As data analytics is playing a more important role in the insurance industry, we are interviewing actuaries and data scientists to share their perspectives on working together. A series of Q&A's will be posted, and we welcome any additional comments.
What questions do you have for actuaries in predictive analytics? To continue the conversation, we interviewed Rachel Hunter, who works with other actuaries and data scientists at a large insurer. She recently wrote for Future Fellows about the CAS announcement last Fall of a new credential in predictive analytics and data science. (“Actuaries and Data Science: Highlights from the 2015 CAS Annual Meeting”, June 2011, Vol. 17, No.2).
Future Fellows: What is the relationship between actuarial and analytics teams in your company?
Rachel: I work for a large insurer that has many teams using the tools of Data Science to address business problems. Many of these teams include both data scientists and actuaries. The team I am on is also a mix. The actuaries often provide a bridge between the modeling tools and the business customers, but we also play an important role in educating modelers with less insurance knowledge about the problems we are trying to solve and putting them in contact with other business partners if we do not know the answer ourselves.
Future Fellows: In your recent Future Fellows article, you mentioned that actuaries range from the “technical actuary” to the “business actuary”. How should actuarial candidates find their place on this broad spectrum? What steps should they take to develop the appropriate mix of technical and business skills?
Rachel: I would encourage candidates to get as wide of a range of experience as possible. Not only does the content of work vary between roles, but the process and flow of projects and deadlines can be very different. For example, more technical, research-oriented projects can be very long and have a risk of not resulting in implementable solutions. And there is value for the non-technical actuary to understand the mathematics and assumptions of models well enough to ask model builders the right questions, or to understand how the models the company implements may impact the data that the actuary is reviewing.
Future Fellows: Would you please share an example of where an actuary can use modeling techniques outside of traditional pricing and reserving?
Rachel: One example could be building a probability model to help focus claim adjuster efforts on open claims that are more likely to result in high settlements. This may not be the same type of model that would help with setting aggregate reserves, but it could be used to make sure that the right claim adjusting staff are assigned to more complex claims. Actuaries and other modelers help insurers build many kinds of models that can help streamline internal operations.
Future Fellows: The CAS recently announced a partnership with The Institutes to provide a predictive analytics and data science credential. What are the key benefits of this credential to students and employers?
Rachel: I think the credential will be of benefit to actuaries who want to understand models well enough to involve themselves in modeling projects more deeply. I am hoping it can also benefit data scientists in giving them an overview of key insurance knowledge that is particularly relevant to model builders. Those of us in the insurance industry already may often forget to tell modelers some key features of our data, such as how data may be impacted by limits, internal case reserving practices that can change over time, changes in policy form, etc.
We thank Rachel for her insights, and we invite you to share your thoughts in the comments section below.