Modified Cost-Risk Analysis as a Bridge Between Target-Based and Trade-Off-Based Decision-Making Frameworks

Abstract

Anthropogenic climate change creates a decision-making problem between near-term mitigation costs and long-term risks of severe impacts. Traditional frameworks such as cost-benefit analysis (CBA) and cost-effectiveness analysis (CEA) represent two ends of a spectrum: the former trades off mitigation costs against expected damages, while the latter seeks the least-cost path to meet climate targets, assuming damages cannot be reliably quantified. Cost-risk analysis (CRA) emerged to reconcile CEA’s temperature objectives with utility-based flexibility; however, improved damage quantification calls for a more integrated approach. This thesis develops Cost-Benefit-Risk Analysis (CBRA)—a novel framework that integrates explicit (partial) damage estimates into a reduced-weight risk function for temperature targets, bridging target-based and trade-off-based methods. The thesis first examines the carbon budget concept, which posits a near-linear relationship between cumulative $\mathrm{CO_2}$ emissions and global temperature rise (TCRE)—a key condition for CRA–CEA equivalence. Scenario-dependent deviations are analyzed using an optimization program, demonstrating maximal deviations of less than 10\% of total temperature rise, which rapidly diminish thereafter. Using FaIRv2 and a one-box model, it is shown that scenario dependence can be explained by the shape of the pulse response, interpreted as Green’s function. This approach generalizes to models of any complexity and helps characterize nonlinearity in the carbon budget. Combined with projections for agriculture and mortality, and implemented in the CGE model GTAP-INT 2, the results reveal substantial regional disparities and significant global damages, particularly in poorer, populous regions—with long-term GDP reductions of up to 50\% for certain countries, while a few specific regions may experience marginal GDP gains (less than 1\%) under high-emission scenarios. Finally, CBRA is implemented within the MIND-L integrated assessment model, incorporating hereby modified FaIR climate module with probabilistic climate sensitivity and a partial damage function derived from CGE-informed economic damages. It confirms that, under a carbon budget-consistent module, CRA and CEA yield equivalent outcomes. Moreover, we show that, by calibrating the CBRA framework to meet the predefined temperature target, one can assess how much of the unknown risks embodied in the climate target are explicitly quantified by the implemented damage function. These findings position CBRA as a consistent, flexible tool for climate policy design.