Essays on the macroeconomic effects of climate and energy price shocks

Abstract

This dissertation consists of three essays. The first essay estimates the difference in the response of non-gasoline consumption spending to gasoline expenditure shocks across recessions and expansions. Using a nonlinear structural VAR (SVAR) model estimated on quarterly U.S. data for the period 1973-2018, we find that shocks to gasoline expenditures are followed by a much larger reduction in consumption if the shock occurs during a recession than if it occurs in an expansion. We also find a big difference in the response of personal savings to a gasoline expenditure shock depending on whether the economy is in a recession or expansion. This points to precautionary saving behavior as a source of the asymmetry in the response of consumption. We conclude that consumption forecasts should account for this asymmetry. The second essay asks whether El Nino-Southern Oscillation (ENSO) has asymmetric impacts on U.S. food and agricultural stock prices. Using weekly data from 1990 through 2019, we find support for the hypothesis that food and agricultural stock prices respond asymmetrically to ENSO shocks. In particular, we provide evidence that El Nino shocks typically decrease or have no effect on U.S. food and agricultural stock prices, whereas La Nina shocks generally increase prices. Our results are robust to a variety of robustness checks. The third essay estimates the effect of the 1999-2007 oil price spike on U.S. manufacturing employment. Using local labor market data, we estimate the change in U.S. manufacturing employment that can be explained by oil price movements that can plausibly be considered exogenous with respect to the U.S. economy. Our estimates suggest that the oil price spike may have been responsible for the movement of more than two million workers out of the manufacturing sector. When controlling for the effect of the oil price spike on manufacturing employment, the effect of import competition from China found by Autor, Dorn, and Hansen (2013) falls by more than 20%, with the estimated effect on less-educated workers declining by 26%.