One of the key issues with failure to deliver is that market-makers get an exemption under Reg SHO whereby they don't have to locate borrowable shares before selling short for "bona fide" MM activity like delta hedging or offering stock. If an investor were to own every share of a stock, and were to refuse to lend shares to other market participants, a market-maker would still be allowed to sell that stock short. If he were to cover it the same day, it wouldn't even trigger a failure to deliver. And as long as he keeps buying and selling shares on a brief time horizon (This would be a "reset transaction" if anyone else were to do it, but for a market-maker it's arguably part of a day's business), he will never end up at the front of the queue of short-sellers who need to buy in shares at open market to make good on delivery three days after a failed settlement.
One of the vulnerabilities of the way US markets implement short-selling is that the ability to sell shares before borrowing them can temporarily create phantom shares, and the buyer has no idea. Plus, borrow/rehypothecation is recursive. This means that you can lend a share to the same person who sold it to you so that he can sell it to you again. The risk of this in a FTD situation, in my mind, could be reminiscent of a stack overflow...think of the float as memory. Fortunately the CNS system should be able to catch these alterations in the perceived number of tradable shares, but what happens when the FTD rate becomes substantial and persistent, as with $GME? Do those fake shares simply create artificial supply until the short-sellers eventually get forced to buy in, five days later?
Another weakness is that it gives market-makers the power to put downward pressure on prices during upside breakouts. Then, once they generate calm, the liquidity comes back and they can cover the artificial supply. Surely some would view this as a force for good, a control on market exuberance. However, imagine what this would mean in a commodity market -- buyers are unaware that the thing they bought doesn't exist, and the clearinghouse is responsible for ensuring that they get it. Also, it allows market-makers to create situations where they get paid free money, because all speculators who were betting on a surge by buying shares must then turn around and cross the bid/offer spread the other way to close out.
In a market with a total capitalization in the tens of trillions, it doesn't make much sense to deposit a thousandth of that much money into a collateral fund for a clearinghouse and say "you take care of it." And make no mistake, they are the ones who do the rejiggering, because FTD means that the clearinghouse is short to the buyer until the seller acquires and delivers the security.
Here’s my take on what happened to Robinhood this week.
The FTD rate on GME has been extreme even leading up to the massive price spike. When the price rose to $300+ these FTDs could have only increased, and started carrying some very heavy liabilities, into the billions of dollars.
If a short seller was going to get called and failed to cover, anyone who held those shares long were at risk. But actually the clearinghouse is the one who has to make the buyers whole. First, any collateral held in the clearinghouse from the short selllers would be used to cover, but it wouldn’t be nearly enough. Then funds are drawn from a collective pool, but not exactly evenly from all participants. As I understand it, the brokerages that bought the shares that FTDd would have to pay a somewhat higher proportion of the shortfall.
I think this is essentially why Robinhood got the 3am call demanding $3 billion. The risk of a hedge fund failing was a multi-billion dollar contagion about to the hit the clearinghouse. They needed to cover that risk and they hit up the brokerage who was buying the majority of the shares.
I don’t see how Robinhood itself was at any risk of default. I think they got caught up in it because their counter-party was insolvent and the clearinghouse knew it.
It wasn’t a risk of Robinhood not being able to pay for its Buy orders. If that was the case they could just shut of margin trading on GME and be done (no buying GME on margin, or using GME as collateral for other margin buys). Basically buying the mandatory “FTD insurance” on GME became prohibitively expense.
I’d love to see a public accounting of the reserve fund algorithms, and who else they asked to pony up billions of dollars that morning.
I think ultimately when the clearinghouse itself became at risk of a multi-billion dollar contagion they pumped the brakes on the stock in order to keep the hedge funds alive. The massive FTD rate is a solid indicator of clearinghouse risk. And it wasn’t Robinhood’s FTD, but they were on the other end of the trade.
> The massive FTD rate is a solid indicator of clearinghouse risk. And it wasn’t Robinhood’s FTD, but they were on the other end of the trade.
Fair point. But ultimately the solution would be to impose a locate requirement on market-makers, then buy all eligible FTD positions at market to cover them and send the bill to the respective short-selling clearing members.
In general, it doesn't make sense for market-makers to sell a threshold security short into demand, even if there's borrow available, because that borrow will be needed to chip away at the FTD queue.
Also, what rate do firms that fail to deliver pay to be short to the clearinghouse? It ought to be in line with the borrow rate in the securities lending market.
> The risk of a hedge fund failing was a multi-billion dollar contagion about to the hit the clearinghouse.
Melvin doesn't directly face the clearinghouse. I'm not sure where Melvin does its PB but I doubt JPM, MS, or GS would go under from failing to quickly cover a single short position. (But it would be very interesting to hear the earnings Q&A after a quarterly loss due to a reddit squeeze.)
I have a few questions on how locates work and shorting:
-I hear that the buy side just sprays locate requests to brokers every day without necessarily having an intention to short a particular stock with that broker on that day. So brokers naturally accept the locate requests
-Fail to deliver from brokers is very common for shorted stocks from what I hear. With associated fines.
Customers have a reasonable basis to ask for borrow pricing daily because rates and availability change frequently, and the cost of the short is one of the factors influencing the decision whether to trade it or not.
A locate request is basically "hey do you have shares to lend out, or can you find some?"
It makes total sense that FTD would affect shorted stocks, because it generally arises from a short sale and the heavily shorted names are the ones where borrow is tight.
Let's say a trader on the cash trading desk buys shares from a short-seller who fails to deliver. Does sec lending know that those shares have not yet settled?
If it is incorrectly assumed that those shares settled properly on T+2, would the bank potentially loan those shares out prior to delivery?
To me, a complete lay person, this is incomprehensible. Not only you can sell something you don't own, but you can sell something that doesn't exist... How did it come to this, why? This feels stupid. What's the point of allowing this, what problem is it solving?
"The lawsuit claims that the creator, Doug Monahan, didn’t use his raised Indiegogo or Kickstarter funds to develop or deliver the backpack, but instead, he put the cash toward “various personal expenses,” including the purchase of bitcoin, ATM withdrawals, and to pay off credit cards. The agency also claims that backers reached out to Monahan, but he ignored their complaints, shut down the company, and ceased communications."
"what Ferguson called “the first enforcement action in the nation against a crowdfunded project that didn’t follow through on its promise to backers” succeeded this July, netting restitution for the Washington-based donors and slapping Asylum with $31,000 in civil penalties."
One of the vulnerabilities of the way US markets implement short-selling is that the ability to sell shares before borrowing them can temporarily create phantom shares, and the buyer has no idea. Plus, borrow/rehypothecation is recursive. This means that you can lend a share to the same person who sold it to you so that he can sell it to you again. The risk of this in a FTD situation, in my mind, could be reminiscent of a stack overflow...think of the float as memory. Fortunately the CNS system should be able to catch these alterations in the perceived number of tradable shares, but what happens when the FTD rate becomes substantial and persistent, as with $GME? Do those fake shares simply create artificial supply until the short-sellers eventually get forced to buy in, five days later?
Another weakness is that it gives market-makers the power to put downward pressure on prices during upside breakouts. Then, once they generate calm, the liquidity comes back and they can cover the artificial supply. Surely some would view this as a force for good, a control on market exuberance. However, imagine what this would mean in a commodity market -- buyers are unaware that the thing they bought doesn't exist, and the clearinghouse is responsible for ensuring that they get it. Also, it allows market-makers to create situations where they get paid free money, because all speculators who were betting on a surge by buying shares must then turn around and cross the bid/offer spread the other way to close out.
In a market with a total capitalization in the tens of trillions, it doesn't make much sense to deposit a thousandth of that much money into a collateral fund for a clearinghouse and say "you take care of it." And make no mistake, they are the ones who do the rejiggering, because FTD means that the clearinghouse is short to the buyer until the seller acquires and delivers the security.