- Banks can't do "proprietary trading" (which is investing in trading securities for their own account).3
- But they can do "market making" (which is investing in trading securities for their own account, but in a customer-facing business).
- And they can do "hedging" (which is investing in trading securities for their own account, but for the purpose of mitigating their risks elsewhere).
Banning proprietary trading is sort of dumb, but it's not the end of the world. On the other hand distinguishing prop trading from "market making" and "hedging" is very difficult, and Congress didn't really attempt it. It's also important: You want banks to be able to hedge their risks and make markets.4 (Or, I do, anyway, and Congress did.) So most of what the regulators have been doing for the last three years and change is trying to figure out how to do that.
There are two main approaches. One is for regulators to look at each trade and say, "Well, hmmm, this sure looks proprietary, I mean, you are using your money." Then they ask the bank to prove that the trade is not proprietary. If you buy some bonds from a customer at 99 at 10 a.m. and sell them to another customer at 99.25 at 10:15, you're probably good. If you call a customer and ask them to sell you some bonds at 99 on Tuesday, and you buy the bonds and hold them for a bit, and then on Thursday you sell a roughly similar amount of DV01 in different bonds in the same industry to hedge, and then the next week you sell your original bonds at 104 because the market moved, and then you buy a third company's bonds to keep your credit exposure roughly flat -- hahaha you didn't even finish reading this sentence, did you? Imagine telling it to a jury. Basically there's no way to know in advance that regulators will sign off on each trade that modern market-making desks actually do.
The other approach is to say, look, guys. You have proprietary trading desks. You know what they are. If you're not a jerk you put them on a different floor from your market-making desk so they can't front-run customer orders. They don't make two-sided markets. When they want to trade, they call your (or someone else's) market-making desk for a market. They have swaggery names like "Process Driven Trading."
Fire those guys. The rest shall keep as they are. Your market-making desk can keep doing what it's doing. The people hedging your loan book can keep doing what they're doing. But don't be an idiot, you know? Don't morph your market-making desk into a prop desk. Don't use the word "hedging" to cover a big prop desk.
The first approach would look at each trade and say, "Is this a proprietary trade, or is it market-making or hedging?" That question is mostly not answerable without knowing the trader's deepest hopes and dreams and probably not even then.
The second approach would look at each trading desk and say, "Is this a prop desk or a market-making desk or a hedging function?" And then it would answer that question in more or less the way a bank would answer that question: If the desk's primary function is to proactively add risk, it is prop; if it is to make the spread by dealing with customers, it is market making. And then, if the answer is "market making" or "hedging," the rule would say, OK, you can keep this desk, but manage it in a sensible way, and don't let it go off and do roguey things.
The 2011 proposed regulations looked a whole lot like the first approach. Today's final rule looks a whole lot like the second. This is particularly clear in market making. Here is a summary of comments on the draft rule, from the final adopting release:
Several commenters expressed general concern that the proposed exemption may be applied on a transaction-by-transaction basis and explained the burdens that may result from such an approach. … Other commenters indicated that it would be difficult to determine whether a particular trade was or was not a market-making trade without consideration of the relevant unit’s overall activities. One commenter elaborated on this point by stating that “an analysis that seeks to characterize specific transactions as either market making…or prohibited activity does not accord with the way in which modern trading units operate, which generally view individual positions as a bundle of characteristics that contribute to their complete portfolio.” This commenter noted that a position entered into as part of market making-related activities may serve multiple functions at one time, such as responding to customer demand, hedging a risk, and building inventory. The commenter also expressed concern that individual transactions or positions may not be severable or separately identifiable as serving a market-making purpose.
That "bundle of characteristics" commenter5 puts it eloquently. A market maker is not in the business of, like, buying 100 IBM calls from Customer X and then selling 100 identical IBM calls to Customer Y. A market maker is in the business of having a book and looking at his delta and gamma and vega and fitting each new trade into his overall risk position, saying, "Hmm I seem to be short a bit of short-dated IBM vega, I should go buy some calls from someone, I wonder what Customer X is up to." Overall the market maker's business involves making markets for customers, but at each second he is thinking at least as much about his book's proprietary risk as he is about customer service.
So the final rule comes down on the side of desk-by-desk rather than trade-by-trade analysis:
For instance, the final market-making exemption does not require a trade-by-trade analysis, which was a significant source of concern from commenters who represented, among other things, that a trade-by-trade analysis could have a chilling effect on individual traders’ willingness to engage in market-making activities. Rather, the final rule has been crafted around the overall market making-related activities of individual trading desks, with various requirements that these activities be demonstrably related to satisfying reasonably expected near term customer demands and other market-making activities.
The activities need to be "demonstrably related to satisfying reasonably expected near term customer demands," but they don't have to just be "buying and selling to customers." So the release specifically discusses a desk that "makes a market in a variety of U.S. corporate bonds and hedges its aggregated positions with a combination of exposures to corporate bond indexes and specific name CDS in which the desk does not make a market." That's all fine! You just have to not be an idiot about it: To qualify as a market-making desk, you have to demonstrate that:
(iii) the trading desk’s exposures to corporate bond indexes and single name CDS are designed to mitigate the risk of its financial exposure, are consistent with the products, instruments, or exposures and the techniques and strategies that the trading desk may use to manage its risk effectively (and such use continues to be effective), and do not exceed the trading desk’s limits on the amount, types, and risks of the products, instruments, and exposures the trading desk uses for risk management purposes; and (iv) the aggregate risks of the trading desk’s exposures to U.S. corporate bonds, corporate bond indexes, and single name CDS do not exceed the trading desk’s limits on the level of exposures to relevant risk factors arising from its financial exposure.
There's a lot of stuff in the rule to the effect of "if a desk exceeds its risk limits maybe you should do something about that." So … yes! If a desk exceeds its risk limits, maybe you should do something about that. Yes, a desk that uses credit-default swaps to hedge its inventory should be using credit-default swaps that, y'know, actually hedge its inventory. A lot of the market-making provisions in the final rule are this sort of good-housekeeping stuff. The proposed rule came from a place of suspicion, using lots of specific measures of risk and profits to try to second-guess whether a trade was prop or market making. The final rule comes from a place of good management: You're supposed to run a primarily market-making desk, and you're supposed to run it in a sensible risk-managed way.
Similarly on hedging. There's a lot of discussion in the press about "portfolio hedging" that I simply don't understand; "portfolio hedging" is not a thing. The thing is that if you want to hedge a risk -- of one position or many aggregated positions -- you now have to jump through some hoops. But those hoops are like, one, you actually have some reasons to think it hedges something, and two, you have to look at it every now and then and see if it's hedging that thing. The rule text requires that the hedge:
At the inception of the hedging activity, including, without limitation, any adjustments to the hedging activity, is designed to reduce or otherwise significantly mitigate and demonstrably reduces or otherwise significantly mitigates one or more specific, identifiable risks, including market risk, counterparty or other credit risk, currency or foreign exchange risk, interest rate risk, commodity price risk, basis risk, or similar risks, arising in connection with and related to identified positions, contracts, or other holdings of the banking entity, based upon the facts and circumstances of the identified underlying and hedging positions, contracts or other holdings and the risks and liquidity thereof;
And that it is subject to "continuing review, monitoring and management" that "Requires ongoing recalibration of the hedging activity by the banking entity to ensure that the hedging activity satisfies the requirements" of the rule.
This probably would not forbid JPMorgan Chase & Co.'s London Whale trades, which were sort of nebulously designed to protect against corporate-credit gap risk and then sort of mutated into confusion. It just would require JPMorgan to actually figure out and write down what they were meant to hedge, and then check on them periodically and ask if they were working. Presumably that sort of discipline would mitigate JPMorgan's losses.
Also it is just a good idea! Sure a lot of the Volcker rule is highly over-engineered checklists and admonitions that boil down to "don't be dumb," and sure there are good theoretical arguments against that sort of regulation, but I don't know, come on. You shouldn't be dumb. You could do a lot worse than a rule that requires you to think about what you're doing.
There is a lot of housekeeping in the final rule -- there's an Appendix A requiring a whole new set of reports and record-keeping on desk risk limits, risk factor sensitivities, value-at-risk, sources of revenue and so forth -- but my sense is that, for all the grumbling on both sides, the final Volcker rule should not cause any significant changes in how big banks run their trading desks. The prop desks are gone, but the market-making desks can still do all the stuff that they normally do: deal with customers, market to customers, anticipate customers, accrue inventories, trade with other dealers, lean in various directions on various sensitivities based on their own sense of the market and generally act as trading desks rather than as pure order-takers.
That's good! That's what you want! The Volcker rule was supposed to shut down proprietary trading desks, and it basically already has. The Volcker rule was not supposed to shut down market making and hedging, which are risky and proprietary and complicated and all that good stuff, but which are also both economically important and specifically allowed by Congress. That's a simple sentence to write, but a hard thing to make happen. The final rule is 978 pages long, but it's not a bad effort at achieving that simple result.
1 That link is to the final rule text at the Davis Polk Volcker rule website, which is a good source for all things Volcker.
2 So does Bart Chilton, though, which is troubling. I dunno, his spin on the market making provisions seems to come from a different rule from the one I read:
Yes, market making is allowed, but only for the benefit of the banks’ customers -- for their customers and not in order to collect market maker fees provided by the exchange or for any speculative reason. The market making is only permitted when a bank is hedging a legitimate business risk for a customer. Full stop.
But risky non-trading securities are fine! Lending money is fine! As Dan Davies puts it, "The Volcker Rule is incomplete; it still allows banks to make risky proprietary investments in illiquid amortising credit instruments." (Those are called loans.)
4 Alternatively you could get rid of bank market-making and just leave that activity to non-banks. This does not seem to be what Congress intended when it wrote a rule saying that market-making is okay. In any case that transition would not be costless. From the release adopting the final rule:
Other commenters, however, indicated that it is uncertain or unlikely that non-banking entities will enter the market or increase their trading activities, particularly in the short term. For example, one commenter noted the investment that banking entities have made in infrastructure for trading and compliance would take smaller or new firms years and billions of dollars to replicate. Another commenter questioned whether other market participants, such as hedge funds, would be willing to dedicate capital to fully serving customer needs, which is required to provide ongoing liquidity. One commenter stated that even if non-banking entities move in to replace lost trading activity from banking entities, the value of the current interdealer network among market makers will be reduced due to the exit of banking entities. Several commenters expressed the view that migration of market making-related activities to firms outside the banking system would be inconsistent with Congressional intent and would have potentially adverse consequences for the safety and soundness of the U.S. financial system.