July 6, 2015
Sent via electronic mail: [email protected]
Mr. Brent J. Fields
Secretary
U.S. Securities and Exchange Commission
100 F Street, NE
Washington, DC 20549-1090
Re: Pay Ratio Disclosure, File No. S7-07-13
Dear Mr. Fields:
On behalf of the American Federation of Labor and Congress of Industrial Organizations (the “AFL-CIO”), I am writing to provide comments to the Securities and Exchange Commission (the “SEC”) on the Division of Economic and Risk Analysis’ memorandum on the potential effect on pay ratio disclosure of exclusion of different percentages of employees at a range of thresholds. In addition, this letter addresses certain issues to supplement the AFL-CIO’s previously submitted comments1 on the SEC’s proposed pay ratio disclosure rule as required to be issued by Section 953(b) of the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010.
Exclusion of Employees from Pay Ratio Calculations
As a threshold matter, the SEC simply does not have regulatory flexibility under Section 953(b) to permit the exclusion of employees from pay ratio calculations. Section 953(b) unambiguously states that the SEC must require that companies calculate and disclose “(A) the median of the annual total compensation of all employees of the issuer” (emphasis added). By any logical definition that a court is likely follow to determine if the SEC has complied with its congressional mandate under Section 953(b), the term “all employees” of an issuer simply means all employees.
Nor is the exclusion of employee groups from pay ratio disclosure calculations needed to reduce the regulatory burden of compliance with Section 953(b). As has been proposed by the SEC, companies will have substantial flexibility to use reasonable estimates to identify the median. A statistical sampling method may be based on the size and structure of their own business.
As shown by the Division of Economic and Risk Analysis memorandum, the exclusion of employees from pay ratio calculations will make the disclosed information materially less accurate. For example, the memorandum estimates that excluding 20 percent of a company’s lowest-paid employees may decrease the disclosed pay ratio by up to 13 percent. This expected variance illustrates why the SEC may not permit the categorical exclusion of employees in foreign countries or part-time, seasonal, or temporary employees. Many companies, for example multinationals and retailers, will have a far greater percentage of their employees in these categories.
The Division of Economic and Risk Analysis memorandum understates the likely impact of excluding employees.2 The memorandum uses several estimates of variability in the log of pay for company employees ranging from a standard deviation of 0.25 to 0.55. As the standard deviation increases, the likely impact of excluding employees also increases. However, there is reason to believe that the dispersion of employee compensation may be much higher than a standard deviation of 0.55, particularly at large publicly-traded companies with global workforces. A recent study of U.K. companies found that within-firm pay disparities increase with firm size.3
The 0.25 and 0.35 standard deviation ranges of wage dispersion that are used by the Division of Economic and Risk Analysis memorandum lack factual support and are likely to understate the real world standard deviation of employee compensation. The only empirical data cited to support a 0.25 standard deviation is federal employee wages, but the federal General Schedule pay classification system is hardly comparable to the private sector. Federal employee pay data is not relevant factual support for assuming a 0.25 standard deviation in private sector companies. Nor does the memorandum provide any data or citations to support a 0.35 standard deviation.
All the academic studies cited by the memorandum suggest a higher standard deviation of 0.45 or 0.55. Yet even these academic studies may understate the real world dispersion of employee compensation within an entire publicly-traded company. For example, the Bureau of Labor Statistics (“BLS”) Monthly Labor Review study measures the dispersion of hourly wages, not annual total compensation.4 Companies with significant part-time, seasonal, or temporary employees are likely to have much higher standard deviations in the annual total compensation of their employees.
In addition, the studies cited by the memorandum measure wage dispersion within establishments, not within firms. The BLS study looks at the Occupational Employment Survey, which defines an establishment to be “the physical location of a certain economic activity, for example, a factory, mine, store, or office.”5 The Barth et al. study estimates within-establishment variance based on unemployment insurance data.6 Finally, the memorandum derives a within-establishment variance from the Leonesio and Del Bene study of economy-wide Social Security data.7 However, firm-wide pay disparities are likely to be higher than establishment-wide disparities.8
Measurement Date for Determining All Employees
Certain commentators have suggested that a fiscal year-end measurement date to determine median employee compensation levels may not provide adequate time to collect and analyze the required data. To address this concern, companies should be given flexibility to determine their measurement date because median employee compensation levels are unlikely to fluctuate significantly from day to day. However, to ensure accurate disclosure for companies that employ a high number of seasonal employees, companies should be required to calculate the median annual compensation of all employees employed at any time over the preceding 365 days.
Privacy Laws and Employee Pay Data Collection
Other commentators have suggested that certain privacy laws may prohibit the collection or disclosure of employee compensation data. The AFL-CIO believes that any privacy law concerns can be easily addressed by anonymizing payroll data sets or by conducting statistical sampling. Furthermore, we believe that for the large, multi-national companies for which the application of international privacy laws may be a concern it would be impossible to reverse engineer the data sets and create the risk of disclosure of individual information. If these privacy safeguards are insufficient, companies should be permitted to exclude employees in such countries only if they disclose the number of employees in the excluded countries and obtain a legal opinion by qualified outside counsel demonstrating that these privacy safeguards are inadequate under local law. Issuers should be required to file these legal opinions as an exhibit to the periodic report in which the pay ratio disclosure appears.
***
In conclusion, we support the SEC’s proposed pay ratio disclosure rule as issued in 2013 and urge the SEC to adopt a final rule as soon as possible to fulfill its legal obligations under Section 953(b) of the Dodd-Frank Act. Thank you for taking the AFL-CIO’s views into consideration regarding this matter. If the AFL-CIO can be of further assistance, please contact Brandon Rees at (202) 637-5152 or [email protected].
Sincerely,
Heather Slavkin Corzo, Director
Office of Investment
1 AFL-CIO letter to the SEC dated December 3, 2013, available at https://www.sec.gov/comments/s7-07-13/s70713-562.pdf; AFL-CIO letter to the SEC dated August 11, 2011, available at http://www.sec.gov/comments/df-title-ix/executive-compensation/executivecompensation-78.pdf; AFL-CIO letter to the SEC dated December 13, 2010, available at https://www.sec.gov/comments/df-title-ix/executive-compensation/executivecompensation-54.pdf.
2 The AFL-CIO expresses its appreciation to Nathan Wilmers, Doctoral Student in Sociology, Harvard University, for his technical assistance in analyzing the Division of Economic and Risk Analysis memorandum.
3 H. Mueller, E. Simintzi and P. Ouimet, “Wage inequality and firm growth,” LIS Working Paper 632 (March 2015) which finds that wage differentials between high and low skill jobs increase with firm size.
4 “Measuring the distribution of wages in the United States from 1996 through 2010 using the Occupational Employment Survey,” Monthly Labor Review, Bureau of Labor Statistics (May 2014).
5 Occupational Employment Statistics Frequently Asked Questions, Bureau of Labor Statistics, available at http://www.bls.gov/oes/oes_ques.htm.
6 Barth, E., Bryson, A., Davis, J., Freeman, R., “It’s Where You Work: Increases in earnings dispersion across establishments and individuals in the U.S.,” IZA Discussion Paper No. 8437 (2014).
7 Leonesio, M., Del Bene, L., “The distribution of annual and long-run US earnings, 1981–2004,” Social Security Bulletin, Vol. 71(1) (2011).
8 See Mueller and Ouimet (2015).