residence may be different than the United States.

        ---------------------------------------------------------------------------

        \655\ See letter from Corporate Secretaries.

        \656\ See letter from COEC I.

        \657\ Compustat segments database reports data on the presence of geographic segments for approximately 63% of registrants potentially subject to the final rule.

        \658\ See U.S. BEA data on Direct Investment & Multinational Enterprises (MNEs) available at http://bea.gov/iTable/index_MNC.cfm.

        ---------------------------------------------------------------------------

        Based on aggregate BEA data on the distribution of employees of U.S. multinational companies by country for 2012,\659\ the majority of employees at foreign affiliates of U.S. multinational companies were estimated to be located in the following regions: Europe (34%, including 30% in the EU); Asia and Pacific (34%); and Latin America (20%). Looking at individual jurisdictions, the most employees of foreign affiliates of U.S. multinational companies were estimated to be located in the following ten countries: China (11.2%); the United Kingdom (10.3%); Mexico (9.8%); Canada (8.4%); India (6.9%); Germany (4.9%); Brazil (4.6%); Japan (3.7%); France (3.5%); and Australia (2.5%), which in aggregate accounted for 65.8% of foreign employees of U.S. multinational firms.\660\

        ---------------------------------------------------------------------------

        \659\ Id.

        \660\ Id.

        ---------------------------------------------------------------------------

        However, these aggregate statistics on the location of employees by country do not capture the distribution of employees by countries at individual registrants. While these aggregate statistics may offer an average perspective across all firms with a non-U.S. workforce in the BEA sample, they do not enable us to draw strong conclusions about the ultimate effects on the pay ratio for those registrants that are subject to final rule and decide to opt for the cost-of-living adjustment. We believe that registrants anticipating an increase in the pay ratio after the adjustment may be less likely to opt for the cost-

        of-living adjustment. Based on 2014 data from the International Monetary Fund (``IMF''),\661\ we note that the E.U. on aggregate had a cost of living similar to the U.S. level. Based on 2014 data on PPP conversion factors from the World Bank,\662\ of the above locations, the United Kingdom, Canada, Germany, Japan, France, and Australia had a cost of living similar to or above the U.S. level. Based on the same measure, of the above locations, India, China, Mexico, and Brazil had a cost of living below the U.S. level.\663\ As we noted above, the actual effects of cost-of-living adjustment on the pay ratio calculation will depend on the countries where employees are located, the actual distribution of employee pay, and the specific cost-of-living measure used.

        ---------------------------------------------------------------------------

        \661\ We lack PPP estimates for the E.U. in aggregate. An indirect proxy is the PPP conversion factor implied by the E.U. gross domestic product (``GDP'') reported in dollar terms and PPP-

        adjusted dollar terms. See IMF World Economic Outlook data by country groups, available at http://www.imf.org/external/pubs/ft/weo/2015/01/weodata/weoselagr.aspx.

        \662\ PPP conversion factor is the ratio of the PPP exchange rate to the nominal dollar exchange rate, available at http://data.worldbank.org/indicator/PA.NUS.PPPC.RF. This ratio makes it possible to compare the cost of the bundle of goods that make up GDP across countries. It measures how many dollars are needed to buy a dollar's worth of goods in the country as compared to the United States.

        \663\ Id. PPP conversion factors calculated based on the 2011 International Comparison Program round and 2014 market exchange rates for the countries above are: China (0.6); the United Kingdom (1.2); Mexico (0.6); Canada (1.1); India (0.3); Germany (1.0); Brazil (0.7); Japan (1.0); France (1.1); and Australia (1.4).

        ---------------------------------------------------------------------------

        Below we illustrate the potential effect of the cost-of-living adjustment on the pay ratio for a hypothetical registrant.\664\ In this example, we make the following assumptions: We assume that the registrant has 70% of employees in the country of residence of the PEO and 30% of employees in another country; the pay of the registrant's employees in the PEO's country, expressed in the currency of the PEO's country, is lognormally distributed (with mean log of pay of 10.5 and standard deviation of log of pay of 0.5); the pay of the registrant's employees in the other country, expressed in the currency of the PEO's country, is lognormally distributed (with mean log of pay of 9 and standard deviation of log of pay of 0.5); and the cost of living is two times higher in the country of residence of the PEO than in the other country. In this hypothetical example, a cost-of-living adjustment would cause the pay ratio to decrease by approximately 6.4%. If the cost of living were three times higher in the country of residence of the PEO than in the other country, holding other assumptions unchanged, a cost-of-living adjustment would cause the pay ratio to decrease by approximately 14.9%.

        ---------------------------------------------------------------------------

        \664\ This is intended only as a hypothetical example for illustration purposes and not as an indication of the effects at an actual or representative registrant. The effects are highly sensitive to the assumptions described. The estimates are obtained numerically.

        ---------------------------------------------------------------------------

        Some commenters \665\ suggested that a cost-of-living adjustment could introduce an element of subjectivity into the pay ratio calculation or permit registrants to alter the reported ratio to achieve a particular objective with the ratio disclosure. In the final rule, the requirements to apply a consistent methodology, to disclose the use of the adjustment, and to provide disclosure of the registrant's pay ratio calculated without the cost-of-living adjustment should address these concerns.

        ---------------------------------------------------------------------------

        \665\ See letters from Prof. Ray and WorldatWork I.

        ---------------------------------------------------------------------------

        e. Frequency of Identifying the Median Employee

        Unlike the proposed rule, which required registrants to identify the median employee every year, the final rule allows registrants to identify the median employee once every three years unless there has been a change in the registrant's employee population or employee compensation arrangements that it reasonably believes would result in a significant change in the pay ratio disclosure. Registrants must still provide annual disclosure of their pay ratio by recalculating the previously identified median employee's annual total compensation each year.

        Under this approach, a registrant may identify its median employee for year one and then use that employee or one who has substantially similar compensation as its median employee in the following two years for calculating the employee's annual total compensation and the registrant's pay ratio. A couple of commenters suggested this approach, noting that it would still result in a registrant providing a pay ratio disclosure on an annual basis while reducing the burden and costs required to identify the median employee annually when there have not been any interim changes in the

        Page 50173

        registrant's workforce or compensation structure.\666\

        ---------------------------------------------------------------------------

        \666\ See, e.g., letters from ABA, COEC I, and COEC II.

        ---------------------------------------------------------------------------

        Choosing this approach is likely to result in lower ongoing compliance costs for affected registrants. We expect that some registrants will identify the median employee every three years, while others may identify it every year or every two years. Thus, depending on how frequently registrants would have to identify the median employee, and on whether the identification of the median employee is the main ongoing cost of compliance, with the change in the final release, the ongoing compliance costs could range approximately from $123 million to $947 million per year.\667\

        ---------------------------------------------------------------------------

        \667\ Ongoing costs without this adjustment are estimated to be between $368 million and $947 million. See Section III.C.2.c. Ongoing costs with the adjustment are estimated as follows. The $123 million is estimated as $368 million divided by three, based on the lowest of the estimates of ongoing compliance costs and the assumptions that all registrants identify the median employee every three years and the identification of the median employee accounts for the entirety of the ongoing cost. This estimate represents the aggregate annual cost in this scenario averaged over three years. The $947 million is based on the highest of the estimates of ongoing compliance costs and the assumption that all registrants incur the full ongoing cost every year.

        ---------------------------------------------------------------------------

        f. Method of Identifying the Median Employee

        In order to allow the greatest degree of flexibility while maintaining consistency with the statutory provision, the final rule does not specify a particular methodology for identifying the median. Instead, it allows registrants a choice of multiple methods, including several with significant flexibility.

        We are adopting this flexible approach because we believe that the appropriate and most cost-effective methodology for identifying the median employee necessarily depends on a registrant's particular facts and circumstances, including, among others, such variables as size and nature of the workforce, complexity of the organization, the stratification of pay levels across the workforce, the types of compensation the employees receive, the extent that different currencies are involved, the number of tax and accounting regimes involved, the number of payroll systems the registrant has, and the degree of difficulty involved in integrating payroll systems to readily compile total compensation information for all employees. We believe that these are likely the same factors that would cause substantial variation in the costs of compliance. By not prescribing specific methodologies that must be used, the final rule allows registrants to choose a method to identify the median employee that is appropriate to the size, structure, and compensation practices of their own businesses, including permitting a registrant to identify the median employee using any consistently applied compensation measure.

        In addition, the final rule's flexibility could enable registrants to manage compliance costs more effectively than a more prescriptive approach would allow.\668\ We also believe that, by allowing registrants to minimize direct compliance costs, a flexible approach could mitigate, to some extent, any potential negative effects of the mandated requirements on competition. We recognize, however, that a flexible approach could increase uncertainty for registrants that prefer more specificity on how to comply with the final rule, particularly for registrants that do not use statistical analyses in the ordinary course of managing their businesses. In light of this potential uncertainty, the final rule establishes certain parameters on the use of this flexibility, such as by specifying that the use of statistical sampling or other reasonable estimates in identifying the median is permitted, as is identifying the median employee based on any consistently applied compensation measure.

        ---------------------------------------------------------------------------

        \668\ For example, one commenter estimated that using all elements of compensation rather than the sum of salary earned, incentive cash earned, and stock awards granted to identify the median employee in a worldwide workforce would increase the initial expense and ongoing workload by a factor of more than five times. See letter from Microsoft.

        ---------------------------------------------------------------------------

        We believe that a flexible approach would not significantly diminish the potential benefits of the disclosure mandated by Section 953(b). As discussed above, we believe that the intended purpose of the pay ratio disclosure is to provide shareholders with a company-specific metric to evaluate the PEO's compensation, rather than a benchmark for compensation arrangements across registrants. Also as discussed above, we are not persuaded that mandating a particular methodology will necessarily improve the comparability of pay ratio disclosure across registrants because of the numerous other factors that could also cause the ratios to be less meaningful for registrant-to-registrant comparison. Even if such comparability could be marginally enhanced by mandating a specific method for identifying the median, we do not believe this marginal improvement in comparability would be justified in light of the costs that would be imposed on registrants by a more prescriptive rule. We also note that some commenters expressed the view that greater comparability across registrants could increase the likelihood that a registrant's competitors could infer proprietary or sensitive information about the registrant's business. This in turn could increase the indirect costs to registrants of the adopted requirements, such as competitive harms in labor markets discussed in the previous section or general costs arising from the mandated disclosure requirement.

        Finally, we recognize that allowing registrants to select a methodology to identify the median, rather than prescribing a methodology or set of methodologies, could reduce the benefits for shareholders if that flexibility results in a pay ratio statistic that is less useful than a more precisely and consistently calculated ratio. In particular, some commenters claimed that permitting flexibility in the rule would allow registrants to manipulate the ratio in their favor.\669\ While we acknowledge that the flexibility we are providing creates some risk that registrants will attempt to use this flexibility to produce a more favorable pay ratio, we think that this risk is mitigated by the disclosures we are requiring with respect to the methodologies and assumptions used to identify the median employee, as discussed below. The final rule specifically discusses two particular permitted methods of identifying the median employee--using a consistently applied compensation measure and using statistical sampling. For all of these reasons, we believe the benefits of the final rule's flexibility outweigh those of a more prescriptive approach.

        ---------------------------------------------------------------------------

        \669\ See, e.g., letters from Bupp, Corayer, Fedewa, Fox, Friend, Grotzke, Hlodnicki, Kizzort, Maly, Petricoin, and Van Pelt.

        ---------------------------------------------------------------------------

        i. Consistently Applied Compensation Measure

        We proposed to allow registrants to use any consistently applied compensation measure, such as amounts derived from the registrant's payroll or tax records, to identify the median employee and then calculate that median employee's annual total compensation in accordance with Item 402(c)(2)(x). We are adopting this approach as proposed.

        Allowing registrants this flexibility is likely to reduce registrants' compliance costs significantly, compared to the alternative of requiring registrants to calculate total compensation in accordance with Item 402(c)(2)(x) for all employees, or for a statistically valid sample, and then identify the median.

        Page 50174

        This view was shared by commenters.\670\ Registrants that choose this approach will be able to identify a median employee from employee compensation data that they may already track or record or that may be less expensive for them to acquire than obtaining and computing all of the Item 402(c)(2)(x) compensation information for each employee. Using one commenter's survey results,\671\ we can estimate the potential savings resulting from our exercise of discretion. According to the survey, which provided both average and median cost increase estimates, costs would increase on average by 4,689% if registrants were required to calculate total compensation in accordance with Item 402(c)(2)(x) for all employees. The median increase in compliance costs would be 175%.\672\ Applying this median percentage to our initial compliance cost estimate of $1,315 million, this alternative would increase the total compliance costs to approximately $3,616 million. Thus, we estimate that allowing registrants flexibility in identifying the median employee could result in a total savings of approximately $2,301 million.

        ---------------------------------------------------------------------------

        \670\ See, e.g., letters from CalPERS (``By offering companies a number of alternatives, companies will be able to determine which methodology works best for their company and/or tailor it for their special circumstances. Moreover, this flexibility will allow companies to select a methodology that is most cost effective for them. Finally, since companies will already have the dataset necessary for this calculation (in order to prepare their financial statements and tax returns), we do not envisage costs will be a barrier to compliance.'') and COEC I (``The value of this flexibility appears to be significant in terms of cost. Indeed, survey respondents were asked how their compliance costs would be affected in the event they were required to calculate median employee compensation using the same method employed in the `Summary Compensation Table.' This method is currently used to calculate total primary executive officer compensation and is equivalent to the first approach offered by the SEC. 99 percent of respondents answered that their costs would increase if they were forced to calculate median employee compensation using the Summary Compensation Table approach. Including all responses, the median increase is reported as 100 percent. The data strongly indicate that adhering to the Summary Compensation Table approach would lead to additional significant increases in compliance costs relative to the Proposed Rule'').

        \671\ See letter from Corporate Secretaries.

        \672\ Another commenter, based on survey results, argued that the average increase in costs would be 4,592%, but did not provide a median estimate. See letter from COEC I.

        ---------------------------------------------------------------------------

        We acknowledge, however, that some registrants will still incur costs if they have to combine or sample from separately maintained payroll systems across segments and/or geographic locations.

        ii. Statistical Sampling

        The final rule, as proposed, also allows registrants to use statistical sampling in their determination of the median employee. The size of the reduction in compliance costs that can be achieved by using statistical sampling or other reasonable estimates in identifying the median employee ultimately depends on a registrant's particular facts and circumstances. Below we provide an illustration of how various registrants' characteristics might affect the sampling size. We note that these numbers are intended to provide examples and should not be treated as recommendations about the appropriate sampling size. For example, in the following figure and tables, we show that the variance of underlying wage distributions can materially affect the appropriate sample size for statistical sampling.\673\ Industries characterized by the BLS as having low wage variances, such as electric power generation, coal mining, and metal ore mining, have estimated minimum appropriate sample sizes for an accurate median estimate of less than 135 employees. In contrast, industries characterized by high wage variances, such as offices of physicians, health and personal care stores, and spectator sports, have estimated minimum appropriate sample sizes of more than 1,263 employees. Figure 2 shows the distribution of estimated minimum appropriate sample sizes for registrants operating in each of the 290 4-digit NAICS industries tracked by the BLS.\674\

        ---------------------------------------------------------------------------

        \673\ The analysis uses average and median wage estimates from the BLS at the 4-digit NAICS industry level (290 industries) and assumes a lognormal wage distribution. We use a 95% confidence interval with relative 0.5% margin of error in the estimate of the average of the logarithm of wage. We estimate the median wage by taking the exponential of the sample average of the logarithm of wage, which is the sample geometric average. This median estimator is the maximum likelihood estimator (``MLE'') of the population median for lognormal distribution, and it is an unbiased estimator when the sample size is large. The 95% confidence interval for the population wage median can be obtained by taking the exponential of the endpoints of the 95% confidence interval for the sample average of logarithm of wage (E.L. Crow and K. Simizu, Lognormal Distributions: Theory and Applications 29, (Marcel-Dekker: New York, 1988), R. Serfling, Efficient and Robust Fitting of Lognormal Distributions, North American Actuarial Journal 6, 95-109 (2002), G. Casella & R. L. Berger, Statistical Inference 320 (Duxbury, 2nd ed. 2002), T. B. Parkin, and J. A. Robinson, Statistical Evaluation of Median Estimators for Lognormally Distributed Variable, Soil Science Society of America Journal 57, 317-323 (1993). The lognormal wage distribution assumption is supported by the following studies: F. Clementi, and M. Gallegati, Pareto's Law of Income Distribution: Evidence for Germany, the United Kingdom, and the United States. Econophysics of Wealth Distributions, New Economic Window. 3-14 (2005), and J. Loacutepez and L. Serveacuten, A Normal Relationship? Poverty, Growth and Inequality. World Bank Policy Research Working Paper 3814 (2006). See also M. Pinkovskiy and X. Sala-i-Martin, Parametric Estimations of the World Distribution of Income, NBER working paper 15433, (2009). It is common in practice to control the relative margin of error (instead of absolute margin of error) to determine the sample size. Accordingly, the sample size depends on the coefficient of variation (CV) of the underlying distribution. The square of the CV, also known as ``relative variance,'' is often more stable and easier to guess in advance than variance (W. G. Cochran, Sampling Techniques 77, (New York: Wiley, 3rd ed. 1977); S. L. Lohr, Sampling: Design and Analysis 46-47, (Cengage Learning, 2nd ed. 2009). This analysis also assumes that when the sampling is implemented, the sampling method would be a true random sampling (i.e., it would not be biased by region, occupation, rank, or other factor).

        \674\ Our analysis excludes the Motor Vehicle Manufacturing and Postal Service industries from our sample because, for these industries, mean wage was lower than median wage in dollar terms, appearing to contradict our lognormal distributional assumption.

        Table 6--The Industries With the Smallest and Largest Appropriate Sample Sizes

        ----------------------------------------------------------------------------------------------------------------

        Example of

        Industry Average wage Median wage registrant

        ($) ($) sample size

        ----------------------------------------------------------------------------------------------------------------

        10 Industries With Smallest Variance in Wage Distribution

        ----------------------------------------------------------------------------------------------------------------

        Electric Power Generation, Transmission and Distribution........ 72,800 70,380 84

        Coal Mining..................................................... 55,740 53,310 116

        Metal Ore Mining................................................ 58,000 55,060 135

        Software Publishers............................................. 96,730 90,390 160

        Wired Telecommunications Carriers............................... 65,580 61,510 162

        Natural Gas Distribution........................................ 74,270 69,350 170

        Rail Transportation............................................. 59,990 56,120 172

        Computer and Peripheral Equipment Manufacturing................. 94,850 88,160 174

        Page 50175

        Interurban and Rural Bus Transportation......................... 35,950 33,690 184

        School and Employee Bus Transportation.......................... 33,440 31,200 199

        ----------------------------------------------------------------------------------------------------------------

        10 Industries With Largest Variance in Wage Distribution

        ----------------------------------------------------------------------------------------------------------------

        Offices of Physicians........................................... 72,040 40,510 1,572

        Health and Personal Care Stores................................. 41,890 27,060 1,290

        Spectator Sports................................................ 42,540 27,680 1,263

        Agents and Managers for Artists, Athletes, Entertainers, and 71,960 45,100 1,251

        Other Public Figures...........................................

        Motion Picture and Video Industries............................. 56,540 36,420 1,226

        Home Health Care Services....................................... 37,780 25,100 1,225

        Cut and Sew Apparel Manufacturing............................... 34,800 23,580 1,180

        Amusement Parks and Arcades..................................... 29,580 20,800 1,095

        Apparel, Piece Goods, and Notions Merchant Wholesalers.......... 52,350 35,800 1,063

        Other Professional, Scientific, and Technical Services.......... 48,140 33,150 1,059

        ----------------------------------------------------------------------------------------------------------------

        GRAPHIC TIFF OMITTED TR18AU15.004

        Because these estimated minimum appropriate sample sizes are based on wage distributions measured by the BLS in standardized industries, they may not correspond to the appropriate minimum sample size at registrants with an employee base that does not correspond precisely to one of these industries. Even for registrants whose operations are wholly within one of these standardized industries, their appropriate sample size may also be different to the extent that their distribution of employee wages is different than that of the industry. In these instances, a registrant's appropriate sample size could be higher or lower than that estimated for its industry.

        In 2014, of the nearly 3,571 registrants that we believe will be subject to the final rule, we estimate that approximately 68% and 63% report business and geographic segments, respectively. Approximately 50% and 65% of the potentially affected registrants that self-report business and geographic segments, respectively, report a single segment of that type.\675\ Of the registrants that self-report a single business segment for which we have industry classifications that match the BLS data, Table 7 shows estimated minimum appropriate sample sizes assuming that each registrant's wage

        Page 50176

        distribution is similar to the BLS-measured industry distribution.

        ---------------------------------------------------------------------------

        \675\ This estimate is based on data from Standard and Poor's Compustat Segment database. We note that the segment information is self-reported by the companies, so it is not based on standardized definitions of geographic areas such as states, countries, or regions.

        Table 7--Number of Registrants According to Sample Size Ranges

        ------------------------------------------------------------------------

        Number of

        Sample size(n) ranges registrants

        ------------------------------------------------------------------------

        n=1000................................................. 20

        ---------------

        Total............................................... 1,214

        ------------------------------------------------------------------------

        The example in Table 7 is simplified by the assumptions that registrants have employees in a single industry and that employee pay is described by a lognormal distribution with parameters based on aggregate statistics for that industry. We recognize that statistical sampling may be more complicated for registrants with different types of pay distributions or multiple business and geographic segments, each of which may have different parameters of the distribution. While one commenter suggested simple random sampling could be used for these registrants,\676\ other approaches, such as stratified cluster sampling,\677\ may yield more efficient estimates in some instances. We also recognize that the implementation of statistical sampling for registrants with multiple payroll systems may require additional steps. While we believe that statistical sampling can produce reasonable estimates of the median for these types of registrants, we lack information on intra-firm employee pay distributions and registrants' costs of sampling to estimate the proportion of registrants for which specific sampling approaches may most efficiently produce reasonable estimates of median pay.

        ---------------------------------------------------------------------------

        \676\ See letter from Ohlrogge II.

        \677\ See, e.g., S. Gross. Median estimation in sample surveys. In Proceedings of the Section on Survey Research Methods. American Statistical Association, 181-184. (1980).

        ---------------------------------------------------------------------------

        While some commenters argued that statistical sampling will not lead to significant reductions in compliance costs,\678\ the majority of commenters supported using statistical sampling for calculating the median employee, implying that this approach can reduce costs for some registrants.\679\ In light of the comments received, we continue to believe that permitting registrants to use statistical sampling will lead to an overall reduction in compliance costs as compared to not permitting this method of identifying the median.

        ---------------------------------------------------------------------------

        \678\ See, e.g., letters from Business Roundtable I, Chamber I, COEC I, COEC II, ExxonMobil, Freeport-McMoRan, FuelCell Energy, Garmin, Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and WorldatWork I.

        ---------------------------------------------------------------------------

        g. Disclosure of Methodology, Assumptions, and Estimates

        The final rule requires registrants to briefly describe and consistently apply any methodology used to identify the median employee and disclose any material assumptions, adjustments, or estimates used to identify the median or to determine total compensation or any elements of total compensation. Registrants also must clearly identify any estimates used. Registrants' disclosure of the methodology and material assumptions, adjustments, and estimates used must be designed to provide information for a reader to be able to evaluate the appropriateness of the methodologies used.

        This disclosure is intended to aid shareholders and investors in their use of the pay ratio disclosure and alert them to any material changes in the methodology that might change the reported pay ratio. Many commenters indicated that the rule should require registrants to provide this narrative information,\680\ but a number of these commenters indicated that the rule should clarify that the narrative be brief.\681\

        ---------------------------------------------------------------------------

        \680\ See, e.g., letters from AFL-CIO I, Barnard, Business Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT State Treasurer, Domini, Hermes, Kasner, LAPFF, Meridian, Microsoft, Somers, Sze, UAW Trust, US SIF, and WorldatWork I.

        \681\ See, e.g., letters from ABA, AFL-CIO I, Business Roundtable I, COEC I, COEC II, Domini, Meridian, Microsoft, and WorldatWork I.

        ---------------------------------------------------------------------------

        Alternatively, we could have required registrants to provide a detailed description of every operational step and methodological assumption used in identifying the median employee. Because we are concerned that disclosure about methodology, assumptions, adjustments, and estimates could become dense and overly technical,\682\ which we believe would limit its usefulness, the final rule asks for a brief overview and makes clear that it is not necessary to provide technical analyses or formulas. We do not believe that a detailed, technical discussion (such as statistical formulas, confidence levels, or the steps used in data analysis) would appreciably enhance shareholders' understanding of how the pay ratio was calculated. We recognize, as commenters noted, that registrants will incur some costs in developing and reviewing the appropriate language to describe the approach taken. However, we expect that the costs of this disclosure will be marginal, as these additional disclosures are intended to simply describe what has already been done or assumed in the calculations, and therefore will not require additional analysis by registrants.

        ---------------------------------------------------------------------------

        \682\ See, e.g., letter from Business Roundtable I (commenting that the disclosure could not be brief because of the issuer's need to use many estimates and assumptions).

        ---------------------------------------------------------------------------

        h. Determination of Total Compensation

        As mandated by Section 953(b), the final rule defines ``total compensation'' by reference to Item 402(c)(2)(x). We received comments supporting the use of estimates in calculating the annual total compensation or any elements of total compensation for employees other than the PEO.\683\ As proposed, the final rule permits registrants to use reasonable estimates to determine elements of ``total compensation.''

        ---------------------------------------------------------------------------

        \683\ See, e.g., letters from ABA, COEC I, COEC II, Davis Polk, Prof. Angel, and Vectren Corp.

        ---------------------------------------------------------------------------

        We acknowledge that, to the extent that the use of estimates causes the disclosure to present a less precise measure of the ``total compensation'' of the registrant's median employee than if we prohibited the use of estimates, it could diminish the potential usefulness of the disclosure. However, commenters did not suggest that allowing for the use of reasonable estimates in determining the ``total compensation'' would diminish the potential usefulness of the disclosure and we likewise believe it is not likely to have such an effect.\684\

        ---------------------------------------------------------------------------

        \684\ But see letter from Dennis T (``No estimates. No sampling. We demand the actual data.'').

        ---------------------------------------------------------------------------

        i. Defining ``Annual''

        As proposed, the final rule defines ``annual total compensation'' to mean total compensation for the last completed fiscal year, consistent with the time period used for the other Item 402 disclosure requirements.

        Some commenters agreed that the pay ratio disclosure should be calculated based on data from the last completed fiscal year.\685\ Other commenters, however, recommended that the rule permit registrants to use another period, such as the fiscal year preceding the registrant's last completed fiscal year.\686\ Commenters also asked that registrants be permitted to choose the period.\687\ We understand that these suggestions are intended to reduce compliance costs for registrants by giving registrants extra time to comply with the rule or the ability to use information in the form

        Page 50177

        that would best suit their particular facts and circumstances. We believe, however, that it is appropriate for the time period used for the pay ratio disclosure to be the same as the time period used for the registrant's other executive compensation disclosures, although the flexibility in identifying the median employee could help to address the concerns raised by these commenters. In particular, using the same time period as for other executive compensation disclosure will avoid any possible confusion for shareholders using this disclosure.

        ---------------------------------------------------------------------------

        \685\ See, e.g., letters from ABA, CalPERs, and UAW Trust.

        \686\ See, e.g., letters from Aon Hewitt, Business Roundtable I, Corporate Secretaries, and Eaton.

        \687\ See, e.g., letters from Davis Polk and WorldatWork I.

        ---------------------------------------------------------------------------

        j. Updating the Pay Ratio Disclosure for the Last Completed Fiscal Year

        The final rule includes instructions to clarify the timing for updating pay ratio disclosure after the end of a registrant's fiscal year. Without this provision, a registrant could be required to include pay ratio disclosure in a filing after the end of the fiscal year, but before it has compiled the executive compensation information for that fiscal year for inclusion in its proxy statement relating to its annual meeting of shareholders. This could impose additional costs on registrants that elect to provide executive compensation disclosure in their annual proxy statement rather than in their annual report and for registrants that are conducting registered offerings at the beginning of their fiscal year.

        To address this concern, we considered the recommendation of commenters that pay ratio disclosure not be required to be updated for the most recently completed fiscal year until the registrant files its proxy statement for its annual meeting of shareholders. The final rule generally follows this recommended approach and also provides a similar accommodation for registrants that do not file annual proxy statements.\688\ It also aligns the final rule to the filing deadlines for providing Item 402 disclosure in annual reports and proxy and information statements. We believe that such an approach will reduce costs to registrants without diminishing the potential usefulness of the disclosure.

        ---------------------------------------------------------------------------

        \688\ Based on a review of EDGAR filings in calendar year 2013, approximately 250 registrants that would be subject to the final rule do not file proxy or information statements in connection with annual meetings of shareholders, including 15D filers (other than SRCs and ABS issuers) and registrants that are not corporate entities required to hold annual meetings of shareholders.

        ---------------------------------------------------------------------------

        We also believe that this approach could reduce costs for registrants in connection with filings made or required to be made before the filing of the proxy or information statement for the annual meeting of shareholders (or written consents in lieu of such a meeting) that would typically contain the registrant's other Item 402 disclosure covering the most recently completed fiscal year. In addition, under the final rule, updating the pay ratio disclosure is not an additional impediment for a registrant that requests effectiveness of a registration statement after the end of its fiscal year and before the filing of the proxy statement for its annual meeting of shareholders. In this regard, this approach could alleviate some of the final rule's potential impact on capital formation.

        k. Status of Disclosure as ``Filed''

        Under the final rule, the pay ratio disclosure will be considered ``filed'' for purposes of the Securities Act and Exchange Act, like other Item 402 information. A number of commenters recommended that the pay ratio disclosure be ``furnished'' rather than ``filed'' \689\ because registrants will have to review a large amount of data and make a significant number of estimates, assumptions, and judgment calls, which will necessarily lead to imprecision.\690\ This, in turn, could subject registrants to potential liability and litigation,\691\ make it difficult to validate the information sufficiently for Sarbanes-Oxley Act certification purposes,\692\ and/or not permit the information to be audited (or greatly increase the costs of the audits).\693\ We recognize that some registrants could have more difficulty in gathering and verifying the information than others. We believe, however, that the flexibility afforded to registrants in connection with identifying the median employee could reduce some of the difficulties of compiling the required information because registrants will be able to tailor the methodology to reflect their own facts and circumstances. In addition, we believe that the final rule's transition periods, which are discussed below, could mitigate some concerns about compiling and verifying the information because they are designed to give registrants sufficient time to develop and implement compliance procedures.

        ---------------------------------------------------------------------------

        \689\ See, e.g., letters from AAFA II, ABA, American Benefits Council, Aon Hewitt, Best Buy et al., Bill Barrett Corp., Business Roundtable I, Business Roundtable II, Chamber I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, Eaton, Freeport-

        McMoRan, General Mills, Intel, Mercer I, NAM I, NAM II, NIRI, NRF, PM&P, RILA, SHRM, and Vectren Corp.

        \690\ See, e.g., letters from AAFA II, ABA, Business Roundtable I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, Eaton, General Mills, NRF, RILA, and Vectren Corp.

        \691\ See, e.g., letters from ABA, American Benefits Council, Aon Hewitt, Bill Barrett Corp., Chamber I, General Mills, Mercer I, and PM&P.

        \692\ See, e.g., letters from ABA, Best Buy et al., Corporate Secretaries, Freeport-McMoRan, Intel, NAM I, and SHRM.

        \693\ See, e.g., letters from COEC I, COEC II, Corporate Secretaries, and NIRI.

        ---------------------------------------------------------------------------

        Requiring registrants to ``file'' their pay ratio information may make the final rule more costly for registrants than the alternative of allowing them to ``furnish'' such information. Treating the pay ratio disclosure as ``filed'' will mean that registrants could potentially be subject to litigation under Section 18 of the Exchange Act, although, as mentioned earlier, Section 18 does not create strict liability for misstatements in ``filed'' information and requires that a plaintiff establish that it relied on the misleading information in purchasing or selling a security and suffered damages caused by that reliance. On the other hand, under the final rule, this potential liability for misleading pay ratio disclosure could make registrants more accountable for the disclosure than if we instead permitted the disclosure to be `furnished' and result overall in fewer inaccuracies in the required pay ratio disclosure. To the extent that registrants perceive there to be a greater likelihood of private litigation under the final rule than if they were permitted to ``furnish'' the information, registrants may decide to apply a more costly process to identify the median employee, or retain additional counsel, thus increasing compliance costs.

        l. Compliance Date

        Section 953(b) does not specify a date when registrants must begin to comply with the final rule. In a change from the Proposing Release, the final rule requires that a registrant must begin to comply with Item 402(u) with respect to compensation for the registrant's first full fiscal year commencing on or after January 1, 2017.

        As discussed above, the change from the proposal provides calendar year-end filers and registrants with fiscal years beginning January 1, 2015 until the day before the final rule's effectiveness one additional year to provide their pay ratio disclosure relative to the proposal. The final rule also changes the compliance schedule for registrants with fiscal years starting on or after effectiveness through December 30, 2015. These registrants will receive two additional years to provide their pay ratio disclosure relative to the proposal. Assuming a hypothetical effective date of November 1, 2015, we estimate that this change will lead to a one-time cost deferral of approximately $147 million for two years and to savings of approximately between $27.3 million and $212 million for 223 registrants subject to the final rule that have fiscal

        Page 50178

        years that end from October 31, 2015, through December 30, 2015 \694\ and a one-time cost deferral of approximately $1,169 million for one year and to savings of approximately between $109 million and $842 million for the remaining 3,348 registrants subject to the final rule.\695\ For this estimation, we identified the fiscal year ends for all affected registrants. We found that 223 registrants have fiscal years that end between October 31 and December 30, of which 130 are multinational registrants and 93 are registrants with U.S.-only segments. We found that of the remaining 3,348 registrants estimated to be subject to the final rule, 1,340 are multinational registrants and 2,008 are registrants with U.S.-only segments.\696\

        ---------------------------------------------------------------------------

        \694\ For these registrants, the initial compliance cost is assumed to be deferred for two years. There are 130 multinational registrants and 93 registrants with U.S.-only segments. The annual cost for the 130 multinational registrants is equal to the product of their cumulative number of employees (3,354,869) and the cost per employee ($38.04), or a total of approximately $128 million. The annual cost for the 93 registrants with U.S.-only segments is equal to the product of their cumulative number of employees (1,018,780) and the cost per employee ($19.02), or a total of approximately $19 million. Thus, the total initial compliance cost for the 223 registrants that is deferred for two years is approximately $147 million. These registrants also will not have to incur two years of ongoing compliance costs. Assuming the ongoing cost as a percentage of the initial cost is between one-third of 28% (9.3%) and 72% of the initial cost, the cost savings are estimated to be between ($147 million*9.3%)*2 = $27.3 million and ($147 million*72%)*2 = $212 million. See Sections III.C.2.c and III.D.2.e.

        \695\ For these registrants, the initial compliance cost is assumed to be deferred for one year. Of the 3,348 registrants, 1,340 are multinational registrants with 24,240,445 employees in total and a total annual cost of $922 million; 1,691 are registrants with U.S.-only segments with 12,530,180 employees in total and a total annual cost of $238 million; and 317 are registrants with U.S.-only segments with missing employee data and a total annual cost of $9 million. Thus, the total cost for the 3,348 registrants is approximately $1,169 million over one year. These registrants also will not have to incur one year of ongoing compliance costs. Assuming the ongoing cost as a percentage of the initial cost is between one-third of 28% (9.3%) and 72% of the initial cost, the cost savings are estimated to be between ($1,169 million*9.3%) = $109 million and ($1,169 million*72%) = $842 million. See Sections III.C.2.c and III.D.2.e.

        \696\ For registrants without segment or employee data, we followed the same approach as in Table 3.

        ---------------------------------------------------------------------------

        m. Transition Periods

        The final rule also includes a transition period for new registrants because we are sensitive to the impact that the rule could have on capital formation. We note that the requirements of Section 953(b), as amended by the JOBS Act, distinguish between certain newly public companies and all other registrants by providing an exemption for emerging growth companies. We also note that the incremental time needed to compile pay ratio disclosure could cause registrants that are not emerging growth companies to delay an initial public offering, which could have a negative impact on capital formation.\697\ In this regard, we expect that registrants that are not emerging growth companies are likely to be businesses with more extensive operations and a greater number of employees than many emerging growth companies, which could increase the initial efforts needed to comply with the final rule. We believe that providing a transition period for these newly public companies could mitigate this potential impact on capital formation.

        ---------------------------------------------------------------------------

        \697\ See, e.g., letters from ABA (``Although we doubt that a company considering an initial public offering of its securities would decide to forego such a transaction simply because of the pay ratio disclosure obligation, in some situations, the time and costs associated with Item 402(u) compliance could certainly weigh in the timing of the offering.''), Lou (``The competitive disadvantages raise the costs of raising capital through public trading markets and thereafter discourage companies to go public. The additional monetary cost obviously makes an initial public offering (`IPO') a less attractive means to raise capital though the negative impact might not constitute a fatal factor that would kill the IPO. However a CEO would feel reluctant to list the company because of the threatening embarrassment of pay ratio disclosure. Therefore the pay ratio exerts a negative effect on IPOs.''), and PM&P (``We agree with the proposed transition period that new registrants should not be required to include pay ratio disclosure in their initial registration statements, and that to do so could significantly delay the IPO.'').

        ---------------------------------------------------------------------------

        To address these concerns, the final rule also includes instructions that would permit new registrants to delay compliance, so that pay ratio disclosure would not be required in a registration statement on Form S-1 or S-11 for an initial public offering or a registration statement on Form 10. Instead, such a registrant would be required to first comply with Item 402(u) with respect to compensation for the first fiscal year commencing on or after the year in which the registrant first becomes subject to the requirements of Section 13(a) or Section 15(d) of the Exchange Act, but no earlier than the year commencing January 1, 2017. Additionally, in response to comments, the final rule provides that a registrant that ceases to be a smaller reporting company or an emerging growth company will not be required to provide pay ratio disclosure until the first fiscal year after exiting such status, but no earlier than the year commencing January 1, 2017. This change from the proposed rule allows registrants exiting smaller reporting company or an emerging growth company status to delay their initial compliance by a year, and will give them additional time to decide how they will identify their median employee and prepare the necessary disclosure. Further, the final rule permits registrants that engage in business combinations and/or acquisitions to not include in the median employee determination employees of a newly-acquired entity for the fiscal year in which the business combination or acquisition occurs. We believe that the exercise of discretion used in allowing these additional transitional periods will result in cost savings for the affected registrants and will further mitigate any effects of the rule on capital formation.

  10. Paperwork Reduction Act

    1. Background

      Certain provisions of the final amendments contain ``collection of information'' requirements within the meaning of the Paperwork Reduction Act of 1995 (the ``PRA'').\698\ We published a notice requesting comment on the collection of information requirements in the Proposing Release for the rule amendments, and we submitted these collections of information requirements to the Office of Management and Budget (``OMB'') for review in accordance with the PRA.\699\ The titles for the collections of information are:

      ---------------------------------------------------------------------------

      \698\ 44 U.S.C. 3501 et seq.

      \699\ 44 U.S.C. 3507(d) and 5 CFR 1320.11.

      ---------------------------------------------------------------------------

      ``Regulation S-K'' (OMB Control No. 3235-0071);

      ``Form 10-K'' (OMB Control No. 3235-0063);

      ``Regulation 14A and Schedule 14A'' (OMB Control No. 3235-

      0059);

      ``Regulation 14C and Schedule 14C'' (OMB Control No. 3235-

      0057);

      ``Form 8-K'' (OMB Control No. 3235-0060);

      ``Form S-1'' (OMB Control No. 3235-0065);

      ``Form S-4'' (OMB Control No. 3235-0324);

      ``Form S-11'' (OMB Control No. 3235-0067);

      ``Form 10'' (OMB Control No. 3235-0064); and

      ``Form N-2'' (OMB Control No. 3235-0026).

      These regulations, schedules and forms were adopted under the Securities Act and the Exchange Act, and in the case of Form N-2,\700\ the Investment Company Act of 1940.\701\ The regulations, forms and schedules set forth the disclosure requirements for periodic reports, registration statements,

      Page 50179

      and proxy and information statements filed by companies to help investors make informed investment and voting decisions. The hours and costs associated with preparing, filing and sending the form or schedule constitute reporting and cost burdens imposed by each collection of information. An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB control number.

      ---------------------------------------------------------------------------

      \700\ 17 CFR 239.14 and 274.11a-1.

      \701\ 15 U.S.C. 80a-1 et seq.

      ---------------------------------------------------------------------------

      Our amendments to the forms and regulations are intended to satisfy the requirements of Section 953(b) of the Dodd-Frank Act, which directs the Commission to amend Item 402 of Regulation S-K to add the pay ratio disclosure requirements specified by that provision. Compliance with the final rule will be mandatory for affected registrants. Responses to the information collections will not be kept confidential, and there will be no mandatory retention period for the information disclosed.

    2. Summary of Information Collections

      In order to satisfy the legislative mandate in Section 953(b), we are adopting new paragraph (u) to Item 402 of Regulation S-K. This new paragraph (u) will require registrants to disclose:

      the median of the annual total compensation of all employees of the registrant (excluding the PEO);

      the annual total compensation of the registrant's PEO; and

      the ratio between these two amounts.

      For this purpose, Section 953(b) specifies that total compensation is to be determined in accordance with Item 402(c)(2)(x). Item 402 already requires registrants to disclose the annual total compensation of the PEO in accordance with Item 402(c)(2)(x).\702\ The median of the annual total compensation of all employees and the ratio are new, incremental disclosure burdens and will require affected registrants to collect compensation information for employees that is not currently required to be disclosed.

      ---------------------------------------------------------------------------

      \702\ As of the date of this release, the requirements for the calculation of total compensation under Item 402(c)(2)(x) are the same as those in effect on July 20, 2010. Therefore, for purposes of this PRA analysis, we have assumed that registrants would not need to recalculate the annual total compensation for the PEO in connection with the pay ratio disclosure.

      ---------------------------------------------------------------------------

      The additional disclosure under new paragraph (u) of Item 402 will be required in any annual report, proxy or information statement, or registration statement that requires executive compensation disclosure pursuant to Item 402 of Regulation S-K.\703\ In addition, the requirements will allow certain registrants to omit the disclosure otherwise required by Item 402(u) from filings made during a specified transition period.

      ---------------------------------------------------------------------------

      \703\ Consistent with the scope of Section 953(b), the new requirements will not apply to the annual reports and proxy and information statements of emerging growth companies, smaller reporting companies or foreign private issuers. In addition, consistent with the instructions J and I of Form 10-K, the new requirements will not apply to the annual reports of issuers of asset-backed securities or to wholly-owned subsidiary registrants.

      ---------------------------------------------------------------------------

      Finally, in order to conform the amendments to current rules for the disclosure of PEO compensation when certain elements of compensation are not yet known, we are adopting a conforming amendment to Item 5.02 of Form 8-K. New paragraph (1) of Item 5.02(f) will also require registrants that are disclosing PEO total compensation in accordance with Item 5.02 of Form 8-K to provide in that filing the updated pay ratio disclosure required by Item 402(u). Because Item 5.02 of Form 8-K provides a delayed method of filing information that would otherwise be required in the registrant's proxy or information statement or annual report, the PRA analysis assumes that the burden and cost of compliance with new Item 402(u) would be associated primarily with those forms and schedules rather than Form 8-K.

    3. Summary of Comment Letters and Revisions to Proposals

      In the Proposing Release, we requested comment on the PRA analysis. We received letters from two commenters that directly addressed the PRA estimates,\704\ as well as a number of other comment letters and submissions that discussed the costs and burdens to issuers that would have an effect on the PRA analysis.\705\ A detailed discussion of these comments is included in the Section III above.\706\

      ---------------------------------------------------------------------------

      \704\ See letters from COEC I, COEC II, and Chamber II.

      \705\ See letters from Avery Dennison, ExxonMobil, FEI and KBR.

      \706\ See Section III.C.2.b.

      ---------------------------------------------------------------------------

      One of the two commenters analyzed data from a survey of 118 companies to conclude that it would take registrants an average of 952 hours per year to comply with the pay ratio disclosure requirement at an average labor cost of $185,600, which we assume refers to external costs only.\707\ The other commenter disagreed with the our assumption in the Proposing Release that ongoing compliance costs in the second and third year would significantly decrease from the initial compliance costs in the first year.\708\ Based on the proposed rule, which did not include several accommodations adopted in the final rule, that commenter estimated that ongoing compliance costs would be 80% (mid-

      range) or 72% (average) of the initial compliance costs for each successive year. This commenter also cited a survey completed by 128 public companies, the majority of which were large companies with assets well over $2 billion, to assert that the average cost of outside securities compliance counsel is $700 per hour, rather than the $400 per hour used in the PRA estimates.\709\

      ---------------------------------------------------------------------------

      \707\ See letter from Chamber II. This commenter estimated that the annual cost of compliance would be $710.9 million and an annual compliance time of 3.6 million hours. The commenter also stated that we may have underestimated costs by more than 870% and underestimated compliance time by 560%. From this information, we infer that the average labor cost of $185,600 refers to external costs, as multiplying the number of registrants estimated to be subject to the proposed rule (3,830) by the average labor cost estimated by this commenter ($185,600) equals $710,848,000.

      \708\ See letters from COEC I and COEC II.

      \709\ As noted in our Economic Analysis, we continue to believe that $400 per hour is the appropriate rate to use for PRA purposes. This is the rate we typically estimate for outside legal services used in connection with public company reporting and is intended to represent an average to cover all registrants of varying sizes. In addition, some commenters indicated that they would retain external advisors such as payroll specialists, human resource consultants, and compensation consultants. See letters from Avery Dennison (stating that it expects to retain two to three external advisors, including legal advisers and HR consultants) and General Mills (indicating that it expects to hire outside advisors, such as compensation consultants and payroll specialists). Generally, we expect the hourly fees for such external advisors to be much lower than those of legal counsel.

      ---------------------------------------------------------------------------

      Several companies submitted estimates of burdens and costs without commenting on the our estimates. One company estimated that compliance with the proposed rule would require 100-150 hours of work by internal staff and 20 to 40 hours of external consulting time at a total ``internal cost'' of $1 million to $1.5 million.\710\ Another company estimated that it would take over 1,000 internal burden hours to develop the database and methodology to derive the pay ratio information, and that ongoing burden hours would be approximately 50% (500 hours) of the initial compliance burden hours.\711\ It also asserted that this represents an approximate cost of over $250,000 on an initial basis and $100,000 on an ongoing

      Page 50180

      basis.\712\ Another large corporate commenter asserted that it would require up to approximately 3,000 internal burden hours to comply with the proposed rule in the initial year of compliance, and that ongoing compliance burdens after the initial compliance year would be approximately 28% of the initial burden hours (850 hours per year thereafter).\713\ Another global issuer estimated that it may take between $500,000 and $1 million to establish and automate the process to comply with the proposed rule.\714\

      ---------------------------------------------------------------------------

      \710\ See letter from Avery Dennison. Although this commenter used the phrase ``internal cost'' of compliance, we assume that this cost includes more than internal staff time. Otherwise, the internal cost of compliance could range from approximately $6,667 to $15,000 per hour ($1 million divided by 150 hours or $1.5 million divided by 100 hours).

      \711\ See letter from FEI.

      \712\ Id.

      \713\ See letter from ExxonMobil.

      \714\ See letter from KBR.

      ---------------------------------------------------------------------------

      We are adopting the final rule as proposed with modifications that may help mitigate compliance costs and burdens. First, we provide two tailored exemptions for non-U.S. employees from the definition of ``employee'': an exemption for circumstances in which a foreign jurisdiction's laws or regulations governing privacy are such that, despite its reasonable efforts to obtain or process the information necessary for compliance with the final rule, the registrant is unable to do so without violating such data privacy laws or regulations and a de minimis exemption. Second, the final rule defines ``employee'' to include only the employees of the registrant's consolidated subsidiaries, instead of all subsidiaries as proposed. Third, to provide consistency and flexibility, the final rule permits registrants to use any date within three months of the last day of their last completed fiscal year to identify the median employee. Fourth, the final rule allows registrants to identify the median employee every three years, instead of every year, if there has been no change in their employee population or employee compensation arrangements that they reasonably believe would result in a significant change in their pay ratio disclosure.

    4. Revisions to PRA Reporting and Cost Burden Estimates

      For purposes of the PRA, in the Proposing Release, we estimated that the total annual increase in the paperwork burden for all affected companies to prepare the disclosure that would be required under the adopted amendments would be approximately 545,792 hours of company personnel time and a cost of approximately $72,772,200 for the services of outside professionals. As discussed in more detail below, we are revising our PRA burden and cost estimates to reflect the responses of commenters, as well as the modifications we have made to the final rule to reduce compliance burdens.

      For purposes of the PRA for the final rule, we estimate the total annual increase in the paperwork burden for all affected companies to comply with the collection of information requirements in our final rule is approximately 2,367,573 hours of company personnel time and approximately $315,390,720 for the services of outside professionals.\715\ These estimates include the time and the cost of implementing data gathering systems and disclosure controls and procedures, compiling necessary data, preparing and reviewing disclosure, filing documents and retaining records.

      ---------------------------------------------------------------------------

      \715\ We describe how we derived the three-year average hour and cost burdens per response below. The portion of the burden carried by outside professionals is reflected as a cost, while the portion of the burden carried by the company internally is reflected in hours. For administrative convenience, the presentation of the totals related to the paperwork burden hours have been rounded to the nearest whole number and the cost totals have been rounded to the nearest dollar.

      ---------------------------------------------------------------------------

      In deriving these estimates, we have assumed that:

      Registrants subject to the final rule would satisfy the new requirements by either including the information directly in annual reports on Form 10-K or incorporating the information by reference from a proxy statement on Schedule 14A or information statement on Schedule 14C. Our estimates assume that substantially all of the burden relating to the new disclosure requirements would be associated with Form 10-K;

      For registrants that would be permitted to provide their pay ratio disclosure in a filing made in accordance with Item 5.02 of Form 8-K, rather than in Form 10-K, the burden relating to the new disclosure requirements would be associated primarily with Form 10-K rather than Form 8-K; \716\ and

      ---------------------------------------------------------------------------

      \716\ Our PRA estimates for Form 8-K include an estimated one hour burden to account for the inclusion of the new pay ratio disclosure.

      ---------------------------------------------------------------------------

      100% of new registrants would use the transition provisions allowing them to omit the required disclosure from their initial registration statements and, for follow-on offerings by these registrants, the burden relating to the new disclosure requirements would be associated primarily with Form 10-K rather than Forms S-1, S-

      11 or N-2 as applicable (because registrants would incorporate the disclosure from Form 10-K).

      We understand from commenters that the burdens and costs of compliance will likely vary among individual companies based on a number of factors, including the size and complexity of their organizations, the nature of their operations and workforce, the location of their operations, and, significantly, the extent that their existing payroll systems collect the information necessary to identify the median of the annual total compensation of their employees. Because the final rule provides additional flexibility in identifying the median and the annual total compensation of employees, the actual burden could be lower if the methodology used is able to reduce the effort needed to collect the data or if the registrant is able to use information that it collects for other purposes.\717\ We believe that the actual burdens will likely vary significantly among individual companies based on these factors. Our estimates in this PRA analysis reflect average burdens, and, therefore, some companies may experience costs in excess of our estimates and some companies may experience costs that are lower than our estimates.\718\

      ---------------------------------------------------------------------------

      \717\ See Section II of this release for a discussion of the requirements.

      \718\ As in our Economic Analysis, we estimated the PRA costs and burdens to reflect a broad range of registrants.

      ---------------------------------------------------------------------------

      1. Estimated Internal Burden Hours

        Commenters estimated that registrants would spend anywhere from 100 burden hours \719\ to 3000 burden hours to prepare and review the pay ratio disclosure.\720\ One commenter estimated that affected companies would on average spend 952 hours per year to comply with the new disclosure requirement.\721\ This estimate was based on a survey of a range of companies, some with operations in more than 50 countries and others with operations in fewer than 10 countries, and did not take into account the modifications that were made to the rule to reduce compliance costs.

        ---------------------------------------------------------------------------

        \719\ See letter from Avery Dennison.

        \720\ See letter from ExxonMobil.

        \721\ See letter from Chamber II.

        ---------------------------------------------------------------------------

        In our analysis of the economic costs and benefits of the rule, we estimated that the total initial compliance costs would be $1,314,694,544 or approximately $368,159 per registrant.\722\ Our estimate did not break down the costs between internal burden hours and external costs, which is how the burdens and costs are described for PRA purposes. As discussed later in our analysis of the estimated cost and hour burdens for each collection of

        Page 50181

        information,\723\ we believe that substantially all of the burden relating to the new disclosure requirements will be associated with Form 10-K. For Exchange Act reports on Form 10-K, we estimate that 75% of the burden of preparation is carried by the company internally and that 25% of the burden of preparation is carried by outside professionals. Using that formula, we estimate that the average registrant will spend 1,105 internal burden hours preparing and reviewing the disclosure for the initial year of compliance.\724\

        ---------------------------------------------------------------------------

        \722\ See discussion in Section III.C.2.c. $1,314,694,544/3,571 = $368,159.

        \723\ See discussion in Section IV.D.3.

        \724\ We did not receive any estimates of the cost per hour related to preparation of disclosures by the company internally, but expect that such costs will be less than the cost of hiring outside professionals. For ease of analysis, we assume that internal hourly costs will be approximately half the cost of hiring outside professionals ($400/2 = $200). Assuming 75% of burden hours are carried internally and 25% are carried externally, the average compliance cost of $368,159 per registrant corresponds to $368,159/

        ((.75)$200 + (.25)$400) = 1,473 hours, of which 1,105 hours (1,473(.75)) are internal and 368 hours (1,473(.25)) are external.

        ---------------------------------------------------------------------------

        In the Proposing Release, we estimated that the internal burden hours would be greatest during the first year of compliance with the rule and would diminish in subsequent years. As discussed above, we received few estimates of ongoing compliance costs from commenters, and the estimates we received varied widely. Some commenters suggested that the internal burden hours and external professional costs would not decrease after the initial compliance year.\725\ As discussed earlier, we believe that part of the initial compliance costs would decline after the first year.\726\ Other commenters provided estimates of ongoing compliance costs that ranged from 28% to 80% of the initial compliance costs. For example, one commenter estimated that its burden would decrease by approximately 72%, from 3,000 internal hours during the first year to 850 internal hours in subsequent years, approximately 28% of the initial burden.\727\ Another commenter suggested, based on results from a survey that it conducted, that ongoing costs could be about 80% (median) or 72% (average) of the initial compliance costs.\728\ Yet another commenter estimated its initial burden at 1,000 internal hours and ongoing burden at 50% of the initial compliance costs (500 internal hours).\729\ This commenter also estimated initial compliance costs at $250,000 with ongoing compliance costs of $100,000, or 40% of the initial compliance costs.

        ---------------------------------------------------------------------------

        \725\ See letters from Chamber II and Intel.

        \726\ See discussion in Section III.C.2.c.

        \727\ See letter from ExxonMobil.

        \728\ See letter from COEC I.

        \729\ See letter from FEI.

        ---------------------------------------------------------------------------

        Because of the limited number of ongoing cost estimates and their wide dispersion, for the purposes of the PRA we assume that ongoing compliance burdens and costs will be approximately 40% (the median of the estimates) of the initial compliance burdens and costs. Thus, we used one commenter's estimated ongoing burden of 40% of the initial burden for our estimated three-year average burden. We utilize an estimated burden of 1,105 hours in the initial year and 442 \730\ hours in the two years thereafter, for a three-year average burden of 663 hours.\731\

        ---------------------------------------------------------------------------

        \730\ 1,105 x 40% = 442 burden hours.

        \731\ (1,105 + 442 + 442)/3 = 663 burden hours.

        ---------------------------------------------------------------------------

      2. Estimated Cost Burdens

        Commenters provided a wide range of estimated external cost burdens. Commenters provided estimates of $185,600 per year on average for each company,\732\ $250,000 on an initial basis and $100,000 on an ongoing basis,\733\ and 20-40 hours of external consulting time.\734\ Another commenter estimated that it may cost between $500,000 and $1 million for each company to establish and automate the process to comply with the proposed rule, although it is not clear whether this includes both internal and external costs and burdens.\735\ As discussed above, we estimate that total compliance burdens for the initial year of compliance will be $1,314,694,544 or $368,159 per registrant. Assuming that 25% of the burden of preparing the disclosure is carried by outside professionals, we estimate that the average registrant will incur $147,200 in outside professional costs in the first year to comply with the disclosure requirement.\736\

        ---------------------------------------------------------------------------

        \732\ See letter from Chamber II.

        \733\ See letter from FEI.

        \734\ See letter from Avery Dennison.

        \735\ See letter from KBR.

        \736\ 368 x $400 = $147,200.

        ---------------------------------------------------------------------------

        As with the estimated internal burden hours, we assume that the compliance costs after the initial year will be reduced because a substantial portion of the costs will be related to establishing systems and processes to collect the payroll data in the initial year of compliance. Applying the same assumption used above that the ongoing compliance costs will be approximately 40% of the estimate for the initial compliance year, we estimate that ongoing compliance costs will be approximately $58,880 per year on average for each affected company \737\ so that the three-year average cost of compliance is $88,320.\738\

        ---------------------------------------------------------------------------

        \737\ $147,200x 40% = $58,880.

        \738\ ($147,200 + 58,880 + 58,880)/3 = $88,320.

        ---------------------------------------------------------------------------

      3. Estimated Cost and Hour Burdens for Each Collection of Information

        For each collection of information, we estimate the following cost and hour burdens:

        a. Regulation S-K

        While the adopted amendments would make revisions to Regulation S-

        K, the collection of information requirements for that regulation are reflected in the burden hours estimated for the forms and schedules listed below. The rules in Regulation S-K do not impose any separate burden. Consistent with historical practice, we are retaining an estimate of one burden hour to Regulation S-K for administrative convenience.

        b. Form 10-K

        Only Forms 10-K that are filed by registrants that are not smaller reporting companies or emerging growth companies will be required to include the pay ratio disclosure. For purposes of our PRA estimates, we have assumed that 100% of asset-backed securities issuers will omit Item 402 disclosure from Form 10-K pursuant to Instruction J of Form 10-K and 100% of wholly-owned subsidiary registrants will omit Item 402 disclosure from Form 10-K pursuant to Instruction I of Form 10-K, and, accordingly, these registrants will also not be subject to the new disclosure requirements. Based on a review of EDGAR filings in calendar year 2014, we estimate that of the approximately 7,619 annual reports filed in that year, approximately 3,571 annual reports are filed by registrants that would be subject to the new disclosure requirements.\739\ We estimate that the new disclosure requirements will add an average of 663 burden hours \740\ to the total burden hours required to produce each Form 10-K that is subject to the new requirements and approximately $88,320 for outside professionals.\741\

        ---------------------------------------------------------------------------

        \739\ Based on a review of EDGAR filings in 2014, approximately 678 annual reports were filed by EGCs, 2,958 by SRCs, and 412 by ABS issuers. See Section III.D.2.b above.

        \740\ As we discuss below, we estimate that 10 of the Forms 8-K filed in a given year would require one additional hour for preparing the disclosure required by the amendments. Thus, substantially all of the internal burden of the pay ratio disclosure is allocated to Form 10-K: 663(3,571)-(110) = 2,367,563 or approximately 663 (2,367,563/3,571) per response. Burden hours are rounded to the next hour.

        \741\ As discussed below, we estimate that the requirement to provide updated pay ratio disclosure on Form 8-K will result in one additional burden hour for that form. We attribute the external costs of the required pay ratio disclosure proportionately between Form 10-K and Form 8-K based on the estimated internal burden hours for each form. 1/(663+1) x $88,320 = $133 per Form 8-K response. The remaining costs have been attributed to Form 10-K: $88,320)3,571--$133(10) = $315,389,390 in aggregate or $88,320 ($315,389,390/3,571) per response. Costs are rounded up to the next dollar.

        ---------------------------------------------------------------------------

        Page 50182

        We estimate that the preparation of annual reports currently results in a total annual compliance burden of 12,198,095 hours and an annual cost of outside professionals of $1,627,400,000. Under the final rule, we estimate that the total incremental cost of outside professionals for annual reports will be approximately $315,389,390 per year and the total incremental internal burden will be approximately 2,367,563 hours per year.

        c. Form 8-K

        As described in this release, the final rule will require a registrant that is filing its PEO total compensation on a delayed basis due to the unavailability of certain components of compensation on Form 8-K (in accordance with Instruction 1 to Items 402(c)(2)(iii) and (iv) of Regulation S-K and Item 5.02(f) of Form 8-K) to provide the pay ratio disclosure at the same time. The final rule also includes a conforming amendment to Item 5.02 of Form 8-K that will require a registrant to include updated pay ratio disclosure in the Form 8-K that it files to disclose its PEO total compensation information.\742\ We estimate that the burden for adding the pay ratio disclosure to that Form 8-K filing will be one hour per registrant.\743\ We also estimate that the Form 8-K amendment will not result in additional Form 8-K filings because registrants who omit disclosure in reliance on Instruction 1 to Items 402(c)(2)(iii) and (iv) are already required to file a Form 8-K. The amendments will, however, add pay ratio disclosure requirements to that Form 8-K filing.

        ---------------------------------------------------------------------------

        \742\ See Section II.B.6.b.

        \743\ As noted above, we have assumed that the burden relating to the new pay ratio requirements would remain associated with the registrant's proxy or information statement or annual report, and, therefore, our PRA estimates for those forms reflect that burden.

        ---------------------------------------------------------------------------

        Based on a review of EDGAR filings for calendar years 2012 and 2013, we estimate that on average approximately 11 Forms 8-K are filed pursuant to Item 5.02(f) annually and approximately 10 of these relate to disclosure of PEO compensation. As a result, we estimate that 10 of the Forms 8-K filed in a given year will require one additional hour for preparing the disclosure required by the amendments, in addition to the total burden hours required to produce each Form 8-K.

        We estimate that the preparation of current reports on Form 8-K currently results in a total annual compliance burden of 507,675 hours and an annual cost of outside professionals of $67,688,700. As result of the rule, we estimate that the incremental company burden will be approximately 10 hours per year and approximately $1,330 in the incremental cost of outside professionals for current reports on Form 8-K.

        d. Proxy Statements on Schedule 14A

        Only proxy statements on Schedule 14A that are required to include Item 402 information, and that are not filed by smaller reporting companies or emerging growth companies, will be required to include the new pay ratio disclosure. For purposes of our PRA estimates, consistent with past amendments to Item 402,\744\ we have assumed that all of the burden relating to the new disclosure requirements will be associated with Form 10-K, even if registrants include the new disclosure required in Form 10-K by incorporating that disclosure by reference from a proxy statement on Schedule 14A.

        ---------------------------------------------------------------------------

        \744\ We took a similar approach in connection with the rules for Summary Compensation Table disclosure required by the 2006 amendments to Item 402. See 2006 Adopting Release, supra note 9.

        ---------------------------------------------------------------------------

        e. Information Statements on Schedule 14C

        Only information statements on Schedule 14C that are required to include Item 402 information, and that are not filed by smaller reporting companies or emerging growth companies, are required to include the pay ratio disclosure. For purposes of our PRA estimates, consistent with past amendments to Item 402, we have assumed that all of the burden relating to the disclosure requirements will be associated with Form 10-K, even if registrants include the disclosure required in Form 10-K by incorporating that disclosure by reference from an information statement on Schedule 14C.

        f. Form S-1

        Because we have assumed that all new registrants will take advantage of the transition period afforded to them under the final rule, so that all of the registration statements on Form S-1 that will be required to include the pay ratio disclosure will incorporate by reference the registrant's disclosure contained in its annual report, we have assumed that all of the burden relating to the new disclosure requirements will be associated with Form 10-K.

        g. Form S-4

        We have assumed that registrants filing on Form S-4 for whom executive compensation information under Item 402 is required pursuant to Items 18 or 19 of Form S-4 will incorporate by reference the pay ratio disclosure contained in the registrant's annual report. Thus, we have assumed that all of the burden relating to the new disclosure requirements will be associated with Form 10-K.

        h. Form S-11

        Because we have assumed that all new registrants will take advantage of the transition period afforded to them under the final rule, so that all of the registration statements on Form S-11 that will be required to include the pay ratio disclosure will incorporate by reference the registrant's pay ratio disclosure contained in its annual report, we have assumed that all of the burden relating to the new disclosure requirements will be associated with Form 10-K.

        i. Form N-2

        Only Forms N-2 filed by business development companies (BDCs) will be subject to the new disclosure requirements. Furthermore, the final rule will apply only to BDCs internally managed such that they compensate their own employees. Rather, such employees are generally compensated by the BDC's investment adviser. Because we assume that all of the Forms N-2 that will be filed by internally managed BDCs will incorporate by reference the registrant's disclosure contained in its annual report, we have assumed that all of the burden relating to the new disclosure requirements would be associated with Form 10-K.

        j. Form 10

        Because we have assumed that all new registrants would take advantage of the transition period afforded to them under the final rule, we estimate no annual incremental increase in the paperwork burden associated with Form 10 as a result of the new requirements.

    5. Summary of Changes to Annual Compliance Burden in Collection of Information

      Tables 1 and 2 below illustrate the total annual compliance burden of the collection of information in hours and in cost under the final rule for annual reports on Form 10-K and current reports on Form 8-K under the Exchange Act. \745\ The burden estimates

      Page 50183

      were calculated by multiplying the estimated number of annual responses by the estimated average number of hours it would take a company to prepare and review the new disclosure.

      ---------------------------------------------------------------------------

      \745\ Figures in both tables have been rounded to the nearest whole number.

      ---------------------------------------------------------------------------

      As discussed above, there is no change to the estimated burden of the collection of information under Forms S-1, S-4, S-11 or N-2 or under Schedule 14A and 14C because we have assumed that the burden relating to the new disclosure requirements would be associated primarily with Form 10-K. In addition, there is no change to the estimated burden of the collection of information under Form 10 because we have assumed that all new registrants would take advantage of the transition period. There is no change to the estimated burden of the collection of information under Regulation S-K because the burdens that Regulation S-K imposes are reflected in our revised estimates for the forms.\746\

      ---------------------------------------------------------------------------

      \746\ Estimates in columns C and E of Table 1 are affected by rounding to the next burden hour. See Section III.D.1. and 2. above for an analysis of the derivation of the estimated incremental burden hours and costs.

      Table 1--Incremental Paperwork Burden under the Final Rule

      ----------------------------------------------------------------------------------------------------------------

      Total Total

      Number of Hour burden incremental Incremental incremental

      annual per response company burden professional professional

      responses (A) (B) hours (C) = costs (D) costs (E) =

      (A) * (B) (A) * (D)

      ----------------------------------------------------------------------------------------------------------------

      Form 10-K....................... 3,571 663 2,367,563 $88,319.63 $315,389,390

      Form 8-K........................ 10 1 10 $133 $1,330

      -------------------------------------------------------------------------------

      Total....................... .............. .............. 2,367,573 .............. $315,390,720

      ----------------------------------------------------------------------------------------------------------------

      Table 2--Calculation of Total PRA Burden Estimates

      --------------------------------------------------------------------------------------------------------------------------------------------------------

      Current Increase in

      annual Current burden Burden Current Increase in Professional

      responses burden hours (C) hours (D) = professional professional costs (G) = (E) +

      (A)\747\ hours (B) \748\ (B) + (C) costs (E) costs (F) (F)

      --------------------------------------------------------------------------------------------------------------------------------------------------------

      Form 10-K..................................... 8,137 12,198,095 2,367,563 14,565,658 $1,627,400,000 $315,389,390 $1,942,789,390

      Form 8-K...................................... 74,911 507,665 10 507,675 $67,688,700 $1,330 $67,690,030

      ---------------------------------------------------------------------------------------------------------

      Total..................................... 83,048 12,705,760 2,367,573 15,073,333 $1,695,088,700 $315,390,720 $2,010,479,420

      --------------------------------------------------------------------------------------------------------------------------------------------------------

  11. Final Regulatory Flexibility Act Certification

    The Regulatory Flexibility Act of 1980 (``RFA'') \749\ requires us, in promulgating rules, to consider the impact of those rules on small entities. The Commission certified in the Proposing Release, pursuant to Section 605(b) \750\ of the RFA, that the proposed rule, if adopted, would not have a significant economic impact on a substantial number of small entities. We received no comments on this certification.

    ---------------------------------------------------------------------------

    \747\ For these forms, the number of current annual responses reflected in the table equals the three-year average of the number of forms filed with us and currently reported by us to OMB.

    \748\ The increase in burden hours reflected in the table is based on the aggregate incremental burden hours per form multiplied by the annual responses that will be required to include additional disclosure under the new rules as adopted. As explained in the discussion above, for purposes of determining the total increase in burden hours, we have reduced the current number of annual responses to reflect that the disclosure requirements will not apply to all forms filed. See Table 1 for estimates per response.

    \749\ 5 U.S.C. 601 et seq.

    \750\ 5 U.S.C. 605(b).

    ---------------------------------------------------------------------------

    The final rule amends Item 402 by adding paragraph (u) to implement Section 953(b) of the Dodd-Frank Act. Specifically, the final rule requires registrants, other than emerging growth companies, smaller reporting companies and foreign private issuers, to disclose the median of the annual total compensation of all employees of the registrant (excluding the PEO), the annual total compensation of the registrant's PEO, and the ratio between these two amounts. The disclosure is required in any filing described in Item 10(a) of Reg. S-K that requires executive compensation disclosure pursuant to Item 402.

    For purposes of the RFA, under our rules, an issuer, other than an investment company,\751\ is a ``small business'' or ``small organization'' if it has total assets of $5 million or less as of the end of its most recent fiscal year and is engaged or proposing to engage in an offering of securities which does not exceed $5 million.\752\ We believe that the final rule will affect some small entities that are business development companies that have a class of securities registered under Section 12 of the Exchange Act. We estimate that there are approximately five of those business development companies that may be considered small entities.\753\ As discussed above, emerging growth companies and smaller reporting companies are excluded from the final rule. An ``emerging growth company'' is an issuer that had total annual gross revenues of less than $1 billion during its most recently completed fiscal year.\754\ A smaller reporting company is an issuer, other than certain classes of issuers (including investment companies), that had a public float of less than $75 million as of the end of its most recently completed second fiscal quarter, or in the case of an initial registration statement under the Securities Act or Exchange Act for the shares of its common equity, had a public float of less than $75 million as of a date within 30 days of the date of filing of the registration statement.\755\ To

    Page 50184

    the extent that a small entity is a registrant, we believe that there are few, if any, small entities that do not qualify as emerging growth companies or smaller reporting companies because it is unlikely that an entity with total assets of $5 million or less would have total annual gross revenues of $1 billion or more, or would have a public float of $75 million or more. Because emerging growth companies and smaller reporting companies are excluded from the new disclosure requirement, we believe that the final rule applies to few, if any, small entities, other than the five business development companies.

    ---------------------------------------------------------------------------

    \751\ For purposes of the RFA, an investment company is a ``small business'' or ``small organization'' that, together with other investment companies in the same group of related investment companies, has net assets of $50 million or less as of the end of its most recent fiscal year. 17 CFR 270.0-10.

    \752\ See Securities Act Rule 157 17 CFR 230.157 and Exchange Act Rule 0-10(a) 17 CFR 240.0-10(a).

    \753\ We estimate that there are 13 business development companies that will be subject to the final rule, five of which may be considered small entities for purposes of the RFA.

    \754\ See Securities Act Section 2(a)(19) 15 U.S.C. 77b(a)(19).

    \755\ See Securities Act Rule 405 17 CFR 230.405. In the case of an issuer whose public float was zero, the issuer could qualify as a smaller reporting company if it had annual revenues of less than $50 million during the most recently completed fiscal year for which audited financial statements are available.

    ---------------------------------------------------------------------------

    For the above reasons, we again certify, pursuant to 5 U.S.C. 605(b), that the final rule will not have a significant economic impact on a substantial number of small entities.

  12. Statutory Authority

    The amendments contained herein are being proposed pursuant to Sections 7, 10, 19(a), and 28 of the Securities Act, Sections 3(b), 12, 13, 14, 15(d), 23(a), and 36(a) of the Exchange Act, Section 953(b) of the Dodd-Frank Act, as amended, and Section 102(a)(3) of the JOBS Act.

    List of Subjects

    17 CFR Part 229

    Reporting and recordkeeping requirements, Securities.

    17 CFR Part 240

    Brokers, Confidential business information, Fraud, Reporting and recordkeeping requirements, Securities.

    17 CFR Part 249

    Brokers, Reporting and recordkeeping requirements, Securities.

    Text of the Amendments

    In accordance with the foregoing, title 17, chapter II of the Code of Federal Regulations, is amended as follows:

    PART 229 --STANDARD INSTRUCTIONS FOR FILING FORMS UNDER SECURITIES ACT OF 1933, SECURITIES EXCHANGE ACT OF 1934 AND ENERGY POLICY AND CONSERVATION ACT OF 1975--REGULATION S-K

    0

    1. The authority citation for part 229 is revised to read as follows:

      Authority: 15 U.S.C. 77e, 77f, 77g, 77h, 77j, 77k, 77s, 77z-2, 77z-3, 77aa(25), 77aa(26), 77ddd, 77eee, 77ggg, 77hhh, 77iii, 77jjj, 77nnn, 77sss, 78c, 78i, 78j, 78j-3, 78l, 78m, 78n, 78n-1, 78o, 78u-

      5, 78w, 78ll, 78 mm, 80a-8, 80a-9, 80a-20, 80a-29, 80a-30, 80a-

      31(c), 80a-37, 80a-38(a), 80a-39, 80b-11 and 7201 et seq.; 18 U.S.C. 1350; Sec. 953(b), Pub. L. 111-203, 124 Stat. 1904; and Sec. 102(a)(3), Pub. L. 112-106, 126 Stat. 309.

      0

    2. Amend Sec. 229.402 by:

      0

      a. In paragraph (l), removing ``(k) and (s)'' and adding in its place ``(k), (s), and (u)''; and

      0

      b. Adding paragraph (u) directly after the Instructions to Item 402(t).

      The addition reads as follows:

      Sec. 229.402 (Item 402) Executive compensation.

      * * * * *

      (u) Pay ratio disclosure--(1) Disclose. (i) The median of the annual total compensation of all employees of the registrant, except the PEO of the registrant;

      (ii) The annual total compensation of the PEO of the registrant; and

      (iii) The ratio of the amount in paragraph (u)(1)(i) of this Item to the amount in paragraph (u)(1)(ii) of this Item. For purposes of the ratio required by this paragraph (u)(1)(iii), the amount in paragraph (u)(1)(i) of this Item shall equal one, or, alternatively, the ratio may be expressed narratively as the multiple that the amount in paragraph (u)(1)(ii) of this Item bears to the amount in paragraph (u)(1)(i) of this Item.

      (2) For purposes of this paragraph (u):

      (i) Total compensation for the median of annual total compensation of all employees of the registrant and the PEO of the registrant shall be determined in accordance with paragraph (c)(2)(x) of this Item. In determining the total compensation, all references to ``named executive officer'' in this Item and the instructions thereto may be deemed to refer instead, as applicable, to ``employee'' and, for non-salaried employees, references to ``base salary'' and ``salary'' in this Item and the instructions thereto may be deemed to refer instead, as applicable, to ``wages plus overtime'';

      (ii) Annual total compensation means total compensation for the registrant's last completed fiscal year; and

      (iii) Registrant means the registrant and its consolidated subsidiaries.

      (3) For purposes of this paragraph (u), employee or employee of the registrant means an individual employed by the registrant or any of its consolidated subsidiaries, whether as a full-time, part-time, seasonal, or temporary worker, as of a date chosen by the registrant within the last three months of the registrant's last completed fiscal year. The definition of employee or employee of the registrant does not include those workers who are employed, and whose compensation is determined, by an unaffiliated third party but who provide services to the registrant or its consolidated subsidiaries as independent contractors or ``leased'' workers.

      (4) For purposes of this paragraph (u), an employee located in a jurisdiction outside the United States (a ``non-U.S. employee'') may be exempt from the definition of employee or employee of the registrant under either of the following conditions:

      (i) The employee is employed in a foreign jurisdiction in which the laws or regulations governing data privacy are such that, despite its reasonable efforts to obtain or process the information necessary for compliance with this paragraph (u), the registrant is unable to do so without violating such data privacy laws or regulations. The registrant's reasonable efforts shall include, at a minimum, using or seeking an exemption or other relief under any governing data privacy laws or regulations. If the registrant chooses to exclude any employees using this exemption, it shall list the excluded jurisdictions, identify the specific data privacy law or regulation, explain how complying with this paragraph (u) violates such data privacy law or regulation (including the efforts made by the registrant to use or seek an exemption or other relief under such law or regulation), and provide the approximate number of employees exempted from each jurisdiction based on this exemption. In addition, if a registrant excludes any non-

      U.S. employees in a particular jurisdiction under this exemption, it must exclude all non-U.S. employees in that jurisdiction. Further, the registrant shall obtain a legal opinion from counsel that opines on the inability of the registrant to obtain or process the information necessary for compliance with this paragraph (u) without violating the jurisdiction's laws or regulations governing data privacy, including the registrant's inability to obtain an exemption or other relief under any governing laws or regulations. The registrant shall file the legal opinion as an exhibit to the filing in which the pay ratio disclosure is included.

      (ii) The registrant's non-U.S. employees account for 5% or less of the registrant's total employees. In that circumstance, if the registrant chooses to exclude any non-U.S. employees under this exemption, it must exclude all non-U.S. employees. Additionally, if a registrant's non-U.S. employees exceed 5% of the registrant's total U.S. and non-U.S. employees, it may exclude

      Page 50185

      up to 5% of its total employees who are non-U.S. employees; provided, however, if a registrant excludes any non-U.S. employees in a particular jurisdiction, it must exclude all non-U.S. employees in that jurisdiction. If more than 5% of a registrant's employees are located in any one non-U.S. jurisdiction, the registrant may not exclude any employees in that jurisdiction under this exemption.

      (A) In calculating the number of non-U.S. employees that may be excluded under this Item 402(u)(4)(ii) (``de minimis'' exemption), a registrant shall count against the total any non-U.S. employee exempted under the data privacy law exemption under Item 402(u)(4)(i) (``data privacy'' exemption). A registrant may exclude any non-U.S. employee from a jurisdiction that meets the data privacy exemption, even if the number of excluded employees exceeds 5% of the registrant's total employees. If, however, the number of employees excluded under the data privacy exemption equals or exceeds 5% of the registrant's total employees, the registrant may not use the de minimis exemption. Additionally, if the number of employees excluded under the data privacy exemption is less than 5% of the registrant's total employees, the registrant may use the de minimis exemption to exclude no more than the number of non-U.S. employees that, combined with the data privacy exemption, does not exceed 5% of the registrant's total employees.

      (B) If a registrant excludes non-U.S. employees under the de minimis exemption, it must disclose the jurisdiction or jurisdictions from which those employees are being excluded, the approximate number of employees excluded from each jurisdiction under the de minimis exemption, the total number of its U.S. and non-U.S. employees irrespective of any exemption (data privacy or de minimis), and the total number of its U.S. and non-U.S. employees used for its de minimis calculation.

      Instruction 1 to Item 402(u)--Disclosing the date chosen for identifying the median employee. A registrant shall disclose the date within the last three months of its last completed fiscal year that it selected pursuant to paragraph (u)(3) of this Item to identify its median employee. If the registrant changes the date it uses to identify the median employee from the prior year, the registrant shall disclose this change and provide a brief explanation about the reason or reasons for the change.

      Instruction 2 to Item 402(u)--Identifying the median employee. A registrant is required to identify its median employee only once every three years and calculate total compensation for that employee each year; provided that, during a registrant's last completed fiscal year there has been no change in its employee population or employee compensation arrangements that it reasonably believes would result in a significant change to its pay ratio disclosure. If there have been no changes that the registrant reasonably believes would significantly affect its pay ratio disclosure, the registrant shall disclose that it is using the same median employee in its pay ratio calculation and describe briefly the basis for its reasonable belief. For example, the registrant could disclose that there has been no change in its employee population or employee compensation arrangements that it believes would significantly impact the pay ratio disclosure. If there has been a change in the registrant's employee population or employee compensation arrangements that the registrant reasonably believes would result in a significant change in its pay ratio disclosure, the registrant shall re-identify the median employee for that fiscal year. If it is no longer appropriate for the registrant to use the median employee identified in year one as the median employee in years two or three because of a change in the original median employee's circumstances that the registrant reasonably believes would result in a significant change in its pay ratio disclosure, the registrant may use another employee whose compensation is substantially similar to the original median employee based on the compensation measure used to select the original median employee.

      Instruction 3 to Item 402(u)--Updating for the last completed fiscal year. Pay ratio information (i.e., the disclosure called for by paragraph (u)(1) of this Item) with respect to the registrant's last completed fiscal year is not required to be disclosed until the filing of its annual report on Form 10-K for that last completed fiscal year or, if later, the filing of a definitive proxy or information statement relating to its next annual meeting of shareholders (or written consents in lieu of such a meeting) following the end of such fiscal year; provided that, the required pay ratio information must, in any event, be filed as provided in General Instruction G(3) of Form 10-K (17 CFR 249.310) not later than 120 days after the end of such fiscal year.

      Instruction 4 to Item 402(u)--Methodology and use of estimates. 1. Registrants may use reasonable estimates both in the methodology used to identify the median employee and in calculating the annual total compensation or any elements of total compensation for employees other than the PEO.

    3. In determining the employees from which the median employee is identified, a registrant may use its employee population or statistical sampling and/or other reasonable methods.

    4. A registrant may identify the median employee using annual total compensation or any other compensation measure that is consistently applied to all employees included in the calculation, such as information derived from the registrant's tax and/or payroll records. In using a compensation measure other than annual total compensation to identify the median employee, if that measure is recorded on a basis other than the registrant's fiscal year (such as information derived from tax and/or payroll records), the registrant may use the same annual period that is used to derive those amounts. Where a compensation measure other than annual total compensation is used to identify the median employee, the registrant must disclose the compensation measure used.

    5. In identifying the median employee, whether using annual total compensation or any other compensation measure that is consistently applied to all employees included in the calculation, the registrant may make cost-of-living adjustments to the compensation of employees in jurisdictions other than the jurisdiction in which the PEO resides so that the compensation is adjusted to the cost of living in the jurisdiction in which the PEO resides. If the registrant uses a cost-

      of-living adjustment to identify the median employee, and the median employee identified is an employee in a jurisdiction other than the jurisdiction in which the PEO resides, the registrant must use the same cost-of-living adjustment in calculating the median employee's annual total compensation and disclose the median employee's jurisdiction. The registrant also shall briefly describe the cost-of-living adjustments it used to identify the median employee and briefly describe the cost-

      of-living adjustments it used to calculate the median employee's annual total compensation, including the measure used as the basis for the cost-of-living adjustment. A registrant electing to present the pay ratio in this manner also shall disclose the median employee's annual total compensation and pay ratio without the cost-of-living adjustment. To calculate this pay ratio, the registrant will need to identify the

      Page 50186

      median employee without using any cost-of-living adjustments.

    6. The registrant shall briefly describe the methodology it used to identify the median employee. It shall also briefly describe any material assumptions, adjustments (including any cost-of-living adjustments), or estimates it used to identify the median employee or to determine total compensation or any elements of total compensation, which shall be consistently applied. The registrant shall clearly identify any estimates used. The required descriptions should be a brief overview; it is not necessary for the registrant to provide technical analyses or formulas. If a registrant changes its methodology or its material assumptions, adjustments, or estimates from those used in its pay ratio disclosure for the prior fiscal year, and if the effects of any such change are significant, the registrant shall briefly describe the change and the reasons for the change. Registrants must also disclose if they changed from using the cost-of-living adjustment to not using that adjustment and if they changed from not using the cost-of-living adjustment to using it.

    7. Registrants may, at their discretion, include personal benefits that aggregate less than $10,000 and compensation under non-

      discriminatory benefit plans in calculating the annual total compensation of the median employee as long as these items are also included in calculating the PEO's annual total compensation. The registrant shall also explain any difference between the PEO's annual total compensation used in the pay ratio disclosure and the total compensation amounts reflected in the Summary Compensation Table, if material.

      Instruction 5 to Item 402(u)--Permitted annualizing adjustments. A registrant may annualize the total compensation for all permanent employees (full-time or part-time) that were employed by the registrant for less than the full fiscal year (such as newly hired employees or permanent employees on an unpaid leave of absence during the period). A registrant may not annualize the total compensation for employees in temporary or seasonal positions. A registrant may not make a full-time equivalent adjustment for any employee.

      Instruction 6 to Item 402(u)--PEO compensation not available. A registrant that is relying on Instruction 1 to Item 402(c)(2)(iii) and (iv) in connection with the salary or bonus of the PEO for the last completed fiscal year, shall disclose that the pay ratio required by paragraph (u) of this Item is not calculable until the PEO salary or bonus, as applicable, is determined and shall disclose the date that the PEO's actual total compensation is expected to be determined. The disclosure required by paragraph (u) of this Item shall then be disclosed in the filing under Item 5.02(f) of Form 8-K (17 CFR 249.308) that discloses the PEO's salary or bonus in accordance with Instruction 1 to Item 402(c)(2)(iii) and (iv).

      Instruction 7 to Item 402(u)--Transition periods for registrants. 1. Upon becoming subject to the requirements of Section 13(a) or 15(d) of the Exchange Act (15 U.S.C. 78m or 78o(d)), a registrant shall comply with paragraph (u) of this Item with respect to compensation for the first fiscal year following the year in which it became subject to such requirements, but not for any fiscal year commencing before January 1, 2017. The registrant may omit the disclosure required by paragraph (u) of this Item from any filing until the filing of its annual report on Form 10-K (17 CFR 249.310) for such fiscal year or, if later, the filing of a proxy or information statement relating to its next annual meeting of shareholders (or written consents in lieu of such a meeting) following the end of such year; provided that, such disclosure shall, in any event, be filed as provided in General Instruction G(3) of Form 10-K not later than 120 days after the end of such fiscal year.

    8. A registrant may omit any employees that became its employees as the result of the business combination or acquisition of a business for the fiscal year in which the transaction becomes effective, but the registrant must disclose the approximate number of employees it is omitting. Those employees shall be included in the total employee count for the triennial calculations of the median employee in the year following the transaction for purposes of evaluating whether a significant change had occurred. The registrant shall identify the acquired business excluded for the fiscal year in which the business combination or acquisition becomes effective.

    9. A registrant shall comply with paragraph (u) of this Item with respect to compensation for the first fiscal year commencing on or after the date the registrant ceases to be a smaller reporting company, but not for any fiscal year commencing before January 1, 2017.

      Instruction 8 to Item 402(u)--Emerging growth companies. A registrant is not required to comply with paragraph (u) of this Item if it is an emerging growth company as defined in Section 2(a)(19) of the Securities Act (15 U.S.C. 77(b)(a)(19)) or Section 3(a)(80) of the Exchange Act (15 U.S.C. 78c(a)(80)). A registrant shall comply with paragraph (u) of this Item with respect to compensation for the first fiscal year commencing on or after the date the registrant ceases to be an emerging growth company, but not for any fiscal year commencing before January 1, 2017.

      Instruction 9 to Item 402(u)--Additional information. Registrants may present additional information, including additional ratios, to supplement the required ratio, but are not required to do so. Any additional information shall be clearly identified, not misleading, and not presented with greater prominence than the required ratio.

      Instruction 10 to Item 402(u)--Multiple PEOs during the year. A registrant with more than one non-concurrent PEO serving during its fiscal year may calculate the annual total compensation for its PEO in either of the following manners:

    10. The registrant may calculate the compensation provided to each person who served as PEO during the year for the time he or she served as PEO and combine those figures; or

    11. The registrant may look to the PEO serving in that position on the date it selects to identify the median employee and annualize that PEO's compensation.

      Regardless of the alternative selected, the registrant shall disclose which option it chose and how it calculated its PEO's annual total compensation.

      Instruction 11 to Item 402(u)--Employees' personally identifiable information. Registrants are not required to, and should not, disclose any personally identifiable information about that employee other than his or her compensation. Registrants may choose to generally identify an employee's position to put the employee's compensation in context, but registrants are not required to provide this information and should not do so if providing the information could identify any specific individual.

      * * * * *

      PART 240--GENERAL RULES AND REGULATIONS, SECURITIES EXCHANGE ACT OF 1934

      0

    12. The authority citation for part 240 continues to read, in part, as follows:

      Authority: 15 U.S.C. 77c, 77d, 77g, 77j, 77s, 77z-2, 77z-3, 77eee, 77ggg, 77nnn, 77sss, 77ttt, 78c, 78c-3, 78c-5, 78d, 78e, 78f, 78g, 78i, 78j, 78j-1, 78k, 78k-1, 78m, 78n, 78n-1, 78o, 78o-4, 78o-

      10, 78p, 78q, 78q-1, 78s, 78u-5, 78w, 78x, 78ll, 78mm, 80a-20, 80a-

      23, 80a-29, 80a-37, 80b-3, 80b-4, 80b-

      Page 50187

      11, and 7210 et seq., and 8302; 7 U.S.C. 2(c)(2)(E); 12 U.S.C. 5521(e)(3); 18 U.S.C. 1350; and Pub. L. 111-203, 939A, 124 Stat. 1376 (2010), unless otherwise noted.

      * * * * *

      0

    13. Amend Sec. 240.14a-101 by adding Item 25 at the end to read as follows:

      Sec. 240.14a-101 Schedule 14A. Information required in proxy statement.

      SCHEDULE 14A INFORMATION

      * * * * *

      Item 25. Exhibits. Provide the legal opinion required to be filed by Item 402(u)(4)(i) of Regulation S-K (17 CFR 229.402(u)) in an exhibit to this Schedule 14A.

      PART 249 -- FORMS, SECURITIES EXCHANGE ACT OF 1934

      0

    14. The authority citation for part 249 is revised, in part, to read as follows:

      Authority: 15 U.S.C. 78a et seq. and 7201 et seq.; 12 U.S.C. 5461 et seq.; 18 U.S.C. 1350; Sec. 953(b), Pub. L. 111-203, 124 Stat. 1904; and Sec. 102(a)(3), Pub. L. 112-106, 126 Stat. 309, unless otherwise noted.

      * * * * *

      0

    15. Form 8-K (referenced in Sec. 249.308) is amended by redesignating paragraph (f) as (f)(1) and adding paragraph (f)(2) to read as follows:

      Form 8-K

      * * * * *

      Item 5.02 Departure of Directors or Certain Officers; Election of Directors; Appointment of Certain Officers; Compensatory Arrangements of Certain Officers.

      * * * * *

      (f)(1) * * *

      (2) As specified in Instruction 6 to Item 402(u) of Regulation S-K (17 CFR 229.402(u)), disclosure under this Item 5.02(f) with respect to the salary or bonus of a principal executive officer shall include pay ratio disclosure pursuant to Item 402(u) of Regulation S-K calculated using the new total compensation figure for the principal executive officer. Pay ratio disclosure is not required under this Item 5.02(f) until the omitted salary or bonus amounts for such principal executive officer become calculable in whole.

      * * * * *

      Note: The text of Form 8-K does not, and this amendment will not, appear in the Code of Federal Regulations.

      By the Commission.

      Dated: August 5, 2015.

      Brent J. Fields,

      Secretary.

      FR Doc. 2015-19600 Filed 8-17-15; 8:45 am

      BILLING CODE 8011-01-P