Emory University accounting professor Shivaram Rajgopal points an accusatory finger at Securities and Exchange Commission employees and proclaims a pattern of selling stocks of companies subject to enforcement actions. His study finds “significant abnormal returns of (i) about 4% per year for all securities in general; and (ii) about 8.5% in U.S. common stocks in particular. The abnormal returns stem not from the buys but from the sale of stock ahead of a decline in stock prices.”
Rajgopal throws the big rock:
Most of these returns stem from the timely sale of these stocks, suggesting that a regulator’s employees are most likely to know about sanctions against companies before the market as a whole.
The study is based on reported securities trades by 3,500 SEC employees during late 2009, and for all of 2010 and 2011. However, the data is severely constrained. Rajgopal merely had a list of transactions and no ability to compute the profits or losses. There is also no data on holdings. The study only looks at what was bought and sold. Rajgopal constructs a synthetic hedge model in an attempt to model the trading.
I’m going to assume his analysis is correct and that SEC employees were more likely to sell in companies that become subject to SEC enforcement actions. Rajgopal claims this is illegal trading before public announcements. I think it’s just SEC employees over-emphasizing SEC actions as a reason to sell the stock.
One fault in Rajgopal’s accusation is how limited the news about an enforcement action may be. The SEC is a huge organization. But even if the news of an enforcement action is leaky, only a small percentage of the 3,500 employees would have that information and be able to make the illegal trade. That small number would not be as statistically significant as Rajgopal finds in his study.
The only meaningful part of the report is its focus in Table 3 of 87 trades of the 7,200 employee trades studied. Those 87 trades are in Bank of America, General Electric, Citi, Johnson & Johnson, JP Morgan, and General Electric in the period prior to the announcement of enforcement actions.
The trades highlight the need for preclearance. An organization may have material non-public information. But the information is only seen by a subset of employees. Clearly, you can track document and email traffic to prove that someone knew that information. That’s what the SEC does in its insider trading investigations. The tough part is defending from an accusation of having the knowledge.
It’s hard to prove that you didn’t know something.
The SEC is stuck with an accusation and little way of proving that those 87 trades were not made on material non-public information. Clearly, the information existed within the SEC. Perhaps the SEC can find a smoking gun that proves that some of those 87 were made by employees who knew about the enforcement action. I would guess that majority were made without that knowledge and no way to prove that they lacked the knowledge.
The SEC should have a pre-clearance requirement or a planned sell window so that the SEC and its employees can avoid the taint of accusations like Rajgopal’s accusation. That’s why public companies have 10b-5 plans and registered investment advisers have pre-clearance requirements.
- Study Finds SEC Staff Sold Shares Before Settlements Made Public by Dave Michaels in Bloomberg
- Stock Picking Skills of SEC Employees (.pdf) by Shivaram Rajgopal