Obviously, an increase in retraining cost has no effects if retraining does not take place (e = 0)
at low c. With retraining, STU is reduced at both ends when c rises. At the upper end of the
education spectrum, some workers who previously considered an STU spell to upgrade their
skills now find that they cannot afford it and stay employed. Retraining becomes also too
expensive for some at the lower end of the education spectrum. For those persons, however, it
is then the best choice to stay out of work permanently.
While retraining costs can be determined by personal characteristics, they may also impor-
tantly depend on labor market policy. For example, retraining costs can be reduced by subsidies
for (re-) education. Our model suggests a short-run increase of unemployment due to subsi-
dies for retraining. The policy only reduces unemployment in the long run, i.e. when workers
re-enter employment who were motivated by the retraining subsidy to shift from LTU to STU
(reflected by the move of hA to the left). Note that retraining subsidies are particularly effective
for persons with rapid human capital degradation since ∂hA/∂c increases and ∂hC/∂c decreases
in absolute terms when d rises.
3.7. Age. Comparative statics of increasing working-life T and, thus, a larger discount factor
δT depend on a number of factors in subtle ways. The derivatives are
∂hA ce 0 ∂hC e — dhC
∂δT = -δTT(i — θ)(i — d) ≤ , aδT = δτ(1 — b) + δτd'
The duration of working life can be analyzed in a cross-sectional perspective distinguishing
between younger and older workers or in a time-series perspective when investigating exogenous
events like an increase in retirement age. In analyzing the effects of a longer work life on STU,
we distinguish two cases.
When retraining does not pay off (e = 0), the effects of an increased duration of working life
are clear. In this case, our model predicts no effect on LTU because low-productivity workers
prefer to stay on the dole permanently anyway. Yet, an increase in the length of working life
induces an increase of employment, i.e. a reduction of STU (∂hC /∂δτ < 0 for e = 0). The reason
is that a longer working life increases the present value of a permanent loss of human capital
during an STU spell. Therefore, STU becomes more threatening and people exit unemployment
earlier.
With retraining being worthwhile (e > 0), an increase in work-life affects STU because it
15