The Economics of Var Models

In the minds of some economic theorists and traditional econometricians, the vector autoregressive (VAR) approach to time-series data is unscientific, obscure, confusing, or simply wrong. Since the publication of Sims’s original contributions (1972, 1980a, 1980b, 1982), the methodology has spurred endless debates. Critics claim that the methodology has very little relationship with economic theory, relies on a set of unsustainable assumptions, and is fundamentally flawed, being subject to the well known Lucas’s critique. But despite the controversies surrounding the use and interpretation of VAR models throughout the 1980s, they appear to have found a permanent position in the tool kit of applied time-series and macroeconomic analysts. VARs are currently used as a tool to summarize data interdependences, to test generically formulated theories, to conduct policy analyses, and, more recently, as a way to compare actual data with the time series generated by artificial economies with calibrated parameters.