Please use this identifier to cite or link to this item: https://repositorio.insper.edu.br/handle/11224/4063
Type: Artigo Científico
Title: Bayesian Instrumental Variables: Priors and Likelihoods
Author: Lopes, Hedibert Freitas
Polson, Nicholas G.
Publication Date: 2014
Abstract: Instrumental variable (IV) regression provides a number of statistical challenges due to the shape of the likelihood. We review the main Bayesian literature on instrumental variables and highlight these pathologies. We discuss Jeffreys priors, the connection to the errors-in-the variables problems and more general error distributions. We propose, as an alternative to the inverted Wishart prior, a new Cholesky-based prior for the covariance matrix of the errors in IV regressions. We argue that this prior is more flexible and more robust thanthe inverted Wishart prior since it is not based on only one tightness parameter and therefore can be more informative about certain components of the covariance matrix and less informative about others. We show how prior-posterior inference can be formulated in a Gibbs sampler and compare its performance in the weak instruments case for synthetic as well as two illustrations based on well-known real data.
Keywords (english terms): Angrist-Krueger data
Bayesian learning
Cholesky decomposition
Demand for cigarettes
Errors-in-variables
Fat-tails
Inverted Wishart
IV regression
Language: Inglês
CNPq Area: Ciências Sociais Aplicadas
URI: https://www.tandfonline.com/doi/full/10.1080/07474938.2013.807146
Copyright: O INSPER E ESTE REPOSITÓRIO NÃO DETÊM OS DIREITOS DE USO E REPRODUÇÃO DOS CONTEÚDOS AQUI REGISTRADOS. É RESPONSABILIDADE DOS USUÁRIOS INDIVIDUAIS VERIFICAR OS USOS PERMITIDOS NA FONTE ORIGINAL, RESPEITANDO-SE OS DIREITOS DE AUTOR OU EDITOR.
Appears in Collections:Coleção de Artigos Científicos

Files in This Item:
File Description SizeFormat 
R_Artigo_2014_Bayesian Instrumental Variables_TC.pdfR_Artigo_2014_Bayesian Instrumental Variables_TC1.19 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.