R A N D E L K. J O H N S O N S E N I O R V I C E P R E S I D E N T L A B O R, I M M I G R A T I O N & E M P L O Y E E B E N E F I T S C H A M B E R O F C O M M E R C E O F T H E U N I T E D S T A T E S O F A M E R I C A 1 6 1 5 H S T R E E T, N. W. W A S H I N G T O N, D. C. 2 0 0 6 2 2 0 2 / 4 6 3-5 5 2 2 M I C H A E L J. E A S T M A N E X E C U T I V E D I R E C T O R L A B O R L A W P O L I C Y October 11, 2011 Ms. Debra A. Carr Director, Division of Policy, Planning, and Program Development Office of Federal Contract Compliance Programs, Room C-3325 U.S. Department of Labor 200 Constitution Avenue, NW Washington, DC 20210 RE: RIN 1250-AA03; Non-Discrimination in Compensation; Compensation Data Collection Tool; Advanced Notice of Proposed Rulemaking Dear Ms. Carr: On behalf of the U.S. Chamber of Commerce, we are pleased to submit these comments in response to the Office of Federal Contract Compliance Programs (OFCCP) advanced notice of proposed rulemaking (ANPRM) relating to non-discrimination in compensation and a compensation data collection tool, as published in the Federal Register on August 10, 2011. 1 The U.S. Chamber of Commerce (Chamber) is the world s largest business federation representing the interests of more than three million businesses and organizations of every size, sector, and region. Chamber members include a large number of federal contractors and subcontractors covered by Executive Order 11246 or otherwise within the OFCCP s jurisdiction. In addition, the Chamber represents many trade associations and state and local chambers of commerce that, in turn, represent a significant number of federal contractors and subcontractors. The issues raised in the ANPRM are of considerable concern to these members and any further regulation in this field will likely have a significant impact on these members. Preliminary Statement Federal contractors are wary of the approach that the OFCCP has taken recently with respect to data collection. The agency is pushing contractors to dramatically expand the types of information and records maintained by contractors and disclosed to the agency. 2 These new procedures will dramatically increase costs to contractors, and consequently to taxpayers, that far exceed any expected benefits. What s more, the OFCCP appears to be making significant policy 1 76 Fed. Reg. 49,398. 2 See, e.g., Affirmative Action and Nondiscrimination Obligations of Contractors and Subcontractors Regarding Protected Veterans, Notice of Proposed Rulemaking, 76 Fed. Reg. 23,358 (April 26, 2011); Proposed Extension of the Approval of Information Collection Requirements Non-construction Supply and Service Information Collection, 76 Fed. Reg. 27,670 (May 12, 2011).
changes that are not concretely based on the statutes and Executive Orders from which the OFCCP derives its authority. The current ANPRM is no different. The Federal Register notice poses 15 specific questions about the possible design of a new data tool. However, the decision over whether to create a new data tool appears to have been already made without public input. Likewise, the ANPRM does not clearly articulate the need for the development of the new data tool, instead merely noting that it would provide insight into potential problems of pay discrimination that warrant further review and suggesting possible uses for the collected data. It is incumbent upon the agency to first clearly articulate the need for a new data collection tool, for example, by detailing why current data collection tools and those already proposed by the agency (such as the proposed revisions to the Scheduling Letter and Itemized Listing) are inadequate for the agency to fulfill its mission. Only after clearly articulating the problem that the agency wishes to solve should the OFCCP seek input on possible solutions. The Chamber and others representing the contracting community would welcome a discussion over how the OFCCP could better fulfill its mission or target enforcement resources. However, it is exceedingly difficult to pose alternative solutions that would serve OFCCP s legitimate needs without understanding the purpose. By its failure to adhere to a more rational process, the OFCCP cannot be surprised that contractors fear that the agency is seeking to engage in a fishing expedition, compelling disclosure of vast amounts of information with no clear purpose that will impose tremendous new burdens. It likewise cannot be surprised that contractors question whether the agency is seeking to implement by fiat provisions of the Paycheck Fairness Act that were ultimately rejected by Congress. These fears are only heightened by the inclusion in the Federal Register notice of a highly misleading discussion regarding the pay gap. Consequently, the Chamber opposes the development of the new compensation data collection tool and urges the OFCCP to first engage stakeholders in a discussion of why the agency believes current processes are not working appropriately and effectively and what purposes would be advanced by changing data collection methodology. Misleading Discussion of the Pay Gap OFCCP s discussion of the need for a new data collection tool is not based on a reasoned analysis of available data, published economic research or its own enforcement data. OFCCP instead appears to justify the new data collection tool on the so-called pay gap. The pay gap represents the raw difference in earnings between two groups, such as between women and men. We find it curious that the OFCCP has chosen not to rely on the Bureau of Labor Statistics for determination of the raw pay gap. In its Federal Register notice, the OFCCP cites to data from the U.S. Census Bureau that the OFCCP admittedly has not even accessed since December, 2010. 3 But the most recent report from the Bureau of Labor Statistics, published in July of this year, shows a raw wage gap that is 4 percentage points less. 4 One is left to wonder why the agency is using a less recent report from another agency. 3 76 Fed. Reg. at 49,399-400 n.5. 4 Compare Table PINC-05; Work Experiences in 2009 People 15 Years Old and Over by Total Money Earnings in 2009, Age, Race, Hispanic Origin, and Sex, available at: (continued on next page)
More troubling, however, is the biased and incomplete discussion of the pay gap. Among other misleading statements, the OFCCP asserts that Ultimately, the research literature still finds that an unexplained gap exists even after accounting for potential explanations. 5 However, while citing to selective studies to bolster its argument, the OFCCP ignores what is perhaps the most complete and authoritative study of the pay gap even conducted a study that the OFCCP itself commissioned from the CONSAD Research Corporation just a few years ago. In a preface to the study, the OFCCP wrote: There are observable differences in the attributes of men and women that account for most of the wage gap. Statistical analysis that includes those variables has produced results that collectively account for between 65.1 and 76.4 percent of a raw gender wage gap of 20.4 percent, and thereby leave an adjusted gender wage gap that is between 4.8 and 7.1 percent. These variables include: A greater percentage of women than men tend to work part-time. Part-time work tends to pay less than full-time work. A greater percentage of women than men tend to leave the labor force for child birth, child care and elder care. Some of the wage gap is explained by the percentage of women who were not in the labor force during previous years, the age of women, and the number of children in the home. Women, especially working mothers, tend to value family friendly workplace policies more than men. Some of the wage gap is explained by industry and occupation, particularly, the percentage of women who work in the industry and occupation. Research also suggests that differences not incorporated into the model due to data limitations may account for part of the remaining gap. Specifically, CONSAD s model and much of the literature, including the Bureau of Labor Statistics Highlights of Women s Earnings, focus on wages rather than total compensation. Research indicates that women may value non-wage benefits more than men do, and as a result prefer to take a greater portion of their compensation in the form of health insurance and other fringe benefits. In principle, more of the raw wage gap could be explained by including some additional variables within a single comprehensive analysis that considers all of the factors simultaneously; however, such an analysis is not feasible to conduct with available data bases. Factors, such as work experience and job tenure, require data that describe the behavior of individual workers over extended time periods. The longitudinal data bases that contain such information include too few workers, however, to support adequate analysis of factors like occupation and industry. Cross-sectional data bases that include enough workers to enable http://www.census.gov/hhes/www/cpstables/032010/perinc/toc.htm with Highlights of Women s Earnings, available at: http://www.bls.gov/cps/cpswom2010.pdf. 5 76 Fed. Reg. at 49,400 (citing Jacobsen, Joyce P., The Economics of Gender (2007)).
analysis of factors like occupation and industry do not collect data on individual workers over long enough periods to support adequate analysis of factors like work experience and job tenure. Although additional research in this area is clearly needed, this study leads to the unambiguous conclusion that the differences in the compensation of men and women are the result of a multitude of factors and that the raw wage gap should not be used as the basis to justify corrective action. Indeed, there may be nothing to correct. The differences in raw wages may be almost entirely the result of the individual choices being made by both male and female workers. 6 Empirical studies that have inferred possible discrimination from comparative compensation data have done so only by demonstrating the existence of a residual earnings difference that is correlated with gender, race or ethnicity after removing the effect of legitimate, non-discrimination factors and characteristics of individuals. The unexplained residual of 4.8 percent to 7.1 percent found in the OFCCP s study is consistent with the unexplained differences found in many other economic studies. In addition, studies of both gender and racial earnings differences have found that the unexplained residual compensation differences have become smaller over time. The relative smallness of unexplained compensation differences and the trend of observed residuals getting smaller over time are facts that OFCCP should consider before proceeding with a new regulatory approach. It is important to recognize, also, that even when studies reveal an unexplained residual compensation difference, it is incorrect to infer that the entire residual reflects discrimination. No available data source identifies the full set of individual characteristics that may legitimately explain and be reflected in compensation differences. Some of the residual is necessarily associated with unobserved explanatory characteristics, with the result that discrimination, if it exists at all, is a smaller proportion than the observed compensation residual. These points were all addressed by the OFCCP s 2009 study that the agency for some reason now chooses to ignore. We know there may be some within the Department who wish that the 2009 study did not exist indeed, when the study disappeared from the Labor Department s website without explanation, the Chamber filed a Freedom of Information Act request with the Labor Department for a copy of the report and related information. E-mails obtained through that request showed that career employees at the Department, acting on an emergency basis, removed the report on the weekend before President Obama was inaugurated. Consequently, to be sure the report is considered and is a part of the record for this proceeding, we have attached a copy as part of our comments. End Run Around Congress Considering the ANPRM in conjunction with proposed changes to the Scheduling Letter and Itemized listing, it appears that the agency has embarked on a backdoor attempt for the 6 CONSAD Research Corporation, An Analysis of the Reasons for the Disparity in Wages Between Men and Women Final Report (January 12, 2009), p. 1 (Foreword).
government to begin creating a massive database of private sector compensation data. Such a provision was a component of the Paycheck Fairness Act (PFA), 7 legislation that was rejected by Congress as recently as last year. The provision of the PFA would have charged the Equal Employment Opportunity Commission (EEOC) with responsibilities for creating the massive database, but the purpose appears the same. The connection between OFCCP s current efforts and the rejected PFA is further evidenced by a report by the National Equal Pay Enforcement Task Force. 8 The Report explicitly states as one of its goals: 2. Collect data on the private workforce to better understand the scope of the pay gap and target enforcement efforts. Private sector employers are not required to systematically report gender-identified wage data to the federal government. This lack of data makes identifying wage discrimination difficult and undercuts enforcement efforts. We must identify ways to collect wage data from employers that are useful to enforcement agencies but do not create unnecessary burdens on employers. 9 The Report explicitly references how the PFA would have helped the government collect such information and that OFCCP may be one agency that could collect such data. 10 The PFA s compensation data requirements were among the many reasons employers opposed the legislation and were among the reasons Congress rejected the bill. 11 Disregarding Congressional rejection of this proposal, OFCCP is now seeking to implement it. It is bad enough for an agency to propose an end-run around Congress through regulation, but the OFCCP is not seeking to make this radical policy change in an up-front manner, but has, with respect to the proposed changes to the Scheduling Letter and Itemized Listing, instead attempted to sneak it through in the Paperwork Reduction Act clearance process. Furthermore, even the National Equal Pay Task Force recognized that wage data should not create unnecessary burdens on employers. We address the inadequate burden estimates below. Privacy Concerns The issues raised by the ANPRM raise the same privacy concerns that are raised by the OFCCP s proposed changes to its Scheduling Letter and Itemized Listing. As the OFCCP well knows, much of the data that the agency would gather through a new collection tool is viewed as extremely sensitive. 12 Indeed, much of the data that would be collected are viewed as proprietary 7 Many substantially similar versions of the bill have been introduced in recent years. S. 3772 failed a cloture vote on November 17, 2010. 8 Available at: http://www.whitehouse.gov/sites/default/files/rss_viewer/ equal_pay_task_force.pdf. 9 Id. at 5 (emphasis in original). 10 Id. 11 For example, see the HR Policy Association s Policy Brief, available at: http://www. hrpolicy.org/documents/positions/10-55%20paycheck%20fairness%20policy% 20Brief%204%2025%2010.pdf; See Statement by Sen. Michael B. Enzi, 156 Cong. Rec. S7926 (daily ed. Nov. 17, 2010). 12 See, e.g., Supporting Statement, Supply and Service Program, OMB No. 1250-0003 (formerly 1215-0072) at 9.
and confidential by many contractors and its disclosure to competitors could decrease a contractor s competitive advantage or even threaten its business model. Further expansion of the types of data collected, especially individualized compensation and benefits, will greatly exacerbate this concern. The OFCCP provides no guarantee that this type of data will be protected from disclosure to competitors or others. Indeed, the OFCCP follows an exceedingly lax view of its responsibilities under the Freedom of Information Act. It requires contractors who submit data to the government to bear the entire burden of opposing improper disclosure of the information. 13 Thus, the burden of protecting the confidentiality of the submitted information is left to the contractor. The provision in the disclosure regulations suggesting that confidential pay information not be disclosed is a slim reed for the contractor to depend upon since the OFCCP policy is to simply notify the contractor of a pending disclosure and require the contractor to go to Court to defend the privilege of its information and to undertake the responsibilities that the OFCCP itself should undertake. Even if the OFCCP were to make such a guarantee, privacy concerns would still remain. Reports are replete with sensitive government data that is lost, stolen, or deliberately leaked. For example, one GAO analysis from 2007 found that 17 agencies reported that they experienced at least one breach and, collectively, the agencies reported more than 788 separate incidents. 14 The inherent risks of disclosure argue against the OFCCP s developing a data collection tool that collects vast amounts of sensitive data. This is not to say OFCCP is never entitled to this information. Clearly such information may be relevant and necessary to help determine whether contractors have complied with their obligations in particular cases. However, the Chamber cannot support massive new disclosure requirements applicable to all contractors. OFCCP must demonstrate some appropriate foundation before collecting such sensitive information. Need for Data Tool and Burden Estimate Creating an Effective Data Tool May Not Be Possible When OFCCP proposed eliminating the EO Survey, it asked for comments on whether the survey could somehow be modified. Just as the EO Survey could not have been modified to achieve important policy goals, it is also unlikely that the OFCCP can create a data tool without facing the same challenges. At the time the agency proposed rescinding the Survey, we made the following observation: In simplest terms, the Survey is fundamentally flawed because it assumes incorrectly that a one size fits all approach to assessing compensation data is viable. No such approach is possible when evaluating systems as complex as compensation systems. As the Agency well knows, contractors maintain a broad range of compensation systems, all unique to their particular workplaces. Those systems encompass a wide variety of jobs 13 See, e.g., 41 CFR section 60-40. 14 U.S. Government Accountability Office, Personal Information: Data Breaches Are Frequent, but Evidence of Resulting Identity Theft Is Limited; However, the Full Extent Is Unknown, at 13 (Jun. 2007).
and reflect different approaches to compensation. When analyzing compensation across a company one must (1) compare similarly-situated individuals, and (2) consider the myriad factors that may impact compensation, such as education, prior work experience, time in the job, and performance. What s more, the determination of who is similarly situated and what factors impact pay necessarily varies significantly from contractor to contractor. There is simply no effective way to gather, through a survey instrument, information that would permit the OFCCP to compare contractors across industries or even within an industry, given the requisite variation in compensation practices. There is likewise no way to revise the Survey to permit the Agency to account for the individualization that is inherent in any system of pay. 15 The same is true today with respect to the development of a new compensation data tool. In order for a data collection instrument to credibly identify possible systemic compensation discrimination, it is essential that data collected include sufficient detail regarding population characteristics to distinguish between compensation differences that reflect legitimate characteristic differences from those that may reflect discrimination. A data collection instrument that ignores differences in educational attainment or training, years of relevant experience (both overall and tenure with the current employer), usual work hours, measures of individual productivity, and so forth is of little utility. To assess compensation differences among Federal contractor employees, it may be particularly relevant to include measures indicating the extent to which individual compensation reflects contribution to acquisition of contract task order and successful completion of work on schedule and within budget. It may also be relevant to consider differences in skills that may not be reflected in simple educational attainment indicators, e.g., knowledge of languages, software, or technical trades. With respect to educational attainment, it may be insufficient, for example, to simply compare employees who have baccalaureate degrees if compensation variation may also legitimately reflect the subject major of the degree, the academic class rank of the degree recipient, or market-reflected differences in compensation of degree holders from institutions of different prestige. With respect to work experience, simple measures of aggregate work experience years may be insufficient to explain compensation differences if compensation also legitimately may be affected by interruptions in experience (periods unemployed or not in labor force) or differences in quality of experience as indicated by types of projects or tasks completed. OFCCP should be aware that age, which is often used as an explanatory variable in compensation studies, is a poor proxy for work experience. In addition, job categories should be defined at the most detailed level possible. The job category agglomerations used in the OFCCP EO survey form and in the EEO-1 form are wholly inadequate for purposes of compensation analysis. For example, the simplistic category technicians may encompass much market variation in compensation across the various detailed types of technicians included. Differences in compensation of technicians by gender or race, may reflect not compensation discrimination but rather concentration of different groups by their 15 U.S. Chamber of Commerce, Comments to the OFCCP in Response to the Notice of Proposed Rulemaking Regarding the Equal Opportunity Survey (Mar. 28, 2006) at 5.
own choice of occupation in different technical specialties which command different market compensation levels legitimately reflecting differences in productivity, which is implicit in labor demand, and in labor supply, which reflects in part the cost and difficulty of training and experience acquisition. The need for such exacting detail in order to create a valid instrument is what doomed the EO Survey it collected irrelevant data and thus produced no meaningful results. The OFCCP now is poised to either repeat the same mistake or to impose staggering new burdens and costs on contractors and subcontractors. It should be noted that some Chamber members seeking to perform compensation analyses of their own workforce, such as in response to large scale litigation, have literally spent millions of dollars and countless hours in order to create a valid assessment. We do not believe it is possible for the OFCCP to construct a valid data collection tool without imposing dramatic costs on contractors. The OFCCP s Data Do Not Support Development of a New Tool OFCCP s own published enforcement data provides cause to question the significance of the purported problem to be addressed and of the need for new regulatory intervention. Since 2006, OFCCP has conducted 25,503 compliance evaluations. Of this number only seven found a basis to allege systemic compensation discrimination and only 58 cases of individual compensation discrimination were found. Including both the systemic and individual cases, OFCCP s own enforcement data shows that compensation discrimination is exceedingly rare amounting to only one-quarter of one percent of audited contractors. Potential Costs May Be Significant To impose on all contractors a significant cost burden to collect and report detailed compensation data in the face of evidence that the probability of finding discrimination is so very small, raises serious doubts that the benefits of new regulation would match the costs. Before proceeding with the contemplated data collection and regulation, OFCCP should conduct further research to estimate the extent of demonstrable compensation discrimination in terms of both the incidence probability and the amount of compensation lost per case. OFCCP should also conduct pilot studies using its proposed data collection instrument to ascertain accurately the labor time and materials cost burden that the data collection will impose on contractors. Such costs may be significant: For example, if the 171,275 service and supply contractors from whom data reports would be mandated required on average just 24 labor hours per year to compile, analyze, review, verify and report compensation data, at a cost of $57.61 per hour for managerial/finance professional labor, 16 the total annual cost burden would be $236.8 million. Benefits are likely to be far lower: If the probability of identifying a case of individual or systemic discrimination were the 0.25 percent probability suggested by OFCCP enforcement data, then the number of cases identified per year among the 171,275 respondents would be 428, and if (as found by OFCCP s own CONSAD study) the unexplained loss of compensation ranges from 7.1 percent to 4.8 percent, then for a typical employee earning $50,000 per year, the compensation loss potentially 17 attributable to discrimination would be less than $3,550 to 16 Bureau of Labor Statistics, Employer Cost of Employee Compensation, 2 nd Quarter 2011. 17 We emphasize that this analysis is a maximum range as it assumes the entire unexplained compensation difference is due to discrimination, a premise that we reject. Nevertheless, we use it here as an upper bound.
$2,410 and the total annual benefit generated by the data collection system would be less than $1.0 million to $1.5 million per year. Even if the incidence and amount of discrimination compensation losses were larger than the rates suggested by OFCCP s own data and studies, the proposed regulation would entail a significant cost-benefit deficit. President Obama s Executive Order 13563 instructs agencies to carefully consider both the costs and the benefits of proposed regulations, to estimate monetary values of costs and benefits, and to choose regulatory or non-regulatory approaches that minimize cost burdens and achieve maximum net benefits. OFCCP has an opportunity to fully estimate both the costs and benefits of any regulatory alternative that it may propose. Costs can readily be estimated by conducting a pilot study in which participating firms record and report their costs for developing or modifying human resource record systems to facilitate compliance with the proposed data compilation requirement and the labor time and costs for compilation, analysis, report preparation, review and verification. A pilot study could test several alternative approaches to data collection requirements so that OFCCP could compare costs of different regulatory. OFCCP can estimate the monetary benefits of a proposed regulation by compiling or estimating a few simple parameters: (1) The annual incidence of compensation discrimination, which can be estimated from OFCCP s own enforcement data, from econometric estimates derived from independent, peer-reviewed studies, or from estimates derived from the alternative data collection tools examined during the pilot tests described above. Incidence of discrimination is simply the estimated probability that an establishment will be indentified and proven as practicing compensation discrimination. (2) The average wage loss per affected worker resulting from compensation discrimination, the upper bound for which is the unexplained residual compensation difference found in econometric studies after controlling for known characteristics of the subject population. Since the residual also includes the effects of unobserved characteristics, the actual effect of discrimination, if any, will be less. Therefore, use of the residual proportion in the benefit computation will provide an upper bound estimate of benefits. This parameter may also be identified from OFCCP enforcement case records. (3) The average number of discriminated-against workers in an establishment identified as discriminating, which may be estimated from OFCCP enforcement data, from independent studies, or from data collected during the pilot testing study described previously. The annual monetary benefit of the proposed rule can be readily estimated as the multiplicative product of the number of reporting establishments and the three parameters described above. OFCCP Should Consult BLS The data collection project envisioned by OFCCP involves a multitude of practical and statistical complexities. The Bureau of Labor Statistics (BLS) has extensive experience dealing with these complexities in the context of its National Compensation Survey program and its Occupational Employment Statistics program. OFCCP should consult with BLS regarding these issues before proceeding to develop and propose any data collection approach. The BLS
experience may be particularly relevant for understanding the potential costs to contractors of data compilation, analysis, verification and reporting. Rather than OFCCP attempt to develop and implement a data collection program on its own, the more efficient approach for both development and implementation may be to rely on the established experience of BLS to develop and implement an approach with funding provided by OFCCP. Conclusion In conclusion, the Chamber believes that the OFCCP should withdraw the ANPRM until such a time as it has clearly articulated the reasons for a new data collection tool. Assuming the OFCCP nevertheless moves forward toward a Notice of Proposed Rulemaking, we have several serious concerns with the approach outlined by the agency, including an erroneous understanding of the pay gap and serious privacy concerns. We also seriously question whether an appropriate tool may be created without imposing very significant burdens on contractors and subcontractors that cannot be justified. Thank you very much for your consideration of these concerns. Please do not hesitate to contact us if the Chamber may be of assistance as you consider these important issues. Sincerely, Randel K. Johnson Senior Vice President Labor, Immigration & Employee Benefits Michael J. Eastman Executive Director Labor Law Policy Attachment