tag:blogger.com,1999:blog-2198942534740642384.comments2015-01-30T16:24:32.944-08:00Econometrics Beat: Dave Giles' BlogDave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comBlogger2533125tag:blogger.com,1999:blog-2198942534740642384.post-43129078696053058032015-01-30T16:24:32.944-08:002015-01-30T16:24:32.944-08:00You can only test for cointegration among variable...You can only test for cointegration among variables that are non-stationary. However, take a look at my posts on ARDL models.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-51516019254036149542015-01-30T16:13:48.607-08:002015-01-30T16:13:48.607-08:00Hi, I'm studying the effects of macroeconomic ...Hi, I'm studying the effects of macroeconomic variables on stock returns and four of my variables are I(1) and the other is I(0), can i run a johansen test of cointegration?vicnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-49106756049581086282015-01-30T13:39:34.633-08:002015-01-30T13:39:34.633-08:00Extending the lag length(s) will usually resolve t...Extending the lag length(s) will usually resolve this - if you have enough degrees of freedom to be able to do so.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-19836321412589290102015-01-30T13:29:18.749-08:002015-01-30T13:29:18.749-08:00Hello Prof.
If I was using VAR model however there...Hello Prof.<br />If I was using VAR model however there is autocorrelation <br />should I change the VAR model or is there some steps to fix it or should I just report my findings and keep it the way it is<br />?<br />thanks Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-45619283805719635202015-01-28T13:53:50.609-08:002015-01-28T13:53:50.609-08:00Yes you do - unless they happen to be cointegrated...Yes you do - unless they happen to be cointegrated.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-31180076453366548842015-01-28T13:51:24.846-08:002015-01-28T13:51:24.846-08:00To regress one macro variable on a constant and an...To regress one macro variable on a constant and another macro variable and then test for structural breaks, do you need to make the data stationary first?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-4040339780359687452015-01-27T09:52:48.510-08:002015-01-27T09:52:48.510-08:00Correct - the error must be different too. I'l...Correct - the error must be different too. I'll fix the wording when I get a chance.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-48026258215232701032015-01-26T20:56:32.621-08:002015-01-26T20:56:32.621-08:00I think there's a minor algebra issue here. J...I think there's a minor algebra issue here. Just before eq. (6) we read, "or, from the definition of __ in (2):" Something is missing; from context it looks like it's 'ut.' But if so, in substituting out ut, instead of simplifying to εt, we get (εt - φ2ut-2). I couldn't figure out a way to simplify further. So could it be that the two lagged forms are identical, except for the intercept *and* the error term? Thanks, JSJames Stodderhttp://www.blogger.com/profile/14179384217139920062noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-56326133024570385392015-01-23T11:03:36.712-08:002015-01-23T11:03:36.712-08:00Fabulous post, Professor! Your post has helped man...Fabulous post, Professor! Your post has helped many students in their dissertation. I have a small question. How would you get the impulse response function in the case with identification restrictions (such as the restrictions on w_t in the Pesaran paper)? Thank you, and looking forward to your 3-variable example :-)Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-2541865947255113662015-01-19T08:25:05.596-08:002015-01-19T08:25:05.596-08:00That's correct.That's correct.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-63644700723338012352015-01-18T21:11:28.609-08:002015-01-18T21:11:28.609-08:00Thank you very much for this wonderful blog Prof. ...Thank you very much for this wonderful blog Prof. Giles! I have questions concerning ARDL(p,q) bounds-testing. <br />Assume I have X-variable(s) which cannot have a unit-root; for instance a dummy-variable for certain events (like macroeconomic shocks) and its lagged values. Can I include this (these) variable(s) in the "ARDL bounds-testing approach"?<br />Such a dummy-variable, name it SHOCK, would be rather similar to your “BREAK” variable, I guess. The coefficient on D_SHOCK would then pinpoint the short-run effect of the shock and the coefficient on SHOCK would represent the long-run impact of the macroeconomic shock (although SHOCK is I(0) and cannot be cointegrated with Y)? The bounds-test for cointegration would be an F-test on joint significance of the lagged Y and lagged X variables (not including SHOCK)?<br /><br />Thank you very much again!<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-66699109934666473442015-01-17T16:35:08.390-08:002015-01-17T16:35:08.390-08:00Thanks for spotting that. Now fixed.Thanks for spotting that. Now fixed.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-10525909641204648682015-01-17T16:24:13.175-08:002015-01-17T16:24:13.175-08:00Trivial typo: In the repeated substitution expres...Trivial typo: In the repeated substitution expression that appears just above equation (4), the B0 term is missing. It is correctly included in equation (4), however, in the form of B0*B1^0.James Stodderhttp://www.blogger.com/profile/14179384217139920062noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-57677636579892387632015-01-17T16:15:47.693-08:002015-01-17T16:15:47.693-08:00Thanks. vt is the (unobserved) error term. zt is t...Thanks. vt is the (unobserved) error term. zt is the corresponding (observed) residual term. You could call it vhat if you wished. Then, it;s lagged by one period to get the error correction term.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-18084933320878932322015-01-17T14:45:05.543-08:002015-01-17T14:45:05.543-08:00Fabulous post, but just a small confusion about yo...Fabulous post, but just a small confusion about your notation in equations (2) and (3), and also in (5) and (6). In (3) and (6) you use vt as the residual of the cointegrating equation. But in (2) and (5) you use zt as the *estimate* of this residual. Is this the correct interpretation? I’m not used to such variable-name changes, but I'm sure you have good reasons. I had expected a ‘hat’ notation on the estimates, or something similar. Thanks, JSJames Stodderhttp://www.blogger.com/profile/14179384217139920062noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-52718244648850936402015-01-16T23:54:46.566-08:002015-01-16T23:54:46.566-08:00The programming in the editor will be fantastic!The programming in the editor will be fantastic!David Barrera Ojedahttp://www.blogger.com/profile/10806642228106715803noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-42531245476092902722015-01-16T10:44:27.184-08:002015-01-16T10:44:27.184-08:00If the data are cointegrated, the regression will ...If the data are cointegrated, the regression will be meaningful, and in fact OLS will be "super-consistent".Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-31590420707083765632015-01-16T10:40:59.891-08:002015-01-16T10:40:59.891-08:00sorry, a follow up question. if the regressors coi...sorry, a follow up question. if the regressors cointegrated with different coefficients, will the original regression still meaningful? thanksAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-6800786963001807402015-01-15T09:44:16.178-08:002015-01-15T09:44:16.178-08:00Both versions 8 and 9 provide standard errors (and...Both versions 8 and 9 provide standard errors (and hence CI's) for the estimated parameters. Version 9 allows forecasting, and provides the forecast standard errors (and hence forecast intervals).Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-53934773512428606512015-01-15T07:58:42.573-08:002015-01-15T07:58:42.573-08:00Does Eviews9 offer a computation of confidence int...Does Eviews9 offer a computation of confidence interval for bayesian VARs?Imenhttp://www.blogger.com/profile/01909464636406148093noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-58376495842436159852015-01-13T09:59:14.653-08:002015-01-13T09:59:14.653-08:00I totally agree!I totally agree!Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-7369061234589835542015-01-13T09:32:41.118-08:002015-01-13T09:32:41.118-08:00If X is cointegrated with Y, then Y MUST be cointe...If X is cointegrated with Y, then Y MUST be cointegrated with X.Dave Gileshttp://www.blogger.com/profile/05389606956062019445noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-35023147515134458482015-01-13T06:49:52.749-08:002015-01-13T06:49:52.749-08:00Well, there are a lot of researchers from economic...Well, there are a lot of researchers from economics, doing some really hard work on the economic foundation and then, while sticking to theory, the statistical part of their research is right from the 101-course, neglecting almost every recent advance from statistical theory. In essence, almost everything is linear in the end ...<br /><br />On the other hand it seems to be tempting to have all those big hammers from statistics:searching for economic nails leads directly to "data-driven" modeling. Often then "fit" or "predictive accuracy" in some sense is used to conclude that the latter models are deemed more suitable compared to theory-driven models.<br /><br />Not only the "statistical sword" should not be yielded blindly. Also there is no excuse for sticking to theory and ignoring modern statistics. The latter also produces garbage.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-70208918252116875672015-01-13T00:51:24.021-08:002015-01-13T00:51:24.021-08:00Sir, Can we apply ARDL bound testing to analyze tw...Sir, Can we apply ARDL bound testing to analyze two co-integration relations between the same variables (x and y). e.g. X c Y and Y c X. JAWAD KAZMIhttp://www.blogger.com/profile/09872901806824420552noreply@blogger.comtag:blogger.com,1999:blog-2198942534740642384.post-8608277862030938842015-01-12T10:49:50.532-08:002015-01-12T10:49:50.532-08:00This was also the conclusion of presentation (ASSA...This was also the conclusion of presentation (ASSA 2015) by Professor Amir Sufi (available at http://events.mediasite.com/Mediasite/Play/7d2207a4b2f049f4a9473b21752591221d after 1 hr 15minute)<br /><br />"Theory and empirics are both necessary, and they should go hand in hand"Anonymousnoreply@blogger.com