Evaluating the Implied Cost of Capital Estimates
Characterizing a firm's true (but unobservable) expected returns as the normative benchmark, we develop a two-dimensional framework for evaluating the relative performance of implied cost-of-capital (ICC) estimates. First, in time-series, variations in ICC estimates should reflect changes in true expected returns rather than changes in measurement errors. Second, cross-sectionally, ICC estimates should predict future realized returns. Using this framework, we compare seven alternative ICC measures and show that several perform quite well along both dimensions, and all do much better than Beta-based estimates. In addition, we provide evidence on the importance of appropriate matching between the earnings forecasting method (analyst vs. mechanical) and the valuation model. Overall, our evidence provides significant support for the broader adoption of ICCs as firm-level expected return proxies.
Keywords: implied cost of capital;
Cost of Capital;
Forecasting and Prediction;