Prediction of stock performance by using logistic regression model: evidence from Pakistan Stock Exchange (PSX)

TLDR: In this article, the authors used logistic regression model to predict stock performance and found that the model was 89.77 percent accurate for prediction good as well as bad performance of stock.

Abstract: The key purpose behind the study is to use logistic regression model to predict stock performance. For this purpose different financial and accounting ratios were used as independent variables and stock performance (either “good” or “poor”) as dependent variable. The result shows that financial and accounting ratios significantly predict the stock performance. Our study consists on the sample period of annual data from 2011-2015 and comprises of 109 listed non-financial firms of Pakistan’s Stock Exchange (PSX). Our sample was shortlisted on the basis of available data of Market Capitalization. Our research examines sales growth, debt to equity ratio, book to price ratio, earning per share, return on equity and current ratio for the prediction of stock performance. The findings indicate that our prediction was 89.77 percent accurate for prediction good as well as bad performance of stock. Although we did not consider macroeconomic variable to forecast stock return performance but our six firm specific accounting and financial ratios were good enough to predict stock performance. This study shows that Logistic regression model can be used by investors, individual as well as institutions or fund managers to enhance their ability to predict “good or poor” stock.