Download entire publication - Social Security

The 1983 Amendments to the Social Security Actincluded, among other measures, the change in theFRA starting with the cohort attaining age 62 in 2000(those born in 1938), for whom the FRA was set at 65and 2 months. The FRA has increased by 2 monthsfor every cohort since then until it reached 66 forthose who attained age 62 in 2005, and it will stay atthat level for a decade. The FRA will increase againby 2 months for the cohort born in 1955 (who reachage 62 in 2017, and it will continue to increase by2-month increments for successive birth cohorts untilit reaches 67 for the 1960 cohort.The changes in the FRA and the DRC were clearlyeasy to anticipate by those nearing retirement age, andit is natural to expect comparatively less pronouncedchanges in behavior resulting from their phasedimplementation. 9 More unexpected was the repealof the earnings test for individuals above the FRA,which withholds benefits for individuals earning abovethe exempt amounts. The legislation was passed inthe spring of 2000, approximately a year after it wasmade a policy objective by President Clinton in early1999, and it affected earnings obtained after January1, 2000. The literature analyzing the effects ofthe earnings test is also quite large and has focusedprimarily on understanding whether people respond tothe exempt amount. 10 Given the data analyzed, we arein a position to infer possible changes in behavior thatare due to the repeal of the earnings test, which wouldotherwise be hard to characterize using householdleveldata.DataWe use the 2004 publicly available release of the Old-Age, Survivors, and Disability Insurance (OASDI)public-use microdata files to analyze the trends inclaiming behavior and level of benefits received in the1994–2004 period. 11 The 2004 Benefits and Earningspublic-use file is a 1 percent random sample of OASDIbeneficiaries who were on the Social Security recordsin December 2004. It contains 473,366 records as ofDecember 2004 and includes information in 16 fieldson OASDI beneficiaries’ characteristics, mainly aboutbenefit entitlements. This more detailed informationallows us to focus only on retired workers whoclaimed benefits on their own earnings history, andbecause it is individual-level data, we can computestandard deviations and therefore statistically comparebenefit levels across ages and years.This microdata has, however, two weaknesses.First, we have not been able to separate disabilityconversions from new entitlements for those personsclaiming benefits at age 65 or the FRA, if higher―adistinction that the public-use file does not allow us tomake. 12 What we have done to overcome this problemis to assume a proportion of Social Security claimantsfrom age-65 samples each year as disability conversions.The proportions used are calculated accordingto the Annual Statistical Supplement to the SocialSecurity Bulletin―referred to elsewhere in this articleas the Supplement (SSA 1995–2008). Second, we arerestricting attention to individuals in the MBR as ofDecember of 2004. The latter method most likelyresults in a selection bias when looking at historicaldata on individuals who claimed benefits in the decadebefore 2004. The reason is that some individuals whoclaimed in the 1990s, or even more recently, might notbe in the sample if they have died in the time sincetheir application for benefits. 13Claiming Behavior and RetirementBenefitsUsing public-use data extracts from the MBR, wenotice the well-known retirement peaks at ages 62 and65, both thinking in terms of proportions of claimersin a given calendar year and as proportions by cohortover different years. 14 It is interesting to highlight,however, that the relative sizes of these peaks, asproportions by calendar year, have changed considerablyfrom previous decades―where the largest peakoccurred at age 65 in the 1970s―or were roughly ofsimilar size during part of the 1980s. In our studyperiod, from 1994 through 2004, we see that the proportionof individuals claiming benefits at age 62 hasremained quite stable in most years (in the 48–52 percentrange) with the proportion of individuals claimingbenefits before the FRA at almost 64 percent by 2004.This is a key development of the past two decades andone that has puzzled economists considerably. 15 Onthe other hand, the proportion of individuals claimingbenefits at age 65 has remained at about 20 percent.Until recently a number of researchers have tried toexplain this with arguments regarding individual preferences(Coile, Gruber, and Jousten 2002; Gustmanand Steinmeier 2002), suggesting that there is a proportionof individuals who seem to be rather myopicand do not quite behave as forward-looking optimizers.Other authors have recently shown that once the80 Social Security Bulletin • Vol. 69 • No. 3 • 2009