CLAM and Local Effect Estimation

Summary

CLAM focuses on a difficult but common causal inference setting: data are only available at a coarse spatial resolution, while the scientific question is local.

Research Question

How can we infer local causal effects from aggregated or coarse spatial observations without pretending the missing resolution does not matter?

Methodology

• Spatial disaggregation for causal effect estimation
• Coarse-to-local inference
• Emphasis on modeling assumptions that make local inference defensible

Key Contributions

• Tackles a practically important mismatch between data resolution and target estimand
• Connects spatial modeling and causal effect estimation
• Opens a path toward more realistic local-effect analysis in observational settings

Outputs

• Publication: (Mukherjee et al., 2025)