I am running a test for cointegration on some data using the R package URCA and the ca.jo function. In that function, the ecdet parameter allows you to include either a constant or trend term, but not both. Is there any way to include both? this seems like pretty obvious use case, and one might assume that including the trend would include both, but it doesn't if your two timeseries have a spread that is trending apart and you want to use the trend term, unless the trend has an intercept of exactly 0, you are out of luck!

I trie to get around this by using dummy variables instead. For example a column representing time, seq(1,nrow(dataMat)). Combined with ecdet="const" I figured this would do the trick. Unfortunately, I am not able to recover the coefficients of these dummy variables using ca.jo. I can't see anywhere in the output that these coefficients are produced.

Does anyone know how I can recover the coefficients of the dummy variables? Or help me find some other way to include both the constant and trend terms in this cointegration test?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.