Modelling Nigeria’s Domestic Output from select Macroeconomic Variables using Variable Coefficient Functions

Abstract

This research dwelt on the application of variable coefficient models on Nigeria Domestic Output (proxy by GDP) taking exchange and inflation rate as two selected macroeconomic variables affecting domestic output in Nigeria. Secondary data spanning the period 1999-2019 monthly as extracted from the Central Bank of Nigeria (CBN) statistical bulletin was used in modelling the economic growth. The natural splines model was first fitted to the data to show the significance of the parameter estimates using R software as findings revealed a significant impact of the duo of exchange and inflation rates on economic growth as shown from the F-statistic p-value < 0.05 level of significance. Basis functions such as natural cubic splines and generalized additive models such as smoothing splines and local regression models were also fitted to the data. Generalized Cross-validation and leave-one-out re-sampling techniques were adopted in determining the tuning parameter, effective degrees of freedom and the span for smooth splines and local regression additive models. The model's accuracy was compared using the residual deviances for the GAMs and the Local Regression model was selected through a nested analysis of deviance on the two postulated models. Findings also indicated from the local regression model that the exchange rate and interaction of exchange and inflation rates cause instability in GDP growth in Nigeria.