Bird Flu A threat to Insurance?

1
Bird FluA threat to Insurance?

• Henk van Broekhoven

2
Preface

• On request of EC Groupe Consultatif started a
  task force to analyse the possible impact on
  insurance because of the Bird Flu
• Actuaries involved
• Anni Hellman (EC)
• Henk van Broekhoven
• Erik Alm
• Tapani Tuominen
• experts (other disciplines) from EC

3
Pandemic

• A Pandemic arises when a disease that affects at
  least 25 of the globe causes high morbidity,
  excess mortality and social and economic
  disruption
• Pandemics cause a sudden explosion of illness
  putting heath services under strain
• Pandemics spread very rapidly around the world

4
Pandemic

• Three pandemics in the twentieth century
• 1918 Spanish Flu
• By far the most deathly pandemic in the last 400
  years ( observation period)
• 99 of the deaths were younger than 65 (!)
• Worldwide 40-50 million deaths

5
Pandemic

• Three pandemics in the twentieth century
• 1957 Asian Flu
• Global deaths 2 million (USA 70,000 excess)
• 90 of the deaths were older than 65
• Looked more like a normal seasonal flu, but with
  more sick people (gt25)
• Started in China Febr. 1957, reached Hong Kong in
  April and the rest of the world in 6 months

6
Pandemic

• Three pandemics in the twentieth century
• 1968 Hong Kong Flu
• Less deaths than the Asian Flu 1957 (USA 36,000)
• Looked similar to the 1957 flu

7
Spanish Flu 1918

• Why was this pandemic so deathly?
• 1918 end of first World War
• Tuberculosis epidemic in same period
• People died within 8 hours after detecting
  condition
• In a normal flu and also in 1957 and 1968 extra
  deaths occur because of complications like
  pneumonia

8
A new Pandemic?

• Experts it WILL happen, only question when (it
  is assumed that chance for a new pandemic in the
  next ten years is above 50)
• Will H5N1 cause a new pandemic?
• Chances are low (article nature)
• Still new viruses can cause a pandemic

9
Would it look like the Spanish Flu?

• Spanish Flu was very extreme
• Unlikely that this happens again nowadays
• Huge medical development since 1918
• Better prepared
• People are in better condition
• No TB epidemic and no WW 1 situations
• Probability similar scenario ltlt 1400

10
How will pandemic look like?

• Scientists simply dont know
• History shows that a pandemic comes in waves with
  a couple of months in between
• Second wave worse than first one
• Gives some time to develop a cure

11
Possible impact depends on ..

• Can new virus easily infect humans
• How easy is the transfer human human
• Power of making people sick
• Incubation period
• How fast can a cure be developed after virus is
  discovered

12
Possible deaths scenarios

• WHO between 2 million and 7.4 million globally
• RIVM, extreme 40,000 in the Netherlands on
  16,000,000 people ( translated Spanish Flu)
• RIVM, more real 0 10,000 in NL

13
At what ages?

• Will the extra mortality be age independent, or
  appear more likely at higher ages?

14
Spanish Flu in NL
15
Spread of pandemic deaths over the ages/gender

• Suppose in extreme RIVM scenario deaths are
  spread age/gender independent
• That will lead to the following overview

16
RIVM extreme scenario
Spread deaths independent of age Spread deaths independent of age Spread deaths independent of age Spread deaths independent of age
Age Group Pandemic death Normal death Extra mortality
0-25 12,201 2,362 517
25-45 12,781 4,631 276
45-65 9,609 20,506 47
65-85 4,848 71,736 7
gt85 561 35,841 2
Total 40,000 135,075 30
25-65 22,390 25,137 89
17
RIVM extreme scenario

• Whole population in case of age independency
  shows an extra mortality of 0.25 (to be added up
  to the qxs)
• Supposing insured population in better health, so
  better protected 60 of 0.25 gives an extra
  mortality of 0.15
• Calamity solvency capital can be calculated in
  this way!

18
RIVM other scenarios

• Suppose 10,000 death in NL age independent extra
  mortality for insured population 0.0375
• Suppose 10,000 deaths 90 at higher ages (gt65)
  xgt65 extra 0.25 extra qx
• xlt65 extra 0.005 extra qx

19
Other risks

• A pandemic has also impact on other risk types
• Morbidity
• PC (Animal insurance)
• Financial
• Operational

20
Financial

• Predicting the impact of Avian flu on global
  economy is impossible
• A re-run of the Spanish flu could strip tens off
  GDP
• In extreme cases goods more useful than cash
• Also temporary impact possible in less severe
  pandemics, simply because of fear following the
  hype

21
Operational risk

• More than 25 of employees are at home
• Partly ill
• Partly surging
• Partly fear
• Precautions
• Stocking medicines for employees?
• Possibility working outside office (at home)

22
Morbidity risk

• Products
• Medical insurance
• Hospitalisation
• Sick leave insurance
• Disability (?)

23
Medical insurance

• Non severe scenario
• High number of extra claims
• Claims low (treatment costs are low)
• Severe scenario
• unclear

24
Hospitalisation

• Non severe scenario
• Some extra claims because of complications
• Severe scenario
• Unclear
• Limited number of hospital beds
• Temporary hospitals
• Costs shared by governments and insurance
  companies (?)

25
Sick leave insurance

• Non severe 15-25 extra claims (?)
• Severe gt25
• what to do with people who are healthy but still
  stay at home (fear)?

26
Disability

• Perhaps but unclear some impact in severe
  situation

27
Severe scenario

• For health care we think that the first goal of
  people and governments will be that the virus is
  beaten ASAP
• Independent on costs
• Independent of insurance

28
Conclusion for insurance

• It is impossible to set up a best estimate
  scenario, only what if scenarios
• Impact unclear for some risk types
• A solvency margin for calamity could be 0.15 x
  NAR (better than something like doubling one-year
  claims)
• Be careful with diversification within calamity
  -gt correlation 1

29
Conclusion Prof. Coutinho

• Be careful in communication
• Try to prevent panic
• In can last another 5-10 years before we have a
  pandemic
• Publications on safety and heath are selling
  good
• A pandemic creates sensation in publications

30
(No Transcript)