tag:blogger.com,1999:blog-2198942534740642384.comments2019-01-20T11:09:22.385-08:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger4017125tag:blogger.com,1999:blog-2198942534740642384.post-91936102276182451732019-01-20T07:31:44.945-08:002019-01-20T07:31:44.945-08:00Dave,
Thank you very much for your blog. I'm...Dave, <br /><br />Thank you very much for your blog. I'm not sure if you take requests, but would you ever consider doing a blog related to time series regressions and interpretation? I feel that this is an often overlooked area of time series in many textbooks, and an area that I (and others) have struggled because of the lack of emphasis of interpreting various coefficients of different lag orders and transformations. <br /><br />Regards,<br /><br />JustinJustinnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-46764530316965293072019-01-16T01:56:38.922-08:002019-01-16T01:56:38.922-08:00Excellent page! Keep up the good work.Excellent page! Keep up the good work.Unknownhttps://www.blogger.com/profile/05661764235053772537noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-61148984955276469942019-01-07T07:28:24.875-08:002019-01-07T07:28:24.875-08:00Yes, that will do it too.Yes, that will do it too.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-2399653072052204712019-01-07T02:15:55.851-08:002019-01-07T02:15:55.851-08:00Dear Dave
What about transforming the VECM into i...Dear Dave<br /><br />What about transforming the VECM into its (restricted) VAR form, and generating the IRF using the traditional IRF routine based on VAR? Isn't that the standard way to generate IRF for VECM?<br /><br />Thanks!<br />Matthieuhttps://www.blogger.com/profile/16014048474736047689noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-26161674758662767382019-01-05T13:39:54.083-08:002019-01-05T13:39:54.083-08:00You are right that a lot of people don't check...You are right that a lot of people don't check the dynamic stability of the model.priti patilhttp://www.bloggrush.comnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-8031902894335812492018-12-29T13:34:03.073-08:002018-12-29T13:34:03.073-08:00I'm not sure what you mean, exactly, by "...I'm not sure what you mean, exactly, by "underwhelming and inconsistent". However - take a look at what I said in the post about Sargan's results. None of the moments of the FIML estimator exist in finite samples, but those of the 3SLS estimator do exist. So, if (for example) you are bootstrapping in an attempt to approximate the finite sample bias of the estimator (and then bias-correcting by subtracting the estimated bias from the original estimator), you are wasting your time in the case of the FIML estimator. The bias isn't even defined (as the first moment - the mean is not defined in finite samples) for this estimator. Ant bootstrap "approximation" to the bias that you come with is an approximation to something that doesn't exist! Not surprisingly, almost anything can then happen if you try to compare the (supposedly) bootstrap-bias-corrected FIML estimator with its 3SLS counterpart. (The latter IS well-defined.) Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-63917261796066854882018-12-29T11:16:41.775-08:002018-12-29T11:16:41.775-08:00Greetings. I was exploring bootstrapping both 3SLS...Greetings. I was exploring bootstrapping both 3SLS and FIML models. I find the results underwhelming and inconsistent. Is there a reason this would be? Any thoughts on it? Cheers, SteveStevehttps://www.blogger.com/profile/02785674096649084777noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-72633500331386534192018-12-24T17:58:57.694-08:002018-12-24T17:58:57.694-08:00thanks Professor Davethanks Professor Davedavid mendyhttps://www.blogger.com/profile/11034119156756708036noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-66721421423937424802018-12-23T21:25:38.613-08:002018-12-23T21:25:38.613-08:00I really like and appreciate your post.Thanks Agai...I really like and appreciate your post.Thanks Again. Keep writing.Kate Lynchhttps://www.clippingpathquick.com/noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-48134622860518624622018-12-11T20:02:36.771-08:002018-12-11T20:02:36.771-08:00Yes, you can use a dummy variable in this case. Th...Yes, you can use a dummy variable in this case. The dummy variable is just "shifting" the intercept - it doesn't matter if the variables are transformed or not.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-24246356167426646652018-12-11T18:12:13.595-08:002018-12-11T18:12:13.595-08:00Dear Professor Dave,
I am working on time dummy v...Dear Professor Dave,<br /><br />I am working on time dummy variable regression, I found that many models in this regression are linear and/or semilog form. Is it possible for time dummy variable models using log-log form (log of variable in both sides)? If not, why? is it because presence of the dummies that disallowed this log-log form? Thank you.<br />Regards Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-57512744681061261402018-12-10T07:12:01.058-08:002018-12-10T07:12:01.058-08:00Dear Professor,
Wishing to read a blog on panel AR...Dear Professor,<br />Wishing to read a blog on panel ARDL using eviews 10 too. Unknownhttps://www.blogger.com/profile/04577017245455610304noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-64355927677785107892018-12-07T10:14:13.630-08:002018-12-07T10:14:13.630-08:00Yes, that would be O.K.Yes, that would be O.K.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-57814019633378228132018-12-07T08:59:40.541-08:002018-12-07T08:59:40.541-08:00Dear Dave,
For the error correction equation, is...Dear Dave, <br /><br />For the error correction equation, is it appropriate to include other predictiors which are stationary, but dont have an impact to dependent on levels basis i.e. no long term impact?<br /><br />ThanksHardik Desaihttps://www.blogger.com/profile/17993077566289569728noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-44673813493335735092018-11-29T04:44:14.074-08:002018-11-29T04:44:14.074-08:00Yes, certainly.Yes, certainly.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-63789914714157574742018-11-29T04:43:21.066-08:002018-11-29T04:43:21.066-08:00Because there can still be a short-run relationshi...Because there can still be a short-run relationship. Cointegration is a long-run equilibrium phenomenon.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-3996146928286883262018-11-28T20:54:14.513-08:002018-11-28T20:54:14.513-08:00If series are not cointegrated why proceed with a...If series are not cointegrated why proceed with anything at all?Unknownhttps://www.blogger.com/profile/06543192220751470513noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-55788213664019723602018-11-28T20:45:08.585-08:002018-11-28T20:45:08.585-08:00Wouldn't it be a good idea to also examine for...Wouldn't it be a good idea to also examine forecasting performance in - sample, using a hold - out sample?Unknownhttps://www.blogger.com/profile/06543192220751470513noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-26050691413779572112018-11-28T10:20:01.846-08:002018-11-28T10:20:01.846-08:00Shoaib - Please clarify: do you have one variable ...Shoaib - Please clarify: do you have one variable that is I(2) and 2 that are I(0); or do you have one that is I(2) and 2 that are I(1); or something else.<br />Dave GilesDave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-49134924663099815762018-11-28T10:03:25.571-08:002018-11-28T10:03:25.571-08:00Dear Professor Dave,
I want to find the following ...Dear Professor Dave,<br />I want to find the following equation "GDP = f(exh, pexe)". "exh" is expenditure on health and "pexe" is the public expenditure on education. In the Unit root test Out of three only one variable is stationary at level-2 (Annual data from 1997-2017 total 21 obs.). What shall I do the make the data stationary from non-stationary. Can I run the Johansen co-integration with one variable stationary at level-2. or not or what will be the appropriate method for my data.<br /><br />Thank you for your guidance and time in advance.<br />Regards,<br />ShoaibShoaibhttps://www.blogger.com/profile/11052049687645364272noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-52084558067274170592018-11-26T02:13:12.460-08:002018-11-26T02:13:12.460-08:00Thanks for article!Thanks for article!Unknownhttps://www.blogger.com/profile/07086862310552183822noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-19687512885561524562018-11-21T23:17:00.395-08:002018-11-21T23:17:00.395-08:00PDL for multiple x is a special case of MIDAS mode...PDL for multiple x is a special case of MIDAS model. R package midasr can estimate such models. Here is the example:<br /><br />library(midasr)<br />x <- rnorm(100)<br />z <- rnorm(100)<br />px <- almonp(c(0.2,0.5,-0.2),10)<br />pz <- almonp(c(-0.1, 1, 0.3, -0.2), 15)<br />y <- 1+ mls(x, 0:9, 1) %*% px + mls(z, 0:14, 1) %*% pz + rnorm(100)<br />mod <- midas_r(y~mls(x, 0:9, 1, almonp) + mls(z,0:14, 1, almonp), data=list(x=x,y=y,z=z), start=list(x=c(0.2,0.5,-0.2), z=c(-0.1,1,0.3,-0.2)))<br />summary(mod)<br /><br />The model estimated here is the following:<br /><br />y_t = c + \beta_0\sum_{i=0}^9 (beta_1+ beta_2(i+1))x_{t-i} + \gamma_0\sum_{i=0}^14(\gamma_1 + \gamma_2(i+1) + \gamma_2(i+1)^2) z_{t-i} +\varepsilon_t<br /><br />It slightly differs from the specification in the text, since lag distribution function almonp has a different definition in midasr package. But you can use any lag distribution you want. Note that the estimation is done via NLS, so you need to supply the initial parameter values.mpiktashttps://www.blogger.com/profile/00263438252335043113noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-83991150533299239992018-11-21T07:42:30.600-08:002018-11-21T07:42:30.600-08:00The only R package that deals with the Almon estim...The only R package that deals with the Almon estimator directly, as far as I can tell, is the dLagM package that you mention. Why anyone would limit the package to allow for only a single regressor is beyond me! However, that does seem to be major limitation of this package. Eviews handle the Almon estimator in a general context. Alternatively you could use the Gretl econometrics package (which is free, and available at gretl.sourceforge.net/). There are some add-ons for Gretl that allow estimation of the Almon DL model - see this link for details: https://userpage.fu-berlin.de/sfu/gretlpkgpdf/lagreg_v07.pdfDave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-1437113651711170382018-11-21T03:57:03.832-08:002018-11-21T03:57:03.832-08:00Hi David,
Thanks for this awesome post. While look...Hi David,<br />Thanks for this awesome post. While looking for R packages that estimate these models, I came across package called dLagM. This package has a function for estimating polynomial distributed lag (PDL) model -- polyDLM. However, this function only accepts a single x driver. <br />I am not sure why the author could not extend the idea to accepting multiple x drivers. Are you aware of other packages in R that do this estimation?aj Jhttps://www.blogger.com/profile/12543012132563550614noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-70766996604341764062018-10-19T10:04:39.429-07:002018-10-19T10:04:39.429-07:00I would suggest that you use the free package, FRO...I would suggest that you use the free package, FRONTIER, which you can download from http://frontier.r-forge.r-project.org/front41.htmlDave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.com