Streetwise Professor

In their never ending search for bad ideas to embrace, regulators are contemplating imposing participation obligations on market makers:

Proprietary trading firms are concerned that last week’s market tumult could see regulators seek to attach more strings to their trading activity, requiring them to buy and sell shares even when they’d rather not.

Securities and Exchange Commission Chairman Mary Schapiro said Tuesday that regulators are weighing obligations that should apply to such firms, after some “professional liquidity providers” stopped trading amid last Thursday’s heated volatility that saw U.S. markets quickly plunge and then rebound.

The command-and-control mindset is running amok in Washington.

How about some basic economics, OK?  Putting aside how you would even define and enforce such obligations, imposing an obligation that firms make a market when they would rather not voluntarily, or to make a market in a size that is larger than they would choose voluntarily, imposes costs on said firms.  This reduces return on capital.  Since capital must earn a normal return, the mandate-caused reduction in return will lead to an exit of some capital from market making.  This will reduce liquidity in periods in which the constraint to make markets is not binding.  Spreads will be higher, and depth lower, during these times.  This will generate higher profits then that will offset the losses market makers must bear when the constraint binds.  (There could be a perverse dynamic here too.  Less liquidity in normal times increases the likelihood (all else equal) of a big price move that may trigger the constraint.)

One can also foresee the potential of a return to the specialist bargain of the old days.  Specialists had an affirmative obligation to make markets, even when it is unprofitable.  They were compensated in the form of entry barriers and information advantages that allowed them to earn rents in times when the obligation constraint was not binding.  But the effect of this was quite similar to what I just sketched out: trading costs are lower when the constraint binds, trading costs are higher when it doesn’t.

In three words: No Free Lunch.

Futures markets almost never impose such obligations.  There are some exceptions, in which there are designated market maker programs, usually implemented to promote trading in new products.

In essence, market making obligations alter the time pattern of liquidity.  Liquidity is higher in some periods, lower in others.  Overall, liquidity is lower, because the constraint necessarily reduces the capital devoted to making markets.

Does Mary Shapiro, or anybody else, know what the “right” pattern of liquidity supply should be?  Do they have any idea of the effects of such a constraint.  Almost certainly not.

Could competition between exchanges ensure the provision of the right amount/pattern?  There’s room for skepticism.  Assume exchange E imposes the constraint, but exchanges X,Y, and Z don’t.  During normal times X etc. will get more business because market makers on those exchanges incur lower costs and don’t have to charge the premium that E’s market makers do to cover the costs of making markets during volatile times.  During the volatile times, E’s market makers will attract business–just when they don’t really want it.  So exchanges could free ride on any other exchange that imposes the obligation.

Is there an externality or other “market failure” that could justify imposing the obligation on all market makers and exchanges?  The Greenwald-Stein story suggests that yes, that is a possibility.  But the mechanism to deal with that is to switch from continuous trading to a call auction market.  With multiple exchanges, that switch must be coordinated across exchanges.  The lack of such coordination exacerbated problems on 6 May.  Before imposing additional obligations that will harm market quality during normal periods, it would be better to improve intermarket coordination.

Moreover, as the CME experience shows, circuit breakers that focus on stop orders is a reasonable, relatively light-touch functionality that can mitigate some of the most serious problems.  Why not try that before imposing difficult to define, hard to enforce, and costly obligations on market makers?