The FX debate a response from RBS March 2006
It seems that participants in the foreign exchange market have a slightly different view of algorithmic trading to those in other asset classes and products. Generally, algorithmic trading in cash equities and futures refers to a type of programme trade designed to achieve better execution. FX players seem to use the term to describe most aspects of programme trading but especially arbitrage in the form of what used to be disparagingly referred to as sniping.
The overwhelming view expressed following last months Euromoney cover story, The problem with foreign exchange, suggested there is very little support for the idea. This should not have been too surprising, given that the one being touted is effectively a big-boys only club. But the feedback received suggests there was a consensus with the view that the market has been the author of its own problems, even if that acknowledgement is reluctant. More than one respondent said that there was no liquidity problem in FX, just a problem of liquidity at a specific price.
With spreads now eroded almost to the point of non-existence, it is unlikely that the big liquidity providers will ever be able to widen them out again. That obviously makes it difficult to run a business based on capturing spread income. But what really appears to be annoying several of the big sell-side players is the belief that they are being arbitraged by certain hedge funds running sophisticated algorithmic trading programmes. In fact a common accusation is that certain buy-side participants and the name of a West Coast hedge fund cropped up regularly are manipulating the market by spoofing it repeatedly.
The rumours suggest that some players are regularly putting in bids and offers into EBS and Reuters, and then whacking several bank platforms and electric communication networks (ECNs) for a far larger amount at the same time. Some are claiming it is this activity that is causing the so-called liquidity mirage where spot prices gap.
Ostensibly, this looks like a modern version of an old theme. Putting in a bid to make the market look stronger than it is to execute a sell order is nothing new. The problem for the liquidity providers is apparently the speed with which these trades can now be executed and the amounts that they entail.
Jack Jeffery, chief executive of EBS, says that, contrary to rumours, this type of spoofing is not taking place over the markets predominant platform. EBS takes the integrity of the market seriously. Well take action if we find out the system is being used improperly. Weve looked into this and weve not seen anything like this happening from algorithmic users, he says.
Another senior figure also says that it is unlikely this is taking place. Its a belief that the sell side has. But the reality is that theyre providing them [the buy side] with prices on portals. The buy side is then using models programmed to search out the best prices. The concept that anonymity and programmes are manipulating the price is a little far fetched. However, several sell-side traders insist that it is and that the culprits doing it are easily identifiable.
What seems to have been forgotten is that the buy side has only limited access to EBS and can trade only in multiples of up to $5 million or A5 million. This is not enough to spoof such a large and liquid market as spot FX one way or another, so it would suggest that a lot of the algorithmic trading that is already taking place is happening at other trading venues. But the allegation is that as the banks auto-pricing machines generally move in line with the prices trading on EBS and Reuters, some players are able to create a false market by entering and withdrawing prices into the markets two dominant brokers.
What is becoming more and more evident is that spot FX is extremely competitive. As was mentioned in last months cover story, EBS is not immune to these competitive pressures. Looking ahead, it is likely that the growing use of algorithmic programmes will increase these.
Recent financial market history has shown that liquidity is extremely hard to poach. The most famous example was the capture by the then Deutsche Terminbörse of the Bund future from the London Financial Futures and Options Exchange. But that was very much a story of the new electronic trading technology versus the old-fashioned and cumbersome open outcry method.
To a degree, the current debate is also about technology, even if the market is predominantly electronic already. If it is not already, the huge data flow into and out of platforms that result from algorithmic trading will highlight any weaknesses in the speed of any trading venue. In the modern era, latency is simply not an option.
From the feedback, it sounds almost as if there is a battle going on between the established incumbents the sell-side liquidity providers and a more nimble, newly emerged and extremely technologically aware buy side who frankly do not care where they trade. That might be exaggerating the situation, but one senior figure at a buy-side firm said that he was recently visited by a very large sell-side liquidity provider after executing a trade through an algorithm provided by FlexTrade.
The bank complained that it had been hit on several platforms simultaneously. Although it is easy to feel some sympathy for the bank, the most obvious question is: really, if it does not like this happening, why stream its prices to so many venues in the first place and expose itself to risk it would appear not to be able to manage? The buy-side player was surprised by the visit, and claims that his system provides no indication who is making the price. Instead, it just shows a single price and indicates how much it is good for.
The inescapable reality is that algorithms are coming to FX and they are coming fast. Although there are obvious lessons to be learnt from the impact they have had in other asset classes which seems to be what the hedge funds that are using them have done there is a major distinction between those markets and FX. As the managing director of electronic execution at a prominent equity player points out, the vast majority of FX does not take place on an exchange and it is not an agency business.
Banks are putting up principal all the time. A tech-aware customer base is used to algorithmic trading from other markets, where they price take when they want to and price make when they want to. Algorithmic trading is a way of timing entry and exit into a market; the problem in FX is that the banks have been used to controlling the spread and they cant do that with algos; also, the FX guys always find themselves on the other side of an algo. There is an element of robotrading and FX guys arent used to it, he adds.
But limiting access to platforms, forming quasi-cartels or simply throwing toys out of prams will not result in the sell side stopping the inevitable. With the steady growth of the FX markets and the increasing adoption of eFX among the market participants, algorithmic trading is emerging as the next level of trading technology for market participants to contend with, says Lee Ratner, vice-president global sales FX at FlexTrade.
We are just at the beginning of a fundamental shift in the way the FX marketplace trades overall. Learning from our 10 years experience in the equity and futures markets, we have recognized very similar market conditions shaping up within FX.
There is significant demand from both model-driven and discretionary clients for faster, more efficient and consolidated ways to trade within the FX markets. When there is a need like this within a marketplace, technology innovation is not far behind, he adds.
Not surprisingly, Ratner says that many sell-side players have also realized the need to embrace this brave new world. With the burgeoning of eFX, liquidity providers are confronted with the issue of providing competitive pricing while effectively managing risk over multiple distribution channels... In order to stay competitive, banks are embracing technology solutions that provide fast execution, multiple order types, intelligent quoting and real-time risk controls, he says.
It could easily be argued that the exchanges, which have far more experience of the impact of algorithmic trading, are well placed to capitalize on this. Many of the ECNs also frequently trumpet the fact that their technology is more modern and, therefore, faster than venues such as EBS and Reuters. The current fragmented nature of FX obviously makes it difficult for liquidity providers to know exactly where they should stream their prices.
The answer in the past has to been to cover all bases and stream it left, right and often centre. That may have worked well five years ago, but the growing sophistication of end users and the ability to pool this liquidity using technology is putting pressure on the existing market model. Whether significant consolidation will occur, or whether liquidity will pool to fewer venues remains very much to be seen. But as mentioned last month, the genie is out of the bottle and nobody is going to get it back in.