Matt Taibbi Turns Software Critic

I've been following the throwdown between Matt Taibbi and John Carney of Clusterstock with some bemusement.  For those who are not devotees of the financial blogs, Taibbi recently posted this video, which he claimed showed evidence of a very large naked short:

Here's Taibbi's description of what took place:

Caught On Tape: A Naked Swindle

Continuing with the theme of naked short-selling, I have a video that was given to me last week that will allow people to see how naked short-selling can take place.

This video is only 31 seconds long (scroll on down to view it), and what it shows is a day-trader trying to sell short shares in a major NYSE-traded stock. To disguise the identity of the trader, I've had to edit out the name of the company in question -- I'll call it BANK X for short. What I can say is that the stock in question is one of America's largest financial companies and the recipient of an enormous amount of public bailout money, so the fact that its stock can be manipulated is something that should be a concern to everyone.

In the video, which believe me doesn't look all that sexy, the trader is using an online trading platform. His clearing firm is a company called Penson Financial Services, which, though not particularly well known, has in recent years suddenly become (by volume anyway) one of the biggest such firms in the country.

[Omit lengthy description of/rant about Regulation SHO and naked shorting]

This is why Reg SHO doesn't really work. If there was a hard pre-borrow requirement that actually forced traders to physically locate and borrow shares before they sold them, you wouldn't have this problem. But what they have instead is a system that leaves three days of wiggle room. As it stands, shares must be delivered within three days after the sale, or else the clearing firm must buy shares to close out the failed trade.

But for a day trader, none of this matters. You're buying and selling within the space of one day. Who gives a damn about three days?

These big clearing firms know this. They also know that competitive advantage in getting the business of these intra-day traders depends on providing locates and providing them fast. A clearing firm that worries about the rules and whines about not being able to locate this or that stock is not going to keep the business of a lot of these day traders. So you see a lot of cut corners.

This video is an example of a cut corner. A second-by-second explanation:

If you look at the display in the first seconds, you will see that the customer is trying to sell 100 shares of BANK X short. You can tell it's a short sale by the purple box near the top marked "Short." To the left of that, you'll see that "order quantity" is 100.

:01 Now if you look at the bottom of the display, you can see that the trade has been rejected, because BANK X that day is on the hard-to-borrow list. It reads REJ - HARD TO BORROW: SHRT 100. Without getting too technical, you don't need a formal locate when a stock is on something called the "easy to borrow" list. But BANK X shares that day were not easy to find (without digressing too much into other complicated realms, there was something going on with BANK X that day that was inspiring lots of people to snatch up its shares). So the trade was rejected temporarily, and the trader was then forced to ask for a formal locate of BANK X stock.

:07 A prompt comes onscreen. Through this box, the customer is going to ask Penson to locate shares in BANK X. But how many shares? This is where it gets interesting.

:11 At eleven seconds, you can see the customer start to fill in the box in the middle of the prompt where it reads, "Locate quantity." To disguise the identity of the trader, I've blocked out the first number in the sequence. But you can see that the number is in the tens of billions of shares. Now, the float for BANK X that day was only five and a half billion, meaning there were only five and a half billion BANK X shares in circulation. Without disclosing the actual number, I can tell you that the customer asked for a locate of shares in an amount that was at least five times the number of BANK X shares actually in circulation. Such a locate, in other words, could not possibly be filled.

:17 At seventeen seconds, at the bottom, you see that the firm Penson has now approved the trade and" located" the multibillion amount of shares. The trade goes through.

This doesn't sound all that dramatic and as video sequences go, it sure as hell isn't the Paris Hilton sex tape. But this is an example of how naked short-selling can happen. If you don't need to actually find the stock before you sell it, there's no real brake on speculative naked short-selling. If a clearing firm will give you a locate no matter how big your request is, there is no real barrier out there to stop this kind of activity.

This set off some alarm bells with John Carney of Clusterstock:

There are plenty of things wrong with this video, and with the conclusion Taibbi draws. Clearing firms will not execute orders without regard to the size of the order, there are real brakes on speculative naked short selling, and real barriers exist to stop "this kind of activity."

Too Speedy. The first thing that rings false the speed with which the trade is executed. The trader apparently manages to short billions of shares in mere seconds. Penson may be a popular and efficient clearinghouse but there is no way they are that fast. There's just no way to have instant execution of a trade of this size.

"It takes minutes to a half-hour for a request to come back from Penson," a trader who clears through the firm tells us.

Too Big Of A Trade.  You cannot sell tens of billions shares without someone wanting and able to buy those billions of shares. This trade involves placing a sell order for more than the total volume of all US equity markets combined for any single trading day.

Too Much Leverage.  More importantly, almost no trader using Penson as his clearing house would have the buying power to put in an order this large. This should sale would require billions of dollars of buying power. The buying power of any trader is a multiple of the money deposited with Penson. That is, it is determined by the maximum margin account available to the trader.

For ordinary, retail traders--the kind of people likely to be called "day trader"--the maximum intra-day leverage is 4 to 1. This is set by FINRA regulation. In order to short tens of billions of shares on any stock worth more than $1, the trader would have to have billions in his brokerage account.

Small hedge funds often use a more sophisticated kind of margin account that allows for more leverage. It's calculated by the clearing firm exclusively for each of their clients who get it based on their style of trading and track record. A risk averse day-trading hedge fund could get leverage as high as 7 to 1 intraday. These typically require a minimum of at least $1 million in the account with the clearinghouse.

Any trader with the billions of dollars necessary to get execution on the trade Taibbi describes would be unlikely to be clearing through Penson. He would be trading through Goldman Sachs or JP Morgan.

It's too risky. There is a good reason for Penson and other clearing houses to limit leverage in this way, even beyond the regulations.  They face counter-party risk when allowing traders to short stocks. If the stock increases in value and the trader cannot pay for the stocks he is short, the clearinghouse will have an obligation to make good with the Depository Trust Company.

This trade would be a disaster. If tens of billions of shares of a stock were sold like this, it would make headlines and single handedly wipe out the stock. Likewise, when the trader bought up the shares to close the trade and deliver the shares to the buyers, he'd face a self-imposed massive short squeeze. Buying shares pushes the price up. He'd get crushed.

Taibbi is reading too much into the video. Although Taibbi concludes his description of the video by writing "The trade goes through," this isn't what happens.  This screen shot is nothing more than the purported "ok" from some trading system that an attempt is made to locate shares.  The quantity is requested and there is no indication that the entire amount was approved.

Many trading systems will accept any amount for the request but only return the approval for the quantity that the firm was actually able to locate. There is no evidence that this trading screen is somehow linked in real time to Penson. Had this order actually been attempted to be entered into the Penson system, multiple risk checks and parameters would rejected the order.

 In short, there's just no way the trade represented in this video is real. It certainly doesn't reflect the reality of short selling.

A later post pointed out that at 0:27 a pop-up notifies the user that his trade failed because the position is too large.  Taibbi retorted that he hadn't been woried about the trade, just the huge locate:

The idiots at Clusterstock, just one week removed from making the uniquely asinine (even for them) claim that there is no difference between short-selling and naked short-selling, have struck again, proving once again that it is always best to actually put down your paper bag full of airplane glue fumes before you make blog posts.

Business Insider writer John Carney here seems to have read my recent post on Penson and taken from that that I was reporting that someone had executed a short sale of tens of billions of shares in a company whose float was only five and a half billion. This would, indeed, be ridiculous. Except that is not at all what I reported.

What I published was a tape of a trader asking for a locate of tens of billions of shares. It is the size of the locate, and the speed with which it is approved  that is the issue, not the trade. The actual trade, if Carney had bothered to read the text, was only for 100 shares.

Carney then goes on to claim in another post that the "system" worked because a second trade was rejected at :27 on the tape. But this is a second trade for a larger amount of shares that was rejected not because of the locate but because the trader in question had insufficient funds in his account to make the short sale. It has nothing to do with the locate and is completely irrelevant to the story.

So now I'm confused:

1.  I've worked with a fair amount of market data/trading software, because I used to do IT for banks.  Now, it's been a decade, and maybe everything's changed--but there aren't that many exciting ways to lay out trading software.  As far as I can tell, whatever this mysterious second trade is, it's in the same stock.  At any rate, I don't know how he can be so confident as to what it refers to, since the video has pretty clearly been edited.  The execution is too quick.

2.  The locate at the bottom at 0:22 is for an odd number of shares.  The alleged locate order was for a round number.  If the software is so crappy that it will accept any locate order, how come it's got a locate for a different number from what was entered?

3.  The headline of the piece is "Caught on Tape:  A Naked Swindle", which would seem to imply that someone is being, I don't know, swindled.  The video is titled "Penson Approves Billion Share Naked Short", not "Penson Sure Has Some Bad GUI Designers".  Why would Taibbi ramp up this level of outrage over . . . a software glitch?  If the trades don't go through, who cares whether the software erroneously tells customers it can find them five times the float?  I mean, I'm sure the customers do.  But that's not a regulatory issue, it's a problem for the customer service reps.

4.  Alternatively, we can say that Taibbi has discovered evidence of hundreds of shares being shorted without proper locates. But a hundred shares is just not a problematic short.  Are even "hard to borrow" stocks so hard to borrow that no one can locate 100 shares in the next three days?  More importantly, unless this guy was trading some OTC penny stock, which Taibbi says he wasn't, 100 shares couldn't possibly move the price by as much as a penny.  Taibbi has called naked shorting "counterfeiting shares", but if one guy goes into his basement and counterfeits 100 perfect dollars, this is not a problem for . . . well, anyone, really. 

Counterfeiting is a problem either because the counterfeits are imperfect, and someone therefore ends up "losing" the money when the fakery is discovered, or because they increase the money supply enough to decrease everyone's purchasing power.  Similarly, a 100 share naked short is a problem for you and your counterparty--if it's fraudulent rather than accidental (which does happen), perhaps one that should be prosecuted. 

But they're irrelevant to the rise and fall of companies, which is what Taibbi's article is allegedly about. Shorts are systemically important only if you can execute naked trades large enough to move the stock price.  Otherwise it's not a swindle--it's just a very stupid idea, because you'll get in trouble if you FTD.    And for a variety of reasons, many outlined by John Carney, it's a lot harder to execute large trades of any description (but particularly short trades) than very small ones. 

5.  Who films themselves doing something illegal?  Sex acts don't count.  No one slips into their silk pajamas, pours themselves a neat Glenlivet, and fires up the laptop to relive every scintillating moment of the time they got some trading software to claim it could locate a bunch of shares that it really couldn't locate at all.  At any rate, if they do, I sure hope they write in to Dan Savage.

This is all sort of academic, because Penson has now written a letter to the SEC pointing out that it is not their trading software, among other problems.  This is the sort of thing that is easy to check, and, whatever you think about the SEC, the kind of thing they do check.  Either Penson is risking gigantic regulatory hassle by lying to the SEC, or Taibbi got punk'd.  But why would anyone bother?

There's now a growing body of spinoff posts attempting to discover what the video does show.  The consensus is that the stock is Citigroup--the price is certainly right, and it matches Taibbi's description.  Kid Dynamite thinks it shows the audit trail--the software asking for a broker against which it can check the locate.  He says this raises possible red flags about compliance--but I'm not sure how.  We don't know that compliance failed, only that the software may not have done it automatically--or it may, because as I said, we know the video has been edited, because pretty much everyone agrees the approvals are happening too fast. 

But failing to automate compliance is not necessarily a bad thing.  I've worked with teams who were automating trading compliance, and what you learn very quickly is that machines are really stupid about compliance.  Humans are less consistent, but also less easy to game.