Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Appendix B Study of Employment Earnings for the Equal Employment Opportunity Program: A Possible Role for Administrative Data from Three Tax Systems Nicholas Greenia INTRODUCTION The proposed Paycheck Fairness Act of 2009 (H.R. 12 in the 112th Congress) would have required the U.S. Equal Employment Opportunity Commission (EEOC) to issue regulations mandating the provision of earn- ings data from employers to the EEOC classified by the race, gender, and national origin of their employees. According to the proposed legislation, these pay or earnings data are needed to bolster the related employment and other data already collected through the equal employment opportunity (EEO) reports, particularly the EEO-1 reports, for purposes of enforcing compliance with statutory nondiscrimination employment practices. The new data were argued to be critical in continuing to administer Section 709 of the Civil Rights Act of 1964, as amended. This paper explores the feasibility of using existing data from the administrative records of three tax systems for accomplishing the EEO-1 stated goals for new data collection. It discusses the data collected from and the interrelationships among three tax systems: two administered by federal agencies, the Internal Revenue Service (IRS) and the Social Security Administration (SSA), and one by the state agencies, the unemployment insurance (UI) offices that operate as federal-state partnerships under the Employment and Training Administration (ETA) of the U.S. Department of Labor. It continues by discussing how the interrelationships of the three tax systems benefit data quality, including timeliness, for EEO-1 purposes. It also provides an overview of the sources, including the forms that could 111
112 COLLECTING COMPENSATION DATA FROM EMPLOYERS provide the needed data. The paper concludes by presenting major concerns on confidentiality. These systems hold particular promise for a number of reasons. One is the coverage of the taxes reported and collected: federal income taxes for funding many federal programs that benefit all U.S. residents, taxes that help fund the Social Security and Medicare programs for retirees and other qualified recipients, and unemployment insurance taxes that fund the unemployment benefits of workers who are laid off during difficult economic times, particularly for extended periods such as during the recent deep recession. Another is data quality: the data records tend to have, in general, high levels of compliance because of the importance of these pro- gramsâhighlighted by the penalties for noncomplianceâfor the nationâs safety net and in funding congressionally mandated expenditures. A third is the potential for triangulation of firm and worker levels of reporting by the use of all three systems. Although there are some issues with response rates in each system, such as the tax gap for federal income taxes, partial participation is likely to result in detection by one of the three systems. Although each administrative record data set holds promise for supple- menting EEO-1 data, there are also challenges associated with the use of these administrative data. Like any data system, these three administrative record systems are imperfect in terms of response rates, accuracy, and all levels of granularity, such as multiemployer member reporting in the UI sys- tem. In addition, each also has constraints, including purposes and access. How the EEOC decides to approach the enhancement of its data, including any redesign of its own EEO-1 collection system, may be key to determining not only the most useful plan, but also the most viable for purposes of obtaining earnings data classified by gender, race/ethnicity, and nativity. A BRIEF OVERVIEW OF THE THREE TAX SYSTEMS This section presents an overview of the purpose, coverage, data avail- ability, national importance, and interrelationships of the three systems. These administrative earnings data are captured by multiple administra- tive forms, reported in various components, and available across multiple years from the three tax systems. The classifier variables for gender, race/ ethnicity, nativity, and even age, also exist at the employee record level although they are not universally captured in the databases. All of these data could be linked to a specific employer for an employee, including for multiple employers.
APPENDIX B 113 Purpose The three data systems are used primarily to collect taxes for adminis- tering and funding vital mandated programs: the federal income tax system by IRS, the Social Security and Medicare programs by SSA, and the state UI systems, which are operated by state Employment Security Agencies (ESAs) under a federal-state partnership. Related national statistics are produced from all three sets of data by the statistical offices of SSA and IRS, as well as the Department of Laborâs Bureau of Labor Statistics and the U.S. Cen- sus Bureau. In addition, they are used for policy analysis in a wide range of offices, including the Joint (Senate-House) Committee on Taxation, the Congressional Budget Office, and the Office of Tax Analysis at the U.S. Department of the Treasury, and for analytical research by top academics through the Intergovernmental Personnel Act as well as the Census Bureauâs Research Data Centers. Such robustâand visibleâuses of the data have beneficial consequences for the EEO-1 Program because weaknesses, limitations, and inaccuracies in the data systems tend to become known and corrective measures taken in order to ensure the utility and consistency of the data over time. In addi- tion, because the U.S. statistical system is decentralized, it is more difficult for any one systemâs data anomalies to go unnoticed, given the cross-checks implicitly or explicitly built in across these quasi independent systemsâÂ particularly for financial data, including employment earnings. Coverage Across the three systems, as well as the U.S. Census Bureau, establish- ments and workers needed for EEO-1 purposes would be covered. The data are reported on IRS income tax returns (for individuals and businesses), employment tax returns (for both the Federal Income Contributions Act [FICA] and the Federal Unemployment Tax Act [FUTA]), information re- turns (including for tax-exempt nonprofit organizations), applications for Social Security Numbers (SSNs), and on UI-related forms. Several federal agencies play major roles in either funding or helping process the data and payments for these programs: the Department of Labor (DOL)âparticu- larly its Bureau of Labor Statistics (BLS) and ETAâSSA and IRS. In addi- tion, the states play a major role in administering the State Unemployment Tax Authority (SUTA) Program as well as the employment and training administration system funded in large part by ETA.
114 COLLECTING COMPENSATION DATA FROM EMPLOYERS Data Availability Table B-1 summarizes availability of the EEO-1 items needed by source, including the Census Bureau. National Importance The data are critical for funding many federal programs that benefit all U.S. residents, taxes that help fund the Social Security and Medicare programs for retirees and other qualified recipients, and unemployment insurance taxes that fund the unemployment benefits of workers who are laid off during difficult economic times. Inter-Relationships The three sets of data are interrelated, albeit sometimes in subtle ways. For example, all three systems depend on the SSNs assigned by SSA, the employer identification numbers (EINs) assigned by IRS, the reporting of employment and payroll at both the firm and individual worker level for federal and state purposes, and related information to update them, such as changes in name or address. Similarly, the IRS determination of which workers are employees and which are contractors has an impact on the other systems. The IRS decision is obtained by the filing of a Form SS-8 for a firm or worker seeking to have IRS establish officially the employee or independent contractor status of a particular worker. This transaction then has ramifications for the other employee data collection systems, such as SUTA and FUTA, and could also be used to inform and supplement the EEO-1 reports. EEO-1 Utility Because of the coverage, availability, and interrelationships, the three tax systems hold considerable promise for providing the employee earnings data needed by gender, race/ethnicity, nativity, and even age, by employer. In addition, these systems could be useful also because of the other data they contain, in addition to employee earnings, for supplementing the EEO-1 report data currently collected, including across time both retrospectively and prospectively. STATE UNEMPLOYMENT INSURANCE DATA This section presents a brief summary of why and how UI and Quarterly Census of Employment and Wages (QCEW) data are reported, collected,
TABLE B-1 Summary of Available Population/Universe Data by Agency Earnings at Earnings at Employee Employee Employee Source Employee Level Employer Level Gender Race/Eth. Nativity State UIa YES YES NO NO NO State ESAb NO YES NO NO NO IRS YES YES NO NO YESc SSA YES YES YES YES YES Censusd YES YES YES YES YES aBased on household surveys (samples), BLS does publish employment and compensation data by gender, race/ethnicity, and nativity. These could be useful, at a minimum, as benchmark estimates for purposes of EEO-1 expansion. However, BLS does not presently have access to the statesâ detailed employee earnings records. bESA = Employment Security Agency. cOnly from applications for individual taxpayer identification numbers (ITINs). dEarnings data exist annually in extracts of tax data from IRS. The decennial census captures gender, race, and ethnicity, but, only every 10 years. However, the Census Bureauâs American Community Survey (ACS) captures gender, race/ethnicity, and nativity (native/foreign born). Although it is a sample, the ACS is sufficiently largeâapproximately 2 million respondent households annuallyâthat, at a minimum, its data could be useful for benchmark estimates and perhaps for matching to the EEO-1 report population on a statistically significant basis. 115
116 COLLECTING COMPENSATION DATA FROM EMPLOYERS and shared with the federal sector, and the significance for the EEO-1 Program. Purpose In addition to complying with FUTA, employers must also comply with SUTA by withholding and depositing tax or insurance payments from each employeeâs wages with the state unemployment offices. Although federal unemployment taxes serve several purposes, state unemployment taxes are used only to fund unemployment benefits in a particular state or territory (including the District of Columbia, Puerto Rico, and the Virgin Islands). Coverage Tax rates and coverage vary by state, as do the content and format of the records a particular state collects. In general, workers not covered by this system include federal employees, independent contractors, the self- employed, and some agricultural workers. Content A state collects the employment and compensation data in two parts. The first part is detailed earnings data1 collected as part of the UI system. The state UI agency collects reports from each employer that include the SSN, name, and quarterly compensation for each individual employee (as well as the employer name and EIN).2 This collection of detailed employee earnings, often called UI wage records, provides the most frequent and granular information about employee earnings across the three tax systems. For the second part, the state ESA collects aggregate monthly employ- ment (for the pay period containing the 12th of the month3) for each quar- ter and the aggregate quarterly employee compensation from each employer in the state covered by state UI laws and federal workers covered by the Un- employment Compensation for Federal Employees (UCFE) Program.4 This program, administered by the BLS, also includes the collection of monthly 1â ee, S for example, http://detr.state.nv.us/uicont/forms/NUCS-4072.PDF [July 2012]. 2â he T coverage varies by state. For a complete review, see Stevens (2002), available at: http:// lehd.did.census.gov/led/library/techpapers/tp-2007-04.pdf [July 2012]. 3â he 12th of the month is the same date used for reporting of employment on the IRS T quarterly employment FICA tax returns (Form 941 series) that is, March 12, June 12, etc. 4â his quarterly reporting of aggregate compensation provides more commonality with the T IRS Form 941 series, which also reports quarterly aggregate employee compensation: see, for example, http://www.bls.gov/cew/forms/mwr_nm.pdf [July 2012], also see http://www.bls.gov/ cew/cewover.htm [July 2012].
APPENDIX B 117 employment data and provides the most frequent aggregate employment data across the three tax systems. The second part data collection is partly funded by BLS, and after a state edits the data, it transmits electronic summaries to BLS for its sta- tistical needs. Although data are also requested for multiple worksite or multi-establishment employers, there is no disincentive for an employer that does not comply with the multisite request as long as total employment is reported accurately and the appropriate amount of UI taxes is paid to the states. EEO-1 Utility For purposes of expanding the EEO-1 Program, the UI data system provides the earnings data needed and at the employee level, but it also presents three problems. First, because of the lack of a disincentive for nonreporting of multisite employer detail, there may be a disconnect in matching to multi-establishment employer data at the worksite levelâbut not the enterprise levelâfrom the EEO-1 reports. It would be up to the EEOC to determine how big a problem this represents for its enforcement needs. Second, gender, race/ethnicity, and nativity data are not collected for either of the two parts described above. However, if the detailed employee earnings data could be matched to SSA Numident (Numerical Identifica- tion System) data, this problem could be reduced if not resolved. Third, and perhaps most daunting, in order to obtain either of the two data parts provided to the statesâespecially the detailed employee earningsâit would be necessary to obtain separate agreements with each state as was done so laboriously for the Longitudinal Employer-Household Dynamics (LEHD) Program at the Census Bureau starting in the 1990s.5 INTERNAL REVENUE SERVICE DATA This section presents a summary of several tax and information forms, especially Form W-2, Form 941, and Form 940, and why they might be of interest to expand the EEO-1 reports on employment and earnings data. In addition, it discusses the close relationship IRS and SSA have in terms of the first two forms, particularly for validating and reconciling amounts withheld for income, Social Security, and Medicare taxes. 5â he T LEHD program is briefly described in Chapter 2.
118 COLLECTING COMPENSATION DATA FROM EMPLOYERS Purpose In 1976,6 the current simplified Combined Annual Wage Reporting (CAWR) Program was established by law to ensure that employers pay and report the correct amount of tax, including federal income tax withholding and that they file timely all necessary forms with SSA. That same year, Form W-2 (Wage and Tax Statement) was redesigned to include Social Security information, and Form W-3 (Transmittal of Income and Tax Statements), was amended to include cumulative totals of each money field appearing on the associated Form W-2s. Content Detailed annual employee compensation, quarterly and annual ag- gregate employee compensation, and number of employees are provided at both the employee and employer level and are linkable by the SSN/ EIN crosswalk also provided. In addition, other tax forms provide various components of aggregate and even detailed employee compensation, such as compensation to corporate officers. Finally, EIN and ITIN assignment and other transactions enable the tracking of new business births, foreign born workers without SSNs, and even the employee or contractor status of a worker. For purposes of expanding EEO-1 reports, three forms in particular figure prominently in the CAWR process: Form W-2, Form 941, and Form 940.7 Form 940, Employerâs Annual Federal Unemployment Tax Act (FUTA) Tax Return8 Purpose Form 940 is required to be filed annually by an employer for purposes of reporting and paying the federal unemployment taxes required by FUTA. These taxes are used to fund state workforce agencies, pay half the cost of extended unemployment benefits in severe economic downturns, and also for loans to states to help them pay unemployment benefits, including extended unemployment benefits. 6â he Tax Reform Act of 1976 (TRA76) also established the present confidentiality statute T in the tax code, namely, Section 6103. 7â chedule H, filed with Form 1040 to report household employees, is omitted from this S discussion. 8â ee http://www.irs.gov/pub/irs-pdf/f940.pdf [July 2012]. S
APPENDIX B 119 Coverage Filing is requiredâat the aggregate employment levelâfor each nonag- ricultural employee earning at least $1,500 in any quarter of the year or for each employee who was employed for part/all of a day in any 20 different weeks of the year.9 EEO-1 Utility Although Form 940 does report annual total compensation, it does not report the number of employees. However, for purposes of this analysis, the compensation information may be useful for benchmarking compensation data reported on other federal tax forms, say, Form W-2, and Form 941, as well as the UI data. Form W-2, Wage and Tax Statement10 Purpose Form W-2 is required to be filed by both employees, with their in- dividual tax returns (Form 1040), and employers, transmitted under the summary Form W-3. The formâs major tax purpose is threefold: reporting of federal income tax, Social Security tax, and Medicare tax withheld from employeesâ compensation. The W-2 is also required to be filed if these taxes were not withheld but should have been. Coverage Withholding of federal income tax is not required for an employee who had no federal tax liability in the previous year and is expected to have none in the current year. However, because Social Security and Medicare taxes must be withheld, a Form W-2 must be filed for such an employee. Thus, this is an extremely potent building block for employment and wage dataâat the employee level, but cross-referenced to the employer level by the cross-walk of SSN/EINâeven for low-wage employees. In addition, because a different W-2 must be filed by each employer of an employee, these data can provide multiple employer information for an employee with multiple jobs. 9â or 2009 and 2010, agricultural employers were required to file if they paid cash wages F of $20,000 or more to farm workers during any calendar quarter or if they employed 10 or more farm workers during some part of the day (whether or not at the same time) during any 20 or more different weeks in either year. 10â ee http://www.irs.gov/pub/irs-pdf/fw2.pdf [July 2012]. S
120 COLLECTING COMPENSATION DATA FROM EMPLOYERS EEO-1 Utility The industry codes available at SSA (at the full 6-digit level of the North American Industry Classification System) can provide a further source of rich classifier information on employersâ business activities. Earnings de- tail is also rich: wages and salaries, deferred compensation (part of total compensation, even if not taxable currently), and certain fringe benefits are reported, in addition to capped Social Security earnings and uncapped Medicare earnings. Together, the W-2 earnings variables provide a unique and comprehensive window on earnings data at the employee level. Form 941, Employerâs Quarterly Tax Return11 Purpose Form 941 is required to be filed quarterly by an employer in order to report and pay federal income tax withheld for employees, and both the employerâs and employeesâ share of Social Security and Medicare Taxes. Similarly, Form 943, Employerâs Annual Federal Tax Return for Agricul- tural Employees,12 is required to be filed annually for the same reasons, but for agricultural employees. Coverage In general, coverage of FICA employees by the Form 941 series is very similar to that of FUTA employees by the Form 940. EEO-1 Utility Both the Form 941 series and Form 943 contain a number of use- ful fields, especially the total number of employees and their total c Âompensationâquarterly for the Form 941, annually for Form 943. In addition, the forms report taxable Social Security wages (which are capped at the SSA ceiling) and taxable Medicare wages (which are not capped and, thus, equivalent to total wages). Data Quality: IRS and SSA Reconciliation IRS and SSA use a reconciliation process involving the filings of both Form W-2 and Form 941 in order to determine discrepancies and possible 11â ee S http://www.irs.gov/pub/irs-pdf/f941.pdf [July 2012]. 12â ee S http://www.irs.gov/pub/irs-pdf/f943.pdf [July 2012].
APPENDIX B 121 tax delinquencies. Specifically, they compare taxable SSA wages, taxable SSA tips, taxable Medicare wages, and federal income tax withheld. Dis- crepancies result in the direct contact of employers, and consequences for noncomplianceâand even nonresponseâcan be serious. For example, in addition to monetary penalties that may result, so-called bad boy employers have been required to file Form 941 on a monthly, instead of a quarterly, basis. IRS uses a similar cross-check system involving more tax forms, such as the Form 104013 series of individual tax returns, to ensure that an indi- vidualâs total reported income jibes with other reports of the income source; for example, the Form W-2 for earnings and other compensation and Form 1099R for income such as interest, dividends, and pension distributions. The consequences of being noncompliant with the federal income tax system are well known and potentially include not only prison, monetary penalties and interest, but also damage to oneâs credit ratings for both indi- viduals and firms. For a firm, such damage can extend to its reputation in the business community, for example, for partnering and other collabora- tive efforts, and adversely affect attempts to raise capital publicly, say, with an initial public offering, and privately. Because of the adverse consequences of tax noncompliance, firms are generally highly incentivized to comply and provide accurate and timely information to both IRS and SSA. If they are not, IRS enlists an array of tools for enforcing compliance that include a discriminant function (DIF) system scoring14 of individual and some business tax returns and numer- ous auditors and agents to ensure that tax laws are obeyed and corrective measures taken when they are not. Data Quality: IRS and State UI Reconciliation Although a similar relationship exists between IRS and the state work- force agencies15 for purposes of ensuring the timely and accurate payment of both state and federal unemployment taxes, Form 940 earnings dataâ annual employment compensation by employerâmay be less useful for 13â ee S http://www.irs.gov/pub/irs-pdf/f1040.pdf [July 2012]. 14â nder U this system, IRS computer programs assign each return a numeric DIF score rating the potential for necessary changes to the return, based on past IRS experience with similar returns. The unreported income DIF score is used to rate the return for the potential of unreported income. IRS staff screen the highest-scoring returns, selecting some for audit and identifying the items on these returns that are most likely to need review. 15â nder Section 6103 of the tax code (and reciprocating state and municipal laws), IRS, U state, and even municipal tax authorities have long shared data for mutual benefit involving tax administration. For states, such sharing has included data to administer both income taxes and employment or payroll taxes.
122 COLLECTING COMPENSATION DATA FROM EMPLOYERS purposes of expanding EEO-1 reports than the more detailed information on the Form W-2 and Form 941. However, the information sharing between IRS and the state workforce agencies also helps ensure the accuracy of the data reported to the states at both the firm and employee level for purposes of both federal and state unemployment taxes. The importance of the in- teragency relationship for ensuring that these taxes are paid correctly and timely is a major reason these data from all three tax systems may hold such promise for expanding the earnings data on the EEO-1 reports. Form SS-4, Application for Employer Identification Number16 In addition to starting the process for assigning an EIN for an entity (usually, but not always, a business), the Form SS-4 establishes an employ- erâs account on the IRS Business Master File (similar to the business regis- ters at BLS and the Census Bureau, but for tax administration), including filing requirements for income tax returns (Form 1120 series, Form 1065 series, Form 990 series) and employment tax returns (Form 940 and Form 941 series). It also provides the SSN-EIN crosswalk for a sole proprietor- ship converting from nonemployer to employer status, important informa- tion in order to link the Schedule C posting to the Individual Master File on SSN with the accompanying Form 1040, to the sole proprietorshipâs employment tax returns posting on EIN to the Business Master File. IRS also provides SS-4 population data to SSA (and the Census Bureau), which uses the detailed alpha information on business activity to assign full 6-digit industry codes,17 which should be useful industry classification for EEO-1 reports. In summary, this short form initiates actions in several systemsâ both statistical and administrativeâwhich begin the cross-tracking of many events for a central use of the form, the identification of new businesses. Form W-4, Employeeâs Withholding Allowance Certificate18 Form W-4 identifies a new employeeâs withholding status for purposes of the required Form W-2 that is later filed with an employeeâs Form 1040 individual income tax return. Although the W-4 is not required to be filed with IRS, it is required to be filed by federal and state agencies for 16â ee S http://www.irs.gov/pub/irs-pdf/fss4.pdf [July 2012]. 17â he T 6-digit North American Industry Classification System (NAICS) codes replaced the 4-digit Standard Industrial Classification (SIC) codes in 1997, but continuity mappings (from SIC to NAICS) exist at many federal agencies using these codes. IRS uses NAICS-based codes for its tax returns, but only what can fit on the allotted one page of the form instructions. These vary by business entity according to the business activity distribution; for example, Form 1065 codes differ from those for Form 1120. 18â ee http://www.irs.gov/pub/irs-pdf/fw4.pdf [July 2012]. S
APPENDIX B 123 employers, as part of the National Directory of New Hires at the U.S. DepartÂ ent of Health and Human Services (HHS) (see related discussion m below under Confidentiality). One use of this form, in addition to its poten- tial for identifying increases in national employment on practically a real- time basis, is that it individually identifies new employees, something that may be of interest for EEO-1 reports to track employment by employers. Form W-7, Application for IRS Individual Taxpayer Identification Number (ITIN)19 Form W-7 is filed for foreign workers, regardless of immigration status,20 in place of an application for SSN. The ITIN is important not only because of the foreign nativity information it contains, but also because it helps complete identification of the worker universe information, supple- menting and complementing the SSNs reported for more permanent status workers. Thus, it indirectly helps provide detailed worker information on the forms filed with the states as well as a more complete picture of the employee/employer relationships revealed by Form W-2 filings. In addition, most immigrantsâeven in illegal statusâhave incentives to have an ITIN so that they and their employers can file required tax returns. The worker may have the additional incentive of obtaining a tax refund later. Form SS-8, Determination of Worker Status for Purposes of Federal Employment Taxes and Income Tax Withholding21 Although Form SS-8 is not required, it may be filed by either a worker or firm to determine whether a worker should be considered an employee or independent contractor. The resulting determination may have ramifica- tions for not only IRS forms, such as the W-2 and employment tax returns, but also for UI and related record filings with the states for SUTA and their employment training administration programs. One purpose of a related return, Form 1099 Miscellaneous,22 is to report payments to contrac- tor workers. Thus, this information, in conjunction with compensation information reported for employees, can help provide a complete worker compensation picture by employer. In addition to helping capture information for contractors required to complete EEO-1 reports, such information might also be helpful for EEOC 19â ee S http://www.irs.gov/pub/irs-pdf/fw7.pdf [July 2012]. 20â rom F a general policy perspective, IRS has not cared about an immigrantâs legal or illegal status, only that the employee and employer file required returns and withhold and pay all required taxes. 21â ee http://www.irs.gov/pub/irs-pdf/fss8.pdf [July 2012]. S 22â ee http://www.irs.gov/pub/irs-pdf/f1099msc.pdf [July 2012]. S
124 COLLECTING COMPENSATION DATA FROM EMPLOYERS in determining which employers might be avoiding compliance with EEOC requirements and which are evading compliance. To paraphrase IRS com- pliance parlance, avoidance would be considered legal, but not evasion. Additional Income Tax Returns Finally, several returns report earnings at both the individual and firm levels. For the former, Form 1040 and the related Schedule C (for sole pro- prietorships) report individual and self-employment earnings. Moreover, when the Schedule Câs filer is also an employer, the Schedule C will contain compensation information for the firmâs workers; for example, Cost of La- bor. At the firm level, aggregate employment compensationâsalaries and wages, cost of laborâcan be found on the Form 1120 series,23 in addition to an item of possible interest for expanding EEO-1 reports, namely, com- pensation to officers of the corporation. Aggregate employment compensa- tion is also reported on pass-through forms, such as the Form 1065 series24 for partnerships and Form 1120-S25 for subchapter S investors. Income and taxes are reported for the individual partner or investor on Schedule K-126 and the respective Form 1040 (although partners and investors may be businesses, not individuals). An additional sector of employers may also be of interest for the EEOC, namely, nonprofit or tax-exempt organizations that have to file Form 990, Return of Organization Exempt from Income Tax,27 (or the related Form 990-T,28 Exempt Organization Business Income Tax Return). Both forms, especially the former, report a number of earnings items of potential inter- est, including aggregate cost of labor and compensation to officers, as well as detailed individual compensation to officers, directors, trustees, and highly compensated employees. Limitations of IRS Data Although IRS data include a wealth of earnings data by individual em- ployee and employer, they include establishment data only when an estab- lishment is also an enterprise (and has an EIN). Another limitation is that they contain no data by gender (except, sporadically, for the Statistics of 23â ee S http://www.irs.gov/pub/irs-pdf/f1120.pdf [July 2012]. 24â ee S http://www.irs.gov/pub/irs-pdf/f1065.pdf [July 2012]. 25â ee http://www.irs.gov/pub/irs-pdf/f1120s.pdf [July 2012]. S 26â ee http://www.irs.gov/pub/irs-pdf/f1065sk1.pdf [July 2012] and http://www.irs.gov/pub/ S irs-pdf/f1120ssk.pdf [July 2012]. 27â ee http://www.irs.gov/pub/irs-pdf/f990.pdf [July 2012]. S 28â ee http://www.irs.gov/pub/irs-pdf/f990t.pdf [July 2012]. S
APPENDIX B 125 Income [SOI] individual Form 1040 tax sample), race/ethnicity, or nativity Â (except for ITIN applications). SOCIAL SECURITY DATA Although a massive amount of data exist at SSA, the data of most interest for expanded EEO-1 reports are captured from the application for an SSN and the linkable federal tax data shared by IRS. Thus, only these data are discussed below. Purpose The data at SSA are used for administering the Social Security and Medicare programs mandated by law. Nevertheless, a related purpose is the statistical analysis necessary for such administration, conducted by not only the Office of the Actuary, but also the Office of Research, Evaluation, and Statistics (ORES). The latter would most likely be the office with which the EEOC would need to discuss any future work involving EEO-1 report data. Content Form SS-5,29 Application for Social Security Number, is administered by SSA and captures gender, race/ethnicity, and nativityâoften shortly after birth for most U.S. citizens. In addition, it captures citizenship status, which might be used as a proxy for or to supplement nativity information. Although the Form SS-5 data are self-reported, SSA uses supporting docu- mentation for verificationâparticularly for changes, such as a marriage license (name), passport (citizenship), and birth certificate (place of birth). The Form SS-5 data, including updates, are maintained on SSAs Numident file. Because many people, such as nonretirees, have more incentives to up- date their tax information changes, say, name and address due to marriage or divorce, the tax information at IRS may be updated before the Numi- dent data. However, because of the Form W-2/941 reconciliation process partnered by SSA and IRS on withholding for income, Social Security, and Medicare taxes, SSA has these data as an additional source for updating changes to the Numident, and can also query the individuals and firms in case of doubt. 29â ee S http://www.ssa.gov/online/ss-5.pdf [July 2012].
126 COLLECTING COMPENSATION DATA FROM EMPLOYERS Quality Because of the supporting documentation, the SSA-IRS relationship (as well as the SSA-Census Bureau relationship), and penalties for non- compliance, filers should have incentives to provide accurate and timely data, although some limitations may be inherent. For example, although nativity data classified by country might be considered relatively reliable, researchers have noted that some of the âforeign bornâ may be, in fact, the progeny of American citizens, say, for military and other Americans stationed overseas, where birth occurs. In conjunction with the citizenship status, however, the data are probably useful for indicating native vs. for- eign born statusâthe same nativity classifications published by BLS for its household surveys. Like IRS and UI administrative data, SSA data are imperfect, but the interrelationship of these seemingly disparate data sets is usually a strength. For example, incorrect decedent data30 might be passed on from SSA to the IRS (for individuals on its Individual Master File), but ultimately, a tax return, say, the Form W-2, could help rectify the mistake, even if it had not already been corrected through other means. Utility for EEO-1 Together with the detailed earnings data obtained from the IRS Form W-2 and Form 941, SSA classification data of gender, race/ethnicity, and nativity should be considered a potent source of information at the indi- vidual employee level, with the essential crosswalk of SSN (or ITIN) and EIN enabling linkage to the respective employers. U.S. CENSUS BUREAU Although the U.S. Census Bureau is a federal statistical office and does not collect administrative data directly or participate in the administra- tion of the related programs, it may play a unique role in the utilization of administrative data for purposes of expanded EEO-1 reports for several reasons. First, the Census Bureau is an established and long-time user of ad- ministrative data for statistical purposes, and it has developed institutional expertise in the integration of these data with its own statistical survey and census data. For example, the bureau has long considered IRS tax data to be the âlifebloodâ of its business register and related business programs, and 30â or example, see âSocial Security Wrongly Declares 14,000 People Dead Each Year,â F CNN, August 17, 2011, available at: http://money.cnn.com/2011/08/17/pf/social_security_ deaths_mistakes/index.htm [July 2012].
APPENDIX B 127 uses these tax data both for sampling frame purposes and to supplement and improve the quality of its own data programs. Moreover, although IRS data are not necessarily reported at the establishment level, the Census Bureau integrates them in the business data system it maintains at the es- tablishment level. Thus, the bureau could be an important resource in the establishment as unit of measurement. Second, the Census Bureau has partnered with other statistical agencies, such as ORES and BLS, on mutually beneficial programs. For example, SSA and the Census Bureau collaborate on work matching individual record data from the Current Population Survey (CPS) and the Survey of Income and Program Participation (SIPP) to Numident data and Form W-2 tax data at SSA. In addition, the Census Bureau and BLS have collaborated on work involving the record to record comparison of their respective business registers, including IRS tax data on the bureauâs register. Third, the Census Bureau is authorized to access earnings data from federal tax data that most other federal agencies, including SSA, are not authorized to access, for example, from income tax returns for individuals and employers. Fourth, the Census Bureau has a unique vehicle, the special sworn status process, which enables outsidersâfrom other federal agencies and even outside the federal sectorâto access non-anonymized or confidential data. Such access might be granted to experts to help ensure that the work being done by the bureau for an outside sponsor is conducted according to the sponsorâs ultimate needs. Finally, the Census Bureau has not only considerable resources as an advantage over most other statistical agenciesâconferring significant economies of scaleâbut also a wealth of data in its own Title 13 programs, such as individual gender and race/ethnicity data from the decennial census. For purposes of the EEO-1 reports, these factors suggest that the Cen- sus Bureau might play an important role in providing data access to EEOC staff (or its contractors, with special sworn status), under mutually suitable terms, and also in partnering with other agencies, such as SSA, in order to use data that the Census Bureau may not possess on a regular basis for the population of employees. CONFIDENTIALITY AND DATA ACCESS IRS Data In general, individually identifiable tax data at IRS, frequently referred to as federal tax information, are considered confidential and nondisclos- able unless such disclosure or access is authorized by statute, meaning that both bodies of Congress have passed a bill signed into law by the president.
128 COLLECTING COMPENSATION DATA FROM EMPLOYERS De-identification (removing identifier information such as name, SSN, ad- dress, etc.) of the data is considered a necessary but insufficient measure for anonymizing tax data, and statute (e.g., sections 6108(c) and 6103(j)(4) of Title 26 of the U.S. Code) requires that publicly released data be protected from both direct and indirect re-identification (e.g., in conjunction with other publicly released data by other agencies). Also, in general, tax data must be used for purposes of tax administration, which includes both a statistical and research component. Thus, the IRS SOI and the IRS Research Office are statutorily authorized users of identifiable tax data within the tax agency. Non-tax administration uses are discouraged, and IRS policyâ and sometimes statute, e.g., for the Census Bureauâhas been that even authorized tax data access should be to the minimal extent necessary to accomplish an authorized purpose. Over the years the overriding concern manifested by IRS has been the protection of the voluntary tax system, and confidentiality protection is seen as a cornerstone of such protection. Thus, any usage or proposed usage of tax data, even by authorized users, that is perceived by the public or IRS (or third-party scrutinyâCongress, the media, privacy advocates, etc.) to threaten taxpayer confidentiality is unlikely to be entertained and likely to be rejected. However, non-tax administration accesses exist statutorily and certain exceptions have been madeârarelyâto statutory access. Thus, HHS is authorized to access limited federal tax information for purposes of its child support program. The Census Bureau and a few other agencies (such as the Bureau of Economic Analysis) are statutorily authorized to access selected tax data for another non-tax administration purpose, namely, statistical purposes. Sometimes, a statute requires that the Treasury Depart- mentâs regulations specify exactly which tax items may be accessed; e.g., for Census, but not for the Congressional Budget Office and SSA. These latter uses must be argued on a case-by-case basis, a less arduous process than changing statute and the regulations, but not an easy one, as IRS must be convinced31 with a compelling business case that does not compromise even the perception of confidentiality protection and the voluntary compliance tax system. BLS, which is not statutorily authorized to access federal tax informa- tion, has accessed identifiable tax data, both through special sworn status at Census/IRS-approved facilities, say, for the Census-BLS business register comparison project mentioned above, but also as BLS employees. The lat- ter access involves only âminorâ tax data access, such as EINs, for a long- standing procedure enabling the two agencies to compare industry codes in order to better harmonize some national statistics by that classification. 31â urthermore, F Treasuryâs Assistant Secretary for Tax Policy must officially approve an amendment to the regulations.
APPENDIX B 129 Another way to access tax data is to obtain the taxpayersâ waivers of their confidentiality rights for specific tax data through an established, formal IRS documentation process. This method has been employed suc- cessfully for a limited amount of tax data for the Health and Retirement Study (HRS). However, as one might imagine, obtaining these waivers for a sample size similar to that of HRS is probably more plausible than for the EEO-1 universe. SSA Data As indicated above, SSAâs classification dataâgender, race/ethnicity, nativityâon the Numident have been linked to tax data as well as data from important surveys, such as the CPS and SIPP, in a long-standing col- laborative arrangement with the Census Bureau. However, this work has been for exclusively statistical purposes, a factor that EEOC would need to consider in formulating a data request. Nevertheless, statistical needs might be joined to the administrative needs of both SSA and the EEO-1 Program, perhaps under a mutually beneficial agreement involving data from the National Directory of New Hires (NDNH) database at HHS (maintained at the SSA National Com- puting Center). SSAâs National Computing Center stores and maintains the NDNH, which also includes other data that may be of interest to the EEO-1 Program, such as from Form W-4, Employeeâs Withholding Allowance Certificate,32 completed for most employees on the first day of employment. The W-4 can thus be used to identify not only new employees, but also new employment in the aggregate, and on a virtually real-time basis. It may be worth considering whether such an arrangement, perhaps most viable as a statistical research program under the child support statute underpinning the NDNH, might make sense, especially if HHS might derive some benefit as well. One proposal might involve a detailed analysis of EEO-1 employees who had experienced job discrimination and any relationship of such dis- crimination to the status of such employees as absent parents. This is only a simple example for illustration; surely, more sophisticated research propos- als could be devised that might meet some of the needs of all three agencies. State Data Although the collection of UI data is largely funded by ETA, the quar- terly detailed employment and compensation data are retained by states and shared outside the states only rarely. Two important exceptions include the NDNH and the Census Bureauâs LEHD Program. 32â ee S http://www.irs.gov/pub/irs-pdf/fw4.pdf [July 2012].
130 COLLECTING COMPENSATION DATA FROM EMPLOYERS HHSâs Federal Office of Child Support Enforcement (OCSE) operates the NDNH, a database established pursuant to the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. The primary purpose of the NDNH is to assist state child support agencies in locating parents and enforcing child support orders33; however, Congress has authorized a limited number of other state and federal agencies to receive information from the NDNH for authorized purposes, including for statistical research related to the child support mandate. Beginning in the 1990s, the Census Bureau began negotiating individual agreements with the states that resulted in their providing these detailed data to Census for the LEHD Program. Any use of these data by the EEOC should probably include some benefits for the statesâas occurred with the Census Bureau collaboration. The above examples are only illustrative of what is exceptional: in general, the data are not shared, even with a statistical agency such as the Bureau of Labor Statistics. 33â his T program is sometimes referred to as the âdeadbeat dadsâ program, although either parent may be in its scope.