The Continuous Net Settlement System (CNS). A system implemented by Wall Street to speed transaction processing, which nets sales against purchases, and against the accounts of DTCC participants.
What does it have to do with the magnitude of the Fail To Deliver (FTD) problem?
Turns out, a lot.
Because of the netting, many delivery failures never show up as such at the DTCC, and consequently are never reported. The DTCC literally doesn't know they exist. How can that be?
Here's an example created for me by Dr. Susanne Trimbath, PhD, an authority on the clearing and settlement system:
|
Here’s what happens if you settle trade for trade |
|
Sell |
“Deliver” at customer account level (some level below your DTCC account) |
|
100 |
0 |
|
200 |
200 |
|
500 |
500 |
|
-800 |
+700 |
|
i.e., you are short 100 shares at this level, with 200 shares failed to be delivered and received |
|
|
|
|
Now look what happens at the end of the same day after NSCC Account nets to the -800 shares position: |
|
Seller’s DTCC Free Account |
9,000 |
|
NSCC settlement |
-800 |
|
Balance after settlement |
8,200 |
|
|
|
|
There are no failures to deliver at NSCC that day because there were sufficient shares in the sellers free account to cover their bill |
|
Ta-dah! Netting hides the failure to deliver |
In other words, Prime Broker A could have, say, 1 million long shares of NFI in their account, and during the course of the day, several of their big hedge fund customers could fail 400,000 shares, and those would never show up as FTDs because the CNS system would net the fails against their securities in their account, essentially netting the fails against shares in their DTC account.
Here's how it was further clarified for me by Dr. Trimbath:
"For your question, you need to follow through to the DTC account, where shares are taken automatically for CNS settlement by NSCC. Here, we’re talking about a hedge fund that failed to deliver to Prime Broker A, a specific trade that failed delivery. However, because the “free” shares are taken from Prime Broker A’s account there is no reported failure to deliver at NSCC.
If, for example, the failure is in IBM, and Prime Broker A has a ton of shares hanging around the house account, then those get swept up for delivery and there is no failure in CNS. There is a failure in the system somewhere, but the DTCC never sees it. Prime Broker A should be tracking it, we hope, to be sure they get the shares... But there are no SEC rules about Prime Broker A reporting that. (From what I’ve seen in NYSE audits, the Prime Broker A's of the world aren’t keeping very good records on this sort of thing, frankly).
Now what if those are NFI, and Prime Broker A doesn’t have a ton of shares hanging around the house account? Now, there’s nothing to get swept out for settlement and you get a reported FTD from NSCC. This helps explain why smaller companies show up on threshold lists more often, despite the fact that over-voting of shares occurs in HP/Compaq, for example."
OK, so I get it. CNS nets against the NSCC accounts of the participants, and the only FTDs that are reported are trades over and above whatever each participant has on account, after all trades for the day are netted against each other.
So realistically, delivery failures hidden by the CNS system could be much larger than what shows up as FTDs at the DTCC, given that 90+% of all trades are netted. Literally, most of the issued shares of a company on account at the DTCC via brokers could be used up in netting BEFORE the first FTD showed up as we think of them, or as they would show up on a FOIA request.
Now think about that for a second.
We know that EVERY company has an over-voting problem. So by definition, more votes are being cast than shares exist. Margin lending, without distinguishing lent security entitlements, accounts for some of that - the brokers just allow the lent share investor, who has no right to vote as his shares were lent out - to vote anyway.
But how much is due to the slop in the system - the CNS hidden delivery failures? Nobody knows.
So what does that say about market integrity?
Try this one out. In addition to CNS netting hiding the true level of delivery failures, we also have ex-clearing, wherein failures are moved out of the system and treated as a contractual agreement between two brokers - thus out of, or "ex", clearing system. Nobody really knows how large that is, either. And foreign clearing firms also net behind their own curtain, further minimizing the problem's size.
Are you starting to get it now?
I haven't seen a comprehensible discussion of the CNS system's impact on FTDs, so I thought I would try to explain it - so next time the DTCC or the SEC say delivery failures aren't a "significant problem" (assuming you believe anything they say after the Aguirre revelations, and the Global Links and NFI and OSTK and CALM and TASR, etc. etc. revelations) consider the context, and further consider how much is hidden behind the curtain of CNS.
Hope this explains it so the average person can get their head around it. Sort of a CNS for Dummies.
The short way of saying it is that CNS netting could easily be minimizing the apparent problem by 90+%.
How's that for scary?
Here's a diagram of a 100 share failed transaction, where the ledger is adjusted against the total long shares held by the broker, resulting in no fail at the DTCC level, even though the seller never delivered the shares to the broker.
