clouston tues 3pm ven v

8/12/2019 Clouston Tues 3pm Ven V

http://slidepdf.com/reader/full/clouston-tues-3pm-ven-v 1/18

wInternal Model

David Clouston

© Lloyd’s1



o e a a on

What is meant by “validation”?

Validation is a process of coming to understand whether the externalcat-model provides a valid representation of the risk for your portfolio

Not necessarily for “the industry”, or “the market”

Not necessarily for someone with a totally different portfolioot necessar y someone or w om t e mater a ty o cat-r s or t s

region/peril is quite different from yours

© Lloyd’s3



o e a a on

It is not possible to fully validate every aspect of an external cat model

There is no requirement to do so

“… expectations of external models […] should be appropriate to theirnature, sca e an comp ex ty. s recogn ses t e pr nc p e oproportionality.” (Solvency II Level 3 draft)

© Lloyd’s4



Proportionality & Materiality drive:-

a er a y an ropor ona y

WHAT you do

HOW you do it

The EXTENT of the requirement for documentation & evidence

Risk ranking

© Lloyd’s5



You have a process for systematically assessing the materiality of your

a er a y an ropor ona y

overall catastrophe risk

You have processes for assessing the relative materiality of different

You can demonstrate that you have such a process

You have evidence that ou have followed it

That the process is continuous (i.e. not just a one-off exercise)…

… determines the relative requirements for model-validation,documentation etc.

© Lloyd’s6



The external cat-model “inherits” all the obligations of the Internal

o e a a on

Model, including for validation

“The use of a model or data obtained from a third party shall not be

requirements for the internal model set out in Articles 120 to 125.”(Solvency II Directive Article 126)

ow muc mo e va a on s requ re

© Lloyd’s7



SII requires that you demonstrate:

o e a a on

There is enough understanding about a cat model to make decisionsabout its selection and use for your portfolio (taking into account M &

Evidence that this process has been followed

© Lloyd’s8



o e a a on – a s mp e examp e

An aggregate-level cat-model, that has not been significantly revisedfor five years

Organisation 1

material cat risk for the region/peril, but of relatively limitedsignificance within the overall Internal Model (evidenced by […])

ex osure data for the re ion tends to be rovided at CRESTA-level

exposure derives from a few very large clients, with similarportfolios

© Lloyd’s9




An aggregate-level cat-model, that has not been significantly revisedfor five years

Organisation 2

very material cat-risk for the region/peril (evidenced by […]);

exposure-data for the region is provided at Postcode, Street-Level and Lat/Lon

exposure derives from many diverse clients, very widelydistributed throughout the region.

© Lloyd’s10



Org 1 Org 2


little impact on Internal Model significant impact on InternalModel

Sensitivity

Exposureassessment

ualit of dev. data

More recent-

Conclusion

© Lloyd’s11



Org 1 Org 2



Sensitivity

broadly similar to industryassumptions in c-m

different from industryassumptions in c-m

Exposureassessment

ualit of dev. data

More recent-

Conclusion

© Lloyd’s12



Org 1 Org 2



Sensitivity



similar to our ex osure-data much lower-resolution than our

Exposureassessment

ualit of dev.exposure-data

data

More recent-

Conclusion

© Lloyd’s13



Org 1 Org 2



Sensitivity




Exposureassessment


limited benefit – no hazard geo-coding and financial-module

data

More recent-

Conclusion

© Lloyd’s14



Org 1 Org 2



Sensitivity




Exposureassessment


limited benefit – no hazard geo-coding and financial-module

data

More recent-

seems to be a validrepresentation of our risk,

not a valid representation of ourrisk, without significant

–

Conclusion

© Lloyd’s15

validation



You need to show:

ocumen a on

that you have formal processes for understanding enough about acat-model to make decisions about its selection and use as part of

Materiality)

that you have formal processes governing your use of the externalca -mo e

A separate, but related, obligation is to provide evidence that you havefollowed those processes

© Lloyd’s16



Demonstrating understanding

ocumen a on – n ers an ng

Knowledge about the model from the supplier – “outward”Knowledge about its applicability to your risk – “inward”

Sensitivity of Internal Model to the external cat-model

Demonstrate understanding as it applies to:

- Model validation (incl methodology)

- Model selection

- Model change

- Model limitations

© Lloyd’s17



Demonstrating use

ocumen a on – se

Evidence of competency, consistency, compliance with policy andgovernance requirements

Use required for

- How the cat-model is being used (in-house, broker), and checksit is bein used a ro riatel

- Outsourcing

- Data, including any adjustments

- Integration with Internal Model

© Lloyd’s18

clouston tues 3pm ven v

Documents