Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: unconventional lag length in VAR model?
From
Robert A Yaffee <[email protected]>
To
[email protected]
Subject
Re: st: unconventional lag length in VAR model?
Date
Tue, 5 Mar 2013 15:48:44 -0500
Jason,
You'll be hard pressed to conduct many of the multivariate tests
with uneven lag structures.
- Bob
On Tue, Mar 5, 2013 at 3:37 PM, Jason R Franken <[email protected]> wrote:
> Bob,
>
> Thanks for the suggestions. It doesn't seem that varsoc, varlmar, and varwle are able to investigate as complex of lag structures as the procedures described in Hsiao (1979) [http://www.jstor.org/stable/pdfplus/2286972.pdf?acceptTC=true] or Kaylen (1988) [http://www.jstor.org/stable/pdfplus/1241509.pdf].
>
> The varwle procedure is a little less restrictive in that it allows the lag structure to differ across each equation of the VAR, but it implies the same lag structure for each variable within an equation. I want to allow for the possibility that different lags should be excluded for each variable within an equation.
>
> If anyone can offer further insights/suggestions it would be appreciated.
>
> Jason
>
> ----- Original Message -----
> From: "Robert A Yaffee" <[email protected]>
> To: [email protected]
> Sent: Monday, March 4, 2013 12:30:10 PM
> Subject: Re: st: unconventional lag length in VAR model?
>
> Jason,
> The varsoc, varlmar, and varwle are generally used for this purpose.
> Bob Yaffee
>
>
> On Fri, Mar 1, 2013 at 1:48 PM, Jason R Franken <[email protected]> wrote:
>> I want to determine the appropriate lag structure for a VAR of 3 price series - C, F1, and E. A prior study used these variables and determined the structure using Akaiki's Final Prediction Error (FPE), which can be obtained with the below commands.
>>
>> My problem is that the prior study was able to ascertain how the lag length differed for each variable and in each equation (that is 4 lags of each variable in each equation might not be appropriate), and I'm not sure how to investigate this with the below commands. Specifically, the prior study finds (for a shorter time period) that the F1 equation has lag 1 of F1 and lags 1 and 2 of C; the E equation has only lags 1 through 4 of C; and the C equation has only lag1 of F1.
>>
>> Can I examine this by estimating a VAR with commands for seemingly unrelated regression (reg3, sur; suest; sureg)?
>>
>> Thanks in advance,
>> Jason Franken
>>
>> RESULTS:
>> . var C F1 E, lags(1/4)
>>
>> Vector autoregression
>>
>> Sample: 1976q1 2010q3 No. of obs = 139
>> Log likelihood = -1096.43 AIC = 16.33712
>> FPE = 2502.717 HQIC = 16.6717
>> Det(Sigma_ml) = 1425.583 SBIC = 17.16046
>>
>> Equation Parms RMSE R-sq chi2 P>chi2
>> ----------------------------------------------------------------
>> C 13 5.20584 0.5664 181.5755 0.0000
>> F1 13 4.50839 0.6138 220.9116 0.0000
>> E 13 3.64389 0.7331 381.8178 0.0000
>> ----------------------------------------------------------------
>> ------------------------------------------------------------------------------
>> | Coef. Std. Err. z P>|z| [95% Conf. Interval]
>> -------------+----------------------------------------------------------------
>> C |
>> C |
>> L1. | .8522103 .1119234 7.61 0.000 .6328444 1.071576
>> L2. | -.2105517 .1371686 -1.53 0.125 -.4793972 .0582938
>> L3. | .4614619 .1400664 3.29 0.001 .1869368 .735987
>> L4. | -.0850031 .1404678 -0.61 0.545 -.360315 .1903088
>> F1 |
>> L1. | .1405764 .136883 1.03 0.304 -.1277093 .4088621
>> L2. | -.3039644 .1454131 -2.09 0.037 -.5889687 -.01896
>> L3. | .0981597 .1456564 0.67 0.500 -.1873216 .383641
>> L4. | .0081719 .1410511 0.06 0.954 -.2682833 .284627
>> E |
>> L1. | -.1111197 .1855016 -0.60 0.549 -.4746961 .2524566
>> L2. | .0254513 .1867506 0.14 0.892 -.3405732 .3914758
>> L3. | -.2986691 .1875468 -1.59 0.111 -.6662541 .068916
>> L4. | .0016383 .1617178 0.01 0.992 -.3153227 .3185994
>> _cons | 18.8637 4.441675 4.25 0.000 10.15818 27.56923
>> -------------+----------------------------------------------------------------
>> F1 |
>> C |
>> L1. | .728714 .0969286 7.52 0.000 .5387376 .9186905
>> L2. | -.221222 .1187916 -1.86 0.063 -.4540491 .0116052
>> L3. | .5090438 .1213011 4.20 0.000 .271298 .7467896
>> L4. | -.3071865 .1216488 -2.53 0.012 -.5456138 -.0687593
>> F1 |
>> L1. | .3239146 .1185442 2.73 0.006 .0915722 .5562569
>> L2. | .0113203 .1259315 0.09 0.928 -.2355008 .2581414
>> L3. | .0733694 .1261422 0.58 0.561 -.1738648 .3206036
>> L4. | .3028863 .1221539 2.48 0.013 .063469 .5423036
>> E |
>> L1. | -.3247369 .1606491 -2.02 0.043 -.6396034 -.0098704
>> L2. | -.225406 .1617309 -1.39 0.163 -.5423927 .0915807
>> L3. | -.0965215 .1624204 -0.59 0.552 -.4148596 .2218166
>> L4. | -.2100509 .1400518 -1.50 0.134 -.4845473 .0644455
>> _cons | 19.94553 3.846605 5.19 0.000 12.40632 27.48473
>> -------------+----------------------------------------------------------------
>> E |
>> C |
>> L1. | .5893907 .0783421 7.52 0.000 .435843 .7429384
>> L2. | -.3047364 .0960128 -3.17 0.002 -.492918 -.1165548
>> L3. | .3976193 .0980411 4.06 0.000 .2054622 .5897763
>> L4. | .0127413 .0983221 0.13 0.897 -.1799665 .2054491
>> F1 |
>> L1. | .4016804 .0958129 4.19 0.000 .2138907 .5894702
>> L2. | -.3147535 .1017836 -3.09 0.002 -.5142457 -.1152614
>> L3. | -.1558412 .1019539 -1.53 0.126 -.3556671 .0439848
>> L4. | .0926355 .0987304 0.94 0.348 -.1008726 .2861435
>> E |
>> L1. | -.023602 .129844 -0.18 0.856 -.2780915 .2308876
>> L2. | .2505328 .1307183 1.92 0.055 -.0056704 .5067359
>> L3. | -.0338416 .1312756 -0.26 0.797 -.2911371 .2234539
>> L4. | -.1952189 .1131963 -1.72 0.085 -.4170795 .0266417
>> _cons | 12.52217 3.109002 4.03 0.000 6.428634 18.6157
>> ------------------------------------------------------------------------------
>>
>> . varsoc
>>
>> Selection order criteria
>> Sample: 1976q1 2010q3 Number of obs = 139
>> +---------------------------------------------------------------------------+
>> |lag | LL LR df p FPE AIC HQIC SBIC |
>> |----+----------------------------------------------------------------------|
>> | 0 | -1236.7 11201.5 17.8374 17.8632 17.9008 |
>> | 1 | -1146.61 180.18 9 0.000 3488.08 16.6707 16.7736 16.924* |
>> | 2 | -1125.79 41.648 9 0.000 2942.99 16.5006 16.6807 16.9439 |
>> | 3 | -1111.73 28.119 9 0.001 2737.7 16.4278 16.6851 17.0611 |
>> | 4 | -1096.43 30.599* 9 0.000 2502.72* 16.3371* 16.6717* 17.1605 |
>> +---------------------------------------------------------------------------+
>> Endogenous: C F1 E
>> Exogenous: _cons
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/faqs/resources/statalist-faq/
>> * http://www.ats.ucla.edu/stat/stata/
>
>
>
> --
> Robert A. Yaffee, Ph.D.
> Research Professor
> Silver School of Social Work
> New York University
>
> Biosketch: http://homepages.nyu.edu/~ray1/Biosketch2009.pdf
>
> CV: http://homepages.nyu.edu/~ray1/vita.pdf
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
--
Robert A. Yaffee, Ph.D.
Research Professor
Silver School of Social Work
New York University
Biosketch: http://homepages.nyu.edu/~ray1/Biosketch2009.pdf
CV: http://homepages.nyu.edu/~ray1/vita.pdf
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/