Cory Cassell, University of Arkansas.Image courtesy of the University of ArkansasFAYETTEVILLE, Ark. – Retiring chief executive officers issue earnings forecasts more frequently during their final year of employment, and these forecasts tend to convey good news – to be optimistically biased – than those released during pre-terminal years, according to a new study by an accounting researcher at the University of Arkansas.

The study revealed that optimistic, terminal-year forecasts were more pronounced in the presence of higher CEO equity incentives. However, the forecasts were less pronounced when stronger monitoring or corporate governance mechanisms were in place.

“So the immediate question is, ‘Why?’” said Cory Cassell, assistant professor in Sam M. Walton College of Business. “The short answer is to inflate stock prices during the year leading up to their retirement. Our work suggests that managers use opportunistic earnings forecasts to manipulate analysts’ and investors’ perceptions, presumably in an effort to maximize the value of their stock holdings as they approach retirement.”

This strategy could benefit retiring CEOs because U.S. Securities and Exchange Commission trading rules for CEO post-retirement security transactions are less stringent than those in effect during executives’ tenure with their firms.

Cassell and research colleagues at Arizona State University and Texas Tech University examined several properties of earnings forecasts – issuance, frequency, news and bias – to compare forecasts issued during CEOs’ terminal years versus those issued in pre-terminal years. The important difference is that for forecasts issued during pre-terminal years CEOs will still be with their firms when actual earnings are realized. For forecasts issued during terminal years, CEOs will have already retired when earnings materialize.

The researchers identified all CEO turnovers from 1997 to 2009, as documented by ExecuComp, a comprehensive database that tacks executive compensation — including salaries, bonuses and stock options — in the S&P 1000 index of firms. Their search yielded a sample of 272 CEO retirements that included quantitative forecast information to construct measures of forecast news and bias. Tests using forecast issuance and frequency were performed using a larger sample of CEO retirements because qualitative forecasts could be used in those tests.

Previous research has focused on the so-called “horizon problem,” the theory that CEOs with short horizons with respect to retirement – generally less than a year left with their firm – have weaker incentives to act in the best interest of shareholders.

In this study, the researchers found that CEOs are more likely to issue forecasts of future earnings during their terminal year relative to other years during their tenure with the firm. They also issue forecasts more frequently during their final year of employment. Most importantly, these forecasts are more likely to convey good news and are more optimistically biased relative to forecasts issued during pre-terminal years.

Cassell said the results of the study should be of interest to governing boards and regulators, who are trying to implement incentive mechanisms that mitigate such conflicts or potentially unethical behavior. The findings can also benefit market investors and analysts who rely on information from management earnings forecasts.

The study was published in the November 2013 issue of The Accounting Review.