Common Correlated Effects Estimation of Heterogeneous Dynamic Panel Quantile Regression Models

Moving towards sustainable environmental development for BRICS: Investigating the asymmetric effect of natural resources on CO2

“…The adoption of the PMG estimator is partly due to the limited number of countries which are analyzed over a limited period of time. With the availability of long panel data, alternative estimators for dynamic panel data models such as dynamic common correlated effects mean group estimator (Chudik and Pesaran (2015)), Quantile common correlated effects mean group estimator (Harding et al, 2020) allowing for slope heterogeneity and cross‐section dependence could be used in futures studies.…”

“…There are also only two recent studies that account for cross‐section dependence to estimate a dynamic model. While one of them, advanced by Chudik and Pesaran (2015), adopts common correlated approach (Pesaran, 2006) to estimate a panel ARDL model, the other (Harding et al, 2020) employs the mean quantile group method for the estimation of a dynamic panel data model. However, both approaches require a large T and large N, which is not the case in the current study.…”

“…The adoption of the PMG estimator is partly due to the limited number of countries which are analyzed over a limited period of time. With the availability of long panel data, alternative estimators for dynamic panel data models such as dynamic common correlated effects mean group estimator(Chudik and Pesaran (2015)), Quantile common correlated effects mean group estimator(Harding et al, 2020) allowing for slope heterogeneity and cross-section dependence could be used in futures studies.Second, the study offers policy recommendations with an unbalanced panel dataset because of the lack of enough data for Russia. Future studies may exclude Russia and/or add Turkey to the dataset to explore the relationship between CO 2 and natural resources.…”

Energy Dependency and Long-Run Growth

“…Regime-changes and threshold models are more fitting in those cases. An interesting future development of the main model could be a Dynamic Panel Quantile Model, as in Harding et al (2020), but adapting it to a panel CS-ARDL framework. This could permit to avoid the ex-ante clusterization of the sample, while maintaining the actual long-run multi-country approach.…”

A regularization approach to common correlated effects estimation

“…Motivation . This research is motivated by two empirical problems associated with the cross‐section average‐augmented procedures: First of all, the well documented unsatisfactory statistical properties of the pooled CCE estimator with more observables than factors (e.g., in Juodis et al., 2021; Karabıyık et al., 2017) call for new methods to be considered to address the underlying shortcomings of the CCE estimator in the linear model. Second, there has been a growing interest in the application of cross‐section average‐augmented models in non‐linear panel data models, for example, binary choice (Boneva & Linton, 2017), quantile (Harding & Lamarche, 2014; Harding et al., 2020), count (Desbordes & Eberhardt, 2019), and non‐linear mean (Hacioglu Hoke & Kapetanios, 2021) models. Some of these setups implicitly assume that there are as many unobserved factors as cross‐section averages.…”