Skip to content
DraconiticResearch
Search Sources API
197,768 studies31 sourcesUpdated 56 min ago
Peer-reviewedCC BY-NCEconometric Reviews

Verifying the existence of maximum likelihood estimates for generalized linear models

  • Sergio Correia1
  • Paulo Guimarães2
  • Thomas Zylkin3
  1. 1Research Department, Federal Reserve Bank of Richmond
  2. 2Economics and Research Department, Banco de Portugal and Universidade do Porto
  3. 3Robins School of Business, University of Richmond
Open at sourcePDFDOICite

Abstract

A fundamental problem with nonlinear models is that maximum likelihood estimates are not guaranteed to exist. Though nonexistence is a well known problem in the binary response model literature, it presents significant challenges for other models and is not as well understood in more general settings. These challenges are only magnified for models that feature many fixed effects and other high-dimensional parameters. We address the current ambiguity surrounding this topic by studying the conditions that govern the existence of estimates for (pseudo-)maximum likelihood estimators used to estimate a wide class of generalized linear models (GLMs). We show that some, but not all, of these GLM estimators can still deliver consistent estimates of at least some of the linear parameters when these conditions fail to hold. We also demonstrate how to verify these conditions in models with high-dimensional parameters, such as panel data models with multiple levels of fixed effects. Applying our methods to a gravity model with heterogeneous free trade agreement effects, we show that failing to detect nonexistence can produce misleading numerical estimates.

Related

  1. Restricted Maximum Likelihood Estimation in Generalized Linear Mixed Models
    Luca Maestrini, Francis K. C. Hui, A. H. Welsh
    Statistical SciencePeer-reviewedLicence not statedFull text21 citedStatisticsOpenAlex
  2. Existence and consistency of weighted maximum likelihood estimator for extreme quantile regression
    Lucien M. Vidagbandji, Alexandre Berred, Cyrille Bertelle +1
    HALPreprintCC BY 4.0Full textProbabilityStatisticsHALarXiv
  3. The Geometry of Existence and Uniqueness of Maximum Likelihood Estimation in Categorical Response Models
    Lukas Sablica, Kurt Hornik, Thomas Rusch
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  4. Asymptotic Theory for Restricted Maximum Likelihood Estimators in Generalized Linear Mixed Models
    Zhanzhongyu Gao, Huadong Mo, Ziyang Lyu
    arXivPreprintCC BY 4.0Full textStatisticsarXiv
  5. Strong Consistency and Asymptotic Normality of the Maximum Likelihood Estimator in r-Largest Generalized Extreme-Value Models
    Jihong Park, Yire Shin, Jeong-Soo Park
    arXivPreprintCC BY-NC-ND 4.0Full textStatisticsarXiv

Cite

BibTeXRIS
BibTeX
@article{correia2026verifying,
  title = {{Verifying the existence of maximum likelihood estimates for generalized linear models}},
  author = {Sergio Correia and Paulo Guimarães and Thomas Zylkin},
  journal = {Econometric Reviews},
  year = {2026},
  month = jun,
  doi = {10.1080/07474938.2026.2673989},
  url = {https://doi.org/10.1080/07474938.2026.2673989},
  note = {CC BY-NC},
}
RIS
TY  - JOUR
TI  - Verifying the existence of maximum likelihood estimates for generalized linear models
AU  - Sergio Correia
AU  - Paulo Guimarães
AU  - Thomas Zylkin
PY  - 2026
DA  - 2026/06/03
JO  - Econometric Reviews
DO  - 10.1080/07474938.2026.2673989
UR  - https://doi.org/10.1080/07474938.2026.2673989
AB  - A fundamental problem with nonlinear models is that maximum likelihood estimates are not guaranteed to exist. Though nonexistence is a well known problem in the binary response model literature, it presents significant challenges for other models and is not as well understood in more general settings. These challenges are only magnified for models that feature many fixed effects and other high-dimensional parameters. We address the current ambiguity surrounding this topic by studying the conditions that govern the existence of estimates for (pseudo-)maximum likelihood estimators used to estimate a wide class of generalized linear models (GLMs). We show that some, but not all, of these GLM estimators can still deliver consistent estimates of at least some of the linear parameters when these conditions fail to hold. We also demonstrate how to verify these conditions in models with high-dimensional parameters, such as panel data models with multiple levels of fixed effects. Applying our methods to a gravity model with heterogeneous free trade agreement effects, we show that failing to detect nonexistence can produce misleading numerical estimates.
SN  - 0747-4938
SN  - 1532-4168
ER  -