Stationarity test in GMM

[Naila Jabeen]

greetings...

I am estimating an equation with time series data that consists of 41 annual observations and applying GMM. The Model is: imports share = f(GDP, average tariff rates, Liberalization dummy, RER, TOT etc)
Some of the variavles are nonstationary. Should i make them stationary before applying GMM? Results are very sensitive to the choice of instruments as well. how can i test which are valid among them?

Best Regards