Christian Bontemps and Nour Meddahi, "Testing Distributional 
Assumptions; a GMM Approach", Journal of Applied Econometrics, Vol. 27, 
No. 6, 2012, pp. 978-1012.

The file test-dis.zip contains four data files and two program files 
(written for Gauss). All files as ASCII files in DOS format. Unix/Linux 
users should use "unzip -a".

1) There are four data files. The contents of each  are
listed below:

i) "exchg.txt": 946 observations of weekday close exchange rates from 
1/10/81 to 28/6/85. U.K. Pound; French Franc; Swiss Franc;Japanese 
Yen.(all versus the U.S. Dollar.) We are grateful to Neil Shephard for 
providing us with these data.

ii)"exchg-vol-5mn.txt": 2449 (from December 1, 1986 through June 30, 
1999) realized variance measures based on the 5 minutes returns of three 
exchange rates. DM-US$; Yen-US$; Yen-DM.

iii)"exchg-vol-30mn.txt": 2449 (from December 1, 1986 through June 30, 
1999) realized variance measures based on the 30 minutes returns of 
three exchange rates. DM-US$; Yen-US$; Yen-DM.

iv)"rvbv.txt": 1369 observations. The two columns are the realized 
variance and the realized bipower variation computed from the 
five-minute returns of the S\&P index cash 1997-2002.

2) There are two program files.

v)"test-garch-stu.prg": a GAUSS file used in Section 6.1. It estimates a 
GARCH(1,1) on the data on exchange rates and tests the Student 
distributional Assumption of the innovation process. It produces the 
output file "exchg-garch-stu.out"

vi)"realvar-invgaus.prg": a GAUSS file used in Section 6.2. It estimates 
the parameters of the inverse Gaussian Distribution on realized variance 
measures and tests this distributional assumption. The variance of the 
moment is estimated with a HAC procedure. The output file produced is 
"realvar-invgaus.out".