Thaís Machado de Matos Vilela.
26/03/2015
Orientador: Leonardo Rezende.
Banca: Juliano Assunção. Leonardo Rezende. Marcelo Medeiros. Helder Queiroz Pinto Junior. Claudio Ribeiro de Lucinda.
Three essays on gasoline and automobile markets in Brazil
Nível: Doutorado
This thesis is comprised of three chapters about gasoline and automobile markets in Brazil. In the first two chapters we are interested in the relationship between consumers’ behavior and the flex-fuel technology. In the first chapter, we focus on how sensitive Brazilian consumers are to fuel price changes, especially after the introduction of the flex-fuel technology on March 2003. We estimate the own- and cross-price elasticities of gasoline demand, combining two identification strategies: Instrumental Variables and Dynamic Ordinary Least Square. We combine different identification strategies to account for the non-stationarity of all series used in the model to estimate the demand for gasoline. We find that consumers respond more to fuel price changes than earlier studies suggest. To better capture the dynamic relationship between gasoline and ethanol markets, we estimate an Error Correction Model. Using a non-recursive identification strategy, we find that transitory unit shocks in the price of gasoline and in the price of ethanol have permanent effect on the demand for gasoline. In the second chapter, we attempt to analyze the importance of the flex-fuel technology for consumers in Brazil. We use the methodology proposed in BLP (1995). Based mainly on aggregate data and within an oligopolistic context with differentiated products – like the automobile industry – BLP (1995) use a discrete-choice model with random coefficients to estimate the parameters of the demand and the supply curves. To control for the price endogeneity in the demand curve, the authors proposed using the Generalized Method of Moments and the instruments are linear combinations of the automobile characteristics (except for the price). The results suggest that the flex-fuel technology is not statistically significant when we control for all other automobile characteristics, such as horsepower and air conditioning. The results also suggest that Brazilian consumers are not elastic to changes in automobile prices. Finally, in the third chapter, we calculate the economic and environmental costs of the price of gasoline stability policy implemented by the Federal Government through its majority position in Petrobras. Based on microeconomic theory, we calculate that the deadweight loss resulting from this policy equals R$ 17 billion from January 2002 to January 2013 (prices are in January 2010). When considering separately the effects of this policy on CO2 emissions and on the ethanol market, we find that the cost increases substantially.
