Parametric Risk Transfer for Volcanic EruptionsCatastrophe Risk Analyst at Guy Carpenter Managing Director & Global Head of Catastrophe Risk Research at Guy Carpenter
The design of parametric risk transfer solutions applied to volcanic eruptions is still at a nascent stage in the (re)insurance industry. However, several organizations have recently expressed public interest in adopting such solutions. Guy Carpenter’s specialist teams have joined forces and expertise to explore the topic and propose a parametric design for explosive volcanic eruptions from Japanese volcanoes.
Parametric risk transfer mechanisms such as catastrophe bonds were first issued in the mid-1990s in the aftermath of Hurricane Andrew and the Northridge Earthquake; and the market has grown steadily since then. One of the advantages of these mechanisms is the access they provide into additional capacity for risk transfer in cases of capacity shortage of the global reinsurance market. Another advantage is that parametric risk transfer makes claims payment more transparent, faster and more cost-effective. Recoveries are tied to a set of measurable physical characteristics, such as the magnitude of an earthquake or the category of a hurricane, rather than to actual losses.
Yet, there is an associated level of “basis risk” in these products. Basis risk appears when an insured suffers losses from a catastrophe event, but a commensurate payment is not released due to the measured physical metric not reaching the predefined payment threshold. Conversely, a payment to the insured may be released in the absence of losses if the measured value of the hazard-related parameters meets the pre-established threshold.
Earthquake parametric catastrophe bond transactions and other solutions have flourished partly thanks to the development of techniques to model earthquake risk, apart from a growing need for vehicles to bring more risk-bearing capacity to the catastrophe reinsurance market. However, a similar development in the field of volcanic risks has not yet taken place, which may be partly explained by the lack of fully probabilistic volcano risk models in the insurance industry.
There is one product in the market, offered by Sompo Japan Nipponkoa Insurance, that is tied to volcanic alert levels issued by the Japanese Meteorological Agency. It caters to commercial corporations in Japan that are at risk of business interruption losses from significant volcanic eruptions.
There are recent examples of demand for this type of risk transfer solution, including from humanitarian organizations, such as the Red Cross, which is currently looking at the potential issuance of a parametric catastrophe bond of around $15 million to protect against volcanic risks.
As part of its research work, Guy Carpenter has conceptualized and designed a parametric risk transfer mechanism to offset losses to building structures arising from large, ash fall-producing volcanic eruptions. The solution has been designed on the basis of Guy Carpenter’s fully probabilistic loss model for eruptions from six Japanese volcanoes, including Mt. Fuji, with a potential to impact the highly populated and industrialized Tokyo and Kanagawa prefectures. This volcano model estimates losses arising from the deposition of volcanic ash fall on building structures and considers residential, commercial and industrial lines.
Such an instrument could be interesting to stakeholders such as insurers with large earthquake concentrations in the area (earthquake policies cover volcano risk in Japan) or regional governments interested in ceding losses to dwellings from a severe eruption impacting the area, among others. This type of parametric solution could be applied to other regions in the world at risk of explosive volcanic eruptions, as long as certain conditions are in place, such as the availability of a fully probabilistic volcanic eruption catastrophe loss model.
This product relies on obtainable, observable physical factors, making it applicable to settings worldwide without special equipment.
Designing a Parametric Solution for Volcanic Eruptions
A crucial step in the design of a parametric solution is the selection of a physical parameter descriptive of the event, one that correlates well with potential losses. Subsequently, one or more threshold values are defined for the selected parameter, above which significant damages are likely to occur. When the physical parameter exceeds that threshold for a particular event, it is considered that a risk cedent should receive a payment commensurate to the loss that their portfolio will likely incur as a result of being exposed to the event.
In the case of parametric earthquake risk transfer, for instance, it is common to select the magnitude of the earthquake as one of the main parameters. After researching several physical parameters associated with the phenomenon of volcanic ash fall, we propose a combination of two eruption-related parameters on the basis of which trigger threshold levels are defined: height of the eruptive column and predominant direction of movement of the eruptive plume. Both parameters are explicitly defined for each of the volcano ash fall events featured in Guy Carpenter’s volcano risk model.
Another requisite for the design of a parametric solution is that the selected parameters are reported on a near-real time basis by a reliable, reputed organization. In Japan, the Japan Meteorological Agency monitors volcanic activity throughout the country and issues relevant warnings and information to mitigate damages. When volcanic anomalies are detected, the agency steps up its monitoring activities and publishes regular bulletins on a real-time basis, including the Observation Reports on Eruption. These reports provide information on the ongoing eruption, such as eruption time, eruptive column height, the main direction of movement of the eruptive plume and the maximum plume height recorded from the onset of the eruption. This information can be used during an ongoing eruption to determine whether any payments are to be released to the insured/risk cedent.
Application of Volcano Parametric Risk Solutions
The formulation of the parametric trigger contemplates both a “binary design”—each event can either pay or not pay a fixed monetary amount depending on whether it exceeds the parameter threshold defined by the specific design—and a “multilayer design”—each event can pay one of several predefined payment levels, associated to a series of defined parameter thresholds. Results of the optimization algorithms consist of the definition of eruptive column height threshold values for the different possible directions of plume movement for a specific application—for example, a defined portfolio of insured properties—associated to one or several predefined target payment amounts. The parametric trigger algorithm is applied on a per-volcano basis. For instance, Mt. Fuji to the west of Tokyo, poses the largest risk to the geographical area covered by the volcano risk model.
In summary, the design of this parametric product relies on easily obtainable, observable physical parameters relating to explosive volcanic eruptions, which makes this prototype potentially applicable to many other volcanic settings worldwide without the need for any special monitoring equipment. If the height of an ongoing eruption’s eruptive column and the predominant direction of ash dispersal meet predefined thresholds, an associated payment is immediately released to the cedent. This mechanism circumvents the need for costly damage and loss assessments and ensures prompt release of funds.
However, there are prerequisites that need to be met to implement such a solution elsewhere. First of all, a fully probabilistic loss model for volcanic eruptions must be available in order to obtain modeled losses and calibrate the solution for the specific setting. Furthermore, an official, reputable national or international agency must be able to report eruptive column and the predominant direction of ash dispersal independently.
Both of these prerequisites appear increasingly realistic to meet on a global basis, given the increased capabilities in the area of volcano modeling enabled by the existence of open source hazard simulation models and data, as well as the simplicity in recording the eruptive parameters of interest.