Single-Dealer Platforms - a white paper

Banks need to understand the technical and business challenges of building a successful Single-Dealer Platform to harmonise and optimise their e-trading offerings

Click here to download a copy of the "Single-Dealer Platforms" whitepaper in PDF format (905kB). 
The text of the whitepaper is reproduced below.

Modern capital markets span a huge range of trading activities, carried out in a multitude of ways across a host of trading venues. As electronic trading has swept through asset class after asset class, the variety and complexity of these activities have mushroomed.

It is the responsibility of sell-side institutions to manage this complexity and present pricing, liquidity and related content to the buy side in a useful and accessible way.  Success in achieving this has increasingly become a vital competitive factor for many banks. In particular, firms such as Barclays Capital and Deutsche Bank have benefitted from the skill with which they have addressed this in their BARX and Autobahn family of offerings, respectively.

Equity and futures exchange-based trading has been mainly electronic for many years, and there is extensive e-trading and client connectivity infrastructure in place. But the much more recent advent of e-trading in the over-the-counter (OTC) markets has presented major new challenges for dealers, including automating pricing strategy, managing liquidity across multiple pools, and real-time risk management.

Key among these challenges is how to deliver high-quality trading services to a mass market. As OTC trading has moved from voice to electronic, the size of the addressable market has ballooned from a few hundred counterparties to tens of thousands. Just as Amazon figured out how to bring a personalised, accessible book-buying experience to a huge audience online, so the best foreign exchange and fixed income dealers are learning to use the Web to deliver integrated multi-product research, analysis, pricing and execution to their clients in a truly compelling way.

The rise of the intermediaries

Banks first started to offer their clients electronic trading in the OTC markets around the turn of the millennium. In the early days the channel to the client was almost always either Bloomberg or other entities, such as TradeWeb and FXall, created specifically to mediate the dealer/client relationship. Similar in some ways to ECNs, these new venues are usually termed Multi-Dealer Platforms (MDPs).

At that time there was a general belief (enthusiastically encouraged by the MDPs themselves) that buy-side clients would achieve greater efficiencies by getting simultaneous quotes from two or more dealers before committing themselves to a transaction, and that aggregation of liquidity by an MDP was therefore a necessary step before passing it on to a client.

But it wasn't long before pioneering banks started taking advantage of the ubiquity of the Internet to build their own direct-access client portals, first for FX instruments, then for rates products. Initially these were almost always specific to a single group of products (with one for FX spot and forwards, another for FX options, another for vanilla swaps, etc).

The early single-dealer FX portals were successful well beyond the expectations of most of their providers, revealing that aggregation of quotes was in fact not a primary requirement for a large segment of the market. So important have these single-bank trading portals become that today there is virtually no serious FX dealer that does not offer one. Around 25% of all electronic FX trading now goes through them, outstripping the flow through MDPs by a factor of almost two to one (Fig 1). They now represent a source of FX flow large enough to be tracked explicitly by central banks.

 

The remaining e-FX flow in Fig 1 goes through e-brokerage channels such as EBS and Reuters.

Single-dealer overtakes multi-dealer

The dominance of single-dealer portals over MDPs is even more pronounced for FX forwards and swaps, for which (according to the Bank of England) single-dealer portal flow in London is currently more than three times that through MDPs.

Fig 2 shows the result of a recent survey that asked 4,300 users globally whether they preferred single- or multi-dealer systems for FX trading. It is clear that single-dealer offerings are winning this contest.

Beyond cash FX, there has been rapid growth in single-dealer portals offering trading of FX options, and, more recently, fixed income.
For example, at Deutsche Bank, interest rate swap trading began taking off on the Autobahn platform at the end of the summer in 2007; between 10% and 30% of this dealing is now done online, according to Rhomaios Ram, global head of e-commerce sales at Deutsche. "Up to 85% of US Treasuries and 75% of European government bonds are traded online," he says ^[1].

Why do users choose single-dealer over multi-dealer? The precise reasons cited for this vary from one user segment to another, but all revolve around the better service that can be obtained via a single-dealer offering.

The 2009 SIFMA buy-side e-trading survey shows depth of liquidity and speed of execution as the leading considerations when choosing a fixed income trading platform (Fig 3).

Credit trading remains a laggard in this area. "Credit is very different because it is more of a consultative sale. Secondary business often opens the way for new issues, and one single-A financial is not the same as another. Dynamic pricing engines are therefore far more challenging, resulting in a greater degree of manual quoting," says Michael Noto, director of US BARX sales at Barclays Capital in New York^[2].

Nevertheless, the first credit default swap online trading offerings are now starting to appear, in anticipation of regulatory and market structure changes that promise to make this market more transparent.

The dawn of the Single-Dealer Platform

Trading portals offered by major banks have tended to proliferate in an unstructured way over the last decade, without much central coordination. One major European bank that recently carried out an audit of its trading portals discovered that it had no fewer than 106 separate online trading systems, with almost no integration between most of them.

This fragmentation is not satisfactory (Fig 4). It forces the client to log into many different screens simultaneously to view research, view prices and execute trades, with different parts of a banks' offerings typically being provided via separate and incompatible portals. And it gives the bank many uncoordinated points of contact with each client, hindering the gathering of intelligence about the client and limiting the value that can be delivered.

As a result, the major trend of the last two years has been away from multiple single-dealer portals towards the Single-Dealer Platform: a software layer that allows a bank to combine pricing, research, commentary and execution capabilities from many different parts of the enterprise, and present them to the client in a seamless form tailored to that client's individual workflow.

A Single Dealer Platform (SDP) supersedes this mix of uncoordinated delivery channels to provide a coherent, managed point of contact between a bank and its customers.

This is not to say, of course, that SDPs supersede MDPs: they coexist with them, and probably always will. Rather, SDPs allow banks to use consistent and cost-effective technology to combine and coordinate their direct e-trading offerings across the enterprise.

Why go SDP?

The benefits of an SDP over an uncoordinated collection of portals are clear. But what are the benefits for a bank of building an SDP over channelling flow through MDPs? Those most often cited include:

• Increased deal-flow, thanks to targeted offerings and ownership of the customer desktop
• Better links between market makers and remote sales traders across a banking group, increasing efficiency and promoting internalization of group-wide flow
• Improved margins, thanks to precision per-client pricing
• Stronger client relationships, through differentiation and value-add
• Reduced transaction costs, through enhanced efficiency and reduced brokerage fees
• Better monitoring of client behaviour
• Escape from commoditization

There is a useful analogy here with the travel industry. Many travelers go to Expedia or Travelocity to check out the flight times from many airlines, compare prices and book flights. In this way Expedia is like an MDP, functioning as an aggregator and intermediary between the airlines and their customers.
But the fact that you can book via these aggregators does not mean that the airlines don't offer their own single-airline booking sites. Quite the reverse: it spurs airlines to offer ever more creative inducements to travelers to book direct. These inducements include more flexible booking, seat selection, meal selection, redemption of air miles, partner programs, and so on.

A similar situation exists in the capital markets. As long as a buy-side client trades with a bank via an MDP, that bank is only as good as its last quote. In other words, its trading service has become a commodity, valued on price alone. Only by providing a differentiated online trading service that really adds value for clients can the bank defend its margins and its franchise.

So, all else being equal, there is little question that a bank will prefer a client to trade via an SDP than via an MDP if this can be achieved. But what's in it for the client?

The simple answer is: quality of service. Inevitably, the pricing and execution services offered by an MDP tend to be only those that most of the participating dealers can support. Innovations in areas such as speed, trading workflow and pre-trade integration are unlikely to be passed on by an MDP until most of the other participating banks offer them too, by which time the competitive advantage has vanished.

Thus, for the client, the SDP provides an opportunity for to obtain a far better trading experience - richer, faster, easier and more certain. It is up to the bank to make sure that it delivers this value to customers in its target segments, and does so better than anyone else. And in order to achieve this, one more step is needed in the evolution of the SDP. Having brought together all the bank's content on a single platform, it is now necessary to segregate both content and delivery so as to provide the best possible experience for users in each target segment. (Fig 6).

This targeted, highly differentiated approach is relatively new, and only a few banks have so far made real progress in this direction. But it is clearly the coming trend.

SDPs and the Internet

The Internet is the natural communication medium for SDPs, and is so well suited to the role that the use of private lines for this purpose has almost disappeared. What little disadvantage the public Internet may have in latency and predictability (and it's marginal) it more than makes up for in ubiquity, scalability and ease of use.

The technology that is used to provide the user experience (UX) on the client desktop is currently undergoing a major change, however.
In the early days, the UX was almost always implemented either as a Java applet within a Web browser, or as a traditional Windows installed application.
However, when Sun Microsystems won its ultimately Pyrrhic victory over Microsoft, forcing it to remove Microsoft Java technology from its browser, the Sun technology that replaced it was too unmanageable for serious trading applications. Over the last few years the use of Java applets (as opposed to installed applications) has dwindled, and is likely to disappear altogether before too long.

It has been replaced by two alternative approaches, depending on the needs of the market segment being served.

The first is the installed desktop application, suitable for small groups of users for whom performance and reliability is paramount, and where scalability, deployment and support costs are not an issue. Nowadays this is almost always written in one of two technologies: installed Java, or .NET.

The second is the Rich Internet Application, or RIA. This is the fastest growing technology option, and the one that is almost certain to predominate over the next few years. RIAs are pieces of software that behave like applications but load like Web pages. As Java applets have declined in use, three new technologies have come to the fore: Ajax (a public standard, supported by all browser manufacturers), Flex (proprietary to Adobe) and Silverlight (proprietary to Microsoft).
The great merit of RIAs is ease of deployment and support, which is a major consideration when the target market is large. They also make other content very easy to integrate, and allow a bank to add new content and functionality instantly at any time rather than having to wait for an application upgrade. The performance and functionality of RIAs is rapidly increasing to the point where they are comparable with installed applications, at which point the above considerations dominate.
There is a lot to say about the relative strengths and weaknesses of these technologies, and this is the subject of a forthcoming White Paper from Caplin Systems.

How to build an SDP

There are three major options open to any bank setting out to create an SDP.

• Use a standard white label offering hosted by a vendor
• Build it from scratch, using low-level generic components
• Build it on an SDP framework

The first option is of course the fastest and, usually, least expensive to implement (although often costly in the long term). However it has the disadvantage of being a generic solution that is not under the bank's control and that is likely to offer little flexibility or scope for individuality, which seriously undermines one of the key reasons for building an SDP in the first place.

Further, at the time of writing, the providers of white label solutions are all primarily focussed on a single asset class and provide little if any support for cross-asset trading. This effectively represents a retreat into the pre-SDP fragmented approach.

The second option - build from scratch - has the attraction that it provides a high degree of freedom, which can translate differentiation in the final offering. However, it tends to be costly, time-consuming and risky compared to the other approaches.  Because building from scratch requires so much of the budget to be spent on creating the standard features (business logic, real time permissioning, trading workflow, pricing management and so on) that all SDPs require, it often leaves little time or budget over to add value and create something unique and differentiated.

Building from scratch also results, by definition, in a large amount of bespoke software that has no third party support, and requires a large development team. If this team is kept together once the project is complete, this translates into very high cost of ownership; if not, support and further development capability can be severely compromised.

Using an SDP framework potentially delivers the best of both worlds. It can dramatically reduce costs and time-to-market, because most of the heavy lifting has already been done, and support is provided by the vendor. At the same time it increases differentiation by freeing you to focus your resources on adding value rather than building the foundations.

A good SDP framework will provide these foundations and a flexible set of core domain models. If properly designed, it will greatly accelerate the solution without restricting what can be done or reducing the scope for differentiation.

Where next?

Single-Dealer Platforms are evolving very quickly, and the next few years are going be a time of exciting innovation and rapid development.
Focussing on user needs and preferences is key. The banks that succeed will be those that successfully address an unmet want in the market for greater speed, certainty, convenience or sophistication of trading.

Some key areas of active innovation in this area are:

• Developing an ever-more intuitive and compelling user experience (UX)
• Integrating pre-trade information and execution more effectively, to reflect user workflows and stimulate trading
• Providing streaming executable prices with one-click trading for all liquid assets
• Improving post-trade integration in areas such as reconciliation and real-time risk management
• Adding more integrated and powerful analytics
• Adding newly-electronic asset classes such as credit and commodities, and exchange-based products such as equities and futures
• Adding support for multi-asset trading - for example, allowing a basis trade of a bond against the bond future, and allowing FX hedging of fixed income trades

At Caplin we are fortunate to be working with several banks that are leading the way in online e-commerce, and are constantly working to raise the bar and redefine best practice in this fast-moving area.

Summary

In the early days of OTC electronic trading, multi-dealer platforms were the primary link between dealers and their clients. But now single-dealer portals have decisively overtaken MDPs as the most popular channel to market in FX trading, and appear set to do so in the fixed income market.
This is because higher levels of service can be provided via this route, and because it is more efficient for the banks.

Following the proliferation of single-dealer portals, banks are now merging multiple portals across products and asset classes into a single-dealer platform.  This allows a bank to unify, simplify and improve its e-trading delivery, usually using browser-based technology.

Building a successful SDP can give a bank a major competitive trading advantage, but requires a good understanding of both user needs and technology options.
When building an SDP, using an SDP framework typically provides the best mix of quality, efficiency, flexibility, risk and time to market, while allowing a bank to create a highly differentiated, unique offering.

The winners will be those banks that focus on providing unique and valuable solutions to user requirements in their target segments, while making the right technology decisions.

This is a vital new frontier, and the race has started.

Footnotes and references