Amid a backdrop of Securities and Exchange Commission roundtables, myriad new regulatory proposals and a changing technology landscape, Traders Magazine spoke with Owain Self, global head of algorithmic trading at UBS, about future trends in electronic trading and order routing.
Traders Magazine: Describe the state of algorithms today. What is the next generation of algorithm technology going to look like?
Owain Self: Today’s algos have come a long way from the days of fixed schedules and rudimentary techniques for order book interaction. They are based on complex mathematical formulae taking in a variety of inputs both historical and real-time. These inputs are used in a continuous calculation, so algos no longer decide at 10 a.m. what they are going to do at 3 p.m. Client demand drives us to assess the market situation based on information in a variety of asset classes, compute faster and be first to react. To keep up with this demand, the technology and processes employed need to continually advance. The super low-latency technology, originally used in smart order routing and market connectivity, is combined with the computing power to run long term optimizations quickly and independently. Interestingly, while algorithmic trading logic grows more sophisticated, we still hear from clients a demand for a simpler order interface and decision-making process, leaving most of the advances having minimal effect on the user interface.
TM: Recent reports suggest the percentage of volume trades executed electronically will match those done in a high-touch fashion. Do you agree?
OS: We’d agree that the trend is likely to continue–but not necessarily aggressively. The distinctions between what is high touch and what is low touch can be a little hard to determine (for example, when a client uses a program desk for efficiency reasons rather than for portfolio management). Institutional clients will move orders between their own desktops and their sales traders based on market conditions, the relative complexity of the order or the liquidity they’re looking to source, and the availability of new electronic trading products.
TM: What about the future of high-touch trading?
OS: Exactly where flow will migrate to/from is difficult to predict. For example, as portfolio algorithms grow in functionality, they could potentially gain share from the program desk, but just as easily could take flow from other single stock algorithms which are being used to manage portfolio orders today. There will always be a need for the full-service, worked-order desk. In times of great volatility, we see clients moving orders back to high-touch streams, especially when capital is required. This healthy balance between worked and direct is likely to continue for some time to come.
TM: We hear some buyside firms are building algos or other systems that direct broker smart order routers. What do you know?
OS: Recently, we have seen a trend of some clients utilizing vendor software to route or allocate between brokers’ liquidity pools. The software essentially replaces the order management logic of an algorithm that would normally be market-responsive and objective-based. This seems to be driven by a perceived need to more easily and efficiently access multiple sources of liquidity. While fragmentation is understandably a challenge for the buy-side, it’s hard to see that any efficiency gain during order entry would outweigh the potential decrease in performance quality.
TM: What does this mean?
OS: We think in most cases these labor-saving tools are a bad trade-off for best execution. It remains to be seen whether the vendors can provide the same level of sophistication and anti-gaming techniques that are employed by most sell-side algorithmic providers. The foundation of these techniques must be based on significant investment in information technology and a far larger set of order data and analysis than any single client could resource individually.
TM: One big issue we hear is that the buyside is very, very concerned about where their orders are being routed. They want to pick and choose. They are asking many questions of their brokers–do you find this to be the case at UBS?
OS: Without a doubt, this has been the number one client question for the last few years. Quite rightly, clients should be concerned about what information on their order is being shared and where it is being routed. For example, we’ve observed a practice in some broker ATSs wherein they onward route orders to another venue, then reported back to the client that the execution traded in the ATS. This creates confusion for the client and is something that is categorically against UBS’s principles. We believe in total transparency around our routing practices, we do not have any onward-routing, non-disclosure agreements in place. And we provide our clients with venue destination flags on all execution reports, plus web-based real-time analysis of those executions. The vast majority of our clients, once made aware of such processes, are more comfortable using our approach. And as always, we still offer a fully customizable platform that allows our clients to exclude any venue, if they wish.
TM: How do you feel about the new market circuit breakers? Do they help or hinder the marketplace?
OS: I think they serve for the overall greater good, especially to protect investors in market situations where it is difficult to know whether the stock price move is real or only due to a momentary lack of liquidity. In this regard, circuit breakers and halts are in principle a very good thing. However, I think they need to be fine tuned.
TM: Do you favor a different approach? What would be better?
OS: Limit up/limit down have some benefits. This means a stock is still open to trade, leading to faster corrections to normal market levels. And since they don’t allow order prices either above or below the limits, they can reduce the risks of erroneous trades.
TM: Do firms that co-locate have an unfair advantage when trading?
OS: If you’re going to invest in technology to improve your connections to public information and your ability to act on it, then being co-located gives you an advantage. But that’s the same with any industry. In the past, the problem was that a small group of people were getting a look at the orders before they were executed and floor traders could trade ahead of those orders. That is not the case here. Here, no one is getting a look at an order before it gets to an exchange. The order gets to the exchange first, before anyone sees it, and then the exchange broadcasts it for all to trade against. Getting the best execution quality is encouraged in the market. Someone who is not willing to invest in technology shouldn’t be able to say that others who are willing have an unfair advantage.
