hot topics in us and europe - finity consulting2012. speakers include: aaa, allstate, the hartford,...
TRANSCRIPT
Hot Topics In US and Europe
Keith Almeida, Jon Tindall and Tim Andrews
Finity Personal Lines Pricing Seminar
3rd May 2012
© Finity Consulting Pty Limited 2012
Hot Topics - Overview
Slide 2
Hot Topics
Portfolio Optimisation
Price Optimisation
Regulation
Natural Perils Pricing
External Data
Multivariate Analysis
Hot Topics - Overview
Slide 3
Hot Topics
Portfolio Optimisation
Natural Perils Pricing
External Data
Portfolio Optimisation
Slide 4
What is Portfolio Optimisation?
Slide 5
What is Portfolio Optimisation?
Slide 6
Risk Deconcentration
Slide 7
Source: Luyang Fu, State Auto Insurance
Companies
Risk Deconcentration - Example
Slide 8
Current Profit Mkt Profit CV
Margin Margin
Adelaide 10% 11% 5%
Brisbane 10% 11% 30%
Risk Deconcentration - Example
Slide 9
Current Profit Mkt Profit CV
Margin Margin
Adelaide 10% 11% 5%
Brisbane 10% 11% 30%
Profit margin in Brisbane does not reflect it’s CV…
Therefore, reduce exposure in Brisbane, and increase
exposure in Adelaide
What is Portfolio Optimisation?
Slide 10
Using External Data
Slide 11
Sources of External Data
Some typical sources of external data for modelling
insurance portfolios:
Geographic
• Location
• Proximity information
• Elevation / Weather
Socio-economic
Credit information – Company, directors, debts, court
actions
Crime Statistics
Industry sources – Glasses guide/Redbook
Usage data – e.g. Telematics
Slide 12
Attitudes to External Data
Slide 13
External Data
Applications in Data –Driven Analytics
A presentation at CAS
March 2011
Source: Towers Watson
Attitudes to External Data
Survey of 40 personal lines and 70 commercial lines, US-
based respondents (2010):
Importance of sophisticated risk selection
98% thought it ‘essential’ or ‘very important’ in personal lines
pricing.
Drops to around 80% for commercial lines (small/middle) and to
under 50% for commercial (large accounts).
• Can respond through risk selection or pricing.
Purchase/Incorporate further secondary data elements
into risk selection process
Personal lines - 38% had taken action over past 2 years, 63%
intend to over the next 2 years
Commercial lines - 29% had taken action over past 2 years, 53%
intend to over the next 2 years
Slide 14
Sources of External Data
Slide 15
Geographic - ‘Geo-coding’ of risks
- weather information
- height
- distance to train station,
shopping centre etc.
- Distance to water course
Socio-Economic - Census information – every 5
years
- SEIFA scores – suite of 4
measures
- Purchase pre-made predictive
variables
Credit - Attach information via ABN/CAN
- Director’s history
- Court actions
- Delinquent payments
- Only for commercial lines in
Australia
Industry Information - Redbook/Glasses
guide
- ANZSIC information
Crime Statistics - Numbers and types
of crimes
- Fraud prediction
Case Study- Telematics
Slide 16
Case Study: Telematics
Initially tried in early 2000’s but ran into implementation
issues.
Progressive used as basis for PAYD policies – patented the
first technology for insurance telematics.
Vehicle use information – statistics gathered in real time eg
location and speed.
Information used to better understand the driving behaviour
of the insured.
Implementation Issues:
Cost of the device
Privacy issues
Selection issues – Safer drivers will be more inclined to
accept the device
Slide 17
Case Study: Telematics
Phase II Telematics – “Pay-how-you-drive” (PHYD) Insurance
Significant cost reduction. Being incorporated into new car
computer systems and “Sat-Nav” devices
Removes some selection and privacy concerns.
Lots of recent activity in the UK
15 Feb 2012 - TomTom + insurance broker Motaquote
launch Fair Pay Insurance - a product that rewards "good"
drivers with lower premiums.
28 Feb 2012 – AA launches telematics.
Tiger.co.uk – 14% of all policies sold based on telematics
compared with 10% the previous year.
Gocompare.com survey March 2012 – 57% of respondents
said likely to switch to telematics insurance in next 5 years.
Slide 18
Case Study: Telematics
Lots of talk in both US and Europe
US Insurance Telematics conference – Chicago September
2012. Speakers include: AAA, Allstate, the Hartford, Zurich +
a range of solution providers.
Europe Insurance Telematics conference – London May
2012. Speakers include: AVIVA, The Co-operative Insurance,
insurethebox, Viasat, Zurich, RBS Insurance, Amadeus
Capital
In Australia
Similar cost and privacy issues as with other jurisdictions
‘Big-brother’ issues more poignant with Australian consumers
Relative late-comers to pay-as-you-drive insurance
Slide 19
Case Study: Crime Statistics
ABS report – produced annually
Victims of crime
Victims of assaults
LawLink – suburb level crime statistics for NSW
Australian Institute of Criminology – statistics provided by
LGA, nothing currently available by postcode
Statistics available by type of crime i.e Murder, assault, theft
etc.
Strong correlation between crime statistics and socio-
economic information
Potential in modelling and predicting insurance fraud.
Slide 20
Incorporating and Implementing
External Data Solutions
Data - attaching secondary data requires ‘clean’ data:
• Names, ABN/CAN etc
• Addresses
Modelling
Efficient mining of the available variables
Understanding the residuals left over from primary
explanatory variables
Rating
Rating risks ‘live’ (PoS) incorporating secondary data
elements can be complicated
Significant IT commitment
Explaining the ‘black-box’ to consumers
Slide 21
Natural Perils Pricing
Slide 22
Slide 23
Natural Peril Pricing
1. ISO paper on pricing Home
2. Use of Catastrophe Models for pricing
Slide 24
1. ISO paper on pricing Home
• Predictive Modelling for Homeowners, by David
Cummings (2010)
• Use of North American Regional Reanalysis (NARR)
• “NARR assimilated a lot of (weather) data from
different sources”
• 32km grids across US
• 27 years of data
• Includes measures of daily Rain / Wind / Snow
Slide 25
1. ISO paper on pricing Home
Slide 26
1. ISO paper on pricing Home
Slide 27
1. ISO paper on pricing Home
Bureau of
Meteorology Data
“Storm scores” –
based on weather
station wind & rain
data
2. Use of catastrophe models for
pricing
Use of catastrophe models for pricing the primary cover
Involves use of AAL, rather than PML. eg:
PML for Sydney is a $600m loss
earthquake AAL for a property in Chatswood is $60
AAL can be used to:
allocate reinsurance costs across risks, or
develop technical prices from the ground up
Use of cat bonds rather than RI prices to derive pure
cost
Slide 28
2. Use of catastrophe models for
pricing
Need modelling results by size of event
What are the RI brokers allowed to provide?
Dealing with non-modelled perils
Getting buy-in of reinsurers
Use of modelled standard deviation (as well as
AAL)
Lots of references to data quality in papers
Dealing with rating factors besides location (eg.
age of building, type of construction)
Slide 29
2. Use of catastrophe models for
pricing
Slide 30
Townsville (QLD) Embleton (WA) Blackwood (SA) Chatswood(NSW)
Grafton (NSW)
Co
st
Cost for a $300,000 Building
Storm Bushfire Earthquake Cyclone
Townsville (QLD) Embleton (WA) Blackwood (SA) Chatswood(NSW)
Grafton (NSW)C
ost
Cost for a $300,000 Building for Losses > $500m
Storm Bushfire Earthquake Cyclone
Slide 31
Disclaimer
Disclaimer provided by Finity Consulting Pty Limited (“Finity”)
This presentation has been prepared for the Finity Personal Lines
Pricing Seminar held on 3 May 2012.
Finity wishes it to be understood that the information presented at
the seminar is of a general nature and does not constitute actuarial
advice or investment advice. While Finity has taken reasonable
care in compiling the information presented, Finity does not warrant
that the information provided is relevant to a particular reader’s
situation, specific objectives or needs.
Finity does not have any responsibility to any attendee at the
conference or to any other party arising from the content of this
presentation. Before acting on any information provided by Finity in
this presentation, readers should consider their own circumstances
and their need for advice on the subject – Finity would be pleased
to assist.