I am replicating a 2001 Journal of Economic Perspectives methodology paper on Vector Autoregression by James H. Stock and Adam Watson. The paper can be found here: http://faculty.washington.edu/ezivot/econ584/stck_watson_var.pdf
I'm having a slight problem with the log differencing. In the paper the researchers state that they multiply the ln differenced price levels by 400.
I have no clue why this is done. The price level data is quarterly chain-weighted and seasonally adjusted implicit GDP deflator of the US. I would understand if they would multiply by four to get the annual inflation figures, but this is unclear. Any suggestions?