- Rating
- 3(+1)
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 3(+2)
Read review
Paper summary
This paper proposes BGM-IV for nonlinear instrumental-variable estimation with high-dimensional covariates. It partitions the latent representation into four blocks and incorporates an IV-integrated pseudo-likelihood into a tempered quasi-posterior. Training alternates between updating the covariate, treatment, and outcome generators and the observation-specific latent states. Beyond point estimation, the method constructs predictive intervals for outcomes under intervention. Experiments on three versions of a synthetic demand-design benchmark, including vector and image covariates, report improvements in structural MSE and evaluate predictive coverage.
Critical strengths
Predictive intervals extend the method beyond point estimation. Providing model-based uncertainty information for outcomes under intervention is a useful addition to structural-mean estimation in nonlinear IV problems.
Critical weaknesses
-
The causal justification for the tempered quasi-posterior needs further explanation. The citation in Section 3.4 suggests a generalized-Bayes interpretation. While this can justify a loss-based update, it does not by itself establish that the chosen loss targets the intended causal quantity. The sensitivity analysis for the tempering parameter provides empirical evidence, but does not resolve this theoretical question.
-
The IV-integrated density loss need not recover the correct structural mean. Section 3.2 motivates this objective as a way to address endogeneity, but the usual IV conditions alone do not guarantee this property.
For a simple example without , let be independent standard normal variables and define
The error satisfies , while . The instrument is relevant and excluded from the outcome equation. The moment equation uniquely identifies the true slope .
The true conditional distributions are and . With the correct first stage and the working outcome model , , the paper's IV-integrated pseudo-likelihood in Eq. (3) becomes . Its population negative log-likelihood, up to a constant, is
+\frac{2+(a+2)^2}{2(a^2+s^2)}.$$ Holding $s^2$ fixed gives $$\left.\frac{\partial R}{\partial a}\right|_{a=-2} =-\frac{2}{4+s^2}+\frac{4}{(4+s^2)^2}<0.$$ Thus, slightly increasing $a$ above $-2$ lowers the loss, although this working family contains the true intervention distribution $N(-2x,2)$. This establishes a failure of mean recovery within this linear-Gaussian family, not inconsistency of the full neural estimator. Appendix A analyzes the integrated EGM outcome loss and provides conditional structural-mean guarantees for final prediction. These results do not establish that optimizing Eq. (8) recovers the structural mean. My concern is this missing link, rather than an error in the conditional results: the predictive intervals are useful in principle, but the density objective may introduce structural-mean bias.
Questions for authors
Could the authors explain mathematically under what conditions the IV-integrated pseudo-likelihood addresses endogeneity bias and recovers the structural mean, particularly in light of the example above?
8 comments
Response to Reviewer B67P (2/2): the tempered quasi-posterior (Weakness 1)
5. Weakness 1: what the tempered quasi-posterior is over, and which loss it uses. The quasi-posterior of Section 3.4 (Eq. 5) is a distribution over the unit-level latent state . The causal quantity is carried by the mean of the outcome generator, and in the revision both use the same IV-moment residual. The outcome generator minimizes the precision-weighted IV-integrated squared loss, whose population minimizer is the structural mean (Proposition A.2). The latent quasi-posterior is
that is, the posterior kernel of the generative model for multiplied by a Gibbs factor built from that same IV-moment loss. The loss in the generalized-Bayes update is therefore one whose population target is the structural mean, which is the property required in Weakness 1. With the density objective of the original submission this was not the case, so the two weaknesses had a common cause.
6. The causal content does not rest on the tempering. At the latent update does not involve at all, and the structural mean is still learned through Eq. 8a. We added to the sensitivity study (Table 4, Figure 5): relative to the default , its mean structural MSE is , and times as large at , and , and the four values differ by less than the repetition-to-repetition variability in every cell. Predictions and model selection use latent states re-inferred from the covariates alone (Lemma A.1), never the outcome-informed training states.
7. Effect on the benchmarks. All experiments were rerun with the revised objective. Relative to the original submission, the structural MSE of BGM-IV is lower on the demand design (geometric mean of the paired ratio revised/original: , and at , and ) and on the vector-proxy benchmark (-), and essentially unchanged on the image benchmark (-). This is the direction the analysis in the review predicts, and the comparison with the baselines is unchanged.
Since this concern was central to the assessment, we would be grateful if you would reconsider the rating in light of the revision. We are glad to clarify any remaining point.
Reference. Gourieroux, C., Monfort, A., & Trognon, A. (1984). Pseudo Maximum Likelihood Methods: Theory. Econometrica, 52(3), 681–700. https://doi.org/10.2307/1913471
We hope that our additional experiments and clarifications have addressed your major concerns. I would greatly appreciate your consideration of updating the assessment and score accordingly. We would also be happy to discuss and address any further questions that you may have. Thank you!
Response to Reviewer B67P (1/2): the objective is revised, and the example in the review is solved
Thank you for the precise analysis. We agree that the original objective need not recover the structural mean under the IV assumptions alone. We have revised outcome estimation within BGM-IV’s generative framework by fitting the structural mean through the IV moment restriction and estimating the noise variance separately from structural residuals. With this change, the example in the review is solved exactly, and the revised procedure recovers both the true structural slope and the noise variance at the population level. Appendix A contains the link between the estimation objectives and their structural targets that the review found missing.
1. The example and the objective of the original submission. We agree. For the objective of the original submission (Eq. 8 there), the IV conditions alone do not guarantee recovery of the structural mean. Maximizing the IV-integrated density is a Kullback-Leibler projection of the law of Y∣W,A onto the laws of m(X~,aY)+sξ, with X~ drawn from the first stage independently of the Gaussian ξ. For a fixed sufficient representation, a correct first stage, and completeness, correct specification of the integrated reduced-form density at m0 is sufficient for mean recovery. Matching the reduced-form variance at m0 requires
Var(ϵ∣W,A)+2Cov(m0(X,AY),ϵ∣W,A) ≥ 0.
The IV assumptions do not imply this. In the example the left-hand side is 2+2⋅(−2)=−2, the constraint s2≥0 binds, and the population optimum is a∗=1−7≈−1.65 with s2=0, in agreement with the sign of the derivative in the review.
2. Revision: separate estimation of the structural mean and the noise scale. We retain the Gaussian structural outcome model and revise how its mean and noise scale are estimated. The revised IV term is a Gaussian pseudo-likelihood centered on the structural mean integrated over the first-stage treatment distribution (Eqs. 3-4):
pIV(y∣w,z)=N(y; μˉω,ϕ(w,z), σIV2),μˉω,ϕ(w,z)=∫μω(x,z0,z1)pϕ(x∣w,z0,z2)dx.
The outcome mean network μω minimizes the precision-weighted integrated squared loss (y−μˉω,ϕ(w,z))2/2σω2(z0,z1) with the noise scale held fixed (Eq. 8a). The noise scale σω2(z0,z1) is fitted in a separate step to the residual at the observed treatment, with the mean held fixed (Eq. 8b), and it supplies the precision weights of the mean update. A Gaussian pseudo-likelihood parametrized by a conditional mean is consistent for that mean without requiring the distribution to be correctly specified (Gourieroux et al., 1984), which underlies the mean update.
3. The example under the revised objective. For the working mean m(x)=ax, the revised population objective is Rc(a)=2E[c]+(a+2)2E[cW2]. It is minimized at the structural slope a=−2 for every weight c, and the noise-scale step then returns Var(U+η)=2, so the fitted interventional law is N(−2x,2), which is the true one. Running the complete BGM-IV procedure on this design (n=5,000, ten repetitions, with one independent standard normal covariate added because the method takes covariates) gives:
4. The missing link (Appendix A).
The initialization, the training update and the selection criterion (Eq. 22) now share one population target, and Corollary A.1 gives the conditional consistency of the selected covariate-only predictor as before. Section 3.3 states this in the main text.