Issues to discuss • No consensus on good econometric practise • Reality much more rich and complex than the theory • The ceteris paribus in economic and econometric modelling The specific-to-general (VAR) • • • • • ceteris paribus variables dumped into the error term low ability to recject a theory model when inadeqate Few (constant) parameters (by assumption) No free parameters Known expectations formation The general-to-specific • Combination of induction and deduction • Weak link between theory model and reality • Can adequately account for unit root nonstationarity structural breaks, such as shifts in equlibrium mean and growth rates A discussion of some basic principles for empirical research using as an illustration Friedmans claim that ’Inflation is always and everywhere a monetary problem’ • Imbedding the theory model in a broader empirical framwork: the pulling and pushing forces • Formulating a set of empirically testable hypotheses on different levels of generality • Deductive inference: testing prior hypotheses • Inductive inference: generating new hypotheses • Reformulating the empirical problem/ the theoretical problem, or both Pulling and pushing forces Imbedding Romer's money demand model in the VAR The pushing forces: First level: Second level: The pulling forces First level: Because: Second level: The restrictions on β contain hypotheses on: • money demand • the Fisher parity • the expectations hypothesis The restrictions on α contain hypotheses on: • pushing behavior (weak exogeneity) • adjustment behavior (unit vectors in α) The role of expectations • for the specification of the long-run relations • for the short-run dynamics The Fisher parity The expectations hypothesis To sum up: Deductive inference Inductive inference Conclusions • The deductive part is based on testable implications of a theory model translated into a set of hypotheses in a VAR • The inductive part treats any discrepancies between theory and the data as a useful piece of information, allowing us to adjust our intuition of how the economic and the empirical model work together: generates new hypotheses • Generates a set of ’sophisticated’ stylized facts • The sensitivity of theory based conclusions to the ceteris paribus assumption