The Annals of Statistics

Least absolute deviation estimation for all-pass time series models

F. Jay Breidt, Richard A. Davis, and A. Alexandre Trindade

Full-text: Open access


An autoregressive moving average model in which all of the roots of the autoregressive polynomial are reciprocals of roots of the moving average polynomial and vice versa is called an all-pass time series model. All-pass models generate uncorrelated (white noise) time series, but these series are not independent in the non-Gaussian case. An approximation to the likelihood of the model in the case of Laplacian (two-sided exponential) noise yields a modified absolute deviations criterion, which can be used even if the underlying noise is not Laplacian. Asymptotic normality for least absolute deviation estimators of the model parameters is established under general conditions. Behavior of the estimators in finite samples is studied via simulation. The methodology is applied to exchange rate returns to show that linear all-pass models can mimic “nonlinear” behavior, and is applied to stock market volume data to illustrate a two-step procedure for fitting noncausal autoregressions.

Article information

Ann. Statist., Volume 29, Number 4 (2001), 919-946.

First available in Project Euclid: 14 February 2002

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Primary: 62M10: Time series, auto-correlation, regression, etc. [See also 91B84]
Secondary: 62E20: Asymptotic distribution theory 62F10: Point estimation

Laplacian density noncausal noninvertible nonminimum phase white noise


Breidt, F. Jay; Davis, Richard A.; Trindade, A. Alexandre. Least absolute deviation estimation for all-pass time series models. Ann. Statist. 29 (2001), no. 4, 919--946. doi:10.1214/aos/1013699987.

Export citation


  • Akaike, H. (1973). Information theoryand an extension of the maximum likelihood principle. In Second International Symposium on Information Theory (B. N. Petrov and F. Csaki, eds.) 267-281. Akademiai Kiado, Budapest.
  • Benveniste, A., Goursat, M. and Roget, G. (1980). Robust identification of a nonminimum phase system: blind adjustment of linear equalizer in data communications. IEEE Trans. Auto. Control AC-25 385-398.
  • Bickel, P. J. and B ¨uhlmann, P. (1996). What is a linear process? Proc Nat. Acad. Sci. 93 12128-12131.
  • Blass, W. E. and Halsey, G. W. (1981). Deconvolution of Absorption Spectra. Academic Press, New York.
  • Bollerslev, T., Chou, R. Y., and Kroner, K. F. (1992). ARCH modeling in finance. J. Econometrics 52 5-59.
  • Breidt, F. J. and Davis, R. A. (1991). Time-reversibility, identifiability and independence of innovations for stationarytime series. J. Time Ser. Anal. 13 377-390.
  • Breidt, F. J., Davis, R. A., Lii, K.-S. and Rosenblatt, M. (1991). Maximum likelihood estimation for noncausal autoregressive processes. J. Multivariate Anal. 36 175-198.
  • Brockwell, P. J. and Davis, R. A. (1991). Time Series: Theory and Methods, 2nd ed. Springer, New York.
  • Chi, C.-Y. and Kung, J.-Y. (1995). A new identification algorithm for allpass systems by higherorder statistics. Signal Processing 41 239-256.
  • Chien, H.-M., Yang, H.-L. and Chi, C.-Y. (1997). Parametric cumulant base phase estimation of 1-d and 2-d nonminimum phase systems by allpass filtering. IEEE Trans. Signal Processing 45 1742-1762.
  • Clark, P. K. (1973). A subordinated stochastic process model with finite variances for speculative prices. Econometrica 41 135-156.
  • Davis, R. A. and Dunsmuir, W. T. M. (1997). Least absolute deviation estimation for regression with ARMA errors. J. Theoret. Probab. 10 481-497.
  • Donoho, D. (1981). On minimum entropydeconvolution. In Applied Time Series Analysis II (D. F. Findley, ed.) 565-608. Academic Press, New York.
  • Engle, R. F. (1982). Autoregressive conditional heteroscedasticitywith estimates of the variance of United Kingdom inflation. Econometrica 50 987-1007.
  • Giannakis, G. B. and Swami, A. (1990). On estimating noncausal nonminimum phase ARMA models of non-Gaussian processes. IEEE Trans. Acoustics, Speech, and Signal Processing 38 478-495.
  • Godfrey, R. and Rocca, F. (1981). Zero memorynonlinear deconvolution. Geophysical Prospecting 29 189-228.
  • Hooke, R. and Jeeves, T. (1961). A direct search solution of numerical and statistical problems. J. Assoc. Comput. Mach. 8 212-229.
  • Hsueh, A.-C. and Mendel, J. M. (1985). Minimum-variance and maximum-likelihood deconvolution for noncausal channel models. IEEE Trans. Geoscience and Remote Sensing 23 797-808.
  • Jacquier, E., Polson, N. G. and Rossi, P. E. (1994). Bayesian analysis of stochastic volatility models (with discussion). J. Bus. Econom. Statist. 12 371-417.
  • Huang, J. and Pawitan, Y. (2000). Quasi-likelihood estimation of noninvertible moving average processes. Scandinavian J. Statist. 27 689-702.
  • Kreiss, J.-P. (1987). On adaptive estimation in stationaryARMA processes. Ann. Statist. 15 112-133.
  • Lii, K.-S. and Rosenblatt, M. (1982). Deconvolution and estimation of transfer function phase and coefficients for non-Gaussian linear processes. Ann. Statist. 10 1195-1208.
  • Lii, K.-S. and Rosenblatt, M. (1992). An approximate maximum likelihood estimation for nonGaussian non-minimum phase moving average processes. J. Multivariate Anal. 43 272-299.
  • Lii, K.-S. and Rosenblatt, M. (1996). Maximum likelihood estimation for non-Gaussian nonminimum phase ARMA sequences. Statist. Sinica 6 1-22.
  • Mandelbrot, B. (1963). The variation of certain speculative prices. J. Business 36, 394-419.
  • McLeod, A. I. and Li, W. K. (1983). Diagnostic checking ARMA time series models using squaredresidual autocorrelations. J. Time Ser. Anal. 4 269-273.
  • Ooe, M. and Ulrych, T. J. (1979). Minimum entropydeconvolution with an exponential transformation. Geophysical Prospecting 27 458-473.
  • Rabiner, L. R. and Schafer, R. M. (1978). Digital Processing of Speech Signals. Prentice-Hall, Englewood Cliffs, NJ.
  • Robinson, P. M. (1987). Time series residuals with application to probabilitydensityestimation. Time Ser. Anal. 8 329-344.
  • Rockafellar, R. T. (1970). Convex Analysis. Princeton Univ. Press.
  • Rosenblatt, M. (2000). Gaussian and Non-Gaussian Linear Time Series and Random Fields. Springer, New York.
  • Scargle, J. D. (1981). Phase-sensitive deconvolution to model random processes, with special reference to astronomical data. In Applied Time Series Analysis II (D. F. Findley, ed.) 549-564. Academic Press, New York.
  • Wiggins, R. A. (1978). Minimum entropydeconvolution. Geoexploration 16 21-35.