15 June 2009 Features

 

Market evolution: brokers become MTFs

Dan Barnes

Inter-broker connectivity is creating a growing pool of liquidity that never sees a trading venue. Exchanges and MTFs should be worried.

Increasing levels of trades are off-exchange, driven by cost according to everything brokers and Multi-lateral Trading Facilities have been saying. Reduction of trading costs was the main stated purpose behind broker-funded MTFs; it was one purpose of the Markets in Financial Instruments Directive. But where MTFs where touted as the venue of choice for large brokers, far more liquidity is moving to dark pools, often crossing networks, than to lit venues.

According to the Thomson Reuters report on equity trading market share, dark trading in March 2009 - as reported on Markit BOAT - accounted for 23.35% of the market by turnover and 22.96% of the market by volume. That gives dark trading the greatest market share by turnover for the first time, and makes it second by volume only to the LSE which has 40.32% of the market.

Venues have set up their own dark pools to compete with these crossing networks that are run independently or within brokers. However MiFID has made finding liquidity costly. Fragmentation across markets necessitates multiple connections, as does clearing and settlement, with multiple data feeds reporting trade price data. These add to the costs of trading, dis-incentivising brokers and buy side firms to use the full range of venues. In the words of Richard Balarkas, chief executive of Instinet Europe, "The last thing a bulge-bracket firm wants to do is to trade outside - the foremost objective is to hang on to all the liquidity you've got, including client orders, so that you can either match up client orders and double your revenue, or you have that liquidity pool there for you to have first call on."

Exchange bypass

Historically, brokers lobbied exchanges to reduce the costs of trading but this was a long and laborious process. Technology and regulation have enabled them to bypass the exchange in two ways.

Firstly they backed numerous multi-lateral trading facilities that compete with the incumbents, driving down the cost of trading through competition. They signed an agreement with Turquoise (backed by Citi, Credit Suisse, Deutsche Bank, Goldman Sachs, Merrill Lynch, Morgan Stanley and UBS) to supply liquidity, in effect becoming market makers, although in practice market volatility left it impossible to fulfil.

As a business Turquoise and other MTFs such as BATS Europe (backed by Citi, Credit Suisse, Deutsche Bank, JP Morgan, Lehman Brothers, Morgan Stanley, Merrill Lynch) - have still successfully driven exchange fees down. Miranda Mizen, principal at Tabb Group points out that this in itself gives a return on investment to their backers. "Brokers have been instrumental in MiFID, in bringing fees down. I think some of the investments they've made in the MTFs have been dwarfed by the savings they've made in exchange fees," she says.

Competition has also certainly been surprisingly effective in moving liquidity says Stephane Loiseau deputy global head of execution services of Société Générale Investment Bank, "Normally when you have a sweeping change like MiFID there is a tendency to overestimate the effects of it - some people will be looking for the worst case scenario, others for the best case scenario, so either way you tend to see polarised predictions. The reality is that the movement of liquidity has actually been much more rapid than anyone thought, and this, despite the recent equity crisis."

"Prior to MiFID people were predicting a move of 6-7% of trades shifting off exchange by this point in time. Actual estimates - and they are estimates because there is some disagreement over what constitutes a trade - put the volume of trading that has moved off exchange at 25-35%. I found one article from late 2007, three weeks after MiFID came into play in which the highest number they quoted with a timeframe going to 2011 was 15%," says Loiseau.

Despite the drop in exchange fees, moving trades away from a central point has effectively led to the overall cost of trading increasing - partly because of market volatility - but partly because of the broader market created.

These are significant increases according to Anthony Kirby director for regulatory and risk at Ernst & Young. "Last year we saw some extraordinary developments to the total cost of trading - commission plus spreads, plus market impact plus clearing and settlement began to total 100 basis points, where in 2007 it was around 60 basis points. That means if you're a fund manager you have to look at where capital commitment is helpful, and also be more careful about information leakage," he says.

Brad Hunt managing director, co-head of global algorithmic trading of Goldman Sachs agrees: "The effect of average trade size reduction has been bad for investors from a cost perspective, but revenue enhancing for exchanges. For example, from the period of 2006 to 2009, for an investor trading French stocks on Euronext, costs have doubled."

As price formation must take into account the prices from a huge range of venues, the costs of multiple sources are factored into trading costs. Hunt points out that while MTFs are giving their trading data away for free the incumbents are not playing ball. "MiFID has not yet fostered contestability in real-time market data and as a result European exchange data remains relatively expensive. Real time prices have not declined despite a loss of market share, which demonstrates pricing power. There are other more significant issues which prevent the efficient consolidation of data in Europe however (e.g. lack of common and independent symbology, consistent application of reporting rules) and it is in this context that we hear calls for a consolidated tape in Europe. A consolidated tape could mean many things, but one aspect that may be appealing is a more democratic distribution of data revenue, for example if distributed based on market share."

As Giles Nelson, senior director of strategy at Progress Apama, says: "You have Turquoise and Chi-X unashamedly competing on price. They have low latency networks and so on but really their offering is about competing on price." Which begs the question; if costs are up and MTFs are cheap why aren't they getting a bigger share of the pie?

Inside out

There is a second route to bypassing exchanges. Brokers can process trades internally, by becoming ‘systematic internalisers'. This allows the broker to trade clients' orders against those of other clients or against its own book using crossing networks like Credit Suisse Crossfinder and Goldman Sachs' Sigma X.

The business of crossing trades internally has always been done by brokers, however where before it was done by phone it is now done electronically.

Increasingly people are expressing concerns about the growth in use of internal crossing networks, specifically questioning their regulatory status.

External crossing networks, such as Liquidnet, are registered as MTFs and regulated. Because crossing networks do not display order prices they are considered to be ‘dark pools' of liquidity.

The internal networks are also considered dark pools. Last year Dan Mathisson, managing director and head of Advanced Execution Services at Credit Suisse, said that "CrossFinder is now consistently one of the top three dark pools in the US ... [it] has become a fantastic tool to help our clients achieve high fill rates without displaying bids or offers."

But they are not regulated as MTFs, and privately most senior practitioners in this area believe this will change.

Loiseau says that there is a huge blurring of lines between businesses: "The situation in the market right now is that everyone is competing with everyone. The debate at first was whether the MTFs would threaten the exchanges. The debate has changed completely. There has been a vertical integration and now MTFs are offering dark pool capabilities, dark pools are offering MTF capabilities, and with the launch of projects like Baikal exchanges  seem to be going further with services that could potentially compete with the broker dealers themselves."

He continues, "You have 19-20 players all competing for the same thing - liquidity - using different products. The MTFs are launching dark pools to get a portion of the liquidity that they aren't getting through their lit pools. The definition of dark venues and dark orders has to be refined. Dark orders are already present on lit markets. If you place an iceberg order on an exchange it's a dark order - it's not visible and accessible to everyone. CESR and regulators are looking at not only dark venues but a new order type, the dark order."

The key issue for regulators looking within the MiFID framework is whether or not any organised trading outside a regulated market falls within the definition of systematic internalisation or operation of a multilateral trading facility.

Hunt acknowledges that with new business models growing "It's clear that regulators are concerned about the application of pre-trade transparency waivers for MTFs and exchanges. As the lit books become less relevant to institutional trading due to the proliferation of high frequency trading and the reduction of trade sizes, it's natural that demand for non-displayed trading would grow. Many models exist, from broker-dealer models to MTF and exchanged-based offerings, and we can expect more to come to market."

MiFID defines a systematic internaliser as an investment firm which, "on an organised, frequent and systematic basis, deals on its own account by executing client orders outside a regulated market or MTF." (Article 4.8)

Systematic internalisers are subject to the pre-trade transparency and dealing regulations set out in the MiFID Level 2 Regulation. CESR lists 11 EEA firms that have notified that they are systematic internalisers - most of whom have either made their notification to the FSA (as the home regulator) or operate much of their trading out of London.

MiFID states that ‘Multilateral trading facility means a multilateral system, operated by an investment firm or a market operator, which brings together multiple third-party buying and selling interests in financial instruments - in the system and in accordance with non-discretionary rules - in a way that results in a contract in accordance with the provisions of Title II.' (Article 4.15)

Broker or MTF?

Whether or not any individual system meets the full definition of MTF will depend on whether it can be regarded as multilateral, or operates on the basis of non-discretionary rules. This is reportedly being investigated by regulators. As internal networks become linked - such as the recent hook up by UBS, Goldman Sachs and Morgan Stanley, another by Credit Suisse and Instinet and venues such as Turquoise begin aggregating liquidity from internal crossing, the line is being blurred in the eyes of some.  Where trades are matched with groups of internal networks through an algorithmic system, a loophole is effectively created, allowing the networks to take on volume that might otherwise go to a venue.

"Interconnectivity is interesting," says Hirander Misra, chief operating officer of Chi-X. "Turquoise made a recent announcement about aggregating dark liquidity, equally Morgan Stanley, Goldman Sachs and UBS announced they were linking their systems together, Baikal is working on this and we're looking at this as a firm. There's no RegNMS in Europe but where it's viable you will definitely seeing greater interconnectivity between these venues."

It is reasonable to observe that the European and US approach to the regulation of trading systems is rather different. MiFID applies the same regulations to all functionality falling with the MTF definition. Reg ATS in the US is often seen as capturing a rather broader range of functionality but as operating rules that have more gradation (e.g. in terms of pre-trade transparency, access requirements). The rigid EU interpretation means that brokers could effectively become venues, change the landscape dramatically and stock exchanges would become even more under threat.

John Barker, managing director of Liquidnet Europe believes that their status will be defined in the next quarter, "The FSA is working hard with us and every investment house and MTF out there. They want to get firm rules and requirements on what should and shouldn't be an MTF - I think we'll see some clarity on this over the next three months."

If Rainer Riess, managing director at Deutsche Bourse's hybrid dark pool Xetra, is right then a lack of change would be bad for investors: "The ‘dark' terminology has been misused, in executing orders away from the exchange without honouring the pre-trade obligations that MiFID places. You're seeing an increasing amount of orders matched on order management systems that are very often not regulated at all," he says. "Where there is a proportion of flow being traded without pre-trade transparency then you are entering into dangerous territory for the price formation process. I think we need clear guidance from regulators and look at some of the waivers that MiFID has established and ask ourselves ‘Are they detrimental to market quality?"

The London Stock Exchange is targeting this internalised liquidity with its new dark pool Baikal, "We will attract flows that are internalised now." Said Xavier Rolet, chief executive of the LSE, "I would say on average internalised flows reached 10-15% of total trades, and I think we can take over a quarter of that business. So that would be up to 5% of the total market."

The route of the problem

To match a trade with best execution requirements at a workable pace across the many venues found in Europe requires automated matching via smart order routing systems. To do it manually would be impossible.

These routing systems or SORs are provided by both technology vendors and brokers; however as brokers are able to bundle the technology in with other services they are providing the majority of these systems.

Nelson sees them as a good counter to the threat of disintermediation by the buy side "Will people access the market themselves? Do they need to go through a broker? The brokers see these tools as a good route by which they can be seen adding value. The market is more complex, so smart order routing and total cost analysis tools can be invested in by the broker once and then distributed to clients. It's a clear differentiator."

"You still have the situation where the primary broker controls the vast majority of the business that goes through," says Tony Whalley, investment director of Scottish Widows Investment Partnership. In his view this increases the competitiveness in the market as they strive for best execution for clients, "I think that brokers are using their smart order routing technology to control where that business goes to a greater extent, and in that they have a greater control over what's happening, however the crossing networks and MTFs still have to make sure they're up to speed or the brokers just won't put business there."

And although the smart order routers may be reaching every venue to check prices, according to Mizen at Tabb Group there is always the chance that the crossing networks will stop them ever seeing that bit of the market, "If you look at Goldman Sachs, Morgan Stanley and UBS getting together, you don't necessarily triple the possibility of an execution occurring by putting three dark pools together, but you do potentially increase the chance of a trade occurring which means it wouldn't reach the open market. If you like the ‘dark cloud' before the order hits the ground is getting bigger and that is potentially threatening to those who are relying on the trade getting down into the mainstream."

Barker thinks that having trades routed and matched by the broker is potentially a worry for the buy side. "One thing is trust. If you are routing through a firm with proprietary trading then that is a concern for a number of reasons. One is the issue of risk and another is the issue of what is happening to your flow. So I think trust is important for a number of reasons. Security of the order flow - whether it's going to a dark pool or a lit pool, there is concern on the buy side as to what is really happening with their flow. Security and anonymity are key," he says.

"When this used to happen over the phone and the broker would phone twenty people to place your order, that was dissemination of information and now that's happening electronically so your order is routed to venues, other brokers, dark pools, and I would argue that that is also dissemination of information. The only question I would ask of the MTFs and the dark pools is whether that order is then being routed electronically within the venue or whether there is any human intervention. That would give a level of comfort that no-one knows who's buying and selling, or how large the order is."

"The brokers get great flexibility from having the secondary markets to execute on, and we on our side are able to piggy back on that from DMA, algorithms etcetera but net, net, net, it is still the broker who effectively controls what is happening."

However the sell side are aware of these concerns notes Misra. "Buy side firms want the best price, they want to their orders to check as many tangible venues as possible they don't just want all trades to be internalised. The brokers acknowledge that and that's why internalised volumes are around 10-15 %. That's where the lit markets - whether it's Chi-X or independent dark pools - also play an important part. The broker's algorithms will split large orders with one part being routed internally and the rest run through multiple other venues."

Tape to the rescue?

Systematic internalisation is not the greatest challenge to MTF liquidity. The lack of a consolidated tape of pricing data has been hitting MTFs hard says Whalley, "When the main exchanges go down you find two things happening. Firstly the brokers' smart order routers use the price from the primary market to determine what's going to happen, so they can't cope if the markets go down. The other thing is that the buy side say ‘If the primary market goes down we don't want to trade anymore because we can't be sure what the real price is."

Mizen concurs that trade reporting needs to be fixed on a wider level, "There is uncertainty that the trade data is 100% accurate. There is a lack of faith in the buy side that that is so. There's a belief in double counting, reporting in different ways, and until you have standards to compare like for like - errors aside - this will be the perception."

All this may be about to change. Regulators have made recent alterations to the rules that will allow sources other than the primary market to be used. Misra says that "CESR has acknowledged that for a reference pegged dark book you can use can now use a European best bid and offer, as long as you use primary and other tangible venues. And it's up to you to decide what makes a venue tangible."

He continues: "The problem with a centralised consolidated tape would be that it would affect latency as the data has to go in and then go out again, and in the US people are reconstituting the tape to get around that issue. So this will not be a mandated, regulator-led initiative - we will go ahead as an industry and work on the development of the tape. Progress will be gradual but a lot of the work on developing standards has already been done. The tape will eradicate concentration risk. If the primary market goes down, pegging can continue in a dark book or elsewhere."

Whalley believes that the development of the tape will definitely move liquidity to MTFs and internal networks, "I think it certainly would. Equiduct carried out a survey taking all of the trades, their prices and times, and they found it didn't matter which of the venues you were trading on, exchange or MTF, between 15-20% of the trades that took place took place at sub-optimal pricing."

Primary market separation

If the tape would result in further migration away for exchanges they will need to fight back. The LSE's Baikal and Xetra Midpoint are initiatives designed to regain and retain market share. Publically the exchanges do not seem worried by their smaller rivals. They do not believe that many of the MTFs will last long. Larry Leibowitz, head of US Markets & Global Technology at NYSE Euronext said recently: "Some of them may roll up with each other, some may throw in the towel. When MiFID was on the horizon it seemed like a great opportunity, everybody said we're getting into the game, a year and a half later not all of them have got traction so my guess is some will throw in the towel or consolidate first."

He was also dismissive of the exchange buying an MTF: "What's the value of somebody who has 3% market share? BATS and Edge have 10% of the market in the US, but as Bob Greifeld said recently, why would I buy them - they could just go and do it again."

This is a sentiment echoed by Clara Furse, outgoing chief executive of the London Stock Exchange as she held her farewell press conference: "The question is how much technology do they own? What exactly are you buying? They are losing a lot of money right now."

However, thinking about MTFs as direct competitors to the exchanges' lit pools would be a mistake. Back in 2007, Phillip Hylander, global head of equity trading at Goldman Sachs - one of the backers of Turquoise - said that the project would create a hybrid market, with a public limit order book and a non-public book. "It won't just look like current exchanges," said Hylander, speaking at the 2007 Exchange Forum conference in London. "Turquoise, if structured right, will be an aggregator for dark pools - it doesn't need Turquoise for that to happen, but it's much easier to see how it would work [with Turquoise]." An alternative for the banks would be to set up webs of bilateral arrangements, he said.

On both counts he has been proven right. Liquidity is moving to dark pools and they reside within brokers. If, as most suppose, the brokers do become MTFs and trading occurs not on Turquoise but through it, secondary markets could fundamentally change. At the Exchange Forum 2009, Mark Hemsley, chief executive of BATS Europe, asked whether the dual role of exchanges wasn't already an anachronism, "Just because they are the best place for primary issuance, doesn't mean they are the best place for secondary trading of those issued stocks."

The future of the market rests on three key points: the creation of a consolidated tape, the hoarding of liquidity and technological innovation. The first two will continue to empower the sell side and their venues, and are both dependent upon regulatory decisions. The latter is the main weapon in the exchanges' arsenal for clawing back market share.

One influential voice, so far muted, on all of these will be that of buy side. Riess for one believes that "MiFID started with buy side empowerment as an aim and has ended up empowering the sell side." Perhaps it is time for that voice to speak up.