tag:blogger.com,1999:blog-2198942534740642384.comments2014-04-23T08:35:03.763-07:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger2086125tag:blogger.com,1999:blog-2198942534740642384.post-88215111507289306802014-04-23T08:35:03.763-07:002014-04-23T08:35:03.763-07:00Gul - You'll have to first-difference the I(1)...Gul - You'll have to first-difference the I(1) series.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-8998126278041174872014-04-22T22:26:20.785-07:002014-04-22T22:26:20.785-07:00Respected Dave Once again thanks for answering.If ...Respected Dave Once again thanks for answering.If data is mix stationary i.e. I(0) and I(1) and system is simultaneous with error terms are correlated than what i can do?which estimator is proper? or how can i transform my data into the form which is suitable for 3sls? Gul Khanhttp://www.blogger.com/profile/17147445177870988404noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-52828951635742293492014-04-21T20:08:47.821-07:002014-04-21T20:08:47.821-07:00Ting - thank you!!!!!! I hope I can have a lot mor...Ting - thank you!!!!!! I hope I can have a lot more interaction with Chinese students of econometrics.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-69815068800318710072014-04-21T19:21:21.560-07:002014-04-21T19:21:21.560-07:00Haha. Professor Giles, I'm proud to mention th...Haha. Professor Giles, I'm proud to mention that I'm the one who first marketed your blog on Weibo! I don't think we are less than "six degrees of separation" away but I'm a longtime follower of your blog. I believe were it not because blogspot is blocked by the "Great Firewall" of the Chinese government, you would have received much more traffic from China. Thanks a million for sharing your thoughts on econometrics and others. Looking forward to future episodes of your great MCMC series!<br /><br />-- TingTING JIANGhttp://www.blogger.com/profile/17858383722575056023noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-35050629778387025042014-04-21T10:22:45.968-07:002014-04-21T10:22:45.968-07:00Yes, that shouldn't matter.Yes, that shouldn't matter.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-78529007047575661502014-04-21T10:13:13.839-07:002014-04-21T10:13:13.839-07:00Dear Dave,
Are the results of the Bounds Test va...Dear Dave, <br /><br />Are the results of the Bounds Test valid if in the VEC model the coefficient on the lagged dependent variable is statistically insignificant?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-65410141647681788052014-04-21T06:02:30.800-07:002014-04-21T06:02:30.800-07:00post amazingpost amazingAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-89797121303564307302014-04-20T21:59:07.897-07:002014-04-20T21:59:07.897-07:00No - your data all need to be stationary; or else ...No - your data all need to be stationary; or else they all need to be I(1) and cointegrated.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-75316122498292547772014-04-20T21:53:05.970-07:002014-04-20T21:53:05.970-07:00Respected Dave
your post is too informative. i nee...Respected Dave<br />your post is too informative. i need a clarification. if data have mix i.e I(0) and I(1) in this case can we apply 3sls on data?Gul Khanhttp://www.blogger.com/profile/17147445177870988404noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-51759400201645200992014-04-19T16:08:38.978-07:002014-04-19T16:08:38.978-07:00All I need to say is thank you for this blog and t...All I need to say is thank you for this blog and this post particularly!Olaolu Olayenihttp://www.blogger.com/profile/15396628355317594447noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-40810769245383794282014-04-19T09:35:19.414-07:002014-04-19T09:35:19.414-07:00Ah - the Sounds of Silence!Ah - the Sounds of Silence!Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-22555313486305073192014-04-19T09:17:08.441-07:002014-04-19T09:17:08.441-07:00A+ (squared!)A+ (squared!)Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-18375743439717564852014-04-19T09:14:29.798-07:002014-04-19T09:14:29.798-07:00I have two ways to demonstrate this result:
A. Kr...I have two ways to demonstrate this result:<br />A. Kruskal's condition says that GLS=OLS if I can find a matrix G such that WX=XG, where W is the covariance matrix of the disturbances (which I'll denote S@I(n,n) since I can't figure out how to write Sigma or the Kronecker product in this comment). The condition says X2=X1*C (just post multiply the left and rhs of the condition by X2). Therefore we can write X=diag(X1,X2)=diag(X1,X1)A =(I(2,2)@X1)A where A'=(I(n,n) C'). Then use the property of Kronecker products to show that (S@I(n,n))*(I(2,2)@X1)=(I(2,2)@X1)@(S@I(n,n)), so G=(S@I(n,n))A. If you don't want to use Kruskal's theorem, you can follow the reasoning above and add a few lines to show that GLS is just OLS on each equation.<br /><br />B. Subract sigma(1,2)/sigma(1,1) times the first equation from the second. The transformed error u2-(sigma(1,2)/sigma(1,1))*u1 is uncorrelated with u1. So GLS amounts to minimizing a weighted sum of the squared errors from the first equation and the second equation. Given the GLS estimate b1, the n.s. condition for b2 is that it satisfies<br />X2'(y2-X2*b2-(sigma(1,2)/sigma(1,1))(y1-X1*b1))=0. <br />By symmetry, we also see that the n.s. condition for b1 given b2 is that it satisfies<br />X1'(y1-X1*b1-(sigma(1,2)/sigma(2,2))(y2-X2*b2))=0. <br /><br />The ols estimators satisfy<br />X2'(y2-X2*b2ols)=0 and X1'(y1-X1*b1ols)=0<br /><br />Your condition says that a vector is orthogonal to span(X1) iff it is orthogonal to span(X2). So we also get<br />X1'(y2-X2*b2ols)=0. and X2'(y1-X1*b1ols)=0<br /><br />From this we quickly see that under your condition, the OLS estimators satisfies the n.s. condition for the GLS estimators, so they'll be the same.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-54585268774392484642014-04-15T13:52:44.669-07:002014-04-15T13:52:44.669-07:00Yes. And if you're using EViews, there's a...Yes. And if you're using EViews, there's a place for you to enter them in the VAR specification - indeed, that's precisely where you put the "extra" lags if you're using the Toyoda-Yamamoto method.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-41178399028748023642014-04-15T11:28:23.047-07:002014-04-15T11:28:23.047-07:00Hello Prof Giles!
If we have two exogenous variab...Hello Prof Giles!<br /><br />If we have two exogenous variables in our model, then should we take those variables also into the Granger causality test?<br /><br />Thanks a lot!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-1174451951199620642014-04-15T08:27:52.446-07:002014-04-15T08:27:52.446-07:00Ruliff - As I said in my response to the previous ...Ruliff - As I said in my response to the previous comment - use the POT package in R. There are other R packages too - e.g. evd.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-23763154218166203622014-04-15T04:38:30.528-07:002014-04-15T04:38:30.528-07:00Thanks Mr. Dave Giles,
My Name is Ruliff Demsy (ru...Thanks Mr. Dave Giles,<br />My Name is Ruliff Demsy (ruliffdemsy@gmail.com) from Indonesia, and currently running a research about EVT analysis with POT<br />I have read about your paper according your research about EVT back then in 2007.<br /><br />I have also searched through your code section and there is no code available for EVT in EViews or R.<br />I am also confued in running MSE and Bias test for selecting best-fit parameter<br />Would you mind to let me take a look into your code?<br /><br /><br />Thanks a lot Sir!!Ruliff Demsyhttp://www.blogger.com/profile/15609366062207342862noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-90280465117518419702014-04-14T15:52:05.047-07:002014-04-14T15:52:05.047-07:00So, do you agree with Dan, or not?So, do you agree with Dan, or not?Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-54680623590465157162014-04-14T15:51:10.673-07:002014-04-14T15:51:10.673-07:00Some quotation marks in your blog post would have ...Some quotation marks in your blog post would have made that clear.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-44775088255831252932014-04-14T15:48:12.734-07:002014-04-14T15:48:12.734-07:00Correction: the quote Dave attributes to me is act...Correction: the quote Dave attributes to me is actually Dan Hamermesh.Chris Auldhttp://chrisauld.comnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-80108880701104308512014-04-13T08:52:04.184-07:002014-04-13T08:52:04.184-07:00The answer is that heteroskedasticity DOESN'T ...The answer is that heteroskedasticity DOESN'T necessarily make the OLS standard errors larger. They can be larger or smaller than they would otherwise be. Which way they are distorted will depend on the the relationship between the (changing) variance of the errors, and the pattern of the variability of the regressors in the sample. Likewise, the het-consistent standard errors can be larger or smaller than the regular standard errors in the face of heteroskedasticity.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-79159472873725888712014-04-13T02:29:18.930-07:002014-04-13T02:29:18.930-07:00Dear Prof Giles,
A question on robust standard er...Dear Prof Giles,<br /><br />A question on robust standard errors.<br /><br />Since HETEROSCEDASTICTY makes OLS variances and SE larger, should we expect Robust S.E to be smaller than OLS SE?<br /><br />But actually when we correct heteroscedasticty by White's Robust test, we find OLS SE can be larger or smaller than OLS SE.<br /><br />Why it is so?<br /><br />Kindly help. <br />Thank you.<br />Santosh Dashhttp://www.blogger.com/profile/02016226999263087762noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-39235284950327072862014-04-11T09:48:39.646-07:002014-04-11T09:48:39.646-07:00There is an app called R-instructor that allows yo...There is an app called R-instructor that allows you to run Rstudio server on your tabletÂ´s web browser. Not the most confortable thing in the world, but is still very cool. The app also has some useful quick guides for data analysis with R.Fede C.http://www.blogger.com/profile/06900428072541523243noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-53304975393049640192014-04-10T13:17:19.924-07:002014-04-10T13:17:19.924-07:00Zeba - see my new post, here: http://davegiles.blo...Zeba - see my new post, here: http://davegiles.blogspot.ca/2014/04/proof-of-result-about-adjusted.htmlDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-56925236961214381132014-04-10T09:53:45.312-07:002014-04-10T09:53:45.312-07:00It needs to be I(1)It needs to be I(1)Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.com