Effects of the fiscal policy on macroeconomic outcomes
The dissertation consists in three chapters addressing the effects of fiscal policy on macroeconomic outcomes. In the first chapter, I take advantage of the scarcity of evidence on the effects of public spending on GDP growth in developing countries. Thus, the main focus is in the top five emerging economies, Brazil, Russia, India, China and South Africa, known as BRICS, that accounts for almost a quarter of world GDP have received little attention. Through an Auto-regressive Panel-Vector (P-VAR) model, using quarterly data from 1997Q1 to 2017Q4, I estimate that the government spending multiplier is 0.145 percent on impact (t=0), and a cumulative multiplier of -0.125 percent percent in 5 years. Also, I analyze how the results of the baseline model vary according to specification adjustments, different order of the variables, and the use of alternative trend removal mechanisms. The robustness analysis shows that the multiplier is sensitive to these changes, which provides evidence of the possible causes of the varied results. In the second chapter, I estimate the local fiscal multiplier using the literature on incumbency status and local spending. The literature suggests that regional authorities seek to in uence their voters through increases in government spending during the election years. Given this, I build an instrument based on term limits and electoral laws established in the Legislatures of the 50 US states and the District of Columbia, to estimate the local fiscal multiplier for the period of 2007-2016. My estimate of the fiscal multiplier in the baseline model is 1.274. Through a robustness check, my estimator uctuates between 1.208 and 1.239. My results are consistent with the estimations in the geographic cross-sectional multipliers literature for the United States. In the third and last chapter, I evaluated the fiscal policy effects in a dollarized Latin American economy such as Ecuador. Therefore, I estimate both the linear and state-dependent impact and cumulative multiplier for the period between 1991Q1 and 2019Q4. The main result shows that the linear impact multiplier is 0.79 and fades in the medium and long term. Furthermore, in expansions and recessions, the impact multiplier is 1.42 and 0.91, respectively. However, the effect disappears in the long run when dealing with expansions. On the other hand, in periods of recession, the cumulative multiplier is 0.64, significantly different from zero at a 90 percent confidence interval. The results suggest that the multipliers are distinct between states at a 5 percent level of significance.