An actuarial toolkit for microinsurance Daniel Clarke University of - - PowerPoint PPT Presentation

An actuarial toolkit for microinsurance

Daniel Clarke University of Oxford and UK actuarial microinsurance working party 22 May 2012

• 1. Poverty and Risk
• 2. The UK Actuarial Profession Microinsurance Working Party
• 3. Two topics covered in the agricultural toolkit
• 4. Beyond the microinsurance toolkit

Overview

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 2

• 1. Poverty and Risk

‘One of the least remarked-on problems of living on two dollars a day is that you don’t literally get that amount each day. The two dollars a day is just an average over time. You make more on some days, less on others, and often get no income at all.’ ‘How do you make sure there is something to eat and drink every day, and not just on the days you earn? If that seems hard enough, how do you deal with emergencies?’ Portfolios of the Poor, Collins et al. (2009)

Being poor is risky

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 4

• Exposure to risk leads to cautious behaviour…

– Avoid costly inputs, choose safe activities, etc.

• … and cautious financial arrangements

– Keep unproductive liquid assets and small stocks – Informal / semiformal mutual support

• However, still exposed to shocks…

Protection against shocks is typically incomplete

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 5

Source: Woldehanna (2010), using Young lives data

Lots of uninsured risk

Shock Rural (%) Urban (%) Illness in family 37 35 Death in family 8 14 Theft/crime 17 9 Increased input prices 37 19 Decreased output prices 9 1 Livestock death 38 7 Crop pests 39 5 Natural disaster (including drought) 63 7 Job loss 9 18 Households reporting a particular event or shock affecting their wealth or standard of living considerably in last four years, Ethiopia 2002-2006

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 6

• Pay a premium in good years to smooth out the worst years

1 2 3 4 Net annual income Year No insurance

So insurance is the answer to risk?

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 7

1 2 3 4 Net annual income Year No insurance With insurance

• Pay a premium in good years to smooth out the worst years
• Offering insurance may help people to design incrementally

better risk management strategies

So insurance is the answer to risk?

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 8

• Better insurance could have large benefits but…
• … so could

– Better risk mitigation – Less painful risk retention through smoothing shocks over time

• Better savings
• Better credit markets

– Better social protection/social safety nets

So insurance is the answer to risk?

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 9

Dealing with risk Insurance Insurance Actuarial functions

“…not a specific product or product line. It is also not limited to a specific provider type. Microinsurance is the provision of cover to a specific market segment, i.e., low-income persons.” IAIS Issues Paper (2007)

Microinsurance is…

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 10

• 2. The UK Actuarial Profession Microinsurance

Working Party

• October 2009-September 2010 (5 members)

– Microinsurance Working Party formed following GIRO 2009 to review literature on crop microinsurance from an actuarial perspective – Won the Brian Hey Prize at GIRO 2010

• October 2010-July 2011 (>10 active members)

– Members from UK, USA, West Africa, Mexico, Caribbean, Jordan, Pakistan, India etc – Working out what to do next, in collaboration with partners (IAA and microinsurance organisations)

• July 2011-present (>15 active members + partners)

– Development of an educational tool for microinsurance (the ‘microinsurance toolkit’) – Funding from the UK Actuarial Profession for dissemination

• November 2012

– Launch of microinsurance toolkit at the International Microinsurance Conference

Work by UK actuaries on microinsurance: The Microinsurance Working Party

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 12

What is ?

• Educational resource for technical microinsurance practitioners
• To be publicly available

– not linked to a specific educational program or qualification

• Being developed on a volunteer basis by qualified actuaries
• Covering life, health and agricultural insurance
• Consists of set of spreadsheets and supporting documentation
• For more information see

http://www.stats.ox.ac.uk/actuarialtoolkit

• Or email actuarialtoolkit@stats.ox.ac.uk

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 13

• 1. An overall illustrative calculation to show how different

pieces of the puzzle connect

• 2. Index insurance product design
• 3. Data cleaning
• 4. Detrending
• 5. Overfitting
• 6. Credibility theory and portfolio approaches to pricing
• 7. Capital and reinsurance
• 8. Reserving and margins
• 9. Expenses
• 10. Glossary

For example, the agricultural insurance toolkit covers

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 14

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 15

• 3. Two topics covered in the agricultural toolkit
• Portfolio-based approaches to pricing
• Trends

• 1. Porfolio-based approaches to pricing
• Historical yields vary

significantly from subdistrict to subdistrict

• Statistical question: how

much of this variation is statistically significant

• Actuarial questions: how

much of this variation should be reflected in prices? How much should prices be smoothed? Historical claim payment rates at 90 % coverage level, Rice crop, Andhra Pradesh, India

Smoothing premium rates using Credibility Theory

• Blue rates are for no smoothing

Premium rate 12% 4% 8% 𝑆𝑏𝑢𝑓 1 = 4% × 𝑎 +8% × (1 − 𝑎) 𝑆𝑏𝑢𝑓 2 = 12% × 𝑎 +8% × (1 − 𝑎)

• Green rate is for full smoothing
• Red rates are consistent with Credibility Theory

– Z is between 0 (‘no credibility’) and 1 (‘full credibility’)

• Credibility Factor 𝑎 is an intuitive intermediate calculation

that can be calculated using statistical techniques, and scrutinised by senior management

• These two yield histories have the same mean and standard

deviation but should they be treated the same?

• 2. Trends

50 100 150 1 2 3 4 5 6 7 8 9 10 Yield (kg/ha) Year number 50 100 150 1 2 3 4 5 6 7 8 9 10 Yield (kg/ha) Year number

Allowance for trends can make a big difference to rates

For example

• Use of improved seeds (Bt cotton) led to dramatic increase

in average cotton yields across India

• Ratemaking without allowance for this technological trend

led to high premium rates and low demand

– Trend in yields mistaken for uncertainty

• Application of detrending methodology provided sound

justification for rate reductions of:

Gujarat Maharashtra Madhya Pradesh Percentage reduction 47% 78% 54%

• 4. Beyond the microinsurance toolkit

• 1. Stepping stone for traditional actuaries interested in working

in microinsurance.

– E.g. Two members of the working party have moved from UK traditional actuarial work to full time microinsurance work. – UK Actuarial Profession funding for volunteers to attend the International Microinsurance Conference this November.

• 2. Building awareness of actuarial functions within the

microinsurance and donor community.

• 3. Networking between actuaries with interest in

microinsurance.

• 4. Tool to support microinsurance educators
• 5. Stepping stone for the profession?

– If so, what is next?

Target outcomes of the toolkit project

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 22

Need a wage premium for actuarial services ⇒ need:

• 1. Appropriate accreditation

– Affordable – Inappropriate for traditional insurance

• Need large gap between qualification and Associate/Fellow of UK/US

profession

– Continuing (and local)

• 2. Reserved functions

– Market forces unlikely to be sufficient

How to develop the actuarial function in developing countries?

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 23

Challenge: Design a sub-$1,000 ‘bare bones’ actuarial qualification for microinsurance practitioners (You’re allowed initial fixed costs to be financed by aid

• rganisations)

Daniel Clarke - http://www.stats.ox.ac.uk/~clarke/ 24