Data Analytics for Captive Insurers: Key Considerations

Editor's note: A little over a year ago, we wrote an article, "Big Data: What Captives Need To Understand," based on a session at the Captive Insurance Companies Association's annual conference. We recently stumbled upon the website for a company called TDWI, which directed our thoughts back to big data and data analytics for captive insurers. TDWI is a company that "advances the art and science of realizing business value from data by providing an objective forum where industry experts, solution providers, and practitioners can explore and enhance data competencies, practices, and technologies." Similar to the National Association of Corporate Directors, TDWI can serve as a valuable resource to captive insurers as they wrestle with the question of data analytics.

Will Captive Insurers Use Data Analytics?

As a refresher for captive insurers already looking at data analytics or for those captive insurers new to this arena, we would suggest reading Captive Insurance Company Reports October 2016 edition's lead article, "Captives Capitalize on Analytics," by Robert J. Walling, principal and consulting actuary at Pinnacle Actuarial Resources, Inc. In the article, Mr. Walling seeks to answer the question "Will captives use analytics?" He outlines the numerous opportunities where captives can make effective use of predictive analytics and suggests, "For those with sufficient data (or access to relevant industry data), engaging in a deeper relationship with their actuary offers the opportunity to leverage this data not only in their captive's operations but within its overall risk management strategy." He also notes the numerous risks of operating without using analytics and points out the reasons there are so many missed opportunities for utilizing an actuary's expertise in predictive analysis.

Why Should Captive Insurers Care about Data Analytics?

Here is an excerpt from an interview with Risa Ryan, head of Strategy and Analytics at Munich Re America, Inc.: "Premium and loss data are the underpinnings and fundamental to the insurance business, its strategy, pricing, and profitability. It's everything we do and can't be replaced; it can only be enhanced....

"An organization that develops and maintains proprietary data can evaluate that data to optimize its risk appetite and risk portfolio. Developing and maintaining a storehouse of exclusive data can also play a role in helping the company identify any gaps either as organic growth in the current portfolio or in new products, and, thus, create new business opportunities. In-house data analysts and applications offer companies several strategic advantages. They can unlock insights into the business strategy and help steer the business to better enhance its market position. At Munich Re, we use our in-house resources for even more tailored solutions that better meet the needs of our clients, and positions us to create customized structures and products based on our experience with a particular client."

What Are Key Data Analytics Considerations for Captive Insurers?

TDWI has produced a report, Best Practices Report: Practical Predictive Analytics, which can be downloaded for free from the TDWI website. Readers of this article are strongly encouraged to download the full report because it provides a depth of information not possible to cover in this story. We will touch on some of the highlights presented in the report's executive summary.

The report opening statement: "Predictive analytics is on the cusp of widespread adoption. Many organizations are excited to make use of the power of predictive analytics (including machine learning) because they understand the value it can provide. However, although numerous organizations do use predictive analytics today, adoption remains elusive to many. In fact, TDWI research indicates that if users stuck to their plans for predictive analytics adoption, 75–80 percent of organizations would use the technology already although only 35–40 percent do so currently." So if your captive insurer is not currently using any predictive modeling, you are not alone based on the research.

What are the impediments to more widespread adoption of data analytics? TDWI suggests the following three.

• Skills development. Survey responses indicate that skills development ranked as the biggest barrier to predictive data analytics with 22 percent of the respondents listing it as the top challenge. For captive insurers, this is especially true. Single-parent captives may be more likely to already have individuals with the requisite skills serving in some capacity in the larger corporate entity, but for almost all group captives, the skill set would have to be procured from one of the captive's external vendors, most likely their outside actuarial firm. However, as analysis has become more important within the insurance industry, the competition for employees with the requisite background to build and populate these models has intensified. Group captives may need to rely on a more junior member of the actuarial team to assist them.
• Model deployment. In model deployment, users struggle with adequately understanding and defining the business problem that the model is being built to solve. TDWI suggests that companies first need to understand the issue and then work to hone and refine it. Questions such as "What is the problem statement?," "What is the outcome?," or "What is the impact?" should be explored before any attempt is made at actually constructing the model. For captive insurers, just getting to this stage requires the expenditure of considerable valuable time and effort. The captive board and captive manager may only meet once a quarter, so it may require dedicated meetings outside of the normal routine. It will also require buy-in from all of the likely participants.
• Infrastructure. This is the build versus buy decision referenced in the Munich Re America interview. While large reinsurers similar to Munich Re will have the in-house resources in terms of technology and people to develop data analytics, the vast majority of the captive insurance industry will not. For captive insurers, the question then becomes how to secure access to both the hardware and software necessary to compete in this space. Fortunately, with the growth of cloud computing and storage, this is not the problem it once was. However, captives need to understand any commitment to predictive modeling is not inexpensive. Therefore, before embarking on this journey, a captive would be well advised to have addressed the questions raised under model development.

There is no question that predictive modeling is here to stay. Captive insurers need to understand the concepts and costs inherent in this new technology.