Over the past few months, the bloggers at Future Fellows and the Candidate Liaison Committee have written about the role of actuaries in predictive analytics. We conclude the article series with this final interview, and we hope you have enjoyed and learned from the various perspectives of actuaries on analytics and data science.
We sit down with Ling Tan, an actuary at Zurich North America. He started his career in predictive modeling, then transitioned into a traditional pricing role.
Please log in to leave your comments or questions. In a few weeks, stay tuned for our next article series on networking in the actuarial community!
Future Fellows: How did you get started in predictive modeling?
Ling: When I was in college at the University of Illinois, I had a predictive modeling internship at an insurance company.
Future Fellows: What experiences helped you prepare for that internship?
Ling: What really helped me were my math and statistics classes, and the one or two actuarial exams I had already passed. My courses developed my interest in modeling.
Also, I joined the math club. I participated in math competitions, where I had to “think outside the box” and get used to tackling challenging questions.
Future Fellows: It seems that internship helped you land your first job in predictive modeling.
Ling: Yes, the internship gave me valuable experience. During my internship, I helped build a traditional pricing model for the personal auto line.
When I joined Zurich for my first full-time job, the prior internship helped a lot. I had prior background in data modeling and analytics. By then, I had also taken more courses in statistics, math, and basic programming.
Future Fellows: How long were you in that role? What were your main responsibilities?
Ling: I was in predictive modeling for roughly three-and-a-half years. I mostly focused on pricing models for various lines of business, but I also did a few projects on claims and risk selection.
Future Fellows: How did you transition into a traditional pricing role?
Ling: Most of my experience in predictive modeling was with building pricing models, so it made sense for me to transition to the “other side” of the process. Modeling provides a technical component of the pricing, but turning that model into a practical solution requires something outside of the modeling domain. Actuaries push the pricing model into implementation, and actuarial ratemaking is a fundamental skill. Ultimately I wanted to become a well-qualified actuary, closer to the business and operations.
Future Fellows: How does analytical work differ from traditional actuarial work?
Ling: Modeling requires knowledge of advanced statistical methods and substantial programming skills. We have to be able to handle a large amount of data, whether internal or external. The modeling work is driven mainly by a statistical background, but data mining is also growing in this field.
Implementing our models is also a complex undertaking. We have to rely heavily on our IT colleagues for ongoing support.
As for actuarial work, we can support the communication of the models. As the models are usually quite complex, we need to be able to translate them to a non-technical audience. This should be a joint effort between analytics and actuarial teams. The complexities can be explained by the analytics team, but actuaries can translate the messaging in a way that is easy for others to understand.
Future Fellows: You are now attending a graduate degree program.
Ling: Yes, I am attending the University of Chicago’s Master of Science in Analytics program.
Future Fellows: What have you learned that is not covered by the actuarial exams?
Ling: My graduate program is expanding my statistical knowledge. It also allows me to learn advanced data mining skills. The actuarial exams don’t teach programming or data management, so now I can fill those gaps in school.
The program is also teaching me key leadership and teamwork skills, and it never hurts to become better-rounded.
Future Fellows: When would you encourage other actuarial professionals to go to graduate school?
Ling: It’s a personal choice that each person has to make on their own. For those that are interested, I would encourage them to think about the time commitment, especially when they have actuarial exams or family commitments to think about. Going to graduate school requires you to manage your time really well.
Instead of graduate school, there are other resources that could help you pick up the skills you want, such as online classes or reference materials. You could attend CAS seminars or learn from training sessions. The new CAS Predictive Analytics credential is also an option, but it’s still in development. I would encourage those who are interested to look out for more information from the CAS.
We thank Ling for his time, and we hope you enjoyed our article series on predictive analytics!