The tagline on Rudi Van Delm’s LinkedIn profile is “Improving people’s lives through insurance”. This describes Rudi’s unique approach—while successfully leading pricing for some of the largest global insurers, he has always been guided by how to improve people’s lives.
Rudi has been in general insurance for more than 30 years. He has worked and lived in Belgium, Germany, the U.S., the UK, and now the Netherlands. His career includes positions ranging from an assistant university teacher to a chief actuary, to managing the pricing and underwriting function for the largest personal lines general insurer in the UK and NL.
Rudi is fascinated by pricing where he sees business needs, customer interests, commercial interests, high tech modelling expertise, data science and employee engagement all come together.
It is fascinating to discuss pricing and data with Rudi.
A Quote from Rudi
“Data does not solve everything; expert human judgement is critical! We need to be very careful in how we use data, as there are several pitfalls—including unwanted bias and discrimination of groups. The past is not always the best guide for the future.”
David Schapiro (DS): Could you please tell us a bit about yourself and your career?
Rudi Van Delm (RVD): I have always loved mathematics and teaching. I started as an assistant university math teacher, but I didn’t want to be a teacher for my entire life. I was introduced to the actuarial world by a professor, and I found it to be a wonderful way to translate statistical theory into the real world.
When I started as a professional actuary, the world of pricing intrigued me—it has real impact on customers, shareholder value and is the ultimate application of statistics in real life.
With the opening of the European borders in 1992, I wanted to develop an international career. Luckily, my wife studied international economics, so she was very interested in going international as well. We moved with our two kids to the U.S. in 1995, and to Germany and the UK after that.
In my spare time, I try to spend as much time as possible with my family, which now also includes a granddaughter! It’s great to go back to teaching; now my student is a seven-year-old…
DS: Could you please share with us your view on the importance of data in insurance?
RVD: Data is probably more important than the analytical modeling. Data is the basis of all the calculations for pricing, underwriting and much more in insurance. The analytical modeling can only be as a good as the data it is based on.
Data is a complex dynamic concept, with several dimensions. There is the amount of data—number of policies, customers, prospects, sales, etc. And for each of these there is additional information we want to know—car model, address, claims history, type of business, etc. And this is all limited to what is legally useable and permissible by the regulator in the specific country or region.
The decision of what data to use is not only based on legal and regulatory constraints, but also on ethical considerations—with some grey areas.
DS: Do you see a difference between enhancing our current data set and searching for new data?
RVD: Some insurance companies tend to be ultra-focused on finding more information within their existing traditional data sets. It is compelling for pricing departments to update their models based on more of the same data, like including an extra year in the calculation. But the holy grail is finding new data and new factors—and how to use them in pricing.
We should shift more focus to creatively finding new data and factors. This culture of creativity needs to exist and flourish in the pricing & underwriting department.
Pricing analysts should strive to find new data sets and sources—structured and unstructured. Maybe we could find the “one factor” that on its own accurately predicts the risk, or just significantly improves risk assessment?
DS: If we achieve the ultimate in accuracy and coverage of data—can it do everything on its own?
RVD: Data does not solve everything; expert human judgement is critical! We need to be very careful in how we use data, as there are several pitfalls—including unwanted bias and discrimination of groups. The past is not always the best guide for the future.
Things that seem mathematically basic, like how to deal with inflation in pricing, require human judgment. How do we price to predict a future risk with inflation, based on historic data with no inflation?
Good data is also good for the customer—it provides full pricing transparency, no surprises, and a price that is linked to the risk. It enables improved internal processes, being right the first time, better at setting the cost, thus lowering expenses for the insurer and the customer.
But accurate and extensive data is not a panacea—we need to understand the data we are using. For example, in some countries insurance history shows there are more claims from people that live far away from their agents/brokers. What is causing this? Maybe it is a result of people moving to different locations and changing their livelihoods, losing familiarity with the neighborhood, or having additional stress? As insurers we should strive to understand and be able to explain the reasoning behind every factor. We need to provide pricing explainability.
DS: How can we continuously use this new data while keeping it accurate and up to date?
RVD: Achieving extensive and high-quality data is a joint effort. The pricing team initially gathers the data, and the data engineering team maintains and updates the data. This process must be well defined and managed, as there is a massive amount of data that needs to be gathered, cleansed, and kept accurate and up to date.
For analysis, slightly unclean data sets can be used. But for implementing pricing accurately, clean data is required. The data has a direct impact on the price of the individual insured. Recency of data is critical—recent data is much more accurate than older, less-recent data. For example, claims data from last year compared to claims data from years ago. Customer data must be completely up to date.
Although new data sources require time to integrate into existing insurance IT systems, we should constantly search for new data sources and factors. The pricing team should not restrict itself to IT limitations. The goal should be to make the best model you can, then scale it down to something that is relatively fast to implement while providing IT with a priority of when to implement/integrate new data sources.
DS: Could you please share with us your thoughts on data ethics and the solidarity of people that are insured?
RVD: There are data factors that the insured/customer can influence—driving more safely, for example. And there are other factors that the insured/customer cannot influence, such as medical diseases. An interesting rule of ethics could be: If the insured can influence the factor, then it is okay to use; but if the insured cannot influence the factor, you have to be very careful using it.
This takes us to the concept of solidarity within insurance, and that is that over time the ‘good’ risks pay for the ‘bad’ risks. Good and bad can be defined in many ways—people without claims will always pay for people with claims. However, people with exemplary behavior should not pay for people with poor behavior. Regarding health insurance in this context, healthy people feel okay paying for sick people’s claims, as we are paying insurance to cover us if we become sick. But we don’t have the same solidarity to pay for the claims of the unsafe/bad drivers in car insurance.
If you price well, you can positively influence the insured’s behavior—improving their insurance factors and data. An example of this could be helping people drive more safely using telematics and appropriate insurance pricing. A similar approach could leverage feedback data from the internet of things (IoT) in homes, creating a safer community for everyone.
DS: The tagline on your LinkedIn profile is ‘Improving people’s lives through insurance.’ Could you please elaborate on this in the context of data?
RVD: As insurers we should ask ourselves how we could positively influence the insureds to reduce their risk of having a claim.
This can be done by including the right data in our pricing—e.g., if taxi drivers are bad risks because they drive many miles, then we can improve their driving with telematics pricing and reduce their risk. It is win-win for the taxis, the insurers, and every pedestrian and driver. The same is true for truck drivers and other commercial fleets.
A similar example is for businesses and people that take good care of their properties. This is also true for product liability, cyber, and other commercial and personal line insurance products. Doing it right will result in safer houses and buildings, better products, safer IT, etc.
DS: Are there any final thoughts you would like to share?
RVD: One fairly recent influence in the field of data is what we, as insurers, will know about our clients’ sustainability activity or targets. This is another place where pricing and underwriting may help society deliver better outcomes for citizens.
Rudi Van Delm – Bio
Rudi Van Delm is responsible for pricing & underwriting at Nationale-Nederlanden Schade & Inkomen.
Previously, Rudi Van Delm was director of pricing and underwriting at Direct Line Group, one of the largest motor and home insurers in the United Kingdom, 2011-2018. In this position, he led the pricing and underwriting teams for Direct Line, Churchill, Privilege, and Green Flag, as well as partnerships with Sainsbury’s, RBS, NatWest, and Nationwide, among others.
Prior to the Direct Line Group, Van Delm was chief actuary for non-life insurance at Lloyds Banking Group, where he also sat on the executive committee and the board of directors for non-life insurance. From 2001 to 2009, he worked at EY in London, first as a senior manager and later as director of the actuarial department.
Rudi is married with two children and one grandchild. In his spare time, Rudi likes to spend time with family and friends, do some recreational walking and cycling, or read a good book.