Imperfect Information and Slow Recoveries in the Labor Market

The unemployment rate remains elevated long after recessions, a persistence that standard search-and-matching models cannot explain. I show that noise shocks—expectational errors due to the noise in received signals about aggregate shocks—account for much of this sluggishness. Using a structural VAR, I find that absent noise shocks unemployment would have recovered to its pre-recession level six quarters earlier over 1968–2019. To interpret this evidence, I develop a search-and-matching model with on-the-job search, endogenous search effort, and wage rigidity. Embedding imperfect information generates two channels of persistence: slow learning amplifies the effects of persistent productivity shocks, and noise shocks provide an additional source of sluggishness, further magnified by sticky wages and vacancy posting. The model successfully replicates both the slow recovery of unemployment and systematic forecast errors, highlighting imperfect information as a key mechanism behind post-recession labor market dynamics.