Pith. sign in

REVIEW 4 major objections 3 minor 35 references

Spatial error models with heteroskedastic normal perturbations and joint modeling of mean and variance

T0 review · 4 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Modeling variance with spatial errors yields β estimates at least as efficient as SEM, SAR, and SARAR estimators.

desk verdict The computational method and simulations are worth a look, but the paper's headline efficiency theorem rests on a false matrix expansion and should not be accepted as-is. read the letter →

arxiv 2411.13432 v1 pith:FLD2G3ZM submitted 2024-11-20 stat.ME

J.D. Toloza , O.O. Melo , N.A. Cruz This is my paper · ORCID
classification stat.ME MSC 62F1062J0562M30
keywords variancepredictiongeneralizedadditivemodelsheteroscedasticityspatialregressionmaximumlikelihoodestimationerrormodelschooldesertion
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 a spatial error model in which the error variances are modeled jointly with the regression coefficients, so that spatial dependence and heteroskedasticity appear in one maximum-likelihood framework. The authors claim two theoretical results: the closed-form $\hat{\beta}$ estimator is biased only by errors in estimating $\lambda$ and the variances, and its conditional variance is no larger than the variances of homoscedastic SEM, SAR, SARAR, and robust SARAR estimators. Simulations show that the estimator recovers all coefficients and variance parameters, with bias and variance shrinking as the sample size grows. An application to Colombian municipal school-desertion data illustrates how the model identifies factors that shift both the mean dropout rate and its variability.

What carries the argument

The carrying object is the joint log-likelihood for the heteroskedastic spatial error model, $\ell = -\frac{n}{2}\ln(\pi) - \frac{1}{2}\ln|\Omega| + \ln|B| - \frac{1}{2}(y-X\beta)^\top B^\top \Omega^{-1} B (y-X\beta)$, with $B = I - \lambda W$ and $\Omega$ diagonal with entries $\sigma_i^2 = \exp(z_i^\top \alpha)$. The estimation procedure alternates: given $\Omega$, the closed-form $\hat{\beta}$ in equation (5) is available; given $\beta$ and $\lambda$, the mean and variance are refit with generalized additive models for location, scale, and shape (GAMLSS), and $\lambda$ is updated by maximizing the profile log-likelihood, using a Cholesky decomposition for $\ln|\Omega|$. Lemma 2 is carried by a variance decomposition that rewrites the proposed estimator's conditional variance as the benchmark SEM variance plus nonnegative terms.

What would settle it

Run the paper's simulation design with known $\beta$, $\lambda$, and variance coefficients $\alpha$ on a regular grid; if, across many replications, the conditional variance of $\hat{\beta}$ from equation (5) is ever larger than the smallest variance among the homoscedastic SEM, SAR, SARAR, and robust SARAR estimators, Lemma 2 is false. A separate algebraic check would test the inclusion $C(P_H) \subseteq C(W)$ used in Appendix B with a design matrix and spatial weights chosen so that the inclusion fails.

Watch

Extended reading notes

Core claim

The paper's central claim is Lemma 2: under model (2) and conditions 1--5 of [18], the generalized-least-squares-style estimator $\hat{\beta} = (X^\top \hat{B}^\top \hat{\Omega}^{-1} \hat{B} X)^{-1} X^\top \hat{B}^\top \hat{\Omega}^{-1} \hat{B} y$ satisfies $\operatorname{Var}(\hat{\beta} \mid \hat{\lambda}, \hat{\Omega}) \le \min[\operatorname{Var}(\hat{\beta}_1), \operatorname{Var}(\hat{\beta}_2), \operatorname{Var}(\hat{\beta}_3), \operatorname{Var}(\hat{\beta}_4)]$, where the four comparators are the homoscedastic SEM, SAR, SARAR, and robust SARAR estimators. Lemma 1 gives the companion bias statement: the estimator is not unbiased, and its conditional bias is of order $\max_i(-\hat{\lambda}/\hat{\sigma}_i + \lambda/\sigma_i)$, so bias grows when the variance estimates are poor even if $\hat{\lambda}$ is good. Together these results are offered as the theoretical advantage of the proposed model over the usual spatial econometric models.

Load-bearing premise

The proof depends on two unverified conditions: the estimated spatial parameter must differ from the true value by exactly the same amount in the proposed model and in the homoscedastic SEM, and a geometric inclusion relating the robust SARAR estimator's projection matrix to the spatial weights matrix must hold.

Editorial extensions

If this is right

  • Under heteroskedastic spatial-error data, the proposed estimator should yield tighter confidence intervals for regression coefficients than SEM, SAR, SARAR, and robust SARAR estimators.
  • The variance equation's coefficients $\alpha$ are estimated along with $\beta$ and $\lambda$, so the model can identify which covariates drive the error variance.
  • Simulation evidence indicates that parameter recovery improves with sample size, with bias approaching zero at the largest grid size reported.
  • In the school-desertion application, municipalities with higher victimization and homicide rates show higher dropout rates, while better education-index scores reduce dropout; the variance equation responds in opposite directions to education and homicide.

Reading between the lines

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

  • If the column-space inclusion used for the robust SARAR comparison cannot be justified, the efficiency claim against that particular estimator would need a different proof, while the SEM and SAR comparisons might still hold under the equal-$\lambda$-error assumption.
  • The same mean-variance alternating scheme could plausibly be applied to spatial lag, Durbin, or SARAR structures, since the separation between $\beta$ and $\alpha$ updates does not depend on the error-only spatial specification.
  • A test the paper does not run would compare out-of-sample prediction intervals from the proposed model with those from homoscedastic SEM; if the variance model is right, interval coverage should be better in heteroskedastic spatial data.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The paper proposes a spatial error model (SEM) with heteroskedastic normal disturbances, where the variance is modeled through a GAMLSS-type specification and estimation proceeds by an iterative maximum-likelihood algorithm. The authors claim two main theoretical results: Lemma 1 gives an order bound on the conditional bias of the weighted least-squares estimator for β, and Lemma 2 asserts that the proposed estimator is at least as efficient as homoskedastic SEM, SAR, SARAR, and heteroskedasticity-robust SARAR estimators. The methodology is evaluated in a simulation study and applied to municipality-level school-desertion data in Colombia.

Significance. The proposal is timely, since spatial econometrics software rarely handles heteroskedastic normal errors in a maximum-likelihood framework, and the joint modeling of mean and variance is practically useful. The simulation study and the empirical application are clearly described and the authors provide R code, which is a strength. However, the paper's central theoretical contribution is Lemma 2, and the proof of Lemma 2 contains a mathematically incorrect expansion and an invalid column-space inclusion. Because the claimed efficiency dominance is load-bearing for the paper's message, the theoretical result is not established. If repaired, the algorithm and simulations could still be valuable as a practical contribution, but the current manuscript cannot be recommended for publication.

major comments (4)
  1. [Appendix B] The proof of Lemma 2 relies on the identity B^{-1} - Bhat^{-1} = sum_{j=0}^\infty (\lambda - \hat\lambda)^j W^j. This identity is false in general. Since B^{-1} = sum_k \lambda^k W^k and Bhat^{-1} = sum_k \hat\lambda^k W^k, the difference is sum_k (\lambda^k - \hat\lambda^k) W^k, which is not equal to the stated expansion. The subsequent variance decomposition Var(\hat\beta|\hat\lambda) = (\tilde X^T \Omega^{-1} \tilde X)^{-1} + R_1(\lambda - \hat\lambda) is derived from this incorrect premise, so the comparison with Var(\hat\beta_1|\hat\lambda) does not follow.
  2. [Appendix B] The comparison with the robust SARAR estimator rests on the claim that C(P_H) \subset C(W), justified by item 2) and Theorem 12.3.4 of Harville. This is not valid: P_H projects onto C(H) = C([X, WX]), which is not generally contained in C(W); indeed WX can lie outside C(W). Consequently, the conclusion Var(\hat\beta_4|\hat\lambda, \hat\rho) \ge Var(\hat\beta|\hat\lambda) is unsupported.
  3. [Lemma 2 and Appendix B] Lemma 2 is stated conditionally on \hat\lambda and \hat\Omega, but the proof in Appendix B computes variances using the true \Omega and never uses \hat\Omega. The variance expressions such as Var(\hat\beta|\hat\lambda) = (\tilde X^T \Omega^{-1} \tilde X)^{-1} + R_1(\lambda - \hat\lambda) condition only on \hat\lambda. The mismatch between the stated conditioning set and the one used in the proof makes the lemma's statement ambiguous and unproven. The simulations in Section 3 are also unconditional and therefore do not provide direct evidence for the conditional claim.
  4. [Section 2, information matrix] The expression I_{\lambda\lambda} = tr(WB^{-1})^2 + tr(\Omega(WB^{-1})^T \Omega^{-1}(WB^{-1})) is ambiguous: tr(WB^{-1})^2 could be read as [tr(WB^{-1})]^2, whereas the second derivative of ln|B| involves tr((WB^{-1})^2). The intended meaning should be stated explicitly, and the derived entries of the information matrix should be checked against standard results.
minor comments (3)
  1. [Table 2] In the variance part of Table 2, the rows labelled \hat\alpha_2 appear twice with different values, and there is no row for \hat\alpha_3; this is likely a typographical error that should be corrected.
  2. [Section 2, Lemma 1] The proof of Lemma 1 introduces D_{\hat\lambda,\hat\Omega} and Q, but the final order bound O(max_i(-\hat\lambda/\hat\sigma_i + \lambda/\sigma_i)) is not derived in detail; the dependence on the inversion formula for N+M should be spelled out.
  3. [Section 2, Lemma 2] The lemma refers to conditions 1)-5) of [18] without restating them; a self-contained statement would help the reader verify whether the regularity conditions are met in the simulations.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the estimator and both lemmas are derived from the likelihood and compared with simulated benchmarks, not from a fitted target or a self-citation chain.

full rationale

The paper's claimed derivation chain is not circular. The proposed estimator β̂ in equation (5) is derived as the closed-form maximizer of the Gaussian log-likelihood for model (2), with score equations given just before it. Lemma 1 is proven in Appendix A by expanding (N+M)^{-1} and taking expectations under ε∼N(0,Ω); Lemma 2 is attempted in Appendix B by comparing Var(β̂|λ̂,Ω̂) with variance expressions for SEM, SAR, and robust SARAR estimators. These are algebraic comparisons, not definitions of the target result in terms of itself. The only self-reference is the statement that the GAMLSS-based estimation method is 'adapted from the work of Toloza Delgado [24] for SAR models'; that citation is not load-bearing for Lemma 1 or Lemma 2, since the algorithm and likelihood equations are fully written out and the lemmas do not invoke [24]. No parameter is fitted to the data and then relabeled a prediction; the simulations compare models on the same generated datasets. The proof of Lemma 2 does contain serious non-circular errors: the expansion B^{-1}-Bhat^{-1}=sum (λ−λ̂)^j W^j is false (the correct coefficients are λ^j−λ̂^j), and the final robust-SARAR comparison uses an unproved column-space inclusion C(P_H)⊂C(W). These are mathematical validity defects, not circularity, and so they do not raise the circularity score.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

No free parameters or invented entities are introduced. The main assumptions are the standard spatial econometrics setup, the exponential variance model, and two unproven premises in Lemma 2: the equality of lambda estimation error across models and a column-space inclusion.

assumptions (6)
  • domain assumption Spatial weights matrix W is known and fixed.
    Section 2, equation (1): W is treated as exogenous and known; all inference conditions on it.
  • domain assumption Errors are normally distributed with diagonal covariance Omega, where log(sigma_i^2) = Z_i alpha.
    Section 2, model definition: epsilon ~ N(0, Omega) with Omega_ii = exp(Z_i alpha).
  • ad hoc to paper The iterative GAMLSS algorithm converges to the profile maximum likelihood estimate.
    Section 2, steps 1-8: no convergence theorem is provided; the paper asserts the algorithm estimates all parameters jointly.
  • domain assumption Conditions 1)-5) of Arraiz et al. [18] hold.
    Lemma 2 statement: needed for the regularity of the GMM estimators that the proposed estimator is compared against.
  • ad hoc to paper In Lemma 2, lambda - lambda_hat is the same for the proposed model as for the homoscedastic SEM.
    Appendix B: 'Assuming that the value lambda - lambda_hat is the same for the proposed model as for the homoscedastic SEM'.
  • ad hoc to paper The column-space inclusion C(P_H) subset of C(W) follows from Harville's theorems.
    Appendix B: used to compare the robust SARAR estimator, but the derivation is not justified and appears incorrect.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spatial error models with heteroskedastic normal perturbations and joint modeling of mean and variance." pith.science (2026). https://pith.science/paper/FLD2G3ZM

@misc{pith2026241113432,
  author       = {Pith},
  title        = {Pith review of: Spatial error models with heteroskedastic normal perturbations and joint modeling of mean and variance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FLD2G3ZM}},
  note         = {Machine review of arXiv:2411.13432}
}
read the original abstract

This work presents the spatial error model with heteroskedasticity, which allows the joint modeling of the parameters associated with both the mean and the variance, within a traditional approach to spatial econometrics. The estimation algorithm is based on the log-likelihood function and incorporates the use of GAMLSS models in an iterative form. Two theoretical results show the advantages of the model to the usual models of spatial econometrics and allow obtaining the bias of weighted least squares estimators. The proposed methodology is tested through simulations, showing notable results in terms of the ability to recover all parameters and the consistency of its estimates. Finally, this model is applied to identify the factors associated with school desertion in Colombia.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages

  1. [1]

    Geographical Analysis 54(3), 488–518 (2022)

    Bivand, R.: R packages for analyzing spatial data: A comparative case study with areal data. Geographical Analysis 54(3), 488–518 (2022)

    work page 2022

  2. [2]

    Journal of regional science 44(1), 43–73 (2004)

    L´ opez-Bazo, E., Vay´ a, E., Artis, M.: Regional externalities and growth: evidence from european regions. Journal of regional science 44(1), 43–73 (2004)

    work page 2004

  3. [3]

    Journal of Economic Dynamics and Control 48, 229–245 (2014)

    Basile, R., Durb´ an, M., M ´ ınguez, R., Montero, J.M., Mur, J.: Modeling regional economic dynamics: Spatial dependence, spatial heterogeneity and nonlinearities. Journal of Economic Dynamics and Control 48, 229–245 (2014)

    work page 2014

  4. [4]

    Spatial economic analysis 15(4), 392–412 (2020)

    Martini, B., Giannini, M.: Regional wage and productivity in italy: a spatio- temporal analysis. Spatial economic analysis 15(4), 392–412 (2020)

    work page 2020

  5. [5]

    Journal of Geographical Systems 20(1), 27–55 (2018)

    Montero, J., M ´ ınguez, R., Fern´ andez, G.: Housing price prediction: parametric versus semi-parametric spatial hedonic models. Journal of Geographical Systems 20(1), 27–55 (2018)

    work page 2018

  6. [6]

    Ingenier ´ ıa y ciencia17(34), 23–52 (2021)

    Delgado, J.T., Mart ´ ınez, O.O.M., Romero, J.A.: Determinants of new housing prices in bogot´ a for 2019: an approach through a semiparametric spatial regression model. Ingenier ´ ıa y ciencia17(34), 23–52 (2021)

    work page 2021

  7. [7]

    Springer, ??? (2023)

    Canch´ e, M.S.G.: Spatial Socio-econometric Modeling (SSEM): A Low-code Toolkit for Spatial Data Science and Interactive Visualizations Using R. Springer, ??? (2023)

    work page 2023

  8. [8]

    CRC Press, ??? (2018)

    Plant, R.E.: Spatial Data Analysis in Ecology and Agriculture Using R. CRC Press, ??? (2018)

    work page 2018

Show all 35 references
  1. [9]

    Ecological Monographs 88(1), 36–59 (2018)

    Ver Hoef, J.M., Peterson, E.E., Hooten, M.B., Hanks, E.M., Fortin, M.-J.: Spatial autoregressive models for statistical inference from ecological data. Ecological Monographs 88(1), 36–59 (2018)

    2018

  2. [10]

    Springer, ??? (1988)

    Anselin, L.: Spatial Econometrics: Methods and Models. Springer, ??? (1988). Dordrecht

    1988

  3. [11]

    Available at SSRN 1625588 (2010)

    Yokoi, T.: Efficient maximum likelihood estimation of spatial autoregressive mod- els with normal but heteroskedastic disturbances. Available at SSRN 1625588 (2010)

    2010

  4. [12]

    Comprehensive R Archive Network, Version, 05–83 (2005)

    Bivand, R., Bernat, A., Carvalho, M., Chun, Y., Dormann, C., Dray, S., Hal- bersma, R., Lewin-Koh, N., Ma, J., Millo, G., et al.: The spdep package. Comprehensive R Archive Network, Version, 05–83 (2005). Guelph

    2005

  5. [13]

    Mathematics 9(11), 1276 (2021) 20

    Bivand, R., Millo, G., Piras, G.: A review of software for spatial econometrics in r. Mathematics 9(11), 1276 (2021) 20

    2021

  6. [14]

    Minguez, M.R., Basile, R., et al.: Package ‘pspatreg’ (2023)

    2023

  7. [15]

    International regional science review 20(1-2), 113–129 (1997)

    LeSage, J.P.: Bayesian estimation of spatial autoregressive models. International regional science review 20(1-2), 113–129 (1997)

    1997

  8. [16]

    Revista Colombiana de Estad ´ ıstica45(2), 341–361 (2022)

    Cepeda-Cuervo, E., Sicacha, J.A.: Spatial econometric models: A bayesian approach. Revista Colombiana de Estad ´ ıstica45(2), 341–361 (2022)

    2022

  9. [17]

    Journal of econometrics 157(1), 53–67 (2010)

    Kelejian, H.H., Prucha, I.R.: Specification and estimation of spatial autoregres- sive models with autoregressive and heteroskedastic disturbances. Journal of econometrics 157(1), 53–67 (2010)

    2010

  10. [18]

    Journal of Regional Science 50(2), 592–614 (2010)

    Arraiz, I., Drukker, D.M., Kelejian, H.H., Prucha, I.R.: A spatial cliff-ord-type model with heteroskedastic innovations: Small and large sample results. Journal of Regional Science 50(2), 592–614 (2010)

    2010

  11. [19]

    Journal of Statistical Software 35, 1–21 (2010)

    Piras, G.: sphet: Spatial models with heteroskedastic innovations in r. Journal of Statistical Software 35, 1–21 (2010)

    2010

  12. [20]

    CRC Press, ??? (2017)

    Lee, Y., Ronnegard, L., Noh, M.: Data Analysis Using Hierarchical Generalized Linear Models with R. CRC Press, ??? (2017). Boca Rat´ on

    2017

  13. [21]

    Regional Research Frontiers-Vol

    Lacombe, D.J., McIntyre, S.G.: Hierarchical spatial econometric models in regional science. Regional Research Frontiers-Vol. 2: Methodological Advances, Regional Systems Modeling and Open Sciences, 151–167 (2017)

    2017

  14. [22]

    CRC Press, ??? (2020)

    Haining, R.P., Li, G.: Modelling Spatial and Spatial-Temporal Data: A Bayesian Approach: A Bayesian Approach. CRC Press, ??? (2020)

    2020

  15. [23]

    Journal of the Royal Statistical Society: Series C (Applied Statistics)54(3), 507–554 (2005)

    Rigby, R., Stasinopoulos, M.: Generalized additive models for location, scale and shape. Journal of the Royal Statistical Society: Series C (Applied Statistics)54(3), 507–554 (2005)

    2005

  16. [24]

    Toloza Delgado, J.D.: Modelaci´ on conjunta de media y varianza en modelos semiparam´ etricos autorregresivos espaciales (2020)

    2020

  17. [25]

    Journal of the Royal Statistical Society: Series C (Applied Statistics) 36(3), 332– 339 (1987)

    Aitkin, M.: Modelling variance heterogeneity in normal regression using glim. Journal of the Royal Statistical Society: Series C (Applied Statistics) 36(3), 332– 339 (1987)

    1987

  18. [26]

    R Foundation for Statistical Computing, Vienna, Austria (2024)

    R Core Team: R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria (2024). R Foundation for Statistical Computing. http://www.R-project.org/

    2024

  19. [27]

    Computational Statistics 36(4), 2563–2590 (2021) 21

    Santi, F., Dickson, M.M., Giuliani, D., Arbia, G., Espa, G.: Reduced-bias estimation of spatial autoregressive models with incompletely geocoded data. Computational Statistics 36(4), 2563–2590 (2021) 21

    2021

  20. [28]

    Revista colombiana de Psiquiatr ´ ıa45, 105–112 (2016)

    G´ omez-Restrepo, C., Mu˜ noz, A.P., Rinc´ on, C.J.: Deserci´ on escolar de adolescentes a partir de un estudio de corte transversal: Encuesta nacional de salud mental colombia 2015. Revista colombiana de Psiquiatr ´ ıa45, 105–112 (2016)

    2016

  21. [29]

    Rom´ an, M.: Factores asociados al abandono y la deserci´ on escolar en am´ erica latina: una mirada en conjunto. REICE. Revista Iberoamericana sobre Calidad, Eficacia y Cambio en Educaci´ on11(2), 33–59 (2013)

    2013

  22. [30]

    Ciudad Paz-ando 12(2) (2019)

    D ´ ıaz, P.A.: Reclutamiento forzado: una cara del juvenicidio en colombia. Ciudad Paz-ando 12(2) (2019)

    2019

  23. [31]

    Chapman and Hall/CRC, ??? (2017)

    Wood, S.: Generalized Additive Models: An Introduction with R. Chapman and Hall/CRC, ??? (2017). Boca Rat´ on

    2017

  24. [32]

    Chapman and Hall/CRC, ??? (2017)

    Stasinopoulos, M.D., Rigby, R.A., Heller, G.Z., Voudouris, V., De Bastiani, F.: Flexible Regression and Smoothing: Using GAMLSS in R. Chapman and Hall/CRC, ??? (2017). Boca Rat´ on

    2017

  25. [33]

    John Wiley & Sons, ??? (2019)

    Magnus, J.R., Neudecker, H.: Matrix Differential Calculus with Applications in Statistics and Econometrics. John Wiley & Sons, ??? (2019)

    2019

  26. [34]

    Springer, ??? (2006)

    Arbia, G.: Spatial Econometrics: Statistical Foundations and Applications to Regional Convergence. Springer, ??? (2006). Berl ´ ın

    2006

  27. [35]

    Taylor & Francis (2006) 22

    Harville, D.A.: Matrix algebra from a statistician’s perspective. Taylor & Francis (2006) 22

    2006

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.