Streetwise Professor

November 4, 2015

I’m Not Spoofing You About Judicial Overkill

Filed under: Commodities,Derivatives,Economics,Energy,Exchanges,Regulation — The Professor @ 11:02 pm

Yesterday in a federal court in Chicago, Michael Coscia of Panther Energy Trading was convicted of six counts each of commodity manipulation and fraud for “spoofing.” Coscia faces decades in prison.

I haven’t seen the evidence, so I cannot judge whether Coscia did manipulate. For the purposes of this post I will stipulate that he did. But even given that stipulation, this entire exercise was judicial overkill and a travesty that can do serious damage to the markets.

What is spoofing? A trader puts in a large order (an offer, say) several ticks away from the best price in the market. He also places a small limit order on the other side of the market (a bid, in this example). If the market moves towards the large order, the spoofer cancels and replaces the order several ticks away from the new inside market. It is this cancellation that attracts all the attention. Much of the coverage says that the spoofer submits orders with the intent of canceling them.

That’s not the whole story though. The point of spoofing is to increase the odds that the small order is executed. After all, what would be the point of submitting orders that are never executed?

In the example, the large sell order is intended to convince others that the current price is too high. This may induce some bidders to cancel their bids, and others to cross the market and hit the bid. Both of these actions increase the odds that the spoofer’s small order will be executed.

So how does he make money? It can’t be by driving down the price persistently. The spoofer has bought: to profit, prices must rise subsequently. So, often the spoofer will reverse direction, putting in a big bid away from the market, and a small offer.

If spoofing works, the spoofer will repeatedly buy at the bid and sell at the offer, making the dealer’s turn. This will not cause the price to diverge persistently from where it would be, absent this conduct.

That’s apparently what happened with Coscia. He made a whopping $1K on the six episodes for which he was charged and convicted. He was just making a tick here and a tick there. And crucially, unlike the kinds of manipulation that cause real damage-corners, in particularly-he is not causing the price to be persistently inflated or depressed.

So who is hurt? Some people may be induced to trade when they wouldn’t have absent the spoofing. Their losses are approximately equal to the spoofers gains, on the order of a tick. And since some might have hit the spoofer’s bid even absent the spoofing, only a fraction of those with whom the spoofer trades are damaged.

Others who might be damaged are those who are fooled into canceling orders, and see the spoofer execute a trade they would have liked to if they hadn’t been fooled. The spoofer takes some of the profit they would have earned.

I find it hard to believe that these damages are are all that large. (They would also be hard to estimate because it is virtually impossible to identify who traded because of the spoofing, and who pulled a quote because of the spoofing, and gave up the opportunity to trade.) And regardless, this is exactly the kind of conduct that can be deterred using monetary fines.

This brings me to another bizarre aspect of these spoofing cases. Many of those who scream loudest about spoofing, like Eric Hunsader of Nanex, say: “SPOOFING IS SO OBVIOUS!!!! JUST LOOK AT THE DATA!”

As another case in point, I saw a Tweet embedding a .gif of someone’s TT trading screen, in which the quoted depth a couple of ticks above the inside market would rise and fall by 800 contracts or so. The Tweeter (I can’t find the Tweet) said something to the effect “look at this obvious spoofing.”

Well, I agree with the obviousness of it, but the implication of that is exactly the opposite of what Hunsader, the Tweeter, and presumably the DOJ and CFTC believe: If it so obvious, nobody is fooled. If nobody is fooled, it can’t affect trading behavior or prices. If it doesn’t affect trading behavior or prices, there is no economic harm. If there is no economic harm, it shouldn’t be prosecuted.

There is a law and economics take on this too. Classic Gary Becker analysis shows that draconian penalties-like 25 years per fraud charge and 10 years per manipulation charge-are justifiable if the probability of detection of a harm/crime is small. This is necessary to make the expected cost paid by the offender equal to the cost of the harm. But if the conduct is obvious, even only ex post, the probability of detection should be high, so penalties far greater than any harm are excessive. In this case, grotesquely excessive. (Furthermore, again pace Gary Becker, incarceration of a defendant who can pay the monetary value of the harm caused is a social waste, in the form of the cost of imprisonment, and the lost output of the convict.)

But it gets worse. The Coscia prosecution-and the popular condemnation of spoofing-focuses obsessively on the large rate of order cancellation. But perfectly legitimate market making strategies involve large rates of order cancellation, especially in volatile markets. They also involve buying frequently at the bid and selling at the offer. Given the Javert-like zeal of prosecutors, their dim understanding of trading, and the difficulty of explaining market making to a jury create the very real risk that a market maker could be charged, and convicted, and be punished severely, because he cancelled a lot of orders, and made the dealer’s turn all day long. This huge and very real risk will no doubt lead to less aggressive quoting (a market maker is less willing to quote aggressively if he is reluctant to cancel too often for fear of being accused of spoofing).

And who pays for that? Market users, including both institutional and retail traders, who take the liquidity market makers supply. Thus, you and me are harmed by overzealous prosecution of spoofing that threatens to demoralize legitimate, efficiency-enhancing trading.

This raises the very real possibility that the prosecution of actions that produce little economic harm will inflict a far larger harm. That is perverse.

What is particularly infuriating is that enforcement authorities are apparently incapable of prosecuting much truly damaging market conduct. Federal prosecutors are crowing over getting Coscia’s scalp, and the Chairman of the CFTC is using the verdict to intimidate would-be spoofers, but 6 years ago Federal prosecutors in Houston utterly botched the BP propane corner case (US v. Radley). That was a real manipulation that caused real damage. But the prosecutors totally flubbed the case, and the perps walked. Then there are those obvious manipulations that the Feds haven’t even bothered to prosecute (perhaps to spare themselves the embarrassment of flubbing another one.)

It reminds me of the old joke about the lawyer who said: “I lost the cases I should have won, and won the cases I should have lost. Therefore, on average, justice was done.”

No, actually, Mr. Lawyer: justice is never done if the guilty walk free and the innocent are punished. And sad to say, US manipulation law is perilously close to embodying that cynical joke.

 

 

Print Friendly, PDF & Email

9 Comments »

  1. But this is just part of the broader picture in which the USA has become the world’s first judicial dictatorship. The people who interpret the laws now effectively make them. If Supreme Court judges don’t like laws passed by elected representatives they cancel them. Who will judge the judges?

    Comment by David Blake — November 5, 2015 @ 2:34 am

  2. I think you’ve got the conclusion backwards – spoofing comes with negative externality as market makers will quote wider bid/ask spread as to compensate for the losses incurred to spoofers. Conversely with very draconian prosecution of these sort of gimmicks, market markers will quote tighter as the risk of getting picked off is lower. Your argument about market making operations worrying about being accused of spoofing also doesn’t make much sense – cancellation aren’t the issue here, but rather the use of outsize orders to skew the visible liquidity in the limit book. Virtually all market makers that round turn for 1 tick scalp use the same order size, in order to preserve priority in the limit book.

    Comment by pxm — November 5, 2015 @ 7:29 am

  3. Agreed. These potential penalties seem unusually harsh.

    Comment by WB — November 5, 2015 @ 8:16 am

  4. @pxm-I disagree. If anyone can detect spoofers, it is HFT market makers. Their comparative advantage is filtering out the signal from order flow noise. Further, it is liquidity takers that lose from spoofing (if anyone does). As discussed in the post, the spoofer’s strategy relies on someone hitting his bids and lifting his offers. It is liquidity demanders, not liquidity suppliers, that an effective spoofing strategy hurts.

    Regarding cancellations, you are mistaken. “Intent to cancel” is written into the statute and the regulations as a predicate of the offense. The prosecution focused on cancellations.

    The ProfessorComment by The Professor — November 5, 2015 @ 11:51 am

  5. It seems to me that spoofers are actually doing a service by keeping other traders from detecting large orders as easily. If spoofing is allowed, it is harder to tell if a large order indicates a mutual fund trying to buy or sell a large block of stock or just a spoofer. So whoever wants to make a large trade – such as a mutual fund or pension fund that may have many small investors – could get more executed before anyone catches on. That hurts some traders who were used to using the information in the order book for their own profit, but it doesn’t seem to hurt those who buy and sell for investment.

    Comment by Steve — November 5, 2015 @ 12:32 pm

  6. @The Professor

    The issue is that outsize bona fide orders are actually predictive of future price move and there isn’t a very good way to discern those from people trying to game the market; after all these is a reason why HFT market makers have universally came out against the practice. Functionally and from what I’ve professionally seen, as a market maker you just end up quoting wider to make up for it. To be more specific: when a bid ticks up, or offer ticks down, under the spoofing scenario market makers won’t immediately follow and quote that new bid/ask, as the spoofing strategy makes these price moves ephemeral. Conversely, defensive market makers will also cancel the other on the other side, because there is a non zero probability the order is bona fide and will drive the market against you.

    Also the statue dictates you have to have the intent to cancel the order before you submit it, which is clearly not a case for market making operations. The main compliance impact I’ve personally seen as a result of Dodd Frank and Rule 575 when it comes to market making is being cognizant of your own impact in thin market, as not to give the impression of skewing the limit book.

    Comment by pxm — November 6, 2015 @ 2:09 am

  7. Hi,

    Excuse me if I have got this wrong but HFT traders are putting in and cancelling orders in very much the same way as happened in this case. So based on the same logic, they are also spoofing the markets. I fully agree with the conclusion that legitimate market making could be caught up in this as traders do very similar things.

    Will there be an appeal?

    Comment by Peter Moles — November 6, 2015 @ 3:06 am

  8. Actually I think Europe gets this right – if you spoof on an exchange they report you to the regulator who chooses between civil penalty (balance of probability and fines only) and criminal proceedings (beyond reasonable doubt and jail time). In the US you can be done for the same thing three times over – exchange then CFTC then DoJ.

    Comment by Flatulentia Buttox — November 6, 2015 @ 4:57 pm

  9. @Peter-That’s exactly what I-and many HFT and algo firms-are concerned about.

    I am sure there will be an appeal. The issue will be the grounds of the appeal. It has to be on a matter of law, not fact. When the charges were filed I conjectured that Coscia would challenge the law as being unconstitutionally vague. A district court bought that argument in a manipulation case in Houston, but the judge in Chicago didn’t: the section of the statute under which Coscia was convicted was different than the one at issue in Houston, and that might have made a difference, or maybe it was a matter of different judges. Perhaps the 7th Circuit will be more sympathetic to an appeal on these grounds.

    The ProfessorComment by The Professor — November 7, 2015 @ 11:34 am

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress