Vector Autoregressive (VAR] models which do not rely on a recursive model srtructure are discussed. Linkages to traditional dynamic simultaneous equations models are developed which emphasize the nature of the identifying restrictions that characterize VAR models. Explicit expressions for the Score and Informtion functions are derived and their role in model identification, estimation and hypothesis testing is discussed.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados