Seemingly Unrelated Regressions and robust covariance matrix
Moderators: EViews Gareth, EViews Moderator
-
- EViews Developer
- Posts: 2672
- Joined: Wed Oct 15, 2008 9:17 am
Re: Seemingly Unrelated Regressions and robust covariance matrix
One possibility is that you are using automatic bandwidth selection. The equation by equation automatic selection will probably give you different lags than doing the lag selection on the system of equations.
-
- Posts: 1
- Joined: Wed Sep 26, 2012 10:08 am
Re: Seemingly Unrelated Regressions and robust covariance ma
EViews Glenn wrote:I guess I need to be a bit clearer on what you want (and what I mean ). The seemingly unrelated refers to the fact that you have a set of equations with no apparent cross-equation restrictions, but with non-zero off-diagonals.
For the purposes of this discussion there are couple of ways to proceed:
. You can estimate the specification using a GLS approach which corrects for cross-sectional heteroskedasticity and contemporaneous correlation (but not for general heteroskedasticity and serial correlation). In principle, you could follow this with a robust standard estimator.
. Alternately, you can estimate using system least squares without correlation correction and then compute with a robust standard estimator, for example a system HAC estimator.
EViews does not allow you to take the former approach, but does allow you to do the latter using the GMM tools. Note that the equivalence results from treating all of the explanatory variables in your specification as exogenous. Just as TSLS using the original regressors as instruments yields the least squares estimator, so too does GMM with the appropriate orthogonality conditions and weighting matrix (what is termed in the dialog TSLS weighting), yield the system least squares estimator. Then all you have to do is to select the appropriate robust covariance option.
Glenn, thanks for your suggestion but I have encountered to another problem when I tried to do system GMM. In short, suppose I have a system of two equations (SUR setup) and when I do system GMM (with 2SLS & GMM Robust SE), the p-value of every estimated coefficient exploded to 0.9+. Which is odd to me because when I do either single equation OLS with HAC or single equation GMM with HAC individually (as expected, they matched almost exactly), many coefficients are significant with practically 0 p-value. Does it imply something about the cross-sectional correlation between residuals? Any suggestions of how to proceed?
Re: Seemingly Unrelated Regressions and robust covariance ma
john_stranger wrote:Glenn, thanks for your suggestion but I have encountered to another problem when I tried to do system GMM. In short, suppose I have a system of two equations (SUR setup) and when I do system GMM (with 2SLS & GMM Robust SE), the p-value of every estimated coefficient exploded to 0.9+. Which is odd to me because when I do either single equation OLS with HAC or single equation GMM with HAC individually (as expected, they matched almost exactly), many coefficients are significant with practically 0 p-value. Does it imply something about the cross-sectional correlation between residuals? Any suggestions of how to proceed?
I'm encountering exactly the same issue. As soon as I switch to a GMM system, the p-values explode into insignificance. Whereas every other estimation technique delivers estimates where the coefficients and p-values aren't dissimilar. Very frustrating.
-
- EViews Developer
- Posts: 2672
- Joined: Wed Oct 15, 2008 9:17 am
Re: Seemingly Unrelated Regressions and robust covariance ma
HAC for the system is different than HAC equation by equation. As these posts have all noted, there are now cross-equation covariances to worry about. How many observations do you have?
Re: Seemingly Unrelated Regressions and robust covariance ma
EViews Glenn wrote:HAC for the system is different than HAC equation by equation. As these posts have all noted, there are now cross-equation covariances to worry about. How many observations do you have?
Aah, probably not enough. 30 for 1 equation, 32 for the other.
Re: Seemingly Unrelated Regressions and robust covariance matrix
Hi,
I am trying to estimate a two equation SUR system sys1 as follows
y c1 x1
z c2 x2
how do I specify the system and estimate sys1.sur? I can't find a sample program in the manual to run sur.
thank you so much, Howard
I am trying to estimate a two equation SUR system sys1 as follows
y c1 x1
z c2 x2
how do I specify the system and estimate sys1.sur? I can't find a sample program in the manual to run sur.
thank you so much, Howard
-
- Non-normality and collinearity are NOT problems!
- Posts: 3775
- Joined: Wed Sep 17, 2008 2:25 pm
Re: Seemingly Unrelated Regressions and robust covariance matrix
Did you type "sur" into the help system?
Re: Seemingly Unrelated Regressions and robust covariance matrix
yes. all I see in the help manual are running sur interactively. I have many equations and must write a program.
I am looking for a sample program in the manual illustrating the codes for specifying a simple two equation system of sur. thank you very much.
I am looking for a sample program in the manual illustrating the codes for specifying a simple two equation system of sur. thank you very much.
-
- Non-normality and collinearity are NOT problems!
- Posts: 3775
- Joined: Wed Sep 17, 2008 2:25 pm
Re: Seemingly Unrelated Regressions and robust covariance matrix
Typing sure into the index in the help system gives you the command to estimate a system by sur. The help under system gives the various commands needed to set up the system. Look at Declare and Append specification line.
Who is online
Users browsing this forum: No registered users and 65 guests