Data-driven

pricing

Opportunities in the non-life insurance market

The market competition in the field of Dutch non-life insurance is
fierce. The ability to compare insurance policies online enables drs. Rinze Valkenburg MSc AAG
consumers to find the lowest prices for the products they need. This is a senior manager at KPMG Financial
Risk Management.
compels insurance companies to quote increasingly competitive valkenburg.rinze@kpmg.nl

premiums, whilst claims and operating costs show an increasing

trend, resulting in decreasing profit margins. Therefore, now is the
time for non-life insurers to reform their pricing strategies. We
strongly believe a data-driven pricing strategy will become the new
Peter Bosschaart MSc
market standard, and have developed an open source-pricing is a senior consultant at KPMG
Financial Risk Management.
platform to assist insurers in the transition. bosschaart.peter@kpmg.nl
 INTRODUCTION accordingly. The increasing availability of data facilitates
this. We consider the following criteria to be fundamen-
The Dutch non-life insurance market is under pressure, tal to fully utilize the advantages:
and has been for a while. High combined ratios are com- •• a scalable IT infrastructure;
mon for several insurance products and a few big players •• internal data of high quality;
dominate the market. However, market competition is •• the possibility to combine internal data with exter-
fierce. The ‘old way’ of buying all your policies at the local nal datasets (open and commercial);
insurance office has disappeared. Consumers are now •• updating pricing strategies with modern analysis and
more than ever able to compare insurance policies that modeling techniques, like machine learning.
cater to their needs, often causing them to search for the
lowest available prices. This forces insurers to set compet- These points open the door to set a competitive premium
itive premiums, whilst managing claims and operational on a more individual basis, and updating it directly when
costs. These competitive premiums, combined with the new claim data is obtained, or when changes occur in an
trend of increasing claim levels and operational costs, individual profile. We call this approach a data-driven
result in decreasing margins with an even higher pres- pricing approach, and strongly believe that this approach
sure to maintain a stable portfolio with profitable clients. will quickly render traditional approaches obsolete. In
order to facilitate the change, KPMG has developed an
However, not only consumers have more possibilities to open source analysis platform that meets the criteria
benefit from the increasing availability of information, set out above, which can be used by insurers to clean,
also the insurance companies can profit from these. For combine and analyze their data. We have described this
example, insurance companies have more detailed pricing analysis platform in the box about ‘KPMG open
insight in their consumer’s behavior, and the underlying source analysis platform’.
risks, and are therefore able to transfer from traditional
pricing strategies to more modern ones.
PRICING STRATEGIES
Traditional pricing strategies originate from insurers
knowing their clients personally, and as such the charac- Traditional strategies
teristics of their claims. Current strategies mostly rely on
established (but old-fashioned) statistical techniques that Traditionally, insurance companies used to have unique
set the premium on a group level, based on data provided portfolios, for example due to their strong regional
by the policyholders, and are often not regularly recali- presence in the market. People bought insurance from
brated. Therefore, premiums are not optimized. insurance companies present in their area, resulting in
portfolios with unique characteristics for the region.
The more insight you have in the characteristics of your Insurance companies used to know the people they
policyholders, the better you can estimate the corre- insured personally, and could set adequate premiums
sponding individual risks, and tailor the premium based on this. With the introduction of buying and
comparing insurance policies on the internet, these geo-
graphic dependencies started to dissipate, and competi-
Figure 1. Non-life insurance pricing strategies. tion between insurance companies increased. Nowadays
it is easy to find ‘the best deal’ for one’s specific insurance
Value- needs and buy a policy in a matter of seconds, without
based any human interaction. For the insurers this results in
Dynamic
pricing more geographically dispersed portfolios, with indiffer-
ent characteristics. Therefore, these traditional strategies
evolved to forms that are supported with data provided
Cost- by the policyholders and established statistical methods
Value-based
based pricing using this data: generalized linear models (GLMs), which
were introduced in 1972.

Risk premium
‘augmented data’
While the market is changing, many insurance companies
still apply traditional or old-fashioned GLM-based method-
Risk premium
ologies to derive their premiums. So how can an insurer set
multivariate
‘own data’ a competitive premium nowadays? For this, we distinguish
One-dimensional between cost-based pricing approaches and value-based
risk differentiation
pricing approaches, which are set out in Figure 1.
Pay-as-you-go

10 Data-driven pricing
 Preparation Data exploration Modelling
phase phase phase

Figure 2. Data-driven pricing approach.

Optimi-
zation
phase

KPMG open source analysis platform

Dashboard and adjustment
We consider a data-driven pricing strategy with phase
both the cost-based approach and the value-
based approach as the new market standard
for pricing non-life insurance policies. Viable The platform can easily be deployed by the
premiums arise from combining: insurance companies own data scientist, and
1. high quality data from internal and external can be tailored to the organization. The premium
sources; can then be optimized following the approach in
2. modern day analysis and modeling Figure 2.
techniques; • In the preparation phase initial data quality
3. management information. checks are performed and datasets are
combined for analysis.
To support insurance companies in the transition • During the data exploration phase the
to a data-driven pricing strategy, KPMG has combined data is analyzed to find well-suited
developed an open source-pricing platform: input parameters for the modeling phase.
the platform does not, and will not, bear any • Both traditional and modern models are
licensing fees. On this platform modern day applied to the combined dataset in the
data handling and visualization capabilities are modeling phase.
combined with up to date modeling techniques • In the optimization phase parameters are
to derive an optimal premium. It features a adjusted to find the statistical optimal
functionality to: premium.
• combine datasets; • The results are visualized in the dashboard
• analyze and clean the datasets (data quality and adjustment phase.
checks, pre-production); • Based on the statistical and visual
• derive premiums with traditional GLM interpretation of results and management
techniques to gain insight on a cost-based information, the models can be fine-tuned
approach; iteratively.
• analyze correlation structure of all variables • New data can be added on a real-time basis,
to enrich the GLMs with cross variables and can be used to update the pricing models
higher order terms; frequently, and provides direct insight in
• optimize premiums in a value-based portfolio developments.
approach, e.g. by means of pricing elasticities
or competitor information; KPMG has experience and credentials in
• derive premiums for a cost-based and value- assisting non-life insurers in every step of the
based approach with modern machine way in implementing the approach above,
learning and deep learning techniques; leading to optimized premiums and portfolios.
• visualize portfolio statistics (claims, risk In this matter KPMG offers support from
profiles, etc.); multidisciplinary teams with expertise from both
• visualize pricing performance (old/new a data and analytics perspective, as well as an
premiums versus claims). actuarial perspective.

Compact 2018 4 Transparency of information 11

 Cost-based data-driven pricing strategies variable for motor insurance policies, but very granular
car type and color variables might (partially) act up as a
A cost-based pricing strategy involves investigating the proxy for gender effects.
risk of a policy, and finding a premium that will cover
this risk. This can be done from pooling risk without Value-based data-driven pricing strategies
diversification (pay-as-you-go), to trying to find unique
characteristics that forecast the underlying risk of The methods discussed so far derive the premium
the policy (risk premium augmented data). The latter with a cost-based approach. They all aim to predict
requires reliable data as input for statistical models. future claims of a policyholder based on certain char-
The more unique characteristics can be used, the more acteristics, so that an accurate premium to cover these
individual the risks can be estimated, and therefore the expected claims can be quoted. Another data-driven
more individual the premium can be set to better cover pricing approach is the concept of value-based pricing.
the risks. This approach derives the optimal premium based on a
customer’s willingness to pay for the product. The will-
Insurance companies already possess a large amount of ingness to pay can roughly be divided in product benefits,
data on their policyholders. The digitalization of insur- service benefits, and brand benefits. Outperforming the
ance even enables them to collect more than before, and competition on these elements can attract new clients,
store the data in a more structured and centralized way. In even if the price is higher. Awareness of what the target
addition, the availability of open and commercial datasets audience is willing to pay per element is key in optimiz-
is ever increasing. For example, the Netherlands Vehicle ing profits. Also, for the value-based pricing approach,
Authority (RDW) publishes an open dataset with a the availability of data and modern analysis and mode-
substantial amount of vehicle information, and the Dutch ling techniques are of utmost importance.
central agency for statistics publishes regional and demo-
graphic information. Using such additional datasets can CONCLUSION
enhance the insurers’ data, by missing data imputations
or replacing low quality data in its own set, and enriching The pressure and tight margins in the Dutch non-life
it by adding new variables to the set. This provides the insurance market forces the market to move to a new
insurer extended insights into characteristics of its pricing approach. Many insurance companies still apply
policyholders and insured objects. However, in our traditional methodologies to derive their premiums.
experience insurance companies often face technical Implementing a data-driven pricing strategy can outper-
difficulties with combining datasets due to different data form traditional strategies, and can either be achieved
formats, legacy systems, and an absence of a uniform data by optimizing the cost-based pricing approach by means
analysis platform. If this problem can be overcome and of using modern analysis and modeling techniques, but
additional data can be used in the pricing procedure, even also by further development of the value-based pricing
old-fashioned statistical techniques (like GLM) can be approach.
used to set a more precise premium than premiums solely
based on internal data. By still finding unique characteris-
tics in the insured portfolio, it is possible to outperform
the market with a traditional cost-based pricing approach.

Access to substantial amounts of relevant data also About the authors

opens the possibility to utilize modern analysis and drs. R. Valkenburg MSc AAG  is a senior manager at KPMG
modeling techniques, like machine learning and deep Financial Risk Management. Rinze is a certified actuary
and highly experienced in the insurance industry, mainly
learning. With these techniques an insurer can extract on strategic, pricing and reserving topics.
extensive intelligence from its data in an almost fully
P. Bosschaart MSc  is a senior consultant at KPMG Financial
automated matter, if desired. Where human interpreta- Risk Management. Peter is a financial mathematician
tion of statistics and modeling choices based thereon are with five years of experience in the actuarial sector, and
key in traditional methods, these modern techniques has a strong focus on quantitative advisory projects.
automatically find less obvious, but significant causal
relationships in the data, which are easily overlooked in a
manual assessment. This ultimately results in even more
precise and individual premiums. One should remain
wary of undesired effects when applying a fully auto-
mated approach, like indirect discrimination. For exam-
ple, regulation prohibits the usage of gender as a pricing

12 Data-driven pricing