Essays on Government Spending and Employment

Abstract

In Chapter 1, I examine the "twin deficits" hypothesis in a new light by pooling data from highly developed open economies to create a representative global economy. Previous work on twin deficits has yielded mixed evidence supporting and refuting this theoretical explanation for the comovement of government budget and current account deficits. I estimate a vector autoregression (VAR) model with panel data using Bayesian methods to determine the effects of shocks to government spending on current account balances and real interest rates using various groups of nations and time periods. The results indicate that increases in government spending do not lead to increases in the current account deficit. My findings do not support the predictions of the twin deficits hypothesis. Motivated by the results in Chapter 1, Chapter 2 investigates the effects of government spending shocks on output, consumption, employment, and interest rates. To address empirical challenges related to fiscal policy lags, data availability, and simultaneity, Chapter 2 uses the narrative approach to record announcements of changes in military, national security, and counterterrorism spending in the form of a defense and civil defense news variable called defnews. This variable captures variation in defense and civil defense spending since September 11, 2001, a period that provides a natural experiment setting in which frequent and meaningful changes in defense and civil defense spending occurred independently of global and national business cycles. I implement a unique and extensive data gathering process to combine data from six nations and use a panel VAR model with Bayesian estimation methods to show that defnews correctly identifies changes in total government expenditures. The paper proposes a new set of stylized facts describing the effects of changes in government purchases that support the predictions of the neoclassical model and describe a value of the fiscal multiplier that is non-positive and close to zero. Chapter 3 investigates changes in the relationship between unemployment duration and unemployment rates in the United States. Although unemployment rates have trended downward in the United States since the 1970s, mean unemployment duration has risen. Over this same period, the dynamics of female labor supply have changed dramatically: large numbers of women have entered the labor force, women's wages have risen, and wage elasticities have fallen. I use Current Population Survey data to match individuals to their spouses and look at how family labor supply (specifically, spouse wages and spousal employment) has affected unemployment duration. This paper is coauthored with Professor Donna Ginther from the University of Kansas and Melinda Pitts from the Federal Reserve Bank of Atlanta.