I recently had an interesting conversation with Arzhang Kamarai from Tradeworx about the role of high frequency trading (HFT) in equities versus foreign exchange (FX or forex), which prompted me to write about the complexities in migrating HFT strategies to this asset class.
Wikipedia defines HFT as “the execution of computerized trading strategies characterized by brief position-holding periods, in many cases taking advantage from [sic] microstructure inefficiencies. In high-frequency trading, programs analyze market data to capture trading opportunities that may open up for only a fraction of a second to several hours.”
Historically, HFT has been predominately focused on the equities market. It emerged as market structure changes created opportunities for arbitrage and as the market making and specialists businesses became less profitable. HFT grew quickly, and in 2010 accounted for as much as 70% of all equity trades in the US.
When HFT first emerged, the cost of entry was prohibitive for most players. But in the ensuing years, the cost for hardware, software, co-location, low latency connectivity and hardware acceleration have all declined. As the technology and infrastructure become more commoditized, more players enter the space, and that results in reduced profitability for everyone.
So the most sophisticated firms are increasingly looking to other instruments as profitability in equities declines. At the same time, foreign exchange (FX) increasingly has been traded as a distinct asset class. The FX market has grown very quickly in the past few years, growing from an average daily volume of $1.5 trillion USD in 2001 to more than $4.5 trillion in 2010 according to Aite Group. Sang Lee from Aite says, “High frequency trading will represent 35% of the FX trading volume this year (2010).”
The FX market is an interesting space for HFT. It’s traded over the counter, and no single bank controls a substantial percentage of the market. The top 10 global banks control more than 77% of the market. However, according to EuroMoney, the market share controlled by the top three banks actually declined in 2010 over their 2009 market share. There are a number of multi-dealer platforms or ECNs, but these combined control less than 15% of the market. A fragmented and inefficient market is an ideal target for effective HFT strategies.
Each bank or ECN publishes its own prices — generally derived from an aggregated book that sources liquidity from Tier 1 banks and from EBS and Reuters. This creates some interesting opportunities for algorithmic arbitrage strategies. A hedge fund or HFT firm could aggregate liquidity from multiple banks simultaneously to find attractive prices for aggressive algorithms.
But it’s much more complicated than equities trading. For one thing, banks have to manage their risk, and so they watch order flow carefully. They may widen spreads or even stop publishing prices to a given customer if they sense predatory trading practices. So if an HFT firm wants to play in the FX market, it will need to develop flexible technology that allows it to respect relationships with the banks and play nice, allowing the banks to effectively hedge their risk. Firms that play fairly will be rewarded with better prices and deeper liquidity.
In addition, each bank and ECN offers different types of pricing and order types. Where one bank might stream firm prices, another might send indicative prices that are not executable. Most FX data is delivered in pulse intervals rather than continuously, and by the time the order arrives, the price may have moved substantially. So HFT firms will have their work cut out for them in developing liquidity aggregators, determining pricing, and creating smart order routers that are sensitive to distinctions in the various price feeds and to the relationships with the counter-parties.
HFT in the FX space requires substantial investment in technology and infrastructure. In FXCM’s S1 filing for Initial Public Offering, they state, “FX brokers cannot rely on standardized and inexpensive infrastructure solutions that are available to online equities brokers… This requires large investments of time and money but can result in points of competitive differentiation.”
MarketFactory.com has a helpful infographic that describes the differences between technology needs for HFT trading in equities and FX.
I’d like to hear your thoughts. Do you think HFT will increasingly target the FX space? What do HFT firms need to think about as they consider this asset class? What do the dealing banks need to consider as they adapt their systems to handle this order flow?
Wikipedia defines HFT as “the execution of computerized trading strategies characterized by brief position-holding periods, in many cases taking advantage from [sic] microstructure inefficiencies. In high-frequency trading, programs analyze market data to capture trading opportunities that may open up for only a fraction of a second to several hours.”
Historically, HFT has been predominately focused on the equities market. It emerged as market structure changes created opportunities for arbitrage and as the market making and specialists businesses became less profitable. HFT grew quickly, and in 2010 accounted for as much as 70% of all equity trades in the US.
When HFT first emerged, the cost of entry was prohibitive for most players. But in the ensuing years, the cost for hardware, software, co-location, low latency connectivity and hardware acceleration have all declined. As the technology and infrastructure become more commoditized, more players enter the space, and that results in reduced profitability for everyone.
So the most sophisticated firms are increasingly looking to other instruments as profitability in equities declines. At the same time, foreign exchange (FX) increasingly has been traded as a distinct asset class. The FX market has grown very quickly in the past few years, growing from an average daily volume of $1.5 trillion USD in 2001 to more than $4.5 trillion in 2010 according to Aite Group. Sang Lee from Aite says, “High frequency trading will represent 35% of the FX trading volume this year (2010).”
The FX market is an interesting space for HFT. It’s traded over the counter, and no single bank controls a substantial percentage of the market. The top 10 global banks control more than 77% of the market. However, according to EuroMoney, the market share controlled by the top three banks actually declined in 2010 over their 2009 market share. There are a number of multi-dealer platforms or ECNs, but these combined control less than 15% of the market. A fragmented and inefficient market is an ideal target for effective HFT strategies.
Each bank or ECN publishes its own prices — generally derived from an aggregated book that sources liquidity from Tier 1 banks and from EBS and Reuters. This creates some interesting opportunities for algorithmic arbitrage strategies. A hedge fund or HFT firm could aggregate liquidity from multiple banks simultaneously to find attractive prices for aggressive algorithms.
But it’s much more complicated than equities trading. For one thing, banks have to manage their risk, and so they watch order flow carefully. They may widen spreads or even stop publishing prices to a given customer if they sense predatory trading practices. So if an HFT firm wants to play in the FX market, it will need to develop flexible technology that allows it to respect relationships with the banks and play nice, allowing the banks to effectively hedge their risk. Firms that play fairly will be rewarded with better prices and deeper liquidity.
In addition, each bank and ECN offers different types of pricing and order types. Where one bank might stream firm prices, another might send indicative prices that are not executable. Most FX data is delivered in pulse intervals rather than continuously, and by the time the order arrives, the price may have moved substantially. So HFT firms will have their work cut out for them in developing liquidity aggregators, determining pricing, and creating smart order routers that are sensitive to distinctions in the various price feeds and to the relationships with the counter-parties.
HFT in the FX space requires substantial investment in technology and infrastructure. In FXCM’s S1 filing for Initial Public Offering, they state, “FX brokers cannot rely on standardized and inexpensive infrastructure solutions that are available to online equities brokers… This requires large investments of time and money but can result in points of competitive differentiation.”
MarketFactory.com has a helpful infographic that describes the differences between technology needs for HFT trading in equities and FX.
I’d like to hear your thoughts. Do you think HFT will increasingly target the FX space? What do HFT firms need to think about as they consider this asset class? What do the dealing banks need to consider as they adapt their systems to handle this order flow?
Postingan
