Essays in this dissertation address three research questions. (1) What types of persons hold dual jobs and what are their motives for doing so? In essay 1, I investigate multiple factors that affect the decision to hold more than a single job. Using data from the Current Population Survey (CPS), the first essay documents the characteristics of second jobs and multiple job holders in the U.S. I characterize the types of people who hold dual jobs and use additional information from the BLS to find out workers’ motives for holding multiple jobs. I examine how multiple job holding differs with respect to age, education, race and ethnicity, sex, foreign-born status, marital status, public-private worker status, broad industry and occupation. (2) How does dual job holding vary with the business cycle and state of the labor market? Essay 2 explores a large micro data set for 1998-2013 that covers most U.S. urban labor markets. We find clear-cut evidence that multiple job holding across labor markets and over time is weakly cyclical, thus (slightly) exacerbating rather than mitigating the severity of business cycles. Much of the cyclicality in multiple job holding seen across labor markets, however, is not causal, dropping sharply after accounting for MSA fixed effects. Using longitudinal worker data, there is minimal response to unemployment changes within labor markets over time. Our large CPS sample size produces precise estimates, albeit ones close to zero, helping explain conflicting results in prior studies based on far smaller data sets. (3) How might one explain the persistent geographical differences in multiple job holding? Essay 3 documents what are systematic (i.e., long-run) differences in multiple job holding across labor markets (MSAs) and explores possible explanations for these differences. Geographical differences in multiple job holding rates have received little attention, although the multiple job holding rates in some regions of the country are substantially higher than in other regions, and these differences have been persistent over time. Examining correlates of these labor market differences in multiple job holding provides us with a better understanding of the determinants of labor supply and how local labor markets work.

<pre>This dissertation consists of three independent chapters at the intersection of macroeconomics and labor economics. The first chapter studies the job-search trade-offs between full-time employment, part-time employment, and multiple job holdings. The second chapter explores the macroeconomic relationship between property crime and output in a dynamic stochastic general equilibrium framework. The third chapter studies the causal effect of property crime on output.</pre> <pre>The first chapter develops a search-matching model of the labor market with part-time employment and multiple job holdings. The model is calibrated to data from the CPS between 2001 and 2004. Workers are able to choose their search intensity and are allowed to hold two jobs while firms can choose what type of worker to recruit. When compared to the canonical Diamond-Mortensen-Pissarides model, this model performs quite well while capturing some empirical regularities. First, the model generates recruiting and vacancy posting rates that move in opposite directions. Second, part-time employment is up to 10 times more responsive than full-time employment. Third, the model suggests that multiple job holding rates are more flexible than observed in the data with the rate changing by as much as 4 percentage points compared to 0.1 percentage points in the data. Finally, the full model is able to capture compositional changes during recessions with the full-time rate declining and the part-time rate increasing. It also produces an empirically consistent increase in the unemployment rate as well as a decrease in output. The DMP model is more muted than in the data for both.</pre> <pre>The second chapter explores how property crime can affect static and dynamic general equilibrium behavior of households and firms. I calibrate a model with a representative firm and heterogeneous households where households have the choice to commit property crime. In contrast to previous literature, I treat crime as a transfer rather than home production. This creates a feedback loop wherein negative productivity shocks increase property crime which further depresses legitimate work and capital accumulation. These responses by households are particularly important when thinking about the effect of property crime on the economy. Household and firm losses account for 24% of compensating variation (CV) and 37% of lost production. This suggests that behavioral responses are quite important when calculating the cost of property crime. Finally, on the margin, decreasing property crime by 1% increases social welfare by 0.19%, but the effect is diminishing suggesting that reducing crime entirely may not be optimal from a policymakers perspective.</pre> <pre>The third chapter estimates the causal effect of property crime on real personal income per capita. Running system GMM on an unbalanced panel of MSA-year pairs suggests that property crime reduces real personal income per capita by a highly statistically significant 13.3%. This implies that the average person loses $4,869 (2009 dollars) per year with real annual personal income per capita totaling $36,615. The effect is driven primarily by larceny-theft and burglary with highly statistically significant coefficients of -0.179 and -0.110 respectively. Estimates for the effect of robbery are unstable, and the effect of motor vehicle theft is statistically significant, but smaller with a coefficient of -0.060.</pre>