Vol. 6, Issue 9, Part F (2020)

Abstract
Over the years, interest has been on estimating best regression models that minimizes error when several estimated models contain irrelevant independent variables. This study deals with analysis of macroeconomic predictive variables of gross domestic product. The data used for this study were obtained from a secondary source. A multiple regression model containing six independent variables of economic data were fitted. The variables included in the model are external debt, exchange rate, foreign direct investment, net export, consumption and debt payment. The data were analyzed through the use of stepwise regression with the help of statistical software (SPSS). The best model excluded both the debt service payment and the external for the single variable model and excludes only the debt service payment for the two variable models. Based on the analysis, consumption index and exchange rate influences the gross domestic product better; therefore, the use of stepwise regression is really essential in determining the predictive power of a regression model.