A rapidly changing world produces rapidly emerging risks that can flummox even the most comprehensive of risk management frameworks. Neil Cantle proposes an approach to improving how insurers respond to today’s threats and opportunities
Modern business is already complex and challenging but, as digitalisation progresses, it is also getting a lot faster and even more inter-connected. This presents a significant challenge for risk managers as most risk frameworks are simply not geared up for risks appearing at high velocity or in a very non-linear way.
Financial markets can move quickly and sometimes unpredictably, so financial risk monitoring is often an element of the framework which is able to identify rapid movement. But other areas like operational and strategic risk still tend to rely on relatively slow-moving frameworks involving components such as registers and escalation processes.
So, how do you manage risks which can flash out of nowhere? The answer is, essentially, that you have to get skilled at anticipating them and maintain operational resilience to give you the flexibility to react if they occur.
This requires a consistent self-challenge by looking at emerging trends which could lead to particular outcomes and asking, “Would we see that coming?” If the answer is “no” or “maybe” then improvements can be made to the risk management framework, or actions taken to remove yourself from its path.
A common mistake with emerging risk processes is that they start with a trend and ask, “What could this do to us?” The problem with this is that the answer is often, “It depends!”
It is almost never the case that a particular trend will evolve in isolation. Rather, it will co-evolve with a number of other trends, each influencing the others, to produce an emergent final outcome which you cannot know by looking at each trend in isolation.
You must therefore decide which outcomes you are interested in first and develop narratives, themed with the trends you are trying to understand, which plausibly take you from today to those outcomes.
So, for example, if you are concerned about whether consumers will continue to purchase long-term savings policies and you have identified emerging trends involving the adoption of new technologies and changing social behaviours, then you should develop scenarios which explore how those trends might influence attitudes to long-term saving and where the tipping points are.
For each scenario, your challenge is to consider whether the identified features would be appropriately identified by the risk framework and whether the required actions to mitigate the risk would be flagged. The new trend might involve a combination of factors which, individually, appear rather benign but their combined effect is highly unfavourable. Risk frameworks which look at indicators individually would miss this.
In addition to building narratives based on input from subject matter experts, it is increasingly possible to build adaptive monitoring tools which can identify new trends from data and reveal complex relationships which are otherwise hard to spot.
This involves capturing information about items which are potentially related to the outcomes you are interested in and whether the relationships present are consistent with your theories.
Increasingly, this information need not be structured or numerical. Visualisation techniques can help to present large complex information sets in a way that makes it easier for experts to spot patterns. And, non-linear relationship measures, such as entropy, can be used to determine networks of relationships which can then be evaluated to help experts describe their meaning.
A further recent advance is the ability to deploy artificial intelligence (such as deep neural networks) to elicit trends from text, even if you don’t really know what you’re looking for, and identify emerging trends, insights about the communication network structure and the strength of feeling about the topics.
Helping your experts to describe compelling narratives about the future and testing the organisation’s ability to prosper under those circumstances provides a robust ongoing process which enables the organisation to constantly test its ability to see and manage new events.
As robotisation increasingly frees the risk management team up from traditional oversight tasks, they can reorient themselves towards more of an “insight” role, providing regular information to the business about what is going on and how to recognise the onset of new trends. This helps the business to anticipate and prepare for an uncertain, high velocity, environment and remain agile.
Neil Cantle is principal and consulting actuary at Milliman in London