This appendix contains 12 tables from Section 5A revised to present statistics by race. These tables were published with race breakdowns in the 2010 Supplement and prior editions, but the race breakdowns were removed beginning with the 2011 Supplement because of concerns about their statistical reliability. Those concerns are summarized in “Why Researchers Now Rely on Surveys for Race Data on OASDI and SSI Programs: A Comparison of Four Major Surveys.”

The Social Security Administration (SSA) recognizes the importance of identifying and addressing systemic inequities relating to race and ethnicity in the programs it administers. Reporting program-related statistics by race is a critical step toward fulfilling that objective. The tables in this appendix respond to that imperative; however, the concerns about the collection of data by race still apply.

Factors Underlying the Concerns About Race Data

Reporting racial identification to SSA is voluntary because it is not required for program administration. Historically, race data have been collected primarily at “enumeration”—the creation of a Social Security number (SSN)—using the racial categorizations in place at that time.

Before 1980, the only racial identifiers available to SSN applicants were White, Black, and Other. Beginning in 1980, new applicants could choose from five race and ethnicity options—Asian, Asian-American, or Pacific Islander; Hispanic; Black (not Hispanic); North American Indian or Alaskan Native; and White (not Hispanic).

Beginning in 1987, the focus of the enumeration process shifted from applications filed by individuals directly with SSA to registration at birth, completed by hospitals; or upon entry for qualifying immigrants, completed by the Department of Homeland Security (DHS). Hospitals and DHS do not include data on race in the information they send to SSA. As a result, the collection of data by race has dropped sharply for recent birth cohorts. Individuals currently of retirement age are not yet widely affected, as race data are generally available for at least 90 percent of individuals born before 1970 and for at least 80 percent of those born before 1987.

How These Factors Affect the Tables in This Section

These tables use only four race categories: White, Black, other, and unknown.

Because several of the racial categories that SSN applicants can select were not established until relatively recently, the counts of individuals in those categories are artificially low. For this reason, along with data disclosure and statistical reliability standards, all applicants who did not identify as either White or Black are grouped in the “other” category in these tables. Further, although an individual of Hispanic origin may appear in any of the four racial categories, these tables do not account separately for Hispanic origin.

Beneficiaries who identify as multiracial are classified as persons of unknown race. Because the enumeration processes introduced beginning in 1987 have sharply reduced the reporting of race data for recent birth cohorts, the shares attributed to “unknown” races are growing and will continue to grow in coming years.

For statistics on dependent and survivor beneficiaries, race is imputed from the race of the wage earner on whose earnings record the benefit is based.