Turns out you can’t regulate accuracy.
More than a decade after a government crackdown on conflicts of interest on Wall Street, a new study says stock analysts are no better now than they used to be at predicting corporate earnings. Actually, they’re worse, according to the paper, which reviewed how close profit estimates came to what companies ended up reporting from fiscal year 1994 to 2013.
The study, to be published in the CFA Institute’s Financial Analysts Journal, is the latest to assess the reliability of Wall Street research and find it lacking. Several explanations are possible for the deterioration, the authors speculated. Among them: analysts got lazy after the spotlight on their work faded, or companies don’t release enough data to make forecasting results possible.
“Even though conflicts of interest are being reduced because of rules, we’re still back to square one because we don’t have accurate forecasts,” Reza Espahbodi, the Washburn University professor who co-authored the report, said in a phone interview. “If the quality of the information coming out of the financial reports analysts have access to is not where it should be, they’re not going to be as accurate and unbiased as they should be.”
After the dot-com crash and Enron scandal, regulators passed laws to improve corporate disclosure and ensure analyst biases weren’t harming investors. The Sarbanes-Oxley Act in 2002 was aimed at improving transparency in accounting. Regulation Fair Disclosure in 2000 compelled companies to give everyone the same information at the same time.
In a study titled “Did Analyst Forecast Accuracy and Dispersion Improve Following the Increase in Regulation Post 2002?” Espahbodi, his brother Hassan and sister-in-law Pouran tried to assess whether the rules resulted in better company research by examining the accuracy of earnings predictions.
Two values were measured: how far analyst projections veered from actual results, or forecast error, and how much analysts differed from each other, known as dispersion. Both worsened by more than 2 percentage points between 2000 and 2013, with the average estimate missing its target by 35 percent, compared with 33 percent, according to the authors.
The results indicate that analysts are “unable or unwilling to accurately predict earnings” and have trouble agreeing. They’re consistent with past studies showing a brief improvement in accuracy that later faded, they said.
While no tests were run to isolate the cause of the worsening, the authors offered speculation, including a tendency for changes in work environments to produce productivity improvements that don’t last, something known as the Hawthorne effect.
“Analysts knew they were being watched and expended ‘extra efforts’ on metrics on which they were being measured,” they wrote. “Over time, as the attention waned, the analysts may have reduced their extra efforts and their forecast properties could have suffered.”
Widening disagreement among analysts suggests another possibility, that something about the data they get from companies is preventing better predictions.
“Greater dispersion indicates less agreement among analysts regarding expected earnings due to higher uncertainty, or lower availability or reliability of information, reducing the value of forecasts,” they said. “The continued problem with the information environment, therefore, seems to be largely due to the quality of financial reports.”