AIR Worldwide’s Shane Latchman highlights some of the challenges in climate risk assessment.
The writer Will Storr states, “The unexpected signifies a change, and a change is where a story begins.” Perhaps this is one of the reasons climate change is such a gripping topic –the unexpectedness of the story. The other reason, of course, is regulation. Actually, regulation is probably the big driver. It increases our focus on physical climate change quantification and hence changes from natural catastrophes. But there are several misunderstandings that sometimes trip us up.
The following are five I’ve seen:
Misunderstanding #1: When natural catastrophe losses increase around the globe, this is due to climate change.
The default assumption shouldn’t be that an increase in historical losses (e.g. from hydro-meteorological perils) is due to climate change alone. Many factors impact loss increase; exposure growth is a key driver as are short-term (e.g. decadal) cycles such as teleconnections. Research by AIR shows that increased historical loss trends were removed when factoring in exposure growth and that most loss change trendlines we see over time in the insurance industry account for inflation but not changes to exposure. Natural catastrophe loss increases over the last few decades have been observed in some places for some perils, but not everywhere; the source may likely be attributable to climate change in some regions but not others.
Misunderstanding #2: We know what climate change will do to all natural catastrophes in the future.
We have more certainty that sea levels are rising and will rise, and that droughts, wildfires and flooding will get worse. We can’t say, however, that a higher global temperature will lead to more storms making landfall. Certainly there is some evidence that the frequency with which stronger storms form will increase and some observation already of more storms in the Atlantic (but not Central and East Pacific) in recent decades compared to further in the past. Attribution of the reason for the change is ongoing.
On a global scale, most tail risk is driven by US hurricane – 60 percent of AIR’s global exceedance curve’s 100-year insured tail value at risk comes from US hurricane, so any changes here can be material, which is why it’s vital to quantify what type of impacts are possible, even though we lack certainty.
Misunderstanding #3: Current natural catastrophe models are invalid due to climate change.
In general most hydro-meteorological models are built on the last few decades of data and hence the recent past. Thus, broadly speaking, catastrophe models represent the current climate. There is not yet a full scientific consensus on trends for natural catastrophes from climate change that’s already happening. Sea level rise, which determines tsunami and surge heights, is accounted for as models use current sea levels. One trend being watched is the poleward migration of tropical cyclones and we’ve been doing scenario analyses on this. In some cases we do see trends in precipitation, for example, more precipitation recently compared to further in the past and adjustments are made as models are updated.
Misunderstanding #4: Percentage changes from climate change studies fully account for increased loss.
Most studies on climate change alter the hazard from catastrophe models and keep other components, e.g. vulnerability, defences, land use, vegetation and exposure constant. So if we see a 30 percent increase by 2050 under RCP 4.5 for a region/peril we need to remind ourselves that other factors will change as well. The impact of these factors are competing, so the overall impact could be greater/less than the value quoted (and keeping in mind this is just a model of the future). People migrating to areas of increased risk such as near rivers or the coast is an example of increasing risk. Building out the modelling for these other factors has begun, but is fairly nascent.
Misunderstanding #5: Natural catastrophe loss changes (increases) from climate change represent a step change compared to current risk.
Specifics here are important. Most studies that run scenario impacts of climate change on constant exposure for hydro-meteorological perils (e.g. tropical cyclones) show impacts across the exceedance probability curve on the order of tens of percent in the coming decades. The impact tends to be even greater for water-based perils, such as surge and flood, as well as wildfire; running these on their own could give greater impacts. So climate change can really change the landscape of some perils in the future; perils not in your top five list of most material perils could move into that list, although your most material perils may not necessarily experience the same level of change.
Shane Latchman is a vice president and managing director of AIR Worldwide UK, a Verisk company.