Dr. Tom Fletcher, VP, Data Analytics, North America Life, discusses the benefits of multidisciplinary teams.
It’s my opinion that without fresh thoughts and methods from outside perspectives, we grow stale. While some progress requires incremental contributions from actuarial science and medicine, other progress and transformational shifts will only happen with input from other disciplines.
My background, for example, is in industrial/organizational psychology and statistics. As I lead a team in predictive analytics, I’m constantly scanning other fields such as economics, epidemiology, the hard sciences, and engineering, to name just a few, each of which can add value.
The next time you hear ‘We can’t solve that’, consider looking to other disciplines for inspiration.
On the “content” side, for example, consider the life insurance application process. There are design issues which underwriters have long known and addressed. What questions should be in the application? What order should these questions be in? How should the questions be worded? All of these are addressed by underlying research that has occurred in the behavioral sciences, such as psychology. The art and science of survey design therefore has a lot to offer.
Indeed, if we move across aspects of the customer journey and ask specific business questions, then look outward at other disciplines, we will see that many disciplines have content to contribute.
Experts from other disciplines bring content from their respective disciplines, and they also know how those “facts” were derived in the first place. They bring unique problem-solving skills. The best managers will leverage this functional diversity.
In insurance, we often have this expertise in-house, but unfortunately tend to overlook it. A predictive analytics team, for example, is often built with such diversity in mind, even so, data scientists are frequently pigeon-holed as math or engineering people. We must remember that predictive analytics does not “happen” in a vacuum; just as models should be developed in collaboration with underwriters, so should the dialogue around substantive be collaborative and inclusive of underwriters and predictive modelers, including openness to their wider discipline know-how.
Speaking across disciplines can in itself add value, because it forces us to be so much clearer with our thoughts.
On the “methods” side, research and analytic methods often evolve in response to solving a specific problem in a given discipline (e.g. time-series in economics, factor analysis in psychology and discrete-choice experiments in marketing). We can therefore learn a lot if we reach across to other disciplines to see how they went about solving their problems. It stands to reason that a multi-disciplinary analytics team will add value to underwriting.
Predictive analytics serves to improve organizations by creating efficiencies, consistencies or finding opportunities including improving risk selection. I often categorize our work into conducting research (gathering data) and analytic methods (deciphering the data). Methods that facilitate hypothesis testing, field experimentation, program evaluation and otherwise ruling out plausible alternatives given the data, have evolved differently. Double-blind, placebo-controlled methods are great, but when they are not feasible, the need to increase confidence in analytic results does not simply go away.
Activities within insurance that impact underwriting include propensity to buy, the inclination to be (dis)honest, and uncovering risks associated with mortality or morbidity. The field of medicine has relied on underlying core sciences, as well as developed techniques specific to addressing statistical challenges. Likewise, other disciplines have developed statistical techniques to address each of these activities: Consumer psychologists have developed and improved upon techniques for creating personas; statistical learning has evolved techniques to ensure rigor is applied to prevent overreacting to variances in data (risk assessment could be erratic in some samples).
The list goes on. The main point I want to make is this: Methods to address new issues are sometimes insufficient due to limitations in the analytic techniques, smart people will develop new methods in response, but many new methods have in fact been evolving for centuries. So, the next time you hear “We can’t solve that”, consider looking to other disciplines for inspiration.
Another benefit of functional diversity is the many bodies of ethics that come together. The principles of objectivity, integrity, and fairness are common to most; in the US for example, whether these be from the American Psychological Association, American Statistical Association, or even the US Office for Human Research Protections.
And finally, speaking across disciplines can in itself add value, because it forces us to be so much clearer with our thoughts, something that can bring to light issues that we weren’t previously aware of.
So, there is certainly much to gain from outside perspectives if we can (1) recognize this, (2) reach out, and (3) welcome multidisciplinary expertise into our discussions.
There’s a big mindset component to all of this. A mindset that’s open to change, challenge, and differing perspectives. A willingness and ability to engage and have the tough conversations. Recognition that we’re better together.
VP, Underwriting Strategy & Innovation, US Life
Please contact the author if you’d like to discuss this topic
VP, Data Analytics, North America Life
This article is an abbreviated version of an article recently published in Hot Notes (Volume 21, Issue 3, March 2021).
Opinions expressed are solely those of the author. This article is for general information, education and discussion purposes only. It does not constitute legal or professional advice and does not necessarily reflect, in whole or in part, any corporate position, opinion or view of PartnerRe or its affiliates.
Q1. What level of underwriting automation can reasonably be achieved in your target market? Why is this question important? A straight through processing (STP) rate […]