Henry Blodget, a former securities analyst, lives in New York City. Read his full-disclosure statement about his potential conflicts of interest in covering the Martha Stewart trial
Chance of conviction
Wednesday, March 3, 2004: 40 percent
Judge Cedarbaum read her jury charge this morning, and the jury began deliberating. Stewart's chances of conviction remain at 40 percent, Bacanovic's modestly higher.
After reading two hours' worth of jury instructions, Judge Cedarbaum placed the immediate future of Martha Stewart and Peter Bacanovic in the hands of 12 American citizens. The jury deliberated all afternoon, occasionally asking to see testimony, summary charts, phone records, and other evidence. In the early afternoon, the consensus in the marble hallways of the U.S. courthouse was that the verdict would arrive "Thursday or Friday." By late afternoon, the consensus was "Friday or early next week."
Over the course of the last five weeks, the Martha Meter—my best guess at the chances that Stewart will be convicted on at least one charge—has moved in a range of 18 percent to 60 percent. After closing arguments, I set the Meter at 40 percent, representing estimated chances of conviction of slightly more than one in three.
The Martha Meter, of course, is not a precision instrument. (The national ink expert, for example, would no doubt hurl it out of his lab.) The odds of accurately predicting what 12 diverse individuals will conclude after more than a month in a jury box and days or weeks in a deliberation room are decidedly worse than the odds at a craps table, and I don't pretend to have any special insight here. Suffice it to say that, for most of this trial, I have believed that the odds are against a Stewart conviction (with non-conviction including both acquittal and hung jury). I also believe, however, that the government put on a compelling case and that the deliberations are going to be contentious. Judging from a random, verbal, and entirely unscientific poll I conducted this morning, the consensus of the press corps is that both defendants will be found guilty. I won't be shocked, therefore, if the Martha Meter is wrong and the real jury returns a verdict of guilty on all counts.
Although I can't predict what the 12 jurors are going to do, I can figure out what I would do if I were a juror—and, at least on a preliminary basis, I have. As of this writing, I have deliberated with myself for approximately 10 hours. And I have reached a preliminary verdict.
Before revealing this verdict, I should say that, for me, this was an unsettling exercise. When I began my deliberations, I didn't know what conclusions I would reach, and I wasn't eager to discover that I believed the defendants should be convicted. As I have described, I have met Peter and Martha, and I like them, and I don't want either to go to jail. My discomfiture was heightened by the sense of irony and tragedy that permeates this case, a sense that, as a juror, I knew I wouldn't be able to factor into my deliberations. To reiterate: All of the crimes that Stewart and Bacanovic allegedly committed were a direct result of an investigation into a crime that they didn't commit (criminal insider trading based on a tip about the Erbitux rejection); the bloodthirsty political environment that spawned this indictment was, in its own way, as rabid, crazy, and ephemeral as the boom that preceded it; the only people hurt by Stewart and Bacanovic's alleged crimes—aside from a few bruised egos at the Justice Department—were Stewart and Bacanovic themselves. (Everyone else was hurt by the investigation, prosecution, and publicity surrounding the alleged crimes.) Again, legally, this doesn't excuse anything. It just muddies the moral waters.
I should also, once again, refer readers to my full-disclosure statement: My potential biases are so salient that I wouldn't have made it through the first 10 minutes of jury selection. As a result of my own experiences as a defendant, for example, as well as a decade of working on Wall Street, I have a "rooting bias": I regard corporate America as my home team and the government as the visitors (although less than I did at the beginning of this trial). Moreover, because I believe this case falls within the bounds of "prosecutorial discretion"—i.e., that the government could have just absorbed the blow to its ego, given Martha and Peter a pass, and rushed off to pursue some real securities fraud (there are undoubtedly other WorldComs out there)—I wouldn't mind seeing the government get egg on its face. I should also say, however, that in the last five weeks, I have developed great respect for—and acquaintances with—many of the particular representatives of the government in this courtroom, and I have been impressed with not only their professionalism and integrity but their ability to shape my perceptions. The case was clearly strong enough to have been brought, and I would take no pleasure in seeing the government lose.
Finally, I should make clear that this is my preliminary verdict. In her jury instructions, Judge Cedarbaum emphasized that jurors should stick to their guns unless and until they are convinced that they are wrong, at which point they should change their minds. Although the hallways, cafeterias, and bathrooms of the courthouse are not formal jury rooms, plenty of deliberation takes place in them, and in the hours, days, or weeks until the real jury returns a verdict, I will remain open to changing my mind.
So, without further preamble, here is my verdict:
Defendant Martha Stewart: not guilty on all counts.
Defendant Peter Bacanovic: not guilty on all counts.
In the next day or two, I hope to detail my reasoning and conclusion with regard to all eight counts. (If the jury returns before I finish, I won't bother.) In this dispatch, however, I'll stick to the macro issues and one of the four key questions that weaves its way through multiple charges: whether the government proved beyond a reasonable doubt that Bacanovic and Stewart did not agree that Stewart would consider selling ImClone if the price broke $60.
First, the macro view. From a high level, in my opinion, the government's theory makes sense. When considered as a whole, the evidence supports the conclusion that, on Dec. 27, 2001, Martha Stewart was given privileged, non-public information, that she traded while in possession of it, that she and Peter Bacanovic knew that she should not have been given it, that she and Bacanovic endeavored to avoid revealing that she had been given it, and that such endeavors may have included the commission of some or all of the alleged crimes. In the case of some of the allegations, moreover, the government's theory is more plausible than the defense's theory.
That said, as Judge Cedarbaum emphasized in her jury instructions, the government's allegations must be analyzed not just from the top down (do they fit the macro theory?), but from the bottom up (do they meet the criteria and burden of proof for each specific crime?). At the heart of the government's burden, of course, is the presumption of innocence: To warrant a guilty verdict, each charge must be determined not just to be possible or plausible, but to have been proven beyond a reasonable doubt. Judge Cedarbaum:
What is a reasonable doubt? The words almost define themselves. It is a doubt based on reason and common sense and arising from the evidence or lack of evidence. It is a doubt that a reasonable person would have after carefully weighing all of the evidence. It is a doubt that would cause a reasonable person to hesitate to act in a matter of importance in his or her personal life. Proof beyond a reasonable doubt must, therefore, be proof of such a convincing character that a reasonable person would not hesitate to rely and act upon it in the most important of his or her own affairs.
"Beyond a reasonable doubt" is not the same as "beyond any doubt." Regardless of how convincing the evidence, some uncertainty remains. (As Judge Cedarbaum is fond of observing, anything is possible.) Experts differ on quantitative measures of reasonable doubt; for the purposes of this trial, I am defining it as more than 85 percent to 90 percent certainty. This level of certainty exceeds that which I would place on some of my own "beliefs." After five weeks of listening to arguments and evidence, I have developed beliefs about what really happened on Dec. 27 and in the ensuing weeks, beliefs that, in many cases, fall somewhere between the black-and-white scenarios described by the government and the defense. Some of these beliefs exceed the "reasonable doubt" threshold; some don't. In some cases, therefore, the beliefs match the government's, but the level of certainty is not high enough to exceed reasonable doubt.
On to the charges. Now that Stewart has been acquitted of securities fraud, most of the remaining eight counts in the indictment depend on one or more of the following assertions, each of which the government must prove beyond a reasonable doubt:
- Stewart and Bacanovic did not have a conversation on Dec. 20 or 21, 2001, in which they agreed to discuss selling Martha's remaining ImClone shares if the price fell to $60.
- Stewart and Bacanovic lied about (and/or concealed) what happened on Dec. 27, 2001.
- Stewart and Bacanovic lied about not discussing the government's investigation into Stewart's ImClone trades prior to their interviews with investigators.
- Stewart, Bacanovic, and Douglas Faneuil, or at least Bacanovic and Faneuil, conspired to mislead investigators.
The $60 Agreement
In my opinion, the government did not prove beyond a reasonable doubt that the $60 agreement did not exist. Similarly, the defense did not prove that the agreement did exist. For the purposes of determining a verdict, however, only the former matters. Here are the factors that, in my mind, added up to reasonable doubt:
- The entries on Peter Bacanovic's "allegedly altered worksheet" that weren't allegedly altered—the check marks, circles, and stock symbols—suggest that Bacanovic and Stewart discussed Stewart's ImClone stock during a review of Stewart's portfolio on Dec. 20 or 21, 2001, a review conducted to finalize plans for that year's tax-loss selling. All of the stocks that Stewart sold on Dec. 21 were marked with check marks, presumably signifying that they were to be sold immediately. Three of the stocks on the sheet—Apple, Nokia, and ImClone—were marked with circles, presumably signifying that they were to be sold later. Apple and Nokia were sold on Dec. 24. ImClone wasn't. The future sale of ImClone, therefore, was presumably subject to a different trigger than the sales of Apple or Nokia. This trigger may have been, as Bacanovic and Stewart claimed, a fall in the price of the stock to $60.
- The October sale of Stewart's 51,800 shares of ImClone in her pension funds occurred at a price of approximately $60-$61. This suggests that Stewart was willing to sell at this price level. This, in turn, suggests that she might have agreed to consider selling if the stock again fell to this level. She also previously indicated a desire and/or willingness to sell all of her personal shares of ImClone—5,000—by tendering them in October. The 3,928 she had left were the ones she couldn't sell in the tender.
- The fact that Stewart and Bacanovic did not use a formal "stop-loss" order does not mean they didn't intend to consider selling if the stock hit $60. In fact, the lack of a stop-loss is irrelevant. Many investors don't use stop-loss orders (or, for that matter, any kind of automatically executed limit orders). Many investors who do occasionally use them don't always use them. Stewart may not have been sure she wanted to sell ImClone at $60, and given ImClone's extreme volatility, if she had used a stop-loss, the likelihood that she would be stopped out by routine, meaningless volatility was high.
- Bacanovic's testimony at the SEC about how he and Stewart arrived at the $60 agreement was convincing. He wanted her to sell; she didn't want to. He pushed her; she pushed back. He said, "How low does it have to fall until you sell it?" She said, "I don't know." He said, "$60?" She said, "Sure." This is the kind of banter brokers and clients engage in all the time.
- Douglas Faneuil did not testify that the $60 agreement did not exist. He was also in no position to know whether it existed or not—he was on vacation when Stewart and Bacanovic allegedly entered into it—and neither Stewart nor Bacanovic would necessarily have told him about it. Faneuil did describe the $60 agreement as a "cover story," but he appears to have taken dramatic license with some of his testimony (support for this later), and he might have done the same here. Faneuil testified that Bacanovic told him the $60 story was a true story, not a cover story, and Bacanovic would have known. At best, therefore, Faneuil suspected the $60 agreement was a cover story. (This issue is separate from whether Bacanovic was trying to pressure Faneuil to keep his mouth shut about the Waksals' sales).
- According to Faneuil's testimony, on Dec. 27, 2001, after Faneuil told Stewart the Waksals were selling, but before she placed her sell order, Stewart asked for a price quote. This suggests that the stock's price might have been important to her. The price Faneuil would have given her, moreover—approximately $58—was below $60. This is at least consistent with the idea that the stock's price was relevant to Stewart's decision to sell and that she might previously have discussed the possibility of selling if the price fell below $60.
- It is true that the phone message Bacanovic left for Stewart with Ann Armstrong on the morning of Dec. 27, 2001 ("Peter Bacanovic thinks ImClone is going to start trading downward") does not mention the stock's price. It is also true that, at the time Bacanovic called Stewart, the stock's price was not yet below $60. However, the question as to what, on the morning of Dec. 27, prompted Bacanovic to call Stewart is a different question than whether, on Dec. 20 or 21, he and Stewart agreed to consider selling the stock if it fell to $60.
- The lack of an ImClone "@60" notation in Bacanovic's ACT notes, in which he kept detailed records of communications and "to dos," does not mean the agreement didn't exist. Many of Bacanovic's client communications and activities did not find their way into his ACT notes, including "to dos." For example, after the tax-loss selling discussion on Dec. 20-21, Bacanovic did not make an ACT note entry to remind himself that he had to sell Apple and Nokia at a later date. So, the ImClone omission may not be significant.
In my opinion, all of these factors add up to reasonable doubt about the assertion that the $60 agreement did not exist.