Environmental and climate-related decisions often have significant equity consequences. Those consequences are increasingly central to policy debate and can affect the legitimacy and durability of the decisions themselves. Yet the analytical models used to inform such decisions rarely incorporate equity in a transparent or systematic way. Even when equity is included, it is usually reduced to distributional equity, with much less attention given to contextual and procedural equity.
This dissertation addresses these limitations by developing and applying an analytical architecture for integrating equity into quantitative decision-support models. The architecture is applied to develop a phaseout plan for California’s 96 natural gas power plants to meet the state’s environmental goals. The dissertation developed a method for using a retrieval-augmented large language model pipeline to extract and synthesize literature-grounded judgments about indicator importance and equity type, retaining disagreement across models as a source of uncertainty. The resulting model evaluates phaseout priorities under utilitarian, sufficientarian, and egalitarian justice principles, which serve as proxies for stakeholder values and perspectives.
The analysis shows that incorporating equity more fully can produce meaningfully different phaseout rankings than an approach limited to distributional equity. It also shows that equity-informed phaseout rankings can distinguish phaseout decisions that are relatively robust to competing stakeholder values and perspectives from decisions that are much more sensitive to those values. Rankings converge most clearly at the top and bottom of phaseout ranking and diverge most in the middle. This pattern helps identify where stakeholder deliberation is most needed.
