Spatial Panel Event Study
Spatial Panel Event Study Design · Also known as: spatial event study, spatial DiD event study, geo-panel event study, spatial panel ES
Spatial panel event study extends the classical panel event-study design to settings where units are geographically located and outcomes may spill over across space. By combining event-time indicators with spatial weights matrices, it estimates dynamic treatment effects while explicitly accounting for spatial autocorrelation, geographic spillovers, and cross-unit contamination that would bias conventional event studies.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use spatial panel event study when your panel units are geo-referenced, you suspect treatment effects may spill into neighboring units, and you want to trace how effects evolve before and after an event across multiple periods. It is appropriate for county- or city-level policy evaluations, environmental shock analyses, or any setting with potential geographic spillovers and staggered or simultaneous treatment. It is not appropriate when units are not geographically or spatially defined, when only a single post-period is available (event-time variation is required), or when the spatial weights structure is theoretically ambiguous and researchers face excessive researcher degrees of freedom.
Strengths & limitations
- Explicitly models spatial spillovers, preventing contamination of the control group from leakage of treatment effects across borders.
- Produces dynamic treatment-effect profiles (event-time plots) that reveal pre-trends and the trajectory of impact over time.
- Combines the identification rigor of panel fixed effects with the geographic realism of spatial econometrics.
- Conley standard errors and spatial cluster inference properly account for cross-sectional dependence, yielding valid confidence bands.
- Can separately estimate direct (own-unit) and indirect (neighbor-unit) causal effects, enriching policy interpretation.
- The choice of spatial weights matrix (contiguity, distance, k-nearest) is subjective and can influence estimates; results should be checked for sensitivity across plausible alternatives.
- Requires sufficiently many treated and control units spread across space — sparse geographic coverage weakens the spillover analysis and inflates variance.
- Staggered adoption adds complexity; heterogeneous-robust estimators (e.g., Callaway-Sant'Anna with spatial adjustment) are needed to avoid negative weighting bias.
- Data requirements are substantial: geo-referenced panel with pre-event periods long enough to test parallel trends.
- Computational burden rises with large spatial panels, especially when inverting large spatial lag matrices.
Frequently asked
How does this differ from a standard panel event study?
A standard panel event study assumes units are independent and the control group is clean. Spatial panel event study relaxes both: it models geographic spillovers so that treatment effects bleeding into neighbors can be estimated separately, and adjusts inference for spatial autocorrelation in errors.
Which spatial weights matrix should I use?
Theory should guide the choice. Use contiguity if effects spread via shared borders (e.g., labor markets), inverse distance if attenuation with distance is expected, or k-nearest neighbors for robustness checks. Always report sensitivity of key estimates to alternative weight specifications.
What if treatment timing is staggered across units?
With staggered adoption, use heterogeneous-robust estimators such as Callaway and Sant'Anna or Sun and Abraham, adapted to account for spatial correlation. Standard TWFE event-study regressions can yield negative-weighted averages of treatment effects in heterogeneous settings.
How do I test for spillovers?
Include spatially lagged treatment indicators (W times D for each event-time bin) alongside own-unit indicators. If the spatially lagged coefficients are jointly significant, spillovers are present. Plotting their event-time trajectory shows whether spillovers build over time or decay.
What standard errors should I use?
Conley standard errors that account for spatial autocorrelation up to a distance cutoff are the standard choice. Spatial block clustering (e.g., by region or commuting zone) is a pragmatic alternative; using OLS standard errors or unit-level clustering is generally inappropriate when spatial dependence is present.
Sources
- Sun, L., & Callaway, B. (2021). Difference-in-differences estimators of intertemporal treatment effects. arXiv:2109.10157. link ↗
- Gibbons, C. E., Serrato, J. C. S., & Urbancic, M. B. (2019). Broken or Fixed Effects? Journal of Econometric Methods, 8(1), 20170002. DOI: 10.1515/jem-2017-0002 ↗
How to cite this page
ScholarGate. (2026, June 3). Spatial Panel Event Study Design. ScholarGate. https://scholargate.app/en/causal-inference/spatial-panel-event-study
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Difference-in-DifferencesEconometrics↔ compare
- Panel Event StudyCausal inference↔ compare
- Spatial Difference-in-DifferencesCausal inference↔ compare
- Spatial Regression Discontinuity DesignCausal inference↔ compare