Policy Learning with $\alpha$-Expected Welfare

Gaoqian Xu; Yanqin Fan; Yuan Qi

REVIEW 4 major objections 3 minor 74 references

Policy Learning with $\alpha$-Expected Welfare

T0 review · 4 major / 3 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read For any $\alpha\in(0,1)$, the debiased $\alpha$-EWM policy achieves regret of order $\sqrt{\mathrm{VC}(\Pi_n)/n}$ and supports valid inference for the optimal lower-tail welfare.

desk verdict Solid theoretical extension of AVaR policy learning to unknown propensity and general policy classes; the main gap is the unverified nuisance-rate assumption behind the numerical pipeline, plus a few fixable write-up defects. read the letter →

arxiv 2505.00256 v1 pith:IPNEOASI submitted 2025-05-01 econ.EM

Yanqin Fan , Yuan Qi , Gaoqian Xu This is my paper · ORCID

classification econ.EM MSC 62G0562G2062P20

keywords AverageValueatRiskconditionalpolicylearningdoublyrobustestimationcross-fittingregretboundsoptimalwelfareinferencedistributionallyoptimization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes to choose treatment-assignment policies by maximizing the $\alpha$-expected welfare: the average outcome of the worst-off $\alpha$-fraction of the post-treatment outcome distribution. It shows that this criterion is identified under unconfoundedness and can be written in a dual form as a supremum over a cutoff $\eta$ of a penalized shortfall, which makes it amenable to doubly robust estimation with cross-fitting. The main theoretical results are an asymptotic upper regret bound of order $\sqrt{\mathrm{VC}(\Pi_n)/n}$ with an explicit constant depending on $\alpha$, matching the best known rate for mean-outcome policies, and asymptotically valid Wald-type and bootstrap inference for the optimal $\alpha$-expected welfare. The paper also shows how the rule interpolates between Rawlsian welfare and expected welfare, and how the same estimator evaluates the lower-tail welfare of any policy for any $\alpha$, enabling welfare-loss comparisons.

What carries the argument

The central object is the dual representation $W_\alpha(\pi)=\sup_{\eta\in B_Y} (1/\alpha)\mathbb{E}[(Y_i(\pi)-\eta)_-]+\eta$, where $\eta$ is a cutoff and $(y-\eta)_-$ is the shortfall below it. The paper constructs a doubly robust score $g_\theta$ for the inner value $V_\alpha(\pi,\eta)$, estimates it with $K$-fold cross-fitted nuisance functions, and bounds the regret by controlling the empirical process over the score class with a new conditional semi-metric and a chaining argument. Inference is carried through the influence-function expansion of the debiased estimator, with the numerical delta method used when the set of optimal policies is not a singleton.

What would settle it

Simulate a DGP satisfying unconfoundedness and strong overlap, choose a policy class with $\mathrm{VC}(\Pi_n)=n^c$ for $c\in(0.5,1)$, and use nuisance estimators converging at only $n^{-1/4}$; if the realized regret of the debiased $\alpha$-EWM rule does not shrink at the claimed $\sqrt{\mathrm{VC}(\Pi_n)/n}$ rate, the theorem's rate condition is violated. On the numerical side, compute the empirical sup-norm convergence rate of the paper's random-forest nuisance estimators on the synthetic JTPA superpopulation: if it does not vanish at the required rate, the simulations do not exercise the theorem's assumptions.

Watch

Extended reading notes

Core claim

The paper's central discovery is that replacing the mean outcome with the $\alpha$-expected welfare does not change the fundamental statistical difficulty of policy learning. Maximizing the cross-fitted doubly robust score over a policy class with VC dimension $\mathrm{VC}(\Pi_n)$ yields regret bounded by an explicit constant times $\sqrt{\mathrm{VC}(\Pi_n)/n}$, and the same debiased objective supports inference: if the optimal policy is unique, $\sqrt{n}(\hat V_n(\hat\theta_n)-V(\theta_o))$ converges in distribution to $N(0,\sigma_o^2)$, while non-unique settings are handled by multiplier bootstrap and the numerical delta method. This means that distributionally robust, lower-tail targeting is feasible in observational studies with unknown propensity scores and outcome regressions.

Load-bearing premise

Everything rests on Assumption 4.2, which requires the estimated propensity score and conditional-outcome regressions to converge to their true functions uniformly and at rates faster than the policy class grows; the paper validates these rates for kernel and sieve estimators but does not verify that the random-forest and simulated-annealing pipeline used in its numerical work satisfies them.

Editorial extensions

If this is right

A policymaker can target the bottom $\alpha$ of the outcome distribution with the same policy classes and the same regret rate as mean-outcome welfare maximization; the targeting choice only changes the constant.
For any candidate policy, the same estimator consistently evaluates $W_\alpha(\pi)$ for every $\alpha$, so one can quantify how much welfare the most disadvantaged lose when a policy is selected for a broader group.
The debiased estimator permits flexible machine-learning nuisance models, as long as they satisfy Assumption 4.2; this is what makes the method operational in observational studies.
Valid inference for the optimal $\alpha$-expected welfare is available under uniqueness via Wald intervals and without it via uniform intervals, covering the exceptional-distribution cases that break mean-outcome inference.
The first-best lower-tail policy has the explicit form $\pi^*_{\mathrm{FB}}(x)=1\{\tau(x,\eta^*_{\mathrm{FB}})>0\}$, so the optimal rule is a threshold on a modified treatment effect evaluated at the optimum cutoff $\eta^*_{\mathrm{FB}}$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because $W_\alpha$ has a distributionally robust interpretation as a worst-case expectation over a risk envelope $U_\alpha(F_\pi)$, the same dual formulation could be used to derive policies that transfer to a target population whose distribution differs from the study population only in tail mass; the paper does not develop this transfer-learning direction.
The debiasing construction may extend to other rank-dependent welfare criteria, such as the extended Gini family, which the paper shows can be written as weighted averages of $W_\alpha$; a testable extension would be regret and inference theorems for those criteria obtained by integrating the $\alpha$-EWM results.
A direct way to stress-test the theory is to measure the actual sup-norm and $L^2$ convergence rates of the random-forest nuisance estimators on the synthetic JTPA superpopulation at $n=2{,}000$, $5{,}000$, and $10{,}000$; unless those rates satisfy Assumption 4.2, the simulation coverage rates are not evidence for the theoretical claims.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know about this paper before you read it. First, the core theory is sound: the doubly robust score, the regret bound, and the inference for the optimal lower-tail welfare all hang together as far as I can tell. Second, the empirical section never connects the random forest / simulated annealing pipeline to the assumptions that the theorems require, and there are a couple of distracting editorial defects. Both are fixable, so I would treat this as a good paper that needs another round rather than a rejection.

The genuinely new pieces are the cross-fitted doubly robust estimator for the AVaR-based policy under an unknown propensity score, the explicit regret bound for general VC-type policy classes, and the inference procedures that cover nonunique optima. The paper is honest about its relation to Qi et al. (2023): the AVaR welfare criterion is theirs, and what is added here is the observational-data machinery, the general policy class, and the inference theory. The JTPA application and the WGAN simulation are careful and the code is posted.

The main substantive soft spot is Assumption 4.2. It requires sup-norm consistency and L2 convergence rates for the nuisance estimators, plus a margin condition between VC dimension and those rates. The paper's Remark 4.1 validates kernels and sieves; for deep neural nets it cites the literature. But the actual implementation uses random forests with default tuning and simulated annealing with no convergence certificate. The theorems are conditional on Assumption 4.2, and the numerical results are not covered by that assumption. I don't think this is a fatal flaw - the theory stands on its own - but the empirical claim that the debiased estimator performs satisfactorily is heuristic unless the authors either verify the condition for RF/SA or soften the claim.

The other two issues are mechanical. Appendix B.2 contains an unresolved author note ('GX:') that makes the faster-rate section incomplete. Lemma G.1 swaps the labels of mu_0 and mu_1 in the decomposition of the score; that's a typo, but it makes an already dense appendix harder to follow.

In sum: the paper deserves a serious referee. I would send it out and ask for a major revision that (a) explicitly addresses Assumption 4.2 for the numerical pipeline or re-frames the simulations as heuristic, (b) removes the GX note, and (c) fixes the Lemma G.1 labels. The central contribution - debiased AVaR policy learning with inference - is solid and likely to be cited by people working on policy learning with distributional objectives.

Referee Report

4 major / 3 minor

Summary. The paper introduces an α-expected welfare criterion, defined as the average outcome of the worst-off α-fraction of the post-treatment outcome distribution, and studies policies that maximize it. Using the dual representation of this AVaR-type functional, the authors construct a cross-fitted doubly robust score, propose a debiased estimator of the optimal α-EWM policy, and derive an asymptotic regret bound of order sqrt(VC(Π_n)/n) with an explicit constant. They also develop Wald and uniform bootstrap inference for the optimal welfare, with and without a uniqueness assumption on the optimal policy, and illustrate the method on the JTPA data and on WGAN-based and stylized simulations.

Significance. If the formal results are correct, this is a useful and natural extension of empirical welfare maximization to distributionally sensitive objectives. The dual formulation yields an orthogonal score that makes the extension non-obvious, the explicit regret constant is a genuine improvement over rate-only statements, and the inference results cover both unique and non-unique optima. The paper also ships data and code and uses a WGAN-based simulation design, which is a strength. The main risks are not circularity or internal inconsistency in the central derivation, but rather verification of the high-level nuisance and optimization conditions in the numerical pipeline and a few proof details that need to be completed.

major comments (4)

[Section 6.2; Assumption 4.2; Remark 4.1] The main regret and inference results, including Theorem 4.1, Theorem 5.1, and Lemma 4.1, require Assumption 4.2, in particular sup_{η∈B_Y} E|μ̂_a(X,η)-μ_a(X,η)|^2 = O(n^{-2ζ_μ}) with ζ_μ∧ζ_e > b_o/2. Section 6.2 estimates μ_a with regression_forest() using default tuning and maximizes the objective with simulated annealing, but the paper provides no rate verification or diagnostics for the uniform-in-η L2 condition. Remark 4.1 justifies the rates only for kernels and sieves, so as written the numerical implementation does not establish that the implemented pipeline satisfies the assumptions behind the stated guarantees. Please provide rates or diagnostics for the random-forest estimator covering uniform-in-η control, or clearly separate the formal theory under Assumption 4.2 from the numerical pipeline as heuristic validation.
[Section 5.2, Eq. (5.1)] The decomposition in Eq. (5.1) labels V(θ̂_n)-V(θ_o) as -Reg(π̂_n,Π), but this is not an equality in general: W(π̂_n)=sup_η V(π̂_n,η) can exceed V(θ̂_n) when η̂_n is not the η-maximizer for π̂_n. The subsequent proof of part (ii) and of Theorem 5.1 uses this identification, so the proof needs repair. The theorem is likely true, but the argument should replace the equality by the correct inequality and bound the gap V(θ̂_n)-W(π̂_n), for instance by using that θ̂_n maximizes V̂_n together with the uniform closeness of V̂_n to V_n.
[Appendix B.2, Theorem B.1] The first paragraph of Appendix B.2 contains the unresolved author note "GX: Based on Eq. (D.2), ...", indicating that the reduction of the regret bound to an oracle regret plus a nuisance-estimation or coupling term is not supplied. Lemma E.1 alone does not close the argument, so Theorem B.1 is not fully self-contained as written. The derivation should be completed, or Theorem B.1 should be explicitly stated under a high-level coupling condition until that step is written out.
[Remark 4.3; Appendix H, Algorithm 1] Theorem 4.1 is stated for an exact maximizer, and Remark 4.3 permits approximate maximizers only if V̂_n(θ̂) ≥ sup_θ V̂_n(θ) - o_P(r_n) with r_n = o_P(n^{-1/2}). Section 6 runs simulated annealing without a convergence certificate or a bound on the optimization error, so the numerical results do not verify the near-optimality condition used in the theory. Please report a verifiable stopping rule or diagnostic for the SA step, or weaken the claims made for the SA-based implementation.

minor comments (3)

[Section 4.3.1, Lemma 4.2] The condition in Lemma 4.2 is stated as "If b_o/2 > ζ_e∧κ_μ"; the correct condition should involve ζ_μ and the inequality should be b_o/2 < ζ_e∧ζ_μ. The symbol κ_μ is not defined anywhere.
[Appendix A, Lemma A.1] In the proof of Lemma A.1, the text says "Since χ_0 and χ_2 are continuous" but only χ_0 and χ_1 are defined; χ_2 should be χ_1.
[Appendix D.1, proof of Lemma 3.1] The displayed equality involving E[(Y_i(π)-η)_- (Y_i(π)-η)_+] is missing the plus sign in the algebraic manipulation; as written it asserts a product identity that is false. Please correct the formula.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the α-expected welfare criterion is externally defined as AVaR, the debiased estimator uses standard doubly robust moment conditions, and the regret and inference theorems are conditional on stated regularity assumptions rather than fitted inputs renamed as predictions.

full rationale

The paper's central derivation chain is self-contained. Definition 2.1 defines Wα(π) externally as the normalized lower-tail integral of Fπ^{-1}, and Lemma 3.1 obtains the dual representation Wα(π)=sup_η Vα(π,η) by appeal to the standard AVaR representation (Shapiro et al. 2021), not by defining the objective in terms of the estimator. Theorem 3.1 and Eq. (3.2) construct the doubly robust score from identified conditional means and the propensity score, with explicit Neyman orthogonality; no parameter is fitted to force a target conclusion. The regret bound in Theorem 4.1 is proven from Lemma 4.1 (controlling nuisance estimation error under Assumption 4.2) and the empirical-process covering-number Lemma 4.3, with constants depending on α, Ξ, and Ξ† rather than on estimated policy performance. Theorem 5.1 derives asymptotic normality from uniqueness, Donsker properties, and Lemma 4.1; the variance sigma_o^2 is the variance of an influence function, not a renormalized goodness-of-fit statistic. The numerical section uses random forests and simulated annealing but does not verify Assumption 4.2 for that pipeline; this is an assumption-verification gap, not a circular reduction. Appendix B.2 also contains an unresolved editorial note, 'GX: Based on Eq. (D.2), the regret bound can be upper-bounded by the sum of the oracle regret bound combined with the nuisance parameter estimation error or the uniform coupling error, as established in Lemma 4.1,' which flags an omitted proof detail for the faster-rate result, but does not make the main derivation equivalent to its inputs. Self-citations, such as Fan et al. (2023), appear only in the literature review and are not load-bearing for the paper's theorems. The empirical benchmarks and WGAN simulations are external validations, not inputs to the proof. No circular step can be exhibited from the paper's own equations or from a self-citation chain.

Assumptions & free parameters 0 free parameters · 9 assumptions · 0 invented entities

No free constants are fitted: alpha is a user-chosen policy parameter, and the compact set B_Y is only an optimization domain. The central claim rests on standard causal assumptions, nuisance regularity, VC complexity, boundedness, and uniqueness conditions. The most fragile premise is Assumption 4.2 for machine-learning nuisance estimators, which the numerical section does not verify for the specific random forest implementation.

assumptions (9)

domain assumption SUTVA and potential outcomes: Yi = AiYi(1) + (1-Ai)Yi(0)
Section 2 opening defines the realized outcome from potential outcomes; this causal model underlies the definition of W_alpha(pi).
domain assumption Assumption 2.1 unconfoundedness plus strong overlap
Identifies F0(·|x) and F1(·|x), hence identifies W_alpha(pi); strong overlap with e0 bounded away from 0 and 1 is used in the debiased score and regret bound.
domain assumption Assumption 4.2 nuisance convergence: uniform consistency and L2 rates n^{-zeta_mu}, n^{-zeta_e} for mu_hat_a and e_hat
Lemma 4.1 equates sup|bV_n - V_n| = O_P(n^{-1/2}) under these rates; all regret and inference results inherit this premise. The random forests used in Section 6 are not shown to satisfy it.
domain assumption Assumption 4.1 VC dimension bound VC(Pi_n) <= n^{b0} with b0/2 < zeta_e and b0/2 < zeta_mu
Bounds the entropy of the policy class and feeds the covering number bound in Lemma 4.3; the linear, tree, and ReLU network examples satisfy the condition.
domain assumption Assumption 4.3 L2(P)-bounded potential outcomes
Ensures envelope functions for g_theta and the covering number bound in Lemma 4.3.
domain assumption Assumption 5.1(1) bounded outcome support in the inference section
Restricts eta to a compact set, makes g_theta uniformly bounded, and gives the Donsker property needed for Theorem 5.1.
domain assumption Assumption 5.2 uniqueness and well-separated maximizer
Delivers asymptotic normality of bV_n(btheta_n) in Theorem 5.1; Appendix C relaxes this assumption for uniform inference.
standard math Known CVaR dual representation sup_eta E[(Y(pi)-eta)_-]/alpha + eta
Lemma 3.1, cited to Shapiro et al. (2021) and Rockafellar et al. (2000); this is the starting point of the debiased score, not a new postulate.
standard math Hadamard directional differentiability of the supremum functional and validity of the numerical delta method
Used in Appendix C for uniform inference; cited to Belloni et al. (2017), Fang and Santos (2019), and Hong and Li (2018).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Policy Learning with $\alpha$-Expected Welfare." pith.science (2026). https://pith.science/paper/IPNEOASI

@misc{pith2026250500256,
  author       = {Pith},
  title        = {Pith review of: Policy Learning with $\alpha$-Expected Welfare},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IPNEOASI}},
  note         = {Machine review of arXiv:2505.00256}
}

abstract

This paper proposes an optimal policy that targets the average welfare of the worst-off $\alpha$-fraction of the post-treatment outcome distribution. We refer to this policy as the $\alpha$-Expected Welfare Maximization ($\alpha$-EWM) rule, where $\alpha \in (0,1]$ denotes the size of the subpopulation of interest. The $\alpha$-EWM rule interpolates between the expected welfare ($\alpha=1$) and the Rawlsian welfare ($\alpha\rightarrow 0$). For $\alpha\in (0,1)$, an $\alpha$-EWM rule can be interpreted as a distributionally robust EWM rule that allows the target population to have a different distribution than the study population. Using the dual formulation of our $\alpha$-expected welfare function, we propose a debiased estimator for the optimal policy and establish its asymptotic upper regret bounds. In addition, we develop asymptotically valid inference for the optimal welfare based on the proposed debiased estimator. We examine the finite sample performance of the debiased estimator and inference via both real and synthetic data.

Figures

Figures reproduced from arXiv: 2505.00256 by the authors.

**Figure 2.** Optimal policies from the linear class Π [PITH_FULL_IMAGE:figures/full_fig_p025_2.png] view at source ↗

**Figure 3.** Optimal policies from the linear class Π [PITH_FULL_IMAGE:figures/full_fig_p026_3.png] view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: Between-quantile differences in outcomes for the 0.25-EWM, 1-EWM, and equality [PITH_FULL_IMAGE:figures/full_fig_p029_4.png]

**Figure 5.** Figure 5: Marginal histograms for JTPA and WGAN-JTPA data. [PITH_FULL_IMAGE:figures/full_fig_p062_5.png]

**Figure 6.** Figure 6: Between-variable correlations for JTPA and WGAN-JTPA data. [PITH_FULL_IMAGE:figures/full_fig_p062_6.png]

**Figure 7.** Figure 7: Between-quantile differences in outcomes for the 0.25-EWM, 1-EWM, and equality [PITH_FULL_IMAGE:figures/full_fig_p065_7.png]

**Figure 8.** Figure 8: Between-quantile differences in outcomes for the 0.25-EWM, 1-EWM, and equality [PITH_FULL_IMAGE:figures/full_fig_p065_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 53 canonical work pages

[1]

and Christensen, T

Adjaho, C. and Christensen, T. (2022). Externally valid treatment choice. arXiv preprint arXiv:2205.05561 , 1

arXiv 2022
[2]

and Chen, X

Ai, C. and Chen, X. (2003). Efficient estimation of models with conditional moment restrictions containing unknown functions. Econometrica , 71(6):1795--1843

2003
[3]

Andrews, D. W. (1994). Empirical process methods in econometrics. Handbook of econometrics , 4:2247--2294

work page 1994
[4]

W., Metzger, J., and Munro, E

Athey, S., Imbens, G. W., Metzger, J., and Munro, E. (2024). Using wasserstein generative adversarial networks for the design of monte carlo simulations. Journal of Econometrics , 240(2):105076

work page 2024
[5]

Athey, S., Tibshirani, J., and Wager, S. (2019). Generalized random forests. The Annals of Statistics , 47(2):1148--1178

2019
[6]

and Wager, S

Athey, S. and Wager, S. (2021). Policy learning with observational data. Econometrica , 89(1):133--161

2021
[7]

L., Harvey, N., Liaw, C., and Mehrabian, A

Bartlett, P. L., Harvey, N., Liaw, C., and Mehrabian, A. (2019). Nearly-tight vc-dimension and pseudodimension bounds for piecewise linear neural networks. Journal of Machine Learning Research , 20(63):1--17

2019
[8]

Bartlett, P. L. and Mendelson, S. (2002). Rademacher and gaussian complexities: Risk bounds and structural results. Journal of Machine Learning Research , 3(Nov):463--482

2002

Show all 74 references

[9]

Belloni, A., Chernozhukov, V., Chetverikov, D., and Kato, K. (2015). Some new asymptotic theory for least squares series: Pointwise and uniform results. Journal of Econometrics , 186(2):345--366

2015
[10]

Belloni, A., Chernozhukov, V., Fernandez-Val, I., and Hansen, C. (2017). Program evaluation and causal inference with high-dimensional data. Econometrica , 85(1):233--298

2017
[11]

and Dunn, J

Bertsimas, D. and Dunn, J. (2017). Optimal classification trees. Machine Learning , 106:1039--1082

2017
[12]

and Dupas, P

Bhattacharya, D. and Dupas, P. (2012). Inferring welfare maximizing treatment assignment under budget constraints. Journal of Econometrics , 167(1):168--196

2012
[13]

S., Orr, L

Bloom, H. S., Orr, L. L., Bell, S. H., Cave, G., Doolittle, F., Lin, W., and Bos, J. M. (1997). The benefits and costs of jtpa title ii-a programs: Key findings from the national job training partnership act study. Journal of Human Resources , 32(3):549--576

1997
[14]

Blundell, R., Chen, X., and Kristensen, D. (2007). Semi-nonparametric iv estimation of shape-invariant engel curves. Econometrica , 75(6):1613--1669

2007
[15]

and Christensen, T

Chen, X. and Christensen, T. M. (2015). Optimal uniform convergence rates and asymptotic normality for series estimators under weak dependence and weak conditions. Journal of Econometrics , 188(2):447--465

2015
[16]

Chernozhukov, V., Chetverikov, D., Demirer, M., Duflo, E., Hansen, C., Newey, W., and Robins, J. (2018). Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal , 21(1):C1--C68

2018
[17]

Chernozhukov, V., Chetverikov, D., and Kato, K. (2014). Gaussian approximation of suprema of empirical processes. 42(4):1564--1597

2014
[18]

C., Ichimura, H., Newey, W

Chernozhukov, V., Escanciano, J. C., Ichimura, H., Newey, W. K., and Robins, J. M. (2022). Locally robust semiparametric estimation. Econometrica , 90(4):1501--1535

2022
[19]

and Han, S

Cui, Y. and Han, S. (2023). Individualized treatment allocations with distributional welfare. arXiv preprint arXiv:2311.15878

2023 arXiv
[20]

Duchi, J., Hashimoto, T., and Namkoong, H. (2023). Distributionally robust losses for latent covariate mixtures. Operations Research , 71(2):649--664

2023
[21]

Durrett, R. (2019). Probability: Theory and Examples , volume 49. Cambridge university press

2019
[22]

Fan, Y., Park, H., and Xu, G. (2023). Quantifying distributional model risk in marginal problems via optimal transport. arXiv preprint arXiv:2307.00779

2023 arXiv
[23]

X., Wang, Z., and Wang, L

Fang, E. X., Wang, Z., and Wang, L. (2023). Fairness-oriented learning for optimal individualized treatment rules. Journal of the American Statistical Association , 118(543):1733--1746

2023
[24]

and Santos, A

Fang, Z. and Santos, A. (2019). Inference on directionally differentiable functions. The Review of Economic Studies , 86(1):377--412

2019
[25]

H., Liang, T., and Misra, S

Farrell, M. H., Liang, T., and Misra, S. (2021). Deep neural networks for estimation and inference. Econometrica , 89(1):181--213

2021
[26]

F., and Parker, T

Firpo, S., Galvao, A. F., and Parker, T. (2023). Uniform inference for value functions. Journal of Econometrics , 235(2):1680--1699

2023
[27]

and Geman, D

Geman, S. and Geman, D. (1984). Stochastic relaxation, gibbs distributions, and the bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence , 6(6):721--741

1984
[28]

and Guillou, A

Gin \'e , E. and Guillou, A. (2002). Rates of strong uniform consistency for multivariate kernel density estimators. Annales de l'Institut Henri Poincare (B) Probability and Statistics , 38(6):907--921

2002
[29]

and Nickl, R

Gin \'e , E. and Nickl, R. (2021). Mathematical Foundations of Infinite-Dimensional Statistical Models . Cambridge university press

2021
[30]

and Zitikis, R

Greselin, F. and Zitikis, R. (2018). From the classical gini index of income inequality to a new zenga-type relative measure of risk: A modeller’s perspective. Econometrics , 6(1):4

2018
[31]

and Li, J

Hong, H. and Li, J. (2018). The numerical delta method. Journal of Econometrics , 206(2):379--394

2018
[32]

Husmann, K., Lange, A., and Spiegel, E. (2017). The r package optimization: Flexible global optimization with simulated-annealing. CRAN citation

2017
[33]

Kallus, N. (2018). Balanced policy evaluation and learning. In Proceedings of the 32nd International Conference on Neural Information Processing Systems , pages 8909--8920

2018
[34]

and Zhou, A

Kallus, N. and Zhou, A. (2018). Confounding-robust policy improvement. In Proceedings of the 32nd International Conference on Neural Information Processing Systems , pages 9289--9299

2018
[35]

Kennedy, E. H. (2016). Semiparametric theory and empirical processes in causal inference. Statistical Causal Inferences and Their Applications in Public Health Research , page 141

2016
[36]

and Pollard, D

Kim, J. and Pollard, D. (1990). Cube root asymptotics. The Annals of Statistics , 18(1):191--219

1990
[37]

and Zubizarreta, J

Kim, K. and Zubizarreta, J. (2023). Fair and robust estimation of heterogeneous treatment effects for policy learning. arXiv preprint arXiv:2306.03625

2023 arXiv
[38]

D., and Vecchi, M

Kirkpatrick, S., Gelatt Jr, C. D., and Vecchi, M. P. (1983). Optimization by simulated annealing. Science , 220(4598):671--680

1983
[39]

and Tetenov, A

Kitagawa, T. and Tetenov, A. (2018). Who should be treated? empirical welfare maximization methods for treatment choice. Econometrica , 86(2):591--616. Supplementary materials and online appendix available at https://doi.org/10.3982/ECTA13288

2018 doi
[40]

and Tetenov, A

Kitagawa, T. and Tetenov, A. (2021). Equality-minded treatment choice. Journal of Business & Economic Statistics , 39(2):561--574

2021
[41]

and Langer, S

Kohler, M. and Langer, S. (2021). On the rate of convergence of fully connected deep neural network regression estimates. The Annals of Statistics , 49(4):2231--2249

2021
[42]

Kosorok, M. R. (2008). Introduction to Empirical Processes and Semiparametric Inference . Springer

2008
[43]

Lei, L., Sahoo, R., and Wager, S. (2023). Policy learning under biased sample selection. arXiv preprint arXiv:2304.11735

2023 arXiv
[44]

and Chambaz, A

Luedtke, A. and Chambaz, A. (2020). Performance guarantees for policy learning. Annales de L'Institut Henri Poincare Section (B) Probability and Statistics , 56(3):2162--2188

2020
[45]

Luedtke, A. R. and van der Laan, M. J. (2016). Statistical inference for the mean outcome under a possibly non-unique optimal treatment strategy. The Annals of Statistics , 44(2):713--742

2016
[46]

Luedtke, A. R. and van der Laan, M. J. (2018). Parametric-rate inference for one-sided differentiable parameters. Journal of the American Statistical Association , 113(522):780--788

2018
[47]

and N \'e d \'e lec, \'E

Massart, P. and N \'e d \'e lec, \'E . (2006). Risk bounds for statistical learning. The Annals of Statistics , 34(5):2326--2366

2006
[48]

Newey, W. (1994). The asymptotic variance of semiparametric estimators. Econometrica , 62(6):1349--82

1994
[49]

Qi, Z., Pang, J.-S., and Liu, Y. (2023). On robustness of individualized decision rules. Journal of the American Statistical Association , 118(543):2143--2157

2023
[50]

and Murphy, S

Qian, M. and Murphy, S. A. (2011). Performance guarantees for individualized treatment rules . The Annals of Statistics , 39(2):1180--1210

2011
[51]

Rai, Y. (2018). Statistical inference for treatment assignment policies. Unpublished Manuscript

2018
[52]

Rawls, J. (2001). Justice as Fairness: A Restatement . Harvard University Press

2001
[53]

M., Rotnitzky, A., and Zhao, L

Robins, J. M., Rotnitzky, A., and Zhao, L. P. (1994). Estimation of regression coefficients when some regressors are not always observed. Journal of the American statistical Association , 89(427):846--866

1994
[54]

M., Rotnitzky, A., and Zhao, L

Robins, J. M., Rotnitzky, A., and Zhao, L. P. (1995). Analysis of semiparametric regression models for repeated outcomes in the presence of missing data. Journal of the American Statistical Association , 90(429):106--121

1995
[55]

T., Uryasev, S., et al

Rockafellar, R. T., Uryasev, S., et al. (2000). Optimization of conditional value-at-risk. Journal of risk , 2:21--42

2000
[56]

T., Uryasev, S

Rockafellar, R. T., Uryasev, S. P., and Zabarankin, M. (2002). Deviation measures in risk analysis and optimization. University of Florida, Department of Industrial & Systems Engineering Working Paper , (2002-7)

2002
[57]

Rubin, D. B. (1978). Bayesian inference for causal effects: The role of randomization. The Annals of statistics , 6(1):34--58

1978
[58]

Rubin, D. B. (1990). Comment: Neyman (1923) and causal inference in experiments and observational studies. Statistical Science , 5(4):472--480

1990
[59]

Schmidt-Hieber, J. (2020). Nonparametric regression using deep neural networks with relu activation function. The Annals of Statistics , 48(4):1875

2020
[60]

and Smola, A

Sch \"o lkopf, B. and Smola, A. J. (2002). Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond . MIT press

2002
[61]

Shapiro, A., Dentcheva, D., and Ruszczynski, A. (2021). Lectures on Stochastic Programming: Modeling and Theory . SIAM

2021
[62]

Shi, C., Lu, W., and Song, R. (2018). A massive data framework for m-estimators with cubic-rate. Journal of the American Statistical Association , 113(524):1698--1709

2018
[63]

Shi, C., Lu, W., and Song, R. (2020). Breaking the curse of nonregularity with subagging---inference of the mean outcome under optimal treatment regimes. Journal of Machine Learning Research , 21(176):1--67

2020
[64]

Shorrocks, A. F. (1983). Ranking income distributions. Economica , 50(197):3--17

1983
[65]

Tsybakov, A. B. (2004). Optimal aggregation of classifiers in statistical learning. The Annals of Statistics , 32(1):135--166

2004
[66]

Van der Vaart, A. W. (2000). Asymptotic Statistics . Cambridge University Press

2000
[67]

van der Vaart, A. W. and Wellner (2023). Weak Convergence and Empirical Processes: With Applications to Statistics . Springer

2023
[68]

van der Vaart, A. W. and Wellner, J. A. (2011). A local maximal inequality under uniform entropy. Electronic Journal of Statistics , 5(2011):192

2011
[69]

and Bradic, J

Viviano, D. and Bradic, J. (2024). Fair policy targeting. Journal of the American Statistical Association , 119(545):730--743

2024
[70]

Wang, L., Zhou, Y., Song, R., and Sherwood, B. (2018). Quantile-optimal treatment regimes. Journal of the American Statistical Association , 113(523):1243--1254

2018
[71]

A., Davidian, M., Zhang, M., and Laber, E

Zhang, B., Tsiatis, A. A., Davidian, M., Zhang, M., and Laber, E. (2012). Estimating optimal treatment regimes from a classification perspective. Stat , 1(1):103--114

2012
[72]

and Cui, Y

Zhao, P. and Cui, Y. (2023). A semiparametric instrumented difference-in-differences approach to policy learning. arXiv preprint arXiv:2310.09545

2023 arXiv
[73]

J., and Kosorok, M

Zhao, Y., Zeng, D., Rush, A. J., and Kosorok, M. R. (2012). Estimating individualized treatment rules using outcome weighted learning. Journal of the American Statistical Association , 107(499):1106--1118

2012
[74]

Zhou, Z., Athey, S., and Wager, S. (2023). Offline multi-action policy learning: Generalization and optimization. Operations Research , 71(1):148--183

2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

and Christensen, T

Adjaho, C. and Christensen, T. (2022). Externally valid treatment choice. arXiv preprint arXiv:2205.05561 , 1

arXiv 2022

[2] [2]

and Chen, X

Ai, C. and Chen, X. (2003). Efficient estimation of models with conditional moment restrictions containing unknown functions. Econometrica , 71(6):1795--1843

2003

[3] [3]

Andrews, D. W. (1994). Empirical process methods in econometrics. Handbook of econometrics , 4:2247--2294

work page 1994

[4] [4]

W., Metzger, J., and Munro, E

Athey, S., Imbens, G. W., Metzger, J., and Munro, E. (2024). Using wasserstein generative adversarial networks for the design of monte carlo simulations. Journal of Econometrics , 240(2):105076

work page 2024

[5] [5]

Athey, S., Tibshirani, J., and Wager, S. (2019). Generalized random forests. The Annals of Statistics , 47(2):1148--1178

2019

[6] [6]

and Wager, S

Athey, S. and Wager, S. (2021). Policy learning with observational data. Econometrica , 89(1):133--161

2021

[7] [7]

L., Harvey, N., Liaw, C., and Mehrabian, A

Bartlett, P. L., Harvey, N., Liaw, C., and Mehrabian, A. (2019). Nearly-tight vc-dimension and pseudodimension bounds for piecewise linear neural networks. Journal of Machine Learning Research , 20(63):1--17

2019

[8] [8]

Bartlett, P. L. and Mendelson, S. (2002). Rademacher and gaussian complexities: Risk bounds and structural results. Journal of Machine Learning Research , 3(Nov):463--482

2002