Sept 11, 2006

1
Risk, Uncertainty

• BUSH 689.601
• September 4, 2006

2
Variability and Knowing Cause and Effect
3
Classical Risk Analysis

• Risk can be clearly defined
• Garrick and Kaplans triplets
• Impacts can be clearly specified
• Agreement on what impacts are
• Agreement on value of impacts
• Probabilities can be bounded
• The Engineering Approach to risk analysis

4
Sources of Uncertainty

• Statistical variation
• Systematic error and subjective judgment
• Linguistic imprecision
• Variability
• Inherent randomness
• Disagreement
• Approximation

5
Evolution of Uncertainty Over Time
6
Decision Analysis
7
Decisions about Interrogation of a Terrorist
Suspect
8
Valuing Prospective Losses
9
Contingent Valuation (CV)

• Measuring value of non-market goods
• Grounded in market prices and utility
• Needed for benefit/cost analyses (required by
  OMB)
• CV strengths
• Direct measures tradeoffs, market-like,
  conservative
• Heavily researched (NSF, EPA) and broadly used
• CV vulnerabilities
• Hypothetical wording sensitivity grew from
  controversy
• Opposed on philosophical grounds
• Payoff valuable element of larger benefit/cost
  analysis

10
Anatomy of a CV Measure

• Definition of the Good to be valued
• Dichotomous take it or leave it pricing
• Random assignment of prices importance of the
  tails
• The payment vehicle
• Plausibility and avoidance of baggage
• Preference for advisory referenda mechanisms
• Budget reminder
• Follow-up questions of certainty of willingness
  to pay (WTP)
• Modeling correlates
• Provide construct validity tests
• Provide individual-level explanation of variance
  in WTP

11
Example Addressing the Risks of Energy Imports
Over the next 20 years it is expected that
crop-based and renewable resources could
re-place a substantial portion of fossil fuels
used currently in the US. These changes would
reduce dependence on unstable sources of oil,
while also reducing emissions of greenhouse
gasses and other pollutants. While the benefits
of such research and development efforts may take
many years to significantly reduce reliance on
fossil fuels, the investment in the research must
be made much earlier. These kinds of energy
research and development would be expensive,
requiring new sources of funding. Suppose that a
national advisory vote or referendum was held
today. You could vote to advise the federal
government whether to develop a new National
Energy Research and Development Fund from
additional fees on fossil energy use. The fees
would apply to purchases of electricity and
products and services that rely on coal, oil,
and natural gas.
12
CV Payment Vehicle
If you were confident that this new fund would
help create new energy sources and reduce US
dependence on foreign oil, even if creating this
National Energy Research and Development Fund
would cost your household cost per year in increased energy prices for
such things as electricity and gasoline, would
you vote for or against creating the National
Energy Research and Development Fund? Keep in
mind that the per year that you
spend on increased energy prices could not be
spent on other things, such as other household
expenses, charities, groceries, or car payments.
13
Measuring WTP Certainty
Asking the same question in another way, suppose
that a national advisory vote or referendum was
held today, and you could vote to advise the
federal government on whether to create a
National Energy Research and Development Fund,
but the fund would cost your household randomly selected cost per year in increased
energy prices. Where would you place yourself on
a scale from zero to 100, where zero means you
are absolutely certain that you would vote
against the creation of the fund and 100 means
you are absolutely certain that you would vote
for it?
14
Calculating Willingness to Pay

• Estimating a latent willingness to pay function
• Where x is a vector of independent variables, ?
  is a vector of coefficients, ? is a scale
  parameter, and e is an error term
• WTP is inferred from yes votes
• Where t is the payment amount, and G is the
  distribution function for the probability
• G is defined to permit positive and negative WTP
• A negative WTP implies one is less well off with
  the creation of the ERDF by the WTP amount

15
CV Estimates of WTP for ERDF

• Results satisfy construct validity tests
• WTP increases with income, environmental worry,
  and concern about future energy supplies
• WTP decreases with political conservatism, and is
  lower for women than for men
• Split experimental design
• The effect of inclusion of nuclear RD is
  nominally negative, but statistically
  insignificant
• This provides a base-line for monitoring change
• Counting only highly certain yes votes (80
  certain), preliminary average WTP per household
  is 432 per year
• Nationally about 48.9 billion annually

16
Willingness to Pay by Certainty
National Energy RD Fund
1,200
1,000
800
600
Standard is 80 Certain
Household WTP in / Year
400
200
0
70
75
80
85
90
-200
-400
Certainty
17
Challenges to QRA

• Does model complexity permit reasoned decisions
  by elected decision makers?
• How do we decide what impacts matter?
• E.g., How do we trade security and liberty?
• How do we decide the weights of the different
  dimensions of impacts?
• Given the trajectory of uncertainty over time,
  how can we be confident that the QRA captures the
  essential elements?
• Why should we trust the experts?
• E.g., Complaints about radiation experts
• Is it all really just politics?