The level of corporate earnings manipulation is similar to that of past pre-recessionary periods, according to research by professors at the University of Missouri and Indiana University.

Their finding is based on the M-Score, a screening model that catches fraud in corporate earnings reports. Messod Daniel Beneish, a professor at the Indiana University Kelley School of Business, created the M-Score in the 1990s. The “M” stands for manipulation, and the measure is also sometimes referred to as the Beneish M-Score.

Based on known examples of past financial misreporting, the M-Score combines eight ratios on a company’s balance sheet to assess its fraud risk. A higher M-Score means a company is more likely to be manipulating its earnings.

“It allows us to assess fraud risk in real time,” said Matt Glendening, an accounting professor at the University of Missouri. “The advantage of using a measure such as the M-Score is that if you use actual instances of accounting fraud, not all cases are caught, especially the less severe cases. And also, there is a delay between the misreporting period and the time at which the fraud is actually revealed.”

One notable M-Score success came in 1998, when a group of Cornell students used the M-Score to flag Enron as having an elevated fraud risk. This was three years before the public learned that the company was inflating its profits, resulting in what was then the largest corporate bankruptcy in history and several executives going to jail.

“Now if a firm has a high risk of fraud, the measure doesn’t say that is a fraudulent firm,” Glendening said. “It just says it’s pointing in the direction of a higher probability of fraud.”

Corporate earnings are traditionally manipulated either by overstating revenues or understating expenses. How companies do this varies, but it could include recognizing sales revenues early or understating inventory.

“There are all sorts of capital market pressures on firms to maintain stock price, maintain earnings growth,” Glendening said. “There could also be some compensation incentives at play.”

In 2019, Beneish expanded the M-Score, creating a new measure that goes beyond individual companies to the economy as a whole. With the help of Glendening and two other co-authors, Beneish created the aggregate M-Score, which now compiles the M-Scores of 2,004 companies to measure the likelihood of earnings manipulation across the economy. Earlier in 2023, the aggregate M-Score was at its highest level in 40 years.

“Accounting manipulation matters for the economy at large,” Glendening said. Companies use other business’ earnings data to inform hiring, purchasing, and production decisions. “What we are finding is that the level of aggregate misreporting is very similar to what we’ve observed in pre-recessionary periods.”