摘要:Maize is a major staple food in Sub-Saharan Africa. Monthly maize prices in Tanzania are analyzed since the country is an important maize producer and exporter in East Africa. We analyze price transmission between the five most important urban regions of Tanzania between 2000 and 2008 which correspond to major maize production or consumption areas. We propose a novel method for the analysis. The semiparametric vector error-correction model allows the partial impact of the past deviations from price equilibria on current price changes to be potentially nonlinear. The nonparametric estimates of these partial influences suggest that they can be adequately modeled by linear functions.
关键词:cointegration;maize;nonlinear time series model;price transmission;semiparametric model;Tanzania