Estimating Beta: Interpreting Regression Statistics

This chapter provides a discussion on beta estimation. For a publicly traded stock, beta is often estimated by regression analysis (ordinary least squares [OLS] regression). Understanding beta estimation requires an examination of the nature of beta and what it is intended to measure. There are two general ways for estimating betas: direct beta estimate (sometimes called a top‐down estimate) and a proxy beta estimate. All statistical analysis software and applications (Microsoft Excel add‐ins) generate tables of regression output values. These output values are segregated into three common tables: regression statistics table; ANOVA table; and regression coefficient table. The chapter addresses interpretation of the OLS data elements in the context of beta estimation by using TIBCO Software, Inc. (TIBX) example.