Biometrika Advance Access originally published online on April 29, 2009
Biometrika 2009 96(2):427-443; doi:10.1093/biomet/asp018
Article |
Double block bootstrap confidence intervals for dependent data
Department of Statistics and Actuarial Science, The University of Hong Kong, Pokfulam Road, Hong Kong smslee{at}hkusua.hku.hk pylaipy{at}graduate.hku.hk
Received for publication 1 May 2007. Revision received 1 October 2008.
The block bootstrap confidence interval for dependent data can outperform the conventional normal approximation only with nontrivial studentization which, in the case of complicated statistics, calls for specialist treatment and often results in unstable endpoints. We propose two double block bootstrap approaches for improving the accuracy of the block bootstrap confidence interval under very general conditions. The first approach calibrates the nominal coverage level and the second calculates studentizing factors directly from a block bootstrap series without the need for nontrivial analytical treatment. We prove that the two approaches reduce the coverage error of the block bootstrap interval by an order of magnitude with simple tuning of block lengths at the two block bootstrapping levels. Empirical properties of the procedures are investigated by simulations and application to an econometric time series.
Key Words: Coverage calibration • Double block bootstrap • p-value • Studentization • Weakly dependent
References
-
Andrews D. W. K. Heteroskedasticity and autocorrelation consistent covariance matrix estimation. Econometrica (1991) 59:817–58.[CrossRef][Web of Science]
Andrews D. W. K., Monahan J. C. An improved heteroskedasticity and autocorrelation consistent covariance matrix estimator. Econometrica (1992) 60:953–66.[CrossRef][Web of Science]
Bertail P., Politis D. N. Extrapolation of subsampling distribution estimators: the i.i.d. and strong mixing cases. Can. J. Statist. (2001) 29:667–80.[CrossRef]
Bhattacharya R. N., Ghosh J. K. On the validity of the formal Edgeworth expansion. Ann. Statist. (1978) 7:434–51.
Bühlmann P. Sieve bootstrap for time series. Bernoulli (1997) 3:123–48.[CrossRef][Web of Science]
Bühlmann P. Bootstraps for time series. Statist. Sci. (2002) 17:52–72.[CrossRef]
Choi E., Hall P. Bootstrap confidence regions computed from autoregressions of arbitrary order. J. R. Statist. Soc. (2000) B 62:461–77.[CrossRef]
Davison A. C., Hall P. On Studentizing and blocking methods for implementing the bootstrap with dependent data. Aust. J. Statist. (1993) 35:215–24.[CrossRef]
Götze F., Hipp C. Asymptotic expansions for sums of weakly dependent random vectors. Z. Wahr. verw. Geb. (1983) 64:211–39.[CrossRef]
Götze F., Künsch H. R. Blockwise bootstrap for dependent observations: higher order approximations for Studentized statistics. Ann. Statist. (1996) 24:1914–33.[CrossRef]
Hall P. Resampling a coverage process. Stoch. Proces. Appl. (1985) 19:259–69.[CrossRef]
Hall P. The Bootstrap and Edgeworth Expansion (1992) New York: Springer.
Hall P., Horowitz J. L., Jing B.-Y. On blocking rules for the bootstrap with dependent data. Biometrika (1995) 82:561–74.
Horowitz J. L., Lobato I. N., Nankervis J. C., Savin N. E. Bootstrapping the Box-Pierce Qtest: a robust test of uncorrelatedness. J. Economet. (2006) 133:841–62.[CrossRef]
Künsch H. R. The jackknife and the bootstrap for general stationary observations. Ann. Statist. (1989) 17:1217–41.[CrossRef]
Lahiri S. N. Edgeworth correction by `moving block' bootstrap for stationary and nonstationary data. In: Exploring the Limits of Bootstrap—LePage R., Billard L., eds. (1992) New York: Wiley. 183–214.
Lahiri S. N. Resampling Methods for Dependent Data (2003) New York: Springer.
Lahiri S. N., Furukawa K., Lee Y.-D. A nonparametric plug-in rule for selecting optimal block lengths for block bootstrap methods. Statist. Methodol. (2007) 4:292–321.
Liu R. Y., Singh K. Moving blocks jackknife and bootstrap capture weak dependence. In: Exploring the Limits of Bootstrap—LePage R., Billard L., eds. (1992) New York: Wiley. 225-48.
Paparoditis E., Politis D. N. Tapered block bootstrap. Biometrika (2001) 88:1105–19.
Paparoditis E., Politis D. N. The local bootstrap for Markov processes. J. Statist. Plan. Infer. (2002) 108:301–28.[CrossRef]
Politis D. N., Romano J. P., Wolf M. Subsampling for heteroskedastic time series. J. Economet. (1997) 81:281–317.[CrossRef]
Politis D. N., White H. Automatic block-length selection for the dependent bootstrap. Economet. Rev. (2004) 23:53–70.[CrossRef]
Rajarshi M. B. Bootstrap in Markov-sequences based on estimates of transition density. Ann. Inst. Statist. Math. (1990) 42:253–68.[CrossRef]
Romano J. P., Wolf M. Improved nonparametric confidence intervals in time series regressions. J. Nonparam. Statist. (2006) 18:199–214.[CrossRef]
Zvingelis J. On bootstrap coverage probability with dependent data. In: Computer-Aided Econometrics—Giles D., ed. (2003) New York: Marcel Dekker. 69–90.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||