betaFeedback

WorksEconomics, Econometrics and Finance

Z-Estimation under Unignorably Large Clusters

Abstract

Cluster-size heterogeneity can invalidate conventional cluster-robust inference when a small number of clusters dominate observation-weighted estimating equations. This paper develops asymptotic theory for cluster-balanced ZZ-estimators under independent clusters, arbitrary realised cluster sizes, and unrestricted within-cluster dependence. Building on the size-adjusted reweighting of Chiang, Sasaki & Wang (2025), the analysis distinguishes target definition, identification, and inference. A common cluster-level zero, imposed as an identifying condition or derived from individual-level exogeneity or design-based mechanisms, makes the structural target invariant to deterministic cluster reweighting; separation of the reweighted population moments then ensures identification. The pointwise results establish asymptotic normality while allowing the covariance sequence to vary along the cluster array. The results for functional parameters establish weak convergence and validity of a cluster multiplier bootstrap for uniform inference. The framework covers smooth and non-smooth scores with distribution, quantile, and expectile regressions as concrete applications. Monte Carlo experiments illustrate pointwise and uniform inference under heterogeneous cluster sizes, and a supplementary empirical application illustrates the consequences of cluster weighting in municipal-school data.

Introduction

Clustered sampling is ubiquitous in empirical research in economics and related fields, where observations within clusters, such as states, schools, or industries, may exhibit arbitrary within-cluster dependence. Ensuring the validity of inference under clustered sampling has hence become a sine qua non. Inference procedures based on cluster-robust standard errors or cluster bootstrap methods, typically coupled with tt- and Wald tests, are routinely employed to account for such dependence [12] (White 1984; Liang and Zeger 1986; Arellano 1987; Cameron, Gelbach, and Miller 2008; Cameron and Miller 2015). Additionally, bootstrap methods are often preferred when the number of clusters is small and asymptotic approximations may be unreliable, or when the estimation of cluster-robust standard errors requires tuning parameters, as in regression quantile models [7]. The validity of these inference procedures, however, ultimately relies on the asymptotic normality of the underlying estimator. This property can fail in the presence of extreme cluster size heterogeneity. Specifically, letting GG denote the number of clusters, NN the total sample size, and NgN_g the size of cluster g∈{1,…,G}g \in\{1,\ldots,G\}, asymptotic normality may fail when max⁡1≤g≤GNg2/N↛0\max_{1\le g\le G} N_g^2/N \not\to0 as N→∞N \to\infty [8] (Chiang, Sasaki, and Wang 2025). Recent theoretical advances, including Djogbenou, MacKinnon, and Nielsen (2019) and Hansen and Lee [8], have relaxed restrictions on cluster growth and within-cluster dependence but continue to require the ignorability of cluster size heterogeneity.

As documented by Chiang, Sasaki, and Wang (2025), the cluster size heterogeneity condition max⁡1≤g≤GNg2/N→0\max_{1\le g\le G} N_g^2/N \to0 likely fails in a wide range of empirical applications. Substantial cluster-size heterogeneity arises in settings such as schools, firms, and municipalities. The supplementary empirical illustration revisits the municipal-school design of Akhtari, Moreira, and Trucco [1], in which students are clustered within Brazilian municipalities of substantially unequal size. The Unignorably Large Cluster (ULC) problem does not require a cluster to retain a positive limiting sample share: the conditions

max⁡g≤GNgN→0,max⁡g≤GNg2N↛0\max_{g\le G}\frac{N_g}{N}\to0,\qquad\max_{g\le G}\frac{N_g^2}{N}\not\to0

can hold simultaneously. Under sufficiently strong within-cluster dependence, cluster scores can then remain first-order influential, so consistency may coexist with failure of conventional Gaussian approximations. To assess the empirical relevance of this issue, Chiang, Sasaki, and Wang (2025), in a pioneering effort, examines 31 articles published in the American Economic Review and Econometrica between 2020 and 2021 that implement cluster-robust inference using publicly available data. Rejection of normality in 24 of the 31 studies (77%) therein raises concerns about the validity of conventional inference in those studies.

Chiang, Sasaki, and Wang (2025) highlights several theoretical challenges arising when cluster size heterogeneity is unignorable. Estimators may be inconsistent, may fail to be asymptotically normal, their conventional cluster-robust variance estimators may fail to consistently estimate asymptotic variances, standard tt- and Wald tests may become invalid, and the wild bootstrap offers no general remedy. The formal asymptotic analysis in Chiang, Sasaki, and Wang (2025) treats cluster sizes as random and potentially heavy-tailed, so that observation-weighted sums can have non-Gaussian stable limits. The present paper adopts a different, complementary asymptotic framework. Throughout the theoretical analysis, cluster sizes are treated as a realised triangular array, and the probabilistic arguments are finite-moment cluster-level weak law of large numbers (WLLN), central limit theorem (CLT), and empirical-process arguments in the tradition of Hansen and Lee [8]. Thus, this extension borrows and generalises the size-adjusted cluster-robust (SACR) reweighting device of Chiang, Sasaki, and Wang (2025) without extending the stable-law asymptotic theory therein.1 Its application to ZZ-estimating equations is termed ZZ-SACR.

Cluster reweighting, as proposed by Chiang, Sasaki, and Wang (2025), replaces observation-weighted averages in the estimating equations with averages of cluster-level scores. Conventional observation weighting gives equal weight to individuals, whereas cluster balancing gives equal total representation to each cluster. These schemes can define different estimands. Conventional observation-weighted inference can fail under ULC, whereas the ZZ-SACR framework supports inference under its stated conditions while accommodating arbitrary realised cluster sizes. The maintained common-zero restriction preserves the structural zero under reweighting. Indeed, in several applications, economic theory or the treatment assignment design motivates observation-level exogeneity conditions that centre each individual score at a common structural parameter (e.g., Parente and Santos Silva (2016); Hagemann [7]). Averaging within clusters then gives the common cluster-level zero, so SACR reweighting preserves that structural zero. Observation-level exogeneity is a sufficient route; the generic theory permits the weaker cluster-average restriction to be imposed directly. Separation of the reweighted population moments establishes identification, while the moment and regularity conditions support consistency and asymptotic normality.

Inverse-cluster-size estimating equations have antecedents in the literature on informative cluster size. Hoffman, Sen, and Weinberg [9] motivates within-cluster resampling by distinguishing selection of an individual from the pooled population from selection of a cluster followed by selection of an individual within it. Cluster-weighted generalised estimating equations implement inverse-size weighting directly (Williamson, Datta, and Satten (2003); Benhin, Rao, and Scott (2005)). Cluster-weighted marginal distributions, quantiles, and regression functionals have also been studied by Nevalainen, Datta, and Oja [13]. The corresponding distinction between individual- and cluster-weighted treatment effects in cluster-randomised designs is discussed by Su and Ding (2021) and Bugni et al. (2025).

Building on these precedents, this paper develops pointwise and functional ZZ-estimation theory for independent, potentially non-identically distributed clusters when conventional cluster-size negligibility conditions fail. Primitive individual-level moment and score function class conditions in this paper support cluster-average laws of large numbers and empirical-process arguments, leading to consistency, asymptotic linearity, covariance-normalised pointwise normality, weak convergence, and functional multiplier-bootstrap inference. The common-zero restriction distinguishes preservation of a structural parameter from a change in population representation. The framework accommodates smooth and non-smooth estimating equations. Distribution Regression (DR), Quantile Regression (QR), and Expectile Regression (ER) serve as concrete examples.

The remainder of the paper is organised as follows. Section 2 formalises the ULC problem within the ZZ-estimation framework. Section 3 develops the general asymptotic theory, and Section 4 verifies the theory for distribution, quantile, and expectile regression as concrete examples. Section 5 presents pointwise and uniform Monte Carlo evidence, and Section 6 concludes. The Supplementary Appendix contains proofs, simulation implementation details and additional experiments, and the empirical illustration.

Notation: There are NN observations in GG clusters. G:={1,…,G}\mathcal{G}:=\{1,\ldots,G\} denotes the set of cluster indices, Gg\mathcal{G}_g is the set of observations in cluster gg, Ng:=∣Gg∣N_g:=|\mathcal{G}_g|, and Wg:={Wi:i∈Gg}\mathcal{W}_g:=\{W_i:i\in\mathcal{G}_g\}, where WiW_i is a vector of unit-ii-specific observed data, with N=∑g=1GNgN=\sum_{g=1}^{G}N_g. g(i)g(i) denotes the unique cluster containing observation ii, and pθ:=dim⁡(θ)p_\theta:=\dim(\theta) for a vector θ\theta. As in empirical-process notation, Pf:=∫f dP\mathbb{P}f:=\int f\,\mathrm{d}P denotes integration under the marginal target population, while P[A]P[A] denotes event probability. The operator EG\mathbb{E}_G denotes expectation under the stage-GG triangular-array law; conditioning is displayed explicitly when needed. For individual-level functions ff and cluster-level functions hh, define P^Nf:=1N∑i=1Nf(Wi)\widehat{\mathbb{P}}_Nf:=\frac{1}{N}\sum_{i=1}^{N}f(W_i), P^Gh:=1G∑g=1Gh(Wg)\widehat{\mathbb{P}}_Gh:=\frac{1}{G}\sum_{g=1}^{G}h(\mathcal{W}_g), and PGh:=EG[P^Gh]\mathbb{P}_Gh:=\mathbb{E}_G[\widehat{\mathbb{P}}_Gh]. Thus PG\mathbb{P}_G is the cluster-averaged stage-GG population operator. For functional parameters, T\mathcal{T} is the index set, and ℓ∞(T)\ell^\infty(\mathcal{T}) denotes the space of bounded functions on T\mathcal{T} equipped with the uniform norm; for vector-valued paths, ∥h∥T:=sup⁡τ∈T∥h(τ)∥\|h\|_{\mathcal{T}}:=\sup_{\tau\in\mathcal{T}}\|h(\tau)\|. ⇝\rightsquigarrow denotes weak convergence, and ⇝P\rightsquigarrow_P denotes conditional weak convergence in probability.

Unignorably Large Clusters and Size-Adjusted ZZ-Estimation

This section formalises the Unignorably Large Cluster (ULC) problem within the ZZ-estimation framework and introduces the ZZ-SACR estimator as a remedy.

The ULC problem in ZZ-estimation

Consider a population moment restriction of the generic form

Pψ(W;θo)=0,(1)\mathbb{P}\psi(W;\theta_o)=0, \tag*{(1)}

where WW is the vector of observable data, θo∈Θ⊂Rpθ\theta_o\in\Theta\subset\mathbb{R}^{p_\theta} uniquely solves the moment restriction, and Θ\Theta is a compact parameter space. Given the moment restriction above and a sample of NN observations, a natural ZZ-estimator of θo\theta_o is obtained as a solution to

P^Nψ(W;θ^)=op(1).\widehat{\mathbb{P}}_N\psi(W;\widehat{\theta})=o_p(1).

In addition to identification, a key ingredient for establishing consistency of θ^\hat{\theta} is a suitable uniform weak law of large numbers (UWLLN),

sup⁡θ∈Θ∥(P^N−P)ψ(W;θ)∥=op(1),\sup_{\theta\in\Theta} \left\|(\widehat{\mathbb{P}}_N-\mathbb{P})\psi(W;\theta)\right\|=o_p(1),

for which pointwise convergence in probability of the estimating equations is necessary, namely

(P^N−P)ψ(W;θ)=op(1)(\widehat{\mathbb{P}}_N-\mathbb{P})\psi(W;\theta)=o_p(1)

for each θ∈Θ\theta\in\Theta. The sampling scheme becomes crucial at this point; it determines whether consistency, asymptotic normality, and valid inference can be established. Under independently sampled clusters, unknown within-cluster dependence, and unit-level marginal laws governed by P\mathbb{P}, suppose that sup⁡θ∈ΘP∥ψ(W;θ)∥2≤C<∞\sup_{\theta\in\Theta}\mathbb{P}\|\psi(W;\theta)\|^2 \le C < \infty, it follows that

∥V[P^Nψ(W;θo)]∥=∥V[1N∑g=1G(∑i∈Ggψ(Wi;θo))]∥≤C∑g=1G(NgN)2.\left\|\mathbb{V}\left[\widehat{\mathbb{P}}_N\psi(W;\theta_o)\right]\right\|=\left\|\mathbb{V}\left[\frac{1}{N}\sum_{g=1}^{G}\left(\sum_{i\in\mathcal{G}_g}\psi(W_i;\theta_o)\right)\right]\right\|\le C\sum_{g=1}^{G}\left(\frac{N_g}{N}\right)^2.

Under unrestricted within-cluster dependence, a sufficient size-negligibility condition for pointwise convergence of the observation-weighted moment is ∑g=1G(NgN)2→0\sum_{g=1}^{G}\left(\frac{N_g}{N}\right)^2\to0, which is implied by max⁡1≤g≤GNgN→0\max_{1\le g\le G}\frac{N_g}{N}\to0, since ∑g=1G(NgN)2≤max⁡1≤g≤GNgN\sum_{g=1}^{G}\left(\frac{N_g}{N}\right)^2\le\max_{1\le g\le G}\frac{N_g}{N}. If this negligibility fails severely, the observation-weighted estimating equation may not even obey the pointwise WLLN needed for consistency.

Even when the law-of-large-numbers restriction is satisfied, Gaussian inference additionally requires control of cluster contributions relative to the covariance normalisation. A familiar sufficient size-negligibility benchmark is motivated by the bound ∥V[NP^Nψ(W;θo)]∥≤C∑g=1GNg2N\left\|\mathbb{V}\left[\sqrt{N}\widehat{\mathbb{P}}_N\psi(W;\theta_o)\right]\right\|\le C\sum_{g=1}^{G}\frac{N_g^2}{N}, which may diverge as N→∞N\to\infty. The condition

N−1max⁡1≤g≤GNg2→0N^{-1}\max_{1\le g\le G}N_g^2\to0

is a sufficient size-negligibility benchmark for standard CLT arguments under suitable moment and covariance conditions. When this condition fails, the estimator may still be consistent in favourable designs, but conventional Gaussian, cluster-robust variance, and bootstrap approximations can be invalid.

Large-sample justifications for clustered estimators typically rely on cluster-size restrictions of precisely this form. [8 Assumption 2], for example, imposes

References

  1. [1]Akhtari, Mitra, Diana Moreira, and Laura Trucco. “Political Turnover, Bureaucratic Turnover, and the Quality of Public Services”. American Economic Review 112.2 (Feb. 2022), pp. 442–93. URL: https://www.aeaweb.org/articles?id=10.1257/aer.20171867.DOI
  2. [2]Andrews, Donald WK. “Generic uniform convergence”. Econometric Theory 8.2 (1992), pp. 241–257.DOI
  3. [3]Chernozhukov, Victor, Iván Fernández-Val, and Alfred Galichon. “Quantile and Probability Curves Without Crossing”. Econometrica 78.3 (2010), pp. 1093–1125.
  4. [4]Chernozhukov, Victor, Iván Fernández-Val, and Blaise M 영어. “Inference on counterfactual distributions”. Econometrica 81.6 (2013), pp. 2205–2268.
  5. [5]Chiang, Harold D, Yuya Sasaki, and Yulong Wang. “On the Inconsistency of Cluster-Robust Inference and How Subsampling Can Fix It”. arXiv preprint arXiv:2308.10138v3 (2023). URL: https://arxiv.org/abs/2308.10138v3.
  6. [6]Frandsen, Brigham R., Markus Frölich, and Blaise Melly. “Quantile Treatment Effects in the Regression Discontinuity Design”. Journal of Econometrics 168.2 (2012), pp. 382–395.DOI
  7. [7]Hagemann, Andreas. “Cluster-robust bootstrap inference in quantile regression models”. Journal of the American Statistical Association 112.517 (2017), pp. 446–456.
  8. [8]Hansen, Bruce E and Seojeong Lee. “Asymptotic theory for clustered samples”. Journal of Econometrics 210.2 (2019), pp. 268–290.
  9. [9]Hoffman, Elaine B., Pranab K. Sen, and Clarice R. Weinberg. “Within-cluster resampling”. Biometrika 88.4 (2001), pp. 1121–1134.DOI
  10. [10]Kosorok, Michael R. Introduction to Empirical Processes and Semiparametric Inference. Springer Science & Business Media, 2007.DOI
  11. [11]Kosorok, Michael R. “Bootstraps of Sums of Independent but Not Identically Distributed Stochastic Processes”. Journal of Multivariate Analysis 84.2 (2003), pp. 299–318.DOI
  12. [12]MacKinnon, James G, Morten Ørregaard Nielsen, and Matthew D Webb. “Fast and reliable jackknife and bootstrap methods for cluster-robust inference”. Journal of Applied Econometrics 38.5 (2023b), pp. 671–694.arxiv.org/abs/2301.04527
  13. [13]Nevalainen, Jaakko, Somnath Datta, and Hannu Oja. “Inference on the Marginal Distribution of Clustered Data with Informative Cluster Size”. Statistical Papers 55.1 (2014), pp. 71–92.pubmed.ncbi.nlm.nih.gov/25878396
  14. [14]Newey, Whitney K. and James L. Powell. “Asymmetric least squares estimation and testing”. Econometrica: Journal of the Econometric Society 55 (1987). Publisher: JSTOR, pp. 819–847. URL: https://www.jstor.org/stable/1911031.DOI
  15. [15]Pollard, David. Empirical Processes: Theory and Applications. Vol. 2. NSF-CBMS Regional Conference Series in Probability and Statistics. Hayward, CA and Alexandria, VA: Institute of Mathematical Statistics and American Statistical Association, 1990. ISBN: 0-940600-16-1.

Paper details

vlnc:W_v6e86shv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Economics, Econometrics and Finance

ChatGPT, Codex

View statement
Looking for related papers…
Contents