Bookmark and Share

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: bootstrap vector autoregression (var)?


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: bootstrap vector autoregression (var)?
Date   Thu, 26 Jan 2012 12:26:49 -0500

The bootstrap methods for dependent data, such as VAR, are not
trivial. Typically, block bootstrap schemes are used where short
contiguous blocks (say with four lags, just to be on the safe side in
case your model did not specify enough lags) of observations are kept
together as a part of the bootstrap subsample. You have to go to the
literature for this, and I would expect to be able to find
user-written block bootstrap commands, but this is not my area, and I
cannot recommend anything specific (other than to Google it).

On the technical side, you need to tell -var- how your data are
-tsset-. When you do the plain bootstrap, you get repeated and missing
values of -tsset-, so -var- does not want to run: it does not have a
sequential -year- values to operate with.

On Thu, Jan 26, 2012 at 10:46 AM, Saunders,Kyle
<Kyle.Saunders@colostate.edu> wrote:
> Greetings statalisters:
>
> I have a small time series dataset, 1972-2010, two year delta (tsset year, yearly delta(2)), so 20 strobes.
>
> I have run a vector autoregression (var) successfully, everything turns out just fine.  The Grangers come out as expected, etc.
>
> I would like to bootstrap the standard errors, however when I run the command, I get the following error message:
>
> bootstrap: var lcdiff avgdiff, lags(1/2)
>
> (running var on estimation sample)
>
> Bootstrap replications (50)
> ----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    50
> insufficient observations to compute bootstrap standard errors
> no results will be saved
> r(2000);
>
> This occurs no matter how I specify the reps(#) option.
>
> Is this something I am missing inherent to the VAR, or is this a software limitation?  Any ideas on how to address it?
>
> Thanks for your time in advance,
> Kyle Saunders
> --
> Kyle L Saunders
> Department of Political Science
> Colorado State University
> w: http://lamar.colostate.edu/~ksaun
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index