tag:blogger.com,1999:blog-2198942534740642384.comments2015-05-27T03:21:16.548-07:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger2712125tag:blogger.com,1999:blog-2198942534740642384.post-7899183778440080642015-05-26T16:19:58.873-07:002015-05-26T16:19:58.873-07:00You'd add them in as additional "exogenou...You'd add them in as additional "exogenous variables".Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-35560879646840121392015-05-26T09:43:23.641-07:002015-05-26T09:43:23.641-07:00No, certainly not exactly. If |x| < 1 you could...No, certainly not exactly. If |x| < 1 you could use the Taylor series approximation for log(1+x), namely, log(1+x) = x - (x^2)/2 + .......... and go from there.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-91978362695686174492015-05-26T08:44:34.251-07:002015-05-26T08:44:34.251-07:00This comment has been removed by a blog administrator.mar skhttp://www.blogger.com/profile/10987234799563055579noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-29950118344716129542015-05-26T08:43:11.868-07:002015-05-26T08:43:11.868-07:00Can I cointegrate a Cobb Douglas function with an ...Can I cointegrate a Cobb Douglas function with an ARDL(or with any other method for cointegration), without using its logarithmic form? mar skhttp://www.blogger.com/profile/10987234799563055579noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-60931347177624639822015-05-26T04:10:35.509-07:002015-05-26T04:10:35.509-07:00Gerard Debreu also used to write very short articl...Gerard Debreu also used to write very short articles.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-26793535680613204102015-05-25T14:23:59.941-07:002015-05-25T14:23:59.941-07:00Yes, that's right.Yes, that's right.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-42372689769402867232015-05-25T13:08:45.370-07:002015-05-25T13:08:45.370-07:00One more question...What about this model? y=c(1)+...One more question...What about this model? y=c(1)+c(2)*x*w+c(3)*(z^2) ? Do I re-label x*w as, lets say q, and proceed in the usual way? ΠΑΛΑΙΟΣ ΠΑΝΑΓΙΩΤΗΣhttp://www.blogger.com/profile/10024506056938331963noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-57970773511063100942015-05-25T13:02:44.608-07:002015-05-25T13:02:44.608-07:00Thank you very much.Thank you very much.ΠΑΛΑΙΟΣ ΠΑΝΑΓΙΩΤΗΣhttp://www.blogger.com/profile/10024506056938331963noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-26508500380473938832015-05-25T12:31:19.846-07:002015-05-25T12:31:19.846-07:00Eric - no idea why they don't. The SHAZAM pack...Eric - no idea why they don't. The SHAZAM package used to have (maybe still has) the BAYES command that allows you to apply John Geweke's procedure for imposing inequality constraints. See Geweke, Journal of Applied Econometrics, 1986.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-48411171449759692702015-05-25T12:23:17.328-07:002015-05-25T12:23:17.328-07:00Your model is, in fact, linear - linear in the par...Your model is, in fact, linear - linear in the parameters, and that's all that matters. Just re-label X1=X^2, and X2=Z^2, and then proceed in the usual way.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-90161773078440400482015-05-24T18:11:08.804-07:002015-05-24T18:11:08.804-07:00I have a model which is not linear...y=c(1)+c(2)*(...I have a model which is not linear...y=c(1)+c(2)*(x^2)+c(3)*(z^2). Can I apply the ARDL method for cointegration in this model ? If yes, how can I transfrom this model in a ARDL model ?ΠΑΛΑΙΟΣ ΠΑΝΑΓΙΩΤΗΣhttp://www.blogger.com/profile/10024506056938331963noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-50792080490102768242015-05-23T23:09:42.869-07:002015-05-23T23:09:42.869-07:00I would like to thank you for the efforts you have...I would like to thank you for the efforts you have made in writing this article. <br />verhuisbedrijf zoetermeerhttp://zoetermeer-verhuisbedrijf.nlnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-11321129408390392752015-05-23T14:52:42.003-07:002015-05-23T14:52:42.003-07:00By estimating a VAR model - not an ARDL model. I h...By estimating a VAR model - not an ARDL model. I have several posts about this.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-27642646218844218592015-05-23T12:18:02.692-07:002015-05-23T12:18:02.692-07:00Nice discussion. I have a follow-up question, if ...Nice discussion. I have a follow-up question, if you will. You say:<br /><br />"Most econometrics/statistics packages that allow you to set up an arbitrary likelihood function, and then maximize it, don't allow you to constrain the signs or magnitudes of the resulting point estimates."<br /><br />Which I agree with. My question: why is this the case? Why don't most software packages build these features? Is it a philosophical rationale, wherein the software developers at Eviews, Stata, etc. allow for the tricks but don't want to encourage the use of inequality constraints by people who don't know exactly what they are doing?Ericnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-1720085420854286732015-05-23T12:14:08.899-07:002015-05-23T12:14:08.899-07:00Thanks for this post, it's really illuminating...Thanks for this post, it's really illuminating. However, I would like to know how we can do a causality test in the ARDL framework. Specifically, if Bounds test shows that X and Y are cointegrated, then (we recall from your earlier post that when variables are cointegrated, there would be at least a unidirectional causality between them) how can we do the Granger causality test in Eviews 9? gazmanhttp://www.blogger.com/profile/04212687883430389936noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-11928601878788727672015-05-21T11:03:46.157-07:002015-05-21T11:03:46.157-07:00Funny, indeed!Funny, indeed!Daumantasnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-66640715967144590682015-05-21T02:31:34.900-07:002015-05-21T02:31:34.900-07:00Pretty funny, it made my day. Thanks for cheering ...Pretty funny, it made my day. Thanks for cheering me up.Jannoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-38128673734502972162015-05-19T10:01:19.359-07:002015-05-19T10:01:19.359-07:00This comment has been removed by a blog administrator.Jazmín Lubina Avaloshttp://www.blogger.com/profile/12073219386764006122noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-44890460079831460852015-05-15T08:09:52.893-07:002015-05-15T08:09:52.893-07:00There is no contradiction. Cointegration implies t...There is no contradiction. Cointegration implies the presence of Granger causality (in one direction, or both ways). However, the presence of Granger causality does not necessarily imply cointegration.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-64455726429318756482015-05-15T04:17:32.495-07:002015-05-15T04:17:32.495-07:00Dear prof. Giles,
If the results from Johansen te...Dear prof. Giles,<br /><br />If the results from Johansen test show no cointegration between two variables (i.e. no long run relationship between them), but causality test (Toda-Yamamoto procedure) shows one-direction causality, does the second result contradict the first one? In other words, my question is whether Toda-Yamamoto test checks for long run or short run causality. <br /><br />Thank you in advance!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-76601783635555874172015-05-13T07:43:08.805-07:002015-05-13T07:43:08.805-07:00can you please tell me about the step by step proc...can you please tell me about the step by step process of johanson co integration and its basic requirements?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-35352601536738368422015-05-12T19:25:54.150-07:002015-05-12T19:25:54.150-07:00We should be teaching kids to code as early as 4th...We should be teaching kids to code as early as 4th grade. Math and Science teachers should have responsibility for this and quite frankly, we could start with $35 (US) calculators if every student couldn't afford an R prompt which can loaded on most smart phones with ARM processors by this point. By the time they are in high school, R or Matlab, or Scilab or Gretl or whatever should be consoles they are used to along with mathematics software like Microsoft Math, Octave, Julia, whatever is useful. 80 years ago John Von Neumann realized the field of math, statistics, and prediction was forever changed by electronic computation. Somehow, our educational systems take a long time to catch up...Ryan M. Ferrishttp://www.blogger.com/profile/03122603266808854365noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-73144746998469900862015-05-09T09:28:11.778-07:002015-05-09T09:28:11.778-07:00Look at the (maximized) log-likelihood values that...Look at the (maximized) log-likelihood values that are reported in each case. Which one is highest?Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-74335642946124278212015-05-09T04:51:42.340-07:002015-05-09T04:51:42.340-07:00Dear Professor,
When I tried to run a MLE for a GA...Dear Professor,<br />When I tried to run a MLE for a GARCH model with conditional skewness, the estimation results changed anytime I changed the starting values for the coefficients. Although in all cases convergence was achieved. Could you please help explain to me, and how can I know which is the most reliable estimation.<br />Thank you very much,<br />Lan Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-64754339192294636412015-05-08T00:04:48.462-07:002015-05-08T00:04:48.462-07:00Dear Professor Giles, Please let me know the p-val...Dear Professor Giles, Please let me know the p-value for rejecting the null hypothesis of 'No Serial Correlation' for LM Test. View the below Results:<br /><br />F-statistic 6.939430 Prob. F(1,26) 0.0140<br />Obs*R-squared 6.109500 Prob. Chi-Square(1) 0.0134<br /><br />@qchisq(.95,1)=3.84 <br /><br />Can I reject H0Muhammad Akramhttp://www.blogger.com/profile/03017698780837416156noreply@blogger.com