retraining, insert (7) and (9) and e = 0 into hA < hC to get:
λ I s δτλ0(1 - d)
(10)
τ + 1 - θ < δτ (1 - b) + δτ d
Thus, STU is more likely to be observed when the opportunity costs of being unemployed are
low (i.e. when labor markets are currently tight (large λ0) but little tightness is expected in the
future (small λτ)), when the direct cost in terms of human capital degradation are low (low d),
when the benefits of STU are high (i.e. unemployment benefits (b) are generous and granted for
a long period (i.e. δτ is large)), and when exiting unemployment after some time is relatively
attractive (i.e. welfare assistance (s) is low, and taxes on labor income (θ) are not too high).
Ceteris paribus, STU without re-education is more prevalent among older workers (low δT).
3. Comparative Statics
We now discuss how labor market institutions (in particular, the generosity of unemployment
benefits b, duration of eligibility for unemployment benefits τ , labor market taxes θ, generosity of
welfare assistance s), the characteristics of the unemployed (in particular, human capital h, skill
loss d, age T -τ, retraining costs c) interact with adverse contemporaneous (λ0) or expected (λτ)
labor market shocks to explain retraining and the duration of unemployment. By differentiating
the functions (7) and (9) with respect to the model’s institutional and individual parameters
we obtain how hA and hC change their position in the career diagram, providing comparative
statics of individual career decisions. If we imagine a country’s (or region’s) workforce ordered
along the h-line according to their human capital endowment, we can infer the comparative
statics of aggregate short-term and long-term unemployment.
3.1. Unemployment Benefits. In line with empirical evidence, we find that a higher replace-
ment rate b increases overall unemployment (e.g. Nickell et al., 2005). Our model predicts an
increase STU, but not in LTU. In Figure 1, an increase in b leaves hA unaffected and shifts hC
to the right.
∂hA ∂hc _ δτ hC
∂b , ∂b δτ (1 — b) + δτ d > .
Higher unemployment benefits increase STU because they make it worthwhile to let human
capital erode for a while or to use an unemployment spell for the accumulation of new skills.
Unemployment benefits increase STU at the expense of employment and leave LTU unaffected.