G13DSF (PDF version)
G13 Chapter Contents
G13 Chapter Introduction
NAG Library Manual

NAG Library Routine Document


Note:  before using this routine, please read the Users' Note for your implementation to check the interpretation of bold italicised terms and other implementation-dependent details.

+ Contents

    1  Purpose
    7  Accuracy

1  Purpose

G13DSF is a diagnostic checking routine suitable for use after fitting a vector ARMA model to a multivariate time series using G13DDF. The residual cross-correlation matrices are returned along with an estimate of their asymptotic standard errors and correlations. Also, G13DSF calculates the modified Li–McLeod portmanteau statistic and its significance level for testing model adequacy.

2  Specification


3  Description

Let Wt = w1t,w2t,,wktT , for t=1,2,,n, denote a vector of k time series which is assumed to follow a multivariate ARMA model of the form
Wt-μ= ϕ1Wt-1-μ+ϕ2Wt-2-μ++ϕpWt-p-μ +εt-θ1εt-1-θ2εt-2--θqεt-q, (1)
where εt = ε1t,ε2t,,εktT , for t=1,2,,n, is a vector of k residual series assumed to be Normally distributed with zero mean and positive definite covariance matrix Σ. The components of εt are assumed to be uncorrelated at non-simultaneous lags. The ϕi and θj are k by k matrices of parameters. ϕi, for i=1,2,,p, are called the autoregressive (AR) parameter matrices, and θi, for i=1,2,,q, the moving average (MA) parameter matrices. The parameters in the model are thus the p (k by k) ϕ-matrices, the q (k by k) θ-matrices, the mean vector μ and the residual error covariance matrix Σ. Let
Aϕ= ϕ1 I 0 . . . 0 ϕ2 0 I 0 . . 0 . . . . . . ϕp-1 0 . . . 0 I ϕp 0 . . . 0 0 pk×pk   and  Bθ= θ1 I 0 . . . 0 θ2 0 I 0 . . 0 . . . . . . θq-1 0 . . . I θq 0 . . . . 0 qk×qk
where I denotes the k by k identity matrix.
The ARMA model (1) is said to be stationary if the eigenvalues of Aϕ lie inside the unit circle, and invertible if the eigenvalues of Bθ lie inside the unit circle. The ARMA model is assumed to be both stationary and invertible. Note that some of the elements of the ϕ- and/or θ-matrices may have been fixed at pre-specified values (for example by calling G13DDF).
The estimated residual cross-correlation matrix at lag l is defined to the k by k matrix R^l whose i,jth element is computed as
r ^ i j l = t = l + 1 n ε ^ i t - l - ε - i ε ^ j t - ε - j t = 1 n ε ^ i t - ε - i 2 t = 1 n ε ^ j t - ε - j 2 ,   l =0,1,,i​ and ​j=1,2,,k ,
where ε^it denotes an estimate of the tth residual for the ith series εit and ε-i=t=1nε^it/n. (Note that R^l is an estimate of Eεt-lεtT, where E is the expected value.)
A modified portmanteau statistic, Q m *, is calculated from the formula (see Li and McLeod (1981))
Qm * = k2 mm+1 2n + n l=1 m r^ lT R^ 0 -1 R^ 0 -1 r^ l ,
where  denotes Kronecker product, R^0 is the estimated residual cross-correlation matrix at lag zero and r^l=vec R^lT , where vec of a k by k matrix is a vector with the i,jth element in position i-1k+j. m denotes the number of residual cross-correlation matrices computed. (Advice on the choice of m is given in Section 8.2.) Let lC denote the total number of ‘free’ parameters in the ARMA model excluding the mean, μ, and the residual error covariance matrix Σ. Then, under the hypothesis of model adequacy, Q m *, has an asymptotic χ2-distribution on mk2-lC degrees of freedom.
Let r^̲=vec R1T ,vec R2T ,,vec RmT  then the covariance matrix of r^̲ is given by
where Y=ImΔΔ and G=ImGGT. Δ is the dispersion matrix Σ in correlation form and G a nonsingular k by k matrix such that GGT=Δ-1 and GΔGT=Ik. The construction of the matrix X is discussed in Li and McLeod (1981). (Note that the mean, μ, plays no part in calculating Varr̲^ and therefore is not required as input to G13DSF.)

4  References

Li W K and McLeod A I (1981) Distribution of the residual autocorrelations in multivariate ARMA time series models J. Roy. Statist. Soc. Ser. B 43 231–239

5  Parameters

The output quantities K, N, V, KMAX, IP, IQ, PAR, PARHLD and QQ from G13DDF are suitable for input to G13DSF.
1:     K – INTEGERInput
On entry: k, the number of residual time series.
Constraint: K1.
2:     N – INTEGERInput
On entry: n, the number of observations in each residual series.
3:     V(KMAX,N) – REAL (KIND=nag_wp) arrayInput
On entry: Vit must contain an estimate of the ith component of εt, for i=1,2,,k and t=1,2,,n.
  • no two rows of V may be identical;
  • in each row there must be at least two distinct elements.
4:     KMAX – INTEGERInput
On entry: the first dimension of the arrays V, QQ and R0 and
Constraint: KMAXK.
5:     IP – INTEGERInput
On entry: p, the number of AR parameter matrices.
Constraint: IP0.
6:     IQ – INTEGERInput
On entry: q, the number of MA parameter matrices.
Constraint: IQ0.
Note: IP=IQ=0 is not permitted.
7:     M – INTEGERInput
On entry: the value of m, the number of residual cross-correlation matrices to be computed. See Section 8.2 for advice on the choice of M.
Constraint: IP+IQ<M<N.
8:     PAR(IP+IQ×K×K) – REAL (KIND=nag_wp) arrayInput
On entry: the parameter estimates read in row by row in the order ϕ1,ϕ2,,ϕp, θ1,θ2,,θq.
  • if IP>0, PARl-1×k×k+i-1×k+j must be set equal to an estimate of the i,jth element of ϕl, for l=1,2,,p and i=1,2,,k;
  • if IQ0, PARp×k×k+l-1×k×k+i-1×k+j must be set equal to an estimate of the i,jth element of θl, for l=1,2,,q and i=1,2,,k.
The first p×k×k elements of PAR must satisfy the stationarity condition and the next q×k×k elements of PAR must satisfy the invertibility condition.
9:     PARHLD(IP+IQ×K×K) – LOGICAL arrayInput
On entry: PARHLDi must be set to .TRUE. if PARi has been held constant at a pre-specified value and .FALSE. if PARi is a free parameter, for i=1,2,,p+q×k×k.
10:   QQ(KMAX,K) – REAL (KIND=nag_wp) arrayInput/Output
On entry: QQij is an efficient estimate of the i,jth element of Σ. The lower triangle only is needed.
Constraint: QQ must be positive definite.
On exit: if IFAIL1, then the upper triangle is set equal to the lower triangle.
11:   ISHOW – INTEGERInput
On entry: must be nonzero if the residual cross-correlation matrices r^ijl and their standard errors ser^ijl, the modified portmanteau statistic with its significance and a summary table are to be printed. The summary table indicates which elements of the residual correlation matrices are significant at the 5% level in either a positive or negative direction; i.e., if r^ijl>1.96×ser^ijl then a ‘+’ is printed, if r^ijl<-1.96×ser^ijl then a ‘-’ is printed, otherwise a fullstop (.) is printed. The summary table is only printed if k6 on entry.
The residual cross-correlation matrices, their standard errors and the modified portmanteau statistic with its significance are available also as output variables in R, RCM, CHI, IDF and SIGLEV.
12:   R0(KMAX,K) – REAL (KIND=nag_wp) arrayOutput
On exit: if ij, then R0ij contains an estimate of the i,jth element of the residual cross-correlation matrix at lag zero, R^0. When i=j, R0ij contains the standard deviation of the ith residual series. If IFAIL=3 on exit then the first K rows and columns of R0 are set to zero.
13:   R(KMAX,KMAX,M) – REAL (KIND=nag_wp) arrayOutput
On exit: Rlij is an estimate of the i,jth element of the residual cross-correlation matrix at lag l, for i=1,2,,k, j=1,2,,k and l=1,2,,m. If IFAIL=3 on exit then all elements of R are set to zero.
14:   RCM(LDRCM,M×K×K) – REAL (KIND=nag_wp) arrayOutput
On exit: the estimated standard errors and correlations of the elements in the array R. The correlation between Rlij and Rl2i2j2 is returned as RCMst where s=l-1×k×k+j-1×k+i and t=l2-1×k×k+j2-1×k+i2 except that if s=t, then RCMst contains the standard error of Rlij. If on exit, IFAIL5, then all off-diagonal elements of RCM are set to zero and all diagonal elements are set to 1/n.
15:   LDRCM – INTEGERInput
On entry: the first dimension of the array RCM as declared in the (sub)program from which G13DSF is called.
Constraint: LDRCMM×K×K.
16:   CHI – REAL (KIND=nag_wp)Output
On exit: the value of the modified portmanteau statistic, Q m *. If IFAIL=3 on exit then CHI is returned as zero.
17:   IDF – INTEGEROutput
On exit: the number of degrees of freedom of CHI.
18:   SIGLEV – REAL (KIND=nag_wp)Output
On exit: the significance level of CHI based on IDF degrees of freedom. If IFAIL=3 on exit, SIGLEV is returned as one.
19:   IW(LIW) – INTEGER arrayWorkspace
20:   LIW – INTEGERInput
On entry: the dimension of the array IW as declared in the (sub)program from which G13DSF is called.
Constraint: LIWK×maxIP,IQ.
21:   WORK(LWORK) – REAL (KIND=nag_wp) arrayWorkspace
22:   LWORK – INTEGERInput
On entry: the dimension of the array WORK as declared in the (sub)program from which G13DSF is called.
Constraint: if LWORKkn+KMAX+2+mk2NPAR+mk2+1+ 3k2+NPAR+1NPAR, NPAR=p+qk2.
23:   IFAIL – INTEGERInput/Output
On entry: IFAIL must be set to 0, -1​ or ​1. If you are unfamiliar with this parameter you should refer to Section 3.3 in the Essential Introduction for details.
For environments where it might be inappropriate to halt program execution when an error is detected, the value -1​ or ​1 is recommended. If the output of error messages is undesirable, then the value 1 is recommended. Otherwise, because for this routine the values of the output parameters may be useful even if IFAIL0 on exit, the recommended value is -1. When the value -1​ or ​1 is used it is essential to test the value of IFAIL on exit.
On exit: IFAIL=0 unless the routine detects an error or a warning has been flagged (see Section 6).

6  Error Indicators and Warnings

If on entry IFAIL=0 or -1, explanatory error messages are output on the current error message unit (as defined by X04AAF).
Note: G13DSF may return useful information for one or more of the following detected errors or warnings.
Errors or warnings detected by the routine:
On entry,K<1,
orLIW is too small,
orLWORK is too small.
On entry, either QQ is not positive definite or the autoregressive parameter matrices are extremely close to or outside the stationarity region, or the moving average parameter matrices are extremely close to or outside the invertibility region. To proceed, you must supply different parameter estimates in the arrays PAR and QQ.
On entry, at least one of the k residual series is such that all its elements are practically identical giving zero (or near zero) variance or at least two of the residual series are identical. In this case CHI is set to zero, SIGLEV to one and all the elements of R0 and R set to zero.
This is an unlikely exit brought about by an excessive number of iterations being needed to evaluate the zeros of the determinantal polynomials detAϕ and detBθ. All output parameters are undefined.
On entry, either the eigenvalues and eigenvectors of Δ (the matrix QQ in correlation form) could not be computed or the determinantal polynomials detAϕ and detBθ have a factor in common. To proceed, you must either supply different parameter estimates in the array QQ or delete this common factor from the model. In this case, the off-diagonal elements of RCM are returned as zero and the diagonal elements set to 1/n. All other output quantities will be correct.
This is an unlikely exit. At least one of the diagonal elements of RCM was found to be either negative or zero. In this case all off-diagonal elements of RCM are returned as zero and all diagonal elements of RCM set to 1/n.

7  Accuracy

The computations are believed to be stable.

8  Further Comments

8.1  Timing

The time taken by G13DSF depends upon the number of residual cross-correlation matrices to be computed, m, and the number of time series, k.

8.2  Choice of m

The number of residual cross-correlation matrices to be computed, m, should be chosen to ensure that when the ARMA model (1) is written as either an infinite order autoregressive process, i.e.,
or as an infinite order moving average process, i.e.,
Wt-μ=j= 1ψjεt-j+εt
then the two sequences of k by k matrices π1,π2, and ψ1,ψ2, are such that πj and ψj are approximately zero for j>m. An overestimate of m is therefore preferable to an under-estimate of m. In many instances the choice m=10 will suffice. In practice, to be on the safe side, you should try setting m=20.

8.3  Checking a ‘White Noise’ Model

If you have fitted the ‘white noise’ model
then G13DSF should be entered with p=1, q=0, and the first k2 elements of PAR and PARHLD set to zero and .TRUE. respectively.

8.4  Approximate Standard Errors

When IFAIL=5 or 6 all the standard errors in RCM are set to 1/n. This is the asymptotic standard error of r^ijl when all the autoregressive and moving average parameters are assumed to be known rather than estimated.

8.5  Alternative Tests

R^0 is useful in testing for instantaneous causality. If you wish to carry out a likelihood ratio test then the covariance matrix at lag zero C^0 can be used. It can be recovered from R^0 by setting
C^0i,j =R^0i,j×R^0i,i×R^0j,j, for ​ij =R^0i,j×R^0i,j, for ​i=j

9  Example

This example fits a bivariate AR(1) model to two series each of length 48. μ has been estimated but ϕ12,1 has been constrained to be zero. Ten residual cross-correlation matrices are to be computed.

9.1  Program Text

Program Text (g13dsfe.f90)

9.2  Program Data

Program Data (g13dsfe.d)

9.3  Program Results

Program Results (g13dsfe.r)

G13DSF (PDF version)
G13 Chapter Contents
G13 Chapter Introduction
NAG Library Manual

© The Numerical Algorithms Group Ltd, Oxford, UK. 2012