Spatial Synthetic Control Method
Spatial Synthetic Control Method for Causal Inference · Also known as: spatial SCM, geographic synthetic control, spatial SC, spatial counterfactual control
The Spatial Synthetic Control Method adapts the classic synthetic control framework to settings where treated and donor units are defined by geographic location. By constructing a weighted combination of spatially proximate or comparable control regions, the method estimates what would have happened to a treated area absent the intervention, while explicitly accounting for geographic spillovers, spatial autocorrelation, and contiguity among units.
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 Synthetic Control when a policy, shock, or intervention targets a specific geographic unit and you have pre-treatment panel data on the outcome for multiple comparable regions. It is especially appropriate when standard difference-in-differences assumptions are implausible due to non-parallel trends and when there are few treated units. Do not use it when: the donor pool is too small (fewer than 10–15 donors), when pre-treatment fit cannot be achieved (synthetic control tracks poorly in pre-period), when spatial spillovers contaminate all available donors, or when the treatment is applied simultaneously to nearly all units so no clean control group exists.
Strengths & limitations
- Constructs an explicit counterfactual that can be visually inspected, making the identifying assumptions transparent rather than implicit.
- Accounts for geographic spillovers by carefully selecting contamination-free donor units, a major advantage over standard DiD in spatial contexts.
- Does not require the parallel-trends assumption; donor weights flexibly match complex pre-treatment trajectories.
- Permutation-based inference is valid with small samples, unlike asymptotic standard errors.
- Provides a time-path of the treatment effect, revealing onset, peak, and fade-out of the intervention's impact.
- Requires a sufficiently large and geographically diverse donor pool to construct a well-fitting synthetic control; sparse regional data undermines the approach.
- Pre-treatment fit quality is unverifiable by a formal test, so a poor fit may go undetected without careful inspection of the pre-period gap.
- Spatial spillover exclusion criteria involve researcher judgment and can introduce selection bias into the donor pool if applied inconsistently.
- Inference via permutation tests loses power when the donor pool is small, and the method has no standard software implementation covering all spatial extensions.
- The method handles a single treated unit most naturally; generalization to many simultaneously treated spatial units requires staggered-adoption extensions.
Frequently asked
How do I decide which donor regions to exclude due to spatial spillovers?
Common practice is to exclude units that are geographically contiguous or within a specified distance buffer of the treated unit, and any units with strong economic or trade linkages that could transmit the treatment effect. The exclusion threshold is a researcher judgment call; sensitivity analyses varying this threshold are strongly recommended.
What makes this 'spatial' compared to the standard synthetic control method?
The spatial extension explicitly considers geographic proximity and spillover contamination when forming the donor pool. Standard synthetic control ignores whether donors are near the treated unit; the spatial version makes geographic isolation a selection criterion and may also account for spatial autocorrelation in the outcome variable during estimation.
Can I use spatial synthetic control with multiple simultaneously treated regions?
With difficulty. The classic approach handles one treated unit. Multiple treated regions require either applying the method separately to each treated unit (with donor pools excluding all treated regions and their neighbors) or using staggered-adoption generalizations, which remain an active area of methodological research.
How do I assess whether the synthetic control is a credible counterfactual?
Inspect the pre-treatment gap: a good synthetic control closely tracks the treated unit before the intervention (low RMSPE). Also check that predictor means are well-matched. If pre-treatment fit is poor, the post-treatment gap cannot be credibly interpreted as a causal effect.
Is there dedicated software for spatial synthetic control?
Standard synthetic control software (Synth in Stata and R, or the tidysynth R package) can be used with a manually curated spatial donor pool. No single package automates the full spatial workflow including spillover detection; researchers typically combine these tools with spatial analysis packages such as spdep or sf in R.
Sources
- Abadie, A., & Gardeazabal, J. (2003). The Economic Costs of Conflict: A Case Study of the Basque Country. American Economic Review, 93(1), 113-132. DOI: 10.1257/000282803321455188 ↗
- Abadie, A., Diamond, A., & Hainmueller, J. (2010). Synthetic Control Methods for Comparative Case Studies: Estimating the Effect of California's Tobacco Control Program. Journal of the American Statistical Association, 105(490), 493-505. DOI: 10.1198/jasa.2009.ap08746 ↗
How to cite this page
ScholarGate. (2026, June 3). Spatial Synthetic Control Method for Causal Inference. ScholarGate. https://scholargate.app/en/causal-inference/spatial-synthetic-control-method
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
- Propensity Score MatchingResearch Statistics↔ compare
- Spatial Difference-in-DifferencesCausal inference↔ compare
- Spatial Interrupted Time SeriesCausal inference↔ compare
- Spatial Regression Discontinuity DesignCausal inference↔ compare
- Synthetic Control MethodCausal inference↔ compare