Text this: Structural versus unrestricted reduced-form econometric modelling :