PrintPrint CiteCite
Style: MLAAPAChicago Close


Silly Thoughts on SuperMontage, Market Structure

Author: Benn Steil, Senior Fellow and Director of International Economics
June 17, 2002
Securities Industry News


Me, I'm just an academic. I just write equations and crunch numbers. You know, the ones you guys spew out in all that soesing and softing and stopping and spoofing and stuff that you do.

Now, I'm well aware that "Our Markets Are the Envy of the World". You guys chant it at every industry conference. Still, sometimes I think (that's what they pay me to do) that some of this stuff is not, strictly speaking, optimizing my retirement portfolio.

Blasphemous, I know. But sort of like John Nash, my mental life is a constant struggle between what I know to be real and good and true and those "Voices." Oh, those Voices! Why, you know that sometimes I can't even distinguish between price improvement and front-running? Honestly! Of course, I know that price improvement is good, 'cause it's good to have the price improved. I mean, duh! Hel-LO?

I know as well that the floor broker penny-jumps for the good of "investors." Chairman Grasso hath spake it.

And I know that the specialist is simultaneously looking out for the best interests of that 25,000 share buy order at the top of his book, even if the price runs away from it and it never executes, and the institution that placed it yells bad words, and stops sending down those orders, thereby widening the NBBO and reducing market depth. It's all good.

Because if it weren't good, tell me now: how could Labranche earn a 3.6% realization rate during the worst bear market in ages? I mean, that's gotta be a Just Reward for all that capital they're committing towards protecting my retirement, right?

And if part of that 25,000 share buy order was destined for my 401k, well, then, why didn't my idiot fund manager rent a floor broker to hide it, and front-run (sorry! "price-improve") some other schmo's 25,000 share limit order?

Now see, I'm getting myself all confused. I was talking about the specialist, and how good it is that we have him standing between buy and sell orders in the most liquid stocks in the world, while those idiot Europeans cut corners and let a computer do the matching, resulting in implicit trading costs 1/4 of US levels (silly equations!). Because if the specialist weren't doing good to earn a 3.6% realization rate while the markets sink, that would imply something crazy, like that he's some sorta monopolist.

And that maybe he doesn't believe in demutualization because he thinks the exchange's new owners wouldn't fully appreciate his public service-mindedness.

Oh, those damned VOICES!

But my incomprehension is not limited to the listed market. It extends to OTC. In fact, sometimes I can't even understand the difference between "listed" and "OTC." Like, why is it immoral to trade through the NBBO for listed stocks, and not OTC? I say immoral rather than illegal because if it were illegal then the SEC would probably try to stop it. Maybe the Catholic church is handling it. So it's under control, then. I don't know.

Or why does the SEC say that my pension fund can buy Microsoft direct on Arca, but that it must pay a broker to buy GM on ArcaEx? I guess that broker must be doing all sorts of investor-protection type stuff on GM that I don't need on Microsoft, or else the rules would be costly and stupid and the SEC would change them. As I said, I don't understand a lot of stuff.

Like SuperMontage. Now I know that something called "Super" has got to be an improvement on something called "Level II," but I still hanker for the bad old days, before the SEC went zero tolerance on dealer markets. Life was simple then. Dealers were dealers (and slow ones were soesed). ECNs were Instinet. The world was working like my models (what a great world!). Dealers make markets in illiquid stocks, the stocks become liquid, automaters step in, dealer margins fall, automation takes over. Been there, done that in London, now the show comes to Wall Street. Just let competition work its wonders . . .

Now, the SEC believes in competition (I've read all their concept releases). It's just not the sort of competition described in economics texts. It's more like "managed competition." Sorta like Hillary's health care plan.

But ever notice that whenever the SEC manages competition, to keep our markets "integrated," the markets come out looking more fragmented?

Ever wonder why the number of US stock exchanges was over a hundred a century ago, fell relentlessly to seven by the 1970s, and only stopped dropping after the SEC gave us the ITS-CQS-CTA?

(I didn't think you did, but this is the sort of stuff that keeps me up at night.)

And who actually believes that the regionals, kept on life support by the ITS, pose more effective competition than Island, which refuses to participate? Or that the CQS and CTA give us anything that Reuters and Bloomberg don't provide in Europe, which somehow staggers through the internet age without an all-knowing SEC?

But I digress.

You see, it seems that the SEC can hardly go a decade without some fundamental and arbitrary redistribution of property rights. This is invariably called "market integration," which was bestowed upon the OTC markets in 1997 by way of the order-handling rules. Here's how they work . . .

First, no more dealer markets. Dealers would henceforth become plain vanilla brokers, required dutifully to fork over their client limit orders to the Level II Montage.

Second, Instinet-like things, christened ECNs, would share in the joys of integration by forcibly donating their best quotes to Nasdaq as well. Bad news for Instinet - one of the most liquid order books in the world at the time, with over 20% of Nasdaq volume and 99% of ECN volume - but great news for regulatory arbitrageurs.

Why? Order books need lots of orders to make money, and folks don't like to give them orders unless they already have a few. A problem for ECNs that weren't Instinet. Unless, of course, the SEC tells Nasdaq to shop ECN orders, and lets ECNs charge a second commission (or, in polite parlance, an "access fee") to any broker unfortunate enough to have a best execution requirement to hit it.

Et voilà, almost overnight there were 10 ECNs. Or, more accurately, brokerage systems registered as ECNs.

Be a broker, get one commission. Be an ECN, get two. Which would you choose? (Hint: no calculus required.)

No matter, says the SEC, because investors benefited. Nasdaq quoted spreads declined by 40% - testimony to the success of its order-handling rules.

Well, that's their storyline. But how many of you out there believe that dealers compete on the basis of quoted spreads? Hands up.

No hands? Well, the SEC believes it.

In nationalizing Instinet's order flow, all the SEC did was to redistribute it to competitors. Comparing Instinet's top of book on their 10 most liquid stocks in December 1996 (pre-order handling rules) with the best bid and offer in the entire Nasdaq market in March 1997 (post-order handling rules), I found that Instinet's spread was, on average, better on half those stocks. Thus, these spreads did not decline - the SEC was just ignoring Instinet pre-'97, because it wasn't legally "the market." But the resulting fragmentation imposed a real cost on institutions, as new systems and traders were now needed to re-aggregate liquidity. No wonder that institutional direct trading via ECNs actually declined from 24% in 1996 to only 15% in 1998.

Having fragmented the Nasdaq market, the SEC decided to blame it on "technology" and collude with Nadaq to re-integrate it. And so we arrive at SuperMontage, the ultimate testimony to the tyranny of compromise. No one is stupid enough to like it, yet no one is powerful enough to oppose it.

For brokers, no need to post a better price to get an execution. SuperMontage will be the only electronic order book in the Milky Way with no price-time priority.

For ECNs, no need to give up their two-commission order books. SuperMontage will be the only order book in the world with little order books inside it. Sorta like those Russian Matrioshka dolls. You know, take off Stalin's head, Lenin appears.

The ultimate irony is that whereas ECNs were the only ones offering guaranteed instant executions on the Level II Montage, they're the only ones guaranteeing delayed or failed executions on SuperMontage. They have 30 second time-outs to decide whether to accept your hit. And it may be a Pyrrhic victory if they do, 'cause they'll whack you with an access fee. Refuse to pay, and the ECN can refuse you as a customer. And you can't trade through them.

But, hey, it's a free market, right? Well, not exactly. The SEC has ordered the NASD to provide a single obligatory alternative for those who can't quite stomach SuperMontage: the charmingly named "Alternative Display Facility" (ADF).

Most people go through their lives without ever being conscious of having an appendix. Or, if they do become conscious of it, they have it removed. Yet the SEC, in its infinite wisdom, has mandated the installation of an appendix into our markets. The ADF.

Does anyone - outside of the SEC or comparably secluded institution - actually believe that the NASD ought to spend millions of its members' dollars building a 1970s style quote display system? For what, "transparency"? Can anyone name a single European exchange stupid enough to refuse to sell real-time quotes and prices to a major market data vendor? The notion that an ECN operating outside of SuperMontage would be a "closed" system is about as quaint as the notion that the NYSE, with its multi-million dollar "seats" and technology designed to prevent investor order interaction, is an "open" system.

If you're still reading at this point, the odds are a lot higher than they were at the outset that you're a buy-side trader. But since all the Ivy League traders went to the sell-side, I can't be too flattered by your love. So I'll pick on you guys too.

My number crunching shows that no matter how you slice it - hard trades, soft trades, red trades, blue trades - brokers, on average, don't add value. Savings on listed and OTC trading costs, including market impact, are roughly a third on execution-only electronic platforms. But you guys are still handing the bulk of my retirement trades to brokers in exchange for soft goodies. Irony of ironies: you will even send my retirement money to brokers to pay for third-party trading cost studies that tell you that brokers cost too much.

And if the Knicks land a 7'5 Chinese guy, it will only get worse next year.

Look, we economists clearly have a better reputation for quantifying inefficiency than for eliminating it. You know, ask an economist to raise farm productivity, and he'll model you a spherical cow. But in that fine tradition of deriving optimal bovines, let me just rattle off a few easy fixes for the SEC:

- SuperMontage. No, don't ask Nasdaq to submit a tenth amendment. Just tell them to sever the umbilical cord to the NASD, and then set them free to fix the stupid stuff. No concept releases, no comment periods. The first thing Nasdaq will do is make all orders instantly accessible and free of charges levied by the order supplier. Sound unfair to ECNs? Ask yourself: how many ECNs allow their subscribers to put "maybe orders" on their books, or to charge "access fees" for hitting them? The second thing they'll do is flick the on switch for price-time priority. A commercial Nasdaq won't cave in to broker tantrums, since price-time priority means more aggressive limit orders, more executions, and more money for Nasdaq.

- ITS. Subject it to an antitrust investigation, and watch it disappear. With the regionals gone, we can get down to the real event: The NYSE vs. Modernity.

- Softing. Tell the buy-siders to pay their own trading commissions. This will raise national savings and free up seats at the Garden.

Just some silly thoughts.

The following was written exclusively for Securities Industry News by Benn Steil, who is the Andre Meyer senior fellow in international economics at the Council on Foreign Relations.