podlaha wed 130pm ven ii
TRANSCRIPT
-
8/12/2019 Podlaha Wed 130pm Ven II
1/33
European Floods
Cat Modeling 2014, February 12, OrlandoAdam Podlaha
-
8/12/2019 Podlaha Wed 130pm Ven II
2/33
1Proprietary & Confidential
Agenda
Section 1 Introduction
Section 2 Data availabilitySection 3 Flood modelling
Section 4 Selected chapters of flood modelling in Europe
Section 5 Close
-
8/12/2019 Podlaha Wed 130pm Ven II
3/33
Flood losses in Europe
Flood insurance coverage in Europe
Section 1: Introduction
-
8/12/2019 Podlaha Wed 130pm Ven II
4/33
3Proprietary & Confidential
Flood losses are second highest after windstorm
Source: http://catastropheinsight.aonbenfield.com
http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/ -
8/12/2019 Podlaha Wed 130pm Ven II
5/33
4Proprietary & Confidential
Flood losses are regularly present in Europe
Source:http://catastropheinsight.
aonbenfield.com
And there is no reason
why this shouldnt
continue
Often cross country
events
Catastrophe modelling is
needed
http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/ -
8/12/2019 Podlaha Wed 130pm Ven II
6/33
5Proprietary & Confidential
Different flavours of flood coverage in Europe
Different flavours and obscurities present due to Europes fragmentation
Country DescriptionCzech Rep.
(+ SK, PL)
All covered, limits for Ind & Com, no limits for Res, commercial rates for
primary, frequent recent experience with flooding
Austria All covered, very small (few 1,000s )for Res, almost no top-up
available, doesnt affect the primary rate as much, rating system
available
Netherlands No Res cover available, very limited Com & Ind cover, cars covered,
talks about pool in 2013, not happening as of now
Switzerland 2 insurance pools, mandatory cover in some, in some no commercial
rate for primary,no reasons for rating systems
Germany Low insurance penetration in the former West. Germany, higher in
former East. Germany, primary rating system available
UK High penetration in all occupancies, commercial rates, good rating
systems, talks about a pool to happen soon
-
8/12/2019 Podlaha Wed 130pm Ven II
7/33
Portfolio data
Model development data
Section 2: Data availability
-
8/12/2019 Podlaha Wed 130pm Ven II
8/33
7Proprietary & Confidential
Better portfolio data in places with recent flooding
..., flood coverage available & functional rating systems (go hand in hand)
Geo-coding
At least postal code available, longitude & latitude data varies from all to none
Primary and secondary modifiers
Limited number of modifiers can be really used for modelling either due tounavailability or the model not being able to use those
Common issues
Multi-locations for industrial occupancy
Wrongly assigned longitude and latitude Inconsistency in what belongs to which occupancy
Portfolio data are improving with events happening, also due to rating needs
-
8/12/2019 Podlaha Wed 130pm Ven II
9/33
8Proprietary & Confidential
Detailed model develop. data, not always cheaper
Elevation data and other topographic data
Country specific data: from topographical maps or remote sensing (on increase) Continental data: remote sensing based, need to be careful about the specs
Many topographic data still from country specific providers
Hydrological data
The availability is generally very good
The cost varies widely from free to very costly Geo-coding data
The usual: Google, ESRI, Bing, PBS for address geo-coding
Same + others for postal code data, local providers as well
Data for loss estimation from just happening events Remote sensing data (disaster charters, PERILS etc.)
Conclusions: more and better data are available, cost varies widely
-
8/12/2019 Podlaha Wed 130pm Ven II
10/33
Flooding types
Model types
Availability Two basic approaches
Section 3: Flood modelling
-
8/12/2019 Podlaha Wed 130pm Ven II
11/33
10Proprietary & Confidential
Flooding comes in many different flavours
Fluvial (riverine) flooding flood plain and off-
flood plain
Storm surge
Tsunami
Pluvial flooding / cloudbursts / rainfall
Flood plain extent (shades of blue) with
off-flood plain buffers (red)
Conclusion: different flood types need different modelling approaches
-
8/12/2019 Podlaha Wed 130pm Ven II
12/33
11Proprietary & Confidential
Scenario models the 1ststep in risk quantification
1. Scenario models
Historic events Events linked to particular return
period
Hypothetical events
Dam breaks,
2. Fully probabilistic models
Stochastic event set needs to be
generated (assumptions)
Other components same
0.00%
0.01%
0.02%
0.03%
0.04%
0.05%
0.06%
0.07%
0.08%
1 10 100
Loss[%o
fTSI]
Return period [years]
6yr 7yr 56yr
0.00%
0.01%0.02%
0.03%
0.04%
0.05%
0.06%
0.07%
0.08%
0.09%
0.10%
A B C D E
Los
ses[%o
fTSI]
CompanyModelled Observed
Conclusion: Ideally modelling
solution includes both probabilistic
and scenario models
-
8/12/2019 Podlaha Wed 130pm Ven II
13/33
12Proprietary & Confidential
Brokers, RIs, Insurers & others ahead of vendors
Classic vendors offer limited model coverage (UK, DE, BE, AT)
Brokers, Reinsurers, Insurers and other commercial companies farahead
Why?
Harder to do than other perils (higher and often local based data requirements,
tricky vulnerability comp, time consuming 2D modelling) It all adds to more expensive in both cost and time
New models face complex model acceptance (due to existing models, S2)
Need to involve local institutions due to handling complex issues
The European flood modelling scene is not led by classic model vendors
-
8/12/2019 Podlaha Wed 130pm Ven II
14/33
13Proprietary & Confidential
Climate model vs. river gauge based model
1. Global climate model based
2. Measured discharge based
Global Climate Model (rainfall)
Rainfall Run-off model (modelleddischarge)
Hydrodynamic modelling (flood extent withinundation depth / velocity / duration)
Measured discharge
Hydrodynamic modelling (flood extent
with inundation depth / velocity / duration)
Conclusion: In fact both approaches are part of one process, the difference is
the starting point and what is modelled / measured
-
8/12/2019 Podlaha Wed 130pm Ven II
15/33
14Proprietary & Confidential
Have a guess which approach is better suited to:
... capture global patterns in flooding?
... reconstruct historical events well?
... take into account existing hydrological regulations?
...take into account real climate trends?
... take advantage of the best measured data?
... represent best the nature of flooding in a given country?
Conclusion: each approach has its pros and cons, combination is possible
GCM
Flow
Flow
GCM / Flow
Flow
GCM / Flow
-
8/12/2019 Podlaha Wed 130pm Ven II
16/33
Hazard uncertainty in the Dutch flood model
Cloudburst modelling
How to test and validate a flood model
Vulnerability curves development
Location uncertainty quantification
Section 4: Selected chapters of flood
modelling in Europe
-
8/12/2019 Podlaha Wed 130pm Ven II
17/33
16Proprietary & Confidential
Hazard uncertainty in the Dutch flood model
How many are there? ...
Many
To mention just few:
1. Safety standard of dyke
rings (affecting the
frequency of dyke ring
being flooded)
2. Correlation of dyke rings(affecting the scale of
events)
3. Maximum flood extent
4. Evacuation (Hazard?)
All based on experts
opinion
Conclusion: various uncertainties can be quantified/ implemented by creating various event sets
where the particular uncertainty varies
Event sets
Probability of flooding
Optimistic Realistic Pessimistic
Correlatio
nmatrix Low (isolated
events)Event set 1 Event set 4 Event set 7
Best estimate Event set 2 Event set 5 Event set 8
High (large
scale events)Event set 3 Event set 6 Event set 9
-
8/12/2019 Podlaha Wed 130pm Ven II
18/33
17Proprietary & Confidential
Cloudburst modelling
Extreme rainfall events, very localized
Pluvial flooding often (not-correctly) mixed with flash-floods Processing extreme rainfall intensities in a very short time steps (~5min)
Two types of assessment:
Rainfall intensity based
Flood depth based
-
8/12/2019 Podlaha Wed 130pm Ven II
19/33
18Proprietary & Confidential
Testing and validation of a flood model
Level 3 Guidelines on External Model and Data
The undertaking should perform their own validation
of the material assumptions of the external model that
are relevant to their risk profile and of the process for
incorporating the external model and data into their
own processes and internal model.
In addition to their own validation, undertakings
may also leverage on th e validat ion do ne by the
vendors o rother third party...
1. Sensitivity testing (input change)
2. Back testing(historical events)
3. Stability testing (effect of randomness)
4. Stress testing (model change)
-
8/12/2019 Podlaha Wed 130pm Ven II
20/33
19Proprietary & Confidential
Sensitivity testingaggregation (Thai flood)
Analysis resolution
Postal code
CRESTA (canton)
Analysed aggregated to the
analysis level, means:
More rows in the postal code run
Less rows in the CRESTA run
Two effects
1. Change of spread
2. Change of mean / median
Which causes what?
-
8/12/2019 Podlaha Wed 130pm Ven II
21/33
20Proprietary & Confidential
29.8%
26.8%
29.8%
8.3%
1.0%
3.6%
0.1%0.7%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
A B C D E F G H
Shareofmarketloss
Back testingas if 2013 event (Austria flood)
As if 2013 data analysis
Modelled vs. real loss for the 2013 flood for the market and 8 portfolios
Background: To test the model results against values from observed historical events
from hazard and loss point of view
34%
16%
8%
-42%
7%
34%
80%
-15%
14%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
A B C D E F G H marketDifference(%o
frealloss)
-
8/12/2019 Podlaha Wed 130pm Ven II
22/33
21Proprietary & Confidential
Stability testingNumber of samples (Swiss flood)
No. of samples can be changed in ELEMENTS interface
Depends on the portfolio size
Background: To test the stability of the outputs with different number of samples
-5.0%
-4.0%
-3.0%
-2.0%
-1.0%
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
100005000200010005002502001501005025201052
Differencefrom2
00D
S[%]
Return period [years]
10 DS
25 DS
50 DS
100 DS
200 DS
-
8/12/2019 Podlaha Wed 130pm Ven II
23/33
-
8/12/2019 Podlaha Wed 130pm Ven II
24/33
23Proprietary & Confidential
0
50
100
150
200
250
300
350
400
450
0%
5%
10%
15%
20%
25%
30%
35%
0 200 400 600
Lo
ssdatacount
Lossratio
Flood depth (cm)
Loss data Engineering curve Loss data counts
Vulnerability curves development
Two main approaches
1. Claims data
2. Engineering
Assumptions
Claims data present realistic view on risk
Sometimes no sufficient amount of dataLack of construction types differentiation
Possible solution
Use claim information as a basis
Use shape of the engineering curve
Use relations between engineering
curves for different construction types
-
8/12/2019 Podlaha Wed 130pm Ven II
25/33
24Proprietary & Confidential
Location uncertainty quantification
2 insured locations
Sum Insured 34 mil CZK
Limit 8.5 mil
0
5,000,000
10,000,000
15,000,000
20,000,000
25,000,000
30,000,000
35,000,000
FGU SIR GR OTHP RREC NTY NFAC
Perspective Box Plot Statistics for Event 123002013
with 5000 samplesThere are 328 zero samples
0%
2%
4%
6%
8%
10%
Histogram for Event 123002013 forperspective FGU
with 5000 samples
0%
10%
20%
30%
40%
50%
60%
70%
80%
Histogram for Event 123002013 forperspective GR
with 5000 samples
Section 5: Close
-
8/12/2019 Podlaha Wed 130pm Ven II
26/33
Summary
About Impact Forecasting
Questions Contact
Section 5: Close
-
8/12/2019 Podlaha Wed 130pm Ven II
27/33
26Proprietary & Confidential
Summary
Flooding is the second major peril in Europe but first in some countries
Classic 3 vendors offer limited flood model coverage however many models
available from alternative vendors
Different types of flooding, different data available require innovative
modelling approaches
The ability to quantify (and implement) various uncertainties is the key to
understanding the risk
Better synchronisation needed between primary underwriting systems andcatastrophe modelling for reinsurance purposes
-
8/12/2019 Podlaha Wed 130pm Ven II
28/33
27Proprietary & Confidential
About Impact Forecasting
Impact Forecasting Team
Catastrophe model developers
Independent, transparent, open,
modularand bespokemodels
Natural and man-made perils
Filling gaps and main perils
Global team (60+)
ELEMENTS platform
Runs our models
Runs 3rdparty models
Visualisation: uncertainty & mapping
20+ programmers in last 4 years
Distributed to re/insurers and AB
Integrated with ReMetrica, IWB Lite
-
8/12/2019 Podlaha Wed 130pm Ven II
29/33
28Proprietary & Confidential
-
8/12/2019 Podlaha Wed 130pm Ven II
30/33
Q
-
8/12/2019 Podlaha Wed 130pm Ven II
31/33
30Proprietary & Confidential
Questions
C t t
-
8/12/2019 Podlaha Wed 130pm Ven II
32/33
31Proprietary & Confidential
Contact
Adam Podlaha
55 Bishopsgate, EC2N 3BD
London, United Kingdom
+44 (0)20 7522 [email protected]
Impact Forecasting
www.impactforecasting.com
Catatrophe InsightJoin our LinkedIn group
Di l i
mailto:[email protected]://www.impactforecasting.com/http://catastropheinsight.aonbenfield.com/http://www.linkedin.com/groups/Impact-Forecasting-6586802http://www.linkedin.com/groups/Impact-Forecasting-6586802http://catastropheinsight.aonbenfield.com/http://www.impactforecasting.com/http://www.impactforecasting.com/http://www.impactforecasting.com/mailto:[email protected] -
8/12/2019 Podlaha Wed 130pm Ven II
33/33
32Proprietary & Confidential
Disclaimer
Legal Disclaimer
Aon Limited trading as Aon Benfield (for itself and on behalf of each subsidiary company of Aon Corporation) (Aon Benfield) reserves all rights to the content of this report (Report). This Report
is for distribution to Aon Benfield and the organisation to which it was originally delivered only. Copies may be made by that organisation for its own internal purposes but this Report may not be
distributed in whole or in part to any third party without both (i) the prior written consent of Aon Benfield. and (ii) the third party having first signed a recipient of report letter in a form acceptable to
Aon Benfield. Aon Benfield cannot accept any liability to any third party to whom this Report is disclosed, whether disclosed in compliance with the preceding sentence of otherwise.
To the extent this Report expresses any recommendation or assessment on any aspect of risk, the recipient acknowledges that a ny such recommendation or assessment is an expression of AonBenfields opinion only, and is not a statement of fact. Any decision to rely on any such recommendation or assessment of ris k is entirely the responsibility of the recipient. Aon Benfield will not in
any event be responsible for any losses that may be incurred by any party as a result of any reliance placed on any such opinion. The recipient acknowledges that this Report does not replace the
need for the recipient to undertake its own assessment.
The recipient acknowledges that in preparing this Report Aon Benfield may have based analysis on data provided by the recipient and/or from third party sources. This data may have been
subjected to mathematical and/or empirical analysis and modelling. Aon Benfield has not verified, and accepts no responsibility for, the accuracy or completeness of any such data. In addition, the
recipient acknowledges that any form of mathematical and/or empirical analysis and modelling (including that used in the preparation of this Report) may produce results which differ from actual
events or losses.
The Aon Benfield analysis has been undertaken from the perspective of a reinsurance broker. Consequently this Report does not constitute an opinion of reserving levels or accounting treatment.
This Report does not constitute any form of legal, accounting, taxation, regulatory or actuarial advice.
Limitations of Catastrophe Models
This report includes information that is output from catastrophe models of Impact Forecasting, LLC (IF). The information from the models is provided by Aon Benfield Services, Inc. (Aon Benfield)
under the terms of its license agreements with IF. The results in this report from IF are the products of the exposures model led, the financial assumptions made concerning deductibles and limits,and the risk models that project the pounds of damage that may be caused by defined catastrophe perils. Aon Benfield recommends that the results from these models in this report not be relied
upon in isolation when making decisions that may affect the underwriting appetite, rate adequacy or solvency of the company. The IF models are based on scientific data, mathematical and
empirical models, and the experience of engineering, geological and meteorological experts. Calibration of the models using actual loss experience is based on very sparse data, and material
inaccuracies in these models are possible. The loss probabilities generated by the models are not predictive of future hurricanes, other windstorms, or earthquakes or other natural catastrophes,
but provide estimates of the magnitude of losses that may occur in the event of such natural catastrophes. Aon Benfield makes no warranty about the accuracy of the IF models and has made no
attempt to independently verify them. Aon Benfield will not be liable for any special, indirect or consequential damages, including, without limitation, losses or damages arising from or related to any
use of or decisions based upon data developed using the models of IF.
Additional Limitations of Impact Forecasting, LLC
The results listed in this report are based on engineering / scientific analysis and data, information provided by the client, and mathematical and empirical models. The accuracy of the results
depends on the uncertainty associated with each of these areas. In particular, as with any model, actual losses may differ from the results of simulations. It is only poss ible to provide plausible
results based on complete and accurate information provided by the client and other reputable data sources. Furthermore, this information may only be used for the business application specified
by Impact Forecasting, LLC and for no other purpose. It may not be used to support development of or calibration of a product or service offering that competes with Impact Forecasting, LLC. The
information in this report may not be used as a part of or as a source for any insurance rate filing documentation.
THIS INFORMATION IS PROVIDED AS IS AND IMPACT FORECASTING, LLC HAS NOT MADE AND DOES NOT MAKE ANY WARRANTY OF ANY KIND WHATSOEVER, EXPRESS OR
IMPLIED, WITH RESPECT TO THIS REPORT; AND ALL WARRANTIES INCLUDING WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE HEREBY
DISCLAIMED BY IMPACT FORECASTING, LLC. IMPACT FORECASTING, LLC WILL NOT BE LIABLE TO ANYONE WITH RESPECT TO ANY DAMAGES, LOSS OR CLAIM WHATSOEVER,
NO MATTER HOW OCCASIONED, IN CONNECTION WITH THE PREPARATION OR USE OF THIS REPORT.