The rulemaking process for implementing the Volcker rule, which will ban proprietary trading by banks, seems to be proceeding by periodic leaks to the media about the draft rules - which are always described as having 174 pages but also always discussed one small bit at a time. Today’s leak is from Bloomberg and is about pay:
The rule, which aims to ban most proprietary trading by banks with federally insured deposits, would exempt trades related to market-making as long as the activity met at least seven standards, or principles. One principle would be that traders get paid from fees and the spread of the transactions rather than the appreciation or profit from their positions, according to a copy of the draft reviewed by Bloomberg News.
As a quick aside, I don’t really know what a “spread” is in this context. If you buy $10MM bonds from Client X at 101 and sell $10MM of the same bonds to Client Y at 101.25 a second later, then your spread is pretty clearly a quarter on $10MM bonds. Easy enough. If however you wait 3 days and sell the bonds at 103, is the 2 points “spread” or “price appreciation” or a combination that the payroll department would have to disaggregate?
That quibble aside, though, this is actually quite clever and diabolical. It is difficult to separate “permitted” inventory and hedging activities from “not permitted” proprietary position-taking. But if you just tell traders “all you can do is get paid on volume of transactions times commission that you can charge,” then there’s no incentive to do anything but flow transactions. Making money on hedges, inventory, etc. can’t help the trader and so almost by definition is irrelevant to him.
It is predictably easy to find someone to say “people who aren’t paid giant gobs of money will quit and go to hedge funds and do dark evil things in unregulated corners of the world.” Specifically:
A forced change to pay structure “could have the effect of driving people out of the regulated industry to the unregulated industry,” said Douglas Landy, a partner at Allen & Overy LLP who once worked at the Federal Reserve Bank of New York.
That doesn’t strike me as the big problem. The big problem to me runs something like this. We – someone – seem to think that it is a good thing to have market makers. There are probably places where it isn’t that important, where a robot could just match buyers and sellers. As far as I can tell that is how the stock market actually works. But there are lots of places - e.g. the corporate bond markets - where you need market makers to see any secondary liquidity, and where most people like the idea of having some secondary liquidity so that teachers can get their pension money, Phil Falcone can cash out investors, etc.
The market maker puts its capital at risk to take the other side of unmatched trades so that later (probably not much later) it can get out of its position at a profit (probably not much profit). A customer comes in with an offer to sell bonds at 101, the market maker buys them, and then it tries to sell at 101.25. If it can't, it holds them and hopes it ends up selling at 101.25 or 102 or 105, not 100.75 or 100 or 95 or zero.
To be good at that job, you have to first know the customers and the order flow - because the focus of your job is to sell the bonds as quickly as possible at 101.25. But second - and also quite important - you really ought to have a view on what will happen to the bonds over time, where "time" may not be "decades" but is probably longer than "the next ten minutes." Because if you are stuck holding the bonds, you'd much rather have them go to 105 than 95. So you evaluate your inventory to try to make sure you have a high chance of it appreciating and a low chance of it depreciating - and, to the extent you're worried about risks you are not really paid to take, you hedge them. So if you like the credit and buy it at 101, you might put on a rates hedge so you don't lose money just because rates go up.
Using incentives to prevent market makers from being risk takers runs the risk of preventing them from being risk managers. If your investment performance cannot influence your comp, then you have no incentive to do anything to maximize it – no incentive to hedge risks, no incentive to make smart choices in buying or declining to buy inventory, no incentive to do anything but maximize unhedged inventory to have a bunch of stuff to sell to clip “commissions.” If you can only get paid for volume, you’ll only maximize volume.
Now on balance, just maximizing volume is probably not the most likely result of the Volcker rule's comp changes. You could construct a model where traders (1) have strict limits on capital at risk, (2) can’t get paid for inventory appreciation, but (3) can get fired for big inventory depreciation. In that case, when offered a bond at 101 that you can’t immediately sell at 101.25, you’ll face the prospect of making some nominal volume commission if the bond goes up, or down only a little, and you sell it, or getting fired if it goes down a lot. That will certainly incentivize conservatism. It will also make you unlikely to make much of a market in anything - and, to the extent you do, you'll be much less sensitive to the quality of your inventory than you would be if you could get paid for it.
But that's all if "you" work at an FDIC insured bank. In that light Mr. Landy’s worries about driving people out of the regulated (for which read "FDIC insured") industry into the unregulated industry (SAC? Citadel? Lazard? Paine Webber?) may be less troubling. If the Volcker rule makes it all but impossible for big deposit-taking banks to operate as effective market-makers, leaving them to a more-or-less agency business, then you’d expect not only hedge funds but also non-FDIC sell-side investment banks to make up the difference and gain a bigger share of market making. Which might not be that terrible. The Volcker rule could be a sort of reinstatement of Glass-Steagall in disguise.
Of course the next leaked version of the draft rules will take back all this stuff about comp, so probably everything’s fine.