Stata and Eviews different results

shakyyy87 · Postby **shakyyy87** » Sun Jul 29, 2012 4:21 pm

Hello ,

I have a confusion regarding results. I used to work in stata for estimating the results, but since i started using Eviews software i get a different result compare to the result of stata 10. For example :

one of the independent variable is showing stationary in level when i use stata 10 , while the same variable in eviews software 7 gives me non stationary in level. This is one of the example i shared. Let me know what could be the reason.

Thank you.

EViews Gareth · Postby **EViews Gareth** » Sun Jul 29, 2012 5:14 pm

We'll need a lot more information. For a start, telling us what routine you are performing would be useful.

shakyyy87 · Postby **shakyyy87** » Sun Jul 29, 2012 7:05 pm

Thank you very much for the reply. But could you tell me in more detail , what do you mean by routine ?

EViews Gareth · Postby **EViews Gareth** » Sun Jul 29, 2012 7:14 pm

What are you actually doing?

EViews and Stata have thousands of calculations built in. Hard to say what is causing the difference if you don't tell us what you're actually calculating!

shakyyy87 · Postby **shakyyy87** » Sun Jul 29, 2012 7:38 pm

Oh okay. Well as i said before, i was comparing the t-stats with the critical value , that is - i was doing the unit root test.
Doing in the stata with the command : Dfuller (var) i find one of the variable as stationary in level. Since Eviews gives you the direct result , i found out the same variable with obviously the same data as non stationary in level.

With Stata i got the value of my variable as -4.282 that is when i performed dicky fuller test for unit root. This makes my variable or series as stationary in level. While Eviews showed me as -1.6607 which makes the variable as NON-stationary in level. All other variables (five) are completely different with stata.

Is there anything you could suggest me ?

EViews Gareth · Postby **EViews Gareth** » Sun Jul 29, 2012 7:56 pm

Ah. You didn't mention you were doing a unit root test.

Could you provide your workfile?

shakyyy87 · Postby **shakyyy87** » Sun Jul 29, 2012 8:24 pm

You mean you required my data ? if that is so , then please let me know where do i send ?

EViews Gareth · Postby **EViews Gareth** » Sun Jul 29, 2012 8:34 pm

Just attach it to a post here on the forum.

shakyyy87 · Postby **shakyyy87** » Sun Jul 29, 2012 8:48 pm

There you go !

shakyyy87 · Postby **shakyyy87** » Mon Jul 30, 2012 7:26 am

Any Updates ?

startz · Postby **startz** » Mon Jul 30, 2012 7:53 am

Any Updates ?

I don't know what time zone you're posting from, but in California you posted late Sunday evening and asked if there has been an update early Monday morning. That's not reasonable.

EViews Gareth · Postby **EViews Gareth** » Mon Jul 30, 2012 7:58 am

You still haven't given us enough information to replicate what you're doing.

Tell us exactly what you have done.

shakyyy87 · Postby **shakyyy87** » Mon Jul 30, 2012 8:13 am

Hello ,

I am Posting this from London. I was not aware of the time in US. Anyhow , as i said before " I am testing the Unit root" To know my variables if it is stationary or non stationary using Dickey Fuller test. The data is time series. Testing the variables in stata is completely different while doing the test in Eviews Software. I have also send you the xls. file to see if you guys can do anything about this.

Hope that makes it clear , let me know what else you need.

Thank you.

shakyyy87 · Postby **shakyyy87** » Mon Jul 30, 2012 8:22 am

Also if you could test the variable "Export" in both stata and eviews , using Dickey fuller test you will find different answers. Export is showing "stationary" in level while using stata , but "non stationary" in level in Eviews software.

startz · Postby **startz** » Mon Jul 30, 2012 8:29 am

I get identical results in EViews 7.2 and Stata 11.

EViews.com