Panel data estimation
Posted: Wed Jul 25, 2012 6:49 am
hi, this is my first encounter with panel data and I really need your help.
So I'm estimating the export flow of country i to 30 countries in 30 years of observations.
The model is as follows:
lnXijt = a1 + a2lnYjt + a3lnYit + a4lnPOPjt +a5lnPOPit + a6lnEXRijt + uijt
where,
Xijt = export flows from country i to country j at year t
Yit = GDP of country i (exporter) at year t
Yjt = GDP of country j (importer) at year t
POPit = population of country i at year t
POPjt = population of country j at year t
EXRijt = exchange rate/exporter's curency units per unit of an importer's currency.
as you might notice, the variables GDPit and POPit are time series because country i is not in the sample (the samples are the countries that receives exports from country i) and it cannot be estimated as panel (Eviews shows near singular matrix message).
how can I solve this problem and proceed the estimation?
your help is really really appreciated.
sara
So I'm estimating the export flow of country i to 30 countries in 30 years of observations.
The model is as follows:
lnXijt = a1 + a2lnYjt + a3lnYit + a4lnPOPjt +a5lnPOPit + a6lnEXRijt + uijt
where,
Xijt = export flows from country i to country j at year t
Yit = GDP of country i (exporter) at year t
Yjt = GDP of country j (importer) at year t
POPit = population of country i at year t
POPjt = population of country j at year t
EXRijt = exchange rate/exporter's curency units per unit of an importer's currency.
as you might notice, the variables GDPit and POPit are time series because country i is not in the sample (the samples are the countries that receives exports from country i) and it cannot be estimated as panel (Eviews shows near singular matrix message).
how can I solve this problem and proceed the estimation?
your help is really really appreciated.
sara