Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Targeted Synthetic Control Method

Yuxin Wang, Dennis Frauen, Emil Javurek, Konstantin Hess, Yuchen Ma, Stefan Feuerriegel

TL;DR

Targeted Synthetic Control (TSC) addresses bias and instability in single-treated-unit causal inference by combining classical SCM with a TMLE-style one-dimensional targeted update of the synthetic-control weights. In a two-stage procedure, TSC first obtains initial weights and a flexible outcome-regression model, then updates the weights along an exponential-tilting submodel to balance residuals, yielding a final counterfactual $\hat{\psi}^{\textsc{tsc}}_{\tilde{t}}=\sum_{j=2}^{N} \hat{w}^\star_j Y_{j\tilde{t}}$ that remains a convex combination of observed control outcomes. This preserves interpretability and ensures bounded predictions, while enabling the use of arbitrary ML models for nuisance components. Across extensive synthetic and real-world experiments, TSC consistently improves estimation accuracy over state-of-the-art SCM baselines and provides robust, bounded counterfactuals suitable for policy analysis and causal inference in panel data with a single treated unit.

Abstract

The synthetic control method (SCM) estimates causal effects in panel data with a single-treated unit by constructing a counterfactual outcome as a weighted combination of untreated control units that matches the pre-treatment trajectory. In this paper, we introduce the targeted synthetic control (TSC) method, a new two-stage estimator that directly estimates the counterfactual outcome. Specifically, our TSC method (1) yields a targeted debiasing estimator, in the sense that the targeted updating refines the initial weights to produce more stable weights; and (2) ensures that the final counterfactual estimation is a convex combination of observed control outcomes to enable direct interpretation of the synthetic control weights. TSC is flexible and can be instantiated with arbitrary machine learning models. Methodologically, TSC starts from an initial set of synthetic-control weights via a one-dimensional targeted update through the weight-tilting submodel, which calibrates the weights to reduce bias of weights estimation arising from pre-treatment fit. Furthermore, TSC avoids key shortcomings of existing methods (e.g., the augmented SCM), which can produce unbounded counterfactual estimates. Across extensive synthetic and real-world experiments, TSC consistently improves estimation accuracy over state-of-the-art SCM baselines.

Targeted Synthetic Control Method

TL;DR

Targeted Synthetic Control (TSC) addresses bias and instability in single-treated-unit causal inference by combining classical SCM with a TMLE-style one-dimensional targeted update of the synthetic-control weights. In a two-stage procedure, TSC first obtains initial weights and a flexible outcome-regression model, then updates the weights along an exponential-tilting submodel to balance residuals, yielding a final counterfactual that remains a convex combination of observed control outcomes. This preserves interpretability and ensures bounded predictions, while enabling the use of arbitrary ML models for nuisance components. Across extensive synthetic and real-world experiments, TSC consistently improves estimation accuracy over state-of-the-art SCM baselines and provides robust, bounded counterfactuals suitable for policy analysis and causal inference in panel data with a single treated unit.

Abstract

The synthetic control method (SCM) estimates causal effects in panel data with a single-treated unit by constructing a counterfactual outcome as a weighted combination of untreated control units that matches the pre-treatment trajectory. In this paper, we introduce the targeted synthetic control (TSC) method, a new two-stage estimator that directly estimates the counterfactual outcome. Specifically, our TSC method (1) yields a targeted debiasing estimator, in the sense that the targeted updating refines the initial weights to produce more stable weights; and (2) ensures that the final counterfactual estimation is a convex combination of observed control outcomes to enable direct interpretation of the synthetic control weights. TSC is flexible and can be instantiated with arbitrary machine learning models. Methodologically, TSC starts from an initial set of synthetic-control weights via a one-dimensional targeted update through the weight-tilting submodel, which calibrates the weights to reduce bias of weights estimation arising from pre-treatment fit. Furthermore, TSC avoids key shortcomings of existing methods (e.g., the augmented SCM), which can produce unbounded counterfactual estimates. Across extensive synthetic and real-world experiments, TSC consistently improves estimation accuracy over state-of-the-art SCM baselines.
Paper Structure (22 sections, 1 theorem, 34 equations, 4 figures, 4 tables, 2 algorithms)

This paper contains 22 sections, 1 theorem, 34 equations, 4 figures, 4 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related work
  3. Preliminaries
  4. Setup
  5. Background on existing methods
  6. Targeted synthetic control method
  7. Overview of TSC
  8. Intuition behind TSC
  9. Derivation of the targeted update
  10. Properties of TSC
  11. Differences to augmented SCM
  12. Experiments
  13. Synthetic data
  14. Real-world data
  15. Extended related work
  16. ...and 7 more sections

Key Result

Theorem 4.1

Assume $Y_{j{\tilde{t}}}\in[a, b]$, where $a,b \in \mathbb{R}$, $a<b$, for all control units $j=2,\ldots,N$ and the targeted weights satisfy Then our TSC is bounded, $a \le \hat{\psi}_{{\tilde{t}}}^{\textsc{tsc}} \le b$.

Figures (4)

  • Figure 1: Synthetic control setting. Panel data with outcomes for single-treated unit (red) and multiple control units (green). Treatment is assigned at $T_0$, so that the pre-treatment history of the treated unit is used to learn synthetic-control weights based on other pre-treatment histories of control units (gray), while the post-treatment period is used to estimate the treatment effect between the treated unit and the synthetic control (dark arrow).
  • Figure 2: Overview of TSC. Our method consists of two stages: a nuisance component estimation and a targeted update for debiasing. In the nuisance stage, we (i) compute initial synthetic-control weights $\hat{w}$ by matching pre-treatment histories and (ii) fit an outcome-regression $\hat{m}_{\tilde{t}}(x)$ on control units using any machine-learning backbone. In the debiasing stage, we construct residual scores $S_j$ of control units and perform a one-dimensional exponential-tilting update $\tilde{w}(\varepsilon)$. The parameter $\hat{\varepsilon}$ is obtained by solving the condition $f(\hat{\varepsilon})=0$. The resulting targeted weights $\hat{w}^\star=\tilde{w}(\hat{\varepsilon})$ yield the final counterfactual $\hat{\psi}_{\tilde{t}}=\sum_{j=2}^N \hat{w}_j^\star Y_{j\tilde{t}}$ (and thus the treatment effect) while preserving interpretable synthetic-control weights.
  • Figure 3: Insights of weights. Bars show the weights of control units used to construct the counterfactual for the treated unit (the first row for continuous outcomes and the second row for binary outcomes). Shown are the initial synthetic control weights produced by TSC (light colors) and the final weights (dark colors).
  • Figure 4: Turnout data. The trajectories of the control units are shown in gray, and the treated unit in red. We plot counterfactual estimates from augmented SCM (blue), classical SCM (orange), the plug-in estimator (purple), and TSC (green). The vertical dashed line indicates the treatment time $T_0$. TSC matches the pre-treatment history closely and suggests a decline in turnout rate after treatment.

Theorems & Definitions (2)

  • Theorem 4.1: Boundedness
  • proof