When Did HFT Start?
We initially spotted HFT at work early in our careers, when we were sales traders at Institutional Network, otherwise known as Instinet, the world’s first electronic brokerage firm. Like the NYSE, Instinet was an order-driven market, but it was anonymous, with no specialists to facilitate order flow. In the 1990s, Instinet captured a huge market share in the block trading of NASDAQ stocks. Institutions and fund managers loved the ability to trade in between NASDAQ market maker spreads, which, at the time, were averaging more than 32 cents. Instinet filled a real need. Institutional fund managers wanted to trade their orders with more control and freedom. They wanted to trade their orders themselves rather than through a broker, and they wanted to do so at reduced costs. Each day, they routinely negotiated and traded blocks ranging from 25,000 to 1,000,000 shares. Each year, their number of trades increased.
Then Instinet began courting a new type of trader. These traders promised large volume, albeit in small size trades of a few hundred shares at a clip. In exchange, these traders wanted insanely low commission rates and access to Instinet’s top-of-book and depth-of-book data. These traders wanted to feed their computer models all Instinet’s order flow information, including best bids and offers, their sizes, as well as information about the other orders in each stock’s limit book. It was a win for Instinet in terms of increased volume. Instinet could pitch that volume as an attribute in attracting new buy-side institutions that would pay higher commission rates.
Some might say it was wise for Instinet to court these automated trading firms. However, the flow these firms pumped into Instinet was predatory. We had never seen anything like it before. The institutional clients complained immediately. If an institutional client placed a bid to buy 5,000 shares at 24 1/8, the automated trading firms instantly placed a bid for a few hundred shares at 1/64 higher (the minimal allowable increment). They did this with every stock. If the institutional clients canceled their bids, the automated traders instantly canceled their higher bids as well.
Allowing these automated traders into the network ran counter to the philosophy Instinet had pioneered since its inception, which was helping institutions trade blocks electronically. It also felt plain wrong. These automated trading firms were nothing more than high-speed scalpers and the first high frequency traders. Instinet enabled them for fear that its order flow would go to the newer upstart electronic crossing networks (ECNs). Eventually, this type of order flow spread to all the other electronic venues in addition to Instinet.
The SEC became concerned, not about the scalping, but that the retail public was seeing one set of prices, while other market players saw better prices on the ECNs. In response, the SEC proposed Regulation ATS (Alternative Trading System), which mandated all orders go to a public quote. Instinet fiercely lobbied against it. Management had employees calling clients and urging them to write comment letters to the SEC. Instinet claimed that such a rule would create a commoditized limit order book, which would “crush innovation.” In reality, Instinet was afraid of losing the special selling appeal of its own private stock market, with meaty institutional orders that the SEC, if they had their way, would forcibly make available to all traders through all systems. No doubt, management was also worried that the automated traders might not pay for the privilege of jumping 1/64 ahead of investor orders if the automated traders could do it to the entire market.
The SEC had its way. Reg ATS passed. Spreads between bids and offers tightened due to the competition. Instinet had to compete by having the fastest platform, and it did well. So, too, did the automated traders. They had a much larger body of water to swim in—and hunt. HFT exploded in volume as more firms entered the space. It would eventually peak in late 2009 with an estimated 70% of all volume in the market.
As HFT grew, so did the structure risks. The market is like an ocean. To the extent that there are many different trading styles and participants interacting with each other, the market is healthy, like a body of water teeming with many species. One player may have a subsecond time horizon, another a minute, another ten years. They might be value investors, chart readers, or earnings momentum players. However, when one participant accounts for so much volume and has eclipsed so many other participants, and its trading styles and horizons prevail, the ecosystem is in disequilibrium. One of its more predatory species, such as a shark, has become overwhelmingly dominant. And it is unsustainable.