Analysing the information embedded in the optimal mean–variance weights: CAPM versus Bamberg and Dorfleitner model

This paper is centred on the analysis of the information embedded in the optimal weights of the assets in the CAPM and the Bamberg–Dorfleitner model. On this basis, first we find a functional relationship between the optimal weights of both models. Next, we find a set of performance indicators that express the contribution of each asset to the reward/volatility ratio measured as the Sharpe ratio or through a utility function. For the Bamberg–Dorfleitner model these indicators also lead to identify the contribution of each independent variable to the reward/volatility ratio. Technically, these connections are obtained through the covariance-normalized portfolio that consists of a transformation of the inverted covariance matrix. The additive property of covariances is transmitted to the indicators. These results enable investors and portfolio managers to obtain a precise knowledge of the causes of the value of the reward/volatility ratio. From the corporate point of view, this approach contributes to a better identification of the features of the different types of investors to whom to focus the corporate financial policy.