CATHERINE RAMPELL

Dollars to doughnuts.

I’ve written before about about how the longer you’re unemployed, the less likely you are to find a job. That’s probably because of some combination of several factors: skill deterioration; better workers are more likely to get hired faster, leaving a pool of less qualified workers as the ones who disproportionately make it to long-term unemployment in the first place; and the stigma of unemployment, or at least the assumption by employers that the longer-term unemployed will be lower quality because of the previous two factors.

A new study tries to measure how big a factor that stigma is.

In the fall of 2009, the authors sent out 12,054 fake résumés for 3,040 jobs posted online, with most showing that the (fictional) applicant had been unemployed somewhere from one to 36 months. The résumés were all variations on a few standard templates that the researchers came up with. The duration of unemployment was randomly assigned to candidates of otherwise equal qualifications.

Over all, 4.7 percent of résumés resulted in the candidate’s being invited for an interview. But a candidate’s chances of being called back depended on how long he or she had been looking for work.

Courtesy of Kory Kroft, Fabian Lange and Matthew J. Notowidigdo, “Duration Dependence and Labor Market Conditions: Theory and Evidence from a Field Experiment.” These figures are generated by computing the average callback rate for each 3-4 month bin in the range of unemployment lasting between one and 36 months. The line shows the five-month moving averages computed at each month.

Candidates unemployed for just a month had a 7 percent chance of being invited for an interview. Those chances dropped off sharply with just a few more months of unemployment. After eight months of unemployment, the callback rate was just 4 percent. In other words, the chances of getting an interview fell about 45 percent from Month 1 to Month 8.

At that point, though, the callback rate pretty much flattened out.

Candidates unemployed for 8 months, 12 months and even 36 months all had about a 4 percent chance of getting an interview.

“A lot of this sorting mechanism happens really early on,” said Fabian Lange, an economics professor at McGill University and one of the paper’s authors, along with Kory Kroft at the University of Toronto and Matthew J. Notowidigdo at the University of Chicago Booth School of Business.

He attributed these patterns to the fact that in a normal economy, most people who are unemployed find jobs within the first several month. After someone has been unemployed for more than a few months, the chances that the candidate is a desirable one start to fall rapidly. But after some point — in this case, about eight months — employers believe the difference in duration of unemployment is less informative about a candidate’s likely quality.

The résumés were submitted for three categories of jobs: customer service, clerical and administrative, and sales. The drop-off in callback rates was largest for sales.

“That makes some sense,” Mr. Lange said. “If you’re in sales, you’re supposed to sell something. If you’ve been unemployed for six or seven months, you’re not that good at selling yourself, so why should an employer take a chance on you?”

Interestingly, the study also found that a candidate who is currently employed is actually less likely to be called back for an interview than one who is newly unemployed.

This may be surprising, given that people with jobs have already demonstrated that they are employable; there are even some job postings that specifically require candidates to be currently employed to be considered.

The authors of the study offered a few theories for why the currently employed had a lower callback rate than the recently employed.

First, they learned from personnel professionals that employers were concerned that people who were already employed were not serious candidates. Second, some jobs require workers to start immediately, and someone who already has a job will probably have to give notice before leaving. Third, a person who already has a job has more bargaining power and can ask for a higher wage in a new position. There’s also the question of loyalty.

“If you’re an employer, you might see that this person has a job and is not being very loyal to that current job,” Professor Lange said. “You might think, ‘Maybe if I give him or her a job, and they go through a phase of training, maybe he or she will leave me  within a few months on the basis of what they did to the last employer.’”

The study also found that in areas with higher unemployment rates, callback rates for even the recently unemployed were very low, and there was not as much of an advantage to being recently unemployed compared with long-term unemployment.

Courtesy of Kory Kroft, Fabian Lange and Matthew J. Notowidigdo, “Duration Dependence and Labor Market Conditions: Theory and Evidence from a Field Experiment.” These figures are generated by computing the average callback rate for each 3-4 month bin in the range of unemployment lasting between one and 36 months. The line shows the five-month moving averages computed at each month. “High” and “low” unemployment refer to whether cities had above or below 8.8 percent unemployment, the median for cities in this sample.

In other words, employers are more forgiving of longer durations of unemployment when they know that the job market is really bad: they recognize that length of unemployment is not as good a proxy for quality of a worker when no one is being hired, not even the really good candidates.

The problem, of course, is that duration of unemployment is always going to be an imperfect signal for quality. You still have a lot of good candidates being passed over for interviews because they got unlucky, and that unluckiness develops an inertia.

“It’s the classic lemon problem,” Professor Lange said. “You have a lot of lemons, but still some good cars among then. The problem is to distinguish the good from the bad when you have a high proportion of bad.”

One possible policy response, he speculates, is having the government provide some sort of additional screening mechanism to help employers distinguish the unlucky-but-good long-term unemployed from the long-term unemployed who cannot find jobs because they are unmotivated or unskilled.

For example, there could be “training” programs whose purpose is just to demonstrate that students have the discipline and motivation to attend and complete them, regardless of what skills are actually taught.

“It could be something that involves 40 hours a week, the regular workday length, and starts early in the morning,” he said. “If you find people who are willing to subject themselves to a 40-hour-a-week program and complete that successfully, that might be the signal the labor market needs to determine they’re good job candidates.”

Mr. Lange and his colleagues are also looking into ways to redesign the unemployment benefits system to account for the patterns they found.

He says one idea might be to start benefits at lower levels, to encourage people to take jobs sooner before their chances for getting interviews fall off sharply. And then to make benefits more generous once applicants pass the eight-month mark, when it becomes much more difficult to get interviews and applicants are in need of assistance.

Article source: http://economix.blogs.nytimes.com/2012/09/27/longer-unemployment-fewer-interviews/?partner=rss&emc=rss