The Great Recession has not been a great vacation. Four years since the recovery officially began, many people are still unemployed not because they don't want to work, but because, with three job-seekers for every job opening, they can't find any work. And they are trying to find work. Indeed, a new paper by Rand Ghayad, a visiting scholar at the Boston Fed and a Ph.D. candidate at Northeastern University, shows that unemployment insurance hasn't otherwise made the unemployed less likely to take a job. If anything, it's made them more likely to keep looking.
For a certain class of conservative economists, the unemployed must be funemployed. They just can't conceive of a world where supply can outpace demand; where excess demand for money and money-like assets can push the economy into a slump. (Paul Krugman's classic essay on the Capitol Hill babysitting co-op shows how it can). Their belief that supply creates its own demand is an age-old fallacy called Say's Law, which isn't a law, and hasn't been one for a long, long time. As Krugman points out, it was discredited enough back in the 1930s that it was thought something of a strawman when Keynes debunked it back then. But yesterday's caricature has become today's conviction. Casey Mulligan, for one, has never met a problem he doesn't think is a supply one. Back in December 2008, he argued unemployment was exploding, because people were choosing not to work so they could get mortgage modifications. By 2012, he decided food stamps were really to blame for joblessness. Needless to say, these arguments don't even pass the laugh test, let alone fit the data.
Now, that doesn't mean unemployment today is all demand and no supply -- just mostly. Safety-net spending does increase joblessness on the margin, and that's the point. It cushions the unemployed so they don't have to take the first job they find, but can wait for a better one. The question, of course, is how generous the safety-net should be. It's a matter of not giving the unemployed forever to take a job, but not giving them too little time when there are few jobs for the taking.
And that brings us to the pertinent question: Was 99 weeks of jobless benefits too much?
Long-term unemployment and long-term unemployment benefits both hit record lengths during this Lesser Depression. The average length of unemployment hit a postwar high of 40 weeks, and unemployment insurance did too, getting extended to 99 weeks (though it's fallen significantly the past two years). But there's no evidence that long-term unemployment benefits have discouraged the long-term unemployed from taking jobs -- because they can't get jobs to begin with.
They can't even get interviews. As Ghayad showed before in a field experiment, employers largely ignore the resumes of people who have been out of work for six months or longer. Firms assume there must be something wrong with people who have unemployed that long, and don't want to spend time finding out what it is, not when they have a stack full of resumes to get through. But the long-term unemployed have kept looking, at least in part because of benefits. Jesse Rothstein, a professor of economics at the University of California-Berkeley, found that extended unemployment benefits have increased unemployment by between 0.1 and 0.5 percentage points, but at least half of this increase came from fewer long-term unemployed dropping out of the labor force. In other words, people who would have otherwise given up looking for work didn't, because they had to keep looking to qualify for benefits. A recent paper by Henry Farber of Princeton and Robert Valletta of the San Francisco Fed found the same: Unemployment benefits -- even long-term unemployment benefits -- encourage the jobless to keep looking for work.
But what about the unemployed who can't get benefits? Remember, only people who lose their jobs are eligible for unemployment insurance. People who leave their jobs or who are looking for their first jobs or who are re-entering the job market are not. If unemployment insurance is making people not want to work, then the labor market should look worse for those with benefits than for those without. It doesn't.
Ghayad shows us just how much it doesn't by breaking out the Beveridge curves for the unemployed who can and can't get benefits. Now, the Beveridge curve just shows the relationship between job openings and joblessness, and it should be just that -- a curve. There should be a pretty stable relationship between the two. If there's not, and the curve shifts out or up, it's a sign that the labor market is somehow broken. As you can see below, it's been badly broken for the unemployed who can collect benefits ever since late 2009.