Dear Statalisters,
I am trying to perform a LR test using -lrtest- on 2 nested models that are
estimated using -arima-. I am actually doing this in a loop. While everything
else in the loop has gone smoothly so far, one run of the loop produces an
error after -lrtest-: "samples differ". I am pasting below the results of both
models. It is not clear to me how the samples differ. Can someone please help
shed light on this error?
Thanks,
Danielle Ferry
. est r
--------------------------------------------------------------------------------
Model v1
--------------------------------------------------------------------------------
ARIMA regression
Sample: 1998q1 to 2006q1 Number of obs = 33
Wald chi2(13) = 77.52
Log likelihood = 65.20513 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
| OPG
D.cdefJUM | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
cdefJUM |
grRate_per~c |
L2D. | 1.050329 .4670402 2.25 0.025 .1349467 1.965711
corr_ue |
L3. | -.0768849 .0869923 -0.88 0.377 -.2473867 .093617
index_herf |
L3. | 3.187155 7.768851 0.41 0.682 -12.03951 18.41382
grRate_hpi |
L3D. | -1.166311 .6600636 -1.77 0.077 -2.460012 .1273904
index_divX~l |
L3D. | -3.582336 2.017353 -1.78 0.076 -7.536275 .3716037
vol_ue |
L3D. | .286281 .2202227 1.30 0.194 -.1453475 .7179096
chg_ue |
L4. | -.0350073 .039897 -0.88 0.380 -.113204 .0431894
index_empStr |
L4D. | -.018157 .0061664 -2.94 0.003 -.030243 -.0060711
index_hai_~l |
L4D. | -.0033723 .0022559 -1.49 0.135 -.0077938 .0010492
ue |
L4D. | .0137407 .016543 0.83 0.406 -.018683 .0461645
grRate_gsp |
LD. | -1.410911 3.674484 -0.38 0.701 -8.612768 5.790945
time | -.0068417 .0030396 -2.25 0.024 -.0127992 -.0008842
_cons | .7078237 1.26332 0.56 0.575 -1.768239 3.183886
-------------+----------------------------------------------------------------
ARMA |
ar |
L1. | .8342549 .1618278 5.16 0.000 .5170783 1.151432
-------------+----------------------------------------------------------------
/sigma | .0329504 .0087411 3.77 0.000 .0158181 .0500826
------------------------------------------------------------------------------
. est stat
------------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+----------------------------------------------------------------
v1 | 33 . 65.20513 15 -100.4103 -77.96264
v2 | 33 . 66.37169 20 -92.74339 -62.81324
------------------------------------------------------------------------------
. est r
--------------------------------------------------------------------------------
Model v2
--------------------------------------------------------------------------------
ARIMA regression
Sample: 1998q1 to 2006q1 Number of obs = 33
Wald chi2(18) = 94.94
Log likelihood = 66.37169 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
| OPG
D.cdefJUM | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
cdefJUM |
grRate_per~c |
L2D. | 1.009204 .6632982 1.52 0.128 -.2908367 2.309245
corr_ue |
L3. | -.0687796 .1115346 -0.62 0.537 -.2873834 .1498242
index_herf |
L3. | 7.672428 11.63437 0.66 0.510 -15.13052 30.47537
grRate_hpi |
L3D. | -.8767668 .7005048 -1.25 0.211 -2.249731 .4961973
index_divX~l |
L3D. | -3.435669 2.060106 -1.67 0.095 -7.473403 .6020643
vol_ue |
L3D. | .2318363 .328985 0.70 0.481 -.4129626 .8766351
chg_ue |
L4. | -.0657352 .0646834 -1.02 0.310 -.1925123 .0610419
index_empStr |
L4D. | -.0166287 .0064652 -2.57 0.010 -.0293002 -.0039571
index_hai_~l |
L4D. | -.0028197 .002613 -1.08 0.281 -.0079412 .0023017
ue |
L4D. | .0246965 .0246577 1.00 0.317 -.0236316 .0730247
grRate_gsp |
LD. | -.7753115 3.590427 -0.22 0.829 -7.81242 6.261797
grRate_ip |
L2D. | .2559729 1.270736 0.20 0.840 -2.234623 2.746569
homebuy |
L2D. | -.058201 .1256407 -0.46 0.643 -.3044523 .1880503
confGap |
L3D. | .0005927 .0007941 0.75 0.455 -.0009638 .0021493
grRate_lei |
L4D. | -.4626627 1.302644 -0.36 0.722 -3.015798 2.090473
grRate_mon~r |
L4D. | -.0367778 .6501572 -0.06 0.955 -1.311062 1.237507
time | -.0070249 .0034281 -2.05 0.040 -.0137438 -.0003059
_cons | .0928921 1.805949 0.05 0.959 -3.446703 3.632487
-------------+----------------------------------------------------------------
ARMA |
ar |
L1. | .8299204 .1840959 4.51 0.000 .4690991 1.190742
-------------+----------------------------------------------------------------
/sigma | .0318132 .0073969 4.30 0.000 .0173155 .0463109
------------------------------------------------------------------------------
. est stat
------------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+----------------------------------------------------------------
v1 | 33 . 65.20513 15 -100.4103 -77.96264
v2 | 33 . 66.37169 20 -92.74339 -62.81324
------------------------------------------------------------------------------
--
National Bureau of Economic Research <http://www.nber.org>
---------- Original Message -----------
From: "Michael S. Hanson" <[email protected]>
To: [email protected]
Sent: Thu, 15 Jun 2006 15:22:34 -0400
Subject: Re: st: RE: Modifying -irf graph- output
> .
>
> [Note: I have edited my original questions and Scott's reply to
> clarify further comments.]
>
> On Jun 15, 2006, at 12:53 PM, Scott Merryman wrote:
>
> > On Jun 15, 2006, at 12:27 PM, I wrote:
> >>
> >> 1. How does one add a title to the whole plot?
> >>
> >> 2. How does one change the text of and/or remove the legend and
> >> the note?
> >>
> >> 3. How does one change/control the axis labels and titles? In
> >> particular, I want to either change or delete the text that reads
> >> "step" by default beneath the final row of subplots.
> >
> > For 1. and 2. take a look at -byopts- option and the axis title can
> > be added at the end.
> >
> > For example:
> >
> > irf graph irf, irf(test_irf) byopts(title(My Title) note("") ///
> > legend(off)) xtitle(Years) ytitle(My ytitle)
>
> Thanks, Scott -- I had misinterpreted what -byopts- was designed to
> do, so I had not explored that option. -byopts(title())- resolves
> my first question above. I appreciate your or anyone else's input
> on aspects of these remaining questions:
>
> 2. -byopts(legend())- seems to support only a subset of the options
> of -legend-: basically, I have been able to get -off- and -pos()-
> to work, but none of the other options of -legend- I have tried
> (e.g., - order()-, -stack-, -size()-, -symx()-). Is this a
> "feature"? Or am I missing something?
>
> 3. The -xtitle- option (*outside* of -byopts-) addresses in part
> what I was after in the third question. However, -xtitle- only
> applies to the whole final row of subplots, as does -xlabel-. Is
> there a way to modify the x-axes on the other rows? (Or the y-axes
> for any column other than the first?)
>
> 4. An additional question, if I may be so bold: I would like to
> change the subtitles at the top of each subplot in a way that still
> conveys which response and impulse are plotted. For example,
> instead of "test_irf, lconsumption, lincome", I would like the (1,
> 2) subplot to read, say, "Response of log(Y) to a log(C) shock".
> The -subtitle- option will allow one to change the text at the top
> of each subplot, but I haven't figured out how to specify the
> response and impulse variables within this option. Perhaps if it
> were possible to extract specific elements of r(response) and
> r(impulse)...? Also, I had hoped that labeling the variables would
> change what was shown in the subtitles, but (apparently) to no avail.
>
> Thanks again for your assistance.
>
> -- Mike
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
------- End of Original Message -------
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
