Professor Phelim Boyle shares some of his favorite memories from his association with ARIA, where his research from one of his most cited papers – “Options: A Monte Carlo Approach”- has led him now, what advice he would give to new academics and how we can help promote diversity in the field. Phelim has been awarded the Centennial Gold Medal of the International Actuarial Association, the Gold Medal of the Institute and Faculty of Actuaries, and was the recipient of the IAFE/SunGard Financial Engineer of the Year in 2005. In 2019, he was named a Fellow of the Royal Society of Canada
1. As you are approaching retirement, what are some of your favorite memories being associated with ARIA?
My first job was with Irish Life Assurance Company in Dublin and I have had an interest in insurance ever since. I have fond memories of interactions with ARIA members. I am also grateful to the Journal of Risk and Insurance for taking a chance on one of my papers when no other journal was willing to do so.
Here is the story behind this paper. When I was working in Dublin, I became familiar with investment linked polices. The returns on these contracts were linked to those on an equity portfolio. In addition, the insurance company guaranteed to make up for any investment shortfall at maturity. The actuarial methods in use at the time were unable to price or reserve for these products in a theoretically sound way. I knew that there was a problem but I had no idea how to solve it. A couple of years later I had moved to the University of British Columbia and during a seminar mentioned these maturity guarantees. Michael Brennan, who was in the audience noted that this guarantee was similar to a put option that could be priced and hedged using the newly minted Black Scholes methodology. We worked on this together with a PHD student, Eduardo Schwartz. We wrote a paper on our solution to this topical and actuarial problem and sent it to several actuarial journals where it was roundly rejected. I was delighted when the Journal of Risk and Insurance accepted the paper, which was entitled: Equilibrium Prices of Guarantees Under Equity-Linked Contracts. This method is now used by insurance companies all over the world to hedge their investment guarantees.
Over the years, I have had many discussions with ARIA members on a wide range of topics and managed to attend a few meetings. I still have warm memories of sipping tequilas with Bob Witt and his wife in the hill country outside Austin several years ago. I have wandered down memory lane with Arnold Shapiro when we meet at conferences. In recent years, I have attended joint ARIA-APRIA international conferences and enjoyed hanging out with David Eckles and my fearless colleague, Mary Kelly. It has been a pleasure to chat with Michael Power of Tsinghua University at recent conferences.
I have been very fortunate to work with some outstanding students during my tenure at the University of Waterloo. Four of these students: Johnny Li, Ken Seng Tan , David Li and Shaun Wang have gone on to become shining stars in the insurance and actuarial science fields. I have also been privileged to know Pat Brockett and have benefitted from his wisdom and insights over the years. I have written several papers with Carole Bernard, including one on the Madoff Ponzi scheme. Weidong Tian and I have had a fruitful collaboration. Weidong is a strong mathematician and uses results like the Lebesgue dominated convergence theorem in his proofs.
I have also got to know Anne Kleffner through our membership of the Actuarial Profession Oversight Board. This independent body oversees professional conduct, discipline and standards of the Canadian actuarial profession. Obviously, this serious Board deals with weighty matters. However, I was wondering if it would lighten the atmosphere of our Zoom Meeting if all the Board members shaved their heads following Anne’s example.
2. When you wrote “Options: A Monte Carlo Approach” did you think it would be as impactful as it was? Where did your research lead you after the Monte Carlo paper?
There is an interesting background to the Monte Carlo paper. When Eduardo Schwartz was working on his PHD thesis, I wanted to check some of his numbers that involved solving a partial differential equation. I wanted a quick method to do so and I had just read a preprint by John Cox and Steve Ross that showed that an option price could be written as an expectation. It was quite natural to use Monte Carlo to compute the expectation. I had no idea that this would become a widely cited paper.
Since then I have carried out research in a number of areas. These include mathematical finance, portfolio theory, insurance and demography. In one of our papers, we analyzed different mortgage contract designs. One problem with the standard fixed rate mortgage is that when house prices fall and incomes decline, the homeowner has a strong incentive to default. This was dramatically illustrated during the 2008-2009 global financial crisis. We analyzed modifications to the mortgage contract design that can reduce the incentives to default and avoid the deadweight costs of foreclosure. I am also involved in applied actuarial work and was the independent actuary for the FHA’s mortgage insurance program for several years. Currently I am working on a problem involving the eigenvalues of correlation matrices. As far as I know, this has no practical application but it is fun. It is something to think about when I walk my dog.
3a. What advice would you give to new academics?
Be curious and do not be afraid to challenge the conventional wisdom having done your homework.
3b. What can we as insurance academics to do promote diversity in our field in response to Black Lives Matter?
Work hard to get rid of bias, especially unconscious bias.
I think insurance academics have a responsibility to critique the industry. Currently big data and artificial intelligence are widely used by insurance companies. Big data raises important ethical issues. The potential abuses of big data have been highlighted by Cathy O’Neil, author of “Weapons of Math Destruction”. She notes that some of these models codify biases and exacerbate inequalities. The embedded algorithms may use factors that act as proxy for race.
A recent paper entitled, Discrimination-Free Insurance Pricing, by Marcus Lindholm, Ronald Richman, Andres Tsanakas and Mario Wuthrich, develops a model for non-discriminatory insurance pricing. Their model assumes that certain information such as race is not allowed to be used for insurance pricing. This could be a step in the right direction.