Transcript Discussant - Princeton University

Comments on Dietz, Groom & Pizer:
“Weighing the costs and benefits of
climate change”
March 13, 2015
Prof. C. D. Kolstad
Stanford University
1
General structure of paper
I. Climate change damages and mitigation
costs
II. Societal discounting over time
III. Representations of social welfare functions
IV. The role of population
2
The evaluation charge: ABCDE
A. The paper is Accurate
B. or at least Balanced
C. The paper is not exactly Comprehensive. In some areas,
there is far too much detail; other clearly connected issues
are omitted. Also, the paper is an early draft -- reads like
several papers stapled together. References are missing for
much of the theory part of the paper.
D. The paper presents little empirical evidence and thus the
issue of Discriminating between weak vs. strong evidence is
not particularly relevant.
E. On the Everyone is a reader test, the paper regrettably fails.
1.
2.
The first section (related to costs and benefits) is widely accessible
or could be made so easily by an editor.
The remainder of the paper is highly technical and not particularly
related to children (though all intertemporal tradeoffs are
ultimately related to children). It is not accessible to the average
educated reader.
3
General Remarks
• The topic of the paper is of central importance
• The first section is accessible and ties the issues in the paper together
nicely – carry it through the rest of the paper
– Don’t overemphasize the credibility of the 2 degree target
• The discounting discussion is largely a review and could benefit from
simplification, perhaps referencing similar reviews (eg, IPCC)
– Suggest adding tabulation of social rates of return on investments in
other arenas which benefit children – eg education, infant health,
vaccination programs, clean water, long-lived infrastructure
• The different types of social welfare functions are incomplete.
Discussion incomplete on implications – such as using welfare weights
to aggregate the costs and benefits discussion in section I.
• The population section is a nice addition (though perhaps not for the
readers of the special issue). As the IPCC articulated, this is a thorny
issue with scant economics literature. Suggest referencing The
Ultimate Externality by Jon Harford. The implications for applied
models of climate change (back to section I) are profound and could
be underscored.
4
Mitigation costs from IPCC incomplete; feasibility of 2 degrees
questionable
Mitigation cost only – residual damage excluded
3o
2o
5
Recent estimates of CC Damage--Sparse
5.0
2.5
Tol 2002a
Rehdanz and Maddison 2005
Fankhauser 1995
Tol 1995
Nordhaus and Yang 1996
Plamberk and Hope 1996
Mendelsohn et al. 2000a
Mendelsohn et al. 2000a
Nordhaus and Boyer 2000
Rehdanz and Maddison 2003
Hope 2006
Nordhaus 2006
Nordhaus 1994b
Nordhaus 1994a
Nordhaus 2007
Nordhaus 2007
Bosello et al. 2012
Maddison and Rehdanz 2011
2.3
0.9
0.0
-0.4
-0.5
0.1
-0.1
-1.5
-1.9
-2.5
%GDP
-2.5
0.0
-0.9
-1.4
-1.7
-2.2
-1.3
-4.8
-5.0
-7.5
-10.0
-10.7
-11.5
-12.5
0.0
1.0
2.0
3.0
4.0
5.0
degree centigrade
6.0
7.0
6
1. Discounting: Positive or Normative
Approach?—from IPCC review
Real returns on financial assets:
Australia
France
Japan
United Kingdom
USA
Government Bills
(maturity <1 year)
1900-2006 1971-2006
0.6%
2.5%
-2.9%
1.2%
-2.0%
0.4%
1.0%
1.9%
1.0%
1.3%
Government Bonds
(maturity =10 years)
1900-2006
1971-2006
1.3%
2.8%
-0.3%
6.6%
-1.3%
3.9%
1.3%
3.9%
1.9%
4.0%
Equity
1900-2006
7.8%
3.7%
4.5%
5.6%
6.6%
1971-2006
6.3%
7.8%
5.0%
7.1%
6.6%
• Some Problems using market rates
– Market may not work with some agents unable to trade (future
generations)
– Markets driven by individuals trading their own consumption
across time, not societies across generations
– We do not observe safe assets with maturities similar to climate
impacts
Source: Chapter 3, WGIII, AR5, IPCC
7
Normative approach to Social Rate of Discount
• Ramsey Rule—largely a tautology
– Social Discount rate: ρt = δ + ηgt [and extended with Var(g)]
– Where
• δ is pure rate of time preference
• g is growth rate of economy
• η is measure of inequality aversion/curvature of social welfare
function (Prioritarian  η > 0)
Author
Cline (1992)
IPCC (1996)
Arrow (1999)
UK: Green Book (HM Treasury, 2003)
US UMB (2003)**
France: Rapport Lebègue (2005)
Stern (2007)
Arrow (2007)
Dasgupta (2007)
Weitzman (2007a)
Nordhaus (2008)
Rate of pure
preference for
present
0%
0%
0%
1.5%
0%
0.1%
0.1%
2%
1%
Inequality
aversion
Anticipated
Growth rate
Implied social
discount rate
1.5
1.5-2
2
1
1%
1.6% - 8%
2%
2%
2
1
2-3
2-4
2
2
2%
1.3%
1.5%
2.4% - 16%
4%
3.5%*
3% - 7%
4%*
1.4%
2%
2%
6%
5%
8
Social Discount Rate Highly Dependent on
Assumed Growth Rate (g) and Var(g)
Country
Discount rate
Ramsey rule
Extended Ramsey
Equation 3.6.4
rule Equation 3.6.5
g
σ (std dev g)
1.74%
2.11%
3.48%
3.35%
OECD countries United Kingdom 1.86%
2.18%
3.72%
3.58%
Japan
2.34%
2.61%
4.68%
4.48%
China
7.60%
3.53%
15.20%
14.83%
3.34%
3.03%
6.68%
6.40%
Russia
1.54%
5.59%
3.08%
2.14%
Gabon
1.29%
9.63%
2.58%
-0.20%
Zaire (RDC)
-2.76%
5.31%
-5.52%
-6.37%
Zambia
-0.69%
4.01%
-1.38%
-1.86%
Zimbabwe
-0.26%
6.50%
-0.52%
-1.79%
United States
Economies in
India
transition
Africa
Country-specific discount rate computed from the Ramsey rule using the historical mean g and
standard deviation σ of growth rates of real GDP/cap 1969-2010, together with 𝛿 = 0, and 𝜂 =
2. Extended Ramsey Rule includes uncertainty.
9
Simple Illustration: Prioritarianism
10
2. Should CC CBA be using welfare weights?
• Is compensation principle ethically justified?
– Costs to a person can be offset by benefits to same person--just
– Not true for costs to one person offset by benefits to another—may be
unjust
• Ethical justification of CBA depends on expression of social welfare
– SWF
• CBA can still be useful even if its findings are ethically ambiguous
– Optimal growth can be viewed as simulating a market response
• CBA could be weighted to have a prioritarian ethical foundation.
– Suggestion that if income distribution optimal then unweighted CBA
correct—generally wrong due to prioritarianism
– Generally for CBA to yield ethical conclusions, need welfare weights—
provided existence of social welfare function is assumed
• How to determine welfare weight?—not easy
– One approach is to use inverse of VSL
– Assume one life of equal social value no matter where
– Then weights used to equate VSL among countries
11
3. The significance of uncertainty to policy
• Suppose climate sensitivity is uncertain
– Median 3oC change from CO2 doubling
– 15% probability change will exceed 4.5oC
• Effect of distributional assumptions:
• What if damage big in the tails?
– Expected value of damages averted from mitigation may
be dramatically different
• Therefore the nature of uncertainty vital to understand
the costs and benefits of carbon mitigation (and the
social cost of carbon)
Table source: Weitzman (2011): REEP
12