Spatial Panel Data Models and Fuel Demand in Brazil

Dr Gervasio Santos, Federal University of Bahia

The objective of this paper is to estimate the price and income elasticity for gasoline and ethanol in the fuel market for light vehicles in Brazil using spatial panel data models. The presence of two main fuels – gasoline and ethanol – in this market makes it considerably competitive. In addition, vehicular natural gas (VNG) also has a marginal presence in the market. Besides diversification, there are three main spatial features of fuel supply that might influence heterogeneity in the behavior of regional consumers. First, sugar-cane ethanol production is concentrated in some states. Second, besides marginal imports of gasoline, its production also is concentrated in some states. And third, along with the small share of VNG, production and logistical problems lead to a concentration of supply in a few states and cities. As a consequence, consumer behavior is heterogeneous and directly and indirectly influenced by different supply gaps and transportation and distribution costs. For this reason, there may be spatial autocorrelation patterns in fuel demand. Thus, spatial panel data models are used to estimate price and income elasticities of gasoline and ethanol, controlling individual and spatial heterogeneity in regional demand. Results showed that, aside strong evidence of the influence of spatial autocorrelation in the consumption of gasoline and ethanol, there also is considerable competition among these fuels, mainly due to the flex-fuel technology.

Comments for BIEE Members only.
Sign in or become a member today.

Sign up to our Events Newsletter

To receive email updates about our forthcoming events and news please sign up here.

Sign Up