A practical method to estimate the cost of equity capital for a firm using cluster analysis

Purpose – The purpose of this paper is to propose and test a method for selecting a portfolio of public firms which can be used for computing the cost of equity capital for a non-public firm or division of a firm. Design/methodology/approach – This method relies on cluster analysis and a large sample of firms. Using the accounting data from the firms, portfolios are formed that are expected to have a high degree of within-group similarity on measures of systematic risk, such as beta. The average beta of the group can be used as an estimate of the beta of the target firm within the group, which is then used to compute the cost of equity capital of the target. Findings – Cluster analysis shows significant potential to group together firms that are found to have high similarity in systematic risk. Cost of equity capital estimates made using the betas of the proxy portfolio firms are predictive of the measured betas of randomly selected target firms. Research limitations/implications – This research has shown significant predictive relationships between accounting data and market based cost of capital estimates, but the relationship is noisy, and the fit of the model could be improved with additional calibration. Originality/value – This approach uses well known data and methodology, but is unique in this application. This suggests many other potentially fruitful applications may be found for cluster analysis in financial research.