Senior Supervisors Group on internal controls
- Source: Regulators call for substantial IT investment to combat risk deficiencies Finextra, October 22, 2009
"Financial services firms need to make substantial and sustained investments in IT infrastructure if they are to overcome severe underlying weaknesses in their risk management capabilities, according to a report by financial regulatory agencies. The Senior Supervisors Group (SSG) that comprises watchdogs from seven countries (United States, Canada, France, Germany, Japan, Switzerland, United Kingdom) says that underlying weaknesses in governance, incentive structures, information technology infrastructure and internal controls require substantial work to address.
The SSG report evaluates how weaknesses in risk management and internal controls contributed to industry distress during the financial crisis. Among other failings, it concludes that inadequate and often fragmented technological infrastructures at most firms hindered effective risk identification and measurement.
One challenge to improving risk management systems has been poor integration resulting from multiple mergers and acquisitions, says the report. One firm suggested that acquisitions over the years have produced an environment in which static data are largely disaggregated. Another firm echoed this view, reporting that certain products and lines of business have not been included in data aggregation and analysis processes. A third firm reported that having two systems for the same business results in duplication of processes.
Almost all struggled to process record-high volumes of product transactions during periods of market stress.
Many firms cited large-scale IT projects planned or under way to address these infrastructure and aggregation deficiencies.
Supervisors remain sceptical, however, noting that firms need to reexamine the priority they have traditionally given to revenue-generating businesses over reporting and control functions.
"In the past, many such projects have fallen behind schedule because of inadequate investment and resources," states the report. "In the current environment, these projects will require a significant dedication of funds, sponsorship, and commitment from the board and senior management during challenging economic times to ensure that technology platforms are constructed to handle unexpected spikes in volumes and to effectively produce aggregated data and appropriate management information for credit, liquidity, market, and other risk metrics.
Read the full report:
Download the document - 1.1 mb (PDF File)
Extreme risk events
- Source: Investors demand extreme risk protection Financial Times, December 19, 2009
Risk management theoreticians have long recognised that extreme market events are not vanishingly rare and that when the going gets tough, the similarity between returns from different asset classes goes up. The tools for incorporating an understanding of these facts into risk management have existed for some time.
Despite this, a lot of asset managers and investors failed to spot the risks they were running in the days leading up to the credit crunch and the ensuing market turmoil. Failure to understand how to use risk management tools, a reluctance to incorporate disaster scenarios into one’s world view and the inadequacy of the tools commercially available seem to have lulled investors into a false sense of security.
“It’s a function of expert risk management to know how to use the risk tools,” says Ron Pananek, head of strategy at RiskMetrics, a provider of risk management services. “A lot of users of risk management tools – I won’t say risk managers – didn’t understand how to use the breadth and depth of analysis available to them.”
Recognising the existence of extreme risk but then ignoring it is not necessarily stupid in all circumstances, according to Michael Burley, head of risk at London-based consultancy Independent Risk Monitoring .
“One could argue that almost everyone on the face of the planet is currently aware that extreme risks exist and the impact they can have. While many will have been hurt in recent years, mankind is strong enough to pick itself up, brush itself down and plough on,” he says. “So yes, the world is aware of extreme risks but unlikely to know when that next meteor is going to strike, and even less likely to sit in a bunker just in case.”
Lisa Goldberg, vice-president of MSCI Barra’s credit research group, adds: “We have been aware for ages of the multi-faceted nature of risk”. Barra’s risk modelling system is about to be upgraded with the addition of BXR, an extreme risk module.
While it may not be news that extreme risks exist, this will be the first time Barra has offered a systematic way to model portfolios taking their existence into account. The new product has been in development for several years.
BXR attempts to answer the question “what’s it really going to be like in turbulent times?” by using data showing how markets behaved historically during periods of turmoil, says Ms Goldberg. “We throw away the calm days, stand in the middle of the storm and ask ‘on average, how many houses blow down?’”.
Barra does not expect this tool to warn asset managers of an impending apocalypse, rather to offer them suggestions as to how their portfolio might behave in such a situation. “It is not a Doomsday approach,” says Ms Goldberg.
RiskMetrics offers clients a similar facility – the ability to pick out days with common characteristics (such as extreme volatility) within its set of historical data and analyse their behaviour.
Risk management in asset management now routinely includes some less high-tech techniques that many commenters welcome as being less of a box-ticking exercise. Stress testing, where a negative scenario is suggested and the likely impact on a portfolio calculated is all the rage.
This concept requires users to ask themselves what a negative scenario might look like. Consultant Watson Wyatt has come up with a list of 15 possible sources of extreme risk, from economic depression to war or a killer pandemic.
The trouble with this approach is that there is no guarantee these events would happen independently.
As Shakespeare said, “when sorrows come, they come not single spies but in battalions”. A global depression could easily lead to sovereign defaults, while the end of capitalism or fiat money (the system whereby money is money because a government says so) could be linked to war.
Watson Wyatt points out in a paper on the topic that “not all of these extreme risks can be hedged, or any hedge used is likely to be very imprecise”. The example they cite is a killer pandemic; even if it were possible to work out how likely such a pandemic is, that tells you nothing about how many people would die or whether some groups would be hit worse than others, so its impact is incalculable.
In events where the outcome is relatively straightforward – the end of fiat money would almost certainly lead to gold being the main store of value – the question arises as to how effective the hedge needs to be and how much an investor is willing to pay for it. A portfolio 100 per cent in gold would be entirely hedged against the collapse of fiat money, but might not seem a precaution worth taking.
Reverse stress testing is coming into fashion as well. It reverses the process of stress testing – instead of asking what would happen if x occurs and coming up with the answer “a loss of y”, it asks “what could trigger a loss of y?” in order to find out what x might be.
In answer to increasing demand for tools to understand extreme risk, the academic world is developing extreme value theory, a specialist branch of statistics that tries to use what data is available. This builds on the work of insurance actuaries over decades, as this is precisely what their task has consisted of.
An interesting point thrown up by EVT is that the left tails of return distributions are significantly fatter than the right. In English, this means extreme events (which happen in the tail of a distribution) are more likely to be negative (occurring on the left hand side of the graph) than positive; econometricians would say kurtosis is often accompanied by heteroskedasticity.
The number-crunching runs out of steam when looking for explanations for this, however, leaving it to behavioural theorists to talk about the irrational behaviour of investors.
- Sound practices for backtesting counterparty credit risk models - final document Basel Committee on Banking Supervision, December, 2010
- Canada: Risk Infrastructure: Lessons from the Trenches KPMG, August 10, 2010