The effect of dividend policy on stock price: new evidence from the Box-Cox models

Abstract

This study re-examines the relationship between dividend policy and stock price through the lens of the Box–Cox transformation. Using panel data of 24,422 U.S. firms from 1992 to 2021, we compare the performance of ordinary least squares, log-linear, and Box–Cox functional forms. Consistent with the dividend relevance hypothesis, both dividends and retained earnings exhibit positive and significant effects on stock prices. The Box–Cox specification, with an estimated transformation parameter of 0.1299, outperforms conventional models under various fixed-effect and random-effect estimations, demonstrating better goodness-of-fit and robustness to heteroskedasticity. Our findings highlight that accounting for nonlinearities and firm-level heterogeneity substantially improves the empirical modeling of the dividend–price relationship. The results provide new evidence that the Box–Cox transformation offers a more flexible and accurate econometric framework for analyzing dividend policy and firm valuation dynamics.