Should the Government Provide Insurance for Catastrophes? By J. David Cummins

1
Should the Government Provide Insurance for
Catastrophes?By J. David Cummins

• Comments by Dwight Jaffee
• University of California, Berkeley
• 30th Annual Economic Policy ConferenceFederal
  Reserve Bank of St. LouisOctober 20-21, 2005

2
Agenda

• I applaud the conference planners for foresight
  to put cat insurance on the agenda long before
  Katrina.
• It is always a pleasure and enlightening to read
  a paper by David Cummins this one is no
  exception.
• My comments follow the lines of Davids paper
• Why do private markets for cat risks fail?
• Should the government pick up the slack?
• And if so, how is this best done.

3
Lack of Capital is One Root of Evil

• The paper begins with a cogent discussion of the
  key role that capital plays for insuring cat
  risks.
• David does use a normal distribution, whereas
  fat-tailed distributions might be the relevant
  ones.
• Still, I agree that catastrophe insurance market
  failures are due mainly to capital market
  frictions.
• But cat insurance failures involve other
  factorsCapital market imperfections are
  necessary, but not sufficient, for catastrophe
  insurance market failure.

4
Capital Market Imperfections are Necessary But
Not Sufficient for Cat Market Failure

• Cases that motivate Not Sufficient
• Only 15 years ago, US had active private markets
  for hurricane, earthquake, and terrorism risks.
• UK still has an active private market for flood
  insurance (and the Thames could flood London).
• Warren Buffett pledged billions of capital to
  insure California Earthquake Authority in 1996.
• Lloyds of London, even today, stands ready to
  provide terrorism insurance (at the right price).

5
Agency Problems For Catastrophe Insurers

• Finance theory suggests no special cat risk
  problems
• Catastrophes reflect mainly idiosyncratic risks.
• Risk sharing achievable as insurers distribute
  equity.
• But insurance firm managers strongly disagree
• Edward Liddy, President Allstate, WSJ,
  9/6/05The insurance industry is designed for
  those things that happen with great frequency and
  don't cost that much money when they do. It's the
  infrequent thing that costs a large amount of
  money to the country when it occurs -- I think
  that's the role of the federal government.
• Buffett and Lloyds avoid these agency issues.

6
Zealous Regulators and Daffy Consumers

• Regulators compound problem by restricting the
  use of reinsurance and similar risk-sharing
  instruments
• Insurers may not pass on full costs of
  reinsurance.
• Capital haircut applied to offshore reinsurers.
• Consumers are not always rational in evaluating
  the contracts and prices offered by insurers
• Aversion to high deductibles is one problem.
• Policyholders seem to have overly optimistic view
  of expected losses (asymmetric information)

7
Other Issues Seem Less Fundamental

• Quantifying risk parameters is not a major issue
• Parameter uncertainty was long ago dismissed as
  issue for CAPM, and Froot confirms for cat.
• Telecommunications satellites were insurable
  from very first launch. Same for liability lines.
• Time diversification is not necessarily
  intrinsically harder than cross-section
  diversification.
• Indeed, less asymmetric information in time.
• Current quantitative estimates of insurer
  reserves are not relevant since firms do not
  offer cat lines.

8
When Private Markets Fail, Is Government
Insurance the Answer?

• Citizens reasonably call on government to fix
  failure.
• Immense social benefits of insurance and risk
  sharing.
• Risk sharing is intrinsically/uniquely a social
  activity.
• Proposition Government plans should mimic as
  closely as possible what private market would do.
  
• No subsidies, because otherwise
• People induced to put themselves in harms way.
• Creates pork barrel requests by
  industries/regions
• Crowds out private market initiatives.
• Risk-based premiums gt proper incentive to
  mitigate.

9
How Government Insurance Actually Works

• We observe wide range of interventions, from free
  reinsurance for terrorism (TRIA) to full
  government insurer entities (flood and
  earthquake). Hurricane is combination.
• These plans are uniformly disappointing
• 14 take-up rates for earthquake (CEA).
• 2 of NFIP properties create 33 of claims.
• Hurricane is barely hanging on in many states.
• TRIA has a variety of flaws.

10
National Flood Insurance Program (NFIP)Whats
Wrong with this Picture?
11
National Flood Insurance Program
(NFIP)Breakdown was Predicted GAO-01-992T

• Congress requires actuarial premiums/high
  standards on new homes, but deep subsidies on
  grandfathered homes.
• 19,600 towns now on Flood Insurance Rate Map
  (FIRM).
• Average new home rate is 310 annual subsidy is
  610.Expected losses on grandfathers are 5 times
  new homes.
• 38 of claims are repetitive on same
  propertyOften these homes had cumulative losses
  gt house valueNFIP may just buy out these owners
  to save money!
• UK has a well functioning private market for
  flood risks.
• Association of British Insurers
  (www.abi.org.uk/flooding)
• A private/public partnership in which government
  guarantees levees and requires good upstream
  practices.

12
What to Do About TRIA?(Terrorism Risk Insurance
Act)

• TRIA has December 2005 sunset. Fierce battle to
  extend
• Government agencies favor sunset.
• Insurance, building, mortgage industries want to
  renew
• Current plan has many unique (perplexing)
  features
• Provides free reinsurance (covers only high risk
  tiers)Firms have substantial deductibles and
  co-insurance.
• Government is to recoup first 15 billion of
  payments with ex-post levy on all commercial
  policyholders.
• Then taxpayers pay bill to 100 billion ceiling.
• Encourages more trophy buildings in high risk
  locations.But same new construction/grandfather
  issue as flood.

13
What to Do About the California Earthquake
Authority (CEA)?

• Key problem is the 14 take-up rate, which seems
  to be a disaster in waiting.
• Consumers feel premiums are too high (Common
  premium base homeowner policy).
• Premiums, by law, are actuarially determined but
  in practice they have been tempered.
• Dispute between Seismologists and Geologists.
• Consumer particularly dislike 15 deductible, but
  also reject new 10 deductible.
• A mandatory insurance requirement is a bad idea.

14
How to Keep the GovernmentOut of the Insurance
Business

• Fully agree with David regarding what should be
  the two key forms for government intervention
• Government must remove regulatory impediments to
  Catastrophe Bonds
• Let government auction reinsurance rights.
• Government as Lender of Last Resort, not Insurer
  of Last Resort, should be considered.
• Proper solution if issue is post-event liquidity
  (and not simply unwillingness to bear risk).