Author Brian Nelson is the President of Investment Research at Valuentum. In his role, he has updated and overseen over 20,000 discounted cash flow models during the past 10 years. Prior to Valuentum, he worked as the Director of Methodology at Morningstar, a large independent research firm in Chicago, developing the company’s discounted cash-flow model used to derive the fair value estimates for the company’s coverage universe.

Key Takeaways

• Based on how we interpret the latest from Fama/French, a new study released a couple weeks ago, the long-advertised quant value factor, the book-to-market (B/M) ratio, may not have even existed.
• By extension, in our view, it reinforces our thesis that many quant value factors that are based on traditional valuation multiples may also not exist. Valuation multiples, themselves, are ambiguous. For example, a high P/E ratio can be an undervalued stock, while a low P/E ratio can be an overvalued stock. Read about this in Value Trap.
• Quant studies that attempt to tie the B/M value factor to the work of Benjamin Graham or studies that try to explain Warren Buffett’s returns with regression models cannot be verified, in our view. Quite simply, any data can explain any set of returns in the past; the data applied, and the relationships established can be spurious, a key peril of the backtest. Read about spurious correlations in Value Trap.
• The list of those that have become concerned about the impact of price-agnostic trading, or indexing and quant, have grown. Not only does it include Valuentum and The Big Short’s Michael Burry, but others as well. We include a link to a Bloomberg podcast in this piece that we think members should tune into.
• We remind readers that 85% of US quant mutual funds have underperformed the market in 2019 and 80% of US quant mutual funds have underperformed in 2018, amid an environment when the public is being sold on empirical, evidence-based terminology, which we believe is a misnomer. Anything that has happened in the past is empirical or evidence-based; empirical and evidence-based studies can be spurious.
• Stock prices and therefore returns will always be based on forward-looking expectations that are realized or not. This is an important consideration; it suggests the very idea of using realized reported data as in most quant applications won’t help explain prices or returns. Read Value Trap for more on this.
• We ask: Is it more reasonable that stock prices and returns can be explained by a combination of a company’s stock price activity relative to the market‘s activity plus some factor based on arbitrary book value as in the ambiguous book-to-market ratio coupled with the size of the company? Or is it more reasonable that stock prices and returns are based on future expectations of a company’s free cash flow, and the net cash position it holds on the balance sheet?
• Following our landmark call on Kinder Morgan and the MLPs in June 2015, we continue to work toward positive change in the industry, and this means helping investors sort through much of the controversial quant nonsense out there.
• Enterprise valuation acts as the confounding variable to statistical studies. Enterprise valuation shows all the drivers tied to stock prices and valuation, a long list that goes far beyond just sales or earnings or even EBITDA, for that matter. Read more about this in Value Trap.

By Brian Nelson, CFA

Before we get started, I wanted to remind you of the fantastic performance of the Best Ideas Newsletter portfolio and the Dividend Growth Newsletter portfolio during 2019. It was another great year for Valuentum members. Read about the Best Ideas Newsletter portfolio here. Exclusive idea success rates have been remarkable, too, and we steered clear of any mistakes in the High Yield Dividend Newsletter portfolio that might have impaired an outsize yield.

Still, perhaps the biggest achievement during the past couple years might have been helping investors avoid the ongoing “quant winter,” where “just 15 per cent of quant mutual funds beat the US stock market last year, according to Nomura.” There are few topics that are not covered in the book Value Trap. The text is an immersion in a variety of topics, from behavioral finance, quantitative theory and stock valuation, in order to provide the context for investors, ranging from the most inexperienced to the most advanced, to start asking some of the right questions, to perhaps start thinking more clearly about the quant discipline.

One of the most difficult things to do, however, is to find out where a learner has deviated significantly off course. Perhaps many are being confused in thinking that Benjamin Graham’s work in the 1930s with respect to net current asset value (NCAV), a cash-flow based process of identifying stocks under liquidation value, is related to today’s quantitative value or a backtest of the B/M? Maybe it’s the view that Warren Buffett’s returns can be mapped to a quantitative model?

We cannot be sure, but we think it is worth emphasizing that fundamental investing and quantitative statistical analysis are two vastly different things--they are not even close. Benjamin Graham’s and Warren Buffett’s principles generally cannot be captured in any regression model that employs reported backward-looking data. At the core of traditional value investing in the Graham and Buffett style is a focus on a company’s future expected free cash flows, and this dynamic cannot be captured in any snapshot one-year multiple, save for the price-to-fair value ratio (which captures future expectations).

We cover a lot of these considerations in the book, Value Trap. But it’s much more than this. Enter the backtest. The backtest might very well be one of the most widely-used items to explain an investment process, but the backtest is just the beginning to any process. Let me repeat: it is the beginning of the study, meaning what happens after the publishing of the backtest is the only thing that matters.

For starters, anybody can create a backtest with spurious correlations and make the chart look pretty. A pretty chart says nothing about whether the data used in the backtest will continue into the future. The bread-and-butter test, and what I think is the only test that matters, is the walk-forward, out-of-sample test. Check this out: As it relates to the walk-forward test, released January 2020, since the novel B/M factor was created reveals the following by Fama/French:

Value stocks in the U.S. produce higher average returns for the full July 1963-June 2019 period than the market portfolio of all listed U.S. stocks (Market). Average value premiums are larger in the 28-year Fama-French (1992) period, July 1963-June 1991, than in the 28-year out-of-sample period, July 1991-June 2019, and we don’t reject the hypothesis that out-of-sample expected monthly premiums are zero. But inferences from average premiums are clouded by the high volatility of monthly premiums, and we also can’t reject the hypothesis that out-of-sample expected premiums are the same as in-sample expected premiums.

This is huge news.

The bottom line is that, in out of sample statistical studies, ones completed by none other than Fama and French, themselves, the value premium, as measured by the B/M ratio might very well be zero. This may come as a shock to the quant community, but for anyone that has built and updated 20,000 discounted cash flow models over the past 10 years and for someone that has overseen the valuation models of a department of dozens of analysts, there’s little chance a statistical value factor such as B/M would hold up…and it did not hold up in walk-forward, statistical studies!

This has been the reality, as I see it. From my perspective, the quant value factor, as in B/M, simply never existed. It was developed by a backtest based on the data that was available at the time, and it was not verified in out-of-sample, walk-forward studies. The takeaway of the failures of the B/M, however, is much more than just the failure of one archaic multiple that arguably served as the foundation of quant factor investing.

Hear me out: Because multiples, themselves, are ambiguous (and enterprise valuation acts as the confounding variable to multiple analysis), most any valuation multiple used in quant analysis, save for the price-to-fair value ratio, might very well be spurious in walk-forward testing. Let this sink in. All of those quant funds and ETFs incorporating ambiguous valuation multiples might very well be charging for randomness. You have to read Value Trap to dig further into what I am talking about, which brings me to another point.

The book also explains how quant and index investing, together, might contribute to a highly irrational marketplace. Michael Burry of The Big Short fame believes index funds are in a bubble and distorting price discovery, and just today, perhaps one of the best podcasts released in some time was made available at Bloomberg, “How the Rise of Passive Investing May Be Creating Huge Distortions in the Market.” Tune in.  

I continue to dedicate my time to helping investors learn about investing and protecting them from a lot of the stuff out there that can get them into a lot of trouble. We’ve helped so many investors with our warning on MLPs, but we think the quant and indexing fallout will be even bigger when the music stops. We continue to update investors on new developments as our thesis in Value Trap continues to play out across the board.

Tickerized for holdings in the DIA.