A while ago I had written a note about business cycles in India using HP and Hamilton filters. That was in the context of slowdown post demonetisation and GST arguing how Hamilton Filter captured the downslide of the Indian economy better. Then someone suggested that I should write a note on this for a journal. I started working on it with a very able research assistant. Eventually I lost the data as I moved jobs and then time came again to talk about this- as part of a course on Dynamic Macroeconomic Modeling for M.Sc. Economics at SRM University AP. Where and how do I get the data and the all the calculations of deviations from trend for my lecture now?
Well between then and now, there was the AI revolution! ChatGPT came to rescue. I directed the AI bots to access data from FRED, dictated what calculations to be made, which filters to be used, then create graphs-individual and comparative ones using both filters, and then summarise. Of course it required checking the numbers and calculations and some sense of what those numbers should be given the developing country context from past calculations, readings, and then the note I mentioned above. Everything seems to be in order- do let me know if something feels amiss.
Of course, this uses quarterly data from 1999 to 2023 which is before we embarked on the GDP revision exercise that is still underway and we will have to wait till December for the backdated series. I do hope the qualitative results will not change but the magnitudes might. We will have to wait and see.
Till then you can access the short article here, lecture presentation for class here, and raw data, graphs, and calculations here.