Multifactor return model based on interim financial statements

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Polytechnic Institute and State University

The purpose of this research is to examine the significance of a market factor, an industry factor, a company factor and a growth factor in explaining security returns. A secondary objective is to determine if interim financial statements--the balance sheet and the income statement--provide useful information in developing the return model.

Market-related and industry-related systematic risks are constructed as surrogate measurements for the market and industry factors. The company factor is composed of one accounting return measure (profitability) and five accounting risk measures (accounting beta, operating leverage, financial leverage, dividend covariability, and cash flow beta). These variables are included as individual regressors in the return model. Also, a company index (the first principal component) is constructed and tested for its significance in the four-factor return model. The compound growth rate in total assets measures the growth of individual companies. Quarterly accounting information is used to measure these company and growth variables, and their significance provides evidence supporting the usefulness of interim financial statements.

A multiple regression analysis is employed to develop the return model. In addition to the market factor, an industry factor, components of the company factor (dividend covariability and profitability), and a growth factor are found to contribute significantly to estimation of the return model. The use of a company index in lieu of individual company variables, however, is not recommended for· developing the return model. Additionally, results indicate that the market model provides the best surrogate measure of the market factor, and Line of Business information is recommended for classifying companies into industry groups.

Major limitations of the study are (i) a self-selection bias of companies for the sample; (ii) measurement errors in interim financial statement data due to accounting allocations; (iii) seasonality of quarterly accounting information; (iv) use of average regression statistics in determining the best return model; (v) a limited number of regression models examined; and (vi) multicollinearity. These may limit the generalizability of the findings beyond the sample data and the interpretation of relationship between security return and its potential determinants.