Munich Personal RePEc Archive

A Two-step System for Hierarchical Bayesian Dynamic Panel Data to deal with Endogeneity Issues, Structural Model Uncertainty, and Causal Relationship

Pacifico, Antonio (2020): A Two-step System for Hierarchical Bayesian Dynamic Panel Data to deal with Endogeneity Issues, Structural Model Uncertainty, and Causal Relationship. Forthcoming in: The European Journal of Health Economics , Vol. NA, No. Health Economics (2021): pp. 1-30.

[img]
Preview
PDF
MPRA_paper_104291.pdf

Download (189kB) | Preview

Abstract

The paper develops a computational method implementing a standard Dynamic Panel Data model with Generalized Method of Moment (GMM) estimators to deal with endogeneity issues, structural model uncertainty, and causal relationship in large and long panel databases. The methodology takes the name of Two-step System Dynamic Panel Data, that combines a first-step Bayesian procedure for selecting the only potential predictors in a static linear regression model with a frequentist second-step procedure for estimating the parameters of a dynamic linear panel data model. An empirical example to the effects of obesity, socioeconomic variables, and individual-specific factors on labour market outcomes among Italian regions is performed. Potential prevention policies and strategies to address key behavioural and diseases risk factors affecting labour market outcomes and social environment are also discussed.

UB_LMU-Logo
MPRA is a RePEc service hosted by
the Munich University Library in Germany.