Cointegration and Error Correction Models

Question 1

Engle-Granger

Answer 1

• two step process to test for cointegration between two or more non-stationary time series

Question 2

What is an error correction model used for?

Answer 2

• to model the short-term dynamics of cointegrated time series, while accounting for the long-run relationship

Question 3

error correction term

Answer 3

• measures the extent to which the previous period’s deviation from equilibrium influences the current period

Question 4

Steps to implement ECM

Answer 4

• Check for stationarity, ensure the series are I(1)
• Test for cointegration, engle-granger
• Estimate the cointegration equation
• Fit the ECM

Question 5

Engle-Granger Procedure

Answer 5

• Test that each series is I(1)
• Estimate the Long-run relationship using OLS regression
• Test the residuals for stationarity

Question 6

Step 1 Engle-Granger

Answer 6

• Perform unit root tests to check if the time-series are I(1)

Question 7

Step 2 Engle-Granger

Answer 7

• Estimate the long-run relationship using OLS regression

Question 8

Step 3 Engle Granger

Answer 8

• Test the residuals of the regression for stationarity
• if they are stationary then the series are cointegrated

Question 9

Box-Jenkins

Answer 9

• Model identification and selection
• parameter estimation
• model diagnostics