My last Advanced Trading column ("Liquidity Begets Liquidity," available online at advancedtrading.com/opinions) discussed the power of new market structures to change age-old liquidity patterns. The corollary to that, however, is not only how do firms connect to these new venues, but how do they trade against a wide variety of liquidity pools, running at different speeds, with differing latencies, and against disparate pricing and cost structures? While first-generation smart order routing by today's standards would be considered dumb, today's smart routers still aren't really able to keep pace.
In today's market, the need for increasing the intelligence of smart routing is driven by the realities of a more complex execution fabric that has outstripped the capacity of many firms' routers. While the current generation of routers can manage traditional lit venues, the proliferation of new market models, dark liquidity pools, new pricing structures, low-latency profiles and inconceivable messaging rates driven by the demands for high-urgency, high-speed, low-impact, large-block, algorithmic and black box alpha-generation models has exacerbated the need for a more sophisticated box.
Smart Router of the Future
Smart routers of the future will need to be increasingly flexible and multifunctional. Instead of simply linking the firm with the marketplace, a router will need to manage connectivity, price, strategy, size, cost, history/probability and opportunity. Many of these functions happen in parallel, but some -- such as history/probability and opportunity -- need to be managed as a feedback loop in which executions are constantly benchmarked and optimized.
Smart routers typically have focused on price. But price, though still important, no longer is the only factor for best execution; to a large extent, price today is governed by Reg NMS. The other variables now provide the real value-add in the execution.
While connectivity always has been critical, in today's market it is not just connectivity to traditional markets, but connectivity to multiple structures within those markets, as well as connectivity to a host of dark pools with a multitude of different matching models and strategies. Smart routers need to know not only how to connect to these liquidity pools, but how to preference, transverse and interrogate them, as well as determine if the pool has hidden size, how to attract size and how to execute against size once it appears. It also must do this without disadvantaging the order.
In addition, intelligent routers need to understand how these liquidity pools are priced, as price not only impacts trading cost/revenue models, it also is an increasingly important aspect of a platform's overall market structure.
Last but not least, intelligent smart routers need to know how to find liquidity and understand the opportunity cost of placing orders of specific sizes into specific markets. While it would be nice to be in all markets all the time, every time an order is placed, there is both an opportunity cost of not placing that order somewhere else and a risk that placing the order will leak information. Placing and taking liquidity is becoming a science that smart order routers need to master. While it was nice to know that Venue A traded Equity Z yesterday, yesterday was just that: yesterday. Where will it trade today? The better a router performs in determining where and how to place liquidity, the better that router will behave and the better the execution.
While smart routing today certainly is better than the smart routing of a few years ago, the increasing complexity of today's market will push routers into places where few ever dared to go. And while it would be great if there were an elegant solution to solve the smart-order-routing mess, there probably isn't one. So get ready -- it just may be time to get your applications in order as your smart order router may need a little continuing ed.



Printer Friendly



