Speed of adjustment in vecm greater than 1 in magnitude

[benlambert18785]

Hi,

I have been estimating a VECM using EViews and have found that I sometimes get values of the speed of adjustment parameter that are greater than 1 in magnitude. The Cointegrating relationship (there are only two endogenous variables) has a time trend within it.

Does anyone know how/why this happens? Also what does it mean? I thought this parameter had to be less than 1 in magnitude.

Best,

Ben

[KerryGMcC]

It should indeed be less that one, is the result significant - if not, nothing to worry about :D