Page 1 of 1

Problems running Pedroni's FMOLS

Posted: Thu Nov 08, 2012 6:28 am
by basher
Hello,
I am facing problems (i.e., error messages) running Pedroni's FMOLS estimator. Here is the simple code I am using:
***
open data C:\Research\data_rats.txt
calendar(panelobs=30,a) 1980:1
data(format=free,org=columns) 1//1980:01 6//2009:01 country year lrm2 lrnogdp ltbill
*
dec vect[strings] labels(6)
input labels
Bahrain
Kuwait
Oman
Qatar
Saudi Arabia
UAE
*
source C:\Research\panelfm.src
*
@panelfm(print=short, lags=2)
# lrm2 lrnogdp ltbill

***
I am attaching the data file for your convenience. I look forward to hearing from you. Thank you.

Re: Problems running Pedroni's FMOLS

Posted: Thu Nov 08, 2012 8:16 am
by TomDoan
Aren't you getting data file errors? The data file you attached has some garbage characters at the start and produces the errors

## IO9. Invalid input "ÿþ1" on line 0 while processing series COUNTRY entry 0
## IO30. There is no series YEAR on the file
## IO30. There is no series LRM2 on the file
## IO30. There is no series LRNOGDP on the file
## IO30. There is no series LTBILL on the file

With no data, it won't get anywhere. Pull the file into Word or something like it, clean it up and re-save it.

Re: Problems running Pedroni's FMOLS

Posted: Wed Nov 14, 2012 12:21 am
by basher
Thank you. I saved the data in Excel and it is working well. Both the 'fmols' and 'dols' estimators report individual+panel coefficients and their respective t-statistics. Could you kindly suggest the syntax that need to be inserted in the code to obtain the respective standard errors (for both fmols and dols estimators)?

Re: Problems running Pedroni's FMOLS

Posted: Wed Nov 14, 2012 11:29 am
by TomDoan
Rather than changing the existing procedure, you might want to look at just repackaging the information. The standard errors are in the %%ISTDERRS vector. See

http://www.estima.com/forum/viewtopic.php?f=8&t=1467

for an example of displaying coefficients and standard errors.