Electronic Market Microstructure

Can banks provide cross-product electronic trading applications to their front office, proprietary desks, trading partners, and prime brokerages. Several years ago one bank developed a strategy just to do that. Their strategy seemed sound at the time; they would conduct a two phased approach, the first would be to consolidated varying instruments into one seamless front-end that had a single-sign-on to all the different systems that supported the varying instruments, fully electronic or not (Meaning some systems eventually passed off the transaction to people to process or execute). To the user it would smell and feel like a cross-product application; it had a nice interface that had all the asset classes needed to service one customers needs, but that was it. When it came to execution it was very disjointed and complex. Trade confirms came back to the front-end in all sorts of methods and formats from all the varying systems. Confirms were sent directly by some systems and others would be received by a physically phone call. There was no concept of cross-product accounts, so there where no concept of using cross-leverage or credit netting. It was a mirage.

The second phase of this project was to fully integrate the various systems using a common message bus that would formalize the transaction formats and communication methods. It was also planned to ferret out as many systems that did not lead to electronic execution as possible. This would enable them to build a “true” cross-asset application that unified risk and financial (P&L) across assets. Alas, the business dream was ahead of were the technology market was, both in terms of underlying infrastructure technology like cluster techniques and high-speed messaging, product technology for electronically trading certain instruments, and finally industry services technology such as exchanges and clearing capabilities. This was eight years ago so people now surly would understand that, at that time, what the bank was taken on was a huge undertaking. Times have changed and the technology and supporting services have changed with it.

To formulate a reasonable opinion on the probability of a useful Cross-Product Electronic Trading system being built, it may be best to first understand the lay of the land. Examining the different asset classes upon the landscape, I believe, one can see that it is clear that electronic trading has penetrated difference sectors very differently. This difference can be, typically, explained by the interaction of several factors; existing market structures, regulatory and competitive factors and the varied needs of traders. These have all affected the integration or renovation of Electronic Trading into mainstream trading. Another important element is the asset type, since standardised, homogeneous products have proved least difficult to migrate to electronic trading because they are often more liquid and demand a fast and efficient systems,. However, electronic trading is still, I believe, at only an early stage of development - The development has been very much implemented in the asset class silos, as mentioned above. To bridge those silos and create multiple assets, or Cross-product trading platforms, either there needs to be a unification of the silo work or some form of rework that uses a common architecture that will make development more robust and common.  Similarly, the direction of developments to date inevitably reflects the current technology – advances will broaden the choices available in next generation trading systems, potentially enabling further waves of change to market arrangements. Knowing all this, are we now ready for a Cross-Product Electronic Trading Platform? To help answer this, lets first look at the state of electronic trading within each asset class.

Fixed Income Electronic Trading

Electronic fixed-income trading continues to gain fast adoption in the U.S marketplace. Competition is expanding into: i) illiquid fixed-income assets; ii) the pan-European market; ii) automated trading; and iv) OTC derivatives products. In Europe there is much intreset as well, again mainly by inter-brokers. Unlike the U.S., and despite tremendous interest in the electronic trading of all these assets, partially due to globalization, the very fact that they are non-uniform make them difficult to automated, Europe is behind the U.S.in adoption. Also the lack of a pan-European clearing housing is a problem in a fulfilling a true electronic transaction. Euroclear and Clearstream were developed to fill this need but because of so much pan-European differences they have not made the impact they originally intended. Otherwise, the US market looks strong to grow and has a lot of the right market infrastructure in place to advance its fixed income market electronic trading capability, I belive it will only be time before Europe catches up.

         –Aite Group (Click here to Enlarge)

The frenzied adoption of electronic markets in the U.S. in the fixed income space is driven by both geographic and product expansion, automated trading, and improved processing services. Ifrom a product point of view applications  like Murex, Calypso and Sophis  continue to help expand both the mid-market and top tier, while both inter-dealer and multi-dealer platforms look to solidify their competitive positions by becoming cross product supported by a stable technology. These prduct applications are often choosen by banks because they are belived to have a lower cost of ownership (LCO). But beware, often firms go with an application and get to dependant on the vendor and suffer greatley for it later.

       –Aite Group (Click here to Enlarge)

In general, fixed Income products vary greatly in their liquidity and complexity. In essence, the more liquid and simple a product is the easier it is to trade electronically. In Fixed Income to date, most of the electronic trading activities have taken place in U.S. Treasuries, a liquid, well defined product. Aite Group estimates that approximately 68% of U.S. Treasuries trading was conducted electronically by the end of 2004. Even though the MBS is fairly liquid and well defined its market was a distant second, with 30% penetration. The corporate bond market currently stands at a single digit number. These instruments have been traditionally very custom, with varying types of default events and maturities, however they have gotten more standardized and now are being listed on major exchanges and ECNs like MarketAxess (and others). However, despite being listed on exchanges, the vast majority of trading volume in corporate bonds in most developed markets takes place in decentralized, dealer-based, over-the-counter markets. We will see more OTC business going to the exchanges and thus becoming electronic.

The adoption of electronic trading in the fixed income space is fragmented mainly due to the complexity of the instruments, standardization and liquidity, but I see this become less of a problem as their becomes more sophisticated industry services to trade and clear. Besides this there is still a fair amount of fixed income being traded electronically already.

Derivatives and SWAPS

I defy anyone to speak about derivatives nowadays with out mentioning credit derivatives and expect to be heard. This is not just a social trend I can assure you. The volume in the credit derivatives has overwhelmed the ability of the traditional progressing methods to handle the vast amount by the banks. With hedge funds, insurance companies, corporations and investment vehicles all concerned with debt defaults or rating downgrading rush to protect themselves. This rush is creating over $17 trillion in notional value being traded, pushing the banks to the brink of their processing capabilities. Imagine, just in simple terms, the number of ISDA agreements that are being produced daily; the mass is piling up and some firms don’t even imaged them into work-flow. The old method of voice-paper-fax-recording model of trading and settling these investments is frail and cracks are starting to be visible in the system. Also the cost of these transactions are enormous, costing $250-$900 a transaction because of the laborious transaction process.

      — Aite (Click here to enlarge)

With these types of transaction cost it begs one to put high on their priority list the task to decrease the cost ASAP by renovating the processing system to move it to a automated process as quickly as possible. And eventually, fully automated for electronic trading so that the volume of trading can scale with global demand. Yet the cost of the transaction alone is enough to precipitate the move to automation, their is also the profound presence of operational risk due to the processing being so strained. This risk appears in many aspects of the business including (i) regulatory capital relief, (ii) ability to do more business, (iii) danger of a risk event occurring causing reputation loss, and finally (iv) economic loss to a “bad trade”.

Market participants are taking an ever increasing interest in electronic trading to address all the issues we outlined above. Participants are so concerned that JPMorgan & Morgan Stanley have fast tracked a solution called TRAC-X, an indices, that allows a participant to get a broad exposure, short or long, to credit risk instead of being a counter party on an individual contract. There are other rival indices as well and they are all rushing to get electronically traded; currently some are traded on Creditex; A New York-based brokerage. Taking the indices electronic is orders of magnitude less complicated than the underlying derivative, the credit derivative. So it is invaluable alternative for a lot of participants and may be all they need for now. But being able to protect oneself from a single specific credit risk is really what most end-users will require, and constructing a electronic market is paramount for scale, risk and business opportunities.

Foreign Exchange Electronic Trading

Electronic FX trading has lifted off! Banks are concentrated on two fronts including; firstly, market-making and proprietary trading side. Determining the best way to execute in the market based on the risk they are given or positions we want to take. Secondly, for the buy side, firms are looking to determine exactly what they are looking for electronically and algorithmically and extending that functionality to them. Internally, banks are constantly looking for better ways to execute in the market and seek additional liquidity and cost efficiencies.

FX seems not to have adopted algorithmic trading to its fullest yet, but firms do use internal tools like our TWAP (trade-weighted average pricing) to determine best liquidity and execution. From the buy side, there have seen some interesting algorithmic trading from individuals that have  migrating over from the more traditional exchange based products to FX; that experience will be valuable. In essence these new comers believe electronic trading has proved itself in other markets like equities and the futures markets, and it’s really the best way to get best execution. So, from their perspective, all traders who care about their P&L and have a duty to get the best price will trade using algorithms on electronic platforms in the next few years. Another aspect of trading that a electronic renovation will advance is; High Frequency trading. Firms are looking to attain direct market access in quiet as well as fast market environments. They want liquidity, best price and best execution. But, obviously competition is heating up.

Clearly the FX market is following the steps of the equity markets with all the banks and vendors that provide platforms or services are gearing up. Some of the problems they face may include the following:

• Unlike other markets like equities and futures, the sell side in FX generally acts as counterparties instead of agents as they do in those other markets.
• FX is largely over-the-counter; there is no one single price. The volume is unknown because it is not published.
• The APIs from the different liquidity providers are proprietary and complex, there is no single standard adopted yet; it’s expensive to integrate.

Some firms believe that algorithmic trading is not ready for prime time yet in FX. There are very few players. Many firms are only interested in execution, not about alpha. Some banks say they are content taken orders buy phone and living with the antiquated processing; it’s where they make the most money.  I believe this view is short sighted and not taking into account many factors including; processing risk, reputation risk, and lack of scale for the business to name but a few.Many experts agree that in five years, most of the FX volumes traded on both the buy side and the sell side will be electronic. The market will look more and more like a traditional exchange. We are just at the beginning stages of this growth as electronic and algorithmic trading has just this year come to the forefront. Algorithmic trading in the near future will touch every piece of the FX transaction from risk/flow management, quote creation and trade execution. Some of this is being done today already and will continue to mature in the next few years.

Equities Electronic Trading

Cross Product Capability