Lecture 3: Empirical evidence on unemployment
The issues
• View #1: Unemployment is the result of cumulated shocks of various nature and persistence
• View #2: Unemployment is a structural phenomenon due to labor market rigidities, as in the models we have seen
Blanchard-Wolfers
• Unemployment has increase in most European countries since 1975
• It was very low before
• And there is an increased dispersion across European countries
The puzzle:
• Shocks-based explanations do not explain the dispersion – Oil shocks and macro conditions too similar
across countries
• Institutions-based explanations fail to account for the fact that unemployment was low while these institutions were still around
The solution
• Shocks interact with institutions
• Some institutions are harmless absent shocks
• But they can increase the magnitude of the unemployment response to the shock
• And they can increase the persistence of the shock
The Pro: it’s plausible
• Coordination in bargaining better response of wages to a productivity slowdown
• High duration of benefits More of LTU during a recessionMore persistence of a shock (as seen)
The Con:
• Hard to think of an institution which does not also increase the natural rate.
• More coordination less negative externalities in wage-settinglower markup of wages on prices less unemployment
• Longer benefitslower search intensity and greater worker outside option more unemployment
The shocks: the medium run
• A slowdown in TFP growth (why does it affect u?)
• An increase in the real interest rate (same question)
• An unexplained and poorly documented fall in labor demand…
TFP:
Real interest rate:
Institutions (their view)
• UB : Higher unemployment and higher duration
• EPL: Higher duration, lower inflows, ambiguous effect on employment
• Taxes: lower wages, lower employment, but little effect on unemployment if « aspirations » adjust proportionally to after-tax wages
Interactions: Specification #1
• We use institutions to explain the change in unemployment rather than its level
• We use a panel and assume the effect is the same across countries
• This leads to the following specification (i = country, j = institution)
Interactions: specification #2
• The time dummy is replaced by a vector of country-specific shock
• The effect of shocks depends on institutions
• But a given institution has the same effect on all shocks
Assessment:
• It is hard to see why institutions could not matter on their own
• The authors present no institutions-only exercise
• If onlys S*I matters, unemployment should go away as shocks are reversed
• Alternative possibility: multiple equilibrium rates of unemployment
Top Related