Streetwise Professor

October 17, 2018

The Harm of a Spoof: $60 Million? More Like $10 Thousand

Filed under: Commodities,Derivatives,Economics,Exchanges,Regulation — cpirrong @ 4:08 pm

My eyes popped out when I read this statement regarding the DOJ’s recent criminal indictment (which resulted in some guilty pleas) for spoofing in the S&P 500 futures market:

Market participants that traded futures contracts in these three markets while the spoof orders distorted market prices incurred market losses of over $60 million.

$60 million in market losses–big number! For spoofing! How did they come up with that?

The answer is embarrassing, and actually rather disgusting.

The DOJ simply calculated the notional value of the contracts that were traded pursuant to the alleged spoofing scheme.  They took the S&P 500 futures price (e.g., 1804.50), multiplied that by the dollar value of a price point ($50), and multiplied that by the “approximate number of fraudulent orders placed” (e.g., 400).

So the defendants traded futures contracts with a notional value of approximately $60+ million.  For the DOJ to say that anyone “incurred market losses of over $60 million” based on this calculation is complete and utter bollocks.  Indeed, if someone touted that their trading system earned market profits of $60 million based on such a calculation in order to get business from the gullible, I daresay the DOJ and SEC would prosecute them for fraud.

This exaggeration is of a piece with the Sarao indictment, which claimed that his spoofing caused the Flash Crash.

And of course the financial press credulously regurgitated the number the DOJ put out.

I know why DOJ does this–it makes the crime look big and important, and likely matters in sentencing.  But quite frankly, it is a lie to claim that this number accurately represents in any way, shape, or form the economic harm caused by spoofing.

This gets to the entire issue of who is damaged by spoofing, and how.  Does spoofing induce someone to cross the spread and incur the bid/ask, who would otherwise not have entered an aggressive order?  Does it cause someone to cancel a limit order, and therefore lose the opportunity to trade against an aggressive order and thereby earn the spread (the realized spread, not the quoted spread, in order to account for losses to better-informed traders)?

Those are realistic theories of harm, and they imply that the economic harm per contract is on the order of a tick in a liquid market like the ES.  That is, per contract executed as a result of the spoof, the damage is .25 (the tick size) times $50 (the value of an S&P point).  That is, a whopping $12.50.  So, pace the DOJ, the ~800 “fraudulent orders placed caused economic harm of about 10,000 bucks, not 60 mil.  Maybe $20,000, under the theory that in a particular spoof, someone lost from crossing the spread, and someone else lost out on the opportunity to earn the spread.  (Though interestingly, from a social perspective, that is a transfer not a true loss.)

But $10,000 or $20,000 looks rather pathetic, compared to say $60 million, doesn’t it?  What’s three orders of magnitude between friends, eh?

Yes, maybe the DOJ just included a few episodes in the indictment, because that is sufficient for a criminal prosecution and conviction.  But even a lot more of such episodes does not add up to a lot of money.

This is precisely why I find the expenditure of substantial resources to prosecute spoofing to be so dubious.  There is other financial market wrongdoing that is far more harmful, which often escapes prosecution.  Furthermore, efficient punishment should be sized to the harm.  People pay huge fines, and go to jail–for years–for spoofing.  That punishment is hugely disproportionate to the loss, under the theory of harm that I advance here.  So spoofing is over-deterred.

Perhaps there are other theories of harm that justify the severe punishments for spoofing.  If so, I’d like to hear them–I haven’t yet.

These spoofing prosecutions appear to be a case of the drunk looking for his wallet (or a scalp) under the lamppost, because the light is better there.  In the electronic trading era, spoofing is possible–and relatively cheap to detect ex post.  So just trawl through the trading data for evidence of spoofing, and voila!–a criminal prosecution is likely to appear.  A lot easier than prosecuting market power manipulations that can cause nine and ten figure market losses.  (For an example of the DOJ’s haplessness in a prosecution of that kind of case, see US v. Radley.)

Spoofing is the kind of activity that is well within the competence of exchanges to detect and punish using their ordinary disciplinary procedures.  There’s no need to make a federal case out of it–literally.

The time should fit the crime.  The Department of Justice wildly exaggerates the crime of spoofing in order to rationalize the time.  This is inefficient, and well, just plain unjust.

Print Friendly, PDF & Email


  1. On another regulatory note, thoughts on the SEC’s decision in favor of SIFMA over NYSE and NASDAQ? Seems like broker-dealer crony capitalism under the auspices of the Exchange Act.

    Comment by Abe Froman — October 17, 2018 @ 4:18 pm

  2. @Abe–I’ve only read the news accounts. It is basically a battle over rents, and yes, the broker-dealers appear to have the upper hand, at least insofar as the SEC is involved. No doubt the exchanges will challenge the SEC in court. We’ll see how it goes.

    The economic issues here are interesting and potentially complex–as is often the case in defining property rights over information (which is basically the issue here). Broker-dealers were perfectly happy to have the exchanges own the property right to data, when they owned the exchanges. Now that they don’t, they are upset, and basically want the right for themselves (or at least, want to attenuate the exchanges’ rights, including most notably the right to choose the price at which they sell it).

    Property rights over information (and intellectual property generally) are complicated because the marginal cost of reproduction is effectively zero. But there has to be an incentive to create it. The efficient allocation of property rights depends on the cost of creation, which are actually even more complicated in the exchange price data situation than in the normal situation (e.g., the creation of a work of art, or a book, or computer software). The price data is a byproduct of transactions. Multiple parties are involved in creating a transaction. The ultimate transactors, and various agents–which include the broker-dealers, and the exchanges. Exchanges incur costs, which need to be covered. Those can be covered with per transaction fees, or other revenue streams–including listing fees (not that big a deal anymore) or data fees. The efficient fee structure is not immediately obvious.

    I’ll have to look at the SEC decision to see if any of these complexities are addressed. I kind of doubt it.

    Good question–perhaps grist for a future post!

    Comment by cpirrong — October 17, 2018 @ 5:09 pm

  3. The Exchange Act would seem to put in place a lack of incentives for SEC-regulated exchanges to innovate. You might create a really unique and useful data tool, but are obligated to provide “fair and non-discriminatory” access to any such innovation, with the definition of those requirements unclearly defined. The SEC’s opinion doesn’t appear to address the lack of incentives, the complexity you spoke of, or clarify exactly what it is looking for from the exchanges.

    Comment by Abe Froman — October 17, 2018 @ 7:24 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress