Collaborative Research: Retirement Decisions After Job Loss

点击翻译按钮获取中文摘要

AbstractSBR-9905275This project investigates the effects of late career job loss on earnings, wealth and employment, then uses these displacement-induced changes in the relative returns to work and retirement to cast new light on the nature of the retirement decision. While previous work on job loss outcomes has often ignored older workers and the burgeoning literature on retirement has similarly ignored involuntary job separations, this project will confront a number of important yet unanswered questions in both the literatures on displaced workers and on retirement decision-making. Proceeding in two stages, the first stage ofthe project involves reduced-form investigations ofthe effects of job loss on older workers' earnings, employment, and wealth, including pension wealth. In the second stage, the estimated effects of job loss are incorporated into a structural model of the retirement decision.In the first stage, data from the Health and Retirement Survey (HRS) are used to estimate the effects of job loss on earnings, financial wealth, pensions, employment and retirement expectations of workers aged 50 and over. While there is a large literature on the effects of displacement, much of the recent research has focused solely on "prime-age" workers. In addition, virtually no research exists on the effects of displacement on wealth, including pension wealth, or on workers' expectations regarding future employment and retirement. These omissions are particularly troubling given the dramatic increase in the incidence of job loss among older workers in recent years. Preliminary results show substantial damage to wages, financial assets, and future employment as a result of late career job loss.The second stage of the project builds on these reduced-form studies, and examines the effects of displacement-induced changes in earnings and wealth on actual and expected retirement, using the framework of a dynamic retirement model. Most empirical implementations of retirement models do not allow an explicit role for assets or saving in the retirement decision, nor for liquidity constraints, features that are particularly important for job-losers given preliminary findings that wealth is significantly damaged by job loss and that this substantially affects subsequent transitions into and out of employment. In this project, the well-known "option value" retirement model is modified to allow for saving, assets and liquidity constraints, and the model is estimated using HRS data on actual and expected employment.Beyond the fundamental interest in job loss itself, the substantial within-person variation in the future returns to work and retirement generated by the job loss provides a unique opportunity to examine more general issues of retirement. In addition to the usual cross-sectional data on observed retirement, the longitudinal data on retirement expectations allows for the estimation of fixed-effects versions of the option value model. Because previous applications of these models have implicitly used differences across individuals in future returns to work and retirement, this project, in identifying this within-person source of variation, offers an important contribution to the empirical literature on retirement decisions.