Exploring the Predictive Power of Valuation Metrics on Subsequent Market Returns

Abstract

This article analyzes several valuation metrics’ predictive power on subsequent 1- and 10-year total returns across multiple periods and developed market countries. I evaluated the ratio of market capitalization to gross domestic product, dividend yield, dividend discount model, and Shiller’s CAPE for US data. Succinctly, this research has demonstrated predictive power at the 10-year horizon across countries and valuation metrics after adjusting for overlapping observations. We highlight that there is no strong predictive power across the entirety of the data set for countries with multiple centuries of data. In fact, there is a definitive structural shift in the data across countries within the 20th and 21st centuries. Detractors have focused on the recent failure of these models to predict subsequent returns, often citing that the shift from dividends to share repurchases has caused a structural shift in the data, and that lower interest rates warrant higher equity multiples. I acknowledge the recent failures, but I find these criticisms to be overstated.