Featured Commentary - HFT Regulation - Michael Aitken, October 2012
|Occupation||Chief Scientist; Professor|
|Employer||CMCRC; University of New South Wales|
Fair and Efficient: A Q&A with Michael Aitken
Michael Aitken is the founder of SMARTS Group, maker of a real-time market surveillance system now in use at over 40 exchanges and regulators worldwide. He is also the chief scientist at the Capital Markets Cooperative Research Center (CMCRC), as well as a professor and leading researcher at the University of New South Wales. Aitken spoke with John Lothian News Editor-at-Large Doug Ashburn about the role of high-frequency trading (HFT), the concepts of fairness and efficiency, and what exchanges and regulators must do to achieve the proper balance.
Q: How would you describe the regulatory climate with regard to trading practices?
A: The surprisingly consistent mandate of every regulator in the world is to ensure that markets are fair and efficient. They are not permitted to authorize changes to markets unless they enhance the fairness and efficiency of markets. The question that I have been addressing is how they do this operationally, as a practical matter. How do they know that authorizing fragmented markets, high-frequency trading and dark pools actually enhances the fairness and efficiency of the market?
If HFT has been authorized already by regulators, as evidenced by the fact that it is there, then surely they should have already made the decision that it passes the test of fairness and efficiency. But here is the key point: they have neither defined HFT, nor have they operationalized the mandate. So it is not surprising that we are in a position where they have no idea whether HFT is good or bad for a marketplace.
Q: So what does “fair and efficient” mean? It certainly covers a lot of ground, does it not?
A: Academics have given us a pretty reasonable understanding of efficiency. When academics look at HFT they will say that it reduces transaction costs and enhances price discovery. Those are the two proxies for efficiency, and academics are saying this all the time.
But what academics do not help with is the fairness part of the “fair and efficient” mandate. Many of the latest stories about high-frequency trading is that it is unfair. So where does this leave us, since regulators have never actually sat down and decided what “fair” means?
Q: So how would you define the issues such that the regulators can address fairness?
A: One issue is that, in a fragmented market, where you are not sure where you are going to get your liquidity from, you may decide to put your order into two or three marketplaces. As soon as you get hit in one of them, you start withdrawing your orders from the others. So one potential issue is that, for the players who are playing in those other two markets, and only those other two markets, the HFT liquidity can seem transitory.
Q: So you are witnessing what we call “the illusion of liquidity?
A: We are doing some work on this now, called “fleeting liquidity.” Say there are three or four main markets, and they have to be positioned in all of them. We are now doing some experiments which look at, when they get hit in one market, do they suddenly withdraw their orders from the other markets? This would be consistent with what the buyside is saying - they see the liquidity and, when they try to get it, it isn’t there. But this is just a manifestation of the fact that we say competition is good, so we fragmented the market.
Q: Do you see any merit to any of the other ideas floated about the industry, such as the 500 millisecond rule that was put into the latest draft of MiFID?
A: There are two sides to the coin, and both sides have unintended consequences. First, if you force every order to stay for a half second, once that half second is over, the high frequency trader is going to win the game anyway - he will be the first to withdraw, and be the first back in. The other side of the coin is that the regulators are thinking that forcing the HFT guys to stay in the market for half a second will lead to more orders getting hit. But if they get hit a lot more, they will be taking on a lot more risk. So they will start repositioning their bids and asks further away from the market so they won’t get hit.
Wider spreads and lower liquidity. How is that good for markets?
Q: How do you see the issue playing out from here? Do you see shared responsibility among regulators, exchanges and the HFT community?
A: The exchanges should require high frequency traders to log on via a particular terminal so that the exchanges know which types of trades they are. So, if you are a market-maker, you log on from a particular terminal. If you are punting with the market, you log on from a different terminal, and maybe a different one for arbitrage transactions, and so forth. They have got to start tracking these things. The marketplace is going to absolutely hate this, and will resist it like crazy, just as they resisted client-identified orders and trades.
Q: Since the industry will be fighting to prevent restrictions, is this where the regulators step in?
A: Yes; but whatever they do they must be clear about what they want and why they want it. For example, they asked for consolidated audit trails, then they asked for everything, without specifying why they needed it in the first place, because they have not asked the fundamental question: “How, practically, do we decide what is fair and efficient?” They have got to define and operationalize these terms. Otherwise we will just go round and round.
Q: What are you seeing around the world in terms of regulatory arbitrage - or regulatory harmony?
A: We have got HFT in the Unites States at about 60-70 percent. In other places, it ranges from zero to about 15 or 20 percent. The U.S., at the moment, has its hands tied. If they try to change things now, it could be catastrophic. But if you try to change it where it is only 15 percent, it would not be nearly as bad. So the idea is to slow it now, before it gets to be too big to stop.
Q: So where does that leave the U.S.?
A: It leaves the U.S. in a bit of a quandary. The evidence we have suggests that HFT is good for liquidity and efficiency, but there are a few questions about fairness. So, how do we define fairness? My definition is, "the extent to which market participants engage in prohibitive trading behaviors." We need to look at whether, since HFT has gone up, has there been any increase in the amount of prohibitive trading behaviors and, in particular, market manipulation and front-running?
Herein lies the problem. There is currently no way to measure front-running, because brokers are not required to tell you when they are acting as principal or as agent. So, if we can’t measure it, how do we know if it is indeed going on? Before we can get anywhere near an answer to this question, we need a surveillance system that actually gives good indications of market manipulation and front-running. If you can get those good indications, we can work out where the stuff is coming from. We already get broker-based information in the order records, so if we see evidence of certain orders from certain brokers are more heavily involved in prohibitive behaviors, we can look into the records. Then we can ask what sorts of strategies are giving rise to such behavior. Is it HFT? Is it the stuff that is co-located, or is it more general across the broker’s business?
We need to get much more systematic about the process, and we need to start with the concept of fairness.