The last one percent

Fort Pitt Capital Diamond

Click here for the Second Quarter 2014 Newsletter

With the recent publication of Michael Lewis’ book “Flash Boys,” a bright light has landed on the murky world of High Frequency Trading (HFT). In the weeks since Lewis appeared on “60 Minutes” to promote his book and announce that the U.S. stock market is “rigged,” numerous charges and countercharges have been lobbed between stock exchanges, investment banks and the HFT community. The New York Attorney General, the FBI and the U.S. Securities and Exchange Commission (SEC) have all launched investigations. With the HFT crowd now scattered to the corners in hopes of avoiding the regulatory broom, we thought we’d offer our two cents on the matter in the form of a Q&A.

What is High Frequency Trading (HFT)?

HFT is automated trading, and it happens at the speed of the world’s fastest computers, faster than humans can imagine. Up until 40 years ago, stock trades required human intervention. The people charged with acting as final intermediaries between buyers and sellers were called specialists. Specialists stood at their posts at the New York or American stock exchanges and facilitated trading. It was also their job to smooth out turbulent markets by taking the other side of a wave of buy or sell orders. Imagine a business where you were required to buy low and sell high, day after day. How did one become a specialist? Generally, your father, your grandfather (or some other very close relative!) was a specialist. Broadly speaking, automated trading is the mechanism which replaced the specialists at the core of our markets.

How did automated trading start?

In 1975, coincident with the demise of fixed trading commissions, Congress authorized the SEC to facilitate a new national trading system. U.S. stock exchanges began to experiment with replacing humans with computers for small trades. Trading volumes were rising, and humans were sometimes having trouble keeping up. The DOT system (short for “direct order turnaround”) was implemented to increase efficiency by routing orders directly to a specialist’s computer, where they could be executed automatically, rather than through a broker on the trading floor. The DOT system was the beginning of the end for human traders. It marked the first step toward disintermediation of all parties other than the ultimate buyers and sellers—and pointed the way toward fully automated trading. It reduced trading costs immensely, and in coming years it allowed visionaries such as Charles Schwab to slash commissions for the general public. By the early 1990s, for example, the cost to trade $100K worth of shares had fallen by 90%, from $800 to $80.

How did it come to dominate markets?

The DOT system was very speedy and efficient, and worked well for smaller orders. Over the years, as computers and communication became cheaper and more powerful, and trading volumes grew, it made sense to automate a much greater portion of trading. So much so that by 2005 the protocols for an entirely new national system of automated trading, one which would finally relegate the specialist to the ash heap of history, were being hammered out between existing stock exchanges, clearing firms, investment banks and regulators. Regulation NMS (short for National Market System) was the name given to these protocols. Implemented in 2007, NMS was designed to promote efficient and fair price formation across markets. The ultimate goal was a real-time auction system, with nothing more than a computer and a phone line connecting buyer and seller. As a buyer, I posted a price and the number of shares I wished to transact. As a seller, you did the same. Where our desires matched, a trade happened. What could be more efficient, transparent and elegant, right?

How did NMS work out?

For the average investor, it worked fantastically well. Like the DOT system 30 years earlier, the National Market System drove trading costs into the ground. Retail trades costing $80 pre-NMS were now tagged at $8, another 90% reduction. Institutional investors paid less than a penny per share. Technology had cut the cost of trading for retail investors by 99% (from $800 to $8) in a generation! Trading volumes rocketed upward. Part of the increase was basic economics—trading was cheaper so people demanded more of it. But another factor was a bit more sinister. It turns out that quirks in the system gave a group of software jockeys the ability to make gobs of money by skimming tiny amounts from each of millions and millions of trades. These were the High Frequency Traders—the Flash Boys—of Michael Lewis’ book.

So the HFT crowd DOES have an unfair advantage!

Our comments on this score are mixed. We encourage you to read both Lewis’ book and the responses from various quarters of the HFT community in order to make up your own mind on the issue. In our view, yes, the stock market is rigged, but only to the extent that it represents a failure to capture the last 1% of cost reductions for investors.

The last 1% is encapsulated in a set of predatory behaviors in the automated stock market. They include various forms of quote-stuffing, which involves promulgating phantom quotes to smoke out other trader’s intentions, with no real intention of transacting. They include using this newly gleaned information to electronically “front-run” other’s trades. They include practices known as rebate- and slow-market arbitrage. Some are the result of fairly obvious flaws in the system, and the most egregious will be undoubtedly be fixed over time. Regulators, exchanges, investors and traders built the system, and they can rebuild it to make it better.

Other problems in the system are more structural in nature. How, for example, are the servers at the center of the various exchanges any less “advantaged” than the human specialists of the past? HFTs have shown that access to and control of the data within these computers, like being born to a specialist, is a license to print money. Who will own these computers? The government? A non-profit? Somebody (or something) has to “be the market.” Who gets the data first? Time is money!

But don’t miss the forest for the trees. Don’t let these (comparatively minor) failures eclipse the quantum advances in market structure which have benefitted millions of investors. The system is far better today than 40 years ago. Focus on the benefits; the problems can and will be fixed. Let’s thank Mr. Lewis for pointing them out, and get going on the next 40 years.