Sunday, February 28, 2010

Unemployment benefit reduces work even in recessions

In 1995 Sweden was in as deep, arguably deeper crisis than the US is now (unemployment went from 2% to 9%, preceded by a financial crisis). In the midst of the crisis the replacement rate from the unemployment insurance was reduced by 5 percentage points, from 80% to 75%, for most, but not all, workers.

A Keynesian would argue that this could have no effect, as there was no “demand” for workers. Indeed, in textbook Keynesian analysis unemployment should have gone up, since this further reduced consumer demand.

But this is not what happened. Supply side reform works even during weak labor markets. The benefit cut put a downward pressure on unemployment. The decrease has been convincingly tied to the benefit cut by Carling, Holmlund and Vejsiu 2001, who show that those workers whose benefits got cut were about 10% more likely to exit unemployment, compared to those whose benefits were not cut. They write:

"Our estimates suggest that the reform caused an increase in the transition rate [from unemployment to employment] of roughly 10%"

The figure shows the probability of exiting unemployment for two groups, those whose benefits were cut (T) and those whose benefits were not cut. For those whose benefits were not cut the probability of exiting unemployment dropped significantly. For those whose benefits were cut the probability of exiting unemployment did not decrees, and actually increased slightly.

"The period 1995-96 is characterized by a weakening of overall economic activity, resulting in a substantial fall in the number of new vacancies notified to the employment offices (Fig. 4). With this evolution of labour demand conditions, some decline in job finding rates should be expected. The surprising part of Fig. 3 is the absence of a decline in job-finding rates among workers in treatment group."

Now, I am not arguing that benefits be cut in the U.S. There is always a trade off between the pain low benefits cause and the pain unemployment causes. The Swedish benefits rates started from a high level, whereas the American are already pretty low.

I just want to make a theoretical point, that Keynesians who claim that supply and demand stop applying in recessions are wrong. The fundamental laws of the economy do not just vanish because the capacity utilization rate decreases.


  1. How is one to read the y-axis in that graph?

  2. More on unemployment benefits and the unemployment rate here:

  3. Likelihood of getting a job.

  4. Excellent post! This suggest that Keynsian-style disequilibrium isn't nearly as strong as most Keynsians say it is.


Google Analytics Alternative