Does unemployment insurance (UI) create unemployment?

Sure it does. If you pay people to be unemployed, odds are that you're encouraging some people to delay seriously looking for a job or encouraging them to hold out for a higher paying job while they collect a government check. In the absence of extended benefits, the unemployment rate would have been about 0.4 percentage point lower at the end of 2009, said the San Francisco Fed.

MORE ON Unemployment Benefits:
  Derek Thompson: FLASHCARD: Unemployment Benefits

Harvard Professor Robert Barro did his own number crunching, and he came to a different conclusion. Twenty-five percent of America's officially unemployed population -- four million people -- are unemployed for one reason: the government is paying them to be unemployed. If jobless benefits hadn't been extended to 99 weeks, he wrote in the Wall Street Journal, our unemployment rate would be 6.8 percent.

Here's his reasoning. In the last deep recession in the early 1980s, long-term joblessness peaked around 25 percent of the total unemployed population. In this this recession, the share of long-term unemployment is twice as high, near 50 percent. What's the only difference? Jobless benefits. Without them, half of the long-term unemployed would have a job.

This seems to leave quite a bit out of the equation, including but not limited to:

(1) Financial crisis tend to deeper, more prolonged recessions;

(2) The last two recoveries have been, rather infamously, jobless recoveries in which long-term unemployment continued to increase for months if not years after the end of the recession;

(3) Not all of the long-term unemployed are collecting UI, or extended UI, because some are in states that don't offer all 99 weeks;

(4) It seems impossible that extending unemployment insurance beyond a certain week is responsible for any share of long-term joblessness that exceeds 1983.

On Monday, I reached out to Professor Barro to chat about the article. His office told me that he was unavailable. I would happy to host Barro's defense of his math, and he should consider this blog post another invitation. This would be my opening question: Do you think that this recession and recovery is so similar to the 1983 economy that we should expect to see the exact same rate of long-term unemployed?