On the Adequacy Model for the Relationship Between Commodity Terms of Trade and Money Supply in Nigeria

Abstract

This research determined an adequate model for the relationship between commodity terms of trade and money supply in Nigeria. The Kwaitkowski Phillips-Schmidt-Shin (KPSS) unit root test was used to establish whether the commodity terms of trade variables for the three sets of data were stationary. Three models were built based on commodity terms of trade’ stationarity. The first Model is the actual series before transformation.  The second Model is the log model and the third model is the square root model. All models contained the variables, Live Animals, Vegetable Products, Prepared Foodstuffs & Beverages, Mineral Products, Chemical & Allied Industries Products, Plastic and Rubber and Cement Products, Leather Products, Wood Products, Paper Making Material Products, Textiles Products, Footwears & Headgears, Ceramics products, Precious Metals Products, Machinery and Chemical Appliances, Vehicles & Parts Products,  Base Metals and Miscellaneous Manufactured Articles. The ordinary least square method was used to estimate the parameters and the actual model has six (6) significant parameters, R² = 99.1%, AIC = 347.66, BIC = 350.96 and DW = 1.60, the square root model has twelve (12) significant parameters, R² = 99.6%, AIC = 116.40, BIC = 119.90 and DW = 2.496, and the log model has one (1) significant parameter, R² = 99.9%, AIC = 25.57, BIC = 28.87 and DW = 2.07. The result showed that the square root model which has the highest significant parameters, (x₁ - Live Animals, x₂ - Vegetable Products, x₃ - Prepared Foodstuffs & Beverages, x₅ - Chemical & Allied Industries Products, x₆ - Plastic and Rubber and Cement Products, x₇ - Leather Products, x₈ - Wood Products, x₁₀ - Textiles Products, x₁₃ - Precious Metals Products, x₁₆ - Vehicles & Parts Products, and x₁₇ - Miscellaneous Manufactured Articles), is the adequate model for identifying the relationship between money supply (MS) and commodity terms of trade (CTOT) since its independent variables have more significant parameters at 1%, 5% and 10% levels. This result contradicts the more works carried out using only logarithm transformations when the variables are large values. It was recommended that more detailed work be carried out using other transformation techniques (i.e another power transformation techniques) to identify a more suitable model to be used for the study or prediction of the variables.