I attended Intel's FasterRisk in London on June 21 at the invitation of SAS's technology partner Hewlett
Packard (HP). The event was hosted by Nigel Woodward, a former GEIS colleague, now Global Director
Financial Services ESS at Intel. Nigel reminded me of our time at GEIS Global Financial Services and our
many discussions of the issues in risk with banks in the '80s and '90s - we were delivering one of the
earliest global risk systems, GLS (Global Limits System). Competition for GEIS in those early days of
limits and exposure management was centred on the Toronto-based firm, IP Sharp. The clock has
moved forward 25 years and Canada's risk management expertise is now well-represented with SAS
RiskAdvisory and a number of other leading risk management firms across the country.
What has also gone unchanged during those 20-plus years is the need to constantly refine your
organization' risk measurements and the underlying data from which decisions in risk are made. In the
'80s and early '90s, GLS and its successor, RXM (Risk and eXposure Management), offered banks the
ability to aggregate their trading positions globally and check against limits in real time. However the
measures used were simplistic in comparison to the pricing and risk models that are deployed today.
With the drive for increased sophistication in risk measurement and the need to bring products to
market more quickly, the introduction of many standalone front and middle office systems started to
break up the unifying, simple consolidated global system. As a stepping stone to a more flexible
reporting approach many firms pushed their GLS/RXM risk data at end of day, or in some cases,
operated a real time feed to a reporting environment. The volume of data and the enrichment
processes required tactical decisions on where to focus the risk analysis, using expert human
judgement and requiring design decisions very early in the configuration process that reduced the
potential for analysing the data in a dynamic and granular manner. Often, a relatively simple query
could take hours to run and made analysis a time consuming process.
With significant advances in high performance computing (HPC), especially around in-memory
processing, the potential to turn numerous and complex sources of data into risk insight has the
possibility to become reality. By efficiently distributing data across multi-threaded processes on multi-
core servers, all valuation information has high availability. This yields the additional benefit of real-
time OLAP querying, whereby the answer to every question is calculated on demand. Gridding and
threading are powerful concepts and have been exploited in the past to efficiently solve large
computational problems. SAS has taken the next step and brought both of these concepts together,
harnessing the power of the latest available industry-standard hardware to allow true distributed
computation; distributed data, distributed results and distributed analysis objects.
There are steps to making the vision of an optimised risk management framework a reality. There are
very few green field sites within which to take an academic approach to improve risk management,
but by taking a well-defined risk framework through the qualitative risk measurement disciplines,
regulatory and business information needs can be met. There may still be a few organisational and
human resource questions to be answered - the structural complexity of an organisation with all its
moving parts may result in a challenging change management and implementation exercises - but that
is the nature of doing business.
At the beginning of the evening Nigel noted to the audience that he was planning to focus on the data
issue in risk, as without good quality data the resulting risk calculations would be flawed. Many in the
room represented businesses and vendors where trading and millisecond response times seem at
odds with the typical end-of-day batch approach to risk. The comments from panellists tended to look
toward the newer technologies to solve the data management questions (which included acquiring
unstructured data) and agreed that the results from analytical processes needed to become simpler to
consume and more user friendly to the business.
For my part, I provided an overview of how a well-defined risk framework, supported by the new
capabilities in HPC, offers the potential for risk decision making on large data sets in a near real time
environment a reality. Taking the "rethinking risk" lead from SAS' white paper Evolving from
Quantitative Risk Management to a High-Performance Risk Management Analytic Framework by
Myron S. Scholes, PhD and Nobel Laureate, and Michael Stefanick, Leader of Global Risk Practice at
SAS, I highlighted that the tangible benefits from a high performance computing risk process should
include:
1. A better ability to react to various market shock events with more precision, and potentially
take advantage of the market shocks and look for better arbitrage opportunities.
2. Provision of a more structured decision capability that will allow a firm to more consistently
evaluate its product structure and expected return against inflections in the market.
3. Offers the firm an analytical framework that will provide a consistent process for making
decisions with the capital optimization/option analysis framework.
4. Expert judgment for principles is augmented by a very sophisticated and analytical view of
business decisions across a very large combination of information.
5. A high-performance risk management framework that will provide a complete picture of
firmwide risk.
By implementing a more inclusive and robust framework that incorporates the various dimensions of
risk, capital, market and liquidity, the firm will be able to not only react to various market shock events
with more precision, but also may be able to better take advantage of the market shocks and look for
better arbitrage opportunities.
Finally, by implementing a combined change in risk-reward methodology along with a high-
performance computational environment as described in the SAS white paper, executive
management can achieve a dynamic view of firmwide risk and a realistic measure of capital
performance. Giving executive management the ability to determine on demand the exposures and
potential impact to market conditions (based on current market conditions) is a requirement for
competitive advantage. As the impact of capital returning to the market is assessed, along with the
constraints on the ability to access capital, leading firms will be strongly considering implementing a
high-performance risk management framework that will provide a complete picture of firmwide risk.
To download the SAS/Scholes white paper and review other risk management topics, including the
latest Economist Intelligence Unit Enterprise Risk Management research briefing and an article by
Keith Collins, SAS CTO, covering the use of HPC technologies please visit the SAS Risk Management
Knowledge Exchange at http://www.sas.com/knowledge-exchange/risk/index.html
Thanks again to Nigel and HP for inviting me to speak at an excellent event and
venue.
David Rogers
Global Product Marketing Manager - Risk ¦ Worldwide Alliances and Product Marketing
Tel:: +44 (0)1628 486 933 ¦ Mobile: +44 (0) 7918 706535 ¦ David.Rogers@suk.sas.com
SAS UK ¦ Wittington House ¦ Henley Road ¦ Marlow SL7 2EB
www.sas.com
SAS(r)... THE POWER TO KNOW