It depends mostly on the subject, a micro analysis won't be a macro one (classical "regression" versus Time Series regression (Baesyan, VARS, etc.)) In the specific case of Moody's, I would go on a time series model with the chosen pertinent output (Is the GDP my goal? Or the public Finance?). When the output is determined I will search to confirm which theoretical model is the most suitable (depending on what we analyze, not all theories are equal to provide a pertinent analysis). *This is a vast (barely too vast) question. I think that you should ask for precision about what your supervisor wants you to determine.