Zurich Risk Report:: pointing at, NOT “pushing the boundary”!

imageIt is reassuring to know that, when “responsible” insurers, such as Zurich Financial Services, publish reports they don’t gloss-over “uncertainty”! But I am still concerned that, whilst they have the considerable insight and wisdom of Doug Hubbard in their corner, they do little more than (re)define the “problem statement” and point in the general direction of theoretical solutions…

It’s like presenting a High School student with a problem from Quantum Physics, giving them a calculator and telling them to get on with it!!! 

Surely, in the current climate, it is too big an assumption that a business has the in-house resource (time, finance and intellectual) to de-code, interpret and address such problems? These are the same problems that have already laid waste to the major institutions and the reputations of banking and finance organisations, whose “investment” in the finest mathematical minds (outwith Academia) has proven what we already know: attempts to model the future in a complex and inter-connected world are, ultimately, futile exercises.

This does not mean that I am totally against building models but let’s get the facts straight first!

We can’t forecast the future based upon the past

Assumptions are just that! They are, by definition, subjective, even if quantitatively based, and

…in complex [non linear] systems, even minor deviations, can have a MAJOR impact.

As a result, and with all due respect to Mr Hubbard (as well as those whose work he cites), I am not convinced that an unreliable model is vastly better than no model. At least, as far as reliable forecasts on a meaningful timeline are concerned.

The pressing need is for, cost effective, “real world” SOLUTIONS to everyday problems, not answers to questions that are topics of research and debate amongst industry experts and Academics. 

It is as very interesting and useful report but that pre-supposes that the reader is willing and able to learn the appropriate skills to define, build and maintain a “cone of uncertainty”, an “early warning system”, scanning and analysis capabilities, etc, etc. Frankly, it ain’t gonna happen unless insurers, complete with emerging and strategic risk teams, evacuate their Ivory Towers and move more freely amongst those upon whom their own economic success is based! I’m sure that isn’t too radical a means of (re)building TRUST – this quote from page 23 of the report:

“Trust facilitates cooperation, making economic transactions easier”

When it comes to “practising what you preach” the industry needs a reality check: WE AREN’T TRUSTED!!!

As the report explains, Zurich already engage in their own scanning, etc. They obviously understand that “risk” is not an exclusively exogenous phenomena but, with this in mind, I am curious to know how they view this report in the context of their own industry; what they are doing about it; do they believe there are financial benefits to be had from earning trust; do trusted companies enjoy a “competitive advantage”?

To this end, I think that, if ever an expression could be applied to an industry, it MUST be applied to insurance (the report does refer to banking): “reputational commons

This is a new expression on me but I get it, like it and can think of many occasions that I have written and spoken about industry-wide reputational damage. In this report it is used in the context of the Oil Industry and their ‘self-similar’ risk strategies but, as if to illustrate the “universality of complex systems” I (and many, many others) have used it, with very good cause, when referring to the Financial Sector i.e. banking AND insurance.

Like the Oil Industry, Financial Institutions are creators and casualties of complexity, that: unmanaged, is communicated and amplified across complex systems and networks adding to uncertainty and volatility…then we ALL stand to be victims!

Report extract:

The presence of uncertainty is always a challenge for insurance and risk management alike. But in the past, insurers have always pushed the boundary and provided solutions for risks that at one time were considered uninsurable. The most recent example for such a transition is supply chain risk insurance, but one can be sure that there will be more to follow in the future. In deliberately pushing the boundary, insurers and risk managers are making uncertainty not only measurable but also manageable. And by adhering to a few basic principles, and by paying appropriate attention to common sense, risk management will preserve and even create value.

5 Lessons for Risk Management

  1. Uncertainty can be assessed and measured. It is a common fallacy to suppose that uncertainty cannot be assessed and measured. However, it is important to define well the object of measurement. Once that’s done, it is typically recognized that the problem – how to grasp uncertainty in our case – is not completely unique and that more data are available than initially thought. Even when faced with the perfect unknown unknowns, scenario analyses will frame the problem and help to reduce uncertainty.
  2. Establish early warning systems. To be aware of uncertainty is one thing, but to spot an emerging risk is another. Emerging risk radars must be built and systems for continuous scanning established. This requires strong communication structures so that information can filter up quickly and easily to decision makers. IT tools such as Web mining and blog mining can be deployed, for example, to capture potential reputation crises.
  3. Prudent forecasting is possible and necessary. Although forward-looking By default, risk assessment should not be confused with forecasting. Nevertheless, forecasts are important to reduce uncertainty. The common sense approach to efficient forecasting assumes that trends cannot be relied upon, and that a good forecast will embrace things that don’t fi t into familiar boxes. It is a first step toward mitigating the risk of being surprised by unknown unknowns, or the now proverbial ‘black swan.’
  4. Contingency planning is indispensable. Since unknown unknowns keep generating surprises it is important to develop contingency plans that cover a whole range of scenarios. When disaster strikes it is usually too late to create effective plans to cover the fallout for production, employees, reputation, supply chains or service disruption. Contingencies for generic adverse outcomes must be in place. Successful contingency planning will also endeavour to map the interaction among emerging risks.
  5. Resilience buffers will dilute adverse impacts. Even the best risk assessment and most efficient forecasting cannot protect against the adverse impact of uncertainty. Reputation risk is particularly treacherous as reputation loss can occur overnight. That’s why creating a ‘resilience buffer’ is vital. As Roland Schatz shows in his contribution, pushing a company’s image above the awareness threshold will help to deflect threats to reputation if a crisis should occur. It gives senior management a stock of goodwill or a resilience buffer to draw from, which helps to reduce the damage to reputation. The challenge for risk managers will be to transfer the idea of resilience buffers to other areas in order to mitigate the adverse impacts of uncertainty.

Chapter 1: Top risks most businesses are currently concerned about

  1. Regulation and compliance
  2. Slow economic recovery
  3. Cost control
  4. Emerging market entry
  5. Social acceptance and corporate social responsibility (CSR)
  6. Taxation risk

Chapter 2: Risk and uncertainty

clip_image001• Uncertainty: the lack of complete certainty, that is, the existence of more than one possibility. The ‘true’ outcome/state/result/value is not known.

• Risk: a state of uncertainty where some of the possibilities involve a loss, injury, catastrophe, or other undesirable outcome (i.e., something bad could happen). From here we make an explicit distinction between how the two are measured:

• The Measurement of Uncertainty: a set of probabilities assigned to a set of possibilities, for example, ‘There is a 60% chance it will rain tomorrow, 40% chance it won’t.’

• The Measurement of Risk: a set of possibilities each with quantified probabilities and quantified losses. For example, ‘We believe there is a 40% chance the proposed oil well will be dry with a loss of USD 12 million in exploratory drilling costs.’

Measuring risk is not possible without measuring uncertainty, and that means applying odds to various possible outcomes. This is the point where many different practitioners of risk management in many industries find their biggest obstacle. But there are some sound methods for applying probabilities that are surprisingly straightforward.


THE FINAL CONTRIBUTION FROM ME is really to attempt, AGAIN*, to illustrate how many of the “issues” identified by this report are catered for by conducting a Quantitative Complexity Analysis by Ontonix.**

*Roads to Ruin: major corporate failures beyond the scope of risk management

“Why did RBS fail?” v “Roads to Ruin”

**Ontonix:: the why, what and how of on-line rating

We have defined, measured and applied solutions for risk associated with “Complexity” across sectors and domains BUT, as far as insurers are concerned, complexity is an “unknown known”…NEW WORLDVIEW REQUIRED TO SECURE AND RETAIN COMPETITIVE ADVANTAGE!?

This graphic may help with distinguishing between knowns, unknowns, unseen and unforeseeable:

One Response to Zurich Risk Report:: pointing at, NOT “pushing the boundary”!

  1. dwhubbard says:


    Thanks for your comments. When you say that you aren’t convinced that a model with error is better than no model, I need to respond with a couple of points.

    1) You always have a model. The question is not whether to have a model but which model you will use. If you aren’t using a quantitative model then your model is something qualitative or even just unaided intuition.
    2) All models have error, including your alterative models like unaided intuition.
    3) Since all models have error, the relevant choice is for which model has less error, not which has none.
    4) The other sources I cite simply show that – for a surprisingly wide variety of decision problems – the model of human intuition has so many errors that even simple quantitative models often outperform them.

    Almost any error you can think of for a quantitative model is certainly also an error in intuition and anything that may be excluded from a quantitative models may also be excluded from intuition. But intuition has a whole new batch of errors beyond quantitative models. Humans have selective memory and they commit inferential fallacies on a regular basis. They make assumptions like quantitative models but unlike quantitative models the assumptions are rarely explicit and available for others to investigate. And unlike quantitative models, researchers have shown how humans are influenced by random, arbitrary external factors that should have nothing to do with the question at hand (like changes in your testosterone levels, a random number you were previously exposed to, and recent anger or fear).

    So, if you want to claim that intuition – your default model – measurably outperforms quantitative models, that would seem an uphill battle given the weight of evidence to the contrary. As early as the 1950’s Paul Meehl was collecting studies showing where human intuition underperformed compared to simple statistical models. More recently, Tetlock showed over 82,000 forecasts from 284 experts on politics, economics, and other macro-trends and concluded that simple statistical models almost always beat the human experts.

    In short, you are GOING to use a model no matter what. The only question is whether you use the model that has additional avoidable error or one that has less error.

    Thanks for the discussion,
    Doug Hubbard

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s