Spatiotemporal Causal Interference Modeling for Cross-Border Policy Spillovers

While Oprescu et al. (2025) pioneered spatiotemporal causal inference with GST-UNet, its application to political science remains underexplored. This research proposes adapting their framework to quantify cross-border policy spillovers—e.g., how a state's voting law changes influence neighboring states' electoral outcomes. Unlike existing DiD approaches (Wang et al., 2024) that assume unit independence, we model interference using weighted networks where edge weights reflect institutional similarity (e.g., shared party control, demographic alignment). By integrating political covariates into GST-UNet’s G-computation backbone, we address Tortú et al.’s (2020) call for interference-aware methods in multi-level governance. This could resolve conflicting findings in policy diffusion studies (e.g., why some policies spread while others stall) by distinguishing causal spillovers from parallel trends.

References:

1. Advances in Difference-in-differences Methods for Policy Evaluation Research. Guangyi Wang, Rita Hamad, Justin S. White (2024). Epidemiology.
2. GST-UNet: Spatiotemporal Causal Inference with Time-Varying Confounders. M. Oprescu, D. K. Park, Xihaier Luo, Shinjae Yoo, Nathan Kallus (2025). arXiv.org.
3. Modelling Network Interference with Multi-valued Treatments: the Causal Effect of Immigration Policy on Crime Rates. C. Tortú, L. Forastiere, I. Crimaldi, F. Mealli (2020).