FX Trades and Hypothetical Deals

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.
( Updated
)
a | A

FX flash crashes.

The pound closed September trading against the dollar at $1.2972; a week later, last Friday, it closed at $1.2434, a drop of more than 5 cents. Here's how it looked:

What do you think about that chart, aesthetically? The near-vertical up-and-down early on Friday -- the pound's "flash crash" last week -- obviously detracts from the overall harmony of the composition. But what about the previous four days? They're fine, boring, normal, moseying along, moving the pound down slowly from $1.30ish to $1.26ish -- very different from Friday's near-instantaneous, jagged whoosh from $1.26ish to the $1.24ish where the pound finished the week.

But how would you feel if you bought pounds at $1.29 on Monday morning? Now your pounds are worth $1.24ish. Whatever information drove the pound from $1.29 to $1.24 over the course of the week, don't you kind of wish it had been discovered and incorporated into the price all at once, so you could have just bought at $1.24? Wouldn't a jagged graph -- an immediate vertical fall to $1.24ish, and then just staying there -- be preferable, for you? Does that slow mosey down represent orderly trading, or does it represent an inaccurate price struggling to incorporate new information? In a sense, isn't the Friday chart better? I mean, sure, there's that horror at the start of the day, but you probably weren't even awake for that. And it was quickly sorted, and then the pound mostly settled around its new price, without the previous days' long delay in getting there.

Here's an article about how it's bad that banks are retreating from currency trading and that electronic market makers are taking their place:

Robert Savage, chief executive of the currency hedge fund CCtrack Solutions LLC and a former Goldman Sachs Group Inc. executive, said when volatility increases to a point that high-frequency trading programs lose money by making markets, “they just shut down.”

Bank trading desks staffed by employees would have kept the selloff from escalating to the degree it did, he said. “You would find out who was selling, if there was any news,” Mr. Savage said. “In the old days, if there was no fundamental reason to sell, you wouldn’t sell.”

Yes sure that is fine but one thing that that chart suggests is that there was a fundamental reason to sell. I mean, far be it from me to tell you what it was -- "something something something Brexit," was my summary on Friday -- but the pound was steadily down over the week (and is down again this morning). The price, to anthropomorphize a bit, wanted to move down. A trading desk that kept the selloff from escalating from $1.24ish down to $1.18ish -- which the pound hit at the worst of the flash crash -- would have done a nice service, keeping prices near the level justified by actual demand. But a trading desk that slowed down the selloff from $1.26ish (or $1.30ish) to $1.24ish would have been artificially inflating prices, slowing down price discovery, keeping the price above the level of actual demand, manipulating the market in the service of a smoother graph.

Here is more on the "liquidity illusion" and flash crashes and so forth. ("If such incidents take place regularly, this is very negative for trust in the financial markets.") And here is David Keohane on the rise of algorithms in foreign exchange trading ("greater efficiency but gappier action"). The good news is that the market reacts quickly to news; the bad news is that it overreacts. That bad news is bad, but that good news is good, and easy to miss. A market that reacts to fundamental news by slowly drifting down over the course of hours or days looks orderly, but it is underreacting to the news; if it takes hours or days to incorporate the news, then the market is not doing its job as efficiently as it could be. A flash crash means that a lot of people are trading pounds at the wrong price -- $1.18 instead of $1.24 -- but a fundamental market move that is cushioned and spread out over time also means that. And more people, for a longer time.

By the way, a good view into high-frequency-trading psychology, as it were, is this paper by Vincent van Kervel and Albert Menkveld:

Liquidity suppliers lean against the wind. We analyze whether high-frequency traders (HFTs) lean against large institutional orders that execute through a series of child orders. The alternative is that HFTs go “with the wind” and trade in the same direction. We find that HFTs initially lean against orders but eventually turn around and go with them for long-lasting orders. This pattern explains why institutional trading cost is 46% lower when HFTs lean against the order (by one standard deviation) but 169% higher when they go with it. Further analysis supports recent theory, suggesting HFTs “back-run” on informed orders.

Their data is from Swedish equities, but the intuition is a broadly useful one: Electronic market makers cushion small market moves, trading sort of like normal market makers -- buying when others are selling and selling when they're buying -- to capture the spread. But once it's clear that everyone is selling, the algorithms are no fools, and they start selling too. That's -- loosely speaking -- how you get four days of orderly moseying down, and then one sharp shock that causes the market to vastly overreact and then quickly recover.

Deutsche Bank.

One of the joys of banking is its deep unreality, its abstraction, its distance from the messy everyday world of goods and services, but I feel like at this point Deutsche Bank's executives would probably be grateful if something could actually happen. They seem to be spending too much time living hypothetically these days. The big event in Deutsche's life is of course the $14 billion fine that the U.S. Department of Justice may or may not extract from it, but that fine remains tantalizingly hypothetical:

Deutsche Bank AG Chief Executive Officer John Cryan failed to reach an agreement with the U.S. Justice Department to resolve a years-long investigation into its mortgage-bond dealings during a meeting in Washington Friday, Germany’s Bild newspaper reported.

The meeting was meant to negotiate the multi-billion-dollar settlement the bank will have to pay to resolve alleged misconduct arising from its dealings in residential-mortgage backed securities that led to the 2008 financial crisis.

But even more purely abstract is this stress-test story. Like the other big European banks, Deutsche Bank had to pass a stress test administered by the European Central Bank and the European Banking Authority. The stress test consists of taking Deutsche's current financial situation, estimating the effects of an assumed economically stressful situation, and seeing if Deutsche would have enough capital left after those stresses. So it is, to begin with, a hypothetical exercise. But the ECB more or less chooses the hypotheses, and one of them is that you don't get to count any deals that you hadn't completed by the end of 2015. ("Any divestments, capital measures or other transactions that were not completed before 31 December 2015, even if they were agreed upon before this date, should not be taken into account in the projections," say the rules.) But Deutsche did:

Deutsche’s results included the $4bn proceeds from selling its stake in Chinese lender Hua Xia even though the deal had not been done by the end of 2015, the official cut-off point for transactions to be included.

The Hua Xia sale was agreed in December 2015. It has still not been completed and now faces a delay after missing a regulatory deadline last month, though the bank is still confident of completion this year.

Counting that as-yet-hypothetical deal improved its stressed common equity Tier 1 capital results from 7.4 percent to 7.8 percent. But either would be fine: Deutsche would have passed the stress test without counting the divestiture. But it counted the divestiture. Which is apparently against the rules, and other banks didn't count similar post-2015 deals (even at least one that closed in early 2016, before the stress test). But the ECB signed off on this one. I don't know what to tell you? It is all a piling of hypotheticals on hypotheticals, and it seems somehow ridiculous to care about any of it, except that if you are a bank investor the only thing you really have to hold onto is hypotheticals, so you might as well care about them. 

Banking salaries.

Frankfurt is trying to win post-Brexit banking jobs from London by perhaps "imposing an upper salary limit on employee protections of €100,000 or €150,000, which would make conditions such as redundancy terms less generous":

One of Germany’s main drawbacks as an international financial centre, often cited by bank bosses, relates to labour laws that make lay-offs difficult and costly. Minimum statutory redundancy terms, for example, are twice as generous in Germany as in the UK. That is a particular issue for banks, which tend to hire and fire more readily than other sectors, in line with cyclical business fortunes.

There is a tendency among banking critics to take a cynical view about variable compensation, but Frankfurt's troubles suggests that banks are serious about wanting the flexibility to pay their employees a lot in good times, less in bad times, and to be able to get rid of them cheaply in really bad times. Generally speaking the post-crisis European model, with its bonus caps, has been to encourage fixed compensation, which might make banks psychologically safer (it's less attractive to take big risks to get bonuses) but financially riskier (it's harder to lower pay in lean years). Lowering severance costs seems even better: It makes banks psychologically safer (it's less attractive to take big risks since you're more likely to get fired if they don't work) and also financially safer (it's easier to lower headcount in lean years). The only problem with it is that, if you believe that banks are socialist collectives run for the benefit of their workers, why would the workers want to reduce worker protections? And yet it seems like the banks do want it.

Twitter.

This quote is about Twitter, but in a sense isn't it about life?

"From a consumer standpoint, it’s a little like a promise that remains unfulfilled -- users are trying to inject some meaning into an experience that remains rather undefined at this point."

That's a marketing professor, quoted in this Bloomberg article about how Twitter hasn't found anyone to buy it and is now condemned to be free, without excuse, responsible for everything that it does. Without a buyer, Twitter must face reality alone and unafraid and ponder how best to achieve meaning in its corporate existence. Probably through live video, that's what most people do:

If a buyer doesn’t appear, Twitter will to try to appeal to more users through a new strategy that emphasizes live video. The company has been entering partnerships for sports, politics and entertainment content -- such as the National Football League’s Thursday night games -- that it can stream alongside tweets related to the video. It may give people without Twitter accounts a new way to use the service, while allowing the company to share revenue on the video ads.

I suppose there are two schools of thought about when you should sell your company, relative to when you should figure out what your company does. One view would be that you should figure out your business, inject meaning into your experience, mature as a company, etc., and then find someone to buy you and give your shareholders the full value of the business you developed. The other view would be that you should sell yourself while there's still some mystery about you, when neither you nor the buyer quite know what it is that you do. The risk here is that, if you sell too early, your shareholders may not capture 100 percent of the value of the business you created. The advantage is that they may capture way more than 100 percent.

Insider trading.

Here is a bizarre story about Julian Rifat, a former Moore Capital trader who pleaded guilty to insider dealing in the U.K., and who was apparently recruited by Peter Nygard, a personal enemy of Moore founder Louis Bacon, to provide evidence that Bacon and Moore were also insider trading. There doesn't seem to have been any such evidence -- Rifat met with U.S. prosecutors, but he says "I deny that Louis Bacon and/or Moore Capital engaged in unlawful activity, or that I provided anyone with any evidence of their having done so." The article goes on to say that "prosecutors didn’t think there was evidence for them to build a case" -- but, still, quite an effort from Nygard, who had become Bacon's enemy because of a dispute over their neighboring vacation homes, as one does. Loosely speaking the bid/ask in U.S. insider trading law right now is whether prosecutors need to provide concrete evidence of a corrupt tipping arrangement, or whether circumstantial evidence that a stock tip didn't look right is enough to send someone to prison. If the Supreme Court chooses the latter, lower standard for insider trading, do you think we'll see an increase in billionaires trying to get each other sent to prison for it?

Governance.

Last night Donald Trump completed his efforts to turn the U.S. presidential election into a grotesque reality television show, with predictably horrifying results. Trump seems to be in "gambling on redemption" mode: His campaign is underwater, so he pursues strategies -- like threatening to imprison his political opponent and discussing her husband's behavior -- with low expected value but high variance. They're unlikely to work, but just doing normal things is also unlikely to work, so you might as well do something risky. 

As in corporate finance, gambling on redemption can be a good strategy for management even as it's a bad one for stakeholders, who would prefer a high chance of some recovery (with zero chance of complete success) over a high chance of no recovery (with a small positive chance of success). In Trump's case, the gambling seems to be bad for his party, and bad enough for donors  that they are asking for their money back. But I am sometimes struck by how, unlike corporate finance, political campaigns seem to come without governance mechanisms. If you give a candidate money, you have no control rights over what he does with that money, or with his office. (Obviously! That would be bribery.) You can't even get your money back. And here is Adam Davidson on political parties as a governance mechanism, which also ... is not working.

Without any formal governance rights, you have to rely on the most basic governance mechanism: the character of the agent. You have to trust someone, with few ex post checks, so all you can do is make the best possible ex ante choice of someone trustworthy. It is striking that so many people who are essentially in the business of choosing good agents to trust, chose Trump. There are no surprises here. As Trump, incredibly, recites at his rallies: "You knew damn well I was a snake before you took me in."

People are worried about unicorns.

"Nobody’s talking about the biggest, most obvious problem with Uber," writes Felix Salmon, and he is of course referring to poor user interface design. I don't really disagree; as a non-millennial, I could handle pushing a button to get an Uber, but having to choose between an UberPool or UberX or UberBlack or UberRush is too daunting for me. Elsewhere in massive unicorns whose user interface frightens me, here are Bloomberg Gadfly's Shira Ovide and Rani Molla with "Five Charts Explaining Why Snapchat's Worth $25 Billion."

People are worried about bond market liquidity.

I spent last Friday at the Notre Dame Conference on Current Topics in Financial Regulation, at which there was a certain amount of worrying about bond market liquidity. For instance, Jens Dick-Nielsen presented a paper on "The Cost of Immediacy for Corporate Bonds," and Stacey Jacobsen presented one on "Capital Commitment and Illiquidity in Corporate Bonds," both of which try to answer the question: If bond market liquidity has gotten so bad, why doesn't it show up in any statistics about bid-ask spreads, price impact, etc.? One short answer is that it does show up if you know where to look: in measures of how much capital dealers will commit to trades, and of how much funds pay to access immediacy around index-drop events. Also Itay Goldstein presented on "Investor Flows and Fragility in Corporate Bond Funds," because if you are worried about bond market liquidity, worries about runs on bond mutual funds can't be far behind.

Elsewhere: "Sweden’s $34 Billion Pension Fund Can’t Be Happier Out of Bonds."

Things happen.

Hart and Holmstroem Awarded 2016 Nobel Prize in Economics. Tyler Cowen on Oliver Hart, and on Bengt Holmström; Alex Tabarrok on "The Performance Pay Nobel." The Dash For Cash: Leaked Files Reveal RBS Systematically Crushed British Businesses For Profit. Wall Street bosses warn on Brexit risks. Cashless Sweden. How the Oracle of Oxford Won a Cult Following. China Intensifies Push to Cut Debt With Steps to Reduce Leverage. Activist funds change tack to fill up board seats. Lawyers are finally converts to technology. Tel Aviv Stock Exchange Loses Second Chief in as Many Months. "Now some Tesla shareholders are saying a Tesla-SolarCity combination could finally bring changes to the boardroom." Musk Says Tesla, SolarCity Don’t Need to Raise Cash This Quarter. UBS exited Wanda deal over compliance. $1.6 Million Bill Tests Tiny Town and ‘Bulletproof’ Public PensionsSuper-silk. Pets on Pot. Even After Criticism, Men Think Highly of Themselves.

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

(Corrects attribution of second quote in "Insider trading" section of article published Oct. 10.)

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