The Design of Monte Carlo Experiments for VAR Models
This paper deals with the design of Monte Carlo experiments in the context of cointegrated VAR models. Such experiments often seek to establish the applicability of asymptotic distributional results for sampled of size 100 to 200, which are typical of macroeconomic times series. Hithertofore, the design of such experiments has relied on certain simple models given in Bannerjee et al. (1986), Engle and Granger (1987), and Phillips (1991). Here we provide the framework for designing experiments based on much more general models, of which the designs above are special cases.
- econ_9596_014.pdf application/pdf 426 KB Download File