Country and border effects in the transmission of maize prices in Eastern Africa: evidence from a semi-parametric regression model

This study uses a rich dataset of 85 market pairs between January 2000 andOctober 2008 for Kenya, Tanzanian and Uganda, the three largest member countries ofthe East Africa Community, to analyze the factors determining national and cross-nationalmaize price transmission. Although the three countries are members of thecommunity’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 todifferent extents. This analysis extends the existing border effects literature in threeways. First, it assesses the magnitude of price transmission, instead of analyzing tradeflows or price variability. Second, distance is shown to have a significant impact on pricetransmission in the region and to be of nonlinear nature, which is modelled using a semiparametricpartially 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 theTanzanian-Kenyan border appears, while no significant effect for the crossing of theUgandan-Kenyan border exists. These results are of high political relevance because theyshow that Tanzania represents a rather isolated and internally fragmented island withinthe East African maize markets. Bilateral maize trade with Nairobi appears to displaysubstantially higher price transmission than with the rest of the markets, confirming itseconomic importance in the East African region and the structural maize deficit inKenya.