Two economists in one!

Compare two quotes:

…public policy designd to help workers who lose their jobs can lead to structural unemployment as an unintended side effect. Most economically advanced countries provide benefits to laid-off workers as a way to tide them over until they find a new job. In the United States, these benefits typically replace only a small fraction of worker’s income and expire after 26 weeks. In other countries, particularly in Europe, benefits are more generous and last longer. The drawback to this generosity is that it reduces a worker’s incentive to quickly find a new job. Generous unemployment benefits are widely believed to be one of the main causes of “Eurosclerosis,” the persistent high unemployment that affects a number of European economies.


Do unemployment benefits reduce the incentive to seek work? Yes: workers receiving unemployment benefits aren’t quite as desperate as workers without benefits, and are likely to be slightly more choosy about accepting new jobs. The operative word here is “slightly”: recent economic research suggests that the effect of unemployment benefits on worker behavior is much weaker than was previously believed. Still, it’s a real effect when the economy is doing well.

But it’s an effect that is completely irrelevant to our current situation. When the economy is booming, and lack of sufficient willing workers is limiting growth, generous unemployment benefits may keep employment lower than it would have been otherwise.

“Slightly” does not appear in the first quote. Yet they were written by the same person: Paul Krugman from, first, his principles textbook and, second, his op-ed this morning.

The paper he cites (and also cited by Ezra Klein last Friday) is a recent research note from Federal Reserve Bank of San Francisco economists Rob Valetta and Katherine Kuang.  They separate the duration of unemployment of those who were involuntarily separated from their jobs (thus eligible for UI) and those who either quit or left their jobs (and ineligible.)  The job losers have an average duration of 1.6 weeks longer.  This is assumed to be entirely due to the difference in UI eligibility.  Valetta and Kuang estimate that the effect on the December 2009 unemployment rate is to add 0.4%, which they call “quite small.”  (Worth noting that the fall to 9.5% in June from 10% back then is called “headed in the right direction” by President Obama.  But that’s to be explained by Obama’s speechwriters, not Valetta and Kuang.)

A hint to non-academic economic readers:  When you see someone say “recent economic research says that …” you can bet they are trying to slide by what the bulk of research before that said.  That is, they may be cherrypicking.  Valetta and Kuang cite a recent study in the prestigious Journal of Political Economy by Raj Chetty,  (Ungated copy here.)  He cites that the original evidence on unemployment is that “a 10% increase in unemployment benefits raises average unemployment durations by 4-8% in the U.S.” Chetty doesn’t dispute that point but rather shows that the effect is more one of families being cash-constrained, that liquidity drives the job search rather than the substitution of subsidized leisure for search and labor. His argument is for a higher unemployment insurance benefit, not an extension of weeks.  That is, Chetty doesn’t necessarily change the 4-8% estimate (which would be consistent with the Valetta and Kuang estimates, by the way) but the reasoning behind it.

So I do not understand why Krugman the textbook writer would argue unemployment benefits is a significant explanation for Eurosclerosis while Krugman the op-ed writer would encourage us to adopt those European policies.  Nothing changed in the results.  They applied for the U.S. in good times and bad (or, at least, they haven’t shown a significantly different result in periods of slack demand, contra Krugman’s assertion.) One of those passages is misleading the reader.  Which is it?