keywords: Vector Autoregressive, Multiple regression analysis, Forecasting, Forecast evaluation, Macroeconomic variables.
This study was used to determine the most suitable model for analysing inter and linear relationships between some macroeconomic variables as the nature of the model do have implications on the forecasted values and accuracy. Then, this study was used to compare the forecast and forecast evaluations of Vector autoregressive (VAR) and Multiple regression analysis models using some Nigerian macroeconomic variables. From the results, the Augmented Dickey-Fuller test showed all series were stationary at the first differencing, I(1) and only inflation rate was stationary at the ordinary level, I(0). Both models revealed the existence of inter and linear relationships but VAR models were better based on the values of the coefficient of determination. The out-sample forecast for both models indicated that Government revenue and Government expenditure exhibited a continuous rise while Inflation rate, Exchange rate and price of crude oil fluctuate on a yearly basis. The forecast evaluations results based on Root mean square forecast error (RMSE), Mean absolute error (MAE) and Mean absolute percentage error (MAPE) showed that the out-sample forecast for Vector autoregressive models was better and this indicated that nature of the model is important when analysing the relationship between macroeconomic variables.