The Effect of Social Security on Retirement

This study examines the impact of social security on the retirement of married men aged 60-70 years. The empirical results are based on a rich file of data from the Social Security Administration (1973 CPS-IRS-SSA Exact Match File). The data permit precise calculation of social security wealth (the actuarial present value of benefits that a person would receive by retiring) denoted SSW. This variable measures social security's effect on retirement. The estimated effects are significant and considerable. When SSW in-creases from $35,000 to $55,000 the probability of retirement rises by .15 for 62-64 year olds relative to a .41 retirement rate. For 65-70 year olds this increase is .22 relative to .78. For 60-61 year olds who are entitled to SSW but not old enough to receive benefits the estimated effect was small and insignificant. This supports the conclusion that the observed effect on men eligible for benefits is a causal relationship. The traditional method of comparing market and reservation wages for analyzing the decision to work provides the basic econometric model. SSW is added to construct a retirement model. A two-step probit analysis is developed to identify structural parameters in the retirement model.