Willinger, Taqqu, Teverovsky(1999)

Use this forum for posting example programs or short bits of sample code.
Post Reply
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Willinger, Taqqu, Teverovsky(1999)

Post by TomDoan »

This is a replication file for Willinger, Taqqu and Teverovsky(1999), "Stock Market Prices and Long-Range Dependence", Finance and Stochastics, vol 3, pp 1-13. This includes use of the @RSStatistic and @Hurst procedures and estimation of an ARFIMA model.
ew.rpf
Program file
(2.71 KiB) Downloaded 861 times
d-vwew.dat
Data file
(366.6 KiB) Downloaded 861 times
prashantj
Posts: 88
Joined: Sun Apr 11, 2010 2:56 am

Re: Willinger, Taqqu, Teverovsky(1999)

Post by prashantj »

Dear Tom,

I have run the code. but the values of H and graphs turn out to be very different. Even graphs do not look like the paper. I attach the output and one sample graph generated by the new code for your reference. Only Fig 1(a)- Trace of EW and Fig 2(b)-equally weighed index log-log plot match with the paper. Previous codes seem to be given better values of H. Value of MA coefficients turn out to be positive 0.40 which was negative -0.40 in the paper.
Looking forward for your reply,
With regards,
Prashant
Attachments
graph1.RGF
(53.31 KiB) Downloaded 878 times
output.RPF
(13.45 KiB) Downloaded 881 times
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: Willinger, Taqqu, Teverovsky(1999)

Post by TomDoan »

The signs of MA coefficients depend upon the parameterization. RATS uses (1+thetaL)u_t while some software uses (1-thetaL)u_t which will give exactly the same magnitude but opposite signs.
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: Willinger, Taqqu, Teverovsky(1999)

Post by TomDoan »

prashantj wrote:Dear Tom,

I have run the code. but the values of H and graphs turn out to be very different. Even graphs do not look like the paper. I attach the output and one sample graph generated by the new code for your reference. Only Fig 1(a)- Trace of EW and Fig 2(b)-equally weighed index log-log plot match with the paper. Previous codes seem to be given better values of H. Value of MA coefficients turn out to be positive 0.40 which was negative -0.40 in the paper.
Looking forward for your reply,
With regards,
Prashant
I fixed an error with that revised @HURST. Download the newest one.
Post Reply