The Brookings Institution, Washington, D.C.www.brookings.edu

1
Comments on You Can Only Die Once

• Peter R. Orszag
• The Brookings Institution
• April 12, 2002

2
Fundamental issue

• Are decentralized market forces sufficient to
  provide proper incentives for homeland security?
• EPA official chemical industry has a very
  powerful incentive to do the right thing. It
  ought to be their worst nightmare that their
  facility would be a target of a terrorist act
  because they did not meet their responsibility to
  their community. Statement of Bob Bostock,
  assistant EPA administrator for homeland
  security, quoted in James Grimaldi and Guy
  Gugliotta, Chemical Plants are Feared as Targets,
  Washington Post, December 16, 2001, page A1.

3
Insight from paper

• Decentralized approach does not necessarily
  provide proper incentives because of
  contamination
• In effect, my security depends on your security
  effort
• In good equilibrium, both invest in security and
  contamination muted/eliminated
• In bad equilibrium Since everyone else is not
  investing in security, your incentives to invest
  are muted. Nothing to prevent bad equilibrium
  from arising and being perpetuated

4
Other motivations for government intervention

• Externality 1 Contamination effects (Kunreuther
  and Heal)
• Externality 2 National sovereignty (just as an
  invasion of the nations territory by enemy armed
  forces)
• Information costs
• Bankruptcy laws
• Moral hazard/bail out
• Incomplete markets

5
Two other effects Endogenous targets/displacemen
t

• Contamination effect suggests that my investment
  in security provides a positive externality for
  you, by reducing the risk of contamination
• But opposite may be true for observable security
  measures, given that the selection of targets is
  endogenous My investment may make it more likely
  that terrorists will target you, rather than me
• Related work on observable and unobservable crime
  prevention steps
• In other words, p (and perhaps q) may depend on
  relative security spending in addition to
  absolute spending
• Requires continuous c, or discrete levels
• Possible to have over-investment rather than
  under-investment

6
Two other effects Incentives from public
programs

• Related work examines the impact of public
  provision of security on private incentives
• Orszag and Stiglitz, Optimal Fire Departments,
  Brookings, January 2002
• Presence of a fire department (or public security
  program) can exacerbate the social costs
  associated with the underlying negative
  externality

7
Types of intervention

• Insurance
• Mandate vs. voluntary
• Pricing
• Reinsurance
• Liability
• Bankruptcy
• Definition of adequate effort
• Tax fines or subsidies
• Fines Political viability
• Subsidies Gold plating and budget outlook
• Regulation
• Coordinating mechanisms
• Anti-trust concerns
• Clusters of high-security groups

8
Slippery slope?

• Especially given the potential for government
  failure and displacement effect, how to draw the
  line?
• Forthcoming Brookings volume (Protecting the
  American Homeland) sets one possible standard
  Loss of thousands of lives, or interruption of
  millions of lives for significant period
• So worry about stadiums, but not small shopping
  malls

9
Possible next steps for research

• Imperfect information and monitoring costs
  crucial for implementing many of the suggestions
• Endogenous target selection/displacement
• Empirical evidence on size of contamination
  effect relative to direct threat/internal
  benefits (fn 7)