Quartz – “A new corporate principle is: Never say you’re restating anything. The machines will hold it against you. Over the past decade, “restatement” is the word companies have most strenuously tried to avoid in the text of their filings with the Securities and Exchange Commission, according to an upcoming paper by academics at Georgia State University’s J. Mack Robinson College of Business and Columbia Business School. The evasion is a response to the army of bots that investors routinely deploy to catch any whiff of implicit market intelligence. The bots flag “restatement” as a negative word, which darkens their outlook on a company’s prospects. The paper, which builds on an earlier version released last October by the National Bureau of Economic Research, analyzes nearly 360,000 10-K and 10-Q documents filed with the SEC between 2003 and 2016. It shows how companies are trying to phase out words that, like “restatement,” are judged to have negative connotations by financial analysis algorithms. “Corporate disclosures and reporting have been reshaped by machine readership,” Baozhang Yang, one of the paper’s authors, said. The dog-eat-dog nature of markets has always tended to favor those with an extra sliver of information. When everyone has access to the same, structured sources of data—published reports, earnings figures, profit-and-loss statements—the appetite for alternative data, of even the least merit, can turn very keen. Sometimes analysts and investors try to latch on to such alternative data using their own intuition. Nils Paellmann, a former vice-president for investor relations at T-Mobile, recalled how, on one earnings call, the company’s CEO—known for cursing freely and often—was far less profane than usual. “I got a lot of calls from investors, asking: ‘What’s wrong? Why wasn’t he his usual self? Is he less confident about the company’s outlook?’” But the real thrust has come with the advance of automation. In the past decade, the channels of obtaining alternative data have grown, as has the raw computing power to crunch and weigh such data. Within the industry, practitioners tell wild stories—of trading equities by using satellites to count cars in parking lots, as a proxy for economic activity; or of planting an infrared camera outside a Tesla plant, to determine how busy it was at night and therefore how full its order book was. Some of this is what one expert calls “innovation theater,” but investors also glean genuine value out several of these strategies…”