8. Non-stationary Time series

Iii consequences of non stationary1 inconsistence

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: III) Consequences of Non- Stationary:1. Inconsistence Estimator: ̄ )( ( ( ̄) ̄) Consistency arises in OLS if data are “weakly persistent”( Stationary) if not then inconsistent. 2. Spurious Regression: a) Deterministic Trend You get huge R2 , huge t’s “OUCH” b) Random Walk with drift mistakenly run You get huge R2 , huge t’s “OUCH” IV) Detection:Look at data & look for trending variable Use correlogram (to see for trending variable) √ It is correlation between X and time t at some different time period. Ex- 1 | | This says it goes away with time. We expect correlation between today and yesterday to be stronger than correlation between today and three time period. Space for figure 1 Ex-2 M A (1) = moving average of errors that goes one period. Space for figure 2 Ex-3 Space for figure 3 This means it hold out for few time periods. This is stationary process because But you can see that past is sticking with time. So you don’t know? whether it is stationary or not. Therefore we have to test whether ther...
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.

Ask a homework question - tutors are online