Complicate, then prevaricate

THE brokers who traded shares in the Tontine coffee house in 18th-century New York often resorted to stronger drink, leaving them “a little addled”, according to one contemporary account. The technology involved in share-trading has changed a bit since then, and at least some of the participants have sobered up. But more than 200 years later, investors in American equities still wonder whether they are really receiving decent service.

On the face of things, they have little to complain about. The cost of trading has declined sharply over the years (see chart). Explicit commissions, which were once levied in percentage points (0.25% in 1792), have largely disappeared. This is thanks mainly to competition. Whereas the New York Stock Exchange (NYSE) dominated the trading of shares listed on it for most of the 20th century, there are now lots of places where they can be bought and sold.

The impetus for the fragmentation was “Regulation NMS”, adopted in 2005 by the Securities and Exchange Commission (SEC), Wall Street’s main regulator. This required share-trading orders to be funnelled to the exchange offering the best price. The intention was to boost competition to NYSE and NASDAQ, which had a near-duopoly in share-trading at the time. It succeeded in that: both now have less than a fifth of the market. (In response, firms running exchanges have branched into other markets—see next article.)

American shares are traded on a dozen exchanges; at least six other exchanges cater to investors in derivatives linked to shares. Shares also change hands on another 40 or so “alternative trading systems”, as well as a number of “single-dealer platforms”. Finally, many trades are now “internalised” by big banks and asset-managers, meaning that they pair up buyers and sellers within their sprawling empires rather than use an outside trading venue.

Yet investors worry that, in many cases, competition has brought down the visible price of trading by adding hidden costs. Two anxieties stand out. One is the worry that the current set-up of the markets allows high-speed traders to anticipate big orders and “front-run” them, moving prices in an unfavourable direction before an order can be executed. The other is the question of how robust the system is, with regulators still unable fully to explain events like the “flash crash” of 2010, when the Dow Jones Industrial Average plunged by 9% in minutes before rebounding.

Start with fears of front-running. Many institutional investors complain that ultra-fast traders spot big orders entering the market, and race ahead of them to adjust their prices accordingly. Attempts to hide from the speedsters can go awry. In January Credit Suisse and Barclays, two big banks, agreed to pay $154m in fines for misleading clients about the workings of their “dark pools”, where offers to sell and bids to buy are not published. In theory, that protects investors from front-running; in practice, several of the firms running such venues had concealed the central role that high-frequency traders played on them. (Credit Suisse didn’t admit or deny wrongdoing in the settlement.)

There is another, less-often-told side to the story. Speed is necessary to knit together a dispersed set of exchanges, so that investors are immediately routed towards the best price available and so that their orders are the first to get filled. And plenty of high-frequency traders are market-makers; it is their job to adjust prices in response to new information. Nonetheless, the idea that markets are rigged is widespread, not least thanks to the publication of “Flash Boys”, a book by Michael Lewis on the evils of high-speed trading.

One proferred solution is to level the field by slowing things down deliberately. IEX, whose founder is the hero of Mr Lewis’s book, is a trading platform that has applied to the SEC to become an exchange. It uses miles of coiled cable to create a “speed bump” that delays trades to the advantage of institutional investors. The SEC has received more than 400 letters in support of its application, but there is a vigorous debate about whether IEX’s system complies with the requirements of Regulation NMS. Some think that the better solution would be to get rid of Rule 611, which in effect requires orders to be sent to the exchange showing the best price, even though such quotes can sometimes be unobtainable in practice. The SEC will vote on IEX’s application by March 21st.

Front-running is not the only concern about America’s market structure. The other is the risk of sudden spasms like the flash crash. Glitches are common. In 2012 two public offerings, for Facebook and BATS (“Better Alternative Trading System”, a firm that runs exchanges and other trading venues, ironically enough), suffered disruption. Later that year faulty software toppled Knight Capital Group, a big trading firm, by vomiting orders to exchanges without tracking those that were filled. In 2013 the primary electronic market and the back-up system both failed at NASDAQ thanks to a software bug, and so on.

Andrew Lo, a professor of finance at the Massachusetts Institute of Technology (MIT), argues that investigations into such events tend to focus on the venue most affected. How they reverberate through the broader system is very little studied or understood. Sometimes, the existence of other venues may help: in July the NYSE briefly went offline and traders barely noticed as other exchanges filled the gap. On other occasions, they may amplify volatility. In August lurches in the future and equities markets caused the value of exchange-traded funds to deviate from the value of the underlying shares they owned.

Mr Lo proposes a simple reform: the creation of a commission that can subpoena witnesses and evidence to look into the causes of crashes, just as the National Transportation Safety Board investigates air disasters. The commission would look as widely as it liked at what went wrong and then publish its findings.

The SEC acknowledges that the rules governing share-trading need amending. Mary Jo White, the chair of the SEC, has mused, for example, about monitoring the controls firms use to prevent their algorithms running amok. Another idea is to provide the SEC with the power to curtail otherwise legal trading when the market is convulsing. The risk is that in addressing market complexity, the regulators only add to it. A single SEC proposal, on a facet of a facet of the overall system, is now up for public comment. It runs to 581 pages.