Defining Algos In Futures Markets

Using algos in finance first emerged in cash equities but the changes sweeping through derivatives markets have ignited demand for more intelligent access to futures and options.

Algorithms affect every aspect of our lives, from deciding which elevator arrives first to forecasting the weather. Yet, in financial markets, the term ‘algo’ rapidly takes on more sinister connotations. The truth is, however, that most algorithms in financial markets provide the same benefits in terms of efficiency and predictability that they bring to our everyday lives.

The idea of using algos in finance first emerged in cash equities, but the changes sweeping through derivatives markets have ignited demand for more intelligent access to futures and options too.

Confusing Fairness with Transparency

There are essentially three categories of algorithms used in financial markets today. The first and most benign are those algos that aim to achieve a certain benchmark, such as volume-weighted average price (VWAP) or time-weighted average price (TWAP), with minimum market impact. Next come those that aim to optimize traders’ workflow and automate reaction to market events, such as triggering orders at specific times around specific market thresholds, or working orders over a specific time period such as the market open and close. The third, and perhaps most contentious, are those that aim to capture alpha explicitly, either through the algo itself or by virtue of the speed or frequency at which it operates. This last category has been the subject of considerable public and political scrutiny, particularly in the U.S.

However, trading – right back to the time of carrier pigeons and semaphore signals — has always been about leveraging speed and the mistake that is often made is to confuse fairness with transparency.

Transparency, not fairness, is really the key issue here so that regulators, other market practitioners and the public at large can see what is going on. They can then decide if, how, and when they wish to participate.

Algos for Derivatives

Although derivatives contracts are more complicated in concept and trading style than cash equities, their non-fungibility simplifies execution logic significantly. The same liquidity cannot be dispersed across multiple lit and dark sources and so the need to look intelligently across different markets, or hit specific benchmarks, has never existed to anything like the same extent. This is all changing, however, as regulations such as Dodd-Frank converge the OTC and exchange-traded derivatives markets so that a broader range of contracts are available to choose from in any given situation. On top of this, the formal concept of ‘best execution’ is being extended by regulators to cover derivatives.

The net effect of all these changes is that complexity in derivatives market structure is rapidly going up and so algos are becoming a crucial part of both buy-side and sell-side armories and look set to become a key competitive differentiator.

Cash Equities — An Obvious Starting Point?

When developing algos for derivatives it may seem like a natural starting point to simply lift out and repurpose what has worked for equities. Experience shows that this approach is likely to come unstuck fairly quickly. Derivatives reflect a broad range of underlying asset classes, from fixed income through to FX and even physical commodities. Each has its own trading characteristics, with differing contract lengths and multi-legged strategies so the ‘one size fits all’ approach that works for cash equities will inevitably be flawed.

In the face of these challenges, the more enlightened Futures Commission Merchants (FCMs) and their buy-side clients are deploying a wide variety of algorithms that have been developed from the ground up. Demand is generating an expanding array of choices and most of these algorithms can be clearly differentiated from those designed to support high-frequency trading and other more controversial trading strategies.

Development is extending beyond the most simple type one algos that were based upon VWAP and TWAP benchmark strategies. For example, algos that mask order types so that trading intention and potential information leakage is minimized offer much greater levels of sophistication, control and discretion. While automation algorithms are removing the need to manually monitor the market, along with those designed to track and hit newer and more relevant performance benchmarks, such as a static price, or even dynamically updated goals.

Many firms also use algos to manage the relationships between orders and trigger pre-defined responses when certain market conditions occur. These can eliminate the errors associated with trading multiple products across multiple regions whilst ensuring that a firm places its order at the right time and it is not rejected by the exchange

The demand for these types of algos has and will continue to increase as volatility returns to the markets and electronic trading volumes rise. That’s not to say there arent clear challenges along the way.

The Science of Compliance

There is growing acceptance of the role algorithms can play in reducing errors and increasing efficiency across the markets. Less widely understood, though, is the need to make them compliant and meet the multitude of regulatory requirements and standards being imposed on the industry. The key issues here are transparency, accountability and the prevention of rogue or runaway behavior.

As regulators have come to understand just how prolific algorithmic trading is becoming, they have naturally sought to protect markets from abusive and potentially damaging practices. This has taken a number of directions, including the introduction of circuit breakers at exchanges and placing formal obligations on market participants to document and test their code better. Regulators are also insisting that participants maintain complete histories of algorithmic behavior so that detailed forensic analysis can take place in the event of a problem.

The obvious aim of all this is to raise the bar in terms of best practice for the development and deployment of algorithms — a sensible objective given the number of high profile algo failures.

The result is that firms need to find ways that industrialize their algorithms and in such a way that their deployment and control can be automated and centralized.

Algo ‘frameworks’ are now emerging that meet all the necessary regulatory best practice requirements but don’t impinge on creativity when it comes to algo creation. This is particularly useful for those firms that operate on an international basis. The framework approach divorces the heavy lifting of differing regional compliance obligations from the actual algorithm itself. Only by achieving this are these firms able to offer globally consistent trading outcomes, maintain client confidence and stay on the right side of the law. It also means that new algorithms can be developed safely and deployed much more quickly, which creates significantly better client outcomes and chances of alpha generation. Those firms that can innovate more quickly in developing and modifying their algos are likely to emerge the winners.

Algos in the Trading Workflow

In isolation, algo frameworks are not the complete answer, as equity markets learned the hard way, with algos misfiring through the execution layer and these problems sometimes not being discovered until it was too late. With todays level of regulatory scrutiny, firms would be wise to take heed of this lesson.

Integration with central workflows and order management systems (OMS) is not only important because in-flight visibility and control over algos is vital, but also because visibility is needed in the same place and at the same time as all the other trading decisions are made, executed and monitored.

You Cannot Manage What You Cannot Measure

Just as we have seen in cash equity markets, the rise in algo trading has spawned a similar growth in analytics. As the regulatory concept of best execution tightens its grip on derivatives it will be essential to be able to prove to yourself, to your client and to the regulator that your algo did what it was supposed to do. If the buy-side is going to take derivatives algos seriously then, it will need to ensure that it has the right tools for analysing the success of different brokers and algo strategies.

Derivative Algos Are Here To Stay

Markets will only ever get more intertwined and asset classes will continue to converge. Derivatives markets are at an exciting inflexion point as structural changes will mean that the complexity of accessing liquidity will soon outstrip the human processing capacity of even the savviest traders. Algorithmic trading, therefore, is becoming an essential part of accessing liquidity but this needs to be done responsibly, measurably and in line with increased regulatory oversight. Those firms that can utilize algorithmic frameworks that are embedded within their order management systems will be able to achieve this and, at the same time, create sustainable competitive advantage for themselves.

Yuriy Shrterk is head of derivatives, product management for Fidessa.