标题：Testing for the Presence of a Random Walk in Series with Structural Breaks - (Now published in 'Journal of Time Series Analysis', 22 (2001), pp.127-150.)
出版社：Suntory Toyota International Centres for Economics and Related Disciplines
摘要：The paper considers tests for the presence of a random walk componentin a stationary or trend stationary time series and extends them to serieswhich contain structural breaks. The locally best invariant (LBI) test isderived and the asymptotic distribution obtained. Then a modified teststatistic is proposed. The advantage of this statistic is that its asymptoticdistribution is not dependent on the location of the breakpoint and itsform is that of the generalised CramJr-von Mises distribution, withdegrees of freedom depending on the number of breakpoints. Theperformance of this modified test is shown, via some simulationexperiments, to be comparable to that of the LBI test. An unconditionaltest, based on the assumption that there is a single break at an unknownpoint is also examined. The use of the tests is illustrated with data on theflow of the Nile and US Gross National Product
关键词：Brownian bridge; CramJr-von Mises distribution; intervention;analysis; locally best invariant test; structural time series model;unobserved components
