MULTIEQUATION TIME-SERIES MODELS
As we have seen in previous chapters, you can capture many interesting dynamic relationships using single-equation time-series methods. In the recent past, many time-series texts would end with nothing more than a brief discussion of multiequation models. Yet, one of the most fertile areas of contemporary time-series research concerns multiequation models. This chapter has four specific aims:
- Introduce intervention analysis and transfer function analysis. These two techniques generalize the univariate methodology by allowing the time path of a dependent variable to be influenced by the time path of an independent or exogenous variable. If it is known that there is no feedback, intervention and transfer function analyses can be very effective tools for forecasting and hypothesis testing.
- Introduce the concept of a vector autoregression (VAR). The major limitation of intervention and transfer function models is that many economic systems do exhibit feedback. In practice, it is not always known if the time path of a series designated to be the independent variable has been unaffected by the time path of the dependent variable. The most basic form of a VAR treats all variables symmetrically without making reference to the issue of dependence versus independence.
- Show that tools employed by a VAR analysis—Granger causality, impulse response ...