An indirect proof for the asymptotic properties of VARMA model estimators☆
Introduction
Vector autoregressive-moving average (VARMA) models are the multivariate generalization of univariate autoregressive-moving average (ARMA) models. ARMA models are used to represent stationary time series in almost all domains where a variable is measured at equidistant times. Its generalizations, ARIMA and SARIMA models, respectively for integrated and seasonal time series, can cope with a stochastic trend and stochastic seasonality. Vector autoregressive (VAR) models are widely used in econometrics. VARMA models are much less used, except in certain specific applications. For example, they are able to represent competitive behaviour between products in economics, see Jenkins (1979). VARMA models have been widely studied since the 1970’s, starting with Hannan (1969) and Tiao and Box (1981). See the books by Hannan and Deistler (2012), first published in 1988, Reinsel (1997), Lütkepohl (2005), Wei (2006), and Tsay (2014). New theoretical results are still being developed for model specification (Poskitt, 2005, Athanasopoulos, Vahid, 2008a, Boubacar Maïnassara, 2012, Brockwell, Lindner, Vollenbröker, 2012, Wegge, 2012), in particular in structural models (Gouriéroux, Monfort, Renne, 2017, Gouriéroux, Monfort, Renne, 2020), for estimation methods (Kascha, 2012, Boubacar Maïnassara, Francq, 2011, Dufour, Jouini, 2014, Hallin, La Vecchia, Liu, 2020, Roy, McElroy, Linton, 2019), also in the case of non-independent but uncorrelated errors (Boubacar Maïnassara, Saussereau, 2018, Dufour, Pelletier, 2019), for diagnostic checking (Boubacar Maïnassara, 2011), and for forecasting (Lütkepohl, 2006, Athanasopoulos, Vahid, 2008b, Kascha, Mertens, 2009). Work is also being carried out on sparse models (Wilms et al., 2019).
More recently, ARMA and VARMA models with time-dependent coefficients have been proposed and asymptotic theories have been developed in order to justify their use. In most cases, a parametric approach has been taken, but there are also semi-parametric approaches. In parametric approaches, it is assumed that there is a relatively small number of parameters. In Alj et al. (2017), the authors have considered time-dependent VARMA (tdVARMA) models with possibly marginal heteroscedasticity (as opposed to conditional heteroscedasticity). Subject to a set of assumptions, they establish strong consistency and asymptotic normality of a Gaussian quasi-maximum likelihood estimator (QMLE) for the parameters of a tdVARMA model.
We present a new proof of the asymptotic properties of standard, i.e. homoscedastic, causal, invertible, and identifiable VARMA models. That proof is indirect and is based on Alj et al. (2017) and a new representation of tdVARMA models under a higher-dimensional tdVAR(1) model. This is done, on the basis of Lütkepohl (1991), by combining features from the two papers of Francq and Gautier (2004) and Francq and Zakoïan (2001); see also Boubacar Maïnassara and Rabehasaina (2020). This makes it possible to check the assumptions made by Alj et al. (2017). It is, therefore, an indirect proof. The paper is of course also applicable to ARMA models (see the Supplementary Material).
Since the proof of the asymptotic properties for tdVARMA models does not exploit ergodicity, a property that is certainly invalid for tdVARMA models, which are not usually stationary, our indirect proof for VARMA model does not use ergodicity. Although there exist elements of spectral theory for non-stationary time series, the Alj et al. (2017) approach does not make use of spectral theory. Consequently, the indirect new proof of the asymptotic properties for VARMA models is provided entirely in the time domain and does not refer to a frequency domain approach. Existing proofs of asymptotic properties in time series models, e.g. Hannan (1973), reported and expanded in Brockwell and Davis (1991) and Yao and Brockwell (2006), for scalar processes, or Dunsmuir and Hannan (1976), Deistler et al. (1978), Rissanen and Caines (1979), Kohn (1979), Hannan et al. (1980), and Ling and McAleer (2003) (the latter with GARCH errors), for vector processes, all make use of ergodicity. Several of these proofs also make use of spectral arguments, starting with Hannan (1973), but not those of Yao and Brockwell (2006) or Rissanen and Caines (1979). Several other comparative aspects will be discussed later, including the ability to handle structural models, the assumptions made about errors and their moments, and the compactness of the parameter space. It should be clear that our results do not require weaker assumptions than previous results. For example, we need the existence of fourth-order moments for the errors, while only the existence of the variance is needed in Yao and Brockwell (2006) for scalar models. The requirement is fourth-order moments for Ling and McAleer (2003) in the wider context of VARMA-ARCH models.
The proof for VARMA models is relatively short; it should be noted, however, that the Alj et al. (2017) paper relies on a law of large numbers and a central limit theorem for martingales and a law of large number for mixingales, and requires a long technical appendix to establish the required lemmas. The latter paper’s technical appendix is, however, not needed for scalar models, as the simpler theory for tdARMA models in Azrak and Mélard (2006) is then applicable. Moreover, in order to reduce the moment assumption on the model errors, the two papers need to be complemented by a result of Azrak and Mélard (2020).
In Section 2, we introduce VARMA and tdVARMA models, and quasi-maximum likelihood estimators (QMLE), and we state the assumptions for homoscedastic tdVARMA models. Section 3 is devoted to preliminaries, where tdVARMA models are put into a tdVAR(1) form that allows for easier treatment of the derivatives of the pure tdVMA representation. Section 4 contains the assumptions we need for standard VARMA models, a few remarks on those assumptions, a theorem of consistency and asymptotic normality of the QMLE estimators of a VARMA model, and its proof, which consists in the verification of the assumptions of Alj et al. (2017). As promised, the proof does not exploit ergodicity (even if the process is ergodic, given the assumptions being made) and does not require spectral arguments.
Section snippets
QMLE for a tdVARMA or a VARMA time series model
We introduce an estimator of the parameters of a zero-mean -dimensional vector mixed autoregressive-moving average (VARMA) model with time-dependent coefficients, denoted by tdVARMA. Then we particularize it to the standard VARMA model.
Putting tdVARMA models in tdVAR(1) form
A representation of an ARMA model in tdVAR(1) form is not new (see Lütkepohl, 1991, and Lütkepohl, 2005). It was used by Francq and Gautier (2004) for tdARMA models and was detailed in a working paper by Francq and Gautier (2003). It was described there using a state-space representation. We will see that the state-space representation can be eliminated from the description and that VARMA models can be handled in the same way as ARMA models. Note that Francq and Zakoïan (2001) propose a similar
Main result
Starting in this section, we consider the standard VARMA model, defined by (2.3), as a special case of (2.2). We now suppose constant matrices with respect to time, hence and for all and where, to simplify the presentation, is composed of the elements of the coefficients of the model, more preciselyIn principle, the vector can be a subset of those elements, but that would complicate the notations. The elements of
Declaration of Competing Interest
None.
Acknowlgedgments
This paper begun when Christophe Ley and I were trying to complete tdVAR(1) theoretical examples in Alj et al. (2017). This lead me to explore an idea in Graeme McRae’s blog (http://2000clicks.com/MathHelp/SeriesPolynomialGeometric.aspx). I found the crucial inspiration for Section 4 in Boubacar Maïnassara and Rabehasaina (2020). I would like to thank Yacouba Boubacar Maïnassara, Christian Francq, and Christophe Ley for their help at the early stages of this research, André Klein and Peter
References (63)
- et al.
On the identifiability of minimal VARMA representations
Statistical Inference for Stochastic Processes
(1998) - et al.
Estimating structural VARMA models with uncorrelated but non-independent error terms
Journal of Multivariate Analysis
(2011) - et al.
Estimation of weak ARMA models with regime changes
Statistical Inference for Stochastic Processes
(2020) - et al.
Practical methods for modelling weak VARMA processes: identification, estimation and specification with a macroeconomic application
Working paper
(2019) - et al.
Resultants of matrix polynomials
Bulletin of the American Mathematical Society
(1976) - et al.
Statistical inference for independent component analysis: Application to structural VAR models
Journal of Econometrics
(2017) - et al.
Center-outward R-estimation for semiparametric VARMA models
Journal of the American Statistical Association
(2020) Practical Experiences with Modelling and Forecasting Time Series
(1979)- et al.
On conditional least squares estimation for stochastic processes
Annals of Statistics
(1978) Asymptotic properties of time domain Gaussian estimators
Advances in Applied Probabability
(1978)
The information matrices of the parameters of multiple mixed time series
Journal of Multivariate Analysis
A Handbook of Time-Series Analysis, Signal Processing and Dynamics
A note on the specification and estimation of ARMAX systems
Journal of Time Series Analysis
Asymptotic properties of QML estimators for VARMA models with time-dependent coefficients
Scandinavian Journal of Statistics
A complete VARMA modelling methodology based on scalar components
Journal of Time Series Analysis
VARMA versus VAR for macroeconomic forecasting
Journal of Business and Economic Statistics
Asymptotic properties of quasi-likelihood estimators for ARMA models with time-dependent coefficients
Statistical Inference for Stochastic Processes
Asymptotic properties of conditional least-squares for array time series
ECARES working paper 2020-12
An algorithm for coprime matrix fraction description using Sylvester matrices
Linear Algebra and its Applications
Selection of weak VARMA models by modified Akaike’s information criteria
Journal of Time Series Analysis
Multivariate portmanteau test for structural VARMA models with uncorrelated but non-independent error terms
Journal of Statistical Planning and Inference
Diagnostic checking in multivariate ARMA models with dependent errors using normalized residual autocorrelations
Journal of the American Statistical Association
Time Series: Theory and Methods
Strictly stationary solutions of multivariate ARMA equations with i.i.d. noise
Annals of the Institute of Statistical Mathematics
Vector linear time series models: corrections and extensions
Advances in Applied Probability
Asymptotic distributions for quasi-efficient estimators in echelon VARMA models
Computational Statistics and Data Analysis
Vector linear time series models
Advances in Applied Probability
Estimation of time-varying ARMA models and applications to series subject to Markovian changes in regime
Working paper
Estimation of time-varying ARMA models with Markovian changes in regime
Statistics and Probability Letters
Stationarity of multivariate Markov-switching ARMA models
Journal of Econometrics
Cited by (5)
Center-outward Rank- and Sign-based VARMA Portmanteau Tests: Chitturi, Hosking, and Li–McLeod revisited
2023, Econometrics and StatisticsAttenuation Law of Tensile Property of PVA-ECC after Freezing-Thawing Cycles
2023, Bulletin of the Chinese Ceramic Society
- ☆
Supplementary Material entitled ”From tdARMA models to ARMA models” is available for ARMA models.
- 1
Université libre de Bruxelles, Faculty SBS-EM & ECARES.