Beginning in 2006, the Census Bureau began a SIPP redesign with the goal of reducing cost and simultaneously increasing quality and timeliness. A recent National Academy study considers the pros and cons of additional uses of administrative data in the redesigned SIPP and includes a discussion of confidentiality and timeliness issues.

Linked administrative and survey data are of vital importance in developing predictive modeling systems that enable SSA and policymakers to understand the broad impact and distributional effects of current program regulations and reform proposals. To address this need, SSA has developed microsimulation models to analyze the current status of its programs and the effect of proposed program modifications.

“Modeling income in the near term” (MINT) is the most prominent model used in OASDI analysis. MINT is a microsimulation model based on linked household data from SIPP with SSA administrative records on earnings, benefit receipt, and date of death. MINT projects the major pillars of retirement income: Social Security benefits, pension benefits, income from assets, and earnings. In addition, based primarily on regression results obtained from survey and linked administrative data, MINT simulates events such as marital outcomes, age at first benefit receipt, and year of death as well as the characteristics of former, current, and future spouses. Although the name suggests near-term modeling, MINT can be used for projection through the end of the century.

According to 75-year projections by the Office of the Chief Actuary, the OASDI program is not sustainable in the long term under current law. Because of this, MINT is an ever more important source of information for policymakers as they attempt to ensure the future program viability, and MINT has become a more complex model that incorporates a wider variety of socioeconomic data. Arguably, there will be an increased role for statisticians to play in the continued development and maintenance of a more complex and more informative MINT.

Future Change in the Method for Issuance of Social Security Numbers

The Social Security number was created in 1936 to track the earnings histories of U.S. workers, for use in determining Social Security benefit entitlement, and computing benefit levels. Since then, more than 455 million SSNs have been issued, and the use of the SSN has expanded substantially. Although the purpose for assigning a number and issuing a card has not changed, over time the Social Security number has become a primary means of identification in both the public and private sectors. As the use of the Social Security number has grown, so has identity theft and Social Security number misuse.

The nine-digit SSN is composed of an area number (three digits) followed by a group number (two digits) followed by a serial number (four digits). The area number is derived from the mailing address (ZIP code) of the applicant. Group numbers are assigned within each area but not assigned in consecutive order, and serial numbers are assigned within each group. SSA has many years’ worth of potential SSNs available for future assignment. However, because of population shifts, given present rates of assignment and existing geographic allocations, two states currently have fewer than five years’ worth of SSNs available for assignment.

In a July 3, 2007, Federal Register notice, SSA solicited public comment on a proposal to change the way SSNs are assigned. After evaluation of the solicitation responses, the SSA commissioner decided to adopt the proposal to randomly assign SSNs from the remaining pool of available numbers with some exceptions. (For example, the 900 series of area numbers will be excluded since it is used by the IRS for employment identification numbers.) After inception of the new method, the first three digits of the SSN will no longer have any geographic significance. This method will ensure a reliable supply of SSNs for years to come while simultaneously reducing opportunities for identity theft and SSN fraud.

A potential risk in the current system of SSN assignment was highlighted recently in a January 2009 publication by Acquisti and Gross (see also Winkler). These authors demonstrated the partial predictability of SSNs using publicly available information—especially for those who were assigned in low-population states in the SSA enumeration at birth (EAB) program in the last two decades.

Currently, SSA statisticians are working with other SSA staff to mitigate any adverse impacts associated with the decision to randomize the assignment of SSNs. While the modifications required by SSN users in government and industry are not as extensive as those required by the year 2000 (Y2K) problem, they may still be substantial. As one example, some organizations have chosen to assign nine-digit sequences that were not available for SSN assignment (such as the 800 series) as temporary personal identifiers. They will have to modify this practice in light of the fact that these numbers may become eligible for SSN assignment in the new system.