Is slower better? The debate around IEX's application to become a public exchange seems to have coalesced around this simple question. If you believe automation and high frequency trading (HFT) are bad because they are fast, then it may seem obvious to slow everyone down. IEX designed its proposed exchange to include an intentional 350 microsecond delay for incoming and outgoing information, which sounds sensible enough to the investors feeling outrun.
There are of course great arguments to be made that fast isn't bad – plenty of academic research, including my own, shows that automation and HFT benefit all market participants with improved liquidity and reduced trading costs. Nevertheless, IEX has marketed its value to investors and the SEC as a safe haven from automation and HFT practices.
And while IEX benefits greatly from the view that its application fight is a story of the fast versus the slow, the merits of automation and HFT are actually largely unrelated to the current debate over its application.
IEX's design isn't problematic by itself, which is indicated by its success as a dark pool. In fact, many who oppose IEX's application to become a public exchange are avid users of its dark pool. That's because IEX's fundamental issues arise specifically from incompatibility with Regulation NMS (Reg NMS), a regulatory framework that largely doesn't affect dark pools.
Reg NMS includes a broad set of rules that govern equity market structure, but one rule, 611, is important to understanding concerns over IEX's application. Rule 611 provides protected quote status for registered public exchanges, and requires orders to be routed to the public exchange displaying the best price. This is an important reason why IEX wants to become a public exchange.
An SEC seal of approval therefore affords IEX entry into a marketplace where business is essentially guaranteed, and it's probably not farfetched to assume that's at least in part what IEX investors are after. But all else being equal, that's not an illegitimate reason to pursue exchange status. If IEX is willing to forgo having protected quote status, by being designated as a ‘slow’ market, I expect opposition to IEX’s exchange application would disappear. The question really is, why is it such a big problem for IEX to have protected quote status?
First, IEX allows specific order types to bypass its speed bump, incentivizing certain liquidity providers to use IEX over other exchanges. These special order types, called hidden-pegged orders, are pegged to a particular price point such as the best bid or ask in the market. When the best bid or ask changes elsewhere in the market, IEX uses that information to update pegged orders on its own exchange. Importantly, that information skips around the artificial delay, while an investor updating a limit order would face an additional 350 microseconds. This structure indisputably creates two tiers of investors.
Secondly, affording IEX protected quote status with an intentional delay creates concerns for stability and fairness across exchanges. There is good reason to believe adding intentional latency, even 350 microseconds, will only increase uncertainty about market prices. Equity markets have changed dramatically in the last ten years, where there are now millions of messages being sent every second and it's impossible to dismiss the significance of any kind of latency. And brokers will be forced to route orders to IEX, despite this uncertainty.
But you don't have to believe one exchange’s 350 microseconds is a serious latency issue to question its legitimacy under Reg NMS. If you believe that 350 microseconds is advantageous to some investors, as IEX does, then you expect other exchanges will be forced to adopt their own latency to compete. This may lead to a race to the bottom with other exchanges offering latency of potentially arbitrary length to gain market share. If all exchanges introduce latency, how can any investor know which exchange currently offers the best prices to buy and sell a stock? We can't ignore these kinds of structural changes without careful analysis on the impact on the market and investors.
This all comes down to the fact that, under Reg NMS, the exchange marketplace is not a free market. In an ideal world, exchanges could design themselves in any fashion and competition would determine who succeeds and fails. But the world we live in is one where exchanges agree to a set of standards for regulatory-granted access to order flow. And IEX's current exchange proposal does not meet those standards.
I sympathize with many IEX supporters who want to see more innovation and competition in the exchange marketplace – I do too. But before we can do that, we have to reexamine Reg NMS.
Hendershott is a professor at UC-Berkeley's Haas School of Business.