ols

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

marta
Posts: 8
Joined: Mon Jul 27, 2009 2:54 pm
Contact:

ols

Postby marta » Fri Jul 31, 2009 2:38 am

Hi,
I am working on my thesis about the hedging effectiveness of the european electricity markets, but I have some problems since it's my first time I use eviews. My version is evews5. First of all, I am working on the returns and not on the prices of the electricity. But when I run an OLS estimation the results forthe prices are quite good (in terms of R square), but when I try with the returns, the reslts are not well at all. My estimation include also the multivariate garch?
what do you think I should use: prices or returns?
Thanks
Marta
eex spot.wf1
(73.33 KiB) Downloaded 367 times

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: ols

Postby trubador » Fri Jul 31, 2009 4:00 am

Results are expected since both your variables are appear to be non-stationary. You can search the forum starting from here.

marta
Posts: 8
Joined: Mon Jul 27, 2009 2:54 pm
Contact:

Re: ols

Postby marta » Fri Jul 31, 2009 4:27 am

trubador wrote:Results are expected since both your variables are appear to be non-stationary. You can search the forum starting from here.


Thanks for your help. If I have understood well, both prices of spot and futures are non-stationary, but when I consider the returns (as the first difference of logarithm), they appear to be stationary, isn't it right?

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: ols

Postby trubador » Fri Jul 31, 2009 5:12 am

Well, I am afraid it will be very difficult to make electricity prices data stationary, since the data will probably be very persistent (e.g. high serial correlation) due to the properties of the market. If I were you, I would consider non-stationary methods and/or multivariate variance modeling techniques. Nevertheless, other members of the forum may have a better solution...

marta
Posts: 8
Joined: Mon Jul 27, 2009 2:54 pm
Contact:

Re: ols

Postby marta » Fri Jul 31, 2009 6:33 am

trubador wrote:Well, I am afraid it will be very difficult to make electricity prices data stationary, since the data will probably be very persistent (e.g. high serial correlation) due to the properties of the market. If I were you, I would consider non-stationary methods and/or multivariate variance modeling techniques. Nevertheless, other members of the forum may have a better solution...


what non-stationary or multivariate variance models do you suggest?


Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 16 guests