Exchanges: a changing world
A quick glance through the 2001 edition of the Handbook of World Stock, Derivative & Commodity Exchanges would suffice to convince anyone, even new to the business, that the world of exchanges is changing, and dramatically so.
Here is a key quote from Martin Scullion, who says in 'Demutualisation: The challenges facing global exchanges':
"Very few exchanges would disagree with the fact that the current competitive climate has never been so intense:
- The barriers to entry for new entrants have been slowly eroded through market pressures and increasing dissatisfaction with the way exchanges have been managed.
- Stakeholders can be easily raided.
- New electronic alternate trading systems seem to be mushrooming.
- Fragmentation in an un-level playing field is rife.
- There will be changes in price discovery forums in response to systemic risk management via central Counterparties."
The problems often mentioned include diminishing margins, disintermediation, and the terrible choice between cannibalisation and being eaten. Today's exchange environment is clearly dynamic and a great concern to those operating and trying to survive in it.
A problem with technological roots
It is also widely recognised (see, e.g., Patrick Young & Thomas Theys, Capital Market Revolution; Prentice Hall, 1999), that the problems the exchanges are now facing have technological roots. No explosion of offshore entities or Electronic Communication Network (ECN) could exist without the new generations of trading systems. There is no better way to reach end-customers than with the public Internet. However, it is not always as easy to do that in emerging economies like India and China. This lack of easy (and open) affordable mass access limits rapid growth of the number of players. Those that do participate in the market will require secure networked environments that include robust wireless multi device communications. The linkages between the economic and technical issues related to exchanges and trading markets are deep. For instance, who would believe that it is pure coincidence that the first exchange to demutualise (the ASX -- Australian Stock Exchange) would also be the first one to launch an all electronic market?
The problems created for exchanges by technology are heightened by the fact that regulatory bodies, together with market pressure, drive exchanges to continually improve performance and hence their efficiency. This pressure provides systemic benefits to the participants such as T+1 settlement cycles, more open, fair and transparent environments, as well as better public disclosure of market information. It is not surprising that some analysts see the problems the exchanges now have from an optimistic viewpoint. They tell us that this technology-led crisis in the market underscores the drive to achieve the three-pronged holy grail of capital markets: More Liquidity! More Accessibility! More Transparency! Young and Theys say that these elements are what the markets want and that technology will make the delivery of them possible.
However, not all exchanges will survive this transition. Besides major players, niche markets have better odds of survival -- especially those which are global in reach, even if rather focused on the list of things they trade. If existing markets cannot provide (or afford to deliver) the liquidity, accessibility and transparency that participants need, they will lose trading volumes to those that can provide these features. We can conclude that technology, by itself and by triggering a variety of mechanisms, has pushed exchanges to global rather than local market strategies.
But will technology allow the exchanges to operate properly in such new conditions? This question of how technology will take on the challenge of global trading systems has been raised by the environment the exchanges operate in. The more basic question of whether technology will be able to take on this challenge is by no means trivial. The issue becomes one of addressing the many competing business requirements that will make liquidity globalisation happen. We often hear expert Information Technology (IT) architects say 'all these requirements cannot be jointly delivered'. This competition of needs and requirements forces impossible choices, such as 'either you get your business process, or you get security', or 'do you want it to work, or do you want it to scale?' The answer of course is that you want it all and in fact you need it all. In the end all will be delivered as technology advances. However, the journey getting there may be very rough on some participants.
From discussing the challenge, to taking it on
To this point our discussion has been primarily at the level of generalities. We will now get more IBM-specific. Clearly, overall problems such as 'The challenge of global trading systems'. are, as a rule, too big for solutions to be described to their entirety in the limited scope of an essay like this one. This is why we will only show how to solve some very specific but significant, hard core embodiments of the challenge. We will also invite the reader to contemplate with us selected parts of the broad horizon of our technological future.
One main point we want to make is that a broad enough technological agenda, if targeted at helping the markets, can provide means to address the main challenge of global trading systems. This is part of a larger scale phenomenon that transcends any special line of business, across all industries: we are, whether we like it or not, in times when technology changes define new business rules instead of causing only quantitative changes to the way business is made. Rather than 'solutions jump' at problems, a visit to some 'elements', technological wonders out of the IBM Research Labs, should stimulate the imagination.
A few words on our Labs are in order. With about 60,000 people in R&D, including 3,000 in the Research Division, IBM has been an industry technical leader for many years: altogether, a web of R&D that produced 3,411 US Patents in 2001. What is less well known is that IBM increasingly focuses on business issues, by endeavouring to solve the major problems faced by our customers and the markets in general, and by combining its own inventiveness with that of its business partners. In a nutshell, if you tell IBM your business problem, IBM Research and its full R&D strength can be brought to bear to help you decide whether this problem has a technological or technology-based solution. If such a solution has not yet been invented but can be assessed to possibly exist without violating any law of nature, well, Research will try very hard to invent it for you.
Focusing on three facets of the challenge
There are three facets to the exchanges/markets challenge that we will address:
- Facing diminishing margins by creating liquidity with electronic brokers
- Creating markets that are both fair and scalable
- Operational resilience (OR).
We will first describe technology elements generated by the Research Labs. Then we will briefly show how some of the elements combine to tackle the three embodiments of the challenge we have singled out. Of these the third one is at a much larger scale and we will discuss it in more detail.
Some technological elements from IBM Research Labs
MQ WEB scale
This is a highly performant, web-oriented, security compatible, content based Pub-Sub messaging system, so performant indeed that it plays a crucial role in the nervous systems of IBM Research's most futuristic autonomic computing projects. Everyone knows what performance and security mean, and why they are essential for a web-oriented messaging system (and who needs justification for making technologies web-compatible). As for 'content based', this replaces the more traditional subject based matching rule. To match a message that is published with a subscriber, one can now go more deeply in the body of the message. For instance, a subscriber can ask for quotes on corporation xxx only on bundles of some arbitrary size, without being limited to a pre-selected set of sizes, with no need to reorganise the nomenclature of what are the permissible matches. Much more flexibility is brought to the matching engine, a fundamental step to elementary intelligence, and a fundamental building block in the architecture for autonomic computing.
4758 PCI Cryptographic Coprocessor
This tamper-sensitive, tamper-responding, tamper-evident, field programmable cryptographic coprocessor protects programs and cryptographic keys against any known attack, with the best certification [FIPS 140-1 overall level 4 certified (hardware and microcode)]. It is indeed unique among programmable cryptocards. We call it the 4758 for short as it will play a big role in discussions to follow later.
Quantum cryptography
This is working already, for instance on a mile-long line at the IBM Almaden Research Labs and at CERN, putting the law of fundamental physics alongside cryptography in charge of protecting secrets. IBM continues to invest in quantum computing research. Quantum computing, if successful, will break cryptography as we know it and can provide the next generation of protection. We also continue to push the envelope of more traditional security technologies.
SASH
SASH is the research name of a web-plication (or web-oriented application development framework. SASH makes it possible to build web-plications that enable the content riches of the WWW to be combined with the ease of use of local handling of folders, files or data on your laptop or other work station.
Audio visual speech recognition
That is lip reading to compensate for high noise, such as found near machinery or on an airport runway. You may also think 'trading room'.
Video streaming
A move toward using all that counts in decision making. Images will be part of the input; no one has any doubt about that. But how you do the indexing, how do you retrieve and distribute? Early usage of such technology includes easy and affordable employee training, which can be deployed across wide geographies if needed.
Pervasive computing
IT will be everywhere, even hidden in places where it cannot be recognised. Hot topics include wireless transaction (a security challenge of the kind IBM Research loves), location-based services, database synchronisation (as one fundamental piece of seamless multi-channel access), and more.
E-Liza
This is IBM's own version of Autonomic Computing (AC) with self-diagnosing, self-optimising, self-protecting, self-healing systems as exemplified by the IBM z-Series servers, and much more to come. This could erroneously be though of as fault tolerant computing under a new guise -- but there is much more than fault tolerance to be achieved. Autonomic computing aims, maybe more than anything else, at controlling the explosive complexity of IT systems. We urgently need to control the costs of maintenance and operation which combine to form an ever growing portion of the total cost of ownership: outsourcing won't do per se, as you need to believe the outsourcer will be able to deliver, and keep up with progress (thus the outsourcer at least must master AC). Just to mention one angle, there is a big difference between fault tolerance and operationally resilient fault tolerance. After all, OR is essentially AC enlarged to contain the human and business layers, and contemplated from the user's point of view.
Deep computing (and deep mathematics)
Good balance needs three pillars. So does computing, to tackle ever harder computational challenges, such as continual optimisation or big portfolio risk assessment. These three pillars are good computers (think cooking ware), good algorithms (think recipes), and the art to execute the best recipes with the best cooking tools.
Services where a lot happens
An ever growing part of IBM is IGS (IBM Global Services: the people whose pictures you see in most IBM ads) and not surprisingly, if you think about what efficient services now entails, an ever bigger part of the research effort goes into helping IGS. This is so big an area that no short description could do it justice. Imagining how next generation services will interact with the other elements we have cited is a fun game we leave to the reader.
Solutions to embodiments of the challenge of global trading systems
Facing diminishing margins, and creating liquidity with electronic brokers
This may seem too good to be true, but it is made possible by combining the 4758 with MQ Web-scale -- all dressed up with some basic cryptography, matching engines, salted and peppered with know how and imagination.
Creating markets that are both fair and scalable
Using MQ Web-scale, basic security technologies (time stamps, non repudiation, ...), rule engines, and matching methodologies makes it possible to create efficient, fair, and scalable call auction markets. No need any more to delay artificially the lines to the customer who are the closest, and making it possible to provide for regulatory requirement while using the best of technology, instead of going against progress.
Operational resilience
This has become a pressing concern after the September 11, 2001 tragedy. Just a few points about OR:
The fractal nature of the statistics of Operational Risk demands redundancies at all scales to control and mitigate the risk. It will be crucial to limit the need for Operational Risk (Op Risk) insurance by using more resilient infrastructures that not only protect better -- from the micro scale of what is inside the workstation, to the corporate network and the inter-firm global infrastructure -- but also allows for more adaptability. Of course, IBM Labs will provide full scale help to IBM's Financial Services Sector and the rest of the corporation to articulate and deliver OR, using skills ranging from core microelectronics to statistics and risk modeling and simulation, through the obvious classical IBM ability to deliver reliable, secure and scalable systems. Because of the deep relationship between OR and AC, and because AC is obviously a tremendous challenge, serious scientific and technological problems will need to be solved to allow successive generations of operationally resilient infrastructures.
Confidential data collection for analysis, modeling, and other aspects of cost saving and efficiency. One main issue in Op Risk is the difficulty in building models. Not only do the relevant probability distributions have fat tails, but the data are scarce because they are confidential. IBM has invented means to use the 4758 to create databases that could hold secrets while allowing performing analysis that are crucial to the industry. We have represented the main architecture component of this system which applies to many privacy and confidentiality issues.
Be sure the insurer can pay; be sure who guarantees services, can serve. Everyone wants to trust the insurer to pay to cover the loss: you need to believe that. Similarly, whoever you rely on to deliver the operations must be trustworthy, especially in hard times. Your strategic outsourcing partner had better be the one offering optimal Operational Resilience; you can guess what we think this implies.
) |
For more information on the ideas and technologies in this article or on any IBM offering, go to www.ibm.com/solutions; or, for a personal response, call or e-mail:
Peter R Bobris, Global Head eMarkets Infrastructure, +44-207-202-6258 peterbobris@uk.ibm.com
Francis Lacan, IBM Financial Markets Strategist, +44-207-202-3019 francis_lacan@uk.ibm.com
Dr Charles Tresser, TJ Watson Research Center,+1-914-714 -5857 tresser@us.ibm.com
Please note that patents are pending on many of the ideas and technologies described in this article.
