Integrated Assessment Models (IAMs) simulate climate change scenarios and substantially contribute to influential reports by supra-governmental institutions (e.g., IPCC). Current models rely on hypothetical policy pathways that largely disregard public preferences, limiting the adequacy of IAM scenarios. Ambitious measures often lack public support and are therefore unlikely to be implemented, resulting in some modeled scenarios being unrealistically optimistic.
We address this shortcoming by linking insights from discrete choice experiments on demand-side policy acceptability with a U.S. focused IAM (GCAM-USA). We assessed acceptability of carbon tax schemes and nine policies across the food, transportation, and household energy domain. Policies were evaluated in isolation and in packages to capture policy bundling effects. Using a large-scale U.S. quota-sample (total N = 5,463 across all U.S. states), we further examined the role of individual- and context-level predictors, applying multilevel regression and poststratification to estimate acceptability at federal and state level.
Results reveal that acceptability declines with more coercive push measures and with less coercive pull measures. Consistent with bundling effects, the effect of policy attributes decreased for policy packages compared to isolated policies. Acceptability was generally low, with few policy packages surpassing 50% (food: 2 of 1,312; transportation: 18 of 1,968; household energy: 8 of 1,312). At the individual level, governmental trust, environmental attitudes, and political liberalism robustly predicted acceptability. At the contextual level, per-capita carbon emissions were negatively and GDP positively associated with acceptability across domains.
Finally, we integrated publicly accepted policy packages into GCAM-USA and demonstrate that realistic scenarios fall short of emission reduction goals and diverge from standard IPCC pathways. While our federal and state-specific scenarios yield similar national emissions, impacts on state-level emissions differ substantially, showing the importance of high-resolution acceptability data. These findings highlight how incorporating behavioral evidence into IAMs can generate more realistic pathways and inform potential levers for policy-makers.