UC Berkeley

In the introductory chapter of this dissertation, I motivate my use of disaggregate, cross-sectional variation for studying the effects of various policies on the aggregate economy. In particular, I argue that the topics empirically studied in this dissertation are historically unprecedented in nature and therefore necessitate a cross-sectional approach. I briefly review the anatomy of the linear least squares estimator to illustrate how standard cross-sectional designs can only recover the relative effects of policies. I then preview the theoretical tools I use in this dissertation to make the mapping between cross-sectional estimates and aggregate effects explicit.

In the second chapter of my dissertation, I generalize a textbook currency union model to incorporate trade in intermediates among labor markets in order to study the local, spillover, and aggregate effects of government spending. This theory provides a theoretical framework for interpreting the empirical, cross-sectional moments estimated in the subsequent chapter. The spillover effects of government spending mediated by trade in intermediates represents a novel and understudied mechanism by which local fiscal multiplier estimates likely represent a lower bound on the aggregate, Zero Lower Bound (ZLB) fiscal multiplier. In my framework, there is both a local and a spillover (relative) multiplier of government spending. Theoretically, summing both multipliers together yields an approximate lower bound on the aggregate, ZLB fiscal multiplier.

In the third chapter, I use geographic variation in government spending under the 2009 Recovery Act and import-export linkages between states from the 2007 Commodity Flow Survey to estimate a local relative multiplier of 1.46 and a spillover relative multiplier of 1.33. Adding both together yields an approximate lower bound on the aggregate, ZLB fiscal multiplier of 2.8, nearly doubling the lower bound implied by the local multiplier estimate alone. A sectoral decomposition of both estimated multipliers strongly corroborates the trade in intermediates spillover mechanism.

The final chapter studies the relative and aggregate economic effects of a prominent policy intervention implemented around the world during the COVID-19 pandemic by similarly interpreting causally identified, cross-sectional estimates through the lens of a currency union model. In particular, the high-frequency, decentralized implementation of Stay-at-Home orders in the U.S. is used to disentangle the labor market effects of SAH orders from the general economic disruption wrought by the COVID-19 pandemic. The analysis implies that each week of SAH exposure increased a state's weekly initial unemployment insurance (UI) claims by 1.9% of its employment level relative to other states. A back-of-the-envelope calculation implies that, of the 17 million UI claims between March 14 and April 4, only 4 million were attributable to SAH orders. A currency union model is developed to provide conditions for mapping this estimate to aggregate employment losses attributable to such orders.