Equity Valuation: Models from Leading Investment Banks

24

Regression-based Valuation

Regressions can be a powerful valuation tool. They help identify and quantify factors that determine company valuation. Because regression lines minimize the distance from all X, Y observation points they generally split the universe of observations into two parts, that being above and beneath the line. This characteristic makes regression valuation inherently relative. Said differently, a regression will always identify some attractive and some unattractive companies. This is very different from DCF or intrinsic-based valuation methods, which can deem all examined companies as attractive or unattractive.

The challenge in properly using regressions for company valuation lies in identifying an independent variable that robustly explains observed valuations of differing companies. Most regression-based attempts to explain observed valuations, typically expressed as PE or EV multiples, use company characteristics such as growth or cost of capital as the independent (explanatory) variable. A classic example is the PEG ratio, which is calculated by dividing the company's observed PE with an estimate of its EPS growth rate. Other company characteristics or factors can be used, one of the more popular being ROIC/WACC. As we will explain, we believe that none of the factors, including ROIC/WACC, commonly used by either fundamental or quantitative investors is capable of robustly explaining observed valuations. For this reason we created a new metric, the Economic ...