Munich Economic Summit
Friday, 29 May 2009
Carlo Carraro
University of Venice, FEEM and CESifo
The road to Copenhagen. Can climate policy be effective? Are there sufficient economic incentives to agree on climate change control?
What are the prospects for an agreement in Copenhagen? Can the world live with a few opt-outs? Are there policy measures to enhance the participation incentives of developing countries? The purpose of this paper is to address these questions by analyzing: i) the incentives for main GHG emitting countries to participate in an international climate policy agreement, i.e. in a "climate coalition"; ii) the effectiveness of such climate coalitions in actually reducing GHG emissions, even when some countries opt-out; iii) the costs and benefits associated with different climate coalitions; iv) the potential for improving the incentives to participate in a climate agreement through international financial transfers and arrangements.
The analysis is carried out using the WITCH model. Incentives to sign a climate agreement will ultimately depend on a wide range of economic and political factors, not all of which can be captured by a climate-economy model. Nevertheless, useful insights can still be gained by focusing on economic incentives, which in the WITCH model include the avoided damages and the abatement costs incurred both within and outside a coalition.
1. Assessing the incentives for the main world regions to participate in an international climate policy agreement
This assessment crucially depends upon the estimates of climate damages.
Two different assumptions on climate damages (low, following Nordhaus' estimates
of climate damages and high, following the estimates in the IPCC FAR) and
two assumptions of the pure rate of time preferences (0.1% like in the Stern
Review and 3% declining as in Nordhaus's work) are considered. Damage functions
are also specified to reflect the following major features of the climate
cooperation game:
- Damages from climate change are unequally distributed across world regions.
Some countries may benefit from climate change (e.g. Transition Economies
or Canada) at least for limited increases of temperature. Other countries,
e.g. some developing countries, are more vulnerable and may suffer large losses.
Sub Saharan Africa, South Asia and Western Europe are likely to suffer the
larger losses from climate change;
- Marginal abatement costs are also unequally distributed across world regions.
Abatement costs are likely to be lower in India, South East Asia, Sub Saharan
Africa and China. They are much higher in Japan, Western Europe and Transition
Economies.
As a result, and considering that what matters for participation incentives is the balance between abatement costs and benefits from avoided damages, countries with the highest incentives to free ride are likely to be China, Transition Economies and Middle East countries. Ceteris paribus, regions with flatter (steeper) marginal abatement cost curves and/or flatter (steeper) marginal damage curves have larger (smaller) participation disincentives, because they contribute more (less) to the coalition's abatement effort and/or benefit less (more).
2. Analysing the size and stability of possible climate coalitions without international financial transfers
As a first important step, the analysis focuses on "potentially effective coalitions (PECs)". A set of n countries is a potentially effective coalition (PEC) if the sum of their technical lower bound emission levels (zero), added to baseline (BaU) emissions of non-participating countries (singletons), results in a concentration level which is below or equal the target by 2100. The existence of at least one PEC is a necessary (but not sufficient) condition for the 550 ppm CO2eq target to be attainable. It is only necessary because a PEC is defined by an emission profile which is lower than the actual emission profile of the coalition. Therefore, if the target is not achieved by the lower emission profile, it cannot be achieved by the equilibrium emissions of the coalitions.
Results show that only seven coalitions (out of 4095 when the world is divided into 12 regions) are PECs with respect to the 550 ppme concentration target in 2100. All these PECs are profitable (they produce a surplus which is sufficient to compensate countries that loose from signing the climate agreements). In these PECs:
- The participation of both China AND India is needed to attain the 2100
target
- When the goal is GHG stabilization in 2100, PEC are subsets of the 12 regions
in which at most three regions are not included (SSA, TE, SEASIA)
- Generally, only SSA or SSA plus another region (LAM, TE, MENA, SEASIA) can
be singletons. All the other countries/regions should sign the climate agreement
for the target to be attainable.
Two additional important policy messages emerge from the analysis:
- When emissions of countries belonging to a PEC and the related emissions
of singletons are the equilibrium ones, i.e. emissions are the outcome of
a cost benefit strategic decision analysis, then only the Grand Coalition
(all countries) achieves the 550 ppme target in 2100. The Grand Coalition
less Sub Saharan Africa gets close to the target. All the other coalitions
cannot achieve the target.
- No PEC is stable, i.e. there is always at least one country/region that
gains from free riding on the other countries' abatement efforts.
These conclusions apply in particular to the following six "politically
important" coalitions: i) "the Grand Coalition" to which all
countries belong; ii) the coalition formed by industrialised countries and
by China, India, Russia and Latin America; iii) the coalition with industrialised
countries, China, India and Russia; iv) the coalition with industrialised
countries, China and India; v) the coalition with industrialised countries
and China only; vi) and, finally, industrialised countries only.
Only (i), the Grand Coalition, can actually achieve the 550 ppme target in 2100. However, the Grand Coalition, as well as all other coalitions, is not stable.
3. Analysing the size and stability of possible climate coalitions with international financial transfers
It becomes therefore relevant to analyse whether there exist transfer schemes that can stabilise the Grand Coalition or other climate coalitions. First, the coalition surplus could be used to transfer resources to countries with the largest incentives to free-ride. If the coalition surplus is large enough to offset all coalition members' incentives to free ride, then the coalition is Potentially Internally Stable (PIS). Unfortunately, there exist no transfer scheme and no coalition that can be shown to be PIS.
Then, we could quantify the amount of resources that developed countries could transfers to developing countries for the Grand Coalition to be stable. The sum that the industrialised countries should be ready to transfer is about 10% of their global welfare (about 3% of their discounted sum of future GDPs).
Finally, we could explore whether an appropriately designed permit market, with an initial allocation of permits that favours countries with the largest incentive to free ride, can provide incentives to form a stable Grand Coalition. Five rules have been analysed: (i) a grandfathering rule (permits are allocated according to each region's share of world emissions in 2005); (ii) a per-capita rule (each human being receives the same amount of emission permits); (iii) an "ability-to-pay" rule (permits are allocated every year to each individual worldwide in inverse proportion to the gap between this individual's GDP per capita and average world GDP per capita in PPP terms); (iv) a "historical responsibility" rule, that grants allowances to each region in inverse proportion to its contribution to cumulative world CO2 emissions over the period 1900-2004; (v) "no-lose" rule under which the amount of allowances given to non-Annex I regions covers their projected baseline emissions, while Annex I regions set their cap - then allocated across them on a per-capita basis - at whatever level is required to meet the 550 ppm CO2eq target.
For all the five allocation rules just described, our modeling results support the same conclusion: no coalition is stable and no coalition is PIS.
4. Sensitivity analysis
PECs previously described are the union of PECs identified in the four scenarios considered in this report: high damage, low discount rate; high damage, high discount rate; low damage, low discount rate; low damage, high discount rate. Therefore, our analysis of potential effectiveness is robust to different crucial parameter specifications. Results on profitability also hold for the four scenarios.
However, the conclusion about the environmental effectiveness of the Grand Coalition, the only coalition that attains the 550 ppme target in 2100, is limited to the scenario with high damage and low discount rate. In all other cases, no coalition attains the 550 ppme target at the equilibrium. Therefore, only if the discount rate is low and future damages, in particular risks of future catastrophic events, are adequately taken into account, can the Grand Coalition yield an equilibrium emission profile consistent with the 550 ppme target in 2100.
Finally, the conclusion that no coalition is stable and cannot be stabilised by a self-financed transfer scheme (no PIS) holds for the four scenarios and is therefore robust.
Let us stress that all results should be interpreted taking into account
the assumptions behind the adopted theoretical framework. This framework,
based on recent advances of game-theoretic analysis of international agreements,
requires immediate, irreversible and self-enforcing participation to mitigation
action, thereby abstracting from other possible negotiation options, including
e.g. delayed participation, sanctions or joint negotiation in multiple areas
(e.g. linking climate and international trade negotiations). Another important
caveat is that the co-benefits from mitigation action, e.g. in terms of human
health, energy security or biodiversity, are not taken into account, even
though, as recently shown by Finus and Rubbelke, this in unlikely to increase
the chances of global cooperation to be stable. From a technological viewpoint,
even though WITCH can model the dynamics of a large set of technologies, including
carbon capture and storage, and considers deforestation as a mitigation option,
it does not include a technology that can yield extensive negative emissions.
Finally, the analysis focuses on economic and technological incentives only,
thus neglecting the political and social dimension of climate policy decisions.
