While the mandate from the G20 countries is bulldozing OTC derivatives onto electronic trading platforms to boost accountability and transparency, it is also creating an environment ripe for algorithmic trading.
Automated trading is already having a significant effect on the futures and options markets although exact figures are hard to come by. According to data from FOW’s data service FOWintelligence.com, the volume of exchange traded derivatives leapt by over 25% in 2010 to 22bn contracts. This year that record looks set to be topped with 20bn contracts already traded in the first ten months of the year.
David Feltes, alliance and venture management, at CME Group estimated recently that between 40-60% of volume in the futures market is generated by proprietary shops’ algorithms. Craig Donohue, chairman of the CME, has been reported as saying that proprietary algorithmic trading accounts for 45% of futures volume on the New York Mercantile Exchange.
The underpinnings of automated trading – standardisation and accessible infrastructure – are now being strengthened in the OTC derivatives market to give regulators and market participants better of trades and positions.
The fallout from the collapse of Lehman Brothers, with the laborious unwinding of OTC positions causing a dramatic loss of confidence in the market, spurred the G20 into action. Its resolution, made in September 2009, to process OTC trades electronically wherever possible and to apply central clearing and reporting to OTC trades by 2012, is being put into effect under new regulatory initiatives around the world.
Automating the processing of OTC derivatives trades will deliver efficiency to the analysis of positions and risk, dramatically improving transparency in the market, which is the regulators’ goal. Mark Goodman, head of European electronic trading at French broker Société Générale, observes that the OTC derivatives market will come to resemble the equities and exchange traded derivatives market as a consequence.
“I think most asset classes will converge towards the equity model as they mature,” he says. “That process has been accelerated as a result of regulation.”
Buy-side consultant, Jeremy Bezant, says that around 30 electronic platforms are expected to compete for order flow of the OTC derivatives, with numerous brokers offering access.
“That will either drive firms down the route of using smart order-routing (SOR) algorithms to trade, using a program to find liquidity, or through a broker-dealer,” he says.
Although there are already multiple derivatives markets in existence, the new rules around clearing will introduce competition that has never existed before. The vast majority of trades will have to be cleared through central counterparties, which take the place of one counterparty on each side of a trade to limit the build up of risk by individual firms.
The regulations stipulate that these clearing houses must be accessible to OTC trades made on any venue. Theoretically, this would make the products fungible, so a trader can open a position on one venue and close it on another. Although these measures are being introduced to reduce systemic risk, they will inevitably deliver advantages to firms seeking to trade algorithmically.
Universally accepted
Fungibility will bring a new dynamic to the derivatives markets. Most clearing houses are tied to trading venues, in a ‘silo’ structure, and will only clear the trades for that venue’s instruments. The silo model allows derivatives exchanges to retain a monopoly over the trading of instruments that they have developed as it is impossible for other venues or clearing houses to process them. To do so they would have to have access to the flow of trade data that the siloed firm uses to price the assets.
Most of the other major derivatives clearing houses – ICE Clear, CME Clearing, Eurex Clearing – operate in a silo with derivatives exchanges. The two exceptions are the International Derivatives Clearing Group which offers clearing in US dollar interest-rate swaps and LCH.Clearnet, which is soon to be taken over by the London Stock Exchange Group.
If contracts are fungible, venues will be able to compete for liquidity in them by offering lower trading fees, fragmenting liquidity across venues.
“Once you can trade product on multiple venues you will have competition to collect liquidity,” says Richard Tibbetts, chief technology officer of Streambase, a supplier of complex event processing technology. “Derivatives trading platform Turquoise has already begun to offer rebate-based structures similar to those used in equities to attract flow and liquidity providers.”
“Mifid II will be to OTC derivatives trading what MiFID was to cash equities trading back in 2007,” asserts Bezant. “It will create competition between venues, and fragmentation of liquidity as a consequence.”
This will have profound consequences on the way that derivatives are traded, he says.
“At the moment you can trade an interest rate swap with a £50 million notional value by making a couple of phone calls,” he continues. “In the future not only will you have to trade this electronically but if the contract is liquid you will need to find that liquidity across various new exchanges, and break your order into smaller-sized trades.”
Goodman adds that this will create ripe conditions for high frequency traders. “Once you have fragmentation you have more opportunity for arbitrage, therefore more high frequency trading activity, a faster trading environment and more liquidity,” says Goodman. “The process of automation becomes self-perpetuating.”
Ramping up
Currently a significant proportion of algorithmic trading in derivatives is performed by proprietary trading shops. They typically operate either ‘market-making’ or quantitative analytical HFT models that make profits intraday and close out their positions by the end of the day. As contracts become standardised these firms will be able to pump increasingly large volumes of contracts through the market, improving liquidity.
“Once you have liquidity being provided in an algorithmic fashion, liquidity takers want to become just as sophisticated and then it becomes an arms race,” says Tibbetts.
Steve Grob of Fidessa also notes that as transparency improves, measurability typically comes to the fore.
“We've had explicit benchmarks in the equities world and people are asking if they could apply those to derivatives trading,” he says. “There is a desire to use algorithms to prove that the traders are hitting benchmarks to improve post-trade transparency. The other dimension is that buy-side firms are taking advantage of some of the extra complexity that will arise in markets to increase competition.”
Although similar benchmarks, such as volume or time weighted average price (VWAP and TWAP), may be used by traders, there will have to be an evolution of the technology enabling them to migrate from existing markets to the derivatives market, predicts Tibbetts.
“Those that naively export their equity algos to a new asset class, whether market making or execution management, will be surprised by how badly it goes,” he says. “It certainly requires careful design and implementation of algos that match the market dynamics. In the early stages of transition the market dynamics will change month to month. So agility in deployment of algos will be critical to success.”
Goodman says that the provision of algos will be familiar to the traders who use them for equities, with the algorithms accessible via third party vendors’ trading systems. But he also notes crucial differences between the markets; some participants in the futures market will trade faster than firms do in the equity market, which affects the trading for all participants.
“Traders might be used to dealing with different types of counterparties in the equities market including HFT, but there is a greater concentration of them in the futures market,” he warns.
With additional dangers there are also additional benefits; the market offers opportunities for complex algos that can interact between different venues. Société Générale is expecting to launch a pairs algorithm that includes a futures leg for clients within the next few months.
“It’s interesting to look at how the standalone products can be brought together,” says Goodman. “Our fundamental way of trading is to look at the equity and what is happening in correlated products so the algorithm is reacting on equity to what is happening on futures. We've been doing this for ten years but now we're making that more explicit for clients and putting the cash equity and the future side by side.”
Although the options market is more complicated than the equities market in some ways, the mathematical relationship between many options products is better defined, allowing more sophisticated algorithms to be used.
“It's mostly a question of understanding the liquidity available to each of these products and running the matrix computation to get you from the price of one product to the price of another product,” says Tibbetts. “There are opportunities for execution algorithms and execution management algorithms that wouldn't be available in equities.
“For example, to achieve a position in one of the illiquid points of the curve then you could take a position in the more liquid strike price or date, and then use that as a hedge against the execution risk in executing these positions.
“Because the products move in lockstep, the more liquid products can be used to hedge the less liquid products. That could be used as part of an execution management strategy, that could be used as part of a market making strategy; there are a lot of ways that can be used.”