Turns out this backdating stuff ain't so black and white to the SEC. Now, about the business press...
Posted by: Peter Burrows on March 9, 2007
There was more evidence today that the options backdating scandal is changing—in ways that make it all seem just a bit less scandalous.
First, there was a story in the Wall Street Journal today, about a debate that's been raging within the SEC for the past year as to whether to fine Brocade Communications for backdating-related problems. While the SEC did charge three former executives last July, including former CEO Greg Reyes, it has failed to take the $7 million set aside by the storage gear maker last March to settle any civil charges. The reason, says the Journal, is that commissioners are debating “whether and how harshly to penalize companies where backdating occurred.” According to the Journal, that’s because of the difficulty of proving how, and to what extent, shareholders were actually harmed by alleged backdating at the company. As the article said:
...backdating cases also present unique questions. Among them: Were executives compensated differently than shareholders believed? Is there a value to the company of using backdated options to attract and retain employees? Did a stock decline triggered by the disclosure of backdated options reflect damage done to the company by the backdating -- or was it caused by fear of litigation or potential impact on corporate officers? Do investors consider restatements involving options, which don't involve an expenditure of cash, relevant?
It’s a cogent summary of a side of the backdating story that many of my sources feel has gotten short shrift, if any shrift at all, in the Journal’s aggressive coverage of the scandal (this excludes, clearly, the work of Holman Jenkins, who has been almost as unrelenting in his efforts to debunk the seriousness of backdating, as his paper has been in proving it. Check out his latest take, entitled "The Backdating Molehill.").
Also, there’s a new study by economists at NERA Economic Consulting that is one of the first that questions some of the academic underpinnings of the scandal. Entitled “Options Backdating: The Statistics of Luck,” the authors take issue with the methodology that has become familiar to anyone that’s followed this story – the tendency to express the likelihood of a company’s ability to grant options at low stock prices, relative to the chances of this occurring randomly. As in, the chances of company x granting options at monthly lows on x occasions is x in a billion, or whatever. Ever since the Journal began using this formulation (with the help of some guidance from University of Iowa professor Erik Lie, according to Lie) when its stories sparked this scandal a year ago, it's become de riguer in articles on backdating. (That includes BusinessWeek, although I’ve resisted the temptation of late. For example, in my recent story on Reyes, I couldn’t bring myself to include that the odds of Brocade issuing 25% of its grants at monthly lows was a staggering 50 trillion to one. Somehow, my gut told me odds that long are meaningless).
Now, NERA authors Renzo Comolli and Patrick Conroy tell me my hunch was right. I spoke with them this morning, and they say that analyzing option grant patterns against the odds of them occurring randomly is misleading. They argue, as have countless Silicon Valley executives and their defense attorneys, that random had little to do with anything during the Net Boom years when most of the wrongdoing is alleged to have occurred.
Indeed, many companies were actively--and unashamedly--looking for opportunities to grant options on days when the stock was low. What better way to incent workers to work hard and stay with the firm until their options vested, than to grant them when there was plenty of opportunity for the stock to rise. On other occasions, options may have been granted as a reward for finishing a big new product or hitting some financial milestone—which could also cause the stock to rise, making the grant date appear to be a “low”, the authors suggest. “Nobody ever thought the granting of options was random. They don’t just throw darts at the calendar,” says Conroy.
Rather, the authors write that “some grant patterns that may appear extremely unlikely at first sight are actually very likely and should be expected.” Whoa, you say? So did I. But they make interesting points. For example, they argue that even if options were granted randomly, many of them would be granted at relatively low prices—just as many would be granted at high prices.
The Wall Street Journal did not account for the fact that there is a large number of directors and officers in the United States who receive grants. With such a large number of D&O, it was practically certain that some of them would receive most of their grants on days in which the stock price was particularly low, even just by chance.
For instance, while the odds of a particular executive receiving grants on the 1st, 5th, 50th and 3rd lowest price in four consecutive years are a million to one by the commonly used method, the odds of some executive, somewhere, having that level of luck is 99.9%, they claim.
For me, the most compelling part of the paper was a chart that showed rankings of companies by the average return in the 20 days after an option grant is made (it's page 5 on the pdf) . Given how many companies are suspected of gaming the system by backdating, one would assume there would be plenty of companies that tended to grant options on days just before the biggest 20-day runups. Instead, NERA's data shows a fairly classic bell shape curve. Most companies tended to be middle of the road. And only slightly more companies showed an ability to routinely grant options just before big stock run-ups, compared with those with the unfortunate ability to always pick the highest price days. Factor in the fact that many of these companies were actively trying to hit lows and that none were trying to hit the highs--not to mention the fact that some companies have in fact admitted that yes, they were backdating--and that slight disparity is easy to understand.
As they write:
Speculation has been rampant about companies that have been very lucky; yet nobody has been paying any attention to companies that have been very unlucky. Probability theory tells us that chance alone can produce both very lucky and very unlucky companies, and indeed we see that there are some of both types.
Now, I'm certainly no statistician, and I do have some questions--for example, would that bell curve have looked so innocuous if they'd only included companies in Go Go sectors such as tech. I'm going to try to reach some other academics tomorrow to do some of my own peer review of the NERA study. I'll report back with what I find.
Regardless, I find it telling that the authors say they were motivated to do the study in the first place by their frustration with the Journal’s continued use of that "random" comparison as a way to point fingers at specific companies--and by the rest of the business press’ willingness to carry on the practice. They don’t blame the academics, such as Lie. They point out that he was careful to point out that his analysis “is designed to uncover evidence of retroactive timing in the aggregate, and might be useless in identifying exactly which firms engage in such activities.” But they say the business press has been a bit too breathless in their pursuit of wrongdoing. Says Conroy, “when something [like all those impossibly long odds of issuing low-priced options) gets repeated enough, it starts to be considered a certainty.”