Value Investing: Why You're Doing It Wrong

Tuesday, May 9 marks the 123rd birthday of the late Benjamin Graham, who is commonly known as the "father of value investing." How would he feel if he could survey the value investing landscape today?

Benjamin Graham in 1950. Source: Wikimedia.

He might be pleased to hear that his disciple Warren Buffett, who has hailed Graham's 1949 book "The Intelligent Investor" as "the best book about investing ever written," shepherded Berkshire Hathaway Inc. (BRK-A, BRK-B) to a 1,972,595% return from 1964 to 2016 (compared to the S&P 500's 12,717%).

But Graham would also have plenty of cause for concern: according to a recent paper by U-Wen Kok, Jason Ribando and Richard Sloan, value investing is in a bad way. (See also, Berkshire Hathaway Annual Meeting Highlights.)

In 1934 Graham and David Dodd wrote that thinking about book value – a firm's net assets – as being the same as intrinsic value is "almost worthless as a practical matter." To be sure, book value is one of value investors' favorite metrics; Buffett lists Berkshires' return in terms of book value (884,319% since 1964) right alongside its share price return. But Kok and her colleagues point out that self-described value investors increasingly content themselves with analyzing a firm's value using book value alone or in combination with a few other narrow metrics, such as trailing earnings per share (EPS) or expected ones. (See also, 5 Must-Have Metrics for Value Investors.)

Such "formulaic value investing" tends to select firms with inflated fundamentals, the authors find, rather than underpriced shares. Value investing has "taken on a meaning we don't think Graham and Dodd intended," Ribando told Investopedia by phone Monday. And people are losing out on potential gains as a result.

Tens of billions of dollars have flowed into products such as the Vanguard Value Index Fund (VIVAX), which has $55.6 billion in assets at the time of writing, according to Morningstar. The fund tracks an index that compares book value, forward earnings, trailing earnings, sales and dividends to the stock's share price – an example of "strategies relying on a few fundamental-to-price ratios rather than on a comprehensive effort to determine the intrinsic values of the underlying companies."

To be fair, it's clear enough that the authors have skin in the game. Kok is chief investment officer of the RS developed markets team at Victory Capital Management. Ribando is senior director of predictive modeling at Arch Mortgage Insurance Company. Sloan is a professor of accounting and international business at the University of California Berkeley's Haas Business School. As investing becomes more and more the province of cheap algorithms crunching market data, a call for a "comprehensive" approach over a "formulaic" one sounds like humans defending humans' continuing worth to investors.

"That's exactly right," Ribando says. He and his colleagues compare value investing to self-driving cars: "you can let the car do the work, but you still want two eyes on the road and your hands near the wheel." 

Metrics are useful to perform screens and generate ideas, Ribando allows, but "the true fundamental analysis" should be left to humans.

The authors' aversion to simple, metric-based value investing is not just a matter of self-interest. According to their calculations, the approach simply doesn't work very well. They compare the returns of "value" indices to their broad-market counterparts to show that simply buying a low price-to-book ratio – or employing some similar formula – is a recipe for underperformance. (That's illustrated by investors who scooped up plunging oil stocks in 2014, which we elaborate on further below.)

Evidence of this misguided approach is that the S&P 500 Value index, for example, lagged the S&P 500 in the year, 5 years and 10 years to December 2015. The same goes for the Russell 3000 Value index, relative to the Russell 3000, and other major indices the authors analyzed.

Value indices' consistent underperformance may come as a surprise. Buffett summed up the central contention of value investing in the maxim, "Price is what you pay. Value is what you get." The implication is that the market is not perfectly efficient, so price and value periodically diverge. When Mr. Market undervalues a stock, an observant value investor will buy it and wait for the market to come to its senses. (See also, Warren Buffett: Most Influential Quotes.)

The only problem is knowing what the real, intrinsic value is. Market value is easy: Berkshire Hathaway was worth $412.5 billion at close on May 5, measured by its A-share price of (exactly) $250,000.00. Intrinsic value is harder. Should you think of the share price as 17.1 times trailing twelve month earnings? Or 20.2 times earnings estimates for the coming fiscal year? Or 1.5 times the company's book value? 

A lot of money is betting that the answer to this conundrum is to combine the resources offered by big data with the supposedly iron law of reversion to the mean. Pick a metric or combination of metrics, apply that formula to enough companies, and on average you should be able to exploit the market's occasional tendency to underestimate firms' value. As value indices' lackluster results show, however, this approach hasn't worked. (See also, Beware False Signals From the P/E Ratio.)

"Not one of these three popular fundamental-to-price ratios" – based on book value, trailing earnings and forward earnings – "has been effective in detecting temporarily underpriced stocks in our sample period," the authors write. "Instead, they have been effective in identifying stocks with temporarily inflated fundamentals." The divergence between a low share price and high forward earnings estimates resolves itself when analysts revise "optimistic" forecasts downwards – rather than the stock rising. A low-seeming price-to-book ratio rises when the company writes down assets. (See also, Why CEOs Aren't as Bullish as Wall Street Over Earnings.)

As inefficient as the market can be sometimes – a fact value investors depend on – it isn't that inefficient. For every mechanical index buying stocks based on uncritical screens of valuation metrics, there are "sophisticated value investors engaging in detailed fundamental analysis" who, the authors write, have "presumably figured out that the fundamentals are temporarily inflated and have set prices accordingly."

Ribando points to 2014 as an example. As the price of oil fell, oil companies' share prices looked attractive relative to their book values. "Something a bit insidious" happened as a result: these companies were rotated into value indices, even though a student of Graham and Dodd might have realized pretty quickly that oil stocks were "cheap for a reason." Write-downs followed, and book values stopped seeming so compelling. 

Ribando also pointed out backward-looking metrics' shortcomings when it comes to changes in earnings potential. Anyone's who's limiting themselves to trailing earnings per share will miss the value offered by a pharmaceutical company with a new blockbuster drug, for example.

"With the advent of computers and large data sets," Ribando says, investors have adopted "simple measures as proxies to doing that due diligence" Graham and Dodd counseled. Big data hasn't uncovered a shortcut to Buffett-like returns. Rather, broad-market indices have consistently outperformed their "value"-oriented counterparts.