Interpreting Johansen Cointegration Test Results

Summary

The document addresses how to interpret coefficients after applying the Johansen cointegration test in a linear time-series setting. The answer advises using the original variables rather than truncating or otherwise altering them for interpretation. It also cautions against assuming that detecting cointegration means one should simply run a regression and treat its result as the long-run relationship.

The response notes that Johansen’s method is needed when there is more than one independent variable, though it does not explain the test’s assumptions, rank selection, normalization, or how to estimate and interpret the resulting cointegrating relations. No empirical example or supporting derivation is provided. The guidance is therefore a brief pointer rather than a full procedure, and readers still need additional detail to conduct and interpret a Johansen analysis correctly.

Key ideas

• The answer recommends interpreting the test with the original variables rather than truncating them.
• Cointegration detection alone does not provide a complete procedure for estimating a long-run relationship.
• The answer identifies multiple independent variables as the setting where Johansen’s method is needed.
• The response does not cover rank selection, model assumptions, or coefficient normalization in detail.

Tags

# How does one use the Johansen cointegration test in a linear time series model?


# How does one use the Johansen cointegration test in a linear time series model?












How does one use the Johansen cointegration test in a linear time series model?

Should I only use normalized coeffients for interpretation? Or, once I know that the variables are cointegrated, do I simply regress the variables and consider it the long-run relationship?

## Answer by 4pie0 (score 1)

https://quant.stackexchange.com/a/7852

no, you should use your original variables, no truncating, normalizing or whatever. And remember that you need Johansen only in case of more than one independent variable.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.