From a trader using their skills and multiple channels to execute client orders to a silent world of APIs and smart order routers, the execution process has changed almost beyond recognition over the past 20 years – but, as Colin Lambert finds out, not everyone sees algos as the next big thing in FX markets.
A direct line can be drawn from the better availability of data and the growth of algorithmic execution tools, but although the noise levels around algos continues to climb, it is not exactly clear to what degree the buy side is embracing the concept. Equally, there are doubts about whether algos offer the next big thing in FX markets or if they have already had their day. What is beyond doubt is that algos have changed how the dealing desk operates and interacts with the market.
“Probably the biggest change in the dealing desk – apart from the absolute lack of noise now compared to 20 years ago – is in the technology that supports you,” says Maxine Dennis, an FX trader with 35 years’ experience who recently left SocGen/Newedge in London. “In 1999, if you weren’t good with a calculator and able to work out some of the strangest crosses out there and execute out of the legs, you weren’t considered for a role. Now the algos work it out for you and get you out of the risk. Trading is an exception-based exercise, the dealer only intervenes where necessary.”
It wasn’t always that way and the journey to the current highly automated world has been a fraught one at times, interspersed with uncertain business structures and flash events. Dennis notes, “The period when the [single-dealer] platform started to quote prices automatically to customers was a tricky one because it didn’t manage the associated risk – that was still the job of the trader. You had to keep your eyes on your front-end system to check your position because a trade would just drop in that you had no idea about.
“In time, of course, the machines started managing the risk and that made things easier,” she continues. “The dealers were free to go back to value-added tasks like quoting trickier trades and taking risk. It was then that we first got exposed to algos for executing larger risk and that really changed the nature of the job again. It was now about selecting the right strategy and parameters for that specific trade.
“I tend to think that the FX market is going full circle,” Dennis adds. “There is a need for traders with a good understanding of the market to help the algos execute. You cannot just point and click the algo, you have to be smart about how fast it goes, where it executes. The new generation of algos are much more flexible and that allows the trader to show their skill in a not dissimilar fashion to how they used to.”
On the Rise…?
So is the use of algos on the rise? Anecdotal evidence would suggest it is, but maybe not at the pace expected by proponents of the strategy.
“I think the rise of algorithmic trading is one of the biggest trends that we’re seeing now, says Tod Van Name, head of FX electronic trading at Bloomberg in New York. “It’s interesting because in the equity market, algos have been around for a very long time, mainly because it’s an exchange-traded market and it’s very transparent. Now in the FX space we’re seeing more and more users, and not just the more sophisticated asset managers, but a broader set of users that are interested in using algorithmic trading, primarily because it reduces market impact and it’s a very quantifiable approach to the marketplace.
“These users are attracted by the ability to actually verify numerically what their performance is and whether it was better using certain strategies or not,” he continues. “It helps desks show stakeholders the job they have done – it’s a nice way to be able to say ‘this is the way that we actually execute trades and we can defend why we do it and here’s a quantitative look at that performance’. This is also happening across more trading pairs as well; we have a lot of clients that are trading emerging market algos. For us, it’s a really interesting segment of the marketplace.”
Steve Flanagan, head of FX e-commerce risk for JP Morgan in New York, also sees a growing demand for algos. “I think the trend for giving clients more choice by putting algos on their desk where they have more control around how the algo interfaces with the market will continue,” he says. “Allied to the market colour we can offer from our pre-trade TCA product it creates better informed clients who are confident the execution will meet their expectations.
“We will also see growth as a new generation of market participants enters FX,” he adds. “New entrants to the market, they only demand these products, they are what they want and they understand their potential.”
The changing nature of banking is also likely to play a role in the growth story, according to a senior source at one non-bank market making firm, who says, “I think the majority of the banks today focus very much on consistency of earnings and fee-based income, and that is incompatible with the assumption of risk taking. Obviously within this there are a couple of exceptions, but the majority of the banks are moving in that direction and they’re going to continue to move in that direction because shareholders demand that. The scars of 2008 are still apparent on those banks and they don’t want to take any of that risk. This means they encourage their clients to adopt more of the risk by using algo execution strategies.
“So the banks are saying ‘okay, fee income is better than trading income because that’s way more volatile’ and it’s hard to disagree with them because if you look at every company that is public and quoted on a stock exchange, they get better performance from a stock perspective if they are earning this prudent and consistent earnings,” they continue. “Proprietary trading and principal trading [are] inconsistent with that statement by definition and therefore it is better to be a private company rather than a public company if you want to be in the business of trading, really trading.
“There is the challenge that FX remains a quote-driven market ather than order-driven and that is maybe why the take up hasn’t been as big as some people thought, but I think it is changing,” the source adds. “To help alleviate some of the risk, clients are starting to ask LPs to execute on their behalf, maybe using a TWAP, but we haven’t fully transitioned yet.”
Jason Woerz, president of 24 Exchange, cites Credit Suisse’s AES division as having had the first mover advantage in the world of FX spot algos, but believes competition is not only hot, it has met rising demand. “The broader FX business in general is moving in that direction, more algos have been developed to the point where pretty much everyone at a tier one or tier two bank has their own algo offering now. It is part of the process of electronification.”
…Or on the Decline?
There are those, however, who do not see the algo world growing – they argue that while there is still room for growth in the use of algos, it will be transitional.
“One of the reasons the banks want to sell algos is because they can get fees, which are at least as much as they could earn from a spread if they were still prepared to take a risk transfer trade,” says a senior industry figure. “They can essentially make a risk-free income by selling that algo and so it’s an important revenue stream for the banks and a way of overcoming their own deficiencies or their own unwillingness to take those large risk transfer trades, and they can dress it up and package it from the user’s point of view.
“This means you have a lot of firms who used to be able to do risk transfer trades and their infrastructure is set up around doing risk transfer trades, but they can’t do them in the same way they used to be able to,” the senior industry figure continues. “So for them to be able to sell an algo to let clients manage that risk for themselves is a very convenient tool.
“The reason it should be transitional is that the users are upgrading their technology because they know they need better technology to risk manage their positions themselves, and as they do that, they will be less and less reliant upon these bank-offered tools and algos because they will have essentially their own in-house algos or tools to be able to understand and manage the risk.”
The head of e-FX trading at a bank in London tends to agree, adding, “I am not sure how robust the revenue stream from these products is. It could very much be like when the HFTs started making markets, smaller, more nimble firms can develop algos for the buy side cheaper than the banks. Of course, they will still rely upon the banks for liquidity – that hasn’t changed in 20 years.”
There is also a very old-fashioned hindrance to the growth of algos – trader reticence. A senior e-FX manager bemoans how their institution can roll out a new algo strategy to the dealing desk only for it to sit there redundant.
“The strategy is then rolled out to clients who love it and you wonder, ‘why isn’t my desk using it?’ Often the answer is they still believe it will take their job away from them, even though in reality it offers them an opportunity to show off their skill in reading markets.”
Another factor that may hinder the growth of algos is client satisfaction, Much as is the case with the FX swaps market, which has resisted dramatic change thanks mainly to customer satisfaction with how things currently work, there is a school of thought that maybe algos are a step too far for many. “Banks have been very good at understanding what their clients want and delivering it,” says Chris Purves, head of UBS’s FRC Strategic Development Lab. “That has, however, made it harder to differentiate between the top five or six banks because their offerings are broadly comparable, but one area in which you can is perhaps algos – and that’s mainly because the banks were slow to catch on to their significance. That said, it’s more than 14 years since the first algo arrived and we don’t as an industry, seem much further advanced in terms of take up – I just don’t think that many people want to use them.”
A Healthy Change?
One consequence of the use of algo execution strategies is without doubt the pushing of the risk from sell to buy side, but is this a good thing?
“I think for firms that are sophisticated enough to do the analysis and reduce their transaction costs by assuming some amount of risk, it is a positive development,” says Van Name. “Most clients still have the option to do a risk transfer trade so they can look at it in a quantifiable manner and say ‘here’s the level of risk that I’m comfortable with, what’s the potential gain that I have by trading in a different fashion and still stay inside my risk parameters?’ With the level of pre-trade analytics tools available, they can evaluate their liquidity profile and evaluate a risk transfer price or, given all the factors we know about the market, choose a strategy with a moderately different risk profile but better result.”
Joe Hoffman, head of currency management at Mesirow Financial in Chicago, feels the shift of risk to the buy side is an inevitable result of regulatory changes and that the algos are their response to help their clients.
“Having the ability to access algos does give the buy side more control over their trading,” he says. “Look at markets like equities, they’re big users of algos. As the market evolves thanks to the regulatory changes, and banks risk warehouse less, having access to algos especially for certain currency pairs where maybe liquidity is more challenging or you have a very large order to work, will be essential.”