Public Employment Agencies and Unemployment Spells: Reconciling the Experimental and Nonexperimental Evidence
研究发现，使用公共就业机构的失业者失业时间更长，但这可能是因为他们在失业一段时间后才使用该机构，而非机构本身导致失业延长。利用英国微观数据，作者提供了与实验证据一致的新计量结果。
According to the econometric literature, unemployed jobseekers who use a Public Employment Agency (PEA) have longer unemployment spells than those choosing alternative search methods. This contrasts sharply with the findings from well-designed US experiments where treatments subject to increased PEA services have shorter spells. We argue that the non-experimental studies are biased towards finding a positive relation as they ignore the possibility that PEA's are chosen after some unemployment has been accumulated. Exploiting UK microdata with information on the timing of PEA usage we present new econometric evidence in line with the experiments. Delays in PEA use can also explain previous econometric findings.