In this dissertation, I build macroeconomic models to answer questions of empirical relevance for the study of labor markets. The dissertation consists of an introductory overview and three research essays. The first essay is devoted to duration dependence in unemployment, namely the fact that recently unemployed workers have a signicantly better chance of finding a job than the long-term unemployed. I build a directed search model to quantify the importance of three common explanations for this fact: (i) unobserved worker differences, (ii) skill loss, and (iii) job-search effort decline. Two novel results emerge: first, the bulk of the effect of unobserved heterogeneity is concentrated in the first six months of the unemployment spell; the drop in job-finding rates observed at longer spells is mostly a result of skill loss and lower search effort. Second, skill loss has a vastly greater impact on job-finding than the decline in search effort. These results have two clear implications for labor market policy: (i) the impact of active labor market programs is expected to be larger for the long-term unemployed; (ii) job-training programs are expected to be more effective than job-search assistance policies at reducing long-term unemployment. In the second essay I study how information obtained by a worker while trying to find a job affects her job-search effort. Specically, I analyze how a worker, who is uncertain about her labor market traits and learns about them while looking for a job, allocates her search effort over the unemployment spell. The main result is that search effort is increasing over time when the worker is optimistic about her traits but decreasing when the worker is pessimistic about her traits. This result can explain discrepant empirical findings from previous literature on search effort. The final essay is devoted to job-search effort as an insurance channel. I build a model in which workers face substantial risk in the labor market but they have two means of self-insurance against this risk: increase their savings and their search effort. The main result is that when labor market risk becomes more severe workers increase both their savings and search effort but the increase in savings is twice as large as the increase in search effort. That is, workers make use of search effort as an insurance channel but much less than the savings channel.