Essays on macroeconomics with microeconomic heterogeneity

Abstract

This dissertation consists of three essays on macroeconomics with microeconomic heterogeneity.

In Chapter 1, I empirically investigate the extent to which regional economic activity responds to fiscal shocks. Exploiting state-level variation of military procurement, I apply an instrumental variable local projection method extended to the panel data context to estimate the dynamic causal effects of a military spending shock. These estimates, which are referred to as “regional impulse responses,” indicate three main empirical findings. First, regional output displays a large and lengthy response over a decade to a regional military spending shock, despite the fact that military spending has returned to a normal level after five years. Second, regional population gradually grows over the decade after the shock. Third, the response of construction to military spending is proportionately much larger than that of total output and also represents an important share of overall output responses. This evidence suggests that labor reallocation across regions can be very important for the impact of fiscal policy.

Chapter 2 quantitatively analyzes the regional and aggregate implications of the empirical findings in the previous chapter. I first study a simple model of regional reallocation to build intuition. Then I develop a multi-region New Keynesian model with labor migration and housing construction and calibrate a U.S. economy with 51 regions. The model reveals that labor reallocation amplifies regional output through a boom in construction spending and amplifies aggregate output through a positive “covariance effect” arising from net directed migration towards booming regions where population and regional output per resident are rising simultaneously. To circumvent high dimensionality, I propose a new method to tractably solve spatial dynamic stochastic general equilibrium (DSGE) models. Using this method, I quantitatively find that in response to a national military buildup that affects regions differentially in a manner consistent with U.S. expenditure, labor reallocation amplifies the aggregate output effect of government spending by 30 percent relative to a model without it.

Chapter 3 proposes a new method to study the macroeconomic impact of microeconomic shocks. I show in what conditions and how to represent a disaggregate stochastic dynamic model into a recursive aggregate system by approximation order. This method provides a sufficient-statistic characterization of “macro state variables or shocks” in terms of heterogeneous micro shocks and structures. The first- and second-order macro shocks can be shaped by the average and the dispersion of the micro counterparts weighted by their micro impact intensities. I apply this method in several applications to illustrate the importance of micro heterogeneity and nonlinearity in macroeconomics. First, I provide a first-order decomposition of the aggregate consumption function when consumers have different marginal propensities to consume. A new redistribution channel—the asset position adjustment channel—is identified and the magnitude of this channel relies on the variability of capital investment and the persistence of shocks. Second, I show that permanent income inequality alters aggregate demand responses mainly at the second-order, instead of the first-order. Non-homothetic preferences and the dispersion of permanent income income jointly determine the impart on aggregate consumption response.