Effect on retirement
decisions
\label{effect-on-retirement-decisions}
Given that we find effects on earnings and accumulated earnings but do
not find statistical significant effects on wealth, it is possible that
reentering the civilian life in periods of high unemployment does not
affect retirement. In this case, we might expect that veterans who enter
the labor market under poor conditions act in ways to mediate those
effects prior to retirement. We documented one instance of this above;
these veterans are more likely than others to attend and complete
college. But to investigate this further, we study two additional
outcomes: work expectations and labor force status in old age.
To estimate the effect on work expectations we estimate the linear model
in equation (5). The outcome \(y_{\text{it}}\) is the veterans’ reported
(stated) probability (from 0 to 100) of working full-time?? or searching
for work at various ages.11In each case, we include only
veterans of appropriate age groups. Need more here?? We include a
quadratic-term in age and similar controls as in the previous
regressions. We cluster the standard errors at the individual level.
\begin{equation}
{}_{\text{it}}=\alpha+\beta_{1}UR_{i}+\beta_{2}\text{Ag}e_{\text{it}}+\beta_{3}\text{Age}_{\text{it}}^{2}+\beta_{4}X_{i}+{\gamma_{t}+\epsilon}_{\text{it}}\nonumber \\
\end{equation}
The estimation results are shown in Table 6. In the OLS model, we find
that an increase of 1 percentage point in the unemployment rate at the
time of entering the civilian labor market increases the reported chance
of working full time at every age tested; in the 2SLS model, the
estimated coefficients are larger than in the OLS model although not all
achieve statistics significance. Finally, we test a model of the
probability of claming Social Security benefits; there is some evidence
that veterans who enter the labor market during poorer conditions are
more likely to claim Social Security, although the results of the 2SLS
model generally suggest there is no statistically signfiicant difference
in rate of claiming benefits.
Table : Effect of unemployment rate at time of exit from active service
on full-time work expectations