We extend the method of indirect inference testing to data that is not filtered and so may be non-stationary. We apply the method to an open economy real businss cycle model on UK data. We review the method using a Monte Carlo experiment and find that it performs accurately and has good power.