Uncertainty
of parameter estimates can affect the results of cost estimates.
Generally, when dealing with uncertainty in cost benefit analyses,
economists perform sensitivity analyses.
This involves varying the values for uncertain parameters and identifying
the result of this variation on the final outcome.
This discussion first considers scenarios in climate change policy where
traditional sensitivity analysis will not work and then looks at the opportunity
to provide more useful decision making criteria than traditional sensitivity
analysis.
Uncertainty and the need for
multiple baselines
In
order to account for the uncertainty associated with underlying development
paths, it is necessary to use multiple scenarios to develop alternative base
cases that represent different, internally consistent patterns of development [1]. The uncertainty that
necessitates this use of multiple baselines is not the uncertainty associated
with parameters such as population or GNP growth rates. The effects from uncertainty with those parameters can be
explored using sensitivity analysis. However,
when dealing with uncertainty regarding different underlying socioeconomic
development paths that would give rise to different emission scenarios and costs
of mitigation, one cannot use probability distributions to determine the most
likely pattern [1]. The common
practice is to use multiple scenarios to account for these uncertainties in
order to get multiple equilibrium scenarios in the long run.
These alternative base cases represent different, but internally
consistent, patterns of development.
Uncertainty and
the use of sensitivity analysis
Two studies conducted by Manne and Richels explore the role of sensitivity analysis in dealing with uncertainty and attempt to expand that tool from identifying alternative outcomes, to identifying most probable outcomes. They contend that the first step policy makers should undertake is to determine the sensitivity of today’s decisions to major uncertainties of the greenhouse debate [10]. This will enable decision makers to identify which uncertainties may have a critical impact on near-term decisions. For these uncertainties, the probability of possibilities can be considered to help identify an “emissions strategy that maximizes the expected discounted utility of consumption in each region over time” [10]. A conflicting school of thought takes the maximin approach method to decision making. This approach focuses on the worst case scenario and makes decisions accordingly [5].
The
following example will illustrate how the expected utility approach can be useful for decision
makers interpreting or analyzing costs. For
details about the methodology used, see Manne and Richels, “The Greenhouse
Debate: Economic Efficiency, Burden Sharing and Hedging Strategies” [10].
The sensitivity analysis Manne and Richels conducted examined six
alternative cases, including a base case; a case which assumed high GDP growth,
a case that included an optimistic backstop cost; a case that assumed high
temperature sensitivity; a case that assumed high willingness to pay to avoid
nonmarket damages; and a case that considered high damages [10]. The results
indicated that, except for the case of high damage scenario, the optimal choice
of emissions is insensitive to the choice of parameter values [10].
This
result identifies the high damages scenario, which combines the assumptions of
high temperature sensitivity and high nonmarket damages, as the uncertainty that
will have a significant impact on near-term decisions.
At this point, analysts can consider the probability that the high damage
scenario will occur and then identify a strategy that limits emissions to a
level between a high probability and low probability case.
Until a point in the future when the uncertainty is resolved, the use of
this “hedging strategy” provides the best policy strategy [10].
Using the maximin approach, a different decision would be recommended. In this approach, society is only concerned with worst possible outcome (i.e. early catastrophe) and seeks to maximize the net benefits to society under this scenario [5]. Naturally, abatement targets are much higher under this approach, which would translate to much higher mitigation costs. Neither approach is ideal. The expected utility approach requires knowledge of the probability distribution for events resulting from climate change, and the maximin model implicitly makes an extreme and unrealistic assumption about risk aversion [5]. The currently preferred method is the expected utility approach.