There's a huge article by Frank Partnoy and Jesse Eisinger in the Atlantic today about how banks are so horribly complicated that even sophisticated investors, meaning basically Bill Ackman, don't trust them any more. I suppose this provides an excuse for me to trot out a toy theory that's been congealing in my head, which is roughly that you can have two of the following in a publicly traded financial company, but not three:
- "market making,"
Like: you could, or someone could, make some educated guesses about what goes on at a bank that takes deposits and makes mortgage and commercial loans, and decide whether or not it's a good business that you should invest in. You could even make such guesses and decisions about what goes on at an old-school broker-dealer that trades securities for a living. (Maybe? I'm less confident of this leg.) But for universal banks I'm kind of with Ackman - and Partnoy and Eisinger - that you have to give up on making intelligent decisions, or failing that have your intelligent decision be "I'm gonna need a 50% discount to book value before I invest in this thing."
You can attribute the opacity of modern banking to like "bankers are eeeeeeevil," which I don't particularly believe, or "regulators are weeeeeeeak," which seems sort of lame. Me I tentatively like "banking + market-making = insoluble complexity."
The heart of the Atlantic article is an attempted deep dive into Wells Fargo's financials, which ends up splattering against the rocks that guard those financials. Here is where things start to get alarming, for some value of alarming:
Suddenly, this folksy mortgage bank starts showing signs of a split personality. It turns out that trading activities, the type associated with Wall Street firms like Goldman Sachs and Morgan Stanley, contribute significantly to each of Wells Fargo’s two categories of income. Almost $1.5 billion of its “interest income” comes from “trading assets”; another $9.1 billion results from “securities available for sale.” One billion dollars of the bank’s “noninterest income” are “net gains from trading activities.” Another $1.5 billion is income from “equity investments.”
Is this shocking? I dunno; it's not really true that Wells Fargo is the folksiest and most mortgage-banky bank of all the banks. There are banks that do not have significant trading activities; here's one, kind of, complete with folksy name. I'm not that shocked that WFC trades derivatives but then I used to do derivatives deals with them.
Next, though, what are these trading activities? Well, they're "market making," but you can dress it up alarmingly:
The bank breaks what it calls “net gains from trading activities” — which doesn’t cover all of its trading income, but is an important part — into three subcategories, leaving the annual-report reader to play a kind of shell game.
Look first at “proprietary” trading — activity a firm undertakes to make money for its own account by buying or selling stocks, bonds, or more-exotic financial creations. Self-evidently, this [is bad but the reported numbers are small, a $14mm annual loss]. ... A second subcategory is “economic hedging.” An activity labeled “hedging” might sound soothing. [But maybe it isn't. Again, though, a $1mm annual loss, so NBD.] ...
Finally we come to a third shell — and there’s unquestionably something to see under this one. It carries an innocuous label: “customer accommodation.” Wells Fargo made more than $1 billion from customer-accommodation trading in 2011. How did it make so much money merely by helping customers? This should be a plain-vanilla business: a broker sits between a buyer and a seller and takes a little cut of the transaction. ... Here is the most helpful of the bank’s disclosures related to customer-accommodation trading:
For the majority of our customer accommodation trading we serve as intermediary between buyer and seller. For example, we may enter into financial instruments with customers that use the instruments for risk management purposes and offset our exposure on such contracts by entering into separate instruments. Customer accommodation trading also includes net gains related to market-making activities in which we take positions to facilitate expected customer order flow.
Bankers, and their lawyers, are careful about the language they use in annual reports. So why did they use the word expected in discussing customer order flow in that last sentence? Is Wells Fargo speculating based on what one of its traders “expects” a customer to do, instead of responding to what a customer actually has done? The language the bank pointed to for answers to our questions only raises more questions.
Well I can answer that one: Of course Wells Fargo is speculating based on what one (most!) of its traders expects a customer to do! That is what market making is.1 If Wells Fargo didn't do that they'd be schmucks! If you're a market maker who only responds to customers, you are not very good at your job.2
There are lots of reasons for opacity in bank financial statements but surely a lot of them have to do with market making. For one thing: derivatives, a major villain in the Atlantic piece. Basically, OTC derivatives market-making doesn't net as cleanly as does, like, buying and selling publicly traded shares of stock. In cash equities, you buy 100 shares of IBM from one customer and sell 100 shares to another customer and clip two cents and are left with zero shares. In derivatives, you buy $100mm of 7-year Libor swaps from one customer and sell $100mm of 6-year Libor swaps to another customer and sell $10mm of 8-year Libor swaps to a third; you're left "flat" - i.e. zeroish DV01 - but have $210mm of derivatives notional for six-plus years. If you were running a directional investing business with Wells Fargo's balance sheet - $1.4-ish trillion - and ended up with $2.8 trillion in derivatives notional you'd be ... aggressive; if you were running a matched book then $2.8 or $28 or $280 trillion are all at least theoretically possible and one is not necessarily riskier than another.
And a market making business changes over time because you're trading every day - and in some loose sense you're trading around a baseline of zero (unlike a lending business where, y'know, you'll always have some loans). A disclosure in your 10Q that said something like "as of the end of the first quarter, we had $5mm of IBM stock on the trading book"3 would be about 40 days stale. By now, you might have $10mm, or zero, or negative $5mm.4
And, of course, being in the market making business makes you want to be opaque. As a retail customer, I know with some certainty that my imaginary local community Bank of Ye Olden Tymes is structurally long duration, since it's basically in the business of borrowing money short via overnight deposits and lending it long via 30-year fixed-rate mortgages. But there's nothing I can do about it because I don't, like, trade mortgages with Bank of Ye Olden Tymes. I just borrow long and deposit short. Maybe if I'm clever I buy a long-term CD to take advantage of BoYOT's exposure to duration. But there's not a lot I can do to pick them off.
But there's tons to be done to pick off universal banks that do a lot of market making. A bank that discloses "we are long a ton of trading exposure X" will find itself getting much lower offers to take X off its hands. You might recall a certain cetacean.
If you believe this toy theory then you need to basically decide: are the advantages of market making by banks (access to capital, customer convenience, etc.) enough to offset the disadvantages (opaque financials, vague Volckery fears)? If you don't, you can do whatever you want, but you'll still run up against that question. Partnoy & Eisinger's prescription is more principles and more jail:
What if legislators and regulators gave up trying to adopt detailed rules after the fact and instead set up broad standards of conduct before the fact? For example, consider one of the most heated Dodd-Frank battles, over the “Volcker Rule” ... Congress and regulators could have written a simple rule: “Banks are not permitted to engage in proprietary trading.” Period. Then, regulators, prosecutors, and the courts could have set about defining what proprietary trading meant. They could have established reasonable and limited exceptions in individual cases. Meanwhile, bankers considering engaging in practices that might be labeled proprietary trading would have been forced to consider the law in the sense Oliver Wendell Holmes Jr. advocated [i.e. by guessing, not what customers, but what judges, will do]. ...
The standard of proof for securities-fraud prosecutions, meanwhile, could and should be reduced from intent, which requires that prosecutors try to get inside the heads of bankers, to recklessness, which is less onerous to prove than intention, but more so than negligence. ... Senior bank executives should face the threat of prosecution the same way businesspeople do in other areas of the economy. When a CEO or CFO sits holding a pen, about to sign a certification that his or her bank’s financial statements and controls are accurate and adequate, he or she should pause and reflect that the consequences could include jail time. If bank directors and executives had to think through their institution’s risks, disclose them, and then face serious punishment if the disclosures proved inadequate, we might begin to construct a culture of accountability.
But that Volcker Rule prescription seems much too quick, doesn't it? Like:
- Either market making is "prop trading" or it isn't.
- If it isn't, then you need some serious convolutions to separate market making from prop trading, because they have meaningful overlaps.
- If it is, then you need to forthrightly ban "banks" from "market making"5 rather than accidentally doing it by banning "prop trading."
Talking about principles and allowing that there will be "reasonable and limited" - but unspecified! - exceptions is a way to sound reasonable without addressing that dilemma. In practice, I'd think that if regulation said (1) "no prop trading," (2) "we won't tell you what prop trading is," and (3) "if you prop trade you go to jail," then that would have the same effect as banning market making, only in a more confusing and lawless way. Which doesn't seem particularly principled.
1.I recently read this book "How the Trading Floor Really Works" and liked it, and while it might be a good idea to read it if you plan to be a trader or banker or what have you, it is definitely a good idea to read it if you plan to write long eyebrow-raisy articles about how bank trading is speculative and gambling and all that bad stuff. Is it speculative and gambling and all that bad stuff? Ehh ... it is just irreducibly what it is.
2.Let's have a little chat about "customer accomodation" and similar phrases, "customer facilitation" and the like. These are phrases that banks use to describe market making and they've always struck me as curiously passive; in fact, only well into the Volcker Rule debate did it click for me that the trades that I used to market to clients - where I went out to try to convince clients to do trades with my bank where we were the principal facing the client on hundreds of millions of dollars of derivatives notional - were "customer facilitation" rather than "proprietary." "Customer accomodation" doesn't really mean that a customer comes to a bank looking for an exposure and the bank passively provides it. "Customer accomodation" just means that the bank receives the bid/ask spread, rather than paying it. This is probably not intuitive to everyone! (If it's not to you, see note 1.)
That is, of course, not at all scandalous. It would be totally weird if, for instance, Apple just built devices in response to written requests from customers saying "oh hey I could totally use a computer in my pocket, could you make one?" "Innovation," a loaded word in finance, and, y'know, marketing, are parts of the job of modern banks. It's just that everyone feels a faint unease about that, so they call the stuff that they market "customer facilitation," which, like, would sound weird from most other businesses that sell stuff to customers.
3.Surely "$5mm deltas" but that's another point. The "sophisticated investors" who lament to Partnoy & Eisinger that they want better bank exposure numbers probably really do want, y'know, exposures: Greeks and DV01s and sensitivities. But imagine the complexities involved in:
- Computing your, like, vega to the S&P 500 or whatevera across your equities portfolio,
- Conveying that to your lawyers to put in your 10-K,
- Delivering that 10-K to retail investors and journalists and whatnot,
- Reading magazine articles about how complicated your disclosures are,
- Getting sued by people who disagree with the model you used to compute that vega,
I dunno it boggles me a bit. I feel like Ackman - and Partnoy, for that matter, a former derivatives salesman - would like that; obviously if you were like going to make rational economic decisions about whether to invest in a trading business you'd probably prefer model sensitivities rather than notionals as a data point. But, I dunno, vegas?
a.Which is like a correlation thing, etc. etc., since a lot of your exposure is single-name or ETF etc. etc., and quite model-dependent.
4.The perhaps-most-useful piece of exposure disclosure that banks do give - VaRs - suffers from this; Wells Fargo doesn't break out VaR in the detail that the bigger trading banks do but does say "The average one-day VaR throughout third quarter 2012 was $19 million, with a lower bound of $12 million and an upper bound of $32 million." So how risky is WFC? I suspect it's riskier overall than it is on a day of average riskiness, no? And today it is ... almost certainly either riskier or less risky than it was on an average day in the third quarter of 2012.
5.Which ... you could do? There are good path-dependency reasons not to do this - it would be a pain in the ass to break up JPMorgan, for one thing - but also arguably good theoretical reasons for doing it?