QUESTION
Okay... I'm doing my dissertation on how the EMU affects trade between Britain and adopting/non-adopting countries. (Basically looking at whether there is trade diversion.) The thing is, most economic models use log-linear models due to ease and the fact a lot of economists don't have a good enough grounding in mathematics/statistical theory to deal with other models. However recent work by Santos Silva and Tenreyro has found a PPML estimation method is much more apppropriate and reduces bias in the estimates.
However there has only been two instances of using this new method for economic modelling.
In your dissertation, would you be penalised for using a new/non-traditional method of analysis if it has a strong mathematical basis? Or is it more appropriate to simply stick with the traditional method and highlight the limitations of it, suggesting this new technique as one they may improve estimates?