Thesis Type: Postgraduate
Institution Of The Thesis: Istanbul University, Turkey
Approval Date: 2020
Thesis Language: Turkish
Student: Adnan Sevinç
Consultant: Ferda Yerdelen TatoğluAbstract:
In dynamic panel data models, the effect of independent variables on the dependent variable can occur in later periods or span multiple time periods. Dynamic panel data can be divided into three groups as autoregressive panel data models, distributed lag panel data model, autoregressive distributed lag panel data models. This study examines autoregressive dynamic panel data models. In dynamic panel data models, it is important for efficient and unbiased estimators to take into account heterogeneity and cross-sectional dependency as well as the endogeneity problem. In this thesis, in addition to all the estimation methods, which proposed for the estimation of dynamic panel data models in the literature, also derived instrumental variable seemingly unrelated regression (IV-SUR) has been performed for estimation. All estimators that used have been grouped according to whether they take into account unobservable individual-specific effect, endogeneity problem, heterogeneity and cross-sectional dependency. In this study instrumental variable seemingly unrelated regression has been derived which take into account heterogeneity, endogeneity and cross-sectional dependency and IV-SUR has been chosen as the well appropriate method for analyzing energy demand model in this study. Obtained result are economically expected.