Analysis of relationships between production of yam [Dioscorea spp.] and its determinants in Nigeria: a cointegration approach

The associations between the production and some of the variously-debated factors that have the tendency to influence it, including cultivated land area, yield, and climatic conditions, were investigated for yams (Discora Roundata), a leading tuber crop in Nigeria. Secondary data collected from two reliable sources covering the 1961-2008 periods were used for the study. Data were analyzed using advanced econometric cointegration technique. Results of the unit roots tests
revealed that production, land area, and yield of yams were each integrated of order one, I(1), compelling their use for cointegration analysis. However, the two included climatic factors, rainfall and temperature integrated only at levels, I(0), making them unsuitable for use in cointegration tests and were consequently dropped. Results showed further that cointegration existed between production and land area on the one hand and yield on the other. The resultant trace- and maximum eigenvalue statistics were 29.83 (p<0.01) and 29.22 (p<0.01) respectively for test involving production and land area, and 29.85 (p<0.01) and 29.23 (p<0.1) respectively for test of production with yield. The Bivariate Granger causality tests could not reveal any causality either from land area to production or vise versa. Also, although causality could not be revealed from yield to production, an inverse, but weak, causality (F=2.83; p<0.05) was observed from production to yield. The implication of the finding is that past values of cultivated land area and yield could not be used as reliable indicators for predicting the future values of yam production in Nigeria. A lot more other factors had contributed to past fluctuations in yam production and there was need to adopt an across-the-board approach in the development of the yams sector in Nigeria.