Using Logistic Regression to Predict the Direction of Nigeria Stock Market Returns (Based on some Selected Sector Indexes, 2015-2022)

Abstract

The direction of the Nigeria Stock performance based on some selected sector indexes was considered, with data of weekly Nigeria Exchange Limited (NGX) indexes (which included closing prices of the NGX All Share Index, Banking, Oil and Gas, Insurance, Consumer Goods and Top 30 Companies for the period, 5th January, 2015 to 31st October, 2022. After transforming the original series to returns, the data were analyzed using correlation analysis and multiple logistic regression analysis. The correlation analysis was carried out to examine the correlations between the NGX ASI returns and each of the selected sector indexes returns, and the result showed that NSE ASI returns were correlated with all the selected sector indexes returns; although Oil and Gas and Insurance showed weak correlations. Logistic regression was used to model the dependent variable, direction (either upward or downward) of NGX ASI returns and independent variables (which are Banking, Oil and Gas, Insurance, Consumer Goods and Top 30 Companies), sector indexes returns of NGX. The findings showed that NGX Banking, Consumer Goods and Top 30 Companies returns contributed significantly to the prediction of the dependent variable, the direction of NGX ASI returns; while Oil and Gas and Insurance were not significant. It was concluded that although the NGX ASI returns were correlated with the five selected sector indexes returns used in this study, only three sectors (Banking, Consumer Goods and NGX 30) were good enough to be used in predicting the direction of the Nigeria Stock Market returns with the overall accuracy of about 70.17 percent. Thus, logistic regression model could be used by investors (individuals or institutions) to enhance their ability to predict the upward or downward direction of NGX ASI.