Climate Clubs:
Designing a Mechanism to Overcome
Free-riding in International Climate Policy
William Nordhaus, Yale University
Presidential Address
American Economic Association
Environmental Seminar
January 4, 2015
1
Outline
• Introduction to current state of climate
economics
• The problem of free riding
• The concept of a Climate Club
• Modeling club formation and equilibrium
• Effectiveness of different Club regimes
2
Four key issues for climate change
1. Climate science: Likely path
Mature science and
of earth system over this century projections
and beyond
2. Estimate costs and benefits of
emissions reductions
Mature estimates of costs;
very rudimentary
determination of damages
3. Instruments for implementing Well understood for national
policy
carbon taxes and national capand-trade systems
4. System to prevent
international freeriding
Zero progress
3
Most recent data
4
CO2 concentrations, Hawaii (ppm)
CO2 concentrations through August 2014
400
380
360
340
320
300
1958
1969
1980
1991
2002
2013
5
Global temperatures (prelim 2014 data)
Global mean temperature, deg C (1900 = 0)
1.2
1.0
0.8
Average
NCDC
Hadley
GISS
0.6
0.4
0.2
0.0
-0.2
1900
1925
1950
1975
2000
6
US decarbonization
2.0
1.6
1.4
1.2
1.0
0.8
0.6
0.4
CO2/GDP
Decarbonization at 2.3% per year
0.2
1930
1940
1950
1960
1970
EPA estimates, real GDP, 5 months for 2014
1980
1990
2000
2010
7
Free-riding equilibrium for public goods
• Free-riding occurs when someone receives the benefits of
a public good without contributing to the costs.
• This syndrome is seen widely for public goods or
“tragedy of the commons” (whales, global warming,
contagions)
• Because of structure of international law, strong
tendency for free-riding in global public goods.
–
–
–
–
Public goods theory from Paul Samuelson
History and treaty theory from Scott Barrett
Kyoto Protocol for climate change (later)
Modeling simulations (later)
8
Free-riding in the Kyoto Protocol:
Share of global emissions covered by binding restraints
Covered emissions as % of world
70
Kyoto participants
60
50
40
30
20
10
0
1990
1995
2000
2005
2010
9
Free riding in NATO
800
United States
Military spending (billions)
700
600
500
400
300
200
UK
Germany
France
100
Luxembourg
0
10
100
1,000
GDP (billions)
10,000
Source: Data
from NATO
10
Free-Riding in International Climate Agreements
• Basic theoretical results:
– Without international agreements, have noncooperative (NC) equilibrium. In simple example,
carbon prices are efficient levels time Hirfindahl
index of country size (≈ 10% of efficient).
– With international cooperation and bottom-up
treaties without sanctions, have small coalition paradox:
Stability can sustain only a small number of countries
(2 or 3).
• Climate Club: Top-down treaty with penalties for nonparticipants: Can lead to high participation with efficient
abatement.
11
International Treaties as “Clubs”
Clubs are agreements where:
1. Have economies of scale or public goods
2. Members pay dues
3. Can exclude non-members (avoid free riders)
4. Stability issues (next slide)
Examples of effective club: Why did Greece stay in EU?
Kyoto Protocol defective club:
membership cost > membership value
12
Should Greece stay in the EU?
The Times (London)
13
International Treaties as Games
• Climate policy without penalties is repeated n-person
prisoners’ dilemma (PD) game.
– Presumption is that high discount rate (or low frequency
of decisions) will lead to PD rather than cooperative
equilibrium.
• By adding penalties for non-participants, payoffs change
so that stage game has (relatively) efficient Nash
equilibrium.
– Presumption is then that the repeated game has the stagegame efficient Nash equilibrium.
– In Scott Barrett’s language, treaties are “self-enforcing” at
efficient level.
• Key issue is “coalition stability” of high-participation
treaty.
14
Penalties are necessary for effective climate treaties
• Need penalties on non-participants to induce
participation in deep abatement treaties
• History and law suggest the most practical penalty
is trade sanctions
• What kind of sanctions?
– Standard approach: Countervailing duties on
carbon content of imports (US and EU legislation)
– Climate Club tariffs: Simple ad valorem tariff on all
imports of non-participants into climate-club
regions.
15
Modeling Climate Clubs with the TRICE model
• Designed a new model to study properties of Climate
Club with realistic country parameters.
– TRICE model (Trade in a Regional Integrated Model of
Climate and the Economy)
• Model assumptions:
– A standard one-period regional model.
– Key variables are the social cost of carbon (SCC), national
carbon prices, tariff rates, and national income.
– Countries can form Carbon Club to set carbon prices
jointly at international carbon price target.
– Clubs can impose penalty tariffs on imports of nonparticipants.
16
Objectives of modeling
1. Examine different Club structures or regimes
(carbon prices and penalty tariffs).
2. Determine whether regimes contain stable
coalitions (Nash coalition stability).
3. Determine effectiveness of regimes (whether actual
carbon price approaches target price).
17
Algorithmics
• Thought to be a NP-hard problem to find optimal
coalition.
• Designed “evolutionary” algorithm to find coalition which
usually find the stable coalition in < 500 mutations.
• Decision criterion is “coalition Nash.” No sub-coalition can
improve its welfare by leaving and/or joining.
18
Data for model for 2011
• Damage functions: Simplified from Nordhaus survey of
estimates (JAERE, 2014)
• Abatement functions: From different models for
aggregate and McKinsey estimates for regions.
• GDP, emissions, population from World Bank
• Trade data from UNCTAD.
• Parameters for trade model from Ralph Ossa (AEA,
forthcoming, 2014).
• 15 regions (US, EU, China, India, Japan, Brazil, Russia,
Canada, and other aggregates).
19
Experiments with the TRICE model
1. Kyoto Protocol
2. Climate Clubs
– Tariff rates from 0% to 10%.
– Carbon tax target is from $25 to $100 per ton CO2
(rough range of proposals).
20
With no penalty, Kyoto regime disintegrates to NC
14
Number participating regions
12
10
8
6
4
2
0
1
101
201
301
401
501
601
701
801
901
1,001
Iteration number
- Assumes carbon tax = $25 and tariff rate = 0%
21
Now look at results for positive tariffs:
What are results for different Climate Clubs?
• Penalty tariffs are uniform on all non-participants
• Rates from 0% to 10%
• Global social cost of carbon $12½ , $25, $50, $100
22
Participation by tariff rate for $50 carbon price
Tariff rate
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
Number
participants (of 15
regions)
0
1
8
13
14
14
14
15
15
15
15
Today’s
free-riding!
23
Participation by tariff rate for $50 carbon price
Number participating regions
15
Bars are tariff rates:
0% at left to 10% at right
12
9
6
3
No (zero)
participants at
0% tariff
0
24
Number participants by tariff and carbon price
25
Carbon price by tariff and target price
26
Net income gain (% of potenital)
Gain from regime (% of cooperative)
100%
0.00
0.01
90%
0.02
0.03
80%
0.04
0.05
70%
0.06
0.07
60%
0.08
0.09
50%
0.10
40%
30%
20%
10%
0%
12.500
25.000
50.000
Target carbon price ($/tCO2)
100.000
27
Where are the votes?
For heterogeneous countries with differing national SCC,
abatement costs, and damages, what level of international
target carbon price would then vote for?
28
What Climate Club would countries prefer?
Region
Global target carbon price
that maximizes domestic
welfare for club of 15
($/tCO2)
Global averages
Non-cooperative price
3
Global SCC
25
Median preferred price
28
Country preferred prices
China
14
US
28
India
31
Canada
34
EU
46
29
Summary
1. Strong international free-riding leads to minimal
abatement with Kyoto Protocol structure of no
penalties.
2. Strong incentive-compatible agreements can be
supported with penalties such as tariffs on nonparticipants.
3. Most important takeaway: With Club structure,
countries acting in their national self-interest can
produce (reasonably) efficient global climate policy.
30