Americans are living longer, yet they are leaving the labor force at younger and younger ages. Possible improvements in health status would suggest the reverse. Firm pension plans, however, tend to encourage early retirement. The central goal of this continuing research is to estimate the effect of private pension plan provisions on retirement. A particular component of this general goal is to estimate the effect of special early retirement incentives, often called """"""""window"""""""" plans. A related goal is to estimate the effect on Social Security net costs of the early retirement induced by private pension plan provisions. To accomplish this long-term objective, the research has several specific aims: (1) To document that pension plan provisions that provide an incentive to retire early do indeed induce departure from the firm. (2) To complete the development and testing of a statistical model that is tailored to the analysis of retirement behavior. The model is forward looking in that it accounts for large jumps or drops in pension wealth that will occur if work continues. At the same time, it accounts for changes in the circumstances and alternatives of individuals as they age. (3) To use the model to estimate the relationship between pension plan provisions and retirement in several different firms. (4) To test the model by predicting the retirement effect of """"""""natural experiments"""""""" that are occasioned by special and unforeseen early retirement incentive programs. (5) To predict the effect on retirement of potential changes in the pension plan provisions. An example is the effect of increasing the age of early retirement in a firm from 55 to 60. Retirement in large numbers might then begin at 60 instead of 55, reducing substantially retirement rates between 55 and 60. (6) To estimate the relationship between retirement induced by corporate pension plan provisions and the net cost to the Social Security system. This involves evaluation of the interaction between Social Security provisions and private pension plan provisions, to estimate the magnitude of the effects of each on retirement. (7) To determine how private pension and Social Security provisions could be made consistent with respect to the retirement incentives that they present. The analysis will be based on personnel data collected from several individual firms. The empirical specification is a continuous time model in which the option value of not retiring plays a key role. The model captures the advantages of both non-linear budget constraint and continuous time hazard models.
