How to account for country business cycles in a panel analysis

6 days ago

My data is in panel format (countries by years): T=33 N=20 (or in another case 100).

The dependent variable is the unemployment rate.

I am interested in controlling for country-specific business cycles. The argument I face is that I would have to average across periods (i.e. taking 5-year averages of all my variables).

Are there other alternatives? Am I right in thinking this is a very crude treatment of the data and underlying information?

Unfortunately, I was not able to find literature which describes the best practices.

I have data on a small sample of OECD countries and a larger sample of very heterogeneous countries (including developing countries). I will include time fixed effects which should account for similar shocks across regions as well as the output-gap which should account for recessions in general.

Evidently, I can, however, not include time-country fixed effects.

Any help is highly appreciated.