Modeling the Effect of Monetary Policy on Commercial Bank’s Lending Operation in Nigeria using Vector Error Correction Model (1991- 2020)

Abstract

The study examines the effect of monetary policy on commercial bank’s lending operation in Nigeria between February, 1991 to October, 2020 using the macroeconomic time series variables of exchange rate, interest rate, maximum lending rate and prime lending rate. The vector error correction model was employed to analyze these interactions as well as the effect and pattern of causality among the variables under investigation. Monthly data spanning from February, 1991 to October, 2020 which covered a period of 29 years and 7 months (357 observations) were sourced from the Statistical Bulletin of the Central Bank of Nigeria (CBN). Preliminary statistical approach such as descriptive statistics and time plot were carried out to test if the data set obey the normality assumption and to verify if there is trend on the series. Pre- estimation diagnosis such as unit root test, lag order selection criteria and co-integration test were carried out and the results shows that at level, all variables had unit root, then, at first difference all variables were stationary. The lag order selection criteria chose lag 3 (14.7844*) of Akaike information criteria, but the vector error correction model (VECM) was done at lag 2 indicating losing a lag. The co-integration test shows the presence of long run relationship among the variables. The vector error correction model (VECM) estimated from the results obtained shows that all the variables had positive effect on commercial bank’s lending operation. The post estimation test on vector error correction model such as normality of the residuals, serial correlation and heteroscedasticity shows that the VEC model was multivariate normal, no serial correlation and homoscedastic. The inverse root of AR characteristic polynomial shows that the estimated VECM satisfy the stability condition of the diagnostic test. The variance decomposition test shows that the variables have a very weak influence in predicting one another. R

Keywords

References