Taking A Comprehensive View of Catastrophe Risk Worldwide

November 26, 2012 / 3 min read

The year 2011 was a significant loss year in terms of natural catastrophes; 175 events—including earthquakes in Japan and New Zealand, flooding in Thailand, and swarms of tornadoes in the United States—produced the second highest insured loss tally, USD 110 billion, since Swiss Re began publishing natural catastrophe data in 1970.

While the total global insured losses from 2011 may seem high, or notable for the number and variety of events that comprise them (15 individual natural catastrophes each exceeded USD 1 billion in insured loss), it falls near the 6.7% exceedance probability (or the 15-year return period) on Verisk's global industry exceedance probability (EP) curve.

Furthermore, although last year's losses were significant, they did not cause the insolvency of a single company—a testament to an improved approach to catastrophe risk management, aided in no small part by the use of catastrophe models.

The events of 2011 did prompt the insurance industry to consider how to put the resulting losses in perspective. Basic questions that surfaced after 2011 and that will be explored in this paper, include: 1) how frequent are loss years like 2011; 2) what percentage of losses were modeled; and finally, 3) are natural catastrophe events becoming more frequent and/or more severe?

Verisk is in a unique position to provide guidance with respect to these questions. First, Verisk develops and maintains detailed industry exposure databases (IED) for each modeled country. These IEDs serve as the foundation for all modeled industry loss estimates and make the generation of a global industry exceedance probability curve a straightforward task. Second, one of the key advantages of Verisk's approach to generating the stochastic catalogs included in its models is that it enables the user to determine the probability of various levels of loss for single- or multiple-event years of catastrophe activity, across multiple perils and regions. Finally, Verisk models the risk from natural catastrophes (and terrorism² ) in more than 90 countries. This gives Verisk a truly global perspective.

The discussion herein promotes awareness of Verisk's global industry EP curve, two key metrics of which are shown in Table 1. To provide additional context around the aggregate risk view on a global scale, the paper illustrates scenarios that could result in losses even greater than those of 2011; namely, scenarios at the 1% exceedance probability (the 100-year return period) and the 0.4% exceedance probability (the 250-year return period). Details about these scenarios, as well as metrics from Verisk's global EP curve and the comprehensive discussion to follow, help risk managers contemplate a critical question: is the industry prepared for losses from natural catastrophes on a global scale?

**Table 1. Key loss metrics from Verisk's global industry EP curve**
Exceedance Probability	Insured Loss (USD)
Average Annual Loss	59 billion
1% Aggregate	206 billion

Note that industry insured losses can and do occur in regions and from perils that Verisk does not currently model; losses from these regions and perils are not included in Verisk's global estimate. As such, the exceedance probabilities in this report can be expected to be higher.

²Terrorism is not included in analyses here, however.