-----Original Message----- From: gretl-users-bounces(a)lists.wfu.edu [mailto:gr etl-users- bounces(a)lists.wfu.edu] On Behalf Of Tim Nall Sent: Friday, May 16, 2014 9:10 PM To: Gretl list Subject: [Gretl-users] sorry, learning hansl question If you can figure out the iffy psuedo-psuedocode below, I'm trying to run 1 ols regression over 4 different date ranges (so total 4 regressions) with 1 dependnet variable and 1 independant variable.. but I want to run a kpss test on both the depvar and indvar before doing the regression for each date range, then utilize either the variable or its fiorst- differenced version depending on the results of the kpss (at 5%, hopefully). Can this be done in hansl? start_yr = 1855 1860 1865 1870 end_yr = 1894 1899 1904 1909 ## get MyDepVar and myIndVar from file, then for i in start_yr, end_yr smpl start_yr end_yr kpss 3 myDepVar ## and capture results if stationary = True: tempDepVar = MyDepVar else: tempDepVar = d_MyDepvar kpss 3 myIndVar ## and capture results if stationary = True: tempIndVar = myIndVar else: tempIndVar = d_myIndVar ols tempDepVar tempIndVar -- Best regards, Timothy M. Nall Assistant Professor National Quemoy University Kinmen, Taiwan _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

-----Original Message----- From: gretl-users-bounces(a)lists.wfu.edu [mailto:gretl-users- bounces(a)lists.wfu.edu] On Behalf Of Tim Nall Sent: Friday, May 16, 2014 10:05 PM To: Gretl list Subject: Re: [Gretl-users] sorry, learning hansl question Thank you... and can kpss output be captured and used within an if statement to select between two possible dependent variables? TMN On Sat, May 17, 2014 at 10:56 AM, Logan Kelly <logan.kelly(a)uwrf.edu&gt; wrote: > You can do this in two ways. > > First: > > smpl 1855 1894 > loop i = 1..3 > #do stuff > smpl +5 +5 > endloop > > Second: > > string st_yr = "1855 1860 1865 1870" > string ed_yr = "1894 1899 1904 1909" > string st ed > loop i = 1..4 > st = strsplit(st_yr,i) > ed = strsplit(ed_yr,i) > smpl @st @ed > #do stuff > endloop > > The second option should only be used if your periods do not change in a regular way as it is slower. > > Logan > >> -----Original Message----- >> From: gretl-users-bounces(a)lists.wfu.edu [mailto:gr etl-users- >> bounces(a)lists.wfu.edu] On Behalf Of Tim Nall >> Sent: Friday, May 16, 2014 9:10 PM >> To: Gretl list >> Subject: [Gretl-users] sorry, learning hansl question >> >> If you can figure out the iffy psuedo-psuedocode below, I'm trying >> to run 1 ols regression over 4 different date ranges (so total 4 >> regressions) with 1 dependnet variable and 1 independant variable.. >> but I want to run a kpss test on both the depvar and indvar before >> doing the regression for each date range, then utilize either the >> variable or its fiorst- differenced version depending on the results of the kpss (at 5%, hopefully). >> >> Can this be done in hansl? >> >> >> start_yr = 1855 1860 1865 1870 >> end_yr = 1894 1899 1904 1909 >> >> ## get MyDepVar and myIndVar from file, then >> >> for i in start_yr, end_yr >> >> smpl start_yr end_yr >> >> kpss 3 myDepVar ## and capture results >> if stationary = True: >> tempDepVar = MyDepVar >> else: >> tempDepVar = d_MyDepvar >> >> kpss 3 myIndVar ## and capture results >> >> if stationary = True: >> tempIndVar = myIndVar >> else: >> tempIndVar = d_myIndVar >> >> ols tempDepVar tempIndVar >> >> >> >> >> -- >> Best regards, >> Timothy M. Nall >> Assistant Professor >> National Quemoy University >> Kinmen, Taiwan >> _______________________________________________ >> Gretl-users mailing list >> Gretl-users(a)lists.wfu.edu >> http://lists.wfu.edu/mailman/listinfo/gretl-users > > _______________________________________________ > Gretl-users mailing list > Gretl-users(a)lists.wfu.edu > http://lists.wfu.edu/mailman/listinfo/gretl-users -- Best regards, Timothy M. Nall Assistant Professor National Quemoy University Kinmen, Taiwan _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

You can do this in two ways. First: smpl 1855 1894 loop i = 1..3 #do stuff smpl +5 +5 endloop Second: string st_yr = "1855 1860 1865 1870" string ed_yr = "1894 1899 1904 1909" string st ed loop i = 1..4 st = strsplit(st_yr,i) ed = strsplit(ed_yr,i) smpl @st @ed #do stuff endloop The second option should only be used if your periods do not change in a regular way as it is slower. Logan > -----Original Message----- > From: gretl-users-bounces(a)lists.wfu.edu [mailto:gr etl-users- > bounces(a)lists.wfu.edu] On Behalf Of Tim Nall > Sent: Friday, May 16, 2014 9:10 PM > To: Gretl list > Subject: [Gretl-users] sorry, learning hansl question > > If you can figure out the iffy psuedo-psuedocode below, I'm trying to run 1 ols > regression over 4 different date ranges (so total 4 > regressions) with 1 dependnet variable and 1 independant variable.. > but I want to run a kpss test on both the depvar and indvar before doing the > regression for each date range, then utilize either the variable or its fiorst- > differenced version depending on the results of the kpss (at 5%, hopefully). > > Can this be done in hansl? > > > start_yr = 1855 1860 1865 1870 > end_yr = 1894 1899 1904 1909 > > ## get MyDepVar and myIndVar from file, then > > for i in start_yr, end_yr > > smpl start_yr end_yr > > kpss 3 myDepVar ## and capture results > if stationary = True: > tempDepVar = MyDepVar > else: > tempDepVar = d_MyDepvar > > kpss 3 myIndVar ## and capture results > > if stationary = True: > tempIndVar = myIndVar > else: > tempIndVar = d_myIndVar > > ols tempDepVar tempIndVar > > > > > -- > Best regards, > Timothy M. Nall > Assistant Professor > National Quemoy University > Kinmen, Taiwan > _______________________________________________ > Gretl-users mailing list > Gretl-users(a)lists.wfu.edu > http://lists.wfu.edu/mailman/listinfo/gretl-users _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

"Serisously" perhaps is too strong a word here. What we mean is that we are not aware of an algorithm which will give you exact p-values for the KPSS test, so we rely on linear interpolation, banking on the fact that the p-value function is monotone by definition. Certainly, the p-value for 0.403951 is between 10% and 5%, and I would be very surprised if it turned out to be very different from 8%.

On Sat, 17 May 2014, Tim Nall wrote: > OK, this is great so far, but one prob: > > The gretl documentation says that the interpolated p-value for kpss > "should not be taken too seriously". So I have the test value... and > now I want to compare that to the *critical value* at the 5% level ... > I want to compare 0.403951 (which I know how to access) to 0.473 (see > output below): > > T = 40 > Lag truncation parameter = 3 > Test statistic = 0.403951 > > 10% 5% 1% > Critical values: 0.352 0.473 0.719 > Interpolated p-value 0.079 "Serisously" perhaps is too strong a word here. What we mean is that we are not aware of an algorithm which will give you exact p-values for the KPSS test, so we rely on linear interpolation, banking on the fact that the p-value function is monotone by definition. Certainly, the p-value for 0.403951 is between 10% and 5%, and I would be very surprised if it turned out to be very different from 8%. ------------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Scienze Economiche e Sociali (DiSES) Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti ------------------------------------------------------- _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

Ah, no, you wouldn't compare $test to $pvalue.... apples and oranges On Sat, May 17, 2014 at 7:25 PM, Tim Nall <tnall.ling(a)gmail.com&gt; wrote: > The original goal ws this: > > via hansl, feed 1 dependent variable and a list of independent variables to kpss > depending on the output of kpss, use either the variable in its levels > or filtered by first difference > > so can the ols paramaters be built up like a string? or as items in a list? > > str='' > kpss 3 63 > if $testvalue < $pvalue: > str += 63 > else > str += 98 # the first differenced form of 63 > kpss 3 32 > > if $testvalue < $pvalue: > str += 32 > else > str += 67 # the first differenced form of 32 > > ols str > > > On Sat, May 17, 2014 at 4:19 PM, Riccardo (Jack) Lucchetti > <r.lucchetti(a)univpm.it&gt; wrote: >> On Sat, 17 May 2014, Tim Nall wrote: >> >>> OK, this is great so far, but one prob: >>> >>> The gretl documentation says that the interpolated p-value for kpss >>> "should not be taken too seriously". So I have the test value... and >>> now I want to compare that to the *critical value* at the 5% level ... >>> I want to compare 0.403951 (which I know how to access) to 0.473 (see >>> output below): >>> >>> T = 40 >>> Lag truncation parameter = 3 >>> Test statistic = 0.403951 >>> >>> 10% 5% 1% >>> Critical values: 0.352 0.473 0.719 >>> Interpolated p-value 0.079 >> >> >> "Serisously" perhaps is too strong a word here. What we mean is that we are >> not aware of an algorithm which will give you exact p-values for the KPSS >> test, so we rely on linear interpolation, banking on the fact that the >> p-value function is monotone by definition. Certainly, the p-value for >> 0.403951 is between 10% and 5%, and I would be very surprised if it turned >> out to be very different from 8%. >> >> ------------------------------------------------------- >> Riccardo (Jack) Lucchetti >> Dipartimento di Scienze Economiche e Sociali (DiSES) >> >> Università Politecnica delle Marche >> (formerly known as Università di Ancona) >> >> r.lucchetti(a)univpm.it >> http://www2.econ.univpm.it/servizi/hpp/lucchetti >> ------------------------------------------------------- >> _______________________________________________ >> Gretl-users mailing list >> Gretl-users(a)lists.wfu.edu >> http://lists.wfu.edu/mailman/listinfo/gretl-users > > > > -- > Best regards, > Timothy M. Nall > Assistant Professor > National Quemoy University > Kinmen, Taiwan -- Best regards, Timothy M. Nall Assistant Professor National Quemoy University Kinmen, Taiwan

From script output:

From script output:

(The regression is incorrect from econometric point of view, just to give an easily reproducible example) open /usr/local/share/gretl/data/misc/denmark.gdt ols LRM const LRY dw=$dwpval (here Durbin-Watson = 0.298798) > From script output: ? dw=$dwpval Imhof integral gave negative value Error executing script: halting > dw=$dwpval May be it worth to include into the function something like this:  if dw<dw0 (or if  Imhof <=0) or pvalue<p0  print( pvalue<e-N )for some N, say e-16? R ouput for the same dataset (dwtest{lmtest}) DW = 0.2988, p-value < 2.2e-16

To reproduce: open /gretl/POE4data/nls_panel.gdt start (interactive) R session in R:  hours, age,  exper are lost renamed variables schmours=hours schmage=age schmexper=exper are read ok. Seems, there must be variables' counter and a warning message for such situations. The predetermined names sets are different in different packages. So, one doesn't know beforehand. Should be a warning.

head(gretldata)

On Sat, 17 May 2014, Tim Nall wrote: > OK, this is great so far, but one prob: > > The gretl documentation says that the interpolated p-value for kpss > "should not be taken too seriously". It actually says, "should not be taken too literally" (since the p-value is an interpolated approximation). > So I have the test value... and now I want to compare that to the > *critical value* at the 5% level ... I want to compare 0.403951 > (which I know how to access) to 0.473 [...] OK, newly added in CVS and snapshots you have the kpsscrit() function. It takes as arguments the number of observations for the test and a 0/1 flag for absence/presence of a trend in the test. The value returned is a row vector giving the critical values at the 10, 5 and 1 percent levels. Simple illustration of use: <hansl> nulldata 100 setobs 4 1960:1 series y = normal() kpss 4 y --quiet if $test > kpsscrit(100,0)[2] print "significant at 5%" else print "not significant at 5%" endif y = cum(y) kpss 4 y --quiet if $test > kpsscrit(100,0)[2] print "significant at 5%" else print "not significant at 5%" endif </hansl> Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users