Maximizing Returns Through Value Investing

Key Points:

The performance drivers of value investing are not unanimously agreed upon by investors and academics.
Performance drivers in finance are never crystal clear due to ever-changing markets and imperfect replicability of historical trends.
A performance driver in value investing is risk sentiment, as investors tend to prefer lower-risk investments during economic downturns.
Three variables to measure risk sentiment in value investing are stock market volatility, stock market skewness, and the yield curve.
Value factors generate positive returns when volatility decreases, skewness is positive, and the yield curve is upward sloping.
A multi-metric risk management framework can be used to time the value factor and avoid significant drawdowns.
The more filters applied to the framework, the more consistent the returns, but with lower exposure to the factor.
The framework can also be used to potentially short the value factor in negative market environments.
Implementation of the framework requires further testing and adjustments, but the results provide some comfort in their significance.

Introduction

Ted Theodore first wrote about value versus momentum stocks way back in 1984, but almost 40 years later, there still is no real consensus among investors or academics on what is driving either strategy.

That’s not due to a lack of research. Thousands of papers have scrutinized equity factors across markets and asset classes, and some have analyzed strategies going back more than 200 years.

Part of the problem is that performance drivers have been identified but lack widespread acceptance from practitioners. That’s understandable. If what’s driving a strategy’s returns is crystal clear, fund managers will be out of work when the environment for their investment style turns unfavorable. They’re better off remaining publicly vague about performance drivers, as that helps to retain their assets under management (AUM).

A second issue is that performance drivers are never crystal clear. Finance is not a hard science with immutable, gravity-like laws. Markets change continuously and historical performance and trends are not perfectly replicable. So when it comes to performance drivers, finance practitioners must live with relatively low standards of proof.

Our framework for determining a performance driver consists of four criteria:

It should have a sound economic basis.
It should work on average, but not all the time.
It should be implementable.
It should hold when tested across time, markets, and asset classes.

So what is the value factor’s key performance driver? On what evidence do we base that determination?

What Is Driving the Value Factor?

The value factor generates positive returns when cheap stocks outperform expensive ones. So when does that happen?

Cheap companies tend to be troubled companies. Otherwise, they wouldn’t trade at low valuations. Their issues might be temporal or structural: an overleveraged balance sheet or being part of an industry in decline, for example. Either way, investors will find these stocks uncomfortable to hold since the associated news flow and broker ratings will tend to be downbeat.

That means investors are most likely to risk buying questionable companies when they’re more confident about the economy and the stock markets. When the economy is heading into recession, investors tend to prefer companies with quality or growth characteristics. Put another way: Risk sentiment is the primary performance driver of the value factor.

There are many variables with which to measure risk sentiment. We focus on three: stock market volatility, stock market skewness, and the yield curve.

The Value Factor and Realized Stock Market Volatility

We constructed a value factor out of the cheapest and most expensive 10% of stocks in the US stock market as measured by price-to-book ratios using data from the Kenneth R. French Data Library. We then calculated the Z-score of stock market volatility using a three-month lookback.

Most of the value factor’s positive returns from 1926 to 2020 can be attributed to decreasing volatility. This relationship is not perfect, however: Between 1931 and 1943, value factor returns fell amid rising volatility. But from then on, returns were consistently negative when volatility was on the upswing.

The Value Factor and Stock Market Skewness

We next analyzed value factor returns in the context of stock market skewness, which we calculated with a 12-month lookback. Stock market skewness is a more abstract metric, but it simply implies that investors may be more cautious in the aftermath of a stock market crash.

With its long upward movements and few but steep downturns, the US stock market is more negatively than positively skewed over time. Almost all of the value factor’s positive returns occur amid periods of positive skewness, when no severe crashes have recently occurred. Investors feel safe and are more willing to bet on troubled firms.

The Value Factor and the Yield Curve

We calculated the yield curve as the difference between the 10-year and two-year US Treasury rates. A downward-sloping yield curve is associated with declining economic growth and an inverted yield curve interpreted as a leading indicator of recession. Unfortunately, the data only goes back to 1976, which limits the scope of our analysis.

We calculated the yield curve’s slope with a Z-score using a three-month lookback. We found that almost all the positive returns of the value factor occurred when the yield curve was upward sloping, or when economic conditions were more bullish.

Combining Metrics for Factor Risk Management

Based on these findings, investors might consider applying these metrics to time the value factor. We recommend approaching this from a risk rather than a return perspective. That is, we advise focusing on avoiding significant drawdowns when the market environment for owning cheap stocks is more negative.

Our multi-metric risk management framework only allocated to the value factor when a combination of stock market volatility, market skewness, and yield curve were favorable. Specifically, we modeled three scenarios in which one, two, or three signals are required for a factor allocation. Without the required signals, zero-interest cash was held instead.

Given the limits of our yield curve data, three positive signals representing a market environment with falling market volatility, positive market skewness, and an upward-sloping yield curve was only possible since 1976.

Our findings are quite typical for multi-metric frameworks: The more filters, the more consistent the returns but the lower the exposure to the factor.

Requiring at least one positive signal generated returns comparable to that of the buy-and-hold value factor. However, with two or even three signals, returns were much more consistent with significantly reduced drawdowns. The overall returns were lower than those of the value factor due to both high cash allocations and the limited yield curve data.

We also evaluated the performance of the value factor when the signals were negative. This could be used to potentially short the factor by buying expensive and selling cheap stocks.

A one-signal requirement generated performance in line with the buy-and-hold value factor. Two or three signals, on the other hand, yielded consistent losses, which represents a market environment of increasing volatility, negative market skewness, and a downward sloping yield curve.

All told, this analysis is far from perfect. We have not thoroughly tested the framework for robustness. We could use price-to-earnings instead of price-to-book for stock selection, change the lookback periods, include transaction costs, apply the framework to international markets and other asset classes, and so on.

But we have used common risk-measurement variables and publicly available data, made few assumptions, and applied our method to more than 90 years of financial history. This provides some comfort in the significance of these results.

One clearly faulty assumption is our same-day application of the trading signals. This is impossible to implement since changes in the variables and stock markets occur simultaneously.

To make the signals more realistic, we analyzed what would occur if the trades were conducted the next day. That led to significant reductions in the CAGRs for the frameworks requiring one or two positive signals, but not for three positive signals.

Further Thoughts

Understanding what drives value factor performance is immensely helpful, but implementing a framework around those drivers is challenging. It will work on average, but not consistently.

And the more filters around investor risk sentiment metrics, the lower the actual allocation to the factor and the more often cash is held. Investors don’t like to be out of the market, especially when the value factor is performing well.

It’s great to know how to improve the odds of generating returns with cheap stocks, but it hardly makes value investing easy.

For more insights from Nicolas Rabener and the FactorResearch team, sign up for their email newsletter.

If you liked this post, don’t forget to subscribe to the Enterprising Investor.

All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.

This article original contains information from CFA Institute, which can be viewed here.

Professional Learning for CFA Institute Members

CFA Institute members are empowered to self-determine and self-report professional learning (PL) credits earned, including content on Enterprising Investor. Members can record credits easily using their online PL tracker.