Comparing the reliability of the intrinsic value estimates derived from three flows-based models and a multiples-based model

Abstract

Different equity valuation models in theory provide identical estimates of equity intrinsic values if they
are implemented with consistent assumptions. In practice, it is not unusual to obtain different value estimates
from different models. Prior research on the comparison between multiples-based valuation model and flowsbased valuation models are quite limited. This research contributes to the literature by comparing the accuracy
and reliability of three flows-based valuation models (discount dividend model, discounted free cash flow model,
residual income valuation model) and multiples-based valuation model using one-year forward forecasted
earning (PE1). For the analysis, all models are tested for bias, accuracy and explainability using t-tests, Wilcoxon
tests and OLS regressions with a large sample of 40.547 observations from U.S public firms during the period
from 2010 to 2020. Results show that PE1 value estimates are more accurate and reliable than that of flows-based
models; and among flows-based models, residual income valuation model performs best. This research provides
evidence for the performance and usage of such models to support sell-side equity analysts while making
decisions of which model to use.