Stephen Mildenhall, Ph.D., Assistant Professor of Risk Management and Insurance, and Director of Insurance Data Analytics, School of Risk Management, Insurance, and Actuarial Science
January 25, 2019
Stephen Mildenhall has been an Assistant Professor at Tobin’s School of Risk Management, Insurance, and Actuarial Science (SRM) since 2016 and teaches courses in predictive analytics and actuarial science for both the undergraduate and graduate programs. He earned his Ph.D. in mathematics at the University of Chicago.
Measuring insurance risk—and then figuring out how to share it—is no easy task. No one understands this better than Dr. Mildenhall. He is currently testing a model that would make it easier to understand and price insurance by breaking down the conventional portfolio model into smaller pricing layers.
“Risk is complicated. There are a host of different measures to quantify it,” said Dr. Mildenhall. “The problem is to figure out how to share risk: what to retain and what to transfer to others through insurance or reinsurance.”
Trained as an actuary, Dr. Mildenhall spent the last quarter century in the insurance industry, working for various insurance carriers, and as a reinsurance broker. His work centered on figuring out how to measure and price risk, particularly catastrophic risk. Risk sharing and allocation and risk pricing are real-world problems with important real-world applications, according to Dr. Mildenhall.
“The real practical question is how you share risk between the risk owner, insurance and reinsurance companies, and how you allocate costs of risk transfer back to the policyholder. That has a very real impact on the pricing of insurance that everyone pays,” said Dr. Mildenhall.
It is difficult to understand and price the different layers of an insurance portfolio, so Dr. Mildenhall’s proposal is to break up that portfolio into smaller layers. This is analogous to how bonds are typically tranched, Dr. Mildenhall explained.
“There is a senior, AAA-rated tranche, which is unlikely to default and which gets the lowest coupon. Below that, you have layers more likely to default with lower ratings and higher coupons. Typically, these are sold to different types of investors,” said Dr. Mildenhall.
Breaking up the total risk into smaller and simpler pieces would make it easier to assess and price the total risk.
“Each slice is completely described by its probability of default and can be priced on that basis,” said Dr. Mildenhall.
A two-time winner of the Casualty Actuarial Society’s Woodward-Fondiller Prize for the Best Research Paper by A New Fellow, Dr. Mildenhall said his theory is nearing completion.
“I have a very solid draft, but it needs more work before I can test some of the predictions against what you actually see in the marketplace. Once I get the draft done, I want to present it and talk about it to gain some traction,” he said.
Dr. Mildenhall believes the model would offer practical guidance to the insurance companies trying to figure out the optimal structure for their policyholders, and the regulators concerned with the solvency of insurance companies.
“Regulators require companies to disclose their estimated risk capacity. They need a risk measure to do that,” he said.