Nigeria’s economic growth is a tremendous concern to policymakers, economics and scholars because of the inherent challenges of economic characteristics, consequences and contradictions. The Central Bank of Nigeria has implemented several policies such as tightening of monetary policy rate and heavy borrowing for infrastructural development to stimulate economic growth in the past few years. The prediction of economic growth from economic variables poses a challenge due to collinearity among the economic variables. In this study, we predict Nigerian economic growth using internal and external debt, interest and exchange rate as well as trade openness in the presence of multicollinearity and outliers. We employed the non-cross-validated and cross-validated partial least square regression method to quarterly data from 1986 to 2021 to predict Nigerian economic growth to simultaneously address the aforementioned problems among the predictor variables. Exploratory data analysis (EDA) and diagnostic test carried out ascertained the presence of multicollinearity and outlier. The fitted non-cross validation partial least square regression model extracted 5 components. The extracted 5 components, namely; 1, 2, 3, 4, and 5 influenced the growth of the Nigeria economy by 28.2%, -11.3%, 38.8%, 10.0% and 50.8%, respectively. The fitted cross-validated partial least square regression model extracted 2 components, namely 1 and 2, which were efficient and optimum for predicting economic growth in Nigeria. The extracted 2 components explained 100% total variation of the predictor variables. The biplot revealed that all the economic growth drivers were concentrated in the region of the two extracted components with positive contributions to the components that efficiently predict the economic growth (RGDP). Results of the variable importance projection showed that the economic growth in Nigeria depends largely on internal borrowing and economic openness. Therefore, it is concluded that the closure of all the international borders during this period affected the openness of the economy for exportation and importation activities and, as such, lowered international patronage, which hindered the growth of the economy. Hence, the need for policymakers to focus on policy formulation and implementation on the direction of internal borrowing and economic openness as essential drivers of economic growth. Also, to the researchers, a cross- validated partial least square method of selecting components required for predicting the economy’s growth was the most efficient technique for dealing with multicollinearity and outliers’ problems in the data set.
Digital Object Identifier (DOI)
Ebiwonjumi, A.; Chifurira, R.; and Chinhamu, K.
"New Perspectives on Predicting Economic Growth in the Presence of Multicollinearity and Outliers,"
Applied Mathematics & Information Sciences: Vol. 17:
4, Article 8.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol17/iss4/8