Evolving Big Data and analytic technologies are putting pressure on insurance companies to mitigate the risks associated with their actuarial and business models.
“Technology is helping insurers get better answers to their actuarial questions, but it’s also dramatically increasing the complexity of insurers’ actuarial models,” according to Henry Essert, U.S. risk and capital management leader at PwC. “That complexity introduces a higher degree of risk into model implementation—which, in turn, drives the business case for more codified model risk management best practices,” he maintains.
Codified best modeling practices might also help insurers mitigate legal and regulatory risk, Essert says. Without such codification, insurers can only make general claims of best-effort diligence if and when they are faced with litigation or regulatory audits. Documented conformance to a recognized set of model risk management best practices, on the other hand, could potentially constitute a legally substantive response to such exposures.
A new paper authored by Essert pushes industry executives to take a closer look at how they ensure the quality of those models—as well as the reliability of the processes they use to create them. Entitled “Model risk management: The next generation for insurers,” the paper argues for better codification of best practices for model governance in the insurance sector. Without such codification, Essert believes insurers individually and collectively may not sufficiently mitigate the business risks associated with the various problems that can crop up in scenario modeling—which include inaccurate or erroneous inputs, flaws in the algorithms used to calculate outputs based on those inputs, and weaknesses in the way models are applied by decision-makers to specific use-cases.
In an exclusive interview with INN, Essert discussed the potential impact of model-related risk on insurers. “A bad model for a single product can cause an insurer to lose money due to under-pricing or inaccurate claims projections,” he says. “The bigger risk, however, is that an insurer will consistently experience such losses more broadly across the entire business. That’s why model risk management best practices are so important.”
Essert’s paper sharply contrasts the state of model risk management in the insurance and banking sectors. The Federal Reserve Board’s Supervisory Guidance on Model Risk Management SR 11-7 provides banks with a fairly detailed description of best practices for model development, use, and governance. Essert asserts that, while SR 11-7 is very useful for banks, it does not address the actuarial modeling issues faced by insurers. He therefore suggests that insurers could benefits from a comparable set of standards.
“Anyone with a stake in the financial performance of an insurance company—including executives, board members, and shareholders—would be able to take comfort in knowing that the company’s use of modeling conformed to a known set of best practices,” Essert says.
Essert also suggests that organizations such as the American Academy of Actuaries, the Society of Actuaries, and the Casualty Actuarial Society could play a lead role in the development of these standards.
