THE IMPACT OF USING ROBUST MM-ESTIMATOR IN REGRESSION MODELS: AN APPLICATION OF NIGERIAN INFLATION ON THE ECONOMIC GROWTH

Abstract

This study was conducted to show the impact of using a Robust MM-Estimator in Econometric regression models, as an alternative to the Ordinary least squares(OLS) technique for analyzing the effects of inflation on the economic growth of Nigeria from 1980-2019, due to the inefficiency of the OLS in the presence of outliers. The study used annual time series data obtained from the Central Bank of Nigerian statistical bulletin. The study also reviewed the performance of OLS and compared it with that of the MM-estimator. The study used GDP as the dependent variable and also as a proxy for economic growth while Inflation rate, exchange rate and interest rate were used as the independent variables. The findings of the study showed that MM-estimation gives a better result than the OLS having the smallest RSE (0.106) and larger R-square (0.78) than the OLS with RSE (0.1205) and R-square (0.77). The result of the MM-estimation model showed that inflation has a negative (-0.07551) impact on the economic growth of Nigeria, while the control variables exchange rate and interest rate have a positive impact on the economy. The study also recommended that to tackle the problem of outlier contamination affecting econometric models, particularly on Inflation-Growth Nexus analysis, is to adopt the use of the MM-Estimation method as an alternative to the traditionally used OLS technique, because it can cope with the presence of outliers in the economic data vertically and horizontally. Likewise, it will also correct the violation of classical assumptions automatically.

Keywords: MM-estimation, OLS, Breakdown point, Efficiency, Inflation, Economic Growth.