Retail trading of currencies is still small as a percentage of the whole foreign exchange market, but by most other standards it is a big market.
Of the daily $3tr turnover of FX trades, retail activity probably accounts for an average of $90bn a day. This compares with estimates by the Bank for International Settlements of around $60bn in 2006 and $10bn in 2001.
Services for private FX traders are mushrooming – newspapers and the internet are full of their advertisements. Even big global banks are making efforts to upgrade their offerings to attract the retail dollar.
Yet it is only just over a decade since the foreign exchange markets opened up to retail traders. The key driver has been the rapid expansion of the internet. Prices which had previously only been visible over Reuters 2000 screens or via the EBS electronic platform could be displayed in real time via the internet.
For the first time, a suitably equipped retail trader could take an informed (or ill-informed) punt on the direction of a currency pair, through an intermediary market maker.
At first, the market makers were typically smaller licensed deposit takers, and poorly regulated in trading terms. This meant that an important risk for a trader was the market makers themselves, still often a characteristic of retail trading today.
According to Aite Group, the financial research consultancy, retail FX trading has gone through three stages.
The first wave were simple trading platforms allowing buying and selling. This was fleshed out later to include research and more sophisticated add-ons.
By the start of the new century, financial institutions such as Citibank and NatWest had opened up their services to customers, offering platforms that were essentially execution-only. Pricing was proprietary, meaning that customers could see, for example, Citi prices and Citi prices only.
Sweet FX
This rapidly evolved into the second generation: multi-user platforms where customers could see other banks’ prices and, if credit lines were in place, execute trades with various parties.
The most successful of these platforms was FXall, owned by a consortium of leading foreign exchange banks, which is still a powerful force in the markets today.
Many of the banks that pioneered the new trading platforms were more than happy to open them up to greater retail trading. Citigroup remains a leader in the field.
Deutsche Bank has expanded its services recently, and now uses social media sites such as Twitter, YouTube, Facebook and LinkedIn to connect with FX traders and attract new visitors to its dbfx.com site.
“It’s a shame that some of the potential of the early ECNs [electronic communication networks] has not been fulfilled, for example in terms of the still limited range of products on offer,” says Howard Tolman, head of strategy at Beta Gamma Research, who has worked in FX for some 30 years. “But part of the history of the FX markets has been the haphazard way technology has developed and been applied.”
Thus, Tolman explains, some parts of the FX market have been doing straight through processing since the 1990s, while in one large US bank until quite recently there were more people working on solving foreign exchange errors than actually trading.
“Basically banks were unable to make rational decisions concerning things like STP [straight through processing] because they were not comparing like with like,” Tolman argues. “STP means less errors. Reduce the errors and you drastically reduce cost, simple as that.”
Meanwhile, software companies had entered the retail forex trading market by launching their own trading platforms. Here, the products offered were miles ahead of the first wave in sophistication.
About three years ago, a third generation of retail forex trading platforms emerged. Here, sophistication with the graphical user interface (GUI) was matched by smarter technology – in particular, the use of algorithmic trading products to produce more sophisticated trading strategies.
“The advantages of algo trading are significant, even if at first casual glance they seem small,” says Tolman. “One of our products for directional trading, for example, should give a $10 to $15 advantage in every million dollars traded. That’s not a trivial advantage, that’s enough to make you an out-and-out winner.”
As part of his firm’s business model, clients that use Beta Gamma’s algo trading models pay a commission on their profits.
But since the mid-2000s the growth has been exponential and, depending on who you speak to, enthusiasts claim that as much as 25% of the forex market will be using algorithms by 2012.
“Given the largely OTC nature of FX markets, it’s hard to put an exact figure on their growth, but nothing short of regulation is likely to hold them back next year and beyond, in terms of the continued rapid expansion of algorithmic execution strategies,” says Tom Griffo, global head of futures sales at Flextrade, which has been involved in electronic trading since 1996.
Algo magic
Although algo trading has similarities with traditional automated FX trading – put simply, a clever black box that does the maths – it also has differences.
Algo FX traders employ four main strategies – algo execution, auto-hedging, liquidity access and statistical trading.
In algorithmic execution, various trading styles are broken down and automated. In practice, this may mean “slicing and dicing” a larger trade into smaller segments and routing it through different channels and crosses to achieve a certain goal.
Auto-hedging means creating orders according to a pre-set formula for managing risks, in a portfolio which may be, say, country-driven, sector-driven or risk-balanced.
This goes beyond a simple static hedge. You create a strategy where market moves in your favour bring disproportionate returns. Positions are continually upgraded to achieve the balancing required.
Liquidity access does what it says. The algorithm tracks which portals or exchanges have liquidity and executes according to models based on factors such as volatility and the cost of trading.
At its simplest, cross-currency trading across various time zones opens up liquidity issues – what happens when one of the crosses has no liquidity at the time you need to execute the trade? Using algorithms, a trader can automatically take advantage of the best route to market.
“When a trader is given CHF/JPY at 4pm on a Friday afternoon in New York, does he trade the CHF/JPY cross or will he trade the USD/JPY and USD/CHF?” says an algo trader. “Based on his position, profit and loss and his order book, a trader can have an algorithm help make the optimal trade decision.”
Statistical trading consists of using statistical tools on historical price data to improve trading returns. If a trader can find even a slight statistical edge, the expected return over a large number of trades can be positive. One of the earliest developers of this approach was Richard Olsen, a founder of Oanda Corp, who set up an early model in the mid-1990s using some two decades of historical FX trading data.
How long it will be before the fourth generation of retail trading platforms, and what they will look like, is anyone’s guess at the moment. But if the past decade and the vibrant interest in FX trading at the moment are anything to go by, retail involvement in the market is likely to increase, bringing ever more opportunities for the savviest intermediaries and technology providers.