Abstract
Portmanteau tests and information criteria are widely used for checking the hypothesis of independence in time series. More recently, data-driven versions were proposed, where the tests are calibrated based on the largest estimated autocorrelation. It seems natural to introduce a double test statistic (M, Q) where Q is the portmanteau and M is the largest squared autocorrelation. Both statistics have been investigated at length in the past decades. We computed under reasonable assumptions the bivariate probability distribution of this double statistic, conditional, in addition, to the lag at which the largest autocorrelation is found. Tests of the null hypothesis of independence based on rejection regions in the plane (M, Q) are proposed, and some methods to select the rejection region in order to maximize power when the alternative hypothesis is unknown are suggested. A simulation study and a thorough comparison with some popular tests have been performed to show the advantages of our proposal. Notice that this latter includes some well-known univariate tests, so we could expect not only an optimal choice but also additional information which may turn useful for a better understanding of the time series for both model building and forecasting.
Similar content being viewed by others
Notes
If more than an integer attains the minimum of (3), the smallest is selected
In practice, the actual top values for M and Q, at which the density vanishes, are often much smaller than N and \(h{\bar{\chi }}_1^2[(1-\alpha )^{1/L}]\), respectively.
Density \(2f(x)F(\beta x)\) where f and F are the standard normal density and cumulative probability functions.
using the method of inversion of the cumulative distribution function on uniform random numbers between 0 and 1.
References
Akashi F, Odashima H, Taniguchi M, Monti AC (2018) A new look at portmanteau tests. Sankhya Ser A 80:121–137
Ali MM (1984) Distributions of the sample autocorrelations when observations are from a stationary autoregressive moving average process. J Bus Econ Stat 2:271–278
Anderson TW, Walker AM (1964) On the asymptotic distributions of the autocorrelations of a sample from a linear stochastic process. Ann Math Stat 35:1296–1303
Baragona R, Battaglia F, Cucina D (2018) Portmanteau tests based on quadratic forms in the autocorrelations. Commun Stat Theory Meth 47:4355–4374
Box GEP, Pierce DA (1970) Distribution of residual autocorrelations in autoregressive-integrated moving average time series models. J Am Stat Assoc 65:1509–1526
Chang Y, Yao Q, Zhou W (2017) Testing for high-dimensional white noise using maximum cross-correlations. Biometrika 104:111–127
Cinkir Z (2014) A fast elementary algorithm for computing the determinant of a Toeplitz matrix. J Comput Appl Math 255:353–361
Coffrey CS, Muller KE (2000) Properties of doubly-truncated gamma variables. Commun Stat Theory Meth 29:851–857
Dufour JM, Roy R (1986) Generalized portmanteau statistics and tests of randomness. Commun Stat Theory Meth 15:2953–2972
Embrechts P, Klüppelberg C, Mikosch T (1997) Modelling extremal events for insurance and finance. Springer, Berlin
Escanciano JC, Lobato IN (2009) An automatic portmanteau test for serial correlation. J Econom 151:140–149
Fisher TJ, Gallagher CM (2012) New weighted portmanteau statistics for time series goodness of fit testing. J Am Stat Assoc 107:777–787
Fisher TJ, Robbins MW (2018) An improved measure for lack of fit in time series models. Stat Sinica 28:1285–1305
Francq C, Roy R, Zakoïan JM (2005) Diagnostic checking in ARMA models with uncorrelated errors. J Am Stat Assoc 100:532–544
Gallagher CM, Fisher TJ (2015) On weighted portmanteau tests for time series goodness-of-fit. J Time Ser Anal 36:67–83
Guay A, Guerre E, Lazarova S (2013) Robust adaptive rate-optimal testing for the white noise hypothesis. J Econom 176:134–145
Guo B, Phillips PCB (2001) Testing for autocorrelation and unit roots in the presence of conditional heteroskedasticity of unknown form. Economics Working Paper 540, University of California at Santa Cruz
Hannan EJ (1970) Multiple time series. Wiley, New York
Harris D, Kew H (2014) Portmanteau autocorrelation tests under Q-dependence and heteroskedasticity. J Time Ser Anal 35:203–217
Hong Y (1996) Consistent testing for serial correlation of unknown form. Econometrica 64:837–846
Hong Y, White H (1995) Consistent specification testing via nonparametric series regression. Econometrica 63:1133–1159
Inglot T, Ledwina T (2006) Towards data driven selection of a penalty function for data driven Neyman test. Linear Algebra Appl 417:124–133
Johnson NL, Kotz S (1970) Continuous univariate distributions in statistics, vol 1. Houghton Mifflin, Boston
Kan R, Wang X (2010) On the distribution of the sample autocorrelation coefficients. J Econom 154:101–121
Katayama N (2008) Portmanteau likelihood tests for model selection. Discussion Paper 1, Faculty of Economics, Kyushu University
Ledwina T (1994) Data-driven versions of Neyman’s smooth test of fit. J Am Stat Assoc 427:1000–1005
Lin JW, McLeod AI (2006) Improved Peña-Rodriguez portmanteau test. Comput Stat Data Anal 51:1731–1738
Ljung GM, Box GEP (1978) On a measure of lack of fit in time series models. Biometrika 65:297–303
Lobato I, Nankervis JC, Savin NE (2001) Testing for autocorrelation using a modified Box-Pierce Q test. Int Econ Rev 42:187–205
Lobato I, Nankervis JC, Savin NE (2002) Testing for zero autocorrelation in the presence of statistical dependence. J Econom 18:730–743
Mahdi E, McLeod AI (2012) Improved multivariate portmanteau test. J Time Ser Anal 33:211–222
Monti AC (1994) A proposal for a residual autocorrelation test in linear models. Biometrika 81:776–780
Newbold P (1980) The equivalence of two tests of time series model adequacy. Biometrika 67:463–465
Peña D, Rodriguez J (2002) A powerful portmanteau test of lack of fit for time series. J Am Stat Assoc 97:601–610
Peña D, Rodriguez J (2006) The log of the determinant of the autocorrelation matrix for testing goodness of fit in time series. J Stat Plan Infer 136:2706–2718
Provost S, Rudiuk E (1995) The sampling distribution of the serial correlation coefficient. Am J Math Manag Sci 15:57–81
Sin CY, White H (1996) Information criteria for selecting possibly misspecified parametric models. J Econom 71:207–225
Taniguchi M, Amano T (2009) Systematic approach for portmanteau tests in view of the Whittle likelihood ratio. J Jpn Stat Soc 39:177–192
Taniguchi M, Kakizawa Y (2000) Asymptotic theory of statistical inference for time series. Springer, New York
Vuong QH (1989) Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica 57:307–333
Wu WB (2009) An asymptotic theory for sample covariances of Bernoulli trials. Stoch Process Appl 119:453–467
Xiao H, Wu WB (2014) Portmanteau test and simultaneous inference for serial covariances. Stat Sinica 24:577–599
Acknowledgements
The authors wish to thank the two anonymous referees for their useful and constructive comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Computational details
Computational details
Throughout this section, we consider the probability distribution under the null hypothesis \(H_0\) of independence, the specification \(|H_0\) is omitted for brevity.
1.1 Test O
The critical points \(m_\alpha , q_\alpha \) must satisfy \(\Pr \{M>m_\alpha \bigcup Q_h>q_\alpha \}= \alpha = 1-\Pr \{M<m_\alpha \bigcap Q_h<q_\alpha \}\) and require computation of the bivariate cumulative distribution function (16). In order to compute \(\Pr \{M<u \bigcap Q_h<v\}\) for any u and v it should be considered that the support is limited to \(u>0, 0<v<hu\) if \(h<p\), and to \(u>0, u<v<hu\) if \(h \ge p\).
We denote for simplicity
The complete form of the conditional distribution may be written as follows
However, for \(h=1\) the exact expression is obtained assuming \(G(v,x,1,\nu ,c)=G(v,x,1,1,1)\) and \(G(v,x,0,\nu ,c)=I_{[v>x]}\) where \(I_{[A]}\) denotes the indicator function of the event A. For \(h=2\) also the exact expression of the conditional distribution may be easily obtained. When \(2 \ge p\) the assumption \(G(v,x,1,\nu ,c)=G(v,x,1,1,1)\) makes formula (21) exact, while when \(2<p\) the correct expression satisfies (20) on assuming
For \(h \ge 3\) the approximation is satisfying. The quality of the proposed approximation has been checked by simulation. For \(3 \le h \le 30\) and \(1 \le M \le 20\), 10, 000 replications of \(S_h=Z_1+Z_2+ \ldots +Z_h\) were simulated, where each \(Z_j\) was generated from the chi-square distribution with one degree of freedom truncated at MFootnote 4. The quantiles at level 0.95 were computed and compared with those of the approximate distribution (13). We found that the differences were uniformly negligible, results for selected values of h and M are shown in Table 2.
On integrating with respect to x and conditioning on p we have:
that may be written
On the other hand, if \(h \ge p\):
that may be written more simply:
Finally, the required probability results:
1.2 Test A
For test A the critical points \(m_\alpha ,q_\alpha \) must satisfy \(\Pr \{M>m_\alpha \bigcap Q_h>q_\alpha \}=\alpha \) and the computation may be based (remembering that \(M=\max \{N{\hat{r}}(j)^2\} \le N\)) on
Thus we have for \(h<p\)
and using (20)
When \(h \ge p\) using (21) we get
Thus the required probability may be written
Rights and permissions
About this article
Cite this article
Baragona, R., Battaglia, F. & Cucina, D. Data-driven portmanteau tests for time series. TEST 31, 675–698 (2022). https://doi.org/10.1007/s11749-021-00794-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11749-021-00794-8