Merrill Lynch Is Fined for Doing Nothing
Among connoisseurs of finance, there are those who admire Nathan Rothschild's trading on Waterloo, or George Soros's bet against the pound, or the great Piggly Wiggly corner of '23, but I grew up as a derivatives structurer, and the trades that I admire most are the ones where nothing happens. If you can build a big complicated derivative that shifts the risks and payoff profiles among the parties to the trade, that's fine, that's the job, good work. But if you can build a big complicated derivative that doesn't shift the risks and payoff profiles among the parties -- that leaves everything exactly where it was, but makes the parties' lives ineffably better -- well, now you are getting somewhere.
Often where you are getting is court, though. Today Merrill Lynch settled a case with the Securities and Exchange Commission, agreeing "to pay $415 million and admit wrongdoing to settle charges that it misused customer cash to generate profits for the firm and failed to safeguard customer securities from the claims of its creditors." There is some boring stuff here -- Merrill Lynch put some customer securities where they weren't supposed to be, and wrote some bad nondisclosure policies -- but the fun part is that Merrill wrote some magical derivatives to reduce the amount of cash that it had to have on hand to meet customer claims. And that's all those derivatives did. There was a complicated dance that allowed the magic to happen, but at the end of the dance, everything was back where it started. Except that Merrill had somehow, through the magic of the dance, lightened its obligation to keep cash lying around for customers.
Here's the deal. There is an SEC rule, Rule 15c3-3, that requires broker-dealers to keep customer cash separate from their own cash, so that if a dealer goes bankrupt, the customers don't lose any money. But "customer cash" is a term of art; the actual requirement is that the dealer needs to keep cash in a separate reserve account that is at least equal to the "net cash owed to customers" calculated under a formula set by the SEC. The formula is basically, add up all the cash you owe to customers, subtract all the cash that customers owe to you, and put the net amount in a box where you can't touch it.
If you are a bank or broker-dealer, you love touching money. Touching money is your business, and money that you can't touch is a personal affront. (By which I mean: Keeping cash in a reserve account is a lot less profitable than using it in your business.) You really want to crack open that box and touch the money. And if you can find ways to make customers owe you more money, that reduces, dollar for dollar, the amount of cash you need to keep in the box.
The easy way to do that is to lend customers money, but this has two problems:
- They may not pay you back. (This is the less important problem.)
- If you lend them the money, then they have the money, and you don't, which means it's of no more use to you than it was in the box. (This is the more important problem.)
Merrill Lynch solved the problems. It would give a customer a loan. (So now the customer owed Merrill money.) The customer would use the loan to buy stock from Merrill. (So now Merrill had the money.) Then it would immediately agree to sell the stock back to Merrill. (So now Merrill had the stock.)
But the customer's sale of the stock back to Merrill would be delayed. The customer agreed to sell the stock, but would just hang on to it for a while. Merrill would have all the economic rights to the stock, and the stock would physically sit at Merrill (in the customer's account), but technically the customer would still own it. And so Merrill wouldn't pay for the stock that it was theoretically buying back from the customer. And so Merrill got to keep the stock and the money, and it got to conclude, for SEC cash reserve purposes, that the customer owed it money.
It's a neat trade! It is sort of too obvious, when I put it like that, that nothing is going on. So the dance involved in a "leveraged conversion" -- which is what these trades were called -- is just a bit more complicated than that. We talked about it last year, when the SEC was investigating these trades; it involves dressing up the customer's delayed sale of the stock back to Merrill not as a "forward sale" but as the purchase of a put option and the sale of a call option with the same exercise price. Those combined transactions are economically equivalent to a forward sale of the stock, but by introducing words like "put option" and "call option," you make it a little less obvious that nothing is going on and that Merrill has reduced its cash reserve requirement without actually lending money to a customer.
Anyway this isn't allowed. I guess. The SEC says it's not allowed, anyway:
Merrill Lynch engaged in complex options trades that lacked economic substance and artificially reduced the required deposit of customer cash in the reserve account. The maneuver freed up billions of dollars per week from 2009 to 2012 that Merrill Lynch used to finance its own trading activities. Had Merrill Lynch failed in the midst of these trades, the firm’s customers would have been exposed to a massive shortfall in the reserve account.
You can see where they're coming from. Merrill was going along, running its business, and that business required it to put (say) $12.6 billion into a reserve account where it couldn't be touched. And then Merrill did a little dance, and when the dance was finished, its business was exactly the same as it was before. Except somehow it only had to put $7.6 billion in the reserve account. That does seem wrong.
But what is most pleasing about this case may be the way that Merrill frog-boiled its regulators into allowing these trades. Merrill Lynch's Structured Equity Financing and Trading desk dreamed up this idea, ran it by compliance, and then, along with William Tirrell, Merrill's head of regulatory reporting, sat down with the Financial Industry Regulatory Authority and the SEC's Division of Trading and Markets to get their blessing. And got it:
Despite being an impetus for the meeting, ML did not inform FINRA and T&M, either at the August 2009 meeting or subsequently, that the primary purpose of the Leveraged Conversion Trades was to finance firm inventory. Tirrell, using the revised handout shown above, presented the Trades to regulators. In addition to explaining the bullets on the slide concerning market exposure and the use of only actively traded large cap stocks, Tirrell told regulators that ML’s customers took on real risk, which in Tirrell’s view made the Trades like any other conversion trade. Based on the revised handout and the discussions at the meeting concerning the presence of real risk, the fact that these were standard conversion trades – which were to be customer and not firm driven – and the perceived economic substance of all of the trade components, the regulators did not object to the version of the Leveraged Conversion Trades that was presented.
The handout explained the trade ... differently from the way I did. It didn't start from the goal of doing nothing. It started at the other end. What if a customer shows up wanting to buy stock? What if he also wants to buy listed options? What if we gave him a margin loan? These are all real transactions, all customer-driven, all subject to various (small) risks. Why shouldn't Merrill be able to combine those real transactions however it wanted?
Once the trades were approved, though, they quickly left behind messy reality for a world of pure abstraction. ("The Trades Morphed," is the SEC's lovely section heading.) Instead of being customer-driven trades, they were initiated by the SEFT desk to finance Merrill's inventory. Instead of using listed options, Merrill started using over-the-counter options, to avoid the risk of exposing these trades to the market. (Merrill ran this change by FINRA; FINRA had some questions, and Merrill ignored them. ) In fact, the desk even created the customers:
In mid-2010, SEFT Trader solicited three former options traders who, at the time, were not ML customers, to participate in the Trades. SEFT Trader advised them to create LLCs to execute Leveraged Conversion Trades, from which they would earn a fixed rate of return of up to 15% on the total amount they deposited to capitalize the LLC. Specifically, SEFT Trader advised each of them to open portfolio margin accounts at MLPF&S in the name of the newly-created LLC and to deposit at least $5 million in cash or securities into the account. The LLCs were offered significant leverage; for every dollar in cash or securities an LLC deposited into a portfolio margin account, ML allowed the LLC to purchase on margin up to $100 of securities.
Some quick math on that. The customer puts in $5 million, say, and is guaranteed a 15 percent return ($750,000). The customer then gets up to $495 million of "margin loans" that flow directly back to Merrill. Merrill can reduce its cash reserve requirement by $495 million, in exchange for paying the customer a $750,000 fee, or about 0.15 percent. If Merrill can earn 0.15 percent more on that $495 million by using it than it can by keeping it in a box, then that's a good trade for Merrill. It's just free money for the customer. (It's not great for other customers, if Merrill fails.)
In its final form, the trade is an obvious, embarrassing nothing: Merrill paid some favored clients a smallish fee to do transactions with no economic substance that allowed Merrill to reduce its customer cash reserves for no particular reason. Of course the SEC hates it. But in its initial form, the SEC explicitly approved the trade. And those two forms are almost identical! They are both "leveraged conversions," both non-economic round-trip trades that don't really change Merrill's risks but do change its accounting. The differences are those of nuance, intention, crassness, the volume of Merrill's gleeful cackling. But once the SEC approved a trade that did almost nothing, it can't be too surprised that Merrill kept refining it, polishing it down to its essential form of doing nothing at all.
No I mean really my ranking is:
- The Piggly Wiggly corner.
- Trades where nothing happens.
Do you know the Piggly Wiggly corner? The canonical version is a chapter ("The Last Great Corner") in John Brooks's "Business Adventures." The guy cornered a stock, then sold it forward (to the public!) to de-risk his position while keeping control of the borrow. An absolutely jaw-dropping move. He loses some points for going bankrupt, but even after that deduction it's amazing.
The fun started in August 2008, before Merrill was bought by Bank of America, though it only really got going in 2009, after the acquisition, and continued until April 2012.
By the way, we talked earlier about another Merrill Lynch trade -- the dividend arbitrage -- that also didn't do much of substance, but also landed Merrill in court. It's a theme!
There is an analogy to another of my favorite trades, a Credit Suisse regulatory capital trade of surpassing loveliness that was eventually killed by regulators. In each case, some essentially non-economic round-tripping allowed a bank to reduce regulatory reserve requirements, though in Credit Suisse's case it was Basel capital, while here it is SEC cash reserves.
Until the trade ended, at which point Merrill would pay the customer, the customer would give Merrill the stock, and the customer would pay back the margin loan. So Merrill would still have the money, and the stock. It's just that now it wouldn't have a margin loan that it could use to reduce its customer cash reserve.
Because of the forward-sale element, this works best with stocks that Merrill was planning to hang on to for a while. The SEC's example is a block of 310,000 Google shares that Merrill had in inventory in February 2011; it did a two-month leveraged conversion that it ended up extending out all the way until November 2011.
- If I buy stock from you for $100 per share, and I simultaneously agree to sell it back to you for $101 per share in three months, that is a "forward sale," and you have paid me, effectively, $1 for the use of my cash.
- If I buy stock from you for $100 per share, and I simultaneously pay you $10 for a put option that expires in three months and has a strike price of $101, and you simultaneously pay me $10 for a call option that expires in three months and has a strike price of $101, then we have a "conversion." Today, I paid you $100 for the stock, and $10 for the put, and you paid me $10 for the call, so net I have paid you $100. In three months, if the stock is above $101, you will exercise the call option and pay me $101 for the stock; if it's below $101, I will exercise the put option and make you pay me $101 for the stock. Either way, you will pay me $101 for the stock. (That is, you'll pay me back $100, plus $1 for the use of my cash.) The economics are identical to the forward sale, but the names and documentation are different.
You can monkey with the prices to get everyone paid the right amount.
From the text of the SEC order:
On February 7, 2011, the SEFT desk executed a Leverage Conversion Trade in which (i) the LLC purchased 310,000 shares of Google, (ii) simultaneously sold them back at $612 with a delayed settlement date of April 6, 2011, and (iii) and purchased from ML an OTC put on the Google shares and sold to ML an OTC call for these shares that both had the same strike price, $612, and expired that day.
Weird, right? Merrill sold stock to a customer (a limited liability company, "LLC") subject to a margin loan, and then bought it back through a forward (the repurchase "with a delayed settlement date" two months later). Normally a conversion would involve the client buying a put and selling a call at the same strike, expiring when you want the trade to expire (e.g. April 6). The put/call combination is equivalent to a forward sale, but you're really supposed to document it as a put/call, not a forward sale. But Merrill didn't do a conversion; it just did a forward sale.
But Merrill also did a conversion, just for fun. Sort of. It bought a put and sold a call, but they expired on the day of the trade, and apparently weren't exercised. They were purely for show. The SEC says that "in fact, there is no evidence that ML or the LLC accomplished the instantaneous sale and repurchase through the exercise of the in-the-money put or call."
So why do them? Here is the SEC's footnote 5:
Due to systems limitations, ML was unable to execute a Leveraged Conversion Trade using OTC options that expired past the trade date. To account for this limitation, SEFT Trader considered a trade structure that simply involved a sale and simultaneous repurchase of stock with a delayed settlement of the repurchase, but he recognized that an instantaneous roundtrip would be inappropriate from a regulatory perspective. The SEFT desk, with the involvement and approval of Tirrell and others at ML, settled on a trade structure that involved these same components but also included a put and a call that expired on the trade date.
That is: Merrill knew that the combination of margin loan + sale + forward repurchase looked like a nothing transaction that couldn't support its regulatory goals. It knew that the right way to make the trade look good was with a margin loan + sale + put/call conversion. But its computers wouldn't let it do that. So it just did the margin loan + sale + forward repurchase, and threw in a phantom one-day put and call so that, if no one looked too closely, it might be mistaken for a conversion.
That is so dumb! Come on.
Those stylized numbers come from the SEC order:
In the OTC iteration of the Trades, which lasted from approximately September 2010 to April 2012, ML reduced the minimum amount required in its Reserve Account by up to $5 billion per week through Leveraged Conversion Trades. During this period, ML’s Reserve Account balance ranged from approximately $7.6 to $12.8 billion; therefore, although no customers were harmed, ML put them at risk by reducing the customer money it was required to deposit into its Reserve Account by approximately 28% to 40%
It had one of those embarrassing boxes-and-arrows derivatives diagrams that don't actually explain anything. Feh:
In particular, the original version of the trade involved listed options that trade on an exchange floor. "ML could take the other side of the Trades, but floor traders could also step in and take all or part of the Trades," in which case Merrill Lynch would no longer be the counterparty to the options and the trade wouldn't quite vanish into itself in the aesthetically pleasing way I set out above.
From the SEC order:
In December 2009, Tirrell emailed FINRA staff to advise them that the SEFT desk “would like to  use unlisted options as it provides greater flexibility.” “I don’t see this as a material change to the current arraingement [sic],” Tirrell stated in the email, “but wanted to ensure you are in agreement.” In January 2010, FINRA staff communicated to Tirrell that the regulators’ non-objection did not extend to unlisted options and requested that ML describe, in writing, the difference in transaction structure, the regulatory impacts, and the rationale for why this version of the Leveraged Conversion Trades should be allowed to have a similar impact on the Reserve Formula.
Tirrell did not provide the requested information. The SEFT desk obtained internal approval from Tirrell and others for Leveraged Conversion Trades using OTC options. However, it is not apparent that, at the time, anyone within ML other than Tirrell knew about FINRA staff’s questions posed in the January 2010 email.
Also don't forget the dumb thing in footnote 6, where Merrill essentially got rid of the options altogether.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Matt Levine at firstname.lastname@example.org
To contact the editor responsible for this story:
James Greiff at email@example.com