OPINION: Big data is key to next-generation life insurance

Published Feb 18, 2019

Share

While most people believe that life insurance underwriting is a hugely complicated process based on very scientific principles, the reality is that it has long been more of an art than a science. While in most cases, the actuarial requirements to build out the various curves related to risk, pricing, and so on, are complex, the need for life insurance applicants to go through onerous medical screening processes adds a significant layer of subjectivity to the entire process.

The results of these health screenings, and the responses provided to the accompanying medical questionnaires are typically used by specialist underwriting assessors, many of whom have backgrounds in medical fields, to evaluate the risk profile of the individual and make decisions regarding their suitability for cover and, if granted, the price of such cover relative to the assessed risk level they present.

While the entire process typically takes place within the context of a risk-assessment “rule book”, it is generally far less scientifically based than one would expect and is not usually built on clearly defined statistical risk indicators. This often ends with one or more insurance experts “making a call” regarding each application, based on the information the applicants provide.

Apart from the fact that this approach is not particularly efficient, it is very administration-intensive - both for customers and insurers - demanding time and human resources.

Taking all of this into consideration, FNB Life invested a massive amount of time and expertise into understanding how the underwriting process could best be transformed to achieve a simpler and more pleasant customer experience; creating opportunities for life insurers to reduce their costs and administrative burdens.

Our point of departure was whether life insurance could ever achieve a level of ease and convenience similar to the “one-click” purchasing experience that Amazon offers its customers.

Essentially, the Amazon platform allows a customer to select and pay for a product, and have it delivered to the address they have provided, all with a single mouse click.

While this is a relatively easy transactional capability to set up in the retail environment, the underwriting demands of the life insurance industry make it significantly more complicated.

The challenge, of course, is if an insurer wants to simplify the process of buying life insurance to the point where the prospective client can do so virtually instantly. This requires some serious thinking around what medical and personal information is essential, and how such information can be obtained without putting the client through the traditional assessment processes.

This naturally also requires access to alternative data sources, preferably with client information that is more predictive than the traditional medical questionnaire. While much of this predictive underwriting information exists, accessing it often requires agreements with third parties, and can also present obvious client protection and privacy implications.

However, our investigations into the viability and value of predictive underwriting revealed that FNB actually holds a significant amount of information on the majority of its clients that they have explicitly given us permission to access and use.

While, at first glance, it didn’t appear as if much of this data would offer value in terms of our efforts to make instant life insurance a reality, we soon realised we could build highly personalised and accurate predictive risk models for many of our banking and investment clients. For example, relatively simple analyses of clients’ credit card spending patterns offer excellent insights into their lifestyle and eating habits and can be a very valuable indicator of their long-term health expectations.

The net result of this intensive statistical analysis of the client data we have at our disposal is that FNB Life is now in a position to offer pre-approved life insurance of up to R1.5million, without the need for any medical questions or screening, to roughly 15% of the FNB client base. In addition, this predictive underwriting approach has allowed us to develop near-instant life insurance offerings for a further 30% of FNB clients, who will now be able to acquire life cover up to R3m from us simply by answering five basic medical questions.

Importantly, these simplified application processes have not added anything to the cost of cover. In fact, we believe that as we further refine our predictive underwriting model and gain access to more client data sources, we could succeed in driving down the overall cost of life insurance for our clients.

While our experience around understanding and implementing this type of predictive underwriting revealed that it is still a relatively unknown concept globally, the potential benefits for insurers and clients are patently obvious.

Instead of a complicated and expensive grudge purchase, the appeal of tailored and flexible life and disability cover as cornerstones of personal financial security could be greatly enhanced, and the barriers to entry will effectively be removed.

Lee Bromfield is the chief executive of FNB Life.

SUPPLIED

Related Topics: