Suppose you want to buy something specific. A digital camera, say. You face multiple venues, and a plethora of different prices, different quality levels, and different levels of reliability.
You need a way of assessing all the information to come up with the right buy for you – a programme with all the information from all the different shops and websites, which will enable you to input what matters to you (price, perhaps, or warranty) and tell you where to buy your camera.
This, says Mark Palmer, vice president and general manager at software house Progress Apama, is exactly what algorithmic trading can do in fragmented financial markets. And one area in which they can play a key role is that of dark liquidity pools.
Behind the somewhat “cloak and dagger” name, as Mr Palmer expresses it, these dark pools, “simply put”, are “the aggregation of multiple liquidity pools”.
Stuart Adams, head of European operations at multi-asset trading system provider Portware, goes into more detail, describing them as “hidden pools of liquidity which would not otherwise be visible other than via electronic means – often with large blocks of liquidity available”.
Paul Scott, director of trading solutions specialist FIXCITY, explains that increased electronic trading means financial institutions can execute trades and secure liquidity from many different venues. “Automated platforms offer the opportunity to match ‘off exchange’ with other buyers and sellers, without showing the available liquidity to the market,” he says.
“Dark pools refer to the non-displayed or hidden nature of the buy and sell orders that reside in a crossing platform. The term dark liquidity can also be applied to all forms of non-displayed liquidity such as the order blotters of buy-side dealing desks.”
According to Mr Palmer, fragmentation of the market is responsible for the growth in these pools – “because there are so many venues to execute trades on”. Mr Scott adds: “In the US, the influx of crossing networks and alternative venues, and the rapid adoption of electronic trading technologies, has driven the growth of dark pools.”
Crossing networks, he explains, “allow dealers to match orders off-market and access ‘hidden’ natural liquidity. Trades can be processed anonymously, without impacting the price.”
Currently, dark pools are primarily a US phenomenon, although Mr Scott expects the use of alternative trading venues, which has driven their growth in the US, to catch up in Europe “as traders integrate electronic trading strategies and streamline their execution facilities”. According to Jerry Lees, head of alternative execution at CA Chevreux, there are 30-40 dark pools of liquidity in the US, but “not many in the European or other markets”.
Richard Balarkas, head of Advanced Execution Services sales at Credit Suisse, adds: “There are only two crossing networks outside of the exchanges in Europe – ITG and Liquidnet. But there is dark liquidity. One area is through the exchanges having iceberg orders – people may have a lot more to trade than is displayed on the screen at any time. The biggest sources of dark liquidity are within the investment banks and major brokers.”
Says Mr Balarkas: “One of the biggest problems for any sizable money manager wishing to trade is lack of liquidity and signalling risk. A dark pool is a very simple way you can hopefully capture lots of liquidity and achieve a large proportion of your order being executed without displaying anything to the market.”
According to Mr Adams, dark pools of liquidity have existed for some time, but it is now that technology has made it easier to access them that the value of accessing as much liquidity as possible has been understood.
“The sell-side has operated as a dark pool in the past,” he says, “but under the name of their proprietary book. With the change in the regulatory environment, the likelihood is that client orders are also increasingly represented in those pools.”
According to Mr Scott, thanks to the success of crossing networks in the US, some buy-side firms choose to have their orders crossed first, before trading through brokers. This, he continues, “has added to more slicing and dicing of orders, where the buy-side splits orders across multiple execution venues (including their brokers) with the aim of matching with the liquidity present in each”.
But the sell-side, he adds, is also seeking to take advantage of dark pools, in an effort to get back order flow lost to alternative venues. “Brokers,” he explains, “are investing in the creation of their own internal crossing platforms and the development of more advanced algorithms.” Moreover, he adds that crossing networks are opening up, allowing brokers to trade through them and so linking liquidity to maximise trading opportunities.
Algorithms are crucial to all of this. Mr Palmer explains: “The role of algorithms comes about precisely because of the dynamic behaviour and changing prices – as they change instantaneously, you need very sophisticated algorithms.”
Accessing pools
Mr Adams adds: “Given that algorithms are a subset of electronic trading, and it is the advance in electronic trading which is driving the access to dark pools, there is no doubt that algorithms will become increasingly part of the way the liquidity is accessed.”
According to Mr Scott, brokers’ algorithms can seek out liquidity by spreading their clients orders across the dark pools available. “They combine different techniques,” he explains, “splitting orders and placing small amounts in the accessible pools. When execution occurs, the algorithm pulls the placements from the other systems and ‘piles’ them into the system where the match occurred in the hope of securing all the available volume.”
He adds that any leftover volume is broken up and placed back across the pools, with the result that the buy-side firm’s order is executed efficiently with minimised risk of overexposure or missed opportunities, and without the firm having to manage the multiple venues itself.
All eyes will be in European financial markets in the wake of the Markets in Financial Instruments Directive (MiFID), with many predicting greater fragmentation as a result.
MiFID’s primary objective, Mr Balarkas explains, “is to improve competition, especially between liquidity venues, mainly established exchanges. So,” he adds, “one would expect to see a burst of liquidity with new entrants. The ability to handle dark liquidity looks like being part and parcel of the increasing competition between liquidity venues. That’s what we see in the US. So MiFID is bound to promote competition.”
He adds that Credit Suisse has seen a shift to so-called “dark algorithms”, such as “Sniper” and “Guerrilla”. Traders, he says, are turning to algorithms which do not display anything to the market. “They look for and identify liquidity opportunities and then go in and capture and trade extremely aggressively.”
According to Mr Palmer, MiFID is driving fragmentation in the European equity market, “because firms do not have to list on only the local market any more – it is opening up competition across multiple electronic exchanges”.
However, Mr Lees suggests that there will only be “a degree of fragmentation” in the wake of MiFID. There remains, he explains, an essential difference between the US and European markets. “The issue we are trying to solve in Europe is different from that in the States. In the US, the markets are still inefficient, in a technological sense, but the clearing and settlement side is all centralised. The European market is already very efficient.”
In Europe, he explains, there is no central counterparty. “So if the business becomes fragmented, for example, if I split an order between multiple entities, I am going to have four or five different settlements, so it will drive my costs through the ceiling, and create an operational nightmare.”
The likely outcome, he adds, is that traditional exchanges, faced with competition from new players such as Project Turquoise, the pan-European exchange being set up by a consortium of investment banks, will respond by reducing costs. “So,” he predicts, “there will not be a significant cost benefit in going to external markets, and then the overriding issue will become the cost of settlement. It will be hard to drive liquidity away from the central points.”
KEEPING UP WITH EXCITING NEW TRENDS
Algorithmic trading is about more than dark pools of liquidity, and there are exciting wider trends set to take off in the industry.
In particular, traders in other asset classes are getting interested. According to Richard Balarkas, head of Advanced Execution Services sales at Credit Suisse, algorithmic trading is becoming mainstream in equities, “and traders in other asset classes are taking a serious look at whether the same benefits could be delivered there.”
“Anonymity is key,” he adds. “People are wising up to measuring execution performance. FX is one that could be particularly exciting. We have had several algorithms running for nearly a year in FX, which our traders have used internally and which we have started rolling out to clients.”
Nor is algorithmic trading restricted to single asset classes. Stuart Adams, head of European operations at Portware, explains: “One of the biggest developments is the advance of cross-asset algorithms.” These, he adds, “may involve someone buying one asset, selling another asset and then adding in an FX component to manage their cash constraints.”
Jerry Lees, head of alternative execution at CAI Chevreux, predicts an increase in buy-side firms using technology firms to develop their own broker-independent algorithmic products, rather than work with all the different models used by all the different brokers.
In the short term, he suggests, the keyword is flexibility: “Our client base is asking us to build specific algorithms for them, and specific solutions.” This, he says, could create problems eventually, with brokers “ending up acting like a software house”. But in the short term, he says, the reaction is to co-operate and develop the client-specific technology.
