Hi, this is my first post on here so excuse any lapse in technical knowledge or rookie statalist etiquette.
I am doing my Economics dissertation on the impact of competition law stringency (measured 0 to 1 on an index) on R&D levels as a % of GDP. I have an unbalanced panel data set of countries from 1981- 2010. I am using a country- year fixed effects approach to account for any time- invariant country characteristics or time specific characteristics which don't vary by country. I am using the competition law index with 1 year lag.
However there of course could still be some omitted variable bias, so I intend on using an IV. I have considered a few options for this IV (based on available data), such as the budget allocated to each competition law regulator each year as a % of GDP, or the number of staff employed by each country's regulator each year. However I am unsure if these could meet the exclusion restriction.
I have found a similar paper which made use of "Internal instruments" which used lagged a independent variable as an instrument.
(https://repositorio.redinvestigadore...=1&isAllowed=y). I didn't even know this was possible.
Both the budget as a % of GDP and the 5 period lag of the competition law index are valid in the first stage for me, however I am worried that they may not meet the exclusion restriction.
Any help would be greatly appreciated!
I am doing my Economics dissertation on the impact of competition law stringency (measured 0 to 1 on an index) on R&D levels as a % of GDP. I have an unbalanced panel data set of countries from 1981- 2010. I am using a country- year fixed effects approach to account for any time- invariant country characteristics or time specific characteristics which don't vary by country. I am using the competition law index with 1 year lag.
However there of course could still be some omitted variable bias, so I intend on using an IV. I have considered a few options for this IV (based on available data), such as the budget allocated to each competition law regulator each year as a % of GDP, or the number of staff employed by each country's regulator each year. However I am unsure if these could meet the exclusion restriction.
I have found a similar paper which made use of "Internal instruments" which used lagged a independent variable as an instrument.
(https://repositorio.redinvestigadore...=1&isAllowed=y). I didn't even know this was possible.
Both the budget as a % of GDP and the 5 period lag of the competition law index are valid in the first stage for me, however I am worried that they may not meet the exclusion restriction.
Any help would be greatly appreciated!
