We employ a pair-wise approach to analyse regional integration in the gasoline and ethanol markets in Brazil. Using weekly price data for these two fuels at the state level over a period of almost ten years, we find that more than half of the fuel price differentials are stationary, which reveals the importance of allowing for spatial considerations when testing for market integration. We also find that the speed at which prices converge to the long-run equilibrium depends upon the distance between states, the differential in sugarcane mills density between states, and the similarity between tax regimes. Other demand and supply factors such as population density, gas stations density, sugarcane mills density and GDP per capita are not statistically significant.
