Blackwell Reference Online

Extract

The chapter addresses the situation in which an economic variable, for which a time series of observations is available, can be seen as the combination of several components. Having only observations on the aggregate variable, the analyst wishes to learn about the unobserved components and, in particular, about the joint distribution of their estimators and forecasts. Unobserved components in time series have been of interest to economists for some time, and a good review of the early developments and applications is contained in Nerlove, Grether, and Carvalho (1979) . Although the treatment and modeling of unobserved components (be that in the context of business cycle analysis or of seasonal adjustment, to quote two important applications) lack an economic theory foundation (see Ghysels, 1990 ), they are nevertheless used extensively by economists and statisticians when dealing with economic time series. The interest has developed along two separate (although related) fronts. First, unobserved component models are used in economic research in a variety of problems when a variable, supposed to play some relevant economic role, is not directly observable. For example, unobserved components have been used in modeling agents’ reaction to (permanent or transitory) changes in the price level ( Lucas, 1976 ), in analyzing the stability of some of the macroeconomic “big ratios” ( Pagan, ... log in or subscribe to read full text