Hidden in Plain Sight: A Powerful Way to Beat the Market

The Wall Street Journal  June 13, 2017

Companies are a little lazy about what they put into regulatory filings. Investors are profoundly lazy about reading them.

Compare a company’s most recent annual report to its previous one, and you will quickly notice the language doesn’t change much. Hot-dog seller Nathan’s Famous NATH 0.88%has talked about the damage done to its flagship Coney Island location in 2012 in the exact same language three years in a row.

Do investors still care?

Nathan’s isn’t taking any chances, nor do other companies. There is, says Harvard Business School economist Lauren Cohen, an incredible amount of inertia in the language companies use in quarterly and annual regulatory filings, known as 10-Qs and 10-Ks respectively. That is partly because corporate lawyers push back against alterations, worried they might expose a company to legal risks.

When the language in filings does change, investors should sit up and notice. Not that they do.

Mr. Cohen and economists Christopher Malloy and Quoc Nguyen downloaded all the 10-K and 10-Q filings with the Securities and Exchange Commission from 1994 through 2014 and used textual-analysis software to create a similarity score showing how the language in corporate filings differed one period to the next.

They then looked at stock performance following filings. The finding: Shares of companies that had significant changes did much worse than those of companies that didn’t. This was particularly true when it came to changes in the risk factors section of 10-Ks.

Summer ReadingNumber of words in the average 10-K filingTHE WALL STREET JOURNALSource: Bill McDonald

Indeed, a strategy of buying shares of companies with no significant risk-factor changes and betting against companies with major changes would have returned more than 22 percentage points more than the overall market annually.

The similarity score for data-storage company NetApp, for example, fell sharply when it released its 10-K in June 2011. The company’s shares had more than quadrupled from their financial-crisis low, but that run ended. Over the next year, its stock fell 41% while the S&P 500 rose 4%.

Tracking risk-factor changes is something even investors without programming chops can do for themselves. Professionals can use services such as FactSet, which has a function called blackline to highlight differences in filings. Individual investors can accomplish as much by pasting text from two 10-Ks into separate Microsoft Word documents and using the compare function, or by using text comparison sites such as

In the 10-K it released in March 2014, copper and fiber-optic wire maker General Cableadded a paragraph to its risk factors stating it was reviewing commission payments to its subsidiary in Angola. That September, it announced it was investigating possible bribe payments in Angola, Thailand and India, and its stock fell sharply.

The stock underperformance of companies that make big changes to their risk factors suggest that, as a group, they aren’t merely updating their filings to reflect risks that investors have already learned about, but providing information about emerging ones.

What is really striking, however, is that the stock market reaction to these risk-factor changes occur gradually. Companies are providing investors with material information, and investors aren’t noticing it.

Some of that probably has to do with the form in which the information is getting provided. Annual reports are big, and getting bigger. University of Notre Dame economist Bill McDonald finds that the average 10-K filed last year weighed in at over 26,000 words, nearly three times as many as 20 year ago. That is the equivalent to the text on around 10 full pages in The Wall Street Journal, but about 100 times duller

Not that investors are trying. Mr. McDonald and colleague Tim Loughran found that the average 10-K gets downloaded from the SEC fewer than 30 times, total, on the day and day after it is filed.