Dummy near matrix error

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

rafrods
Posts: 17
Joined: Mon Apr 05, 2021 5:32 pm

Dummy near matrix error

Postby rafrods » Tue Apr 06, 2021 3:01 am

Hello. I am estimating a regression between the GDP and social spending of municipalities and I am trying to insert a dummy variable for municipalities in the metropolitan region (municipalities in the metropolitan region = 1; municipalities outside the metropolitan region = 0). But Eviews is displaying the near matrix error message when I use a panel regression for fixed effects or random effects, and when I select neither effect, the error message does not appear. However, when I analyze the correlation matrix, it is clear that there is no perfect correlation between the dummy variable and the other variables in the equation. What could be causing this error?

The equation is LogGDP = LogEducation LogHealth DummyGV

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Dummy near matrix error

Postby startz » Tue Apr 06, 2021 7:47 am

Are the fixed effects perfectly correlated with the municipality dummy?

rafrods
Posts: 17
Joined: Mon Apr 05, 2021 5:32 pm

Re: Dummy near matrix error

Postby rafrods » Tue Apr 06, 2021 2:06 pm

I think so because in the none specification effects or random effects format the error message does not appear. Although the panel data tests (chow, breusch-pagan, and hausman) tell me to estimate the panel in fixed effects.
Attachments
random effects.png
random effects.png (30.36 KiB) Viewed 3848 times
fixed effects.png
fixed effects.png (28.71 KiB) Viewed 3848 times
none effects specification.png
none effects specification.png (25.12 KiB) Viewed 3848 times

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Dummy near matrix error

Postby startz » Tue Apr 06, 2021 2:37 pm

I believe that EViews is telling you correctly that you can't have both the municipality dummy and cross-section fixed effects. Dummy variable trap.

rafrods
Posts: 17
Joined: Mon Apr 05, 2021 5:32 pm

Re: Dummy near matrix error

Postby rafrods » Tue Apr 06, 2021 3:39 pm

Okay. So is it just discarding the variable? It is not essential for regression, I was just looking to see if it would make any difference whether the municipality was in the metropolitan region or not. Thank you for your attention.

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Dummy near matrix error

Postby startz » Tue Apr 06, 2021 3:48 pm

You can't estimate an intercept for each region and then also separately estimate the effect of being in a metropolitan region. So yes, you probably want to delete the dummy.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 23 guests