﻿Explanation of Programs and Data for "Do Macro-Forecasters Herd" by Michael P. Clements. 28.8.2017

The programs are written in Gauss 14.

Herd.prg is the main program file.
It reads in data from:
pQvQd.csv
ROUTPUTQvQd.csv.
These are .csv versions of the real-time data sets taken from:
https://www.philadelphiafed.org/research-and-data/real-time-center/real-time-data/


It also reads  in forecast data from:
Individual_PGDP.csv;
Individual_RGDP.csv;
These are .csv versions of the individual-level forecast data taken from:
https://www.philadelphiafed.org/research-and-data/real-time-center/survey-of-professional-forecasters/

The program herd.prg includes procedures which apply the tests described in the paper.

Running herd.prg generates output files. To produce the tables reported in the paper, run tab_gen.prg.

As set up, the programs assume all the files are in a directory c:\jmcb_16-171

The program herd.prg will need to be run a number of times (followed by tab_gen.prg) to produce all the results reported in the paper,
with different settings of the switches to use `current consensus’ or `lagged consensus’, and the minimum number of forecasts per individual.
[These variables are "lagged_consensus" and "min_nobs" and are in the first few lines of herd.prg]
