Bank stock valuation theories: do they explain prices based on theories?

Purpose The objective of this paper is to investigate the validity of stock valuation theories and their forecasting ability by conducting an empirical study. It employs four most commonly used theories which are then tested using 19-year banking-firm market data. The usefulness of these models demonstrates with promising results. Design/methodology/approach This paper conducts a multi-country study using the multi-model testing approach to evaluate validity of theories and forecast accuracy of banking firms. It employs four methodology models used in finance literature; (1) P/E multiples model, (2) accounting-information-based clean surplus model, (3) theoretical model based on Gordon and Shapiro (1956) method and (4) the Damodaran-Kottler Free Cash Flow or FCF theory based on discounting model. Findings The tests show that the four theories under tests have a significant fit with actual price formation. The explained variation ranges from 72 to 92%, so the explanatory power of the theories accounting for variations in bank prices over 19-year period is substantial. The models fit suggest that the P/E model has superior predictive power followed by the RIM, DDM and FCFE. These findings shed new lights on the relative performance of valuation models. Research limitations/implications The study is limited in terms of the sample period size for 1999-2019. The availability of essential financial data prior to 2000 is very limited, so one can understand interpretation of statistical results under certain assumptions. Practical implications The paper suggests that one-factor model is better than the two-factor model. Originality/value The work done in this paper is unpublished and original contribution to banking and finance literature and also not under consideration for publication in any other journal.

Automated summary

The paper investigates the validity of stock valuation theories and their forecasting ability through an empirical study using 19-year banking-firm market data. The findings suggest that the tested theories have significant fit with actual price formation, with the P/E model showing superior predictive power among the four models tested.