Flash Crashes and Insider Trading

Matt Levine is a Bloomberg View columnist. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz and a clerk for the U.S. Court of Appeals for the Third Circuit.
Read More.
a | A

Flash crashes.

In overnight trading, the pound was trading against the dollar at a price a bit north of $1.26. Then it dropped sharply, and a few hours later it "was 1.4 percent lower at $1.2437 as of 9:30 a.m. in London." I don't know why this happened -- something something something Brexit -- but there has definitely been a big fundamental move downward in the pound this week. Traditionally the way a drop like that would happen is that there would be a lot of demand to sell pounds, and market makers who were previously buying pounds for $1.262 would lower their bids to $1.260, and there'd be still more selling, and the market makers would start buying at $1.255, and then $1.250, and then $1.245, and so on down, until the pound eventually settled at its new price. It was an orderly and convenient system, but it was kind of dumb for the market makers. Everyone wanted to sell -- the fundamental demand for the pound was in the $1.24 area, not the $1.26 area -- but the market makers kept buying, at prices that kept getting lower, until when the dust settled they were left long a lot of pounds that they'd bought for prices that were now much too high.

In the new system, the market makers are computers, and when things get hairy they just stop buying pounds and walk away with their computer hands in their computer pockets, whistling a jaunty tune out of their computer speakers. And so instead of a smooth but harrowing path down from $1.26ish to $1.24ish over the course of an hour, the pound fell to $1.1841 -- and "at least one electronic-trading platform recorded a transaction at $1.1378" -- over the course of two minutes, and then rapidly recovered as the computers, instead of buying all the way down to the new level, allowed the pound to fall to an absurd level and then bought all the way back up to the new level. This is a better trade. Buying on the way up is better than buying on the way down. These computers were smart.

This is incredibly schematic and over-simplified -- neither old-school human market making nor modern robotic market making really work quite the way I described, and the thin liquidity was due to it being overnight Asian trading, not just to conscious decisions by algorithms to get out of the way -- but I do think it captures something essential. The old ideal, which was never fully realized, involved market makers who stepped in to cushion big market moves by trading against them. But if the moves are real, trading against them is a money-losing strategy. The new, cruel, coldly logical regime involves electronic market makers who are too timid and capital-constrained and rational to trade against big moves. So they just let them happen -- in fact, let them go too far and too fast -- and then buy on the rebound, making money instead of losing it. Making money is better than losing money.

You can, and everyone does, complain about this system. "Flash crashes" are bad or whatever. But remember that the alternative, traditional system requires the market makers to trade, in a sense, irrationally, to lose money to cushion big fundamental moves. Are you volunteering to lose money to cushion those moves? Are you asking banks -- which are constantly told by regulators to reduce their trading risk -- to lose money to cushion those moves? In a sense, the reason that the algorithms have taken over a lot of market making from the banks is that they are better at it: They know to get out of the way of big moves and buy on the rebound. That is a better, more profitable, more efficient strategy, so of course it dominates. It is also worse service, as market making goes, and leads to flash crashes. But the explanation is not some weird exogenous accidental move to computerization and algorithms, not some failure of regulation or imagination. The explanation is that cushioning big moves is unprofitable and taking advantage of them is profitable.

One other point about "algorithms." Let's assume, schematically, that a lot of electronic market-makers got out of the way of the big move down. Then why did that move go too far down, reaching $1.18ish or $1.14ish before settling back up at $1.24ish? One possible answer is a "fat finger," where someone accidentally sold way more pounds than he meant to. (JPMorgan says, "the move was on high volume so it’s hard to justify this as a fat-finger move.") Another possible answer is that other trading firms' algorithms told them to follow momentum (or just reduce risk) and sell the pound all the way down. But a third answer is that a bunch of traders -- firms or humans or whoever -- had stop orders in place that were triggered when the pound crossed various previously undreamt-of levels. A stop order is in a sense an algorithm: It's an instruction to a computer that, if an event happens, the computer should sell pounds. When we talk about algorithms making trading more volatile, we are not always talking about inscrutable complex algorithms that no human can understand. Sometimes we just mean that people had orders to sell the pound if it got too low, and those orders were triggered, so it got lower.

Anyway it is not much fun to be in the FX trading business this morning.

Just a little more Salman.

Maybe the most important thing that happened in Wednesday's Supreme Court oral argument in the Salman insider trading case was this exchange between Bassam Salman's lawyer Alexandra Shapiro and Justice Sonia Sotomayor:

MS. SHAPIRO: But, Your Honor, that would be true in virtually any instance one could think of where an insider disclosed confidential corporate information, whether it's in a business setting, or, as is often the case, a mixed social and business setting. Analysts talk to company insiders all the time, and it's essential to the free flow of information to the marketplace that that occurs. And if --

JUSTICE SOTOMAYOR: Wait a minute. First of all, that's no longer true. There's regulations that stop that, talking to analysts.

Wait another minute. That's wrong too. Justice Sotomayor is referring to Regulation FD, the 2000 rule limiting selective disclosure of material information by companies. As a shorthand description, "companies can't disclose material nonpublic information to analysts" is a decent summary of Regulation FD. But "companies can't talk to analysts" is not. "Publicly traded U.S. companies held an average of 99 one-on-one meetings with investors apiece" in 2014. Perhaps none of those thousands of meetings are material? But, then, why do the investors bother with them?

Perhaps some of them violate Regulation FD? But Regulation FD is a civil issue, for the company, and rarely enforced: If a company selectively discloses some information to analysts, it might pay a fine, but probably not. No one goes to prison. No one goes to prison because insider trading law doesn't punish people for trading on material nonpublic information. It punishes them for trading on material nonpublic information that is disclosed corruptly, in particular, in exchange for a personal benefit to the person disclosing it. The U.S. does not have a "parity of information" insider trading law: Our markets encourage investors to try to find things out about companies and then trade on that information. As long as they don't bribe anyone for it.

It is no secret that prosecutors and the Securities and Exchange Commission don't love this situation, and would prefer more of a parity-of-information standard. That's why they brought a criminal case against Todd Newman for trading on information he got from a Dell investor-relations employee. And Newman was convicted, because -- and this is also no secret -- everyone kind of assumes that there is a parity of information standard. It just seems right. "Insider trading is not about fairness; it's about theft," I often say, because it needs repeating. Everyone assumes it's about fairness.  

But an appeals court reversed Newman's conviction, and the Supreme Court refused to review it. And now the Supreme Court is reviewing the Salman case, a dumb case of a brother tipping a brother for no obvious personal benefit. It's pretty easy to see why Salman's conviction should be upheld -- giving your brother a financial benefit seems pretty close to getting a benefit yourself -- but it's a troubling vehicle for making insider trading law. Deep down, people -- even Supreme Court justices -- are just suspicious of informed trading. They want everyone to have the same information; they want people who trade with secret information to go to prison. There are good arguments against that view, arguments that investors should be able to talk to companies because they are the companies' owners, and because their conversations make the companies better and the markets more efficient. But Bassam Salman -- convicted of trading on a tip that his brother-in-law got from his brother -- is not the guy to make those arguments.

What is market manipulation?

The Commodity Futures Trading Commission is fighting a pretty bizarre market-manipulation case against DRW Investments LLC. The CFTC argues that DRW manipulated the price of an interest-rate swaps futures contract by "banging the close." DRW does not exactly disagree, but argues that it wasn't manipulation because the contract was priced incorrectly and its trading moved it toward the correct price. This sounds like a pretty hazy philosophical discussion, but it's not particularly: There was a mathematical issue with the contract that DRW understood, and that no one else did, and that caused it to be mispriced; now everyone understands it and agrees that DRW's prices were "correct." (We discussed this before, ages ago, when the CFTC brought its case, and even earlier when DRW preemptively sued the CFTC.)

Now DRW has won a round:

A federal judge in Manhattan ruled Friday the CFTC needed to show that a trader intended to create an “artificial” price in order to prove attempted market manipulation. The CFTC had tried to lower the bar by saying it needed to show only that the trader had an intent to affect market prices.

If you think too hard about market manipulation you'll go mad, and so I advise you not to think too hard about either of these alternatives. DRW's approach, adopted by the court, is that you have to try to create an "artificial" price. But what could that ever mean? The DRW case, premised as it is on a mathematical error, is maybe the only time you'd ever be able to argue convincingly that just wanting the price of your investments to go up is different from wanting to create an "artificial" price. If I own Facebook stock, and I bang the close to make it go up, am I trying to create an artificial price? What, I clearly think Facebook is undervalued (that's why I own it), so what's the problem?

On the other hand the CFTC's definition is even crazier. Any time you trade, you will predictably affect market prices, which means that anyone who ever trades is guilty of market manipulation. (Doing X knowing that it will cause Y is, for most legal and philosophical purposes, the same as doing X with the intent to cause Y.) You can see why the judge rejected that idea, though the "artificial price" idea is only slightly better.

Fees and commissions.

Merrill Lynch announced a more or less inevitable result of the fiduciary rules for financial advisers dealing with retirement accounts:

The Bank of America Corp. brokerage unit told its more than 14,000 brokers on Thursday that after April 10, when the new rules take effect, investors who want a retirement account at Merrill will need to pay a fee based on a percentage of their assets, instead of having the option of being charged for each transaction made in their account.

So that's probably good. Merrill Lynch provides a service -- retirement investment advice -- and will now charge a clearly disclosed fee for that service, one which its clients will either think is fair and be willing to pay, or not (and will move their money elsewhere). The alternative approach of providing the service for "free" and making money by charging commissions, which the clients are less likely to notice and which are more likely to lead to conflicts of interest as Merrill thinks up creative ways to generate more commissions, is clearly less in keeping with fiduciary obligations, which is why Merrill is scrapping it. (For retirement accounts. Not for regular accounts.) 

Still you see the problem. For one thing, the fees will generally be higher: If you have a commission-based account and don't trade much, you can get away with paying less; now that option is gone. "Morningstar said fee-based accounts can yield as much as 60% more revenue than those that charge commissions." So the fiduciary rule will get retirement savers better advice and better alignment of interests with their brokers, but it might cost them more. And the fees will probably be clearer, which means that some people will be less happy about paying them even if they end up coming to the same dollar amount. So some of them will move money out of Merrill and, quite plausibly, get less or worse retirement advice than they otherwise would have. 

Kids these days.

We have talked before about the National Futures Association's enforcement action against teen trader and social media sensation Jacob Wohl, accusing him of, among other things, refusing to submit to a required NFA examination. Well now Wohl has filed his answer in the enforcement case, and it is ... let's say not focused closely on securities law issues? It's more like this:

The week after Wohl's resignation from the NFA, three men arrived at Wohl's home in Los Angeles. As Wohl came to learn over the following day, the men who were reported to the Los Angeles Police Department after peeping through the windows of his home, were from the National Futures Association. A complaint issued weeks later by the NFA admitted that they "looked through the second story window and saw someone". This comes as a disturbing revelation, as there is a serious likelihood that someone on the second floor, would have been someone changing in a bedroom.

Don't you wish every securities enforcement action was like this? Like, the Justice Department would sue Deutsche Bank for $14 billion for fraudulently packaging and selling mortgage-backed securities, and Deutsche Bank would reply "AHHH YOU SAW US CHANGING AHHH!" 

People are worried about unicorns.

No, the news from the Enchanted Forest is great, as Snapchat, the Unicorn With The Cartoon Dog Face (Elasmotherium kynopis), is preparing to go public at a $25 billion valuation. (Actually the company is now called Snap Inc., as I guess the "chat" part was too limiting.) Snap has negative net income, and had about $60 million in revenue last year, but this year has been much better, and it might do "as much as $1 billion" in revenue next year. So, sure, $25 billion valuation, why not. I think you get a valuation premium if none of your investment bankers or investors understand what your business is. I sometimes joke around here about investment bankers pandering to startups when they pitch for initial public offering mandates -- showing up in yoga pants to pitch Lululemon (that happened), or cooking at the pitch for Blue Apron (that probably didn't?) -- but I really want to know how the Snap pitches went. My joke was that they would involve sending nude pictures, but that is a very outdated old-person's view of what Snapchat is for. Because I am an old person and do not understand what Snapchat is for. I imagine many of the bankers are in the same boat. They splash some cartoon dog noses around the pitchbook, crack a joke about how investor demand in the order book won't disappear in 24 hours, and then lapse into a desperate silence. "But what does Snapchat do," they ache to ask, but can't.

Elsewhere, my Bloomberg Gadfly colleague Max Nisen says that "Theranos is no longer the Blood Unicorn," but I don't think that is a title that can be lost just by falling below a billion-dollar valuation and pivoting away from running blood labs. The Blood Unicorn's magic is an old magic, and far more powerful than the generic Silicon Valley idea that any billion-dollar startup is a "unicorn." Theranos, in its heart, and in my heart, will always be the Blood Unicorn. Now it is just a Zombie Blood Unicorn.

By the way, I realize that Theranos's essential problem is that it built a product that didn't work and told everyone that it did, while Snapchat's essential accomplishment is that it has become indispensable to millions of members of an extremely desirable demographic. Still, if you are a rational young entrepreneur, do you take any other lessons from Snapchat's success and Theranos's parallel debacles? Like: Curing disease is hard, building cartoony messaging services is where it's at? Is there any downside to that lesson?

People are worried about bond market liquidity.

Ehh I don't know but issuance is still going gangbusters. For instance in emerging markets:

Corporate-bond issuance in emerging markets hit a record high for the month of September, even as worries grew globally over soaring debt levels amid tepid growth.

No one is worried enough about liquidity to stop buying bonds.

Things happen.

"Deutsche Bank AG is holding informal talks with securities firms to explore options including raising capital should mounting legal bills require it." Aswath Damodaran on Deutsche Bank. Mars Cashes Out Warren Buffett to Take Control of Wrigley. 5 Years Later, Jon Corzine May Avoid Trial With $5 Million Settlement. A.I.G. Ex-Chief Ends Testimony in Fraud Case Unbowed. The Innocent Origins of Today’s Corporate Crimes. Trump’s Tax Strategy Seen Turning Unpaid Debt Into Benefit. Trump’s use of debts and tax laws spurs concerns about his methods. Donald Trump Urged Congress to Loosen Tax Rules in Early 1990s. "The one time Trump’s leadership skills were put to the test as an agent of middle-class people’s economic well-being, he ripped those people off ruthlessly and unapologetically." Calpers Sees Next Headache in Slowing Private-Equity Cash Gusher. Hurricane Matthew to Test Catastrophe-Bond Market. Introducing the 'mutualised database.' Do Battleground States Get a Break from Regulators? Executive Behind Parmesan Wood-Pulp Fraud May Get Jail Time. Academic debates in philosophy. We Have Been Blessed With This Picture of a Very Happy Dog Photobombing the Pope. "The Marquis de Sade and the office novel."

If you'd like to get Money Stuff in handy e-mail form, right in your inbox, please subscribe at this link. Thanks! 

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 mlevine51@bloomberg.net

To contact the editor responsible for this story:
James Greiff at jgreiff@bloomberg.net