CATASTROPHIC EVENTS AND STOCHASTIC COST-BENEFIT
ANALYSIS OF CLIMATE CHANGE
An Editorial Comment
The role of cost-beneﬁt analysis (CBA) in making decisions about climate change
has been a contentious issue for long. The debate was sparked by the publication of
Nordhaus’s DICE model (Nordhaus, 1994). His approach boils down to the view
that emissions should be reduced as long as the marginal beneﬁt of doing so (i.e.,
avoided emissions) is higher than the marginal cost. In doing so, he found that only
modest abatement levels were warranted (from an economic point of view), e.g.,
a 14% reduction by the year 2100 below baseline emissions. But since baseline
emissions grow, his optimal CO
emission trajectory amounts to an increase of
the emissions by a factor of three above current levels. Several other studies using
similar approaches end up with similar results, e.g., Manne et al. (1995) and Peck
and Teisberg (1993).
These results have been challenged by a growing number of studies. Cline
(1992), Azar and Sterner (1996), Rougharden and Schneider (1999) and others
have estimated the marginal cost of CO
emissions or the optimal level of emission
reduction to be much higher than Nordhaus. Others, e.g., Schneider (1993, 1997),
Grubb (1993), Azar (1998), Azar (2000), Schneider et al. (2000), have focused on
the value-laden assumptions that have to be made in the context of cost beneﬁt
analysis of climate change, e.g., as regards the discount rate, low-probability but
catastrophic events, the valuation of non-market impacts (for instance, losses of
lives and biodiversity) and the way distributional issues are dealt with, or rather
not dealt with, in cost-beneﬁt analysis.
In this issue of Climatic Change, Richard Tol (2003) enters the camp of those
pointing towards difﬁculties in applying cost-beneﬁt analysis to the problem of
climate change. He writes that ‘it seems as if the uncertainty about climate change
is too large to apply cost-beneﬁt analysis’. We are critical about the applicability
of cost-beneﬁt analysis in the context of climate change, but for reasons other than
those Tol puts forward.
In this commentary, we will ﬁrst explain why we disagree with his methodol-
ogy (Section 2), and then try to demonstrate why cost-beneﬁt analysis offers little
guidance when it comes to actual decision making in the greenhouse (Section 3).
We will present results from a stochastic version of Nordhaus’s DICE model that
demonstrate that a small probability for a catastrophic event may substantially alter
Climatic Change 56: 245–255, 2003.