Bargaining and Coalition Formation Dr James Tremewan - - PowerPoint PPT Presentation

Bargaining and Coalition Formation

Dr James Tremewan (james.tremewan@univie.ac.at) International Environmental Agreements

International Environmental Agreements

Overview

• International Environmental Agreements (IEAs) typically involve

negotiations between a large number of countries so are an interesting application of coalition-formation theory.

• The global environment is a public good, and IEAs are

vulnerable to two types of free-riding:

• Choosing not to sign an agreement.
• Signing an agreement but failing to comply with its terms.
• We will focus on the first, assuming that all signatories comply

with an agreement.

• We will look at four specific issues: uniform emission reduction

quotas, transfer schemes, single vs multiple agreements, and far-sighted vs myopic agents.

• The first three issues will be analyzed using the same

computable general equilibrium (CGE) model: STACO.1

1See ”Stability of climate coalitions in a cartel formation game”,

Finus, van Ierland, and Dellink (2006).

2/27

International Environmental Agreements

Analytical vs Numerical Analysis

• Economic models can be solved analytically (finding a ”closed

form” solution, possibly involving various parameters) or numerically (as in CGE models).

• Closed form solutions:
• Allow us to find general (especially qualitative) results that are

independent of specific functional forms or precise parameter values.

• But to find closed form solutions we must often make extreme

symplifying assumptions (e.g. identical agents, linear costs...), and if results depend on parameter values, what range of values is ”realistic”?

3/27

International Environmental Agreements

Analytical vs Numerical Analysis continued

• Numerical models:
• Allow analysis of more complex environments and can be

”calibrated” to be close to the real world and give specific policy advice.

• But require estimates of parameters which may be hard to

evaluate, and different estimates may give different results (ameliorated by ”sensitivity analysis”: testing a range of parameter values).

• Note that both types of analysis will mostly use similar

assumptions in the underlying models, such as optimizing behaviour by individuals, Nash equilibrium, etc. so both are equally sensitive to modelling errors in this regard.

4/27

STACO

STAblility of COalitions model (STACO)

• Two-stage game (baseline model):
• Regions decide whether or not to sign agreement.
• Coalitions (and singletons) choose abatement strategies.
• Abatement strategies are chosen based on the following payoff

function: πi(q) =

T

• t=1

(1 + ri)−t(Bit(qt) − ACit(qit)),

• T is the time horizon,
• ri is the discount factor of region i,
• qt =

N

• i=1

qit, the sum of individual abatement.

• Bit(qt) is a country’s benifit from global abatement.
• ACit(qit) is a region’s cost from individual abatement.
• Note incentive to free-ride: private cost to abatement, but

global benefit.

5/27

STACO

STACO: Solution Concepts

• In the baseline model, coalitions are assumed to choose

strategies to maximise the sum of their members’ payoffs (singletons maximise their own payoff), taking the strategies of

• thers as given.
• A coalition structure is considered stable if no country has an

incentive to change its membership strategy, taking the decisions

• f others as given. Requires:2
• Internal stability: no coalition member wants to leave.
• External stability: no non-member wants to join.
• With no agreement (Singleton coalition structure) we have the

classical Nash equilibrium.

• If everyone signs an agreement (Grand Coalition) we have the

social optimum.

2This is just a Nash equilibrium: coalitional deviations are not

considered.

6/27

STACO

STACO: Callibration

• Analysis looks at the 100-year period from 2011-2110.
• World divided into 12 regions: USA, Japan, European Union

(EU-15), Other OECD Countries, Central and Eastern European Countries, Former Soviet Union, Energy Exporting Countries, China, India, Dynamic Asian Economies, Brasil, and ”Rest of the World.”

• Abatement costs and benefits callibrated using pre-existing

estimates (from mid-90’s so probably now out of date).

• Discount factor assumed to be 2%.
• Clearly heroic assumptions, but:
• Authors primarily interested in stability of coalition structures,

so precision not so important.

• Alternative specifications are tested and qualitative results are

robust.

• What is the alternative?

7/27

STACO

Preliminary notes

• In the Singleton coalition structure, individual rationality implies

each country equating its marginal abatement cost (MAC) to its marginal abatement benifit (MAB).

• In a coalition, efficiency implies that the marginal cost of

abatement will be equal for each coalition member: Countries with shallower marginal cost curves will take a larger share of a coalition’s abatement.

• The attractiveness of a coalitions depends upon:
• the country’s MAC relative to other coalition members (the

share of additional abatement it must take).

• the country’s MAB (how much it benefits from the additional

abatement).

8/27

STACO 9/27

STACO 10/27

STACO 11/27

STACO 12/27

STACO

Summary of Basic Results

• Even without cooperation (i.e. in the Singleton coalition

structure) it is in every country’s interest to unilaterally reduce emmisions (a total of 4.6% of global emmisions).

• With global cooperation, abatement is significantly higher

(21.4%).

• The total net benefit of abatement for the Grand coaltion is

$6,031 billion compared with $1,960 in the Singleton coalition structure: large potential gains from cooperation.

• All regions except Japan and the EU have an incentive to leave

the Grand Coalition (high MAC so small share of abatement costs, and high MAB so remaining coalition members would reduce abatement a lot): it is not stable.

• EET and China are worse off under the Grand Coaltion than the

Singleton coalition structure (because of low MAB for EET; and low MAC for China implying they bear much of the abatement costs).

13/27

STACO

STACO: Sensitivity analysis

• Under the baseline calibration, NO coalition is stable.
• Raising the ”global benefit of abatement” parameter:3
• lower emmisions in all coalition structures.
• by 20% results in a stable coalition between EU and Japan (but

little extra abatement compared to the new non-cooperative NE because of their high MACs).

• no other major changes (tested up to 300% of original

callibration).

• An alternative calibration for regional shares of benifits finds
• nly one stable coalition: Japan, Brasil, and ROW (again with
• nly minor improvement over non-cooperation).
• Conclusion: in the baseline model, useful cooperation appears

unlikely.

3Altering benefits from abatement can also be viewed as altering

discount factors as they work in a similar way in this model.

14/27

STACO

Uniform Emmision Reduction Quotas4

• The baseline model assumes that coalitions maximise their total

payoff which implies abatement is allocated efficiently (and probably asymmetrically) across members.

• However, many IEAs specify UERQs:
• Helsinki Protocol (30% reduction in sulphur emmisions).
• Sofia agreement to freeze N2O at 1987 levels.
• ”Protocol Concerning the Control of Emissions of Volatile

Organic Compounds or Their Fluxes” (30% reduction).

• Clearly inefficient: for a given level of abatement, abatement will

not be distributed across countries in a cost-effective way (unless MACs identical across coalition members)... why so attractive?

4See Altamirano-Cabrera et al (2008).

15/27

STACO

UERQs: Alternative Models

• Common Quota: coalition maximises total welfare under the

constraint that all regions reduce emissions by the same percentage.

• Median Quota Proposal: each member proposes a coalition

abatement reduction that is individually optimal, and the median reduction is chosen (corresponds to majority voting).

• Lowest Quota Proposal: each member proposes a coalition

abatement reduction that is individually optimal, and the lowest reduction is chosen (corresponds to unanimity voting).

• For the grand coalition, these imply global welfare of:
• Common Quota: $4,589 bln
• Median Quota Proposal: $4,565 bln
• Lowest Quota Proposal: $3,212 bln
• Much worse than social optimum ($6,021 bln)... but need to

consider stability.

16/27

STACO 17/27

STACO

UERQs

• UERQs are clearly inefficient, but enable stable coalitions to

form.

• All UERQ models result in non-trivial coalitions which improve

global welfare relative to the Singleton coaltion structure.

• The best is the Lowest Quota Proposal (the worst when stability

is not taken into consideration): lower abatement targets are made up for by wider participation.

• Also, abatement costs (i.e. gains from cooperation) spread more

evenly.

18/27

STACO

UERQs with trading

• Allowing trading of emission quotas should reduce inefficiency by

shifting abatement to regions with lowest MAC.

• Potential information problems: pay a country to abate, but how

can abatement be verified?

• For now, assume trading is efficient: price of quota equal to

equilibrium MAC (which will be the same in every country).

19/27

STACO 20/27

STACO

UERQs: Conclusions

• In a second-best world with free-rider incentives and no

third-party enforcement, UERQs can improve on ”Efficient Design” by allowing larger stable coalitions.

• Concensus treaties (Lowest Quota Proposal) result in lower

abatement targets than Majority Rule, but this can be made up for by wider participation.

• Allowing trading of quotas, as in Kyoto Protocol, can improve

participation and abatement targets.

21/27

STACO

Monetary Transfers

• Part of the success of UERQs due to ”fairer” sharing of the

gains from cooperation... an alternative is allowing monetary transfers between countries.

• Transfer schemes are rules for allocations (within the coalition)
• f the gains from cooperation.
• Pragmatic schemes (more or less preserve the status quo and

favour industrialized countries):

• In proportion to GDP.
• In proportion to Current Emissions.
• Morally motivated rules:
• Equal Sharing.
• Ability to Pay: in inverse proportion to per capita GDP.
• In Inverse proportion to Emissions (historical responsibility).
• In proportion to Population.

22/27

STACO 23/27

STACO

Monetary Transfers: Conclusions

• Transfers can improve the possibility of cooperation, e.g. the EU

can compensate China for the extra abatement they take on because of their lower MAC.

• The Grand Coalition is still not stable, so the transfer schemes

considered cannot completely offset free-rider incentives.

• The ”pragmatic” schemes are more effective than the ”morally

motivated” schemes.

24/27

STACO

Multiple Agreements (and Membership Rules)

• So far we have assumed that only one agreement can form, to

which regions sign up or not. The idea of some kind of global agreement is appealing, but why not allow multiple agreements?

• Also, some coalitions in the baseline model where internally

stable, but not externally stable: no region wanted to leave the coalition, but one or more regions wanted to join (and if they did, the additional abatement resulting would cause existing members to leave).

• We now consider allowing multiple agreements under three

different membership rules:

• Open Membership: any region may join any coalition.
• Exclusive Membership- Majority Voting: a region may only join

a coalition if a majority of existing members are made better off.

• Exclusive Membership- Unanimity Voting: a region may only

join a coalition if all existing members are made better off.

25/27

STACO 26/27

STACO

Multiple Agreements (and Membership Rules)

• Stable, non-trivial, welfare-improving coalitions exist under

exclusive membership rules, more so under unanimity voting (again the increase in membership outweighs the lower abatement targets):

• IEA typically have open membership... perhaps this should be

reconsidered.

• Other (stable) international agreements have exclusive

membership, e.g. WTO with unanimity voting.

• Multiple agreements can increase global abatement:
• With only one agreement, choice is to join or free-ride. Multiple

agreements allow middle ground when the single agreement is too stringent.

• Regions with similar MAC and MABs club together as large

differences imply unequal divisions of surplus which is bad for internal stability: US demand for participation of developing countries in Kyoto open to criticism.

27/27