David:

I don't understand your question and hence unable to provide you with a meaningful answer. Can you clarify what you mean by "ethical", and give examples of what you mean by " ... generalised for countries ..."?

I presume that you are trying to estimate a model of exchange rate determination for various West African countries

Most of your data is probably already available in the International Monetary IFS databank. You might also consult the Pen World table database (https://www.rug.nl/ggdc/productivity/pwt/?lang=en). The Pen World Table time series are converted to a common currency using purchasing power parities rather than exchange rates. This should lead to more comparable data for various economic series. You may need to supplement this data from other sources depending on your assumptions.

One of the assumptions underlying standard panel data estimation (e.g. fixed and random effects) is that the coefficient of each explanatory variable is the same in all countries in the model. This might not be true for a model of exchange rate determination in different countries. Thus panel analysis might not be appropriate. Also, you probably have the problem that the residuals in your model are likely to be correlated across countries, (An economic shock will have similar or opposite effects across countries)

You are now facing two problems

Talk to your supervisor and ask what is expected. I don't know how much time you have to complete this project. Make sure that you do something that is consistent with your knowledge of economics, econometrics, and mathematics, and that can be completed within the available time.