Segmenting Nonstationary Time Series via Quantile Autoregressions

• Activity Number: 423
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Ming Zhong*+
• Companies: University of California, Davis
• Keywords: non-stationarity ; change points ; quantile autoregression ; likelihood-ratio-type test

Abstract

Identifying and locating structural breaks has become an important issue in time series, since parameter estimates and forecasts can be severely biased if non-stationarities are not taken into account. In our work, we try to detect structural breaks within quantile autoregression models. As a member of the class of random-coefficient models, quantile autoregressions allow the autoregressive coefficients to take distinct values over different quantiles of the innovation process, and may thus expand the modeling options for non-stationary and asymmetric time series. We propose a likelihood-ratio-type test for multiple structural breaks and will present its asymptotic properties. Numerical results from simulations and economic data applications, based on the principle of dynamic programming, show that the procedure is consistent for estimating the number and location of the breaks.