Analysts Ignore Key Factor When Estimating Earnings Says University of Miami Study
May 28, 2009
Financial analysts seem to ignore a key factor when they estimate earnings of some of the most prominent firms according to a new study from the University of Miami (UM) School of Business Administration. The study indicates that more than 70 percent of analysts do not account for an extra week that many companies include in their earnings quarter every five or six years. These companies normally report earnings on 13-week quarters, four times per year, which leaves an extra day in most years and two extra days in leap years. To make up for those extra days, the companies report earnings based on 14-weeks during one quarter every five or six years.
Research involving more than 600 of those companies and 365 analysts, conducted by Andrew Leone, department chair and professor of accounting, Sundaresh Ramnath, assistant professor of accounting, and Ya-Wen Yang, assistant professor of accounting at the UM School of Business Administration, along with Ohio State University's Rick Johnston, reveals that most of those analysts ignored that extra week, basing their estimates on the usual 13-week quarter.
“I’m not sure what is more surprising about our findings, the fact that analysts fail to anticipate the extra week, which causes them to underestimate earnings and revenue, or the fact that investors haven't tried to make money off of this somewhat rare but often ignored quarter,” said Leone. “If someone buys the stock in its 14-week quarter and holds it until the earnings announcement, he or she will profit on average because of the higher than expected earnings and revenues resulting from the extra week.”
In fact, the study found that buying and holding stocks of such companies results in positive abnormal returns of approximately 2.9 percent over the quarter (11.6 percent on an annualized basis).
The oversight also contributes to overestimation of revenues the following year. The study found that for the same quarter in the year following a 14-week quarter, analysts appear to overestimate revenues and, to a lesser extent, earnings, apparently forgetting that there was an extra week in the previous year’s quarter.
“These results are quite unexpected in light of the ease with which this information can be obtained and incorporated into expectations,” said Ramnath. “We attribute analysts' and investors' apparent failure to do so to a lack of effort rather than ability.