This preview shows page 1. Sign up to view the full content.
Unformatted text preview: endent of time
However Variance of time series of X at time t depends on time t. So variance depends on
time and is not stationary. Key thing
keep past with us. That means error
terms goes to . In result we get bad standard errors.
2) Random walk with a drift:
is a random walk
is a random walk with a drift Stationarity fails because is function of time Space for figure (1) 3) Deterministic Trend depends on time, therefore not stationary. Space for figure (2)
Errors are clustered around a trend line
In random walk with drift
Space for figure (3)
It has increasing trend but no trend line.
4) Making Non Stationary Stationary : a) Difference Stationary
Random walk with drift Difference data
(It does not depend on time)
(It does not depend on time)
This is integrated of order 1, made stationary by differencing the data.
This is written as I(1)
Integrated of order 1
General form of writing time series is ARIMA
Stata
In stata , sort by trend
Sort trend
Gen dX = dX
Br trend X dX
b) Trend Stationary
̂ This is detrending of the data
̂ Define ̃
̃ is “ Trend Stationary”
(̃)
(̃)
Therefore, Stationary....
View
Full
Document
This note was uploaded on 10/18/2013 for the course EC 421 taught by Professor Gaus during the Fall '08 term at University of Oregon.
 Fall '08
 GAUS

Click to edit the document details