Simultaneous Equations System vis-aviz Classical Regression Estimators on Impacts of Some Economic Indicators in Nigeria

Abstract

The choice of models for analyzing the relationship between Total Debt Services (TDS), Real Interest Rate (RIR), Inflation Consumer Price (ICP) and Gross Domestic Product (GDP) variables using recursive models, though produces non-spurious results since the coefficient of determination (R²) is strictly less than the Durbin Watson (DW) statistic. However, the recursive model did not give a good fit for the relationship between TDS and GDP, since macroeconomic variables are usually prone to endogeneity, first order serial correlation, autocorrelation problems among others. This research work x-rays the TDS-GDP relationship on one hand and TDS versus GDP, RIR and ICP indices on the other hand in Nigeria. The TDS, GDP and RIR are   endogenous variables, while the ICP, TRA, and  are exogenous variables. Pre-tests analyses using time series datasets extracted from the repository of World Governance Index showed level stationary series I (0) and a causal relationship between TDS and GDP variables. Furthermore, the results from the estimation techniques showed that the Three-Stage Least Square (3SLS) and Seemingly Unrelated Regression (SUR) outperformed the Ordinary Least Squares (OLS) and Two-Stage Least Squares (2SLS) estimators   when applied to both the exactly and over-identified structural equations in the simultaneous equation system (SES).