In recent years, quantitative stock analysis has taken the investment world by storm while qualitative analysis has been given a back seat. Value investing however — as Benjamin Graham would remind us — is both an art and a science.
A new study in the CFA’s Financial Journal showed the underperformance / diminishing alpha of stocks chosen purely based on their quantitative value, and explains why value factors taken alone aren’t great indicators.
But first, a quick trip into the value investing time machine..
Quantitative Value Investing History
Benjamin Graham’s and David Dodd’s 1934 Security Analysis is the seminal book on value investing. Security Analysis offers investors a comprehensive guide to analyzing companies to find value stocks which are priced below their intrinsic value. Graham & Dodd advise a number of strategies to find value stocks, ranging from qualitative factors like identifying industry trends and a company’s management team to quantitative factors like book value, P/E ratio, and sales-to-price.
As Graham’s value investing ideas gained popularity in the investing community with disciples like Warren Buffet and Mario Gabelli, a ton of portfolio managers and private investors began mining his work to develop their own investment strategies.
With the advent of powerful computers, databases, and stock screeners it has become increasingly easy for investors to implement Graham’s quantitative strategies, while the qualitative study of companies remains nearly as time consuming and research intensive as it was in 1934. This imbalance of effort has generated a dangerous proliferation of quantitative investing without qualitatively studying stock fundamentals.
This view of intrinsic value was quite definitive, but it proved almost worthless as a practical matter because neither the average earnings nor the average market price evinced any tendency to be governed by book value.
– Benjamin Graham, Security Analysis, pg. 17
Facts about Formulaic Value Investing
Recently U-Wen-Kok, CFA, Hason Ribando, CFA and Richard Sloan published a paper, Facts about Formulaic Value Investing which outlines the shortcomings of quantitative investing in the CFA’s Financial Analysts Journal.
Minimal Evidence of Fundamental-to-Price Ratios Outperforming
U-Wen-Kok, Ribando & Sloan back-tested six portfolios split using the Fama-French Value Factor Model, which simply divides up the stock market using two factors:
- Book-to-Market tiers – HIGH, MEDIUM and LOW. As a refresher, book-to-market is the total asset value on a company’s balance sheet divided by the company’s market cap, so the higher a companies book-to-market ratio, the better value the company offers by this metric. In this study the High portfolio is the top 30% of book-to-market companies, the Medium is the next 40% and the Low the remaining 30%.
- Market Cap Tiers – Large Cap (BIG) & SMALL Cap.
To look for a correlation between book value and stock returns, we will take a look at the study’s results for the HML, HML Big and HML Small alpha returns. The HML alpha returns for each period show the returns on stocks with the highest book-to-market values (top 30%) minus the returns of stocks with the lowest book-to-market values (bottom 30%). Comparing the returns between the highest and lowest value stocks gives us a good indicator of how strong a stock’s book value predicts it’s future returns.
A last note before we take a look at the study’s results, HML BIG is the large cap subset of the overall HML performance, and the HML SMALL is the small cap subset. The overall HML alpha return is equally weighted between HML BIG and HML SMALL.
Study Results: Annualized Alpha for High – Low Value Stocks
After reviewing the HML alpha for each period a few conclusions can be drawn.
Firstly, value offered significantly better returns from 1963 to 1981, both large and small cap high-value stocks produced a 6%+ alpha over low-value stocks.
Outside of the 1963-1981 period, value has been a weak factor to predict outperformance.
The second best HML period was 1982-2015, however the bulk of that performance was in HML SMALL (9.15%) while HML BIG only generated a meager 1.41% alpha. Despite the study equally weighting HML BIG and HML SMALL, in reality 90% of NYSE market cap is from HML BIG while only 10% is from HML SMALL.
Digging further into the outperformance of HML SMALL during this period, the study’s authors note that the HML alpha can be tied more to horrible performance by the low-value small cap stocks for the period instead of great performance by the high-value stocks.
Furthermore, value’s alpha falls to nearly flat for the 2002-2015 period.
So why did value only outperform in the 1963 to 1981 period?
The Nifty-Fifty craze of the 60s and 70s helps tell the story of value’s temporary outperformance. As investor confidence soared, a fad surfaced that there were about 50 well-known companies (IBM, McDonald’s, Pfizer, etc) which were labeled can’t miss investments that would continue to grow their dividend and offer healthy stock price appreciation.
The Nifty-Fifty stocks inflated to an average P/E of 42 while the S&P 500 P/E was 19. Extraordinary P/Es were given to names like Polaroid (91), McDonald’s (86) and Walt Disney (82). These growth stocks eventually reached their demise in the 1973-74 crash which dropped the Dow Jones 45% from it’s then all-time high. As the stock market bled and investors fled, the P/Es of the Nifty-Fifty contracted at a rapid pace. Value stocks during the same period were obviously severely hurt by the crisis but weathered the storm considerably better than the Nifty-Fifty growth stocks; helping to explain the value factors outperformance from 1963-1981.
Could FANG stocks — Facebook, Amazon, Netflix, and Google — be today’s version of the Nifty Fifty? I’d love to hear your thoughts in the comments!
Why do some stocks appear significantly undervalued?
Stocks that are significantly undervalued by quantitative measures often experience a reversion to the mean, their price eventually becomes more inline with their fundamental value. This reversion to the mean can occur one of two ways (or a combination of both!):
- Price Increase — the optimistic outlook of the quantitative value investor — that the stock’s valuation will reach a natural equilibrium by the stock’s price increasing more in line with the underlying value of the company.
- Fundamentals, book value, earnings deteriorate – the low price of the stock actually foretold the coming decrease in the underlying value of the company, and the lower earnings of the company reverts the firm’s P/E to a higher level.
To discover whether price increases or fundamental deterioration plays a larger role in stocks returning to a reasonable valuation, let’s examine an attribution graph pulled directly from U-Wen-Kok, Ribando & Sloan’s study.
These graphs show the change in book-to-market (graph A) and earnings-to-price (Graph B) for the highest value stocks from one year to the next. The “beginning spread” is the difference in valuation for these stocks vs the middle tier of value, and the “ending spread” is the difference in valuation after a year.
For instance, Stock A and Stock B both have earnings per share of $2 but Stock A’s price per share is $50 and Stock B’s is $100. In this case the beginning spread of Stock A and B is 100% as Stock A offers double the value of Stock B — an earnings-to-price ratio of 4% vs E/P of 2%. If we take it that Stock A was an exceptional value compared to the rest of the market, after a year we’d expect to see Stock A’s 4% E/P revert towards the mean of 2%, if it ended the year at 3% that would drop the “ending spread” to 50% from our beginning spread of 100%.
The “impact of book value changes” (Graph A) or “impact of trailing earnings changes” (Graph B) represent how much of the difference in valuation change came from the underlying value of the stock. So if in our example Stock A’s price stayed at $50 but it’s earnings per share dropped to $1.5 from 2, the impact of earnings changes was the sole contributor to closing the valuation spread.
As you can see from the chart, on average the impact of changes in the stock’s underlying fundamentals (e.x. book value or earnings changes) makes up more than 100% of the change in valuation spread! Price is actually contributing negatively to the reversion to the mean, so the price is actually DOWN slightly (increasing the value spread) while the fundamentals of the stock are dropping precipitously.
A Sports Analogy
To help illustrate what’s happening when stocks which appear to be high value on paper revert to their mean valuations over time, consider betting on a basketball game.
The championship contending Cleveland Cavaliers are playing the — non-championship contending — Charlotte Hornets. You can use any data-driven analysis to conclude that the Cavaliers should be the favorite by perhaps 6-7 points. Yet you look at the betting line and somehow the Cavaliers are 4 point underdogs! Of all the games tonight, betting on the Cavaliers offers you the most favorable spread.
However, nowhere in your quantitative basketball gambling model is the fact that Lebron James sprained his ankle and he won’t play against the Hornets; and that’s the reason the spread seems to be out of whack.
Looking beyond the numbers is the core concept of the CFA journals’ study. If a stock appears to be cheap based on it’s financial statements, there is probably a qualitative reason for it’s current price level. This is of course not to say that value can’t be found, but value cannot be determined strictly by crunching numbers without assessing a company’s industry, management, growth prospects, and so on; to determine if and by how much a company is undervalued.
A Real World Investment Example
While the CFA study’s data alone is powerful, let’s use the video game retailer GameStop (GME) as a real world example of why seemingly high value stocks can be misleading.
In January of 2015 GameStop traded at $32.50 per share on $3.30 EPS, giving it a PE ratio of 10. As of writing this article in June of 2017, GameStop’s PE ratio has dropped to 6, it’s EPS remains at $3.30 but it’s stock price is at an all-time low of $20. GameStop is now an extremely high value stock on paper. From a stock analysts perspective however, GameStop faces considerable headwinds:
- Internet speeds have increased so quickly that now almost all PC games are downloaded over the web, and in fact GameStop no longer sells PC games in their stores. Similarly, Playstation and Xbox are now offering the option for consumers to download their game directly from their consoles, so people can pick up games instantly without having to trudge down to their local GameStop.
- Online retailers — namely Amazon — are creating tremendous pressure on brick-and-mortar stores. Even beyond the direct competition these online stores offer, overall mall and shopping center foot traffic is falling quickly.
- Video game rental subscriptions are gaining popularity, with Microsoft’s Xbox now offering their own service for $9.99/mo to their most popular games on demand
Through the lens of the CFA study we reviewed, we can make a prediction that GameStop’s P/E will revert closer to it’s mean historical ratio of 10 in the coming years. Unfortunately for GameStop shareholders, it’s more likely that GameStop’s earnings will dry up — similar to BlockBuster after the advent of Netflix (NFLX) — than the company’s share price will rise.
Could GameStop defy gravity and maintain or grow it earnings in the coming years? Of course, there are exceptions to the rule; but I thought it would be fun to take an example stock in the moment today when the outcome is uncertain. It will be interesting to watch GameStop’s stock in the next few years!
In conclusion, using financial ratios alone to assess a company’s value can be extremely misleading and seldom leads to portfolio outperformance. Sometimes there is great wisdom in the valuations of Mr. Market and active investors. However, to truly excel in investing, we need to learn the dance of art and science.
If you’d like to learn more about Facts about Formulaic Value Investing, here’s a CNBC interview with the study’s author U-Wen Kok.