About Us | Search | FAQ | Contact Us
ISSN No:14Enterprise Risk Management
The bank of the future
The efficient bank of the future will be driven by a single, consistent, analytical engine, This engine will power front, middle and back-office functions as well as supply information on enterprise-wide risk. The ability to control and manage risk will undergo a vast improvement.
Most banks resemble a pawnshop with software in all shapes and sizes, inconsistent data formats, inconsistent pricing models, inconsistent market assumptions etc. Indeed, in one bank I visited recently, there were no less than 70 different systems in treasury alone!
With the appropriate technology in place, financial trading will move from the current vertical, product-oriented environment (e.g. Swaps, FX, Equities) to a horizontal, customer-oriented environment in which complex combinations of instrument types will be traded. There will be less need for desks specialising in a single product line. As customised trading portals replace product-oriented trading desks, the focus will shift to customer needs and not instrument type. The management of trading limits will be based on capital, set in such a manner as to maximise the risk-adjusted return on capital for the firm. Traders will be remunerated on their risk-adjusted earnings and not on earnings alone, orienting them in a manner consistent with the firm's goals. A single analytical engine providing consistency between the strategic and tactical level will value all transactions. The firm's exposure will be known and distributed in almost real time. Evaluating the risk of a specific deal will take into account its effect on the firm's total risk exposure, rather than the exposure of the individual deal alone.
Banks that dominate this technology will have a tremendous competitive advantage. Their information technology and trading infrastructure will be cheaper than today's by several orders of magnitude. Some part, the "vanilla", will be outsourced and some, the part essential to maintaining a competitive advantage, will reside in-house. Conversely, banks that attempt to build and run all of this infrastructure in-house will be bogged down in a quagmire of large, expensive IT departments and poorly-supported software.
The successful banks will utilise available software components wherever possible, requiring far fewer systems, most of which will be based on industry standard, re-useable, robust software. More importantly, they will be free to focus resources and capital on their core business and offer services and products more directly suited to their customers' needs. Pricing of transactions will adequately reflect the marginal risk to the firm. Less in-house development will occur - after all, how many banks today design and build databases? Not so long ago, many did.
There is a revolution occurring in Enterprise-Risk Systems that is likely to affect the entire architecture of a capital markets environment and potentially the entire banking system as well. The leading enterprise risk system is now able to price and simulate the risk of banks made up of all the fixed income, money market, derivative, OTC and equity contracts traded in more than 20 of the largest markets world-wide. It can calibrate models to many different pricing standards. It has mapping tools, scenario generation engines and programming interfaces for adding proprietary models and model libraries. It can link the market, credit, liquidity and ALM functions of a bank into one consistent framework. It can simultaneously deal with mark-to-market and accrual-based analysis. It can simulate the risk of dynamically changing portfolios in a dynamically changing market. It can evaluate the riskiness of strategies. It can trade-off the risk and return of arbitrary portfolios, including derivatives and credit portfolios.
So, it is now technically feasible to support the various risk functions, market, credit, liquidity, ALM, in one system. There is no good reason therefore, other than historical, for all of these functions to be separate entities within a bank. Integrating these functions into one entity will improve the quality of risk management and make capital allocation far more effective. The cost implications of a single risk entity, based on a single risk architecture, versus at least four entities and systems that operate in an inconsistent manner, with different support teams, different vendors, different upgrading frequencies and different hardware, are enormous!
As senior officers in the bank become aware of this fundamental shift, these individual distinct business units will disappear, if for no other reason than cost. This is now happening in leading edge institutions, and it will eventually eliminate more than half of the cost of support in today's risk management functions.
Even more exciting is the fact that more and more elements of the banking book are now being traded and, as a result, the distinction between capital markets trading and the traditional loan book, is fast disappearing. There is enormous pressure on banks to cut costs, particularly in light of the wave of bank mergers. A natural place to cut costs therefore will be in the overall risk systems support. In a medium to large global bank, expenditure on risk support on these four functions can amount to over 150 million dollars annually. A savings of up to 90 million dollars or more annually in risk support costs is a good target to aim for, especially when the functions for market, credit, liquidity risk and ALM are combined into one entity. With such enormous savings and the potential for a quantum leap in the quality of risk management, it is likely that we will see these functions aggregated and risk systems architecture unified in an accelerated manner.
Further down the line, even more savings are possible. In fact, the cost savings achievable will dominate the ones we have addressed so far. Where will these additional savings come from?
The typical medium to large bank supports a huge number of distinct legacy systems to deal with the requirements of Capital Markets and retail banking functions. In Algorithmics' client base we have observed banks with over one hundred different environments, running on a plethora of hardware, based on inconsistent data models, with almost no standards to rely upon. We believe that all of this will be eventually be replaced by a single, unified environment with a risk engine at the core of the bank and not as an adjunct, as it is today. The risk engine will support the front and middle office, and support the bank's e-commerce initiatives. With a core risk engine covering all products, across all markets, linked to a unified back office processing backbone and a unified position-keeping framework, there will be no need for the heavy front office applications we see proliferating today.
Probably one of the most important side benefits of the pace of e-commerce is that it has forced banks to deal with the outsourcing of their IT function, since it has been impossible to build the needed systems in house at the pace required. In many instances this outsourcing, for the most part, has been successful and has opened the eyes of management to the benefits of such an approach. Banks are more willing to consider outsourcing the risk measurement function or the software development required to support the risk function.
Algorithmics' biggest competitor of the past - a bank's in-house development - is now losing the battle as banks realise the benefits of rationalising their risk support environment. So, thanks to mergers and e-commerce, banks will begin to attack the very large costs that were assumed as IT departments grew over the 70s, 80s and 90s to support the every growing complexity of trading.
The market is now seeing the development of an ASP (Application Service Provider) model for transaction processing. It is quite telling when one observes a complex bank such as J.P. Morgan outsourcing their entire back office processing to EDS. This sets the pace for the industry. It is only a matter of time before we will see banks outsourcing enterprise risk processing functions as well.
Mark-to-Future(HYPERLINK "http://www.mark-to-future.com" www.mark-to-future.com) is a standard framework for risk and reward that will make outsourcing of risk a reality. Without Mark-to-Future, distributing the processing of risk, without losing any accuracy in the process, was not feasible. For that matter, no other standard that has been proposed can claim to link market, credit, liquidity and ALM risk functions. The linearity of a Mark-to-Future simulation makes distributed processing natural. The focus on instrument simulation and not portfolios breaks down a major barrier to the risk service business over the net, since risk services can be offered without the client ever disclosing its holdings! The fact that one simulated Mark-to-Future cube, or database, can be used to process the reports of thousands of portfolios without any further simulation makes the computational burden feasible and the performance acceptable. It also facilitates BtoB and BtoC risk outsourcing from banks to their clients.
In a discussion with a banking regulator recently, I heard that some regulators feel that one of the biggest barriers to risk management practice improving is the lack of standards. This is true particularly for emerging markets or small, mathematically-unsophisticated financial institutions, where the typical bank's infrastructure does not allow for the huge expense implied by state-of-the art risk systems.
But it's not a trivial task. A system to continuously monitor exposures across an entire organisation, with an integrated view of all risks - accessible not only by the risk control department, but also by the front office, senior managers and, increasingly, customers as well - cannot be delivered by today's rigid and fragmented world without standards. The key enabler in the new economy is standards.
"We are striving for a framework whose underling strategies may remain relatively fixed, but within which changes in application can be made as both bankers and supervisors learn more. It is the framework we must get right we must avoid formulaic approaches"
Alan Greenspan, October 1999
Greenspan is right. Standards based on formulas will not work, because formulas imply assumptions, which are invalidated as soon as the world changes. We need standards based on a framework that can adapt as the world changes, without re-architecting the entire bank's IT infrastructure. A bank that wishes to evolve at e-speed will need to adopt standards soon. Mark-to-Future is the first and only standard framework for risk and reward available today. More than 140 installations world-wide, implicitly or explicitly, now conform to this standard.
We are on the cusp of a major change in banks and a race to prepare for the e-commerce world. The face of risk management in banking will change radically. The good news is that this will lead to far more cost-effective risk management, which must inevitably result in more efficient and consistent pricing. It will also lead to a levelling of the playing field, since Mark-to-Future will enable at least some portion of the risk function in a bank to be commoditised. Small and big institutions will have access to equally good risk measurement and practice. Unified trading systems will appear. The quality of risk management will improve system wide and it will be commonplace to see risk measurement conducted over the Net in the bank of the future.
© 2000 Algorithmics Incorporated
|Home | About Us | Search | FAQ | Contact Us|