Dear Community,
I am new on this forum and hope to proceed in the correct way and according to rules.

I have been learning about STATA for the last days and nights and ask you to forgive my probable lack of knowledge!

- I am in the process of analyzing the effects of exchange rates (the Swiss franc, CHF) and its effect on exports in the mechanical watch industry.

- To do this, I'd like to complete the analysis with a comparaison of the effect on "all exports" and it is affected compared to "watch exports".

- My country variable contains 10 countries and each country has the data from 2006Q1 to 2017Q4

- I'm using the lag operator with n-1, n-2 and n-2 for both "watch exports" and "all exports"

- I'm using time series, as I think I should


So, I was thinking of doing the panel regression by country individually including each lag, in order to get the coefficients. BUT probably not a good idea.. due to lack of data (45 lines per country)

So doubts arise on the way to regress this data; if I'd like to see how differently the "watch exports" and "all exports" get influenced by the exchange rate...

How would you suggest making best use of this data using a random effect panel regression?

I tried using the command

xtreg Exrate Watchexports watchexports_L1 watchexports_L2 watchexports_L3, re
But the coefficients don't seen to make sense (far too small decimals)

Any insight would be highly appreciated!


Array