14 - Dynamic Econometric Modelling of Decisions under Uncertainty

Abstract

Dynamic model specification and testing have received much attention in the econometrics literature. This chapter is concerned with the developments in structural econometric modeling and time series analysis (SEMTSA), which provides a synthesis of econometric and time series methods in modeling economic time series. It further discusses the main features of SEMTSA and their relationships with recent contributions to econometrics and time series analysis. The research on SEMTSA has shown that a linear dynamic simultaneous equation models (SEM) in which the exogenous variables are generated by an autoregressive integrated moving average (ARIMA) process is a special case of a multivariate ARIMA model. With some modification, statistical estimation and testing procedures developed for multivariate time series processes can, therefore, be used to analyze dynamic SEMs. Conditions for identification and criteria for order determination of multivariate time series models also apply to SEMs with exogenous variables generated by ARIMA schemes. Nonlinear dynamic models have also become popular in econometrics. For instance, the autoregressive conditional heteroscedastic (ARCH) model has been applied in various areas, for example, in the estimation of the variance of inflation or to model time-varying risk premia in asset markets.