On the unavoidability of ‘unscientific’ judgment in estimating the cost of capital

Estimating the cost of capital using DCF methods requires forecasting dividends. This paper shows that past dividend behavior may support more than one reasonable forecasting model and that reasonable forecasting models can lead to divergent estimates of the cost of capital. We find no way of selecting the best cost of capital estimate. Any choice must ultimately be based on ‘judgment’ of the analyst.