MEASURING VALUATION UNCERTAINTY: A PCA APPROACH

Abstract

Determining which companies are more difficult to value is a topic of significant interest in finance. Prior studies have employed many univariate proxies for valuation uncertainty to classify firms into high- and low-valuation uncertainty groups. In this study, I employ principal component analysis (PCA), a dimensionality reduction technique, using 11 valuation uncertainty proxies to extract the valuation uncertainty latent factor contained in the first principal component, which is proposed as a new measure of a firm’s valuation uncertainty, and show that it accurately captures a firm’s valuation uncertainty. I contend that using a PCA-derived valuation uncertainty index offers two benefits. First, integrating multiple valuation uncertainty proxies into a single metric improves our ability to quantify a firm’s valuation uncertainty. Second, it can assist in identifying the proxies that are most useful in measuring a firm’s valuation uncertainty.