The rate of unemployment is in general much higher among young workers than for more mature workers with similar levels of education. In Discussion Paper No. 1314, Michael Orszag and Gylfi Zoega attempt to provide a new explanation for this phenomenon. The turnover-training models of Phelps and Salop are modified by allowing for asymmetric information about workers' inherent productivity. Workers differ in terms of their post-training productivity and the employer does not have complete information about the ability of prospective workers before they have been hired, given training and started productive work. Hiring becomes an investment decision under uncertainty. The paper derives the impact of the forecast error in predicting future performance - which is taken as the measure of uncertainty - on equilibrium unemployment.

The main finding is that uncertainty reduces the rate of hiring and hence raises unemployment in the steady state because of firms' aversion to risk. The following implications are drawn from this result: The rate of unemployment among young workers should be higher, other things equal, because of less information being available due to shorter employment histories. Firms' policies of using observable worker characteristics describing possible ability levels are rational in the profit-maximizing sense and, moreover, increase the number of hires. Any form of restrictions on firms' choices increase unemployment. Acquired attributes, such as education, can serve a socially useful signalling role by reducing uncertainty. One policy proposal put forward is to pool risks by introducing industry-funded training schemes. However, the incentive problems caused by such a scheme are recognized.

Hiring Risk and Labour Market Equilibrium
J Michael Orszag and Gylfi Zoega