A well-timed shift from local to global agreements accelerates climate change mitigation

The paper addresses how to accelerate international cooperation on climate mitigation given that the atmosphere’s limited capacity to absorb greenhouse gases is a global public good subject to free-riding. While rapid global mitigation is essential to avoid severe climate impacts and irreversible ecological losses, countries face incentives to delay mitigation due to high near-term costs, competing investments with faster payoffs, expectations of others developing technologies first, and strategic considerations. Attempts to establish a legally binding global agreement with enforceable sanctions (e.g., tariffs) have historically struggled, as seen in the failure of Kyoto, Copenhagen, and Doha to achieve high participation. More recent bilateral and Paris-style pledge-and-review frameworks have advanced voluntary, lower-cost reductions but may be insufficient to limit warming to ≤2 °C. This motivates exploring policy strategies that can achieve widespread, stronger, enforceable commitments. Building-block approaches emphasize local or regional climate clubs and subgroup-focused sanctions to reduce free-riding and coordination challenges. However, while localized agreements can succeed where global efforts fail, they may prolong the timeline for world-wide mitigation because multiple groups must independently adopt mitigation and have limited leverage over late adopters. The central research question is whether, and under what conditions, combining local and global legally binding agreements—either sequentially or concurrently—can both increase the likelihood and speed of achieving global mitigation.

The study builds on literature regarding public goods, climate clubs, and cooperation under international environmental agreements. Prior work indicates that global cooperation is hampered by free-riding and large group sizes, while clubs and coalition-structured governance can improve cooperation by localizing enforcement and reducing coordination burdens. Game-theoretic and common-pool resource studies suggest localized sanctioning within subgroups can achieve cooperation where global enforcement fails. However, these bottom-up approaches may lead to slower global adoption due to the need for multiple, independent group-level transitions and inadequate influence over laggards. Empirical narratives around the failures of global-scale negotiations (Kyoto, Copenhagen, Doha) and the limited strength of voluntary frameworks (Paris) further motivate enforceable mechanisms. Additional strands include the roles of economies of scale in technology diffusion and industrial development, rivalry among regions creating game-of-chicken dynamics, and inequality in emissions and diplomatic clout influencing negotiation outcomes and thresholds for agreement enactment.

The authors develop a stochastic, multi-group decision model of climate mitigation. Players (countries or blocs) are organized into L groups representing coalitions or negotiation forums (e.g., EU, BRICS, bilateral summits). Groups may overlap and vary in size; base simulations use eight regions with 18 players each. Each player chooses between minimal mitigation (consistent with >2 °C) and stringent mitigation (≤2 °C). Decisions are driven by a combination of exogenous stochastic shocks and payoff-based choice: with probability μ = 0.05, a player’s decision is random (capturing chance events such as disasters or crises), while otherwise the player selects the strategy with higher payoff with sensitivity governed by β. The net cost of stringent mitigation is modeled as c = C − B_local − B_global, where C is the direct mitigation cost, B_local captures group-level co-benefits (e.g., reduced pollution, technological spillovers), and B_global represents global climate benefits. In the base ‘worst-case’ scenario, net costs are high so that payoffs favor non-mitigation absent enforcement. Legally binding agreements can be enacted at either the local (within-group) or global scale when a quorum of mitigators supports them (e.g., 50% threshold for global enactment, p_l = 0.5). Agreements impose sanctions (e.g., fines, tariffs) on non-mitigators, with sanctioning incurring costs to mitigators. The model incorporates both active sanctioning costs and constitutive costs f representing resources expended even when attempts to form agreements fail (e.g., political capital). Strategies compared include: local-only agreements, global-only agreements, a sequential shift from local to global agreements triggered when the global share of mitigators crosses a threshold T (often T = 0.4 in simulations), and simultaneous attempts at both local and global agreements. Extensions examine: (1) economies of scale, where mitigation costs decline as more neighbors or group members adopt mitigation; (2) group rivalry, representing competitive economic dynamics that raise mitigation opportunity costs when rival groups do not mitigate; and (3) inequality in emissions and diplomatic weight, assigning higher influence to large emitters in reaching agreement quorums and, in some scenarios, placing major emitters into smaller groups (e.g., bilateral summits). Performance metrics include the rate to reach specified mitigation levels (e.g., time to 80% global mitigation) and cumulative mitigation achieved by a fixed time horizon (e.g., time step 625; some analyses extend to 1500). Results are summarized over multiple replicate simulations (e.g., 10, 30, or 100 replicates depending on figure). Sensitivity analyses explore mitigation and sanctioning costs and the degree of payoff-based decision-making (β).

- Global-only enforcement attempts generally fail when mitigation is initially rare: simulations show mitigator proportions never reach the global quorum (p_l = 0.5), leaving non-mitigators unsanctioned and global cooperation stalled. - Local agreements can capitalize on stochastic, early adoption within groups to establish mitigation as a local norm. Over time, each group may transition, yielding global mitigation, but the process can be slow. - A timely shift from local to global enforcement once mitigators become sufficiently common dramatically accelerates adoption. After local agreements raise global mitigator share above a threshold (e.g., T ≈ 0.4), switching to global agreements more than doubles the rate at which players commit to mitigation compared with local-only strategies. - The optimal shifting threshold aligns with the global enactment quorum (~50%); delaying the shift squanders the benefits of global enforcement. - The benefit of shifting increases with higher mitigation costs, higher sanctioning costs, and greater payoff sensitivity (β), all of which otherwise slow local-only progress. - Quantitative improvements relative to local-only strategies (time to reach high mitigation and/or total mitigation by a fixed time) increase under key frictions: • Economies of scale: ~19% additional mitigation improvement from shifting. • Group rivalry: ~27% improvement, as global enforcement is crucial to overcome holdouts engaged in strategic competition. • Both economies of scale and rivalry: ~34% improvement. - Under high constitutive sanctioning costs (f), focusing efforts first on local enforcement and then shifting to global outperforms simultaneous attempts; premature global attempts can trap systems at low mitigation for extended periods. When f is low, simultaneous local and global attempts yield similar outcomes to shifting because early global attempts fail and later become redundant as local agreements spread. - Emissions and diplomatic inequality: granting greater weight to large emitters in agreement enactment accelerates both local and global agreements. Global mitigation occurs fastest when major emitters participate in small groups (e.g., bilateral high-emitter summits), unless intense economic rivalry between these players negates the advantage. - Across all examined scenarios, sequentially shifting from local to global (or, when constitutive costs are low, overlapping the two) consistently yields the fastest path to world-wide mitigation compared with pursuing only local or only global agreements.

The findings resolve the initial question by showing that scale-specific enforcement roles can be leveraged sequentially to accelerate climate cooperation. Local agreements are vital early on: they exploit localized stochastic increases in mitigation and ‘lock in’ gains through enforceable norms, protecting mitigation against subsequent adverse shocks. However, local agreements alone struggle to influence late-adopting groups, especially under inter-group rivalry where unilateral mitigation reduces competitiveness. Once a sufficient global share of mitigators exists, shifting to global agreements enables system-wide enforcement, pulling in holdouts and overcoming rivalry and persistent high costs. This multi-scale sequencing speeds the transition relative to local-only or global-only approaches and remains effective under economies of scale, rivalry, and unequal emissions/diplomatic power structures. The results imply that policymakers should prioritize forming and deepening local, legally binding climate clubs to create early momentum and resilience, followed by a concerted push for stringent global agreements once political conditions permit enactment. This approach can complement existing voluntary frameworks, reduce free-riding, and harness regime overlaps to minimize required sanctioning while maximizing adoption speed.

The study contributes a multi-scale governance strategy for climate mitigation: prioritize localized, legally binding agreements to initiate and stabilize early adoption, then shift to stringent global enforcement once mitigators are sufficiently common. This sequencing consistently outperforms local-only, global-only, or strictly simultaneous approaches across a range of frictions (costs, economies of scale, rivalry) and under emissions/diplomatic inequality. Policy implications include focusing near-term efforts on enforceable regional clubs and subgroup tariffs, preparing for a timely transition to a global agreement as support grows, and using overlapping coalitions or regime complexes to broaden coverage. Future research should extend the model to continuous mitigation levels and heterogeneous costs/benefits, explore varied institutional forms of sanctions and incentives, incorporate dynamic timelines for policy reconsideration and implementation, couple with climate system feedbacks and risk perception, and examine uncertainty in quorums and commitments. Such extensions would refine quantitative projections while preserving the qualitative advantage of the shift strategy.

The model abstracts mitigation decisions to a binary choice and assumes a net mitigation cost structure that aggregates global and local co-benefits; it omits heterogeneous costs/benefits across players and dynamic changes in benefits as climate impacts intensify. Institutional details of agreements and sanctions are stylized and uniform across groups, and the model does not explicitly represent the speed of policy implementation, administrative delays, or detailed trade network effects of sanctions and counter-sanctions. Timing of decision updates, learning processes, and feedbacks from evolving climate impacts to policy preferences are simplified. Quantitative results (e.g., thresholds, time scales) may vary with richer representations of behavior, institutions, and climate-economy dynamics, though the qualitative advantage of a timely shift from local to global enforcement is robust across tested scenarios.