Selecting Lag Length for the Johansen Cointegration Test

Summary

The document asks how to choose the lag length when applying the Johansen cointegration test to two time series. It notes that test conclusions can vary with the lag specification: some choices may lead to rejection of the null hypothesis while others do not. The question seeks a data-based procedure for selecting the lag rather than choosing a value arbitrarily.

The sole answer points to a likelihood-ratio testing routine in a Matlab spatial econometrics toolbox. The routine performs a sequence of tests from a user-specified maximum lag down to a user-specified minimum lag. This provides one possible way to compare lag specifications, but the document does not explain the decision rule, establish that the procedure is appropriate in every setting, or report an application. It is therefore a brief pointer rather than a complete guide; practitioners must still set the search range and assess whether the selected model is suitable for their data.

Key ideas

• Johansen test results can change when the lag specification changes.
• The question seeks a data-based way to select the lag for a two-series test.
• A cited toolbox routine tests a sequence of lags between user-defined maximum and minimum values.
• The discussion does not provide a full selection rule or evidence from an empirical application.

Tags

# What is the correct procedure to choose the lag when preforming Johansen cointegration test?


# What is the correct procedure to choose the lag when preforming Johansen cointegration test?












When preforming Johansen cointegration test for 2 time series (the simple case) you need to decide the lag you want to use. Doing the test for different lag levels returns different results: for some lag levels the null hypothesis can be rejected but for others it can't.

My question is what is the right method based on the input data to decide what lag I need to use?

## Answer by mapsa (score 2)

https://quant.stackexchange.com/a/12630

If you are using Spatial Econometrics toolbox in Matlab you could use the lrratio function which implements a sequence of such tests beginning at a maximum lag (specified by the user) down to a minimum lag (also specified by the user). (more info in http://fmwww.bc.edu/ec-p/software/matlab/mbook.pdf)

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.