Secondary perils – Not so secondary

According to Swiss Re Institute’s annual sigma review of natural catastrophes and man-made disasters, more than 60 per cent of the USD 76 billion of insured natural catastrophe losses in 2018 were due to 'secondary peril' events.

sigma also reports that the total insured catastrophe and man-made losses in 2018 were the fourth highest on record at USD 85 billion, supporting the increasing loss trend over the past two decades. Total economic losses were USD 165 billion, once again highlighting the large protection gap that exists.

'Primary perils' are defined as those that have the highest loss potentials, are well-monitored, and usually covered by catastrophe models – here we refer mainly to earthquakes and tropical cyclones. Secondary perils, on the other hand, are those that generate small to mid-sized losses, such as hail, flood, storm or bushfire.

What caused the high representation of secondary perils in 2018?

One of the main reasons was a lower representation of primary perils in 2018. Even without mega-catastrophe events, the total insured losses were still the fourth highest on record, suggesting that we are seeing a rise in secondary perils. sigma states that the growth in secondary peril insured losses is mostly due to concentration of exposure growth in larger cities, in coastal areas, and on floodplains. Put simply, when we put more assets in harm's way, we will see more losses.

Secondary perils are typically weather-related, so we might expect that climate change will have an influence on secondary peril loss outcomes in the future. As our world warms, we will experience increased atmospheric energy, i.e. more storms; a warmer atmosphere, which can absorb more moisture leading to more intense rainfall events; and more heatwaves driving an increase in wildfire/bushfire losses.

Are secondary perils going up Down Under?

Australia is heavily exposed to secondary perils. Analysis of the Insurance Council of Australia's Natural Disaster Event List suggests that 67 per cent of all normalised1 insured losses since 1966/7 have been due to secondary perils. Of the top 20 normalised losses, 14 are from secondary perils.

Reflecting on secondary peril loss activity in Australia over the last several years, one might be tempted to ask whether we are seeing more of these losses in Australia in recent times. For secondary perils (excluding earthquake and cyclone losses), our analysis of the ICA's normalised loss data shows a slight and probably statistically insignificant residual trend in normalised losses over the past 50 years (Figure 1). From this we can infer that the main driver of increased losses over time in Australia is growth in underlying exposures, which aligns with our sigma research on global losses.

Graph source: ICA Dataglobe

What about climate change?

Although there's no convincing underlying trend in the normalised loss data, other research tells us that climate change is already having an impact on losses.

In recent years, a number of attribution studies have been published, which seek to determine if our changing climate has had any influence on studied events. Carbon Brief analysed the results of more than 230 peer-reviewed scientific global attribution studies, with mounting evidence that human activity is raising the risk of some types of extreme weather, especially those linked to heat. Sixty-eight per cent of all extreme weather events studied to date were made more likely or more severe by human-caused climate change. Heatwaves account for 43 per cent of these events, droughts make up 17 per cent and heavy rainfall or floods account for 16 per cent2.

Closer to home, the Climate Council3 states that most of the attribution studies in Australia have so far been focussed on heat-related events, i.e. heatwaves; record hot months, seasons and years; and surface ocean heat events, which led to extensive bleaching of the Great Barrier Reef in recent years. A few attribution studies have, however, explored the link between extreme rainfall and climate change. For example, the warming trend in sea surface temperatures (SSTs) to the north of Australia may have contributed, by up to 20 per cent, to the magnitude of the heavy rainfall of 2010-11 in eastern Australia.

The Climate Council report also suggests that we may expect to see other climate impacts in the future in the following areas:

  • Heavy rainfall: Extreme rainfall events are projected to increase in intensity.
  • Hail and thunderstorms: Analysis of conditions conducive to severe thunderstorms in Australia indicates significant increases in potential severe thunderstorm days by the end of the century – 14 per cent for Brisbane, 22 per cent for Melbourne, and 30 per cent for Sydney.
  • Extra-tropical cyclones: Projections for east coast lows suggest a reduction by up to 30 per cent towards the end of the century, which is a continuation of the observed trend. However, there is some indication that the intensity of the most severe east coast lows could increase.

Other studies have suggested a potential increase in both the mean hail size and frequency of larger hail events over North America4. 5A similar shift in Australia would only increase the loss potential from this hazard, which is already responsible for a significant proportion of the normalised historical loss record.

And in New Zealand?

While earthquake and other geological perils will remain the dominant force, we can also expect to see more extreme weather events under a changing climate. More transitioning cyclones from the tropics, such as Gita and Fehi, and an increase in torrential rainfall due to the formation of 'atmospheric rivers', such as following Cyclone Debbie, are examples. Recent wildfires in the Nelson area may be an early local indicator of increasing wildfire exposure –sigma also highlights yet another record wildfire year for California.

What are we to do in the face of climate risk?

The challenge for insurers and reinsurers is in understanding when, and by how much, climate change will impact their businesses. Due to the inherent noise in the loss data, it will undoubtedly take quite some time for a statistically relevant climate change loss signal to emerge in the loss record6. However, we don't have the luxury of waiting 100 years to acknowledge, with hindsight, that climate change has made its mark on insured losses. Nor can we simply ignore the problem – it won't go away.

We are now seeing encouraging movement in terms of quantifying and reporting financial impacts from climate change. Despite some limitations around estimating the impact of climate change on certain secondary perils, such as hail, the United Nations Task Force on Climate-related Financial Disclosure (TCFD) initiative7 could provide a framework for insurers and reinsurers to estimate many other climate risks (cyclone, flood) in their insured portfolios, and price for them.

TCFD may also offer governmental authorities a framework to understand where they can expect the biggest impacts from climate change, enabling appropriate approaches (such as good planning and mitigation) to manage the impacts of climate risk and keep catastrophe risk transfer affordable and sustainable into the future.

1 ICA’s normalised loss data removes trends attributable to exposure growth and can be used to detect other underlying trends.





6 Crompton, Ryan & Pielke, Roger & McAneney, John. (2011). Emergence timescales for detection of anthropogenic climate change in US tropical cyclone loss data. Environmental Research Letters. 6. 014003. 10.1088/1748-9326/6/1/014003.