• 19 Feb 2013 at 3:16 PM

SEC Ruins One Anonymous Swiss Person’s Lucky Day

The SEC’s Heinz insider trading complaint is a delightful epistemological puzzle. The story goes that, on February 13, some person or persons with a Swiss bank account bought 2,553 Heinz June $65 calls (255,300 shares), with the stock at just over $60, for $0.30 – $0.40 per share (around $90,000 total). The next day, Heinz announced that it was being bought for $72.50, the calls went to ~$7.33, and he/she/they made almost $1.8 million on paper. That’s a one-day return of 1,979%, or approximately Error 1 annualized.1

Which is pretty good! Except those gains are on paper, and that paper has been impounded by a federal judge at the SEC’s request, because that story is pretty much too shady not to be insider trading. But what I just told you is the whole story that the SEC appears to know: there’s no proof that it’s insider trading as opposed to just fabulously lucky trading. When I win the MegaMillions next week my one-day return will be considerably in excess of 1,979%, and no one will have leaked the winning numbers to me in advance, as far as the SEC knows anyway. Though it is suspicious that volume in the HNZ $65 calls was up just a smidge that day:

Despite the thin complaint, you can’t really criticize the SEC much here. They’re not throwing anyone in jail without evidence, they’re just impounding the money to look for the evidence, and their suspicions certainly seem reasonable. Here is a good DealBook column describing the SEC’s approach:

If the securities are frozen in a Goldman account in the United States, then anyone who wants to claim them will need to make an appearance in this country and agree to be subject to the jurisdiction of the federal court. That would give the S.E.C. a very big carrot to try to lure the trader out into the open.

Sure, though there is a bit of a logical flaw. If you were looking to prove that you are the sort of big-balls’ed, semi-irrational risk-taker who would gamble $90,000 on out-of-the-money 4-month call options with no inside information, showing up in court to claim this money should do the trick.2 Actually if no one else does I may claim it. Judge Rakoff, I am Certain Unknown Traders in the Securities of H.J. Heinz Company! Now give me my money!

Sometimes I get nervous about SEC data-mining exercises. Is that spike in volume suspicious? Of course it is. But other similar spikes in advance of recent deals do not seem to have caused similar investigations. And, it goes without saying, other similar spikes in out-of-the-money options volume that are not followed by massive price moves aren’t investigated. Here’s some trading in Heinz’s June $55 puts last year:

These puts, which were around $3 or so out of the money, traded just 67 contracts in the whole three weeks before November 30, and then traded 2,054 contracts in one day! And then … and then nothing happened. The stock drifted slowly up for a few months, and then it was LBO’ed. The last trade today was at $0.05.3 If you spent $370,000 buying those 2,054 puts in November you’re just a schmuck. If you spent $90,000 buying the 2,553 calls last week, on the other hand, you are (1) rich but also (2) innocent-until-proven-guilty-but-c’mon.4

Recent SEC pronouncements have made it creepily clear that the agency considers any outperformance to be suspicious, and possible evidence of illegal activity. If you beat the market, you probably cheated, and you’ll be hounded about it until your legal fees devour your alpha.

This is of a piece with the SEC’s increasing reliance on financial-academia-approved methods of rooting out fraud: financial academia is a bastion of efficient-markets-hypothesis fundamentalism, and increasingly, it seems, so is the SEC. The SEC’s EMH fundamentalism, though, is kind of self-reinforcing: if active stock-picking will result either in (1) missing your benchmark or (2) beating it and getting investigated by the SEC, then why do it?

This growing negative convexity to “edge” is probably bad for market efficiency and all that. Somebody’s gotta do the research, go to the management meetings, and generally pick stocks. If all of those efforts are viewed with suspicion, then there’ll be less of them.

In the abstract that’s bad: somebody should be doing the work to allocate capital intelligently. But in the specific … I mean, the trades at issue here are big outright bets in listed options from unnamed accounts. I am fond of Felix Salmon’s argument that “stock picking, at least when practiced by individuals, is best analyzed as an upper-middle-class hobby rather than as purely profit-focused investing activity.” If stock-picking is like building model railroads, day-trading listed options is like sniffing glue. It’s hard to get too upset if the SEC makes it even less attractive.

SEC Freezes Assets in Swiss-Based Account Used in Suspected Insider Trading Ahead of Heinz Acquisition [SEC]
Complaint: SEC v. Certain Unknown Traders in the Securities of H.J. Heinz Company [SEC]
For S.E.C., Obstacles Aplenty in Pursuit of ‘Suspicious’ Heinz Trading [DealBook]
Options Activity Questioned Again [WSJ]

1. According to the TI BA II Plus I use at Dealbreaker HQ. I am aware that I am on record saying that “using a non-RPN calculator is no way to live,” and I stand by that. The way I live is, it turns out, no way to live.

Update: an earlier version of this post said it was a 198% 1-day return, which … which I blame on the calculator? Going with that. The annualized return remains correct.

2. At least in my book. And I’m not alone. As Peter Henning notes:

A case I wrote about last year involving a foreign trader was dismissed when the S.E.C. could not show anything more than suspicious trading. Much like the Heinz trading, the defendant bought out-of-the-money call options right before a deal, generating a 1,000 percent profit in a short period. But the defendant appeared in the United States and fought the case, and the S.E.C. was never able to show any connection to a source of inside information.

3. Which seems exceedingly generous, no?

4. Incidentally I just bopped around Heinz options until I found an analogous 100x increase in volume that did not pan out. Listed options being thinly traded things, I am sure there are zillions of other examples in non-Heinz-y companies, but I have no skills to find them in any systematic way. (Suggestions welcome!) Point is, only the winning lottery tickets get scrutinized.

14 comments (hidden to protect delicate sensibilities)
Show all comments ↓

Comments (14)

  1. Posted by Common Sense | February 19, 2013 at 3:20 PM

    Well, if there wasn't so much cheating, scams and rip offs, then the SEC wouldn't be so trigger happy.

  2. Posted by Relational Investors | February 19, 2013 at 3:29 PM

    Maybe if you opened your mouth and articulated your thoughts through writing less people would like you more.

  3. Posted by Querying Commenter | February 19, 2013 at 3:30 PM

    Why don't you have your HP 12C at the office? Did you loan it to your niece again so she can learn the joys of RPN in kindergarten?

  4. Posted by Matt's Excel 2013 | February 19, 2013 at 3:40 PM

    epistemological ?

    Matt, can I pis on you? It's the only logical thing we have left.

  5. Posted by Confused HP 12C user | February 19, 2013 at 4:08 PM

    $1800k / $90k = 198%?

  6. Posted by PC Police | February 19, 2013 at 4:09 PM

    If it could be credibly claimed that a female woman of color had overcome oppression to invent RPN, then and only then could you be sure that RPN would be taught in kindergarten.

  7. Posted by Guest | February 19, 2013 at 4:25 PM

    The passion for RPN always struck me as being a bit bizarre, given that virtually every modern programming language can handle full expressions, and that stack order is only needed for very primitive languages like assembly. Outside of the testing center, there's no reason not to use a fully programmable calculator or Excel, either of which can handle full expressions. Inside the testing center, anything that can handle TVM computations with a key sequence that you can remember will be good enough.

    On a related note, I am gratified to find that Matt is using the exact same trusty cheap-but-gets-the-job-done TI BA-II+.

  8. Posted by UBS Algo Stud | February 19, 2013 at 4:33 PM

    Huh? Only extremely advanced languages can handle the kinds of expressions your discussing, also, in what world is a TI BA-II+ cheap?

  9. Posted by guest | February 19, 2013 at 5:03 PM

    The first world?

  10. Posted by OECD Quant | February 19, 2013 at 5:17 PM

    So definitely not UBS Stamford, gotcha.

  11. Posted by quant me maybe... | February 19, 2013 at 5:45 PM
  12. Posted by quant me maybe... | February 19, 2013 at 5:49 PM

    RPN is a religion, not an input method.
    It's also really really good for doing long multi-step equations.

  13. Posted by Bbb- | February 20, 2013 at 9:14 AM

    You have a bb terminal but don't have a calculator?

  14. Posted by UBS joke aficionado | February 20, 2013 at 10:14 AM

    comment 10 for the win