Reading Filings  ·  No. 5

The other side
of value.

A 2013 paper out of Rochester proposed that the cleanest signal in equity research is sitting two lines below the top of every income statement.

Good morning. There is a 2013 paper out of the University of Rochester that found a strikingly simple ratio for picking higher-quality public companies. Forty-two years of U.S. equity data. Three of the most prestigious prizes in finance the year it was published. The math takes about thirty seconds to calculate from any 10-K.

Here is what is in it.

The paper.

Robert Novy-Marx was a finance professor at the University of Rochester when he published The Other Side of Value in the Journal of Financial Economics. The title was a little cryptic. The argument was anything but.

For most of the previous two decades, “value” investing, the strategy of buying companies that traded cheaply relative to their book value, had been the dominant signal in academic finance. Eugene Fama and Kenneth French formalized it in 1992. A generation of researchers refined and extended it. By 2013, the field felt close to settled.

Novy-Marx asked a question nobody had really pressed on. What if the most useful number on a company’s income statement was not the bottom line, but something near the top? He took revenue, subtracted cost of goods sold, and divided by total assets. That was the ratio. Gross profits over assets, or GP/A in the literature.

He submitted the paper. It won the Journal of Financial Economics’s prize for the best capital markets paper of the year, plus the Whitebox Prize and the AQR Insight Award. Two years later, when Fama and French expanded their three-factor model to five factors, profitability was one of the two factors they added. The academic absorption was complete and quick.

The Ratio

(Revenue − COGS)  ÷  Assets

Two lines from the income statement, one from the balance sheet. Run across forty-two years of U.S. data, this ratio predicted future returns about as well as the most-cited number in academic finance.

How it tested.

The way you test a return-prediction signal in academic finance is to sort companies into portfolios. High GP/A in one bucket, low GP/A in another, then you measure the spread between them over decades. Novy-Marx ran his sort the same way Fama and French had run their book-to-market sort, value-weighted at NYSE breakpoints, U.S. equities ex-financials, June 1968 through December 2010. Forty-two years of data.

The high-profitability portfolio outperformed the low-profitability portfolio by about as much as cheap stocks outperformed expensive ones over the same period. The signal worked, and at a magnitude comparable to the most famous signal in equity research.

The more interesting result was that the two ratios, value and gross profitability, turned out to be negatively correlated. Profitability had a rough stretch in the mid-1970s and again around 2007. Value had its lost decade in the 1990s. The two factors took turns underperforming, which is the most useful property a pair of factors can have, because a 50/50 mix of the two strategies, by Novy-Marx’s calculation, had drawdowns notably lower than either strategy on its own.

Why gross, not net.

The data on whether the ratio predicts returns is clear. The more interesting question is why such a simple number works at all.

Gross profits sit two lines below the top of the income statement. By the time the accounting walks down to net income, the figure has passed through SG&A, depreciation, stock-based compensation, interest expense, taxes, and a long list of items that vary by management decision. Each of those steps adds noise and gives executives more room to influence what the final number looks like. Gross profits are closer to the underlying economics of what the business does, which is sell things at a margin above the cost of producing them.

Scaling by total assets is the other half of the design. A company financed mostly by equity and a company financed mostly by debt can have very different ROEs but similar GP/A ratios, because the denominator does not care how the assets were paid for. The ratio is comparable across capital structures and across industries, and reasonably stable over long stretches of a company’s history. Novy-Marx’s data went back forty-two years and the signal was still there.

If you wanted to test this on a company sitting in your retirement account, the work is small. Pull up the latest 10-K. Find revenue. Find cost of revenue, sometimes called cost of goods sold. Subtract one from the other. Find total assets at the bottom of the balance sheet. Divide.

That is the ratio. Run it on the company today and you have a number you can compare against any other company in the U.S. equity universe, in any industry, over any year going back to 1968.

Have a good Thursday.

— Marques

About the author

Marques Blank is a finance leader with an FP&A background at Northrop Grumman, Lantheus Medical Imaging, and Citibank. CMA, MBA.

Source

Novy-Marx, R. (2013). The Other Side of Value: The Gross Profitability Premium. Journal of Financial Economics, 108(1), 1–28.

For informational purposes only. Not investment advice. Past performance does not predict future results. Factor strategies discussed are referenced from academic literature, not as recommendations.