Black swan theory

Black swans

Unlike bankers, insurers have so far averted trouble in the recession. But, how much of this is due in part to the effective use of modelling and can this help them stay out of trouble? Rachel Gordon speaks to key market figures about modelling challenges on the horizon.

Bankers are in the dock, seen as greedy and self-serving, condemned for having failed to learn from history. Insurers meanwhile" at least to date" remain the good guys, and would certainly appear more attuned to risk.

Back in April, Lloyd Blankfein, the chairman and chief executive of Goldman Sachs, made a speech in Washington where he blamed a range of factors for the banking sector mess, including the fact that models failed to capture the risk inherent in off-sheet activities.

"It seems clear now that managers of companies didn't appreciate the full magnitude of the economic risks they were exposed to; equally worrying, their counterparties were unaware of the full extent of these vehicles and, therefore, could not accurately assess the risk of doing business. Post-Enron, that is quite amazing," he said.

In contrast, Zurich's chief executive designate Martin Senn, currently its chief investment officer, said last week that prudence had helped his firm weather the financial crisis, saying excess capital would be used to either pursue acquisitions or gain market share rather than seek lucrative but risky investments. The insurer, which has assets of around $200bn (£121.3bn), avoided risky bets in the US mortgages bond market.

And Allianz's UK chief executive Andrew Torrance picked up plenty of media coverage when he said "boring is the new sexy", referring in part to sensible investment strategies. At group level, Allianz's capital of 161% of the regulatory requirement puts the company in a strong position.

So, in terms of new modelling challenges, particularly those on the investment side, are insurers genuinely one step ahead of the game? Patrick Grealy, managing director of software provider Ultimate Risk Solutions, says: "Insurers are generally much more comfortable in their liability modelling. Asset modelling is in the main a recent development. Most feel that there is more volatility in the liability side of their balance sheets, given that many non-life insurers adopt a very conservative bond/cash and currency matched position. Those with more aggressive investment objectives will need to give more time to the process of economic scenario generation."


Better position

Mark Churchlow, Allianz's director of actuarial and planning, oversees much of his company's modelling work and also regularly works with counterparts in the Munich head office. He comments: "Insurers as a whole are in a better position in the investment cycle and, certainly from Allianz's perspective, we have adopted a conservative approach taking account of the current economic turmoil when modelling."

In terms of investment in modelling, he says the biggest focus has been Solvency II. "There is the regulatory pressure and the fact this is a major project for all insurers. But, we are also constantly building in the impact of future risks into our models."

Mr Churchlow says these include areas such as weather catastrophe, the higher inflation rate for personal injury claims as well as rising incidences of fraud, which are tipped to grow as a result of the recession. "We can look back to figures from the 1990s when there was again a downturn and look at our claims experience then. This showed, for example, rising theft in both domestic and commercial, plus more fraud. We are also building in the fact there are more insolvencies, which in turn could mean less premiums, alongside higher credit risks as well as the fact that asset values may be falling."

He adds that Allianz is not being swayed by economists talking up the economy. "It is safer to be realistic; we're probably looking more towards the end of 2010 before we'll see evidence of recovery." Since this is a global recession, he says there is an increased focus on collaborating with Munich and of work across the whole group. "A lot of issues apply to the business as a whole rather than looking at areas in isolation. You also need to ensure there is regular stress scenario testing, where applicable, looking at what we think could go wrong and how much we would need to fix it."

Steven Blake, chief actuary at Groupama, adds: "We are already making good progress with our efforts to utilise financial modelling to help us drive business strategy and aid decision making. We are also making a considerable investment in updating our stochastic model within the actuarial function and developing a much wider internal model, which will allow us to embed a more formal, documented and informed decision-making and risk assessment process across the business. From a governance perspective this is being actively encouraged and supported by both the Financial Services Authority and the framework that surrounds the Solvency II initiative."

He continues: "More generally within the financial sector the ability to communicate the requirements and achieve an understanding of the outputs from what are necessarily complex models will also be a challenge given that this has previously been an area preserved for actuarial teams. From our own viewpoint we are already working hard to identify the most effective ways of communicating information in a manner that is as simple and straightforward as possible. It is certainly encouraging that we are moving in the right direction but we do recognise that we will need to retain this very clear focus as things continue to develop."

He says another area that poses a significant challenge for the industry will be the assessment and quantification of operational risk: "At Groupama this is an area where we have been developing our skills for some time now" but there is always more that can be done and greater levels of sophistication that can be achieved."


Filtering out history

Meanwhile, David Ingram, senior vice president with Wills Re, says that, even within insurers, models are only as good as the modellers and there remains a tendency to "filter out history". And, he emphasises the model is also only as useful as the data the user feeds into it.

However, he agrees they have a crucial role to play, pointing out "humans, who to varying degrees all have a limit to their capacity to juggle multiple inter-connected streams of information, need models to assist with decision-making at all but the smallest and least complex firms".

He continues: "Models can help to answer complex 'what if' questions that are always going through a business manager's mind. But those answers are only reliable if the modeller still takes on board the warning US economist Robert Merton gave during his Nobel Prize acceptance speech in Stockholm in 1997."

This was: "The mathematics of financial models can be applied precisely, but the models are not at all precise in their application to the complex real world. Their accuracy as a useful approximation to that world varies significantly across time and place. The models should be applied in practice only tentatively, with careful assessment of their limitations in each application."

Certainly it seems experts view modelling, while vital, as always having weaknesses and only able to fulfil one part of enterprise risk management. There also needs to be regular assessment of current practices and ongoing investment in both people and technology. As Elliot Varnell, principal adviser in KPMG's actuarial team, comments: "The global financial crisis has certainly challenged the economic and asset models that insurers have been using with movements in market parameters that were more extreme than some models had predicted. This has meant revisiting model calibrations and model structures to include allowance for heavy downside risk and tail correlation."

And referring to the need to take a more comprehensive approach, he adds: "Cutting edge players are also considering the integration of economic modelling across risk types and business areas. For example, the integration of models for different risk types can include integration of market, credit, catastrophe and operational risk. Integration over business lines can involve integration of economic risk not only between life and general insurance, but also including the banks and asset management operations."

Work around scenarios is a key part of horizon scanning and Mr Ingram refers to so-called 'Black Swan' testing, which relates to events occurring outside the realm of normal experience. The Black Swan theory, so called after essayist Nassim Taleb's book, suggests that nearly all major scientific discoveries and major historical events were Black Swan events, including the invention of the computer, the rise of the internet, and the 11 September terrorist attacks in 2001. The term comes from the 18th Century discovery of black swans, which disproved the previously accepted assumption that all swans were white.

While insurers are typically not renowned for thinking outside the box, Mr Ingram says this is essential if they are to start seriously considering the types of emerging risks that could impact on the business.

He says emerging risks often result from changes in the political, legal, market or physical environment" pointing out examples from the past include asbestos or silicone liabilities. "Other examples could be problems deriving from nanotechnology, genetically modified food or climate change. The recent problems experienced by banks and other financial firms resulting from mortgage losses could also be seen as emerging risks."

Mr Ingram adds: "Clearly in general insurance you have long-tail risks that do cause big problems, but insurers have to work towards understanding these and to take a position." He urges insurers to set in process early warnings that will help them anticipate disasters.

Alongside models and actuarial expertise, he says that gaining insight into unknown risks can also be helped by an insurer's culture, whereby the following is encouraged: searches of publicly available information on new risks; multi-disciplinary meetings to share knowledge across silos; 'free thinking' workshops to explore the extremes of current knowledge; an emphasis on logical 'what if' thinking rather than emotional 'that won't affect us'; an emphasis on severity as, by definition, most emerging risks will be perceived as low probability" until they happen; and internal reporting that accepts opinions can be as useful as facts.


Infinite variety

Using models to measure operational risks remains far less sophisticated, not least because of the scale on which these can occur" and their infinite variety. For example, they could range from damage to premises and internal fraud to technology failure. There is also an obstacle in the lack of data. The Association of British Insurers only launched its database of the operational losses of member companies in 2005. The database is run by the operational risk insurance consortium, in conjunction with software company SAS. The ORIC's data is mainly UK-based with only 24 companies" although there are plans to reach 50 members by 2011.

Mr Ingram points out that the industrial sector "tends to be quite advanced in scenario analysis as there has not been the same degree of application of the actuarial/probabilistic techniques well known in the insurance sector. Some of these uncertainty-modelling disciplines are increasingly relevant to the insurance sector."

Advances in modelling techniques clearly put more pressure on boards to ensure they understand the work being carried out and the implications. It is also a major issue for regulators who must ensure they have high calibre employees that can assess and monitor such models. As Mr Varnell says: "Making models more sophisticated to capture more complex market behaviour is not without its challenges, however, because mathematical sophistication creates governance issues as fewer people within an organisation understand the model details. This makes the process and controls around economic models much more important than they used to be."

And, Mr Grealy concludes: "Regulators and other agencies will look and are looking to see that the process is embedded into the culture of the company and that the modelling is a day-to-day process with a feedback loop to improve decision making. Documentation of the stages of parameterisation/implementation and careful generation and version control will also be required. Companies that may not have a strong actuarial focus" internal or external" may struggle to adopt and embed the modelling culture together with the required controls and associated documentation. There may be a push to gear up in this area."

  • LinkedIn  
  • Save this article
  • Print this page  

You need to sign in to use this feature. If you don’t have an Insurance Post account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an indvidual account here: