Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Make Smart Use of Available Data1st Catastrophe Knowledge Exchange
5. July 2016 - B. Aeberhardt & M. Spörri
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
Beat Aeberhardt
Head NatCat Tools Group UW
Martin Spörri
Head P&C IT Info Foundation & Smart Analytics
Who are we?
2
Introduction
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
New Insights
Large Datasets
Visualization
Combine internal with
external data
Where’s the smartness in using available data?
4
processing large and full datasets at reasonable speed and cost
combining internal with external data on a ‘good enough’ basis rather than perfect matches, validated through smart plausibility checks
… deriving new insights by:
investing in value driven and tailored visualization
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri 5
Importance of DataSmart usage of available data boosts the performance of an organisation
Source: "Infographic based on data from Columbia Business School, MIT/Wharton, IBM, Forrester Research, CMO Council/SAS"
The Problem
The Impact
Internal data is often unused!
Tapping into the data boosts financial profit, return on investment and productivity!
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
• Loads of detailed information on property risks available
– used mainly in individual costing process
– received data granularity increasing
– stored in internal NatCat platform used for costing, RM and capital allocation for all BU’s
– NatCat exposure is fully linked with all (re-)insurance structures and contracts
• Granularity of NatCat models is increasing
– Geographic resolution
– Improved handling of financial structures
– New physical properties becoming more important
What does this mean for NatCat?
6
Internal data is often underused!
Demand for ‘smarter’ use of exposure data is increasing
Cluster DB
NatCat Exposure Data Intelligence
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
Goal & Motivation - Current situation
8
Billions of Risks
Stored internally
Clients participating
Incoming channels
•Broker
•Direct
•Global vs R&N
Years
Same risk is stored multiple times in our system
• several clients participate in the same risk
• we participate on same risk via different channels
• through renewals, we receive ‘same’ risks year over year
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
What is a cluster?
9
Key properties of a cluster are:• Distribution of an attribute value (primary modifiers;
Structure types, Occupancy, Design Year, …)
• Hierarchies of clusters
• Life cycle of a cluster
Pre-Cook
Lookup & Enrich
Periodical Updates
Detailed Cluster Aggregated Cluster
Grid Match Fix Confidence
Index Threshold
Include new Exposure data Re-calc clusters
Histogram
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
Costing
•Exposure validation & reporting
•Shadow rating in NatCat models
Accumulation
•Supporting risk clearing
•Accumulation of Non-NatCat LoB
NatCat models
•Granular building stock improves model assumptions for Geo and Vulnerability disaggregation
Analytics
•Benchmark reports
•Protection gaps, neighbourhood analysis
•Temporal exposure development
Where can the Cluster DB be used?
10
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
Validation of derived
Cluster DB
•e.g. sample check of important clusters
•e.g. validation on aggregated basis and benchmark with external datasets
False positives
•e.g. wrong allocation of risk objects to clusters (e.g. inconsistent street address information
Relation of clusters
•e.g. apartment to building, building to complex/campus
Context specific
bias
•e.g. data tuned/optimized for given peril and/or specific model vendor
•e.g. combination with other data sources
Challenges faced
11
Size ofDataset& Bias
All faced challenges are related to the sheer size of the exposure dataset we are working with and the fact that the submitted information is erroneous.
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
•Geographic extension; CA/FL USA Europe Global
•Attribute extensions: more physical properties, insured values and coveragesScope
•Extend to other Lines of Business (e.g. Terror, Fire)
Models
•Owner of physical objects (e.g. company names)
Ownership
•Market building stock and simple “market portfolios”
•CBI networkDerived products
•New data sources from IoT (e.g. connected homes)
•Google 3D
•Sanborn DBEnriching
Roadmap
12
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
• Do you leverage internal data beside using in the context of costing?
• If yes, what are the areas applied?
• Do you track individual physical objects over time and across policies?
DiscussionAre you doing something similar?
13
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri 15
1st Catastrophe Knowledge Exchange | B. Aeberhardt | M. Spörri
Legal notice
16
©2015 Swiss Re. All rights reserved. You are not permitted to create any modifications or derivative works of this presentation or to use it for commercial or other public purposes without the prior written permission of Swiss Re.
The information and opinions contained in the presentation are provided as at the date of the presentation and are subject to change without notice. Although the information used was taken from reliable sources, Swiss Re does not accept any responsibility for the accuracy or comprehensiveness of the details given. All liability for the accuracy and completeness thereof or for any damage or loss resulting from the use of the information contained in this presentation is expressly excluded. Under no circumstances shall Swiss Re or its Group companies be liable for any financial or consequential loss relating to this presentation.