“…If z is not 1(0), i.e., if x, Yt are not cointegrated, then the z. term does not belong in these equations given that the dependent variables are 1(0) and hence at least one of Pi, p2 does not vanish. These models were introduced into economics by Sargan (1964) and Phillips (1957) and have generated a lot of interest following the work of Davidson, Hendry, Srba and Yeo (1978), von Ungern Sternberg (1980), Curry (1981), Dawson (1981) and Salmon (1982) amongst others. The models are seen to incorporate equilibrium relationships, perhaps suggested by an economic theory of the long-run, with the type of dynamic model favoured by time-series econometricians.…”