An examination of the validity of industry class as a proxy for the equivalent risk class: total firm risk approach
Since the Modigliani and Miller's classic study in 1958, the concept of an industry class (IC) has been used widely to refer to a group of firms possessing an equivalent level of risk. However, efforts in previous studies to prove that IC is a realistic surrogate for an equivalent risk class (ERC) have not been conclusive.
This dissertation considers the limitations of previous research and presents evidence that toe typical measure of riskiness in an IC, business risk (BR), may not adequately explain the relationship between an IC and ERC. The primary purpose here is to investigate whether an alternative measure of risk, total firm risk (TFR), will reveal a more clear relationship between an IC and ERC than BR has done. TFR is composed of the three risk terms, BR, financial risk (FR), and miscellaneous risk (MR).
Six testable hypotheses, as well as one inferable hypothesis, were posed in two contexts, "within an industry," and "across industries." The six testable hypotheses were; 1) there is heterogeneity of 3R within an industry, 2) there is heterogeneity of BR across industries, 3) there is heterogeneity cf FR within an industry, 4) there is heterogeneity cf FR across industries, 5) there is homogeneity of TFR within an industry, and 6) there is heterogeneity cf TFR across industries. The last hypothesis concerned the comparative advantage of the use of TFR rather than BP in explaining the relationship between IC and SRC. The variables in the empirical specification or TFR were obtained from accounting and market returns of a common stock.
Several statistical methods were applied to make testing the hypotheses possible. The methods adopted here include autoregressive linear regression, analysis of variance, nonparametric tests, multivariate analysis of variance, multiple discriminant analysis, correlation tests, and F - test. These methods were applied to a sample consisting of 299 firms from 14 four-digit standard industrial classification (SIC ) industries.
The results of these tests supported strongly five hypotheses. But the fifth hypothesis concerning "homogeneity of TrR within an industry" was not supported strongly from the results of the test, and thus, the four-digit SIC groups cannot be considered as excellent proxies for equivalent risk classes on the basis of these results. However, compared to the findings of previous studies, these results show that TFR explains better the relationship between IC and Z2C, and thus, the inference of the last hypothesis is supported.