A loose but useful way to think about U.S. insider-trading law is that it is supposed to encourage research, but discourage cheating.1 So if you read a company’s financial statements, examine its stock chart, inspect its products, survey its customers, compare its competitors and otherwise do the work of obtaining information that could affect its stock price, you can go ahead and trade on that information, and we all hope that you are rewarded for your efforts with an above-market return. But if you just play golf with the chief executive officer, and he tells you that the company is being acquired next week, we don’t want you trading on that information, and if you do we want you to be rewarded with prison.
That is the intuition. It is not quite the law. It is not quite the law because, in the real world, it is not always easy to distinguish research from cheating. In practice, a lot of the stuff that investors do to research a company involves talking to the company.2 Sometimes this occurs out in the open, like on a public earnings call. Sometimes it occurs out in the semi-open, like at an industry conference. Sometimes it occurs in private conversations with the company’s investor-relations employees, whose job is, as the title implies, to have conversations with investors. Sometimes it occurs in private conversations with the company’s executive management. Sometimes, I suspect, it even occurs in private conversations with the chief executive officer on the golf course.
You could imagine a world in which we drew the line by just banning all non-public conversations between companies and their investors: Public sources are research, but private conversations are cheating. It would in some ways be a simpler world. But for the most part, in the U.S., that’s not what we do.3 We’ve concluded — and the ”we” here is pretty abstract — that those private conversations can be good. We want shareholders to be able to have an open exchange of views with the managers they’ve hired to run the companies they own. We want them to develop good hard-hitting questions and demand honest answers. We want companies to be able to meet with investors to try to persuade them to buy stock. We want Bill Ackman to be able to call up Valeant’s CEO and ask, “Mike, is there any fraud going on at the company?” These things feel like research; they make prices more accurate and markets more efficient, so we want to encourage them.4
Still, just getting tips from the CEO on the golf course feels icky.
And so U.S. insider-trading law presents an interesting line-drawing problem. Very loosely, we’re okay with you trading on material nonpublic information that you got from the company, but not so much with you trading on material nonpublic information that you got from your shady buddy at the company. Since companies can only act through humans, the way we say that is that if an insider gives you information as part of his job, you can trade on it, but if he gives you the information in breach of his duty to the company, then you can’t trade on it.
Still, though, how can you tell? Part of the job of an investor-relations person, or of a CEO for that matter, is to develop good relationships with investors. Part of the job of an investment analyst is to develop good relationships with corporate management. A CEO and an investor might chat regularly, exchange small talk, be personally fond of each other, even play golf together. And the CEO might give that investor nonpublic information. And that might still be okay: Warm as their relationship might be, it is still essentially a business relationship, and their exchange of information is still done on the company’s behalf rather than in breach of a duty to the company.
The way you can tell that it’s a breach of duty is this:
The test is whether the insider personally will benefit, directly or indirectly, from his disclosure. Absent some personal gain, there has been no breach of duty to stockholders.
That’s from a 1983 Supreme Court decision called Dirks v. SEC, and it makes a kind of rough sense. If the insider is getting a kickback for giving you the information, he’s probably not doing it on behalf of the company: He’s probably doing it for himself, and in breach of his duty to the company.
But the Dirks test doesn’t entirely correspond to our intuitions. Sometimes it is too harsh: Every insider gets some benefit out of every disclosure, even if it is just the warm feeling of knowing that he helped out a business acquaintance. For a long while after Dirks, prosecutors were making arguments exactly like that: They had to prove a personal benefit to prove insider trading, but that was not a high bar; just the joy of helping out another human was enough of a personal benefit. But in December 2014, a federal appeals court in New York put a stop to that in its Newman decision, ruling that the personal-benefit requirement, “although permissive, does not suggest that the Government may prove the receipt of a personal benefit by the mere fact of a friendship, particularly of a casual or social nature,” and that “in order to form the basis for a fraudulent breach, the personal benefit received in exchange for confidential information must be of some consequence.” Giving someone inside information in exchange for a sack full of money: illegal. Giving someone inside information and getting a vague warm feeling in return: fine.
Prosecutors were very upset about Newman and tried to get the Supreme Court to reverse it, but the Supreme Court ignored them. And rightly so, I think. The Newman case involved hedge-fund analysts who got allegedly material nonpublic information from an investor-relations employee at Dell and passed it on to other investors. In exchange, the investor-relations employee got some vague “career advice.” No one got sacks of money. There were some awkward facts,5 but for the most part the case seemed to be on the research side of the research/cheating divide. Sending Newman to prison would make it harder for other investors to do their jobs of finding stuff out about companies.6
But if you read the Dirks test to require a real benefit, then it might be too lenient for our intuitions. Sometimes an insider tip is obviously cheating, even though there’s no quid pro quo. If a CEO and an activist hedge fund manager discuss ideas for improving the company together over a round of golf, and the hedge fund manager buys or sells stock based on the conversation, that could well be the sort of interaction that makes our capital markets more efficient. But if a CEO and his brother-in-law just enjoy playing golf together, and the CEO tips off the brother-in-law to an upcoming merger not as part of any business-related conversation but just to help a family member make money, there’s no efficiency gain in that. That’s just cheating. It would clearly be illegal for the CEO himself to trade on the information; giving it to a family member seems equally illegal and for the same reasons.