Earnings Per Share Versus Cash Flow Per Share as Predictor of Dividend; Evidence from Non Financial Sector of Pakistan

Abstract

This research study has compared the relative predictive power of earnings and cash flow in order to find out the best predictor of dividend. This study has used panel data techniques and the data has been collected from the balance sheet analysis of 104 KSE-100 index non-financial companies. This study has employed a linear mixed effect model approach to investigate the main problem. The two models (earnings based model and cash flow based model) have been compared by using log likelihood estimator and Akaiki information criteria (AIC). The results have shown that cash flow per share is a better predictor of dividend than earnings per share in term of log likelihood estimator and Akaiki information criteria (AIC). The results have also concluded that both earnings per share and cash flow per share have a significant relationship with dividend. The study also finds out that all the control variables including firm’s size (SIZE), leverage ratio (LR), market to book value (MBV) and liquidity ratio (LIQ) have a significant relationship with dividend.

Keywords: Earning per share (EPS), Cash flow per share (CFPS), Dividend, Linear mixed effect model, Log likelihood estimator, Akaiki information Criteria (AIC) and KSE-100 index