FX And Equities Algos

With David Mechner, CEO, Pragma
What are the fundamental similarities and differences between equity and FX algos?
Both equities and FX are continuous two-sided quote-driven markets, which creates fundamental similarities. Equity and FX algos therefore offer many of the same fundamental benefits. For example, algos reduce market impact by breaking a larger order up into several smaller pieces so that a smaller price concession has to be paid, and thus better execution. In addition, algos can trade passively in a systematic way, making prices in addition to taking, and thus further saving part of the bid-ask spread even for those smaller individual orders. And in both asset classes, one of the major practical challenges of algorithmic trading is managing adverse selection through intelligent routing and order placement policies.
The biggest difference between equities and FX for algorithms is probably the nature of the fragmentation – including the existence of Reg NMS in equities – and the bilateral structure of the FX market – the fact that more than half the spot volume is transacted on a private, disclosed basis. From a trading perspective, this bilateral model provides a lot of flexibility and is in many ways superior to the equity markets. When a dealer knows his client, he’s able to price liquidity more efficiently. In contrast, when dealers have to price orders to be profitable in public markets, they have to price for the worst case scenario. Effectively, in FX directional traders can get better prices by excluding short-term informed traders like HFTs from the equation and transacting directly with dealers.
To what extent is responsibility for FX trading shifting to the buyside?
The basic structure of the FX market is that the buyside trades against their dealers’ P&L. This creates a clear conflict of interest when the dealer is in control of the client’s order – every dollar the client saves is a dollar the dealer loses. This conflict has been starkly illustrated over the past few years by a series of scandals, and is very much on the buyside’s mind.
As a result, the buyside is shifting where trading decisions are made, pulling back control from the dealer. In the context of click-trading, this can be accomplished through aggregation, which is increasingly common workflow. In algorithmic trading, one way to accomplish it is a service bureau model, in which dealers are used as trading counterparties and liquidity providers, not as agents entrusted to control the client’s entire execution.
That said, there will continue to be long-term sustainable and mutually beneficial relationships between dealers and their clients, and indeed this is one of the strengths of the FX market structure. The buyside can achieve this when have the ability to aggregate liquidity across dealers in order to create a competitive environment that eliminates conflicts and keeps everyone honest.
What are the future trends in this area?
Algorithmic trading is still a relatively small fraction of overall FX trading, and it is likely to continue to increase over the coming years for the same basic reasons it has become the dominant way of trading in equities – it adds significant value.
Other than that, short of a regulatory earthquake, the fundamentals of trading don’t look likely to change much. Volatility and spreads will change over time, but there is a natural homeostasis in the markets. Regardless of the market conditions, there is always a spread that is sustainable and provides mutual value for dealers and their customers. Some players may drop out, but as competition decreases and liquidity dries up, spreads will widen, participation will again become more profitable, and new players will step in to keep the markets healthy and efficient.