error autocorrelation

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

error autocorrelation

Postby lisa » Mon Oct 12, 2020 1:43 am

Please I want to estimate a VAR model with 5 stationary variables but the residuals of the model are correlated. How can I do?

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

Re: error autocorrelation

Postby startz » Mon Oct 12, 2020 6:20 am

Add more lags

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: error autocorrelation

Postby lisa » Thu Oct 15, 2020 10:20 am

I have the same problem with all lags (2, 3, ..., 12). The 5 variables of my VAR model (dlog(tax revenue), dlog(public expenditure), log(real gdp), dlog(gdp deflator) and d(money market rate) are stationary; but the residuals of the model are autocorrelated (the probability of the Portmanteau Autocorrelation is "0" for all lags).

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

Re: error autocorrelation

Postby startz » Thu Oct 15, 2020 10:58 am

Perhaps post your workfile, including the VAR equation.

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: error autocorrelation

Postby lisa » Thu Oct 15, 2020 11:45 am

I have a missing variable. Thank you the problem is solved.
Attachments
workfile - Copie.wf1
(49.88 KiB) Downloaded 267 times
Last edited by lisa on Fri Oct 16, 2020 5:00 am, edited 1 time in total.

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

Re: error autocorrelation

Postby startz » Thu Oct 15, 2020 12:21 pm

In the workfile you only have 2 lags. What happens if you include several more?

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: error autocorrelation

Postby lisa » Thu Oct 15, 2020 12:26 pm

I have the same problem with more lags. I estimated the model with 4, 6, 7, 10 and 14 lags and I have the same problem of residual autocorrelation.

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

Re: error autocorrelation

Postby startz » Thu Oct 15, 2020 1:00 pm

You might try running separate OLS regressions. That at least would give an idea of which equations it is.


Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 23 guests