# Including lagged dependent in DOLS to estimate R&D effects

I have estimated a DOLS model which seems to have a good fit. My aim with the model is to test alternative scenario's where I change some of the parameters ceteris paribus and analyse the impact. The only issue is that in my baseline scenario, the model basically produces a horizontal line since there is no autoregressive property in the model and the variables are kept constant (see figure below). Since it is quite unreal that my dependent variable will continue at the same level, I would like to have some sort of autoregressive property. A lagged dependent sucks up all the variability in the model however (R2=1, other coefficients are 0). Is there any way I could include some sort of autoregression in my model without including a lagged dependent?

Many thanks, Erik-Jan