Hi everyone,
I'm analyzing the existence of knowledge spillovers from Country A to the other countries in my panel dataset.
My initial idea is to investigate whether the growth rate of country i (where i different from A) is influenced by technological advancements in Country A.
The dataset consists of both a large N and T.
Since macro panels are a relatively new area for me, I would greatly appreciate your suggestions on relevant literature to guide my research on phenomena like knowledge spillovers.
The simplest approach that comes to mind is using the Dumitrescu and Hurlin test via the xtgcause command to determine whether technological improvements in Country A Granger-cause growth rates in other countries. However, I'm concerned about the potential cross-correlation among the idiosyncratic errors, which is likely present due to the nature of technology diffusion. Would this cross-correlation pose a problem for the validity of the Dumitrescu and Hurlin test?
If the Dumitrescu and Hurlin test is unsuitable, would you recommend testing for co-integration instead? Although co-integration analysis provides valuable insights, it doesn't allow for conclusions about causality.
Thank you for your advice!
I'm analyzing the existence of knowledge spillovers from Country A to the other countries in my panel dataset.
My initial idea is to investigate whether the growth rate of country i (where i different from A) is influenced by technological advancements in Country A.
The dataset consists of both a large N and T.
Since macro panels are a relatively new area for me, I would greatly appreciate your suggestions on relevant literature to guide my research on phenomena like knowledge spillovers.
The simplest approach that comes to mind is using the Dumitrescu and Hurlin test via the xtgcause command to determine whether technological improvements in Country A Granger-cause growth rates in other countries. However, I'm concerned about the potential cross-correlation among the idiosyncratic errors, which is likely present due to the nature of technology diffusion. Would this cross-correlation pose a problem for the validity of the Dumitrescu and Hurlin test?
If the Dumitrescu and Hurlin test is unsuitable, would you recommend testing for co-integration instead? Although co-integration analysis provides valuable insights, it doesn't allow for conclusions about causality.
Thank you for your advice!
