Algorithmic Trading in the FX World
One of the challenges for any organisation considering the use of algorithms for foreign exchange (FX) is to understand the evolving definition of algorithmic trading and how algorithms can be used to better exploit FX opportunities. For many, the term ‘algorithmic trading’ is synonymous with equities execution strategies that focus on achieving benchmarked prices and minimising market impact, such as volume weighted average price (VWAP). This perception makes sense when you consider that algorithmic trading originated in the equities market to automate trade execution. In other words, the decision as to ‘what’ to trade is made elsewhere (generally by a human trader) and the focus of the algorithm is ‘when’ to trade – how to best execute the trade decision in order to achieve best execution and minimise market impact.
However, to better understand the potential role of algorithms in FX, you need to expand the view of algorithms to also include decisions about ‘where’ to trade, not just ‘what’ and ‘when’.
With the fragmentation in the FX market, the focus of FX algorithms will be less on minimising impact within a single liquidity source. Of more importance are algorithms that provide a view into multiple liquidity sources in order to find the right liquidity (both price and depth) in a timely fashion.
In more and more situations, finding liquidity in FX is as much a decision of where to place an order as it is when to place the order. Liquidity in the FX market is spread across diverse pools, including bank-provided liquidity pools, as well as aggregators and electronic computer networks (ECNs), such as EBS, Reuters, Currenex and Hotspot. With the inherent liquidity fragmentation of the FX market, algorithms need access to multiple liquidity venues to identify the best price and depth. In order to deploy effective FX algorithms, traders require a system that provides an aggregated view of the multiple FX pools.
The implications of this requirement are three-fold:
Forward thinking buy-side firms are devising sophisticated trading techniques to make the most of this aggregated view of liquidity. For example, variants of multiple currency crosses, across multiple liquidity pools, are being traded and arbitraged like regular currency crosses. These synthetic instruments can be continuously recalculated, whenever the underlying instruments move to break down orders across the appropriate sources.
One of the key evolutionary trends in algorithms is the movement away from exclusive reliance on first generation ‘black box’ algorithms. Sell-side firms offer black box algorithms to their clients and, while they provide a solid starting point, they are inherently commoditised, offer little real differentiation and limit the ability to achieve alpha returns.
The emerging trend is for more customisable platforms that allow traders to encode their own unique trading ideas in automated algorithmic strategies. In a continually shifting market, traders need the ability to rapidly develop, customise and evolve their algorithms. While standard development tools like C++ and Java can be used to create algorithms, the lengthy development time needed to create and backtest these algorithms result in a lack of agility to rapidly respond to changing market conditions that create new trading opportunities and threats.
For trade execution algorithms like VWAP, this lack of agility may be less of a concern, but for alpha seeking strategies, the ability to quickly deploy and continually re-calibrate algorithms is often critical. ‘White box’ algorithms, on the other hand, provide trade and quant strategists with greater levels of control. These are the antithesis of black box trading and empower members of the organisation to act upon unique trading ideas, incorporating them within the code of an algorithm to generate alpha returns. The ability to customise algorithms according to a firm’s unique requirements and quickly develop algorithms for first mover advantage brings increased opportunity for competitive gain.
Within the white box, the rules-based algorithmic trading technology allows the buy-side to rapidly encode algorithms, in terms of ‘when-then’ rules. The ‘when’ part can monitor for a pattern on incoming market data streams and calculate analytics. When a pattern is detected, the ‘then’ part can specify how orders can be placed. By monitoring real-time market data, analysing the relationships between individual events on a continuing basis, and taking automated actions, firms are able to quickly react to emerging market opportunities for first-mover advantage.
The white box approach has been increasingly adopted by the FX community and applied to a wide range of applications – including high frequency trading, position management and risk management. Unique strategies can now be used at any single stage of the trade, providing a huge benefit to financial institutions looking to gain a competitive edge over their rivals.
FX is also often a component in algorithmic cross-asset class strategies. For example, consider a cross-border trading application where arbitrage opportunities may exist on an instrument that is listed on different exchanges in two countries in two different currencies. Real-time FX data can be used to convert the prices to a uniform currency in real-time, enabling the arbitrage engine to do its work.
A common concern about algorithmic trading is the increased exposure to risk, derived from delegating trading to an automated system. While this concern has substance, it assumes that such risks cannot be managed. Actually they can – often with the same technology that is used for the algorithmic trading. The technology that monitors market conditions and identifies patterns that warrant trading actions can also be used to monitor portfolios and continually appraise value-at risk to ensure breaches of risk thresholds are identified immediately. Corrective actions can be instantly taken, such as trading to take a position back to a more risk-neutral status.
Given the impact of economic or political conditions on currency rates, it is easy to see the potential impact upon FX algorithms of changes or updates to key economic or political indicators. While electronic news feeds have been available for years, the relatively unstructured nature of news information has created significant challenges in the efforts to incorporate news information with an automated trading strategy. Analysing a news feed to identify its real meaning – and therefore its real impact – is difficult to do in an automated fashion.
However, in response to the increasing demand for traders to make decisions, as news is happening, electronic news providers such as Dow Jones and Reuters have introduced more structured news feed capabilities, for example providing XML tags that can identify elements with the news feeds that are computer readable by algorithms. By turning streaming text into ‘textual data’, such news is much more amenable to automated interpretation by a trading strategy. With the right tags, a strategy can analyse and react to news much more quickly than a human trader could.
By combining the power of electronic news and algorithmic trading, automated strategies can now respond to changing economic indicators like employment rates or pricing indices, and anticipate the prospective impact on exchange rates in ways that allow them to pre-emptively act ahead of the rest of the market.
There is no question that the rise of algorithmic trading has impacted the way traders operate. Just as the definition of algorithmic trading has evolved, so has the role of the trader. Today’s traders require more quantitative skills in order to develop and tailor algorithms themselves. Algorithmic trading has enabled the scaling of the trader, so they can work in a capacity far beyond what they could before trading algorithms existed.
An FX trader can now handle many hundreds of complex trades at the same time by initiating an algorithm to manage each trade. This contrasts with managing two or three trades manually. Consequently, more autonomous algorithms will exist, constantly hunting for opportunities and evolving over time.
The application of algorithmic trading in the FX market has significantly evolved from its origins in equities. With the technological issues now addressed, algorithmic trading can enable the development of sophisticated and customised algorithms and the aggregation of multiple pools of liquidity to generate alpha in the fragmented FX market.