Growth at risk
Posted: Thu Feb 21, 2019 5:36 am
Good morning,
I am trying to replicate the approach used by the IMF to project the distribution of GDP growth a few quarters ahead. It entails
1/ running quantile regressions (at the 5th 25th 50th 75th and 95th %ile ) of GDP growth on a lag and an index of financial conditions
2/ make a forecast of the points of the distribution
3/ derive the whole distributon
(the procedure is explained in detail here https://www.newyorkfed.org/medialibrary ... .pdf?la=en)
steps 1 and 2 are straightforward, I wonder wheter #3 is feasible. The original paper uses a specific t-distribution, I'd be happy also with a standard t.
Thanks a lot
Paolo
I am trying to replicate the approach used by the IMF to project the distribution of GDP growth a few quarters ahead. It entails
1/ running quantile regressions (at the 5th 25th 50th 75th and 95th %ile ) of GDP growth on a lag and an index of financial conditions
2/ make a forecast of the points of the distribution
3/ derive the whole distributon
(the procedure is explained in detail here https://www.newyorkfed.org/medialibrary ... .pdf?la=en)
steps 1 and 2 are straightforward, I wonder wheter #3 is feasible. The original paper uses a specific t-distribution, I'd be happy also with a standard t.
Thanks a lot
Paolo