Stanford pair warns of Quadrophobia

Lending insight on potential management manipulation of earnings reports, a new study by Stanford researchers suggests that some companies may slightly tweak their reported earnings per share (EPS) — and that companies that engage in such adjustments are more likely to be charged with accounting violations.

Stanford Law School professor Joseph Grundfest and Nadya Malenko, a doctorate candidate at the Stanford Graduate School of Business (GSB), co-authored the study, “Quadrophobia: Strategic Rounding of EPS,” based on research and analysis they have conducted since the summer of 2007.

The two analyzed annual and quarterly filings for 22,460 companies from 1980 to 2006 and found statistically significant infrequent appearances of the number four in the tenth-of-a-cent decimal place of reported EPS figures.

“Changing the EPS number by .1 cent from .4 to .5 will change the reported EPS figure by one cent,” Malenko explained.

Through statistical analysis, Grundfest and Malenko found that there were very few fours in the hundredth place of unrounded EPS data<\p>–<\p>a phenomenon they dubbed “quadrophobia” and a strong indication that firm managers tend to round up figures.

“In the past, people had tested for rounding behavior, not specifically for quadrophobia,” Grundfest said. “What we did was analytically a little different: we looked at the number four, rather than a clusters of numbers, and we found patterns that had not yet been appreciated in the literature.”

In addition to pinpointing the prevalence of “quadrophobia,” the study also looked at characteristics of companies that exhibited quadrophobic behavior in order to determine potential causes of the phenomenon.

“One of our results is that analyst coverage is a good determinant of quadrophobia,” Malenko said. “The behavior is more pronounced among companies that are covered by analysts, and when a company gains analyst coverage, the behavior is more pronounced than before.”

“It’s likely that one of the incentives is to beat analyst forecasts,” she added.

Furthermore, Grundfest and Malenko found a strong correlation between companies that exhibited quadrophobia to incidence of restatements, class action securities fraud litigation and Security and Exchange Commission enforcement actions alleging accounting violations.

These findings contribute to existing studies on potential management manipulation of reported earnings data, though the most immediate application of the study may be for regulators and audit firms.

Both co-authors cautioned against using “quadrophobia” as a determinant or predictor of individual companies’ characteristics, as a company may have historically exhibited the tendency simply out of chance rather than managers’ active manipulation of EPS reporting.

“Quadrophobia should be one of many factors to consider,” Grundfest said. “Statistically, it is very hard to use quadrophobia as a measure of the likelihood that one individual company will get into trouble.”

“Instead, the study and its conclusions would be better used as a descriptor of a pool of companies,” he said.

About Joanna Xu

  • Skeptic

    Sounds like a fishing expedition instead of a contribution to a solid research program. Too much data and computing power, not enough authentic insight