How textual fingerprints can predict corporate collapse

Research suggests computer-aided financial statement analysis can detect company malfeasance years before markets catch on

How textual fingerprints can predict corporate collapse

While scandals can hit markets and share prices with the suddenness and intensity of a bomb, the downfall of the companies involved usually results from errant behaviour and dishonesty that stretches back for years. And while the majority of investors might be caught unaware, some are able to make prescient divestment decisions through unorthodox corporate analysis.

That’s what Jason Voss, CFA, described in a recent piece published by the CFA Institute.

“We used Deception And Truth Analysis (D.A.T.A.) to examine 10 of the largest corporate scandals in recent history and found that the average lead time between our textual identification of deception and the public recognition of possible scandal was more than six years,” Voss wrote.

The ability of D.A.T.A. to detect deception more quickly than even intensely focused investors and regulators, he said, may come down to the fact that 86.5% of financial information is expressed in text rather than numbers. While the large majority of the market may believe that numbers in financial statements accurately price in all relevant information, the reality is that poorly behaving firms have a limited ability to keep up the appearance of good financial performance by flattering their numbers.

 

Voss highlighted the cases of AIG and Lehman Brothers, which were hidden from public view for the longest period out of all the scandals in the analysis. Both financial companies, the two ill-fated firms released annual reports that ran in the hundreds of pages, with money churning through their balance sheets and income and cash flow statements at tremendous rates. That helped ensure their poor behaviours and choices wouldn’t reflect in the numbers for more than 10 years.

Going beyond the 10 scandals to include more recent controversies such as the Wirecard affair, Voss pointed to five textual fingerprints that distinguished deceptive companies from their more truthful counterparts.

  • Words indicating friendship – based on the analysis, deceptive companies employ terms that imply friendship at a rate 56.1% higher than the average expected from business reports;
  • Risky words – while firms in general proactively excise words such as “averse,” “avoid,” “concern,” and “difficulty” from their annual reports, as they tend to raise red flags among securities researchers, scandal firms favour these types of words at a higher rate;
  • Impersonal pronouns – compared to truthful peers, scandal companies use impersonal pronouns such as “another,” “everybody,” “someone,” and “whichever” 54.1% more frequently, presumably to create emotional distance between them and the readers they wish to mislead;
  • Words that indicate difference – because of the cognitive load of lying, those who do it are less able to make distinctions among competing points of view, making them less able than average to draw comparisons. Since they have a preferred narrative to push onto their targets, deceivers are also loath to highlight distinctions with others;
  • Words that negate a statement – research also indicated that liars are 50.4% more likely to use words such as “not,” “never,” “should not,” “does not,” and “must not” than truthful firms.

Apart from the five fingerprints, Voss highlighted one hallmark that by far gives the strongest indication of corporate deceptiveness:

“Though they are rarities, swear words occur in scandal company annual reports a whopping 277.1% more frequently than the mean.”

 

Follow WP on FacebookLinkedIn and Twitter

LATEST NEWS