Essays on the Simulation-Based Estimation of Dynamic Macroeconomic Models

This dissertation consists of two chapters, both of which approach macroeconomic issues using simulation-based methods. Aside from the fact that each chapter contributes to its narrowly scoped field, the two chapters demonstrate an implementation of simulation-based estimation techniques and identification strategies to examine dynamic properties of unobserved economic shocks. The main objective of two chapters is to understand the properties of shock process, which in turn provides better macroeconomic implications. The first chapter structurally estimates idiosyncratic labor income risks over the life-cycle to obtain implications for a redistribution policy, namely tax and transfer systems. Since a redistribution policy provides a partial insurance to those exposed to income risks, understanding the underlying life-time labor income risks that households face is central to designing better institutional arrangements. The chapter constructs a human capital life-cycle model and structurally estimates the underlying source of labor income risks across age. We find that the estimated shock process is significantly age-dependent even after controlling for the endogenous responses to the exogenous shocks. In particular, young workers encounter a highly persistent (almost unit-root) but relatively small volatility of permanent shocks, while older workers encounter a less persistent but higher volatility of permanent shocks. In addition, we demonstrate that under the age-dependent shock process the self-insurance ability of young workers is 20% lower than that of middle-aged workers. Finally, we find that more benefits, either through a tax exemption or subsidies, to young workers drastically improve aggregate production, welfare, and income inequality. In the second chapter, we structurally estimate the dynamic properties of elasticity of substitution between capital and labor to resolve well-known puzzles in labor market dynamics: Dunlop-Tarshis phenomenon, the labor productivity puzzle, the labor share puzzle including its oveshooting response to productivity shocks, and the hours-productivity puzzle. We propose an aggregate production function that potentially takes a different shape in the short run (SR) from the long run (LR). Specifically, we allow for cyclical fluctuations of the short-run elasticity of substitution between capital and labor, σt, while keeping the Cobb-Douglas shape in the long run. We find that productivity shocks are on average biased toward labor (i.e. σt\u3c1) and that this bias is largest in expansions. Specifically, &sigmat declines (making capital and labor more complementary) during expansions and increases (making capital and labor more substitutable) in recessions with a lagged and slow diffusion. The slow diffusion of σt after a productivity innovation breaks the tie between competitive wages and average labor productivity in a very specific manner by making wages less sensitive to changes in labor productivity at prompt, but more sensitive thereafter. We find that our competitive framework is largely consistent with the business cycle behavior of the labor market and resolves four main labor market puzzles at once. We show that our competitive model outperforms standard New-Keynesian and Diamond-Mortensen-Pissarides search frameworks. Finally, we show direct empirical evidence from KLEM data that supports the behavior of our structurally estimated σt. On the one hand, recent macroeconomic models, as illustrated above, have become too complicated to solve analytically, especially when dealing with uncertainty and disaggregated data. On the other hand, computational methods and capacity are advancing day by day. Therefore, with the aid of simulation-based methods, this dissertation provides methodological advances in estimating the dynamic properties of unobserved shocks and explores the macroeconomic implications