with Dan Bernhardt
We study the wage-setting problem of an employer with private information about demand for its product when workers can engage in costly on-the-job search. Employers understand that low wage offers may convey bad news that induces workers to search. The unique perfect sequential equilibrium wage strategy is characterized by: (i) pooling by intermediate-revenue employers on a common wage that just deters search; (ii) discontinuously lower revealing offers by low-revenue employers for whom the benefit of deterring search fails to warrant the required high pooling wage; and (iii) high revealing offers by high-revenue employers seeking to deter aggressive poachers.
Krueger and Mueller (2011) show that job seekers' search effort fell throughout unemployment during the Great Recession. Exploiting the longitudinal nature of their data, I show that variation in past search effort explains this decline. Furthermore, I document that search effort rises after a job offer is received. These facts are inconsistent with standard models of search. I introduce a tractable model of sequential search in which job seekers are uncertain about the process governing the arrival of job offers and learn through search. I use the model to show that beliefs influence search via two opposing channels: Failing to find work reduces search by reducing the perceived opportunity cost of leisure, but stimulates search by reducing the perceived option value of unemployment. I structurally estimate the model and show that learning can quantitatively account for the measured effects of job offers and cumulative past search in the data.
We estimate a real business cycle economy with search frictions in the labor market in which the latent wage follows a non-structural ARMA process. The estimated model does an excellent job matching a broad set of quantity data and wage indicators. Under the estimated process, wages respond immediately to shocks but converge slowly to their long-run levels, inducing substantial variation in labor's share of surplus. These results are not consistent with either a rigid real wage or flexible Nash bargaining. Despite inducing a strong endogenous response of wages, neutral shocks to productivity account for the vast majority of aggregate fluctuations in the economy, including labor market variables.
Review of Economic Dynamics, January 2013, Vol. 16(1), 120-134
Using a newly created microeconomic archive of US imports at the tariff line level for 1930–1933, we construct industry-level tariff wedges incorporating the input–output structure of US economy and the heterogeneous role of imports across sectors of the economy. We use these wedges to show that the average tariff rate of 46% in 1933 substantially understated the true impact of the Smoot–Hawley (SH) tariff structure, which we estimate to be equivalent to a uniform tariff rate of 70%. We use these wedges to calculate the impact of the Smoot–Hawley tariffs on total factor productivity and welfare. In our benchmark parameterization, we find that tariff protection reduced TFP by 1.2% relative to free trade prior to the Smoot–Hawley legislation. TFP fell by an additional 0.5% between 1930 and 1933 due to Smoot–Hawley. We also conduct counterfactual policy exercises and examine the sensitivity of our results to changes in the elasticity of substitution and the import share. A doubling of the substitution elasticities yields a TFP decline of almost 5% relative to free trade, with an additional reduction due to SH of 0.4%.
I introduce a simple measure of the extent of discouragement among potential job seekers---the discouragement rate---and show that this measure rose sharply during the Great Recession and never recovered to its pre-crisis level. To explain this, I propose a theory in which fears of prolonged joblessness can become self-fulfilling, drawing the economy into a high-discouragement, low-participation state: a discouragement trap. Intuitively, when job losers fear it will be difficult to find work if they remain jobless for too long, they search aggressively, crowding out those at the back of the queue, inducing labor force withdrawal, and rationalizing fears of prolonged joblessness. This mechanism emerges naturally from a model of ranking in the spirit of Blanchard and Diamond (1994) when workers make participation decisions and submit multiple applications. The model gives rise to multiple Pareto-ranked steady states with significantly different rates of labor force participation but similar, and in some cases identical, unemployment rates. The high-participation state is saddle-point stable, while the (Pareto-inferior) low-participation state is a sink---a discouragement trap. I study global dynamics, conditions under which an economy is susceptible to such traps, and implications for unemployment insurance policy.