Spotting the Predictive Dynamics of Cyclically‐Adjusted Financial Ratios in the US Stock Market

Abstract

Predicting returns is a timeless and important economic topic. Fundamental analysts see identifying time‐varying predictive factors as a key goal in asset pricing. This paper provides strong evidence of time‐varying return predictability of the S&P 500 Index from 1929 to 2020, by proposing the construction of altered versions of the classical dividend‐price (dp) and earnings‐price (ep). Our study's primary objectives are to introduce a cyclical adjustment to the simple dp and ep, which would smooth dividends and earnings, to compare the predictive dynamics of these new ratios with their simple counterparts and to modify all ratios based on long‐run equilibrium relationships that arise between the examined variables. We provide solid evidence that the cyclically‐adjusted ratios perform better than the simple predictors in all forecasting horizons, both in and out of sample, and the modified variables capture more predictive components in returns. Through certain robustness checks including multivariate regression settings and evidence on excess and real return predictability, we thoroughly present the predictive components identified in our data set.