EvalCompassEvalCompass
Causal inference & experimentsCausal inferenceComparative casesPolicy evaluationWeighted synthesisFew units

Synthetic Control Method

Synthetic Control Method
Also known as: Synthetic comparison group method、SCM、Synthetic Control

Weights several unexposed areas or organizations to create a comparator resembling the treated unit's past trends. Pre-fit and donor suitability matter. [1]

Synthetic control design worksheet — openReferences and checked sections
Difficulty
Advanced
Cost
Medium
Duration
Illustrative planning durations, not common standards: collection 1–4 weeks, analysis/checks 2–4 weeks
Content updated: 2026-09-13Source checked: 2026-09-13 (includes AI cross-checking)No documented human reviewIncludes fictional examples

How it works at a glance

The basic form uses non-negative donor weights summing to 1 to match pre-outcomes/predictors, then estimates post-differences. [1]

Do not choose weights from post-results. Check spillovers, poor pre-fit and donor dependence, using placebos to examine counterfactual uncertainty. [1]

  1. Define treated units and donor pool
  2. Select outcomes and covariates
  3. Construct synthetic controls
  4. Check pre-fit
  5. Estimate and plot effects
  6. Placebo-based inference
Sequence of implementation steps. See “Implementation steps” for details of each stage.

When to use it and when it is unsuitable

When to use it

  • Evaluate policies affecting one/few areas, countries or firms, such as city industrial policy or national regulation
  • Multiple similar untreated units and long pre-panels available
  • No single parallel-trend DiD comparator but combinations improve pre-fit
  • Present academically/politically persuasive causal policy evidence

When it is unsuitable

  • Insufficient series for pre-comparison or fit assessment
  • Insufficient comparable unaffected donors for synthesis/placebos
  • Donor spillovers undermine counterfactual interpretation

Participants, roles, and deliverables

Participants and roles

1–2 econometric/statistical specialists with skills such as R Synth

Main deliverables

  • Path plot: treated actual versus synthetic series
  • Donor weight list
  • Estimated post-gaps/effect magnitude
  • Placebo plots and MSPE ratios

Implementation steps

  1. Define treated units and donor pool
    Identify one/few treated units and comparable untreated donors. Restrict donors to similar unaffected units.
  2. Select outcomes and covariates
    Choose outcomes and predictors, including covariates and multiple pre-outcomes, for weighting. Fix choices beforehand on theoretical grounds.
  3. Construct synthetic controls
    Find non-negative weights summing to 1 reproducing pre-outcomes/predictors. Distinguish predictor and donor weights and report criteria and pre-prediction errors.
  4. Check pre-fit
    Show observed/synthetic trends, residuals and MSPE. Report counterfactual limits if fit is poor; do not adjust to post-results.
  5. Estimate and plot effects
    Calculate period-specific observed/synthetic gaps. Examine causal interpretation conditions, pre-fit, concurrent shocks and spillovers.
  6. Placebo-based inference
    Treat donors as hypothetical interventions and examine gaps or post/pre MSPE ratios. Explain pre-fit differences, comparison counts and assignment interpretation rather than relying on formal p-values alone.

Strengths and limitations

Strengths

  • Builds candidate counterfactuals from multiple comparators for few treated units
  • Published weights/rationale make comparator composition inspectable
  • Quantifies pre-fit, but good fit does not prove post-counterfactual validity
  • Placebos enable inference where conventional test statistics are unavailable

Limitations

  • Poor pre-fit undermines counterfactual credibility and gap interpretation
  • Small donor pools weaken placebo-test power and make statistical inference unstable.
  • Multiple units or staggered designs require different estimands/methods from this basic form
  • Limited methods for donor spillovers

Common pitfalls

  • Affected donors compromise counterfactuals. State selection grounds and exclude potentially exposed units.
  • Always optimize on pre-data only; using post-variables makes estimation circular.
  • Large gaps alone are not strong evidence. Consider placebo fit differences and donor counts; they do not remove confounding.
  • Large MSPE suggests poor counterfactual representation; review variables/donors before reporting effects.

Conditions of use and evidence

  • Uncertainty about whether combinations reproduce counterfactuals is not automatically resolved by aggregate data. [1]
Fictional example; not actual results

Evaluate a consultation scheme introduced in one city

A fictional calculation combining Cities B and C at 0.5 each. The short table illustrates weighting, not sufficient duration for valid research.

TimepointCity ACity BCity CSynthetic
Pre 11081210
Pre 212101412
Pre 315121614
Post 120141816

Common errors in judgment

Common mistake

Optimize weights using post-results

How to revise

Set weights from pre-data

ReasonDo not use the target post-gap to select comparators.

Common mistake

Call a large post-gap an effect despite poor pre-fit

How to revise

State poor fit and reconsider design

ReasonThe counterfactual may already be poorly approximated before intervention.

Organize your thinking by writing

Synthetic control design worksheet

A practical template is available to fill in and print on A4 paper.

Open worksheet

Related methods

Difference-in-DifferencesInterrupted Time Series AnalysisRegression Discontinuity DesignRandomized Controlled TrialContribution Analysis

References and checked sections

  1. Synthetic Control Methods for Comparative Case Studies: Estimating the Effect of California’s Tobacco Control Program
    Alberto Abadie, Alexis Diamond, and Jens Hainmueller · 2010
    Section consulted
    1. Introduction / Synthetic control methods
    How it relates to the text
    Uncertainty about whether combinations reproduce counterfactuals is not automatically resolved by aggregate data.
    Source checked: (includes AI cross-checking)

Editorial and checking information

Content updated: Source checked: (includes AI cross-checking)No documented human reviewIncludes fictional examplesEditorial policy and scope of checkingCite this pageReport a correction

How to cite this page

EvalCompass Editorial Team. “Synthetic Control Method”. EvalCompass, updated 2026-09-13. https://evalcompass.jun-nakatani.com/en/methods/synthetic-control/
Add the date you accessed the page. Text created by this site is available under CC BY 4.0. Quotations, external materials, and bundled fonts have their own terms of use. When reusing content, credit the source, link to the license, and indicate any changes.

Suggest a correction or improvement to this page

Change history
  • — Checked sources for definitions and conditions and added fictional examples, judgment rationales and completed entries for every field. Revised unsupported historical or uptake claims. Reviewed procedures and cautions, adjusting fixed counts, thresholds and institutional generalizations. Preserved field positions, types and counts while revising fixed conditions and thresholds.