Research

Working Papers

U.S. Electricity Interconnections: Technology, Potential, and Investment 

Job Market Paper

AI-integrated models for assessing agricultural resilience

with Joshua R. Waite, Dana Golden, Kevin Camp, Mojdeh Saadati, Shannon Regan, Pat Schnable, Baskar Ganapathysubramanian, Carlos D. Messina, Suzanne Thornsbury, Soumik Sarkar (Under Review)

Agricultural supply chains are vulnerable to disruptions through linked biophysical and economic systems. We develop a novel AI-powered tool that integrates economic models (GTAP) with biophysical models (APSIM) to analyze supply chain shocks, enabling policymakers and market participants to assess cross-disciplinary impacts through queries and responses written in natural language. We present a conceptual vulnerability framework that categorizes the agricultural supply chain as a social-ecological system with biophysical and economic exposures and a working prototype that couples domain models through a lightweight AI orchestration layer. Our framework is an important first step in orchestrating existing domain models to illustrate the power of model coordination. An empirical example illustrates both the power of the tool and remaining gaps.

Agentic AI orchestration of heterogeneous economic models for rapid, multi-scenario analysis of energy crises

with Dana Golden, Lav R. Varshney, Suzanne Thornsbury 

Rigorous economic models can take months to construct, yet energy crises demand decisions from policymakers within days or even hours. Any disruption in energy markets is not isolated but rapidly disseminates through interlinked global systems. Off-the-shelf models that already exist typically focus only on limited aspects and are distributed across research groups, programming languages and incompatible formats. Integrating these models manually can take longer than the crisis itself, forcing analysts to rely on whichever models are easiest to connect and leaving consequential scenarios unexplored. Policymakers must make rapid decisions with obstructed and limited information. We show that large language models can perform the critical integration directly, revealing connections in a way that is timely and computationally efficient. The system constructs internally consistent scenarios, translates assumptions into model-specific inputs, executes existing economic and physical models in dependency order, and synthesizes outputs for policymakers. The language model generates no quantitative results: every reported value is reproduced directly from an underlying model run, remains traceable to its source and is subject to analyst approval at each stage.

U.S. Clean Electricity Generators: Census Tracts and Income, Poverty, Population Dynamics