Modeling justice without deciding it

Integrated assessment models shape climate policy by estimating mitigation costs and tracing pathways with different consequences across countries, income groups, and generations. As equity and justice become more prominent in this work, a basic question follows: what gives a model-based ranking normative authority? Social-choice theory shows why technique alone cannot answer it. Arrow's theorem rules out a neutral procedure for converting individual rankings into a collective ranking under a standard set of fairness conditions. Later work shows that broader forms of aggregation require substantive choices about what counts as welfare, how welfare is compared across people and generations, and how competing claims are combined. Where those choices enter depends on model architecture. Welfare-based models place them in welfare functions, discounting, and aggregation. Process-based models may introduce them through optimization objectives, allocation rules, scenario constraints, distributional metrics, or interpretation. Modeling and effort-sharing studies already vary many of these normative choices, but further closure occurs when alternative principles or criteria are selected, combined, or used to rank pathways. This paper proposes a two-track disclosure framework that separates choices embedded in an analysis from those introduced when its outputs are interpreted. It makes the basis of policy rankings visible and distinguishes conclusions that hold across ethical positions from those that depend on particular premises.