Advanced Bayesian Analysis Using Panel Data of European Union Consumption Function
Author: Saba Mumtaz

Over a decade the core objective of macroeconomists has been to model consumption function using strong and robust econometric techniques for sustainable analysis of economic behavior. The main objective of this study is to frame out consumption function under dynamic conditions with unique Bayesian techniques. The foremost cause of this effort is to address a unique panel data modeling approach under Bayesian framework which can nest all classical panel data regression models. This study is an attempt to check out the significance of different ad-hoc econometric techniques as well as advanced Bayesian methods of estimation using same data set. We used panel data of eleven European Union countries over the period of 1970 to 2015. By applying Autoregressive Distributed Lag (ARDL) model and Error Correction Mechanism (ECM) approach we examined that a long run relationship exists between final household consumption expenditures and its determinants. For this study, empirically data does not indicate any significant cross country effect. This study deduced that Hierarchical Bayes via Gibbs sampler is most reasonable technique to model economy‘s consumption behavior. It is an extremely effective technique for panel data model estimation with minimum forecast error and a unique framework which has the good ability to nest all classical panel data regression models as a special case, to get more precise and accurate estimates. Supervisor:- Dr. Asad Zaman

Meta Data

Keywords : ARDL and Error Correction Mechanism, Classical Bayes, Consumption Function, Cross Country Effects, Empirical Bayes, General to Specific Approach, Gibbs Sample, Hierarchical Bayes, Panel Models
Supervisor: Asad Zaman

Related Thesis​