Insurance often gets a bad rap. For most of us, dealing with insurance companies is a necessary evil defined by unpleasant tasks such as filling out paperwork, defending claims, or spending money on ever increasing premiums. 

For actuarial science researcher Silvana Pesenti, assistant professor in the Faculty of Arts & Science’s Department of Statistical Sciences at the University of Toronto, insurance—and specifically risk assessment and regulations within the sector—is an exhilarating topic that she loves to spend time researching and talking about. 

“When I tell people that I study mathematics and insurance, they often look at me sceptically, but from a mathematical point of view insurance is exceedingly interesting, because it’s such a complex topic,” Pesenti says. 

“If you price a policy, you need to take a lot of different risks into account; the habits of people, particularly in health insurance, drastic changes in the weather, political tensions, all of it can play a significant role in how to assess risk.”

Much of Pesenti’s work focuses on helping insurance companies to navigate the complex process of understanding risks. Her research vision is to develop reliable statistical tools to assess the precision and accuracy of model results, and which can easily be used by people working in the industry. 

“Many mathematical models which are currently in use in the insurance sector are outdated or are not fit-for-purpose. If you’re interested in calculating your loss, you want to use a model that’s good in modeling loss specifically—and not, let’s say, capital gain. Otherwise, your results might be completely misleading.”

Faulty risk assessments can have dire consequences—for individuals, but also for society at large. 

Accurately calculating risk helps insurance companies treat their customers fairly and equally by, for example, not overcharging on premiums and prevent insolvency, which may leave people stranded without health insurance or pensions. Even more importantly, risk management can help prevent catastrophes, such as the 2008 financial crisis, which led to millions of people losing their life savings, family homes, and jobs. 

“One aspect that led to the financial crisis was companies (mis-)using standard models that clearly didn't work well enough and the inadequacy of risk management,” Pesenti says. “If you use models that aren’t accurate, it can hurt an entire sector. That’s true for calculating risk in the insurance sector but also in finance.”

Recently, Pesenti and her fellow researchers developed a tailored and consistent risk assessment methodology that includes a shareable, open source code and data package, ready-to-use by anyone familiar with the coding language R. 

“Practitioners don't necessarily have the time or mathematical background to keep up with the latest research advances and implement them in a programming language. That’s the gap our R package fills,” she says. “And it’s really one of the aspects I love most about my research. It’s huge potential of industry applications and that it builds bridges between academia and people in the industry.”

In only four months, Pesenti’s R package has been downloaded more than 900 times following its online launch in July, allowing insurance companies to do so called “reverse stress testing” of their insurance portfolios, in addition to more conventional stress testing. 

“Normal stress testing is scenario-based, providing answers to questions such as, how will the entire insurance sector perform in a bad economy? Reverse stress testing is a lot more company-specific. A company can test which major risks will drive the company into insolvency,” she says. “Our methodology is one of the first to provide a mathematically consistent foundation for both stress testing and reverse stress testing.” 

A company insuring, for example, people affected by wildfires, can use Pesenti’s latest research to more accurately assess their performance if a catastrophe strikes—and make necessary changes to their policy based on the results. That way, insurance companies can minimize risk and even plan for scenarios where one adverse event triggers another, such as an earthquakes leading to a tsunami. 

Pesenti’s groundbreaking work in actuarial science, the discipline of applying mathematical and statistical methods to assess risk, recently earned her the prestigious Dorothy Shoichet Women Faculty Science Award of Excellence. The award supports up-and-coming female researchers by offering teaching relief, allowing women in academia to fully focus on their research.

“The Faculty of Arts & Science is pleased to congratulate Professor Pesenti on this terrific award that will allow her to advance her impactful research program that connects world-class fundamental statistical research to insurance practitioners,” says Jay Pratt, vice-dean of research & infrastructure in A&S.

In addition to helping Pesenti focus on her passion for research, early career awards supporting women in STEM (science, technology, engineering, and math) can play a role in diversifying the field, she says. 

“There aren’t many women in my field and it’s especially important to get a boost early on to thrive in your career. You need to network, collaborate, and speak publicly about you work. This award will give me so much more flexibility and time to share my research with the people to make a difference and to put it to good use.”