tag:blogger.com,1999:blog-2198942534740642384.comments2019-02-21T06:36:54.614-08:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger4031125tag:blogger.com,1999:blog-2198942534740642384.post-86205104779679288612019-02-21T06:36:54.614-08:002019-02-21T06:36:54.614-08:00A dummy variable can't be I(1). It is bounded ...A dummy variable can't be I(1). It is bounded to take values of either 1 or 0. So, it will be I(0). But it can still be included in an ARDL model.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-63634627857912508572019-02-20T19:24:45.289-08:002019-02-20T19:24:45.289-08:00If the dummy variable is I(1), is it possible to i...If the dummy variable is I(1), is it possible to include that dummy as a fixed regressor in the ARDL approach? Thanks. <br />Anuruhttps://www.blogger.com/profile/06537609951027415713noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-41507876524932444252019-02-19T05:39:11.497-08:002019-02-19T05:39:11.497-08:00Hi - You're right. Here, things work because o...Hi - You're right. Here, things work because of the connection between the Normal and log-Normal distributions. I can't think of a general reference off-hand. Is there a particular transformation you're concerned about?Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-48601427130334838622019-02-18T20:42:29.317-08:002019-02-18T20:42:29.317-08:00Thank you for the excellent explanation!
Out of c...Thank you for the excellent explanation!<br /><br />Out of curiosity - I suspect that problems caused by other transformations on the dependent variable don't resolve as cleanly. Is there a paper or book you'd recommend that deals with these back-transformations? Studenthttp://anonymous.comnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-56497517883268865662019-02-18T00:05:20.862-08:002019-02-18T00:05:20.862-08:00Dave: thank you for your prompt reply and feedback...Dave: thank you for your prompt reply and feedback.Federico Belottinoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-30938223063324129312019-02-15T14:29:28.131-08:002019-02-15T14:29:28.131-08:00Frederico - I have now amended the EViews code and...Frederico - I have now amended the EViews code and updated the blog post. Again - much appreciated.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-25741585730409782912019-02-14T09:55:57.562-08:002019-02-14T09:55:57.562-08:00Frederico - you are right! How silly of me. The st...Frederico - you are right! How silly of me. The structure I used would have been correct if it had been the Logit model, and U'd used the cumulative logistic instead of the cumulative Normal. I'll have to fix this at some stage, even though this post is ancient history. Second, it's moot as to whether one reports the marginal effect at the mean, or the average of the marginal effects. Of course, you do get different answers. Finally, the X variable was just artificially generated - it's in the EViews workfile alreadt and wasn't generated in the program. DGDave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-2346854863949221972019-02-14T08:03:18.615-08:002019-02-14T08:03:18.615-08:00Dear Dave,
I'd like to jump in here, even tho...Dear Dave,<br /><br />I'd like to jump in here, even though this is an old thread.<br />In particular, I'd like to ask for two clarifications on the Eviews code you used for the Monte Carlo analysis. I might be wrong or missing something since I don't know the Eviews syntax very well but it seems to me that marginal effect of x (at means) in a probit model should be<br /><br />@dnorm(c(1)+c(2)*@mean(x))*c(2)<br /><br />instead of<br /><br />@cnorm(c(1)+c(2)*@mean(x))*(1-@cnorm(c(1)+c(2)*@mean(x)))*c(2)<br /><br />Am I wrong?<br />Second: Why did you consider the marginal effect at mean instead of the average marginal effect? How the regressor x is generated? I wasn't able to find it looking at the code.<br /><br />Many thanks,<br />Federico<br /><br /><br /><br /><br /><br /><br />Federico Belottinoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-77285091802954523452019-02-06T09:12:28.732-08:002019-02-06T09:12:28.732-08:00Excellent work Mr. Giles, thank you for your contr...Excellent work Mr. Giles, thank you for your contribution!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-56502489277046588572019-02-01T06:36:50.926-08:002019-02-01T06:36:50.926-08:00Thanks for your query. Have you read the "rea...Thanks for your query. Have you read the "read_me" text object in the workfile I supplied, and the comments embedded in the program file? In the latter you will see the comment <br /><br />' r = cointegrating rank<br />' p = no. of variables in system<br /><br /><br /> The first column in the output is (p - r), not the cointegrating rank itself.<br /><br />I presume that you have modified the code in the program file to match your particular situation, as mentioned in the various comments? The table that you supplied relates only to the H(L) test. The main table of results that you should get has additional columns, to the right, for the critical values for the H(R) test.<br /><br />Regarding your second question, the answer is "yes". I hope this helps.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-5998392503947825482019-02-01T01:22:31.361-08:002019-02-01T01:22:31.361-08:00Dear professor,
Sorry for writing this, but after...Dear professor,<br /><br />Sorry for writing this, but after running your code I got a table like this,<br /><br />P_MINUS_R CRIT_HL_90 CRIT_HL_95 CRIT_HL_99<br /> 1 20.94518 23.50003 28.80999<br /> 2 42.04003 45.37806 52.09231<br /><br />I don't understand this output. In the first column, Is 1 the number of cointegration vectors? I mean, If my test statistics is 30 I can say that there is only one cointegration vector because its value is larger than the value of all the critical values. On the other hand, I cannot say that there are 2 cointegration vectors because my test statistics is 30 with is lower than 42.04003 (which is the the value for the 10% significance level).<br /><br />Am I right?<br /><br />Finally, for the decision about the cointegration in my data, I just need to take the test statistics resulting of applying to my data the trace test in any software (including in my data the dummies accounting for the breaks, logically)?<br /><br />Kind regards and very much thank you for your great blog.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-35437131721638335432019-01-31T08:09:33.455-08:002019-01-31T08:09:33.455-08:00In my experience, statistics and econometrics espe...In my experience, statistics and econometrics especially is more concerned with putting forth an understandable model of how x influences y. In machine learning, it is less important how x influences y than it is that we can predict it (eg. through some massive poorly understood neural network)Thor P.Nhttps://www.blogger.com/profile/14773159126426249081noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-14000622088814027332019-01-28T14:42:42.108-08:002019-01-28T14:42:42.108-08:00Reynaldo - yes,
Dave GilesReynaldo - yes,<br />Dave GilesDave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-22160896824021649552019-01-27T19:41:13.565-08:002019-01-27T19:41:13.565-08:00Dear professor Giles,
After the Cointegration tes...Dear professor Giles,<br /><br />After the Cointegration test, I need to set up the final VECM model including these dummies, right?<br /><br />Kind regards,<br /><br />Reynaldo SenraReynaldo Senrahttps://www.blogger.com/profile/16267851641132171695noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-38560695081838897682019-01-25T01:41:50.621-08:002019-01-25T01:41:50.621-08:00Dear Professor,
Thank you so much for this very il...Dear Professor,<br />Thank you so much for this very illustrative and pedagogical guided example (and for the excel)! Great content, still relevant nowadays.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-27096978033561344892019-01-23T09:39:02.857-08:002019-01-23T09:39:02.857-08:00Dealing with the autocorrelation is of primary imp...Dealing with the autocorrelation is of primary importance. More so than the (apparent) lack of significance of some of the extra lags. And on what basis are you judging them to be insignificant, anyway? The t-tests won't be valid. I'd keep the extra lags if this sorts out the autocorrelation problem.Dave Gileshttps://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-83127244000843145082019-01-23T06:51:11.215-08:002019-01-23T06:51:11.215-08:00Sir, I am working with Financial time series data ...Sir, I am working with Financial time series data with all variables stationery at I(1), and using ARDL system for this. After using the ARDL and ECM, when I checked residual diagnostics, it shows serial auto correlation. ANd when I try to increase the lag structure, the coefficients are becoming insignificant. Kindly help. Unknownhttps://www.blogger.com/profile/11018115257209246305noreply@blogger.comtag: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.com