Cointegration and market integration: An application to the Indonesian rice market

This article suggests improvements in the use of regression analysis to measure spatial market integration. The procedure pioneered by Ravallion is still widespread but is valid only under certain conditions of exogeneity. The alternative offered here is an error‐correction mechanism which makes it possible to test for exogeneity as well as indicating the direction and strength of causality in price formation between markets. The method is illustrated with data on rice prices in different parts of the Indonesian market. The results confirm, among other things, that supply sources are more important than demand sources in driving prices.