Which Factors Are Over-Owned? Or, Supply and Demand: A Possible Roadmap to Solving the Factor Timing Problem

Abstract

Recent difficulties with certain factor models have increased interest in finding methods to “time” factor investing better. So far, however, the consensus is that factor timing is difficult. As inspiration for a possible solution, this paper reviews one of the best long-term return prediction models for the S&P 500 – the level of equity ownership in investor portfolios – which handily outperforms commonly-cited valuation-based forecast methods by relying on the more fundamental dynamics of supply and demand. Indeed, it has been called the “greatest predictor of future stock market returns” you’ve (probably) never heard of! For example, it can explain the earnings-less bull market of the 1980s, and overcomes the negative-PE problem of the 2008 Financial Crisis for which the traditional methods masked a good buying opportunity.

The first section of the paper recreates and compares the various long-term S&P 500 forecasting methods, using our own robust fitting procedures. This paper then suggests a roadmap for applying this methodology to factor forecasting, since it is already known that standard valuation (or other) methods are not good estimators of factor performance.

Keywords: factor premium, factor timing, factor investing, styles, CAPE, PE ratio, valuation, forecasting

JEL Classification: D46, G11, G12