# 1 look at the the time series for xt 2 look for

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

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ocesses. They found β to be highly signiﬁcant even though it should be close to 0. Basically, one gets spurious results. The standard errors and test statistics of the normal t-tests are invalid. 14 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Spurious Regressions Why do we care about nonstationarity? Because it can mess up OLS! Refer to the famous simulations by Granger and Newbold. Yt = β1 + β2 Xt + εt where X and Y are independent random walk processes. They found β to be highly signiﬁcant even though it should be close to 0. Basically, one gets spurious results. The standard errors and test statistics of the normal t-tests are invalid. 15 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Spurious Regressions Why do we care about nonstationarity? Because it can mess up OLS! Refer to the famous simulations by Granger and Newbold. Yt = β1 + β2 Xt + εt where X and Y are independent random walk processes. They found β to be highly signiﬁcant even though it should be close to 0. Basically, one gets spurious results. The standard errors and test statistics of the normal t-tests are invalid. 16 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Spurious Regressions Why do we care about nonstationarity? Because it can mess up OLS! Refer to the famous simulations by Granger and Newbold. Yt = β1 + β2 Xt + εt where X and Y are independent random walk processes. They found β to be highly signiﬁcant even though it should be close to 0. Basically, one gets spurious results. The standard errors and test statistics of the normal t-tests are invalid. 17 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Autocorrelation function ADF Testing for Nonstationarity There are a few ways of detecting nonstationarity of which were covered in the problem sets. 1 look at the the time series for Xt 2 look for Durbin Watson statistics close to 0 3 look at autocorrelation function of Xt 4 look at the Augmented Dickey Fuller test statistic 18 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Autocorrelation function ADF Testing for Nonstationarity There are a few ways of detecting nonstationarity of which were covered in the problem sets. 1 look at the the time series for Xt 2 look for Durbin Watson statistics close to 0 3 look at autocorrelation function of Xt 4 look at the Augmented Dickey Fuller test statistic 19 / 52 Introduction Stationary Processes Nonstationary Processes Spurious Regressions Testing for Nonstationarity Cointegration Autocorrelation function ADF Testing for Nonstationarity There are a few ways of detecting nonstationarity of which were covered in the problem sets. 1 look at the the time series for Xt 2 look for...
View Full Document

## This document was uploaded on 03/12/2014 for the course ECON 202 at University of London University of London International Programmes (Distance Learning).

Ask a homework question - tutors are online