Estimating individual company costs of capital from cross‐sectional data: A random coefficients approach

Abstract

The purpose of this study is to extend the empirical analysis of Thompson (1979) to generate individual company costs of equity capital from statistical analysis of cross‐sectional industry data. The paper uses a new random coefficients regression approach on 18 cross‐sections of 75 electric utilities for the years 1963–1980. The results show rank stability between years and plots of individual company costs of average and simple facts about companies. The procedure does not allow calculation of standard errors of the estimates but when compared with existing cost of capital methods is shown to be a useful addition to the methodology of estimation.

On the estimation of cost of capital and its reliability

“…Makhija and Thompson (1984) have compared the various cross-sectional models using this method regards to their efficiency as a tool for rate cases. Thompson (1984) used the same technique along with cross-sectional data to estimate the cost of capital for individual utilities, but measures of reliability of the estimates were not obtained. To cope with the "reliability" question, Thompson (1985Thompson ( , 1987 combined the "dividend yield plus growth" method with Box-Jenkins time series analysis of past dividend experience to estimate the cost of capital and its "reliability" for individual Þrms.…”