A question that you may or may not find interesting is: have the U.S. government’s rather strenuous recent efforts to stamp out insider trading actually reduced insider trading? How would you go about answering that, if you really wanted to know? I guess the right approach would be a survey; like, go email every hedge fund manager and ask “have you insider traded in the last 12 months? more or less than you used to?” and see what they say. That has … problems, so you look for proxies. Do stocks tend to go up, on heavy volume, before the announcement of secret good news? Then that at least suggests that someone traded on the secret good news before it was announced. It’s something.
A while back I idly committed some pseudoscience about pre-merger trading and found some indications that (1) stock prices and volumes tend to increase before mergers and (2) that increase has been more pronounced in say 2009-2013 than it was in say 2001-2008. This would seem to be weak evidence of increasing insider trading? This was a little puzzling given:
- those strenuous efforts, lots of people going to jail for long periods, etc.; and
- my assumption, anyway, that traders would be rational and competent judges of risk and reward who would weigh the increased odds of being caught and sent to prison in their decision to insider trade or not.1
But there my pseudoscience was. Anyway I learned today (via) about a recent study where some b-school professors committed some … I dunno, whatever b-school professors do, something between “pseudoscience” and “science” … of their own and got the opposite result, so I figured I’d pass it along. Here’s the abstract:
On October 16, 2009, the U.S. government charged Galleon hedge fund founder Raj Rajaratnam and five others with insider trading, in what was described by a key prosecutor overseeing the case as a “wake-up call to Wall Street and to every hedge fund manager.” We find that the mean abnormal stock price runup of targets (a measure of informed trading) during the 26 months since the inception of the Galleon case declined from 5.12% to 2.84%. The early evidence strongly suggests that the Galleon case has sent a clear signal to the traders, and that the traders are listening.
Their pseudoscience is more science-y than mine, meaning principally that their sample was more robust and their stock price runup statistics are adjusted for market returns over the relevant period, which is sensible. The core finding, though, is something like:2
- In the 30 days before a merger is announced, the target’s stock was up on average 4.6% before the Galleon case was announced, and 2.5% after, and
- In that same period, the target’s stock traded 320% of its normal volume before Galleon, and 290% after.
Obviously other things have changed in the M&A (and trading) landscape since October 2009, so you could take this with some amount of salt, but they do some multivariate analysis controlling for some merger characteristics (size, sector, private-equity deals, rumors, multiple bidders, etc.) and still find a smallish Galleon effect:
When using the abnormal runup of the target over the (-42, -1) window, the coefficient for the GALLEON variable is -5.56% (significant at the 1% level), which implies that the abnormal runup declined by an estimated 5.56% since the inception of the Galleon case when accounting for all other control variables.
Anyway, I pass it along because I wouldn’t want you to rely on my pseudoscience when there’s, like, practically science out there on the matter, and it comes to the opposite conclusion. The conclusion being, specifically, that the SEC’s and DOJ’s five-plus years of work to catch Raj Rajaratnam and his network of insider traders seems to have reduced pre-merger insider trading by about 5%.3
This is far from conclusive, of course. For one thing, merger news is the easiest thing to study, since mergers tend to have a large and predictable effect on price, while a lot of insider trading is of the earnings-preview or product-launch/approval variety, so this tells you only about one subset of the insider trading world.4 For another, it’s a weird subset; M&A market dynamics change faster than, like, earnings-call dynamics, so it’s hard to know if the authors controlled for all of the things that might have reduced the post-Galleon pre-merger runups.
Also: stock volume is still like 3x its normal levels for merger targets in the month before the merger is announced, which makes all of this sort of troubling. Either the large bulk of pre-merger trading is not informed by inside information, and therefore these results are only a vague gesture in the direction of insider trading – or it is informed, and the crackdown on insider trading has left just tons and tons of insider trading still out there. I dunno. I plan to hold out for the “how much do you insider trade?” survey.
Inga Chira & Jeff Madura: Impact of the Galleon Case On Informed Trading Before Merger Announcements [SSRN]
Rajaratnam Prosecution Sends a Signal: Traders Notice [AllAboutAlpha]
Earlier: Crackdown On Insider Trading Seems To Have Led To More Insider Trading
1. It’s fun to think of alternative mechanisms here. Like one possibility is that insider trading is completely irrational: for a markets professional, a small but nontrivial chance of going to jail is never worth the marginal additional profit you can make trading in advance of news. That leaves only the irrational, the psychopathic, the self-destructive, etc., as the only insider traders, and so not really subject to deterrence. On this theory insider-trading deterrence works about as well as it can, and increased publicity, punishments, etc. won’t improve it. That sort of nonlinearity seems pretty unscientific but I’m nonetheless attached to it because, really, insider trading is nuts, isn’t it?
2. There’s a less marked but still significant effect if you take a shorter pre-Galleon period:
3. Ehhh reduced the pre-merger stock price runup by about 5%. (Percent not percentage points.) They don’t run the multivariate tests on the abnormal volume, which I guess would be a more direct test of actual trading.
4. Someone should do a similar study on pre-earnings insider trading (like, T-10 to T-1 change divided by T-1 to T+1 change). The problem there is that knowing what earnings will be doesn’t necessarily mean you’ll trade the right way. “Oh they beat expectations but the price dropped” etc. You don’t see enough about people who insider trade ahead of earnings with perfect information but nonetheless get the direction wrong.