Exchange monopolies, really?

"Let me put it to you this way, Weisberger. There's a reason people call him Don Grasso... With their ability to regulate, he might make us an offer we can't refuse..." said a senior attorney.

This conversation happened when I was planning to start making markets to the Smith Barney retail flow, bypassing the NYSE. It was going to be based on our new quantitative platform and would offer superior execution prices and service to our clients. While we did, eventually, get it started, the most memorable part was how hard it was to achieve. Even though Smith Barney customers would benefit, it was very hard to convince management to do battle against the NYSE legal monopoly. The reason was the combination of regulatory power and the NYSE chiefs', purported, willingness to use it.

Having experienced this earlier in my career I have a keen interest on the topic of exclusive listings. Most of the reading I do, relate to BATS president, Chris Concannon's, idea for exclusive listing for small cap stocks. Concannon's idea is to concentrate displayed liquidity in one place, but is not a call for a pure monopoly. Chris has been outspoken against the idea of a "trade-at" rule, which would prohibit alternative market models, such as dark pools or auctions, from competing with the single listing exchange. Chris's idea is an interesting concept and could probably be achieved without regulation, but rather, an agreement between the three large exchange groups.

Recently, the WSJ published an editorial that called instead for exclusive listings for all stocks. This is a great example of all that is wrong in market structure commentary. Written by Yale's Jonathan Macey and David Swensen, the piece made a variety of unfounded assertions and distorted conclusions about complexity in the US equity market. The only thing worse than the tortured logic they employed to justify their assertions, is how ignorant and misguided their proposed "solution" would be. Their idea is to grant every exchange a complete monopoly over trading their own listings, but try to create competition for listings itself.

Their editorial contains multiple flawed assumptions:

• First, they assume that competition for listings could work to somehow incentivize innovation in trading.
• Second, they assume that complexity, in and of itself, is bad.
• Third, they assume that a single exchange is the only model for trading that can provide stability and inspire confidence.

All three of these ideas are, simply, wrong.

The idea that the mechanics of trading would be a major driver of where companies would list is absurd. Most corporate financiers have little knowledge of equity market structure, but do understand branding and marketing. The NYSE has a powerful brand with global recognition and NASDAQ has spent decades cultivating a brand among technology and high growth companies. Competing with those two exchange groups is so difficult that BATS, despite excellent technology and higher trading market share, has decided not to pursue corporate issuers. They plan to focus 100% on attracting ETF (exchange traded fund) listings, since those issuers care, exponentially, more about trading characteristics. Even if there was some potential to listing competition, the authors conveniently ignore facts.

• Fact: NYSE owns three listing exchanges and has segmented them so as not to compete with each other for listings.
• Fact: NASDAQ's exchanges do not compete with each other for listings.

Frankly, unless they plan to legislate a "centrally planned" cap on how many stocks an exchange owning group can list, there is no way their plan would create real competition.

The editorial's claim, that competition has massively increased the risks to stable markets from technological complexity, is also very misleading. They claim that fragmentation caused the flash crash, which, as we know, is sheer fantasy. The crash of 1987, which was far worse in both magnitude and duration, occurred in an era of almost no trading competition vs the NYSE. Both crashes were caused by sharp declines in the futures market (a single listing market), but the flash crash rebounded far more quickly. While it is true that algorithms and markets should have had better controls, there is no proof that having multiple exchanges was the cause of the crash. In fact, most complaints were in the opposite direction, namely that the markets were linked together too well. The only real example of dangerous complexity was in the regulations governing trading. In 2010, the NYSE's slowing down potential erroneous price moves differed from the other exchanges. It's true that the inconsistency between the NYSE's "Liquidity Replenishment Points" and other markets did not help the situation, but that has since been fixed. The new rules including the introduction of market-wide "Limit Up / Limit Down" and specific rules that govern the provision of market access, make the recurrence of such an event extremely unlikely.

The WSJ article also presents the unsubstantiated claim that complexity must make the market more vulnerable to disruption. This is amusing, given that any 1st year engineering student knows the key to protecting against process failures is redundancy. Today's web of interconnected markets provides much more redundancy, than a single market model. An outage of several hours at NASDAQ or New York would not even affect the average investor, unless it occurred at the open or the close. This is due to the fact that equities can also trade on exchanges, such as, BATS, as well, as the ability for NASDAQ to provide backup for the NYSE and vice versa. As an aside, the only market-wide outage we have experienced in the last couple of years, related to a failure in the NASAQ SIP, which is a legally granted monopoly provider of market data for all their listings.

The other problem with the Macey /Swensen logic is that it's based on the idea that complexity is, inherently, bad. I suppose that's because a human trader would not be able to manually "navigate 13 exchanges, more than 40 dark pools and a handful of "Electronic Communication Networks," that these two criticize. The flaw in that particular reasoning, is in the fact that humans don't have to. As I have previously written, virtually every trading system has access to smart order routing (SOR) technology, which effectively connects traders to all potential trading venues. While I have consistently made the point that routing deserves greater analysis and attention, it is not the technology details that matter. What matters is the ability to measure both routing and the characteristics of trading venues, in their proper context. Based on the extensive analysis I have done, there is a value to each type of venue, depending on the situation. Unfortunately, this particular editorial ignores this, in favor of ad hominem attacks against those venues.

Mr. Macey and Swensen's duplicity goes further when they describe dark pools as sinister and claim that fragmentation necessarily impedes market quality. I won't reprise my discussions of dark pools, but will note that well known facts point out that they are simply wrong. First, transaction costs, as reported by major brokerage houses such as ITG, show that execution costs are dramatically lower in the United States than all other major markets, many of whom have limited to no competition. In addition, the United States stock market trades a much higher percentage of market cap than all other major markets. The result, is that it is as much as 50% cheaper to trade large-cap equities in the United States versus their peers in the rest of the developed markets. The major driver of that difference, in my opinion, has been competition and innovation in market structure. Consider the innovation of so called "inverse venues" in the US. As I pointed out previously, trading costs in stocks such as SIRI would be 47% higher if such venues were not allowed to operate as competition to primary listing exchanges. Last, consider a very simple comparison - futures vs stocks.

Futures markets are single listing and trading markets in precisely the same manner as the WSJ article wants to make mandatory. The S&P E Mini trades on only one exchange. The S&P 500 ETF (SPY), however, trades on many. It is one of the most actively traded stocks on most exchanges. This is particularly instructive because the cost of trading the S&P future is much higher than the S&P 500 ETF (SPY) for orders up to $5 million in value. As recently as 20 years ago, all hedging by proprietary and statistical trading desks was accomplished by S&P futures. Today, most intra-day hedging is done by the SPY, due to the readily available liquidity at a lower cost. In essence, the market has voted and the competitive model wins. [Note - I am not suggesting that there is anything wrong with the E-mini or the futures market in general. It is well suited for trading very large positions and accommodating investors that need leverage. It simply is not as low cost as the equity market for the average investor]

Macey and Swensen have also disregarded recent history. In the 1990s, the proliferation of ECNs such as Island and ARCA drove NASDAQ towards automation and tighter spreads. In the last decade, the entrance of BATS into the market also helped to drive down trading costs by providing competition to the NYSE and NASDAQ. The editorial makes the claim that the NYSE loss of market share is evidence of a problem, when, in reality, the exact opposite is true. Once Regulation NMS ushered in competition to the NYSE, the volumes at the big board did not fall. Rather, the volumes in NYSE listed stocks increased in the aggregate with the new volumes occurring in competing exchanges and ECNs. This directly lowered the cost of trading listed stocks, which is a major benefit to investors.

To be clear, I am not saying that the market is perfect. There are certainly pitfalls in the market and wide disparities between the effectiveness of different strategies and brokers. Knowledge of the plusses and minuses of various routing strategies and venues in the context of particular trading situations is very important. Utilizing this knowledge as part of a quantitative and disciplined approach to trading and trade measurement can lower costs and improve asset management performance. That said, it is extremely rare to see a firm that has invested in building sophisticated technology for clients make claims that the market is "broken" or "rigged".

The bottom line about this article and all of the commentary that supports simpler, less competitive market models, is that they are wrong. Sometimes the author's motivation is to sell newspapers or books, sometimes to prove their own relevance to their bosses and others to create a climate of fear to help them sell their services. In all of these cases, they do the investing public a major disservice. As I wrote in my first commentary, the "good old days" were not necessarily that good and, despite the fear of the unknown, technology and competition generally lead to progress.

*50% is estimated by using a square root function of the ADV per $ of market cap differential between the US and the rest of the G7 in concert with the calculated costs in published reports.

David Weisberger, Managing Director, Trading Services at Markit
Posted 24 June 2015

S&P Global provides industry-leading data, software and technology platforms and managed services to tackle some of the most difficult challenges in financial markets. We help our customers better understand complicated markets, reduce risk, operate more efficiently and comply with financial regulation.

This article was published by S&P Global Market Intelligence and not by S&P Global Ratings, which is a separately managed division of S&P Global.