tag:blogger.com,1999:blog-2198942534740642384.comments2017-05-23T14:28:38.928-07:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger3686125tag:blogger.com,1999:blog-2198942534740642384.post-48630151419734614312017-05-23T13:16:19.777-07:002017-05-23T13:16:19.777-07:00Another problem in keeping up with the latest is t...Another problem in keeping up with the latest is the emergence a lot of new journals. Many of them do publish good articles. More review articles and meta-analyses would certainly help. That's why journals such as the Annual Reviews of ___, and Journal of Economic Survey, and Cliometrica are of great utility. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-61232624729682580132017-05-23T11:01:02.451-07:002017-05-23T11:01:02.451-07:00I've noted repeatedly in my posts about ARDL m...I've noted repeatedly in my posts about ARDL models that they can't be used if any of the variables are I(2).Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-34149384931117269962017-05-23T10:59:37.123-07:002017-05-23T10:59:37.123-07:00Hello sir,
thank you for your job. I have a quest...Hello sir,<br /><br />thank you for your job. I have a question: what should I do if the dependent variable is I(2) and the other variables is I(1)?<br />Thank youAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-80587883857739158802017-05-21T07:12:55.167-07:002017-05-21T07:12:55.167-07:00Dhanusha - with only 39 observations, and nine var...Dhanusha - with only 39 observations, and nine variables (and their lags), you're not going to be able to do very much with either type of model. Just think about the degrees of freedom.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-22139973673207627262017-05-21T04:08:56.627-07:002017-05-21T04:08:56.627-07:00«less excited when they're using an old tool, ...«<i>less excited when they're using an old tool, but they think that it's a new one!</i>»<br /><br />In another field with which I am more familiar old ideas, concepts and tools get rediscovered all the time and published as original under new names.<br />The difficulty is that every PhD student, postdoc (and associate professor) has to find a constant stream of "original" ideas for writing papers.<br />If only truly original ideas were publishable, very few people would get published, and the whole "tenure track" scheme to use PhD students and postdocs as casual low-cost project labour would crash.<br />One interesting statistic I read somewhere is that the number of PhD students and postdocs keeps increasing faster than the number of articles published in "top journals", creating a massive crunch.Blissexnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-63398818432401692652017-05-21T00:51:25.419-07:002017-05-21T00:51:25.419-07:00Dear Professor Giles, I'm trying to chose betw...Dear Professor Giles, I'm trying to chose between VECM and ARDL approaches for modeling inflation (CPI Index) in Sri Lanka. My sample size is only 39 annual data and have altogether nine variables. Is it true that ARDL works better with small samples and higher number of variables compared to VECM? Thank you very much in advance! Dhanusha Pathiranahttps://www.facebook.com/dhanushagihan.pathirana.5noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-69435531059459618912017-05-20T04:36:36.062-07:002017-05-20T04:36:36.062-07:00Should that non-significant variable be interprete...Should that non-significant variable be interpreted and discussed in the results like the other variables. My important variables are non-significant although the model has passed the Bounds test Muhammad Azam Niazihttp://www.blogger.com/profile/06113652375933331044noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-72946633919231142322017-05-20T04:28:14.584-07:002017-05-20T04:28:14.584-07:00Michael - thanks for the interesting question/comm...Michael - thanks for the interesting question/comment. It deserves a full response so I'll allocate a post to it this weekend. DGDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-50482218224750556772017-05-20T04:10:42.523-07:002017-05-20T04:10:42.523-07:00Mark - thanks for the thoughtful comments. All of ...Mark - thanks for the thoughtful comments. All of them right on target. I guess I could summarize my feelings behind this post as follows: 1. (Positive) - it's great to see "old tools getting a new lease of life when new problems come along. 2. (Negative) - I wish that people would be less excited when they're using an old tool, but they think that it's a new one!Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-34344782046686176122017-05-19T14:22:41.995-07:002017-05-19T14:22:41.995-07:00great article Dave. Similarly, one could argue th...great article Dave. Similarly, one could argue that GMM, although clearly not exactly the same , has its origins in the method of moments which was discovered by Karl Pearson, a statistician, in 1894. In the same way, some would argue that ARCH is an AR model (which was discovered by Box et al in the 70's ) where X_t is the volatility. Hansen and Engle are incredibly talented and deserving of their discoveries but, in one way or another, everyone stands on the shoulders of past giants. All the best.<br /><br /><br /><br /><br />mark leedshttp://www.blogger.com/profile/13213841692738932471noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-2509466510112247872017-05-13T10:27:26.807-07:002017-05-13T10:27:26.807-07:00No, I'm not using EViews but, thank you very m...No, I'm not using EViews but, thank you very much, Prof. Dave for your help. Hope I'll be able to get the right resultS.http://www.blogger.com/profile/09529156142256439066noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-74921753424623479752017-05-13T08:58:54.585-07:002017-05-13T08:58:54.585-07:00The asymptotic distribution of the ratio of the ma...The asymptotic distribution of the ratio of the maximum likelihood estimates will be normal. The corresponding ratio of estimates will NOT be normal in finite samples. Why is the asymptotic distribution normal (not STANDARD) normal)? The ML estimator has an asymptotic normal distribution. By the invariance property of MLEs, a ratio of MLE's is the MLE of the ratio of the parameters being estimated. So, the ratio estimator, being an MLE itself, will have an asymptotic distribution that is normal. In addition, if the model's errors are normal, then MLE = OLS, so the above comments apply to the ratio of the OLS estimates. To get the variance (and ultimately the standard error) of this variance ratio, you would use the Delta method (which can be used for any non-linear transformation of random variablnes, such as estimators). Are you using EViews? If so, I can suggest a trick to get the asymptotic standard error with no effort. Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-65112849505881315342017-05-13T07:44:09.118-07:002017-05-13T07:44:09.118-07:00Hello Prof. Dave,
I have a question about the rat...Hello Prof. Dave,<br /><br />I have a question about the ratio of coefficients (α/β) estimated from a simple regression model - can we say that the asymptotic distribution of (α/β) is normal? <br /><br />Thank you very much.S.http://www.blogger.com/profile/09529156142256439066noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-34733513844417111992017-05-13T02:28:26.657-07:002017-05-13T02:28:26.657-07:00Prof. Giles,
Is it required to accommodate structu...Prof. Giles,<br />Is it required to accommodate structural breaks when one deals with long (time periods) and large (no. of countries) panel?<br /><br />In case of large and long panel data, if we cointegration without accounting for structural breaks is established, the results is “robust”. However, if cointegration cannot be established then it calls for accommodating structural breaks in the cointegrating test equation. <br /><br />Further, In a long time period, Bai-Perron structural break test will suggest more than 2 endogenous breakdates. It would be difficult to accommodate all break points (if they are country specific and dates are different).<br /><br />Thank you.<br /><br />SanthoshAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-29497529649105547612017-05-11T14:51:09.991-07:002017-05-11T14:51:09.991-07:00So ARDL is basically a one-equation version of a V...So ARDL is basically a one-equation version of a VAR model?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-31205542781929840012017-05-11T04:38:05.678-07:002017-05-11T04:38:05.678-07:00I should have said "You interpretation is cor...I should have said "You interpretation is correct". Sorry. DGDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-86928190474699864152017-05-11T04:18:07.798-07:002017-05-11T04:18:07.798-07:00Reynaldo - your interpretation. My earlier remarks...Reynaldo - your interpretation. My earlier remarks are specifically in reference to code for MONTHLY cointgegration. In the case of seasonal cointegration at the quarterly level you don;t need any special "code" the steps are straightforward and can easily be implemented with regular data transformations in EViews, Gretl, Stata, R, etc. See my comments and references towards the end of the above post.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-49624615371707179152017-05-10T22:36:21.869-07:002017-05-10T22:36:21.869-07:00Dear professor Giles. After reading this article, ...Dear professor Giles. After reading this article, can I conclude that when dealing with quarterly or monthly series, the first step is to perform seasonal unit root tests to assess if the seasonality is deterministic or not? Is this correct? If yes, I have two possible outcomes. One is that the series has deterministic seasonality and in that case I can seasonally adjust it.<br /><br />However, if the outcome of the unit root test is tha the seasonality is stochastic, What is next (considering that you reply to a comment in this post that regarding to the availability of codes for cointegration "Nothing comes to mind - sorry")?<br /><br />Kind Regards,<br /><br />Reynaldo<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-11076944996473242072017-05-10T05:02:43.027-07:002017-05-10T05:02:43.027-07:00Take a look at this EViews blog post...... http:/...Take a look at this EViews blog post...... http://blog.eviews.com/2017/05/autoregressive-distributed-lag-ardl_8.htmlDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-85731035852345344302017-05-10T04:58:47.702-07:002017-05-10T04:58:47.702-07:00Yes - especially tests for serial correlation. The...Yes - especially tests for serial correlation. The presence of the latter could signal a mis-specification of the lag length(s). And the bounds test critical values are only valid if the errors are independent,.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-79547724715446487412017-05-09T23:40:13.223-07:002017-05-09T23:40:13.223-07:00Dear Professor,
mentioned below a table from a pub...Dear Professor,<br />mentioned below a table from a published article. Is it mandatory for ARDL model to have only ONE co-integration? such tables are reported in research article.<br /><br />Dependent variable AIC lags F-statistic Decision <br /><br />FF (F\Y, K, L, T) 2 6.701 Cointegration<br />FY (Y\L, K, F, T) 2 2.365 No cointegration<br />FL (L\Y, K, F, T) 1 0.762 No cointegration<br />FT (T\Y, K, F, L) 1 2.736 No cointegration<br />FK (K\Y, L, F, T) 1 2.552 No cointegration<br />Lower-bound critical value at 1% 3.06 <br /><br />Upper-bound critical value at 1% 4.15 <br /><br />Lower and Upper-bound critical values are taken from Pesaran et al. (2001), Table CI(ii) Case II.<br /><br />Thank you<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-16574711289378517942017-05-09T23:16:24.189-07:002017-05-09T23:16:24.189-07:00Dear Professor,
Should I apply tests lik serial co...Dear Professor,<br />Should I apply tests lik serial correlation (Breusch–Godfrey) and heteroskedasticity(ARCH) even after using the HAC(Newey-West) option in Eviews 9?<br />Thank youAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-11525937662299819482017-05-08T06:56:31.768-07:002017-05-08T06:56:31.768-07:00Hi Dave,
Thanks for sharing this interesting list...Hi Dave,<br /><br />Thanks for sharing this interesting list of articles! I'm wondering, how do you go about finding these types of articles to read? Are you a subscriber to these publications/do you regularly check for new updates online? I'd like to start keeping more up to date with academic articles, but I'm not sure where to start. Really love the site!<br /><br />Best,<br />MichaelMichaelhttp://www.blogger.com/profile/12631186488420390412noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-43427200888375547552017-05-06T20:36:07.276-07:002017-05-06T20:36:07.276-07:00Thank you very much for the advice. I'll check...Thank you very much for the advice. I'll check it.Çağlar Önalhttp://www.blogger.com/profile/11998252927899233697noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-41104008223461184122017-05-01T07:18:14.841-07:002017-05-01T07:18:14.841-07:00Thank you Professor.
Thank you Professor. <br />Burcu Ozcanhttp://www.blogger.com/profile/06077111864603959268noreply@blogger.com