Location: Blogs Bob O'Brien's Sanity Check Blog
Posted by: bobo 10/17/2006 12:27 PM
------------
Sign the Market Reform Petition now. Click petitiononline.com to view it.
--------------
I inevitably get emails from folks who have tried to explain the scope of the fail to deliver problem, and get shut down by their audience. In this latest one, the skeptic, reasonably, said something like, "I don't trust that bunny character - he's anonymous, and could be lying through his teeth. Show me some evidence that this isn't just a tiny issue he's blowing out of proportion, and show me some credible comments from real people that aren't nutcases, and maybe I'll believe we have a crisis."
Now, I don't have a problem with honest skepticism. I think it is healthy. Rather than write private emails in perpetuity, I thought I would present my response here, so that everyone with a doubting friend or relative can read my words, research my sources, and decide for themselves. So here's how I responded to the latest query:
"It's actually all available now in the public forums - it isn't like it was a few years ago, before we had FOIA data and a better understanding of all the pieces. You ask how big is the problem, and how can I believe your numbers, or that this isn't all the rantings of a lunatic fringe. Fair enough. I'll walk through the evidence for you on how big the problem is, and how we know what we know, and can extrapolate with reasonable certainty the rest.
From the DTCC's website:
"@dtcc: Just how big is the fail to delivers, and how much of those fails does the Stock Borrow program address?
Thompson: Currently, fails to deliver are running about 24,000 transactions daily, and that includes both new and aged fails, out of an average of 23 million new transactions processed daily by NSCC, or about one-tenth of one percent. In dollar terms, fails to deliver and receive amount to about $6 billion daily, again including both new fails and aged fails, out of just under $400 billion in trades processed daily by NSCC, or about 1.5% of the dollar volume. The Stock Borrow program is able to resolve about $1.1 billion of the "fails to receive," or about 20% of the total fail obligation."
$6 billion per day of fails to deliver and receive.
But wait. There's more.
CNS netting nets buys and sells against each other. 96% of all trades are netted this way. If a broker has a client that fails to deliver, say, 100,000 shares of NFI, but that broker has 500,000 shares it is holding long on behalf of all its clients, CNS netting "nets" that 100,000 failed trades against the 500K, and presto, no fails are reported - the broker just now shows 400K long. 96% of all trades are handled via netting. The $6 billion per day of fails to deliver, and to receive, are what happens OVER AND ABOVE all the trades that NET!!!
(I haven't been able to find definitive info on how shares held long are differentiated in CNS netting - is there some mechanism to differentiate margin shares, and only allow those on margin to be used to offset the fails? Theoretically that would make a kind of sense, however with netting we aren't talking about borrowing or lending or any of the other 15c3-3 issues - we are just talking pluses and minuses, netted against each other. If a fail (minus) is offset by a plus (share held long) nothing has been lent or borrowed. A simple book-keeping entry has taken place. So until otherwise clarified, I assume most or all shares held long can be used to offset the fails. Oh, and it is unclear who polices compliance in CNS, or whether anyone does.)
Are you starting to get it? But wait. There's still much, much more. Before trades go into CNS netting, most independent brokers go through clearing houses, that "pre-net." That works like this. Today, customer A sells and fails to deliver 1,000 shares. You, customer B, place a buy order for 500 shares. Pre-netting nets your buy against his failed sell, and produces 500 fails. Which then go into CNS netting, and are likely netted against other accounts at the broker, again, out the end, producing no fail.
For a fail to be reported as a fail at the NSCC, the problem has to have gotten so bad that all the shares held long are used up and netted, and no more available to cancel out the fails. Then, and only then, does a fail hit. Now, once that fail enters the system at the NSCC, we know that $1 billion or so are satisfied by the Stock Borrow Program - leaving the remaining $6 billion.
So after pre-netting and CNS netting and the SBP, those fails amount to $6 billion a day. But what does that number mean? Well, first of all, the $6 billion isn't the sale price the fails happened at. It is today's price for the stock that is failed. Thus, if a $20 stock is failed, and then gets crushed down to $1, instead of $20 of fail, now we see $1.
Then how big is the problem, really? $6 billion per day of FTDs and FTRs. Which represents about 4% of all trades that DON'T NET - and after the SBP gets done handling around $1 billion per day.
Of the $400 billion per day that are processed, all but $16 billion net via CNS. The DTCC also says that only about $1 billion of new fails happen every day - over and above what the Stock Borrow Program satisfies, netting conceals, and covering reduces. Let's take them at their word. $1 billion sounds small, per day, with $400 billion being processed, right? Let's ignore that if a bank processed $400 billion per day of cash deposits, and $1 billion per day went missing, that the bank would be shut down, and everyone involved put behind bars. For the sake of this argument, we'll just say that $1 billion per day is not that big a deal.
Not so fast. Netting handles 96%. So that is $1 billion out of $16 billion that netting doesn't take care of. Hmmm. But what about the stock borrow program? It handles another $1 billion post-netting, so isn't the daily number actually more like $2 billion? Or around 12% of the $16 billion? That would make sense, because netting handles the other approximately 96%. So far so good. But let's do the math right. If 12% of what doesn't net out fails, isn't it reasonable to assume that around 12% of those netting also fail?
Now the number gets way bigger. 12% of $400 billion per day is about $48 billion per day failing, with 96% being concealed by netting effects, and $1 billion being satisfied by the SBP. The DTCC chooses to use the $400 billion number, not me. If you don't like that math (the percentage), what's wrong with it? Got some feeling that less than 12% of the netted trades also fail? Fine. Based on what data? Nobody has successfully made that claim. They can't. Which is why they don't.
But it gets even worse. How, you ask?
The number the DTCC throws around to make the problem seem small - $400 billion? That isn't only stock. That is commercial bonds and such too. The actual number for stock is around $150 billion per day - a little more than a third of that number. So, if we apply the DTCC's own math, it gets really ugly. Follow along. The DTCC doesn't say much in terms of this, for good reason.
If only 38% of $400 billion is stock trades, and we know that netting covers 96% of stock trades, then realistically what the DTCC could very probably be saying is that $48 billion per day of the $150 or so billion of stock trading pre-netting is failing, or as much as 33% of all stock trades per day - that's why they use the funny math and blend bonds with stocks when talking about percentages, and then move to transactions rather than shares. They don't use the $150 billion number, they choose to use the $400 billion number, so we don't know which one to use to calculate the final numbers - which is strange in and of itself. Dr. Trimbath speaks to that in the SEC comment letter below. The point is that we don't know whether to use $400 billion to calculate the netting and thus the fails, or $150 billion.
Take your pick - $48 billion, or $18 billion, depending upon which number you favor for the big number. I tend to think that the DTCC is trying to be technically accurate, which is why they use the $400 billion number. But since the DTCC won't answer direct questions about this, including queries to clarify the $6 billion number from a year and a half ago, we have to guess as to their motives and veracity.
Now, some could say I am overstating things using $400 billion rather than $150 billion of equity trades, but here's my point: WE DON'T KNOW WHICH IS RIGHT.
And the reason we don't know is because the DTCC doesn't tell us. Which makes me deeply suspicious and cynical. Just as their switching from dollars to "transactions" mid-stream makes me suspicious. "Only" 24,000 "transactions" fail per day. But how many shares are there per transaction? 100? 1000? It isn't said.
We now know from FOIA data that the total fails on the NASDAQ and NYSE are in the 500 MILLION to 1.2 BILLION shares per day. We can probably take the changes per day, and average that change per day, divide by 24,000, and come up with a number - say, 150 shares per transaction. Suddenly we are talking tens of millions of shares per day.
But that is POST NETTING.
Bluntly, it could be 20-30 times worse due to what netting masks. Likely is. Again, if anyone has a reason it wouldn't be, bring it on. I have yet to hear it.
And here's a cheery thought: None of this includes ex-clearing fails, nor international clearing house netting.
So, we have confirmation from the DTCC of the dollar size of the problem. We have confirmation from FOIA requests of the share size of the problem. We know, from the DTCC, that CNS netting handles 96% of the trades, leaving what we see from the FOIA requests as the 4% netting doesn't handle, and minus the SBP $1 billion. Or maybe we can be really generous, and say that the SBP handles $1 billion of the $6 billion every day - which would then mean "only" $5 billion of fail to deliver and receives per day. No matter how you parse the math, it gets confusing pretty quickly, I believe, by design.
What percentage of netted trades are fails to deliver? We don't know. What percentage of fails pre-net before they ever get to CNS? Again, nobody seems to know. What percentage of all trades are failed, and then pushed off into ex-clearing? Dunno. Nobody does. So what is the actual size of the problem: How many bogus trades are being processed every day, where you pay your money, but nothing is delivered?
The truth is that nobody actually knows.
And that is beyond frightening.
We can parse my math all you want - change the assumptions, move some decimals, whatever. But reality is that so many loopholes and exceptions are in the mix that it is impossible to quantify the figure, beyond saying it is somewhere between big, and catastrophic.
Speaking of obfuscation and twisting of figures and logic, if you really want to understand how conniving the DTCC (which is owned by the brokers and has every reason to distort the numbers to try to make this all seem small) actually is, read Dr. Susanne Trimbath's comment letter to the SEC - she worked at the DTC for years, and is an internationally acknowledged expert on the problem - she was hired to set up the Russian stock market a few years ago. Or read Robert Shapiro's letter to the SEC - he was the Undersecretary of Commerce for Clinton. Or Wayne Kleine, the Securities Regulator for Utah. Or one of the most comprehensive and ugly indictments of the system, from NASAA - the North American Securities Administrators Association.
In closing, it isn't that the information isn't available from reputable sources to understand the problem. Most just don't want to believe the conclusions that are inevitable from studying the data. And an entire industry is devoted to keeping the problem under wraps. If the average Joe got it, there would be rioting on Wall Street. Fortunately for those robbing the country, the average Joe knows what he knows, and isn't about to put any energy into knowing anything more, or anything different.
As to credibility, don't believe the Bunny. These are powerful people and organizations putting their names on these letters. They aren't doing it for fun. Guys like former SEC Chairman Harvey Pitt aren't writing expose commentaries in Forbes as a gag. This is the next big meltdown crisis. Honest. If someone can read my message and those letters and not go, "Holy Sh#t" they either can't read, didn't read them, or are deliberately ignoring the info.
Don't take my word for it. Go educate yourself. Instead of blindly parroting the, "It could never happen in this country" panacea that you are being spoon fed, take some time to run the numbers, read the FOIA data, read the comment letters, and figure this out. Your financial future and the future of the country as a viable entity are at stake, and that isn't hyperbole. Most people don't want to understand anything that challenges their comfortable worldview. Most people are followers, not leaders. Most people don't get it, or are too lazy or stupid to do the work to get it.
So now my question:
Are you most people?
thesanitycheck.com |
