Testing For and Dating Common Breaks in Multivariate Time Series

摘要: This paper develops methods for constructing asymptotically valid confidence intervals the date of a single break in multivariate time series, including I(0), I(1), and deterministically trending regressors. Although width asymptotic interval does not decrease as sample size increases, it is inversely related to number series which have common date, so there are substantial gains inference about dates. These applied two empirical examples: mean growth rate output three European countries, U.S. consumption, investment, output.