uradf code

[bizi_rats_10]

Hi all,

I have a question about my uradf code (see attachment). Every time when I run it, an error shows up:## SR10. Missing Values And/Or SMPL Options Leave No Usable Data Points. I can not figure out the reason. I checked the loop in the code and source file, everything seems normal. Can any of you help me find why? I appreciate.

Regards,

JC

[TomDoan]

Code: Select all

do i=1,10
source(noecho) G:\uradf.src
@uradf(Criterion=AIC,DET=None) i 1999:1:13 2003:10:15
end do i

do i=1,10
source(noecho) G:\uradf.src
@uradf(Criterion=AIC) i 1999:1:27 2003:10:15
end do i

do i=1,10
@uradf(Criterion=AIC,DET=Trend) i 1999:1:14 2003:10:15
end do i

Don't put the

Code: Select all

SOURCE

instruction inside the loop. Do it once, outside. That is,

source(noecho) G:\uradf.src

Code: Select all

do i=1,10
@uradf(Criterion=AIC,DET=None) i 1999:1:13 2003:10:15
end do i

do i=1,10
@uradf(Criterion=AIC) i 1999:1:27 2003:10:15
end do i

do i=1,10
@uradf(Criterion=AIC,DET=Trend) i 1999:1:14 2003:10:15
end do i

[bizi_rats_10]

Thanks very much for your quick reply. It worked finally!!!

Regards,
JC

[shimarats]

hi,
i have one question about uradf test. i did this code for one variable and i taken this variable is unit root. for stationary i should do first and second difference of variables that used in uradf test.i need give me example code related to this topic. i wait you reply .
thanks so much.

[TomDoan]

@URADF was written many years ago by one of our users. We don't use it in any of our examples, because we have our own @DFUNIT procedure which now has effectively all the features of @URADF. @DFUNIT is used in dozens of textbook and other examples for many different reasons. Sometimes it's merely informational, sometimes it's a preliminary to running a cointegration test, sometimes it's preliminary to doing an ARIMA model. If you accept that there's a unit root, you might want to further test whether the process is I(2) by doing a UR test on the first difference. You also might want to do a specification test on your acceptance by using one of the unit root tests with breaks. What you do with the data after accepting that it has a unit root depends upon what is the goal of your analysis.

[shimarats]

@URADF was written many years ago by one of our users. We don't use it in any of our examples, because we have our own @DFUNIT procedure which now has effectively all the features of @URADF. @DFUNIT is used in dozens of textbook and other examples for many different reasons. Sometimes it's merely informational, sometimes it's a preliminary to running a cointegration test, sometimes it's preliminary to doing an ARIMA model. If you accept that there's a unit root, you might want to further test whether the process is I(2) by doing a UR test on the first difference. You also might want to do a specification test on your acceptance by using one of the unit root tests with breaks. What you do with the data after accepting that it has a unit root depends upon what is the goal of your analysis.

thanks so much

in analysis ,i should get stationary variables.i gotten stationary 8 variables from 10 variables but 2 remain of variables are unit root .i used 5 difference of these variables ,but i can not get their stationary. i donot know what is the problem . i want apply svar method with use of these stationary variables. please guid me .thanks so much

[TomDoan]

I don't know where you got the idea that an SVAR requires stationary variables---it doesn't.

You might want to read the thread

http://www.estima.com/forum/viewtopic.php?f=6&t=1524

However, if differencing a variable that seems to have a unit root still gives you a "non-stationary" series, you may have a structural break. That's the phenomenon that Perron first described---the structural break causes a false acceptance of the unit root.