Correlated Synthetic Controls (CSC)

Updated 6 July 2026

Correlated Synthetic Controls (CSC) is a synthetic-control estimator for settings with many treated units and short panels, using correlated donor weights to harness observable similarity.
The method employs a correlated random coefficients structure on donor weights, balancing the overfitting risk of separate controls and the rigidity of pooled estimators.
Simulation studies and empirical applications, such as the Mariel Boatlift analysis, reveal that CSC outperforms standard DiD and pooled approaches under treatment assignment correlated with unobservables.

Searching arXiv for papers on Correlated Synthetic Controls and closely related synthetic control methodology. Correlated Synthetic Controls (CSC) is a synthetic-control-type estimator for settings with many treated units and short panels. Rather than estimating a fully separate synthetic control for each treated unit or imposing one common pooled synthetic control for all treated units, CSC models donor weights as a correlated random coefficients structure that varies systematically with treated-unit observables, so that treated individuals with similar observables receive similar synthetic controls (Moev, 11 Jul 2025). The estimator is designed for panels with many treated units, many donor units, a single treatment date, and treatment assignment correlated with unobservables; within that design, the method is positioned as an alternative to difference-in-differences (DiD) when parallel-trends-style logic is fragile (Moev, 11 Jul 2025).

1. Formal setup and estimand

The CSC framework in (Moev, 11 Jul 2025) considers a panel with total units $N$ , treated units $n_1$ , donor units $n_0=N-n_1$ , total periods $T$ , and pre-treatment periods $T_0$ . Treatment begins at $T_0+1$ , so the setup is non-staggered adoption with a common treatment date. Potential outcomes are denoted $y_{it}(0)$ and $y_{it}(1)$ , with individual treatment effects

$\tau_{it}=y_{it}(1)-y_{it}(0),$

and the average treatment effect on the treated at time $t$ ,

$n_1$ 0

The observed outcome matrix is partitioned into donor and treated blocks, pre-treatment and post-treatment blocks. The inferential target is the missing untreated post-treatment block for treated units,

$n_1$ 1

from which treatment effects are obtained by comparing observed treated outcomes to the predicted untreated outcomes (Moev, 11 Jul 2025).

The paper emphasizes a specific empirical regime: many treated units, many donors, short pre-treatment histories, treatment assignment correlated with unobservables, and time-invariant discrete covariates used to structure similarity across treated units. It explicitly notes that CSC currently does not allow continuous covariates in the weight structure, although continuous variables may be discretized or preprocessed (Moev, 11 Jul 2025).

2. Correlated random coefficients and the CSC estimator

The defining feature of CSC is that donor weights are not unit-specific free parameters and are not common across all treated units. Instead, they are parameterized as

$n_1$ 2

subject to the synthetic-control constraints

$n_1$ 3

Here $n_1$ 4 is an individual fixed effect, $n_1$ 5 is the donor-specific invariant component of the weight, $n_1$ 6 is a row vector of treated-unit covariates, and $n_1$ 7 are donor-specific coefficients on those covariates (Moev, 11 Jul 2025).

This structure implies that treated units with identical observables receive identical donor weights, while treated units with similar observables receive similar donor weights. The paper’s intuition is that CSC “creates synthetic controls that are correlated across individuals with similar observables” (Moev, 11 Jul 2025). It therefore occupies an intermediate position between two extremes. Estimating separate SCs for every treated unit can generate multiplicity of solutions and overfitting when $n_1$ 8 is short, while a pooled synthetic control can be too restrictive and miss heterogeneity. CSC is intended to balance those two problems by allowing heterogeneity that is disciplined by observables (Moev, 11 Jul 2025).

The paper gives the estimation problem as a constrained quadratic optimization: $n_1$ 9 subject to

$n_0=N-n_1$ 0

and

$n_0=N-n_1$ 1

A matrix-form expression is also given, and implementation is reported in R using CVXR (Moev, 11 Jul 2025).

A simple illustration in the paper uses small and big cities. If treated units are partitioned by observables into “small” and “big,” CSC can impose

$n_0=N-n_1$ 2

In that case all small cities share one donor-weight pattern and all big cities share another. This suggests that CSC uses observables not merely as matching covariates but as a structure on the weight map itself (Moev, 11 Jul 2025).

3. Identification logic and theoretical properties

The theoretical analysis in (Moev, 11 Jul 2025) is conducted under an interactive fixed effects data-generating process,

$n_0=N-n_1$ 3

where $n_0=N-n_1$ 4 are time-invariant covariates, $n_0=N-n_1$ 5 are time-varying coefficients on observables, $n_0=N-n_1$ 6 are common factors, $n_0=N-n_1$ 7 are factor loadings, and $n_0=N-n_1$ 8 is an idiosyncratic shock. The paper states the treatment-not-at-random condition

$n_0=N-n_1$ 9

along with $T$ 0, $T$ 1, independence of covariates from $T$ 2 and $T$ 3, $T$ 4, stochastic $T$ 5, fixed $T$ 6, and a single post-treatment period $T$ 7 (Moev, 11 Jul 2025).

The core identification device is an Exact Fit assumption. There exists a matrix $T$ 8 with weights

$T$ 9

whose entries are nonnegative and whose rows sum to one, such that for all pre-treatment periods

$T_0$ 0

and for covariates

$T_0$ 1

The paper remarks that this condition is stronger than the original Abadie-style condition because it effectively rules out multiplicity of solutions by imposing a unique matching structure (Moev, 11 Jul 2025).

Under the interactive fixed effects DGP, Exact Fit, and invertibility of $T_0$ 2, Lemma 1 generalizes Abadie’s representation of synthetic-control error to the many-treated-unit setting. The decomposition shows that CSC estimation error is driven by pre-treatment and post-treatment idiosyncratic shocks transformed through the factor structure (Moev, 11 Jul 2025). Building on that decomposition, the paper proves a high-probability upper bound for the CSC estimation error under iid subGaussian $T_0$ 3 errors, invertibility of $T_0$ 4, and Euclidean geometry. The bound depends on $T_0$ 5, factor dimension $T_0$ 6, idiosyncratic noise $T_0$ 7, and the extrema of the eigenvalues of

$T_0$ 8

The paper interprets the result as implying that larger $T_0$ 9 improves CSC, while larger $T_0+1$ 0 and larger $T_0+1$ 1 worsen the conservative error bound (Moev, 11 Jul 2025).

The same paper derives an asymptotic expression for feasible DiD under the same DGP: $T_0+1$ 2 This is presented as an explicit asymptotic bias term. If treated and donor units differ in unobservables $T_0+1$ 3, DiD is asymptotically biased. By contrast, CSC’s theoretical error characterization does not depend on the covariance between treatment and unobservables in the same way, because CSC can reweight donors and discard irrelevant controls (Moev, 11 Jul 2025).

For benchmarking, the paper also defines infeasible DiD (iDiD) by removing the interactive fixed effects using unobserved terms: $T_0+1$ 4 and proves consistency under the interactive fixed effects DGP: $T_0+1$ 5 This serves as a gold-standard comparison in the simulation study (Moev, 11 Jul 2025).

4. Relation to synthetic control, DiD, and adjacent research

CSC is best understood against the background of standard single-treated-unit synthetic control. In (Moev, 11 Jul 2025), standard SC is written as

$T_0+1$ 6

with weights chosen by

$T_0+1$ 7

That formulation is inherently a single treated unit method. CSC generalizes the weight system to many treated units through the observable-indexed structure $T_0+1$ 8 (Moev, 11 Jul 2025).

Theoretical work on standard SC with many periods and many controls helps clarify what CSC inherits and what it changes. “On the Properties of the Synthetic Control Estimator with Many Periods and Many Controls” shows that, under a linear factor model, SC can remain asymptotically unbiased if the treated unit’s factor loadings can be reconstructed by a convex combination of control loadings with diluted weights, formalized by

$T_0+1$ 9

(Ferman, 2019). That result concerns the geometry of the donor pool for a single treated unit; CSC instead imposes structure across many treated units by correlating their synthetic controls through observables.

Inference is another adjacent issue. “Bayesian and Frequentist Inference for Synthetic Controls” characterizes when the population risk-minimizing linear predictor lies in the simplex and develops a Bayesian alternative whose posterior predictive distribution becomes asymptotically equivalent to the frequentist sampling distribution in total variation under a Bernstein–von Mises-style result (Martinez et al., 2022). CSC, by contrast, focuses on multi-treated short panels and emphasizes prediction under endogenous treatment assignment rather than a full inferential theory.

A separate but related direction appears in “Adaptive Experiment Design with Synthetic Controls,” which studies exploratory clinical trials over many subpopulations. That paper states that different subpopulations should not be treated as statistically isolated and uses synthetic controls that combine control samples from other subpopulations; it explicitly describes this as the “correlated synthetic controls” intuition in an adaptive trial design (Hüyük et al., 2024). The shared theme is information borrowing across correlated units or subpopulations, although the design problem there is adaptive experimentation rather than panel counterfactual imputation.

Another distinct use of “correlated” arises in “Distributionally Robust Synthetic Control: Ensuring Robustness Against Highly Correlated Controls and Weight Shifts,” which addresses highly correlated donors and post-treatment weight drift by replacing point identification of post-treatment weights with an uncertainty class

$y_{it}(0)$ 0

and targeting a conservative estimand $y_{it}(0)$ 1 (Koo et al., 4 Nov 2025). That paper treats donor correlation as an instability problem in the weight solution; CSC in (Moev, 11 Jul 2025) uses “correlated” to describe synthetic controls that move together across treated units with similar observables.

5. Simulation evidence and the Mariel Boatlift application

The simulation study in (Moev, 11 Jul 2025) uses the interactive fixed effects model

$y_{it}(0)$ 2

with treatment assignment generated by

$y_{it}(0)$ 3

When $y_{it}(0)$ 4, treatment assignment is correlated with unobservables. The paper compares CSC, feasible DiD (fDiD), PSC, and infeasible DiD (iDiD) (Moev, 11 Jul 2025).

The reported simulation findings are consistent with the theory. iDiD dominates, as expected. Among feasible estimators, CSC usually outperforms fDiD and PSC when treatment is nonrandom. Under random assignment, fDiD can do better relative to CSC. Increasing $y_{it}(0)$ 5 improves all methods substantially, while changes in $y_{it}(0)$ 6 and treatment share have limited empirical effect in the reported designs. Representative baseline average estimation errors are reported as CSC $y_{it}(0)$ 7, fDiD $y_{it}(0)$ 8, PSC $y_{it}(0)$ 9, and iDiD $y_{it}(1)$ 0, and RMSE results likewise favor CSC over fDiD and PSC in nonrandom-treatment scenarios (Moev, 11 Jul 2025).

The empirical application revisits the 1980 Mariel Boatlift using PSID rather than CPS. The sample uses PSID waves 1974–1984, male heads of households, working-age individuals, and Florida residents as the treated group, with U.S. residents outside Florida as donors. Outcomes are total hours worked per year and hourly wages transformed as $y_{it}(1)$ 1. Covariates include race, education, marriage status, industry dummies, occupation dummies, age restriction, and illness indicator (Moev, 11 Jul 2025).

For prediction, the paper compares CSC and PSC using cross-validation with pre-treatment years 1975–1979, post-treatment years 1980–1984, testing years 1978–1979, and varying training window length $y_{it}(1)$ 2. CSC performs slightly better than PSC for several training lengths. The best wage prediction occurs with $y_{it}(1)$ 3 and CSC, and the best hours-worked prediction occurs with $y_{it}(1)$ 4 and CSC (Moev, 11 Jul 2025).

The heterogeneity analysis divides workers into low-skilled and high-skilled groups, with low-skilled defined as workers without a college degree. The reported findings are no meaningful labor supply effects for either group, negative wage effects for low-skilled workers, and no wage effects for high-skilled workers. The paper interprets this as consistent with stronger competition between Mariel immigrants and low-skilled Florida workers, while also stressing that sample sizes are very small and the confidence intervals do not account for uncertainty in estimated synthetic-control weights (Moev, 11 Jul 2025).

6. Limitations and methodological significance

The paper identifies several practical limitations. CSC naturally supports only discrete time-invariant covariates in the main weight structure. Continuous covariates require discretization or preprocessing. The method does not use donor covariates in the main specification. Inference remains difficult because uncertainty arises both from estimated weights and from treatment effects conditional on those weights, and the application-level confidence intervals only partially account for that uncertainty (Moev, 11 Jul 2025).

These limitations clarify CSC’s methodological role. The estimator is not a generic replacement for DiD or for single-unit SC. It is a structured prediction method for multi-treated short panels in which treatment assignment may be correlated with latent heterogeneity. A plausible implication is that CSC is most attractive when separate SCs would overfit because $y_{it}(1)$ 5 is small, while pooled estimators or DiD would be too rigid because treated units differ systematically in observables and unobservables. The method’s significance lies in turning heterogeneity across treated units into a source of regularization: donor weights vary, but only through a constrained observable-indexed map.

In that sense, CSC extends the synthetic-control logic from “one treated unit matched to many donors” to “many treated units whose synthetic controls are linked through observables.” The resulting object is neither a collection of unrelated unit-specific SCs nor a single global counterfactual. It is a many-treated synthetic-control estimator in which cross-treated-unit dependence is built directly into the weighting scheme (Moev, 11 Jul 2025).