Captive solutions for complex risks in the absence of extensive loss history: an actuarial approach

Sol Feinberg and Enoch Starnes of SIGMA Actuarial Consulting Group review some creative approaches to viewing risk management.

The captive insurance space serves as an enticing option for entities faced with emerging or complex risks which often have limited amounts of historical data points. While this presents obstacles for actuarial analytics using traditional methods, it doesn’t mean actuarial techniques are impossible to use. Creative approaches with regard to research and analytics may offer solutions for covering risks in the absence of an extensive history of unique losses or even industry losses. This article addresses how such a challenge may be approached from an actuarial perspective and contrasts the approach with more common actuarial methods.

For a traditional risk such as workers’ compensation, when a significant history of losses is available, loss “triangles” can be compiled from company-specific loss runs at successive annual valuations. Each loss run contains a list of claims, their date of loss, the cumulative amounts paid on the claim, and case adjusters’ estimates of remaining future payments. The loss development triangles are a starting point for traditional methods. However, for complex emerging risks, the requisite history of a sufficient number of claims developing over time to provide statistically credible patterns may not be available.

Another common traditional technique is for an actuary to compare ultimate losses to an exposure measure—payroll is usually used for workers’ compensation. This gives a pure loss rate which can be used to project losses for an upcoming period. The concept of a pure loss rate or amount of ultimate losses per an exposure measure may still be useful for a complex emerging risk, but a history of ultimate losses for prior periods, which can be trended to the desired projected period, may not be available.

For workers’ compensation, there are bureaus that aggregate state-specific loss data for the industry and publish statistics, including loss development factors and pure loss rates by classification code. These bureau statistics can be used to supplement the available company-specific data.

In addition to looking at total loss dollars paid or incurred for the group of claims with dates of loss in a given accident year, actuaries might also look at the number of claims, the number of claims with loss payment, or the number of open claims. Looking at the number of claims per exposure gives the frequency, and looking at the losses divided by the number of claims gives the severity (or average loss size per claim). 

Thus, we can think of our pure loss rate as a product of frequency times severity:

Pure Loss Rate=Losses/Exposure

=(Number of Claims)/Exposure∙Losses/(Number of Claims)

=Frequency∙Severity

Useful information

Even when a risk has little historical loss data available, a company may still have other useful data points that an actuary might consider in their analysis, eg, commercial rates, quotes, or other internal company information. Qualitative discussions between company management and its actuary will often help guide the way these alternative data points are gathered. In some cases, a questionnaire may help streamline the collection of qualitative information, or the company management may be familiar enough with the risk in question that they can suggest sources of relevant information or provide information they may have access to.

A company that is looking to self-insure, for a complex or emerging risk, likely has no loss history. If this potentially large risk is currently emerging, there are likely no industry statistics available either. In this case, concepts from the traditional measures, such as exposure, or frequency and severity, are still useful, but employing them may require some creativity. 

An actuary researching liability (or suit-focused) risks might look to a legal database for related cases. This could be important, at least qualitatively, to demonstrate the relevance of the coverage and the existence of the exposure. That said, successful suits resulting in damages may not have yet penetrated the legal system. In this case, similar suits in related industries might be relevant.

If damages in suits are based on cumulative operations, it may be that the potential severity of a suit scales with the size of the company or with its cumulative operations over time. It could also be that the severity of a claim is not proportional to the size of the business overall, but the number of claims may be. For example, consider two construction companies of different sizes but with otherwise comparable operations. If the process giving rise to a claim is related to actions of an individual worker in one place and time, severities for either company could be similar, but the larger company would have proportionally more claims.

Different sources of data can be looked at as well. In addition to legal databases, a general internet search can be useful. Related news articles can be very beneficial in terms of introductory information and could be used to enhance the scope of research as relevant terms, and better sites and authors, are discovered.

Regardless of the approach used, an actuary should always apply Actuarial Standard of Practice number 23, Data Quality, in determining whether data is sufficient, appropriate, and reasonable to use. Data should be reviewed to determine whether its obvious characteristics appear reasonable and consistent. Significant data limitations or data adjustments should be disclosed.

Captives have historically been sought as a solution to address risks that have never been addressed (or potentially not been seen) before. While this may present the types of obstacles or complexities outlined in our article, it also offers an opportunity to broaden the capabilities presented in the captive insurance landscape. Complex risks may require new approaches, and new approaches almost always lead to a stronger industry.

Sol Feinberg is a consulting actuary at SIGMA Actuarial Consulting Group. He can be contacted at: sol@sigmaactuary.com 

Enoch Starnes is a captive and complex risk consultant at SIGMA Actuarial Consulting Group. He can be contacted at: enoch@sigmaactuary.com

Did you get value from this story?  Sign up to our free daily newsletters and get stories like this sent straight to your inbox.