Can someone help me test for Granger causality after a VECM?
After a search of the archives, I still cannot find a means of testing for Granger causality after a VECM. -vargranger- can only be used after a var or svar, which excludes models with cointegration. -test- cannot be used after vec I'm not sure how to write a standard F-test for a vecm, due to the need to include the alpha-beta cointegration matrices. At my present stage of understanding, an F-test needs to be conducted on the full model, which would include the cointegration equation(s) for a vecm. Any help would be greatly appreciated, Brandon. * * 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/ |
From Lonnie Magee's lecture notes:
http://socserv.mcmaster.ca/magee/761_762/Stata%20Examples/VEC%20handout%20full.pdf T On Sun, Jan 24, 2010 at 4:35 AM, b.tracy <[hidden email]> wrote: > Can someone help me test for Granger causality after a VECM? > > After a search of the archives, I still cannot find a means of testing for Granger causality after a VECM. > > -vargranger- can only be used after a var or svar, which excludes models with cointegration. > > -test- cannot be used after vec > > I'm not sure how to write a standard F-test for a vecm, due to the need to include the alpha-beta cointegration matrices. At my present stage of understanding, an F-test needs to be conducted on the full model, which would include the cointegration equation(s) for a vecm. > > Any help would be greatly appreciated, > > Brandon. > > > > * > * 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/ > -- To every ω-consistent recursive class κ of formulae there correspond recursive class signs r, such that neither v Gen r nor Neg(v Gen r) belongs to Flg(κ) (where v is the free variable of r). * * 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/ |
Thank you!
Quoting from the mentioned pdf (to capture the answer to the question): " The final two test commands are testing for Granger causality. By having already concluding that log income and log consumption are cointegrated, we have implicity concluded already that there is a long-run causal relation between them. So the causality being tested for in a VECM by these tests is sometimes called “short-run Granger causality”. vec lconsumption lincome, lags(3) test ([D_lconsumption]: LD.lincome L2D.lincome) test ([D_lincome]: LD.lconsumption L2D.lconsumption) " * * 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/ |
On Sunday, Brandon <[hidden email]> posted a question on Granger causality
containing the following statement: > -test- cannot be used after vec Tirthankar [[hidden email]] provided a link that includes comments and also code with one solution to Brandon's question. The code uses the -test- command after fitting a model with -vec-. I would like to emphasize that -test- is a valid postestimation command after using -vec to estimate the parameters of a vector error correction (VEC) model. Here is a short code where I use -test- to perform a Wald test after fitting the VEC model in the first example of the entry for -vec- on the time-series manual: use http://www.stata-press.com/data/r11/rdinc vec ln_ne ln_se test [D_ln_ne]LD.ln_ne + [D_ln_ne]LD.ln_se = 0 And the output for the last command: . test [D_ln_ne]LD.ln_ne + [D_ln_ne]LD.ln_se = 0 ( 1) [D_ln_ne]LD.ln_ne + [D_ln_ne]LD.ln_se = 0 chi2( 1) = 0.13 Prob > chi2 = 0.7167 --Gustavo Sanchez [hidden email] * * 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/ |
This post has NOT been accepted by the mailing list yet.
And what command should you use when you have many independent variables with granger causality after a VECM. My VEC regression is:
vec d.lstp d.lexr d.lsint d.lip d.lur d.lyie d.loil d.lcon d.lgold d.ltb d.linf, lags(3) Can anybody try to give me the exact command in this example which I should use to test for Granger causality of the dependent variable (d.lstp) on all the different independent variables? Thanks in advance! |
This post has NOT been accepted by the mailing list yet.
In reply to this post by Gustavo Sanchez
And what command should you use when you have many independent variables with granger causality after a VECM. My VEC regression is: vec d.lstp d.lexr d.lsint d.lip d.lur d.lyie d.loil d.lcon d.lgold d.ltb d.linf, lags(3) Can anybody try to give me the exact command in this example which I should use to test for Granger causality of the dependent variable (d.lstp) on all the different independent variables? Thanks in advance! |
Powered by Nabble | Edit this page |