Powerful Tests of Structural Change That are Robust to Strong Serial Correlation

This paper proposes powerful and serial correlation robust test statistics that can be used to test for the presence of structural change in the trend function of a univariate time series. Four models are analyzed, each model corresponding to a different way in which a trend break might occur. Given a model, the proposed tests are designed to detect a single break at an unknown date. The tests do not require the knowledge of the form of serial correlation in the data, and they are made robust to the presence of highly persistent serial correlation and a unit root in the errors by using a more comprehensive version of the scaling factor approach of Vogelsang (1998b). The tests utilize the popular nonparametric kernel variance estimators. The fixed-bandwidth asymptotic framework, proposed by Kiefer and Vogelsang (2003), is used to approximate the effects of the variance estimators on the test statistics. The fixed-bandwidth framework makes possible the choice of kernel and bandwidth that deliver tests with maximal asymptotic power within a specific class of tests. For each of the proposed tests, concrete and specific recommendations are made for the bandwidth and kernel to be used in practice. The recommended tests are shown to have good finite sample size and power properties.

[1]  D. Andrews.  Tests for Parameter Instability and Structural Change with Unknown Change Point , 1993 .

[2]  P. Phillips,et al.  Testing the null hypothesis of stationarity against the alternative of a unit root: How sure are we that economic time series have a unit root? , 1992 .

[3]  Katsuto Tanaka.  Testing for a Moving Average Unit Root , 1990, Econometric Theory.

[4]  D. Andrews,et al.  Optimal Tests When a Nuisance Parameter Is Present Only Under the Alternative , 1992 .

[5]  P. Perron,et al.  The Great Crash, The Oil Price Shock And The Unit Root Hypothesis , 1989 .

[6]  Timothy J. Vogelsang,et al.  Testing for a Shift in Mean Without Having to Estimate Serial-Correlation Parameters , 1998 .

[7]  J. Bai,et al.  Least squares estimation of a shift in linear processes , 1994 .

[8]  Timothy J. Vogelsang.  Testing for a Shift in Trend when Serial Correlation is of Unknown Form , 2001 .

[9]  Victor Solo,et al.  Asymptotics for Linear Processes , 1992 .

[10]  Chia-Shang James Chu,et al.  A Direct Test for Changing Trend , 1992 .

[11]  R. Frisch,et al.  Partial Time Regressions as Compared with Individual Trends , 1933 .

[12]  Helle Bunzel,et al.  Powerful Trend Function Tests That Are Robust to Strong Serial Correlation, With an Application to the Prebisch–Singer Hypothesis , 2005 .

[13]  Timothy J. Vogelsang,et al.  Wald-Type Tests for Detecting Breaks in the Trend Function of a Dynamic Time Series , 1997, Econometric Theory.

[14]  P. Phillips.  Time series regression with a unit root , 1987 .

[15]  T. Vogelsang,et al.  Are U.S. regions converging? Using new econometric methods to examine old issues , 2002 .

[16]  Jörg Breitung,et al.  Nonparametric tests for unit roots and cointegration , 2002 .

[17]  Nicholas M. Kiefer,et al.  A NEW ASYMPTOTIC THEORY FOR HETEROSKEDASTICITY-AUTOCORRELATION ROBUST TESTS , 2005, Econometric Theory.

[18]  Timothy J. Vogelsang,et al.  Trend Function Hypothesis Testing in the Presence of Serial Correlation , 1998 .

[19]  G. C. Tiao,et al.  Random Level-Shift Time Series Models, ARIMA Approximations, and Level-Shift Detection , 1990 .

[20]  W. Krämer,et al.  Testing for structural change in dynamic models , 1988 .

[21]  Ozgen Sayginsoy.  Powerful and Serial Correlation Robust Tests of the Economic Convergence Hypothesis , 2004 .