EViews.com

Hello,

I have used the state space object in Eviews 6 to derive the time varying NAWRU (non-accelerating wage rate of unemployment). The NAWRU is integral to other equations in my main model object, so I would like to refer to it within that model. Is it possible to merge a state space object into a model object so that it can all be treated endogenously? If so, how do I do it ? The standard merge and append tags for the model object do not seem to work in this case.

Thanks,
Phil

The problem is that the model object can't handle the unobserved components in the state space specification. The algorithms for forecasting from a state space object are pretty specific to that object and the model doesn't understand how to deal with the states. The best you can do is to do the forecasts for the state space first, then incorporate the relevant terms into your model. Unfortunately, you won't be able to build system dependencies between the observed in your sspace and the other endogenous in the model in this way...

Ah, shame. I've used the two step approach instead. Thanks Glenn.