During the Flash Crash, most high-frequency traders slowed or stopped quoting. And some took liquidity out of the market by hitting the bids of others. Thus, as a group, they did exactly what the specialists on the New York Stock Exchange were not allowed to do, which violated both their positive and negative obligations -- or would have, if they had had such obligations.
So couldn't regulators solve the Flash Crash problem by just giving high-frequency traders some market maker obligations? If only it were that simple.
The Fate of Specialists
Investors might be surprised to learn that specialists don't formally exist anymore. The name was dropped, along with some obligations. And the NYSE's market share has dropped from around three-quarters to under a third, as Reg NMS has ignited massive competition with the Big Board since 2005.
Given these body blows to its traditional structure and dominance, the NYSE could not stem the Flash Crash, as its specialists of old might have done, or at least would have been expected to try to do. But while the NYSE did not stem the crash, it did sidestep it.
This was no small feat, and a significant vindication of its regulatory stance, which enabled the exchange to retain some manual functions in the face of its competitors' objections and the SEC's clear sympathy with their all-electronic-all-the-time view. But New York stood fast, and the SEC allowed the exchange's liquidity replenishment points, or LRPs, to remain intact. These partially manual auctions permitted the Big Board to momentarily disengage from, and largely avoid, the mayhem on May 6, as evidenced by the fact that it was the only market that did not need to break any trades.
Old World Charm
In the old world, the specialists oversaw the bulk of trading in an NYSE-listed stock and participated in significant portions of it. There was only one specialist in each stock, and "owning" that book was an opportunity no specialist would want to risk losing. Specialists' reputations mattered greatly to them, and their businesses were built on trust.
It was a very personal business that depended on the success of countless interactions with numerous traders who needed to trade a specialist's stock. Success meant dealing fairly and effectively with all of them. Any unfair treatment on the part of the specialists -- be it different treatment of one trader verses another or taking excess profits for themselves, would be noticed and talked about.
Rather, specialists would try to exceed expectations whenever possible, because that would both help to expand their business in a given stock and make them more likely to be selected as a specialist in other stocks. Exceeding expectations meant demonstrating skill at handling complex and difficult situations, and a willingness to risk capital, especially by providing liquidity when stocks were falling.
Another critical dimension of specialists' performance -- and the contractual source of their obligations -- was that they were members of a particular exchange. Their reputations were part of what formed the reputation of the exchange itself and its list of securities. So when the market was falling, specialists were expected to take one for the team and would be rewarded with an enhanced reputation if they did.
Back in those days, the Nasdaq dealer market was the other team. Just as the NYSE's specialists wanted to show that the auction market was better than the dealer market, a Nasdaq dealer wanted to show that the dealer market was better than the auction market. They both had their points. But the key for purposes of this discussion is that market makers on both markets were human beings with names and a need to uphold their personal reputations, the reputations of their firms and the reputations of their markets. The competition between the two markets, with their distinctly different structures, provided a material inducement for market makers on both exchanges to risk capital in a stabilizing fashion.
Enhancing the visibility of their performance was the fact that market maker trades of the sort that would stabilize a market were much larger than the common fare today. They didn't disappear in a blizzard of small, algorithmically distributed trades hitting all the markets randomly so as to avoid being noticed. They were done on the market maker's home exchange, meant to impress, meant to be noticed.
A Brave New World
The high-frequency trading world is different. Trading today is anonymous, so reputation is not a factor. Trading is done in thousands or millions of small quotes and prints that are, by design, not detectable in aggregate or traceable to the trader on the tape or by any other means. Traders spread around orders and trades to all the venues -- exchanges, ECNs, ATSs, dark pools -- with no loyalty to any particular market or structure.
All of the formal exchanges are more or less the same now. The old auction and dealer markets are gone. Differentiating features like LRPs are copied by others and sometimes mandated by regulators in the name of coordination. Take one for the team? What team?
Laying stabilization obligations on today's high-frequency traders -- perhaps sweetening them with exemptions for short selling, as has been suggested -- would not stabilize markets with anything like the force of the old, reputation-based inducements to perform stabilization feats that went far beyond legal requirements.
HFT quotes would have trivial stabilization effects even when active because they are too small to matter. While standing en masse would theoretically work better, high-frequency traders will never do so. The risk to each trader that his fellows will desert him would dictate a cut-and-run algorithmic trigger at the first sign of trouble. Why would a high-frequency trader stand with his fellows when he doesn't know who or where his fellows are, and wouldn't get credit for his efforts anyway, since they don't know who or where he is?
Moreover, quoting obligations would not require HFTs to be active all the time, at least not as some have proposed them. Requiring market makers to provide quotes 90 percent of the time, for example, would leave them free to forego quoting for 10 percent of the day. Ten percent of a 390-minute day is 39 minutes. The Flash Crash took less than 7 minutes.
Obligations will raise barriers to entry for high-frequency firms, thereby cutting off our main hope for stabilization from them: their sheer numbers. The small horde of high-frequency traders today is likely to continue to grow rapidly and become a huge horde, eventually solving the stabilization problem on its own, or at least as much as it is possible to solve it with traders.
But that won't happen if stabilization requirements arrest that growth or, worse, kill off all but the biggest HFT firms. This would put the SEC in the position of picking market makers and dictating their behavior in its National Market System, effectively taking over that role from the New York Stock Exchange and Nasdaq. As much as we might regret having authorized the breakup of the old market structure, with its naturally stabilizing features, authorizing an attempt to reassemble it now will only compound the error.
Steve Wunsch, an inventor of the ISE Stock Exchange and founder of the Arizona Stock Exchange, is a market structure consultant living in New York City.