Score-driven non-linear multivariate dynamic location models€¦ · 09/11/2017 · SZABOLCS...
Transcript of Score-driven non-linear multivariate dynamic location models€¦ · 09/11/2017 · SZABOLCS...
Score-driven non-linear multivariate dynamic location modelsSZABOLCS BLAZSEK (FRANCISCO MARROQUIN UNIVERSITY)
ALVARO ESCRIBANO (CARLOS III UNIVERSITY OF MADRID)
ADRIAN LICHT (FRANCISCO MARROQUIN UNIVERSITY)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 1
Contribution
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 2
Contribution
In this paper, we extend the dynamic conditional score (DCS) model of the multivariate �-distribution that was introduced in the work of Harvey (2013, Chapter 7.2.2).
Motivated by Harvey (2013, Chapter 3.2.1), we name the new model quasi-vector autoregressive (QVAR) model.
QVAR with lag-order �, denoted as QVAR(�), is a score-driven non-linear multivariate dynamic location model, in which the conditional score vector of the log-likelihood (LL) updates the dependent variables.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 3
Contribution
QVAR(�) is an extension of the DCS model for the multivariate �-distribution that is QVAR(1) under our notation.
For QVAR, we present the details of the econometric formulation, the computation of the impulse response function, and the maximum likelihood (ML) estimation and related conditions of consistency and asymptotic normality.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 4
Contribution
We compare the statistical performance of QVAR and that of two benchmark multivariate dynamic location models:
VAR and VARMA (vector autoregressive moving average).
We estimate QVAR by using the ML method.
We estimate VAR and VARMA by using the quasi-ML (QML) method.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 5
Contribution
The likelihood-based model performance metrics suggest that the statistical performance of QVAR is superior to that of VAR and VARMA.
The residual and conditional score diagnostic test results suggest that each residual and conditional score variable of QVAR(2) forms a multivariate i.i.d. time series.
The conditions of consistency and asymptotic normality of ML are satisfied for QVAR.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 6
QVAR model
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 7
Quasi-vector autoregressive modelReduced-form representation of QVAR( )
�� = � + �� + �
�� = Φ����� + ⋯ + Φ����� + Ψ�����
where ��, �, ��, � and �� are (� × 1) and
Φ�, … , Φ� and Ψ� are (� × �)
For the first � observations, we initialize �� by using its unconditional mean �� = � �� = 0��.
For QVAR(�), the conditional expectation of �� is � �� ��, … , ���� = � + ��.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 8
Quasi-vector autoregressive modelReduced-form representation of QVAR( )
�~��(0, Σ, ν) is the multivariate i.i.d. reduced-form error term that updates ��. Σ is positive definite and ν > 2.
The log conditional density of �� is
ln " �� ��, … , ���� = ln Γ$%�
&− ln Γ
$
&−
�
&
−�
&ln Σ −
$%�
&ln 1 +
)*+,-.)*
$
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 9
Quasi-vector autoregressive modelReduced-form representation of QVAR( )
The partial derivative of the log of the conditional density with respect to �� is
/ 01 2 �� ��, … , ����
/3*=
$%�
$Σ�� × 1 +
)*+,-.)*
$
��
� =
=$%�
$Σ�� × ��
where the last equality defines the scaled score function ��, which is the representation of �� by using the reduced-form error term.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 10
Quasi-vector autoregressive modelReduced-form representation of QVAR( )
Harvey (2013, Chapter 7) shows that �� is multivariate i.i.d. with mean zero and covariance matrix
Var �� =$%�
$%�%&��
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 11
Quasi-vector autoregressive modelStructural-form representation of QVAR(�)
For the reduced-form error term we have � � = 0 and Var � = Σν/(ν − 2)
We factorize Var � as
Var � =$
$�&
�/&�� �� 9 $
$�&
�/&
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 12
Quasi-vector autoregressive modelStructural-form representation of QVAR(�)
We introduce the multivariate i.i.d. structural-form error term :�:
� =$
$�&
�/&��:�
where � :� = 0�×�, Var :� = ;�, and :�~�� 0, ;� ×$�&
$, ν .
Therefore, the representation of the score function �� by using the structural-form error term is
�� = (ν − 2)ν �/&Ω�� ×<*
$�&%<*+<*
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 13
Quasi-vector autoregressive modelSummary
Reduced-form representation:
�� = � + Φ����� + ⋯ + Φ����� + Ψ����� + �
where � � = 0�×�and Var � = [ν/(ν − 2)]Ω�� Ω�� 9.
Structural-form representation:
?@�� = ?@� + ?@Φ����� + ⋯ + ?@Φ����� + ?@Ψ����� + :�
where ?@ = [ν/(ν − 2)]��/&Ω��:�, and
� :� = 0�� and Var :� = ;�.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 14
Non-linear vector MA representations and impulse response functions
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 15
Non-linear vector MA representations and impulse response functions
The first-order representation of the reduced-form QVAR(A) is
B� = C + D� + E�
D� = ΦD��� + ΨF���
where
B� =
��
����
⋮����%� (��×�)
C =
��⋮� (��×�)
D� =
��
����
⋮����%� (��×�)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 16
Non-linear vector MA representations and impulse response functions
E� =
�
���
⋮���%� (��×�)
Φ =
Φ� Φ& … Φ��� Φ�
;� 0�×� … … 0�×�
0�×� ;� 0�×� … …… … … … …
0�×� … 0�×� ;� 0�×� (��×��)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 17
Non-linear vector MA representations and impulse response functions
Ψ =
Ψ� 0�×� … 0�×�
0�×� 0�×� … 0�×�
… … … …0�×� … … 0�×� (��×��)
F��� =
����
0⋮0 (��×�)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 18
Non-linear vector MA representations and impulse response functions
The corresponding structural-form non-linear vector MA(∞) representation of �� is:
�� = � + ∑ JΦKJ′Ψ� ν − 2 ν �/&Ω�� <*-.-M
$�&%<*-.-M+ <*-.-M
NKO@ +
+$
$�&
�/&��:�
where J = (;� , 0�×� , … , 0�×�) (� × ��).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 19
Non-linear vector MA representations and impulse response functions
The impulse response function IRFK = S��%K/S:� for T = 0,1, … , ∞is given by:
IRF@ =$
$�&
�/&��
IRFK� = JΦKJ′Ψ� ν − 2 ν �/&Ω��U����K for T = 1, … , ∞,
where
U� =/
V*
W-XYV*+V*
/<*(this derivative is evaluated in closed form)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 20
Non-linear vector MA representations and impulse response functions
As IRFK� for T = 1, … , ∞ depends on �, we evaluate its
unconditional mean and use:
IRFK = JΦKJ′Ψ� ν − 2 ν.
X���(U����K)
If all elements of U����K form covariance stationary time series,
then �(U����K) can be estimated by using the sample average.
We test the covariance stationarity of U����K by using the
augmented Dickey-Fuller (1979) (ADF) unit root test.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 21
Maximum likelihood (ML) estimation
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 22
Maximum likelihood (ML) estimation
The parameters of QVAR are �, Φ�, … , Φ�, Ψ�, Ω�� and ν.
We estimate these parameters by using the ML method:
Θ[\] = arg maxa
LL ��, … , �c =
= arg maxa
∑ ln "(��|��, … , ����)c�O�
We use the inverse information matrix to estimate the standard errors of parameters (Harvey 2013). We use Harvey (2013,
Chapters 2.3, 2.4 and 3.3) to find the conditions under which the ML estimates of QVAR(�) are consistent and asymptotically Gaussian.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 23
Conditions of consistency and asymptotic normality of ML
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 24
Condition 1
If all eigenvalues of Φ are within the unit circle, then �� is covariance stationary.
Let C� denote the maximum modulus of all eigenvalues.
C� < 1 supports Condition 1.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 25
Condition 2
We use Condition 2 from the work of Harvey (2013, p. 35, Condition 2).
Condition 2 is that the score function �� (� × 1) and its first derivative S��/S�� (� × �) have finite second moments and covariance that are time-invariant and do not depend on ��.
We test Condition 2 by using the ADF test (C& metric).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 26
Condition 3
In order to obtain Conditions 3 and 4 for QVAR, we use the arguments of the proof of Theorem 5 from the work of Harvey (2013, p. 49).
Recall the first-order representation of QVAR(�):
D� = ΦD��� + ΨF���
We consider the representative element ΨfK from the
matrix Ψ.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 27
Condition 3The derivative of the previous equation with respect to ΨfK is:
/g*
/hiM= Φ
/g*-.
/hiM+ Ψ
/j*-.
/hiM+ kfKF���
Where the element (l, T) of kfK (�� × ��) is one and the rest of the
elements are zero. From here we express:
/g*
/hiM= Φ + Ψ
/j*-.
/g*-.+
/g*-.
/hiM+ kfKF��� = m�
/g*-.
/hiM+ kfKF���
Condition 3 is that all eigenvalues of E m� are inside the unit circle (Cometric). We estimate this expectation by using sample average (ADF test).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 28
Condition 4
From
/g*
/hiM= m�
/g*-.
/hiM+ kfKF���
we express:
vec/g*
/hiM
/g*+
/hst= m� ⊗ m� vec
/g*-.
/hiM
/g*-.+
/hst+
vec m�/g*-.
/hiMkfK
9 F��� + vec F���9 kvw
/g*-.+
/hstm�
9 +
vec kfKF���F���9 kvw
9
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 29
Condition 4
Condition 4 is that all eigenvalues of � m� ⊗ m� are inside the unit circle (Cx metric: maximum modulus of all eigenvalues).
We estimate this expectation by using sample average.
We use the ADF test for each element of m� ⊗ m�.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 30
Benchmark models
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 31
Benchmark models
We compare QVAR with the benchmark VAR(�) and VARMA(�, 1) models.
We estimate both benchmark models by using the QML estimator (Gourieroux, Monfort and Trognon 1984a-b).
We use the Gaussian distribution as auxiliary distribution for QML.
Covariance stationarity and invertibility of VARMA ensures consistency and asymptotic normality of QML.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 32
Data
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 33
DataWe use macroeconomic data from the book of Kilian and Lütkepohl (2017). This dataset includes the following variables:
(i) monthly West Texas Intermediate (WTI) price of crude oil for period December 1972 to June 2013;
(ii) quarterly US GDP deflator for period 1959Q1 to 2013Q2;
(iii) quarterly US real GDP level for period 1959Q1 to 2013Q2.
The use of these variables is motivated by several works from the body of literature, which study the question of how oil price shocks affect US real GDP and inflation (e.g. Blanchard 2002; Barsky and Kilian 2004; Kilian 2008; Kilian and Lütkepohl 2017).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 34
DataWe define:
(i) variable ��� as the quarterly first difference of log real price of crude oil (hereinafter, crude oil);
(ii) variable �&� as the quarterly first difference of log US GDP deflator (hereinafter, inflation);
(iii) variable �o� as the quarterly first difference of log US real GDP level (hereinafter, GDP growth).
We define �� = (��� , �&� , �o�)9, hence, � = 3 for all models in this paper. Furthermore, we use data for period 1987Q1 to 2013Q2 (Kilian and Lütkepohl 2017, Chapter 9.2.1).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 35
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 36
INTERNATIONAL CONFERENCE IN HONOR OF LUC BAUWENS, 19-20 OCTOBER 2017 37
Seasonality effects in variables
For each variable we estimate a linear regression:
��� = z�,{�U{�,� + z�,{&U{&,� + z�,{oU{o,� + z�,{xU{x,� + �
�&� = z&,{�U{�,� + z&,{&U{&,� + z&,{oU{o,� + z&,{xU{x,� + η&�
�o� = zo,{�U{�,� + zo,{&U{&,� + zo,{oU{o,� + zo,{xU{x,� + ηo�
We test the significance of parameter differences.
For crude oil and GDP growth, we find significant differences, which indicates seasonality effects in those variables.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 38
Identification of QVAR, VAR and VARMA
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 39
Identification of structural forms
The QVAR, VAR and VARMA models used in this paper are recursively identified structural models (Kilian and Lütkepohl 2017, Chapter 9).
This identification method is supported by the argument that oil price shocks may act as domestic supply shocks for the US economy (Kilian and Lütkepohl 2017, p. 239).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 40
Identification of structural forms
Therefore, we specify
�� =
���� 0 0
Ω&��� Ω&&
�� 0
Ωo��� Ωo&
�� Ωoo��
with Ω���� > 0, Ω&&
�� > 0 and Ωoo�� > 0.
That is Σ is factorized by using the Cholesky decomposition.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 41
Identification of model parameters
For both QVAR and VARMA we use the restriction:
Ψ� = Ψ�,�� × ;�
This implies that � is diagonal and �,�� = �,&& = �,oo.
Without this restriction, the ML procedure did not converge.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 42
ML estimates and model diagnostics
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 43
ML estimates and model diagnostics
Firstly, we find that some elements of Φ and Ω�� are significantly different from zero for all.
This suggests significant dynamic and contemporaneous interaction effects, respectively, among crude oil, inflation and GDP growth.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 44
ML estimates and model diagnostics
Secondly, for both VAR specifications, the Ljung-Box (1978) (LB) test suggests that :&� and &� each form a non-independent time series.
For all VARMA and QVAR specifications, the LB test suggests that each reduced-form and structural-form error term forms an independent time series.
For QVAR(1), the LB test suggests that the score function ��� is a non-independent time series. For QVAR(2), we find that each score function forms an independent time series.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 45
ML estimates and model diagnostics
Thirdly, for all models, we find that conditions of consistency and asymptotic normality of ML and QML are supported.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 46
ML estimates and model diagnosticsFourthly, we compare the statistical performance of QVAR, VAR and VARMA, by estimating the following likelihood-based performance metrics:
(i) mean LL; (ii) mean Akaike information criterion (AIC); (iii) mean Bayesian information criterion (BIC); (iv) mean Hannan--Quinn criterion (HQC).
All model performance metrics suggest that the statistical performance of QVAR is superior to that of VAR and VARMA.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 47
Impulse response functions
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 48
Impulse response functions
We find that Φ�,�& = Φ�,�o = Φ&,�& = Φ&,�o = 0.
Since by the Cholesky decomposition we impose that the oil
price changes are predetermined with �&�� = �o
�� = 0(Kilian 2008), therefore, we conclude that oil price changes are strictly exogenous (Kilian 2008) for the parameters of interest in the inflation and the GDP growth rate equations.
Furthermore, since Ω&��� = 0, therefore, the inflation rate is
also predetermined (Kilian 2008).
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 49
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 50
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 51
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 52
Robustness analysis
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 53
Robustness analysis
Perhaps, the most surprising result of the present paper is that for QVAR(2) stochastic annual cyclical effects are identified for the impulse response function.
One may argue that these cyclical effects may be spurious, because they may be due to the non-significant parameters within Φ�, Φ& and Ω��, or they may be due to the order of variables in QVAR. We investigate this concern.
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 54
Restricted QVAR(2) and restricted VARMA(2,1)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 55
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 56
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 57
Variable order: crude oil, GDP growth and inflation for QVAR(2)
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 58
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 59
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 60
Thank you for your [email protected]@[email protected]@UFM.EDU
[email protected] [email protected] [email protected] [email protected]
GESG SEMINAR, UNIVERSIDAD FRANCISCO MARROQUIN, 9 NOVEMBER 2017 61