Measures of conditional linear dependence and feedback between time series. (English) Zbl 0553.62083

Geweke, John F.

J. Am. Stat. Assoc. 79, 907-915 (1984).

The author extends the scope of his previous works by introducing measures of linear dependence and feedback for two multiple time series conditional on a third one. Such extension is suitable for various situations in macroeconomics. The measures are non-negative and can be expressed in terms of measures of unconditional feedback between various combinations of the three series.
Moreover, the decomposition of directional conditional linear feedback is suggested. The conditional measures can be estimated by combining the canonical factorization algorithm of Whittle with a parametric bootstrap (point estimates with a first order correction for small sample bias and approximate confidence intervals for these estimates are obtained). The theory is demonstrated by means of an empirical example with real output, money and interest rates.

Reviewer: T.Cipra

MSC:

62M15	Inference from stochastic processes and spectral analysis
91B84	Economic time series analysis
62M10	Time series, auto-correlation, regression, etc. in statistics (GARCH)
62P20	Applications of statistics to economics

Keywords:

conditional linear dependence; resampling; macroeconomics; spectral density; measures of linear dependence; feedback; multiple time series; decomposition; factorization algorithm of Whittle; bootstrap; point estimates; approximate confidence intervals