摘要：This study uses a rich dataset of 85 market pairs between January 2000 and October 2008 for Kenya, Tanzanian and Uganda, the three largest member countries of the East Africa Community, to analyze the factors determining national and cross-national maize price transmission. Although the three countries are members of the community’s customs union and they each claim to pursue maize trade without borders, their agricultural trade policies still differ, thus affecting prices and trade flows to different extents. This analysis extends the existing border effects literature in three ways. First, it assesses the magnitude of price transmission, instead of analyzing trade flows or price variability. Second, distance is shown to have a significant impact on price transmission in the region and to be of nonlinear nature, which is modelled using a semiparametric partially linear model. Third, the border effect is found to be heterogeneous, that is, it matters which national border is crossed. A strongly negative effect of the Tanzanian-Kenyan border appears, while no significant effect for the crossing of the Ugandan-Kenyan border exists. These results are of high political relevance because they show that Tanzania represents a rather isolated and internally fragmented island within the East African maize markets. Bilateral maize trade with Nairobi appears to display substantially higher price transmission than with the rest of the markets, confirming its economic importance in the East African region and the structural maize deficit in Kenya.
关键词：border effect;spatial market integration;cointegration;semi-parametric regression;partially linear model;Eastern Africa;maize
