Solving the Chicken-or-Egg Job Problem

How can you gain job experience when all jobs require previous experience?

By SHARE
EC_073113_Wasserman.jpg

You know the complaint: "I'm trying to find a job, but everywhere I look I see experience required. How am I supposed to accumulate experience if I need experience to gain experience?" I've heard the story twice this week alone. It is of particular importance during what is still a slow recovery, with high rates of unemployment and, especially, youth unemployment. Why does it happen? And how do you get yourself a chicken or an egg if you start out with neither?

Harvard economist Amanda Pallais has, in a new research paper, tried to provide answers to these questions. "How?" you may ask. Well, by hiring lots of people without experience and seeing whether that helps them, of course. To avoid having to meet the capital requirements necessary to set up a Terry McAuliffe-style "manufacturing" facility, she hired workers on oDesk, a global online marketplace. On oDesk, employers post jobs, and workers can then apply for those jobs by setting a wage. Employers can see some demographic info, test-based certifications and, importantly, workers' previous oDesk experience and associated evaluations.

[See a collection of political cartoons on the economy.]

Professor Pallais posted 10-hour data-entry jobs, received more than 3,500 applications, and hired about 1,000 of them, randomly. She then rated their performance, and gave half of them detailed evaluations and the other half coarse evaluations. What happened to the luckily employed workers, about three-quarters of whom had no previous oDesk experience or evaluations? Did detailed evaluations help? They had been handed a chicken – did it bring them eggs?

Turns out it did, indeed - and without harming anyone else much. Workers who were hired found more new jobs after their evaluations became public than the control group of applicants who were randomly rejected, and their wages went up. For inexperienced workers, total earnings over the next two months tripled. Workers whose evaluations were more detailed (and positive!) benefited even more. Workers who performed less well were hurt a bit by their evaluations, but not by enough to undo the benefits accruing to their high-performing peers.

Pallais suggests that what we see here is the value of firms gaining access to information about workers' ability. Her evaluations followed the existing distributions of ratings on oDesk, so it's not rating inflation that helped her workers. And it was only after evaluations became public that her employees started becoming successful, not when it became clear that they had been hired: it's information, not just experience as such, that made them more attractive.

[Read the U.S. News debate: Does Stimulus Spending Work?]

And they weren't just stealing jobs from other oDesk workers, even though Pallais' experiment was quite sizable compared to the market as a whole (there were about 200,000 hours of work per week being done on oDesk at the time). The benefits to employers and workers from all the new jobs created thanks to the new information available outweighed the full cost of the experiment, including the hours workers spent entering data for no other reason than to facilitate the experiment, and the slight drop in the prevailing wage among credentialed workers. Imagine how beneficial it all would have been if they had done something productive!

Now, if one is willing to extrapolate from this online marketplace to the broader labor market, and I'm more than happy to do so, there are important implications to Pallais' findings. Why do employers not hire more people? Because they cannot tell if they are any good, and no one wants to be stuck with a terrible worker. Of course, if all employers behave that way, no employee is ever going to look good to anyone.

What can the worker do to avoid the vicious cycle of eggless non-chickenhood? He could lower the wage he demands, or even offer to work for free. If that makes him attractive to employers, they will hire him. This, in turn, will generate information that is valuable on aggregate, and inefficiently low levels of unemployment will disappear in entry-level labor markets.

What keeps that from happening? Minimum wages, for example. Bans on unpaid internships. The liquidity constraints that young workers face also spring to mind. Potential policy solutions to these problems aren't hard to come up with – they're actually clamoring from right behind me as I write this: lower the minimum wage for young, unexperienced workers. Allow them to gain experience in unpaid internships. And if you're into that kind of thing, subsidies for firms that hire unexperienced workers but can't capture the value of the information that their employment would produce.

Stan Veuger is an economist at the American Enterprise Institute.

  • Read Jim Lardner: Wall Street's Weak Arguments Against a New Glass-Steagall
  • Read David Shulman: Republicans Shouldn't Dismiss Obama's Latest Economic Plan
  • Check out U.S. News Weekly, now available on iPad