Page 1 of 1
Time series and non-stationary variables?
Posted: Mon Dec 15, 2008 3:33 am
by nisse
I have time series data for the oilprice. When I do a Dickey Fuller test it shows that the serie is non-stationary. I wonder how I can make the serie stationary? Don´t know if this is the forum to ask in, but I will be very grateful for help.
/Nisse
Re: Time series and non-stationary variables?
Posted: Wed Dec 24, 2008 5:52 am
by nightwalk
Your time series can be made stationary by differencing. The number of times a series needs to be differenced in order to achieve stationarity depends on the number of unit roots it contains. For example suppose you have a y non-stationary series with d unit roots, you need to difference your series d times to induce stationarity. I suggest you take the first differences and normaly you will get stationary series. Under the Dickey-Fuller function in Eviews, try to use first-differences instead of levels.
Re: Time series and non-stationary variables?
Posted: Sat Aug 01, 2015 9:55 am
by sbkb
can you tell me the way of including the first differences of data - ex. y c x1 x2 x3 are levels, and dy c dx1 dx2 dx3 are first differences?
And if i use log of data will i get stationary data form non stationary data integrated of (I)1?