growth equally distributed among all sectors and all jobs is indicated. According
to Aghion, Howitt (1994) technical progress is not equally distributed implying
that technical progress generates and destroys jobs at the same time. This causes
a creative destruction effect, which generates that an increase of technical progress
induces via the net destruction of jobs unemployment, and a capitalization effect
representing that an increase in technical progress leads to more employment by the
net creation of jobs. Aghion, Howitt shows that, at small rates of productivity
growth the creative destruction effect dominates the capitalization effect and a re-
duction in employment is induced; whereas at high rates of technological progress
the opposite holds causing an increase in employment. Thus, a hump shaped rela-
tionship between productivity growth and unemployment is implied.
While in both approaches productivity growth generates unemployment, Bean,
Pissarides (1993) point to the other direction of causation requiring that unem-
ployment determines technological progress. Using this hypothesis they attain a
positive respectively negative interdependence between growth and unemployment
depending on whether the point of view is Classical or Keynesian. Postel-Vinay
(1998) and Mortensen, Pissarides (1998) confirm the ambiguous link between
growth and unemployment in endogenous growth-matching-models. They show in
stochastic matching models with heterogenous productivities that, via the creation
of new jobs, increasing productivity growth leads to an increase in employment
when renovation costs are low and it induces unemployment when renovation costs
are high.
Some recent studies analyze the effects of skill-biased technological shocks on
unemployment respectively long-term unemployment. Coles, Masters (2000)
examine in a search-matching model the effect of skill depreciation on the equilib-
rium level of unemployment when long-term unemployment emerges as endogenous
phenomenon. Their model implies that today’s recession, which leads to longer un-
employment spells, impacts on the distribution of market skill levels in the future.
Due to this result, they conclude that subsidizing retraining to reduce long-term
unemployment is inappropriate; a better way is to subsidize vacancy creation. The
argument of Ljungqvist, Sargent (1998) points in the same direction and iden-
tifies generous welfare schemes as obstacles for reducing long-term unemployment.
Their point is that generous welfare schemes tend to be more prone to generate
high levels of unemployment when economies undergo rapid structural change. If
redundancy is associated with the loss of skills, abundant welfare payments make
10