Food Prices Transmission In Rwanda: Econometric Analysis

Abstract

This paper analyses the food prices transmission using econometric techniques where vector autoregressive models were formulated. Price transmission means the change of one price caused by the change of another price. Spatial price transmission is used in this study and price transmission is affected by different factors; among them we have transport and transaction costs and market power. Monthly average prices data of potatoes and banana collected from Musanze, Huye, Kibungo, Muhanga and Nyabugogo markets from January 2002 to July 2013 are used. Potatoes are mainly produced in Musanze District and Banana in Ngoma District where the Kibungo market is located. Analysis was conducted in two ways, first concerns relationship between retail prices of potatoes in Musanze market and retail prices of the other markets and second on relationship between retail prices of banana in Kibungo market and retail prices of banana of other markets. Graphical analysis shows that there is a positive trend for all retail prices. Retail prices of all markets are not stationary at levels but the first differences are stationary. This implies that series are integrated of order one [I(1)]. Engle and Granger co-integration tests were done and the results show that the variables are co-integrated. This implies that there is a long run relationship. Results of Error Correction Model (ECM) show that the coefficient of error term is significant in all models. This implies that there is a short-run relationship. Granger causality tests confirm the existence of causality between analyzed variables. These results provide useful information for forecasting the retail prices of potatoes and banana in different markets.

Keywords: Price transmission, Granger Causality; Co-integration; Vector Autoregression (VAR) and Error Correction Model (ECM).