Multivariate Analysis and Modeling the Effect of the Gdp of Nigeria on the Petroleum Product Prices (1987-2018)

Abstract

The study presented multivariate analysis and modeling of the effect of the GDP of Nigeria on the Nigerian petroleum product prices from 1987 to 2018. The petroleum products considered as the response variables were the Premium Motor Spirit (PMS(Y1)), Automotive Gas Oil (AGO(Y2)) and Dual Purpose Kerosene (DPK(Y3)) while the predictors were GDP(Z1), Total Reserve(Z2), External Debt(Z3), Gross National Expenditure(Z4) and GDP/Capita(Z5). These predictors were studied in pairs on the responses and also studied jointly with all the five predictors on the responses. Comparisons were made among the pairs, also, each pair was compared with the joint analysis. SPSS software was used in the analysis in which Pillai's Trace, Wilks' Lambda, F-value, P-value, coefficient of determination and sum of square errors were applied to determine the contributions of each predictor variable in the models built, to the petroleum product prices. Correlation and covariance analysis were also applied to know the joint effects of the variables. It was observed that PMS was greatly affected by the economic policies of Nigeria, same to the AGO and then DPK. PMS is insignificantly impacted in an economy with two indicators where GNE is involved. PMS and AGO proved better than DPK in the economy of Nigeria. The relationship between GDP on Total reserve or External debt is positive. That is to say, any increase in these variables will result to an increase in the petroleum product prices. Correlation and Covariance analysis revealed that the analysis between GNE and External debt proved to be the worst pair. The analysis on all the five predictors, GNE and External Debt, Total Reserve and External Debt, Total Reserve and GNE and Total Reserve and GDP had no negative correlation, while GDP and GNE had negative correlations between AGO and DPK and AGO and PMS, also, GDP and GDP/Capita and GDP and Total Reserve recorded negative correlation respectively between PMS and AGO and AGO and PMS.