Nonlinear price transmission in the rice market in Senegal: a model-based recursive partitioning approach

This paper analyzes the nonlinear effects in price transmission from international markets to the local rice market of Dakar (Senegal) focusing on asymmetries through threshold effects. We use recent machine learning methods (model-based recursive partitioning trees) to detect asymmetries in the price transmission mechanism. Using a model based recursive partitioning algorithm, we identify a threshold and confirm the asymmetry in the price transmission. Local retail prices are more sensitive to world price increases than to declines. Only 11.80% of positive deviations (international prices go down) are eliminated at the end of the subsequent month, while 39.50% of negative deviations (world prices go up) are eliminated after one month. These results highlight the role of transaction costs and the market power of commercial intermediaries in price transmission in the sense that margins are corrected more rapidly when they are squeezed relative to their long run level than when they are stretched. Our results are confirmed by the traditional Threshold Autoregressive (TAR) model.