On the Relationship between Systematic Risk and the Degrees of Operating and Financial Leverage

In order to use the capital asset pricing model (CAPM) to make operating and financial decisions, financial managers must confront the problem of estimating a security's systematic risk, or what commonly is termed beta (3). One approach to this problem is to estimate / by regressing the time series of a security's realized financial return (dividend plus capital gain or loss) on the contemporaneous realized financial return on a market portfolio. Such a procedure is suspect for three reasons. First, there is evidence that ,Ss of individual securities are not stationary 3, 4, 7, 9. Second, this procedure cannot be undertaken when historical information is not readily available, as might be the case in evaluating new product decisions. Third, it masks the important fact that firms make decisions about how to operate in the factor and product markets of the real economic sector. As Myers 20 points out, the actions taken as a result of these decisions generate a real return composed of an immediate cash flow plus any change in the present value of future investment opportunities. The regression procedure, though, uses the realized financial returns generated in the financial sector of the economy to calculate /. It therefore provides no knowledge of the real determinants of/ from the underlying characteristics of the real assets. Moreover, financial managers cannot look to the CAPM to find clues for estimating ,B on the basis of real variables. As a theory of financial market behavior, the CAPM states only a necessary equilibrium relationship between the prices of securities given their stochastic characteristics over a period of time. It says little about how stock prices are determined by the real variables that financial managers must consider in evaluating strategic, operating, and financial alternatives. The studies that seek to provide financial managers with knowledge of the real determinants of systematic risk are numerous. There are those that empirically test intuitive specifications between real variables and / [2,