Symmetries of Counterfactuals: Identification and Testability of Counterfactual Maps from Covariate Heterogeneity
Authors: Marc Braun, Adel Daoud, Jose M. Peña
Organizations: Department of Computer and Information Science, Linköping University · Institute for Analytical Sociology, Linköping University · Chalmers University of Technology
In general, counterfactuals are not identified from data alone, because many causal models produce the same observed distributions but different counterfactuals. A common approach to circumvent this problem is to restrict the mechanism to a class in which it becomes unique, such as triangular monotone maps or optimal transport maps. However, whether the assumed class is correct is usually not examined. In this work, we study what the data can tell us when covariates are available that influence the outcome only through its latent cause. First, we characterise all counterfactual maps that are consistent with the data. They differ from the true map by a group of symmetries that preserve the conditional outcome distributions given the covariates. The more these distributions vary across covariate values, the smaller the group. Second, we show that for the classes above, consistency with the data is equivalent to a single conditional independence, which yields a test of the assumed class. We illustrate both results on synthetic data and on the ACTG 175 HIV trial.
Figures & tables
Figure 1: Causal graph of our setup. The latent variable U is drawn as a dashed node. The dashed edge between U and V denotes that they may be dependent.
Figure 2: The transport map T0→1f=f1∘f0−1 pushes Y∣\doop(A=0) forward to Y∣\doop(A=1) (a). Conditioning on V does not change the map, but it must now also push each conditional distribution Y∣\doop(A=0),V=v to Y∣\doop(A=1),V=v for every v (b, shaded regions). The more these conditional distributions vary with v , the fewer maps satisfy the constraint.
Figure 3: Intuition for (ii) of Theorem 11 . (a) Since U⊥⊥A∣V , both conditionals Y∣A=a,V=v are images of the same latent distribution. (b) Pulled back with h , they must therefore coincide. Requiring this for every v is the independence of Theorem 11 (ii). In the example in (b) they do not coincide, so an eventual hypothesis test will likely be rejected.
Figure 4: Observed distributions for the three different settings when K=4 coloured by v . (a) shows ν0,v which is equal in all settings. (b–d) show ν1,v for the three settings. Larger points denote the means of νa,v .
Figure 5: Identified sets in the synthetic example (Setting 2) for an increasing number of sectors K . The background shows the conditional distributions ν0,v , coloured by v . Three factual points y , and the identified set of the counterfactual of each y : the whole plane for K=1 , two points for K=2 , and the single point S2y for K=4 .
Figure 6: (a) Rejection rate of the ci test against ρ for three combinations of true outcome function and hypothesised model class and three sample sizes. (b) The conditional distributions ν0,v for two values of ρ .
Figure 7: Pulled-back conditional means of the observed ZDV outcome (circles) and the observed ZDV+ddI outcome (crosses), coloured by V , for the three hypothesised classes (a–c) and the reflected Brenier map TR (d). A data-consistent map must pull back each treated mean onto its untreated counterpart (compare Figure 3 (b)). The reflection does not achieve that.
Appendix figures & tables3 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 8: ACTG 175 outcomes. (a) Standardised log CD4 and CD8 counts at week 20 by arm. (b) The ZDV arm coloured by V . Circles mark the conditional means E[Y∣A=0,V=v] . (c) Shift of the conditional means from ZDV ( A=0 ) to ZDV+ddI ( A=1 ).
Figure 9: Estimated counterfactual maps T0→1hC from ZDV to ZDV+ddI for the three hypothesised model classes, drawn as arrows on a grid over the ZDV sample.
Figure 10: p -value of the ci test for the hypotheses hθ , which transport η to ν1 for every θ but rotate the treated latents by θ . Outside the shown window p<10−8 for all three classes.
Previous work establishing completeness results for counterfactual identification has been circumscribed to the setting where the input data belongs to observational or interventional distributions (Layers 1 and 2 of Pearl's Causal Hierarchy), since it was generally presumed impossible to obtain data from counterfactual distributions, which belong to Layer 3. However, recent work (Raghavan & Bareinboim, 2025) has formally characterized a family of counterfactual distributions which can be directly estimated via experimental methods - a notion they call counterfactual realizability. This leaves open the question of what additional counterfactual quantities now become identifiable, given this new access to (some) Layer 3 data. To answer this question, we develop the CTFIDU+ algorithm for identifying counterfactual queries from an arbitrary set of Layer 3 distributions, and prove that it is complete for this task. Building on this, we establish the theoretical limit of which counterfactuals can be identified from physically realizable distributions, thus implying the fundamental limit to exact causal inference in the non-parametric setting. Finally, given the impossibility of identifying certain critical types of counterfactuals, we derive novel analytic bounds for such quantities using realizable counterfactual data, and corroborate using simulations that counterfactual data helps tighten the bounds for non-identifiable quantities in practice.
Arvind Raghavan, Elias Bareinboim
Causal Artificial Intelligence Lab, Department of Computer Science, Columbia University
We develop theory for the identification of complex causal queries from non-IID data. We propose a formal description of stochastic systems based on symmetries in data leaving causal mechanisms invariant. These symmetries become the source of statistical repetition for identification and mandate knowledge-transfer from observations to queries. The result is an abstract, simple and very general language for causal reasoning. We give descriptions of models and queries, and the formal infrastructure for their mathematically rigorous identification. The proposed structure enables a number of simplifying principles: A shared description of mechanisms and interventions, a duality of observed and queried structure, and the sub-division of the identification task into three steps - extraction, computation and query-expression. Our approach reproduces and matches standard theoretical results on IID data and transport of experimental and non-experimental data. By design, it additionally applies beyond IID data and captures complex causal queries not expressible via do- or soft-interventions.
Martin Rabel, Jakob Runge
Department of Computer Science, University of Potsdam, Potsdam, Germany
We consider the problem of learning a counterfactually fair regressor. We adopt a causal uncertainty view in which counterfactual fairness is defined with resampled noise. We focus on obtaining theoretical fairness guarantees for a new post-processing estimator. We begin by showing that counterfactual fairness is equivalent to satisfying demographic parity conditional on the latent variable. This allows us to provide a closed-form expression of the optimal fair regressor via a barycentric quantile map. In order to handle continuous latent variables, we propose a discretized post-processing method. Then, under mild regularity assumptions, we prove high-probability finite-sample fairness guarantees for our estimator, providing an unfairness decay at rate O~(n−1/3), and establishing a matching risk bound of order O~(n−1/3). We provide a matching lower bound on the excess risk of almost fair predictions. Finally, we extend our results to the setting of relaxed counterfactual fairness. We validate our approach on real-world and synthetic data.
M. Generali Lince, S. Gaucher, J-J. Vie +1
Inria, Soda team · CREST, ENSAE, IP Paris · CMAP, CNRS, Ecole Polytechnique, Institut Polytechnique de Paris +1