Climate change spells death of certainty

Global warming threatens to upend everything risk models take for granted

After years of complacency, financial firms are finally getting serious about measuring their exposure to climate change, and taking action to mitigate its effects.

This is no easy task.

Beyond the usual intricacies and pitfalls of modelling, climate change brings with it the erosion of long-held certainties: predictable weather patterns in developed markets; steady sea levels in heavily insured jurisdictions; stable governments capable of maintaining fiscal discipline while spending trillions on climate defences.

All of which makes the job of risk managers – those tasked with measuring, modelling and putting a dollar value on climate risks, and enacting a plan to mitigate the impact – rather difficult.

Insurers and re-insurers know this first hand. The industry just witnessed its costliest back-to-back years, with losses totalling $227 billion in 2017 and 2018, according to data compiled by Aon. Last year, insured catastrophe losses hit $90 billion, the fourth-highest on record. Weather disasters, such as hurricanes Michael and Florence and Typhoon Jebi, accounted for nearly $89 billion of the total.

Usually, when underwriting a risk of this type, an insurer would rely on a catastrophe or cat model to estimate the frequency, intensity and possible damage footprint of a particular hazard. A nat cat model for hurricanes, for instance, looks at recorded instances of historical hurricanes in a particular geographic area and generates a range of estimates as to the extent of the damage a hypothetical future hurricane could inflict.

But cat models struggle to offer any kind of accurate gauge for events that are far more extreme than those witnessed before, or in a location they were never expected to occur – wildfires or typhoons that may owe their increase in frequency and severity to changing weather patterns being a prime example – because, at the outer limits of the tail, there’s no historical data to feed the model.

For the five largest cat events of 2018, the average loss estimates of the two main modelling firms – AIR Worldwide and RMS – came in at $14.25 billion, roughly 65% below the true loss figure of $40.3 billion.

Cat models struggle to offer any kind of accurate gauge for events that are far more extreme than those witnessed before, or in a location they were never expected to occur

The industry is now searching for more consistent and accurate ways to model losses arising from weather-related disasters. Several lines of inquiry are converging on the idea of combining decadal forecasting techniques – which compute climate fluctuations over multi-year periods – with orthodox stochastic models.

“The firm that merges decadal climate models into traditional stochastic natural catastrophe models the most quickly and credibly will be the winner,” says Alison Martin, chief risk officer at Zurich. “They will be able to say: ‘We can attribute X storm, X flood, X wind event to climate change’ – and the modelling would support it: ‘Here is the economic cost of climate change.’ No-one has done that yet, successfully. It’s a trillion-dollar question.”

Forecasting climate change is a disaster

Financial firms are also applying so-called ensemble techniques – an umbrella term for quantification methods that employ multiple models at once – to quantify climate exposures. The most common approach is to run cat models alongside general circulation models, or GCMs, which can be used to simulate various climate scenarios.

At the extreme, of course, insurers can stop underwriting risks they cannot accurately gauge: Argo, a large California-based reinsurer, decided it did not want to write casualty business for utilities in the state – just before 2018’s deadly wildfires struck. And banks could cease financing such risks.

But simply unbanking whole sectors and uninsuring whole jurisdictions is not what agents of risk transfer are supposed to do; if someone is willing to pay, mitigation should have a price. New modelling techniques could help financial firms to more accurately set that price.