Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures

We propose a Bayesian nonparametric instrumental variable approach that allows us to correct for endogeneity bias in regression models where the covariate effects enter with unknown functional form. Bias correction relies on a simultaneous equations specification with flexible modeling of the join...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Wiesenfarth, Manuel, Hisgen, Carlos Matías, Kneib, Thomas, Cadarso Suárez, Carmen
Formato: Working Paper
Lenguaje:Inglés
Publicado: Georg-August-Universität Göttingen. Courant Research Centre. Poverty, Equity and Growth 2025
Materias:
Acceso en línea:https://www.econstor.eu/bitstream/10419/90568/1/CRC-PEG_DP_127.pdf
http://repositorio.unne.edu.ar/handle/123456789/57160
Aporte de:
id I48-R184-123456789-57160
record_format dspace
spelling I48-R184-123456789-571602025-10-20T12:18:46Z Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures Wiesenfarth, Manuel Hisgen, Carlos Matías Kneib, Thomas Cadarso Suárez, Carmen Endogeneity Markov Chain Monte Carlo methods Simultaneous credible bands We propose a Bayesian nonparametric instrumental variable approach that allows us to correct for endogeneity bias in regression models where the covariate effects enter with unknown functional form. Bias correction relies on a simultaneous equations specification with flexible modeling of the joint error distribution implemented via a Dirichlet process mixture prior. Both the structural and instrumental variable equation are specified in terms of additive predictors comprising penalized splines for nonlinear effects of continuous covariates. Inference is fully Bayesian, employing efficient Markov Chain Monte Carlo simulation techniques. The resulting posterior samples do not only provide us with point estimates, but allow us to construct simultaneous credible bands for the nonparametric effects, including data-driven smoothing parameter selection. In addition, improved robustness properties are achieved due to the flexible error distribution specification. Both these features are extremely challenging in the classical framework, making the Bayesian one advantageous. In simulations, we investigate small sample properties and an investigation of the effect of class size on student performance in Israel provides an illustration of the proposed approach which is implemented in an R package bayesIV. 2025-08-05T14:09:22Z 2025-08-05T14:09:22Z 2012 Working Paper Wiesenfarth, Manuel, et. al., 2012. Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures. Germany: Georg-August-Universität Göttingen. Courant Research Centre. Poverty, Equity and Growth https://www.econstor.eu/bitstream/10419/90568/1/CRC-PEG_DP_127.pdf http://repositorio.unne.edu.ar/handle/123456789/57160 eng Discussion papers;127 application/pdf Georg-August-Universität Göttingen. Courant Research Centre. Poverty, Equity and Growth
institution Universidad Nacional del Nordeste
institution_str I-48
repository_str R-184
collection RIUNNE - Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
language Inglés
topic Endogeneity
Markov Chain Monte Carlo methods
Simultaneous credible bands
spellingShingle Endogeneity
Markov Chain Monte Carlo methods
Simultaneous credible bands
Wiesenfarth, Manuel
Hisgen, Carlos Matías
Kneib, Thomas
Cadarso Suárez, Carmen
Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
topic_facet Endogeneity
Markov Chain Monte Carlo methods
Simultaneous credible bands
description We propose a Bayesian nonparametric instrumental variable approach that allows us to correct for endogeneity bias in regression models where the covariate effects enter with unknown functional form. Bias correction relies on a simultaneous equations specification with flexible modeling of the joint error distribution implemented via a Dirichlet process mixture prior. Both the structural and instrumental variable equation are specified in terms of additive predictors comprising penalized splines for nonlinear effects of continuous covariates. Inference is fully Bayesian, employing efficient Markov Chain Monte Carlo simulation techniques. The resulting posterior samples do not only provide us with point estimates, but allow us to construct simultaneous credible bands for the nonparametric effects, including data-driven smoothing parameter selection. In addition, improved robustness properties are achieved due to the flexible error distribution specification. Both these features are extremely challenging in the classical framework, making the Bayesian one advantageous. In simulations, we investigate small sample properties and an investigation of the effect of class size on student performance in Israel provides an illustration of the proposed approach which is implemented in an R package bayesIV.
format Working Paper
author Wiesenfarth, Manuel
Hisgen, Carlos Matías
Kneib, Thomas
Cadarso Suárez, Carmen
author_facet Wiesenfarth, Manuel
Hisgen, Carlos Matías
Kneib, Thomas
Cadarso Suárez, Carmen
author_sort Wiesenfarth, Manuel
title Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
title_short Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
title_full Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
title_fullStr Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
title_full_unstemmed Bayesian nonparametric instrumental variable regression based on penalized splines and Dirichlet process mixtures
title_sort bayesian nonparametric instrumental variable regression based on penalized splines and dirichlet process mixtures
publisher Georg-August-Universität Göttingen. Courant Research Centre. Poverty, Equity and Growth
publishDate 2025
url https://www.econstor.eu/bitstream/10419/90568/1/CRC-PEG_DP_127.pdf
http://repositorio.unne.edu.ar/handle/123456789/57160
work_keys_str_mv AT wiesenfarthmanuel bayesiannonparametricinstrumentalvariableregressionbasedonpenalizedsplinesanddirichletprocessmixtures
AT hisgencarlosmatias bayesiannonparametricinstrumentalvariableregressionbasedonpenalizedsplinesanddirichletprocessmixtures
AT kneibthomas bayesiannonparametricinstrumentalvariableregressionbasedonpenalizedsplinesanddirichletprocessmixtures
AT cadarsosuarezcarmen bayesiannonparametricinstrumentalvariableregressionbasedonpenalizedsplinesanddirichletprocessmixtures
_version_ 1849012467648167936