Professional Documents
Culture Documents
Download pdf Modeling Dependence In Econometrics Selected Papers Of The Seventh International Conference Of The Thailand Econometric Society Faculty Of Economics Chiang Mai University Thailand January 8 10 2014 1S ebook full chapter
Download pdf Modeling Dependence In Econometrics Selected Papers Of The Seventh International Conference Of The Thailand Econometric Society Faculty Of Economics Chiang Mai University Thailand January 8 10 2014 1S ebook full chapter
https://textbookfull.com/product/semantic-technology-4th-joint-
international-conference-jist-2014-chiang-mai-thailand-
november-9-11-2014-revised-selected-papers-1st-edition-thepchai-
supnithi/
https://textbookfull.com/product/advances-in-web-based-learning-
icwl-2018-17th-international-conference-chiang-mai-thailand-
august-22-24-2018-proceedings-gerhard-hancke/
https://textbookfull.com/product/data-mining-and-big-data-4th-
international-conference-dmbd-2019-chiang-mai-thailand-
july-26-30-2019-proceedings-ying-tan/
https://textbookfull.com/product/formal-modeling-and-analysis-of-
timed-systems-12th-international-conference-
formats-2014-florence-italy-september-8-10-2014-proceedings-1st-
edition-axel-legay/
https://textbookfull.com/product/multimedia-modeling-20th-
anniversary-international-conference-mmm-2014-dublin-ireland-
january-6-10-2014-proceedings-part-ii-1st-edition-sema-alacam/
https://textbookfull.com/product/birds-of-thailand-michael-
webster/
Van-Nam Huynh
Vladik Kreinovich
Songsak Sriboonchitta Editors
Modeling Dependence
in Econometrics
Advances in Intelligent Systems and Computing
Volume 251
Series Editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
The series “Advances in Intelligent Systems and Computing” contains publications on theory,
applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually
all disciplines such as engineering, natural sciences, computer and information science, ICT,
economics, business, e-commerce, environment, healthcare, life science are covered. The list
of topics spans all the areas of modern intelligent systems and computing.
The publications within “Advances in Intelligent Systems and Computing” are primarily
textbooks and proceedings of important conferences, symposia and congresses. They cover
significant recent developments in the field, both of a foundational and applicable character.
An important characteristic feature of the series is the short publication time and world-wide
distribution. This permits a rapid and broad dissemination of research results.
Advisory Board
Chairman
Members
Jun Wang, The Chinese University of Hong Kong, Shatin, Hong Kong
e-mail: jwang@mae.cuhk.edu.hk
Van-Nam Huynh · Vladik Kreinovich
Songsak Sriboonchitta
Editors
Modeling Dependence
in Econometrics
ABC
Editors
Van-Nam Huynh Songsak Sriboonchitta
Japan Advanced Institute of Science Faculty of Economics
and Technology Chiangmai University
Ishikawa Chiangmai
Japan Thailand
Vladik Kreinovich
Department of Computer Science
University of Texas at El Paso
El Paso, Texas
USA
4.4 Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Residual Based Cusum Test for Parameter Change in AR-GARCH
Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Sangyeol Lee, Jiyeon Lee
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
2 Residual Based Cusum Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3 Empirical Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.1 Simulation Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.2 Real Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Dependence and Association Concepts through Copulas . . . . . . . . . . . . . . 113
Zheng Wei, Tonghui Wang, Wararit Panichkitkosolkul
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
2 Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3 Invariance of Affiliation of Subcopula through Bilinear
Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4 Average and Local Measures of Affiliations . . . . . . . . . . . . . . . . . . 121
5 Conditions on Affiliation in the Bivariate Skew
Normal Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Pairs Trading via Three-Regime Threshold Autoregressive GARCH
Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Cathy W.S. Chen, Max Chen, Shu-Yu Chen
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
2.1 Threshold AR Model with GARCH Effect . . . . . . . . . . . 129
2.2 Priors and Likelihood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
2.3 Pairs Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Testing Dependencies in Term Structure of Interest Rates . . . . . . . . . . . . . 141
Kian-Guan Lim
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
2 Deriving the Spot and Forward Rates . . . . . . . . . . . . . . . . . . . . . . . . 142
3 Term Structure of Interest Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4 Tests of Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
X Contents
6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
An Integration of Eco-Health One-Health Transdisciplinary
Approach and Bayesian Belief Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
Chalisa Kallayanamitra, Pisit Leeahtam, Manoj Potapohn,
Bruce A. Wilcox, Songsak Sriboonchitta
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
3.1 Population and Sampling Design . . . . . . . . . . . . . . . . . . . . 465
3.2 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
3.3 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
4.1 Benefit of Integration of Eco-HealthOne-Health
Transdisciplinary Approach and Bayesian Belief
Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
4.2 Constraint of Bayes Risk Minimization Using
Symmetric Loss Function . . . . . . . . . . . . . . . . . . . . . . . . . . 471
4.3 Lack of Optimal Sample Size Determination . . . . . . . . . . 472
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
Factors Affecting Hospital Stay Involving Drunk Driving and
Non-Drunk Driving in Phuket, Thailand . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Jirakom Sirisrisakulchai, Songsak Sriboonchitta
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
2 Switching Regression Model for Hospital Stay . . . . . . . . . . . . . . . . 481
3 Copula Approach for Modeling Switching Regression . . . . . . . . . 482
4 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484
5 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
5.1 Binary Choice Equation for Alcohol Consumption . . . . . 485
5.2 Binary Outcome for Fatality or Injury . . . . . . . . . . . . . . . 486
5.3 Zero-Inflated Negative Binomial Models for the
Length of Stay in the Hospital . . . . . . . . . . . . . . . . . . . . . . 486
6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
How Macroeconomic Factors and International Prices Affect
Agriculture Prices Volatility?-Evidence from GARCH-X Model . . . . . . . 491
Gong Xue, Songsak Sriboonchitta
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
2.1 China Agricultural Commodity Prices,
Macroeconomic Variables and International
Price Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
XVIII Contents
6 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
Valuation of Interest Rate Derivatives under CSA Discounting . . . . . . . . 551
Amy R. Daniels, Coenraad C.A. Labuschagne,
Theresa M. Offwood-le Roux
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
2 Classical vs OIS Swap Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 554
2.1 Classic Approach to Price and Value Interest
Rate Swaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
2.2 CSA Approach to Pricing and Valuing Interest
Rate Swaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 556
3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 558
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
Systemic Knowledge Synthesis for Product Recommendation . . . . . . . . . 561
Yoshiteru Nakamori
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
2 Theory of Knowledge Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 562
3 Product Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563
4 Data Collection and Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565
4.1 Correspondence Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 566
4.2 Fuzzy-Set Theoretical Data Processing . . . . . . . . . . . . . . . 567
5 Information Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
6 Knowledge Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571
7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574
1 Introduction
Under the regulations introduced in Finance since Basel 2, bank capital is risk-
sensitive. Financial institutions are required to measure the riskiness of their assets
and, for instance, to hold more capital to compensate more risk. While the Value-
at-Risk (VaR), defined as a quantile of some loss distribution, continues to play
Christian Francq
CREST and University Lille 3 (EQUIPPE),
BP 60149, 59653 Villeneuve d’Ascq cedex, France
e-mail: christian.francq@univ-lille3.fr
Jean-Michel Zakoı̈an
EQUIPPE (University Lille 3) and CREST,
15 boulevard Gabriel Péri, 92245 Malakoff Cedex, France
e-mail: zakoian@ensae.fr
The authors gratefully acknowledge financial support of the ANR via the Project
ECONOM&RISK (ANR 2010 blanc 1804 03). The second author gratefully thanks the
IDR ”Risques systemiques” for financial support.
Corresponding author.
where (ηt ) is a sequence of iid random variables, ηt is independent of {εu , u < t},
θ 0 ∈ Rd is a parameter belonging to a parameter space Θ , and σ : R∞ × Θ → (0, ∞).
The most widely used specifications of volatility belong to this class, in particular
the GARCH(p, q) model of Engle (1982) and Bollerslev (1986),
εt = σt ηt ,
(2)
σt2 = ω0 + ∑qi=1 a0i εt−i
2 + p b σ2 ,
∑ j=1 0 j t− j
where θ 0 = (ω0 , a01 , . . . , b0p ) satisfies ω0 > 0, a0i ≥ 0, b0 j ≥ 0. For this model, if
the lag polynomial β (L) = 1 − ∑ pj=1 b0 j L j has its roots outside the unit disk, we
have a representation of the form (1) given by
∞
σt2 = β (1)−1 ω0 + ∑ γi εt−i
2
,
i=1
to the class (1) are the EGARCH, GJR-GARCH, TGARCH, QGARCH, APARCH,
Log-GARCH, models introduced, respectively, by Nelson (1991), Glosten, Jagan-
nathan and Runkle (1993), Zakoı̈an (1994), Sentana (1995), Ding, Granger and En-
gle (1993), and for the log-GARCH, under slightly different forms, by Geweke
(1986), Pantula (1986) and Milhøj (1987). See Francq and Zakoı̈an (2010) for an
overview on GARCH models.
The conditional VaR of a process (εt ) at risk level α ∈ (0, 1), denoted by
VaRt (α ), is defined by
Pt−1 [εt < −VaRt (α )] = α ,
where Pt−1 denotes the historical distribution conditional on {εu , u < t}. When (εt )
satisfies (1), the theoretical VaR is then given by
6 C. Francq and J.-M. Zakoı̈an
Remark 1. It can be noted that in the standard GARCH(p, q) model, the conditional
VaR at level α automatically satisfies the stochastic recurrence equation
q p
VaRt2 (α ) = ω0 ξα2 + ∑ a0i ξα2 εt−i
2
+ ∑ b0 j VaRt−
2
j (α ).
i=1 j=1
Direct modelling of the conditional VaR has been proposed in several papers, for
instance Engle and Manganelli (2004), Koenker and Xiao (2006), Gouriéroux and
Jasiak (2008). A difficulty in this approach is to constrain the model so as to guar-
antee the monotonicity of the conditional VaR as a function of the risk level. Mono-
tonicity is automatically satisfied in our approach.
θ n = arg min Q
n (θ ), (4)
θ ∈Θ
with
n
n (θ ) = n−1 ∑ t (θ ), εt2
Q t (θ ) = t2 (θ ).
+ log σ
t=1 σt2 (θ )
The following assumptions are required to derive the asymptotic properties of the
QMLE θ n .
A1: (εt ) is a strictly stationary and ergodic solution of Model (1). Moreover,
E|ε0 |s < ∞ for some s > 0.
A2: For any real sequence (xi ), the function θ → σ (x1 , x2 , . . . ; θ ) is continu-
ous. Almost surely, σt (θ ) ∈ (ω , ∞] for any θ ∈ Θ and for some ω > 0.
Multi-level Conditional VaR Estimation in Dynamic Models 7
Note that Assumptions A2, A3, A5 and A6 can be simplified for specific forms of
σt : for instance if the model is the GARCH (p, q) Model (2), A2 reduces to standard
assumptions on the lag polynomials of the volatility and A3, A5, A6 can be directly
verified. Note also that the only moment assumption on the observed process is the
existence of a small moment in A1, which is automatically satisfied for standard
models such as the classical GARCH(p, q).
Now let the residuals of the QML estimation
εt
η̂t = , t = 1, . . . , n,
σ̃t (θ n )
and let ξn,αi denote the empirical αi -quantile of η̂1 , . . . , η̂n . Let α = (α1 , . . . , αm ) ,
ξ n,α = (ξn,α1 , . . . , ξn,αm ) and let ξ α = (ξα1 , . . . , ξαm ) denote the vector of popula-
tion quantiles.
Remark 2. The derivation of the joint asymptotic properties of sample quantiles
goes back to Cramér (1946) in the iid case. Different articles have extended these re-
sults for the marginal quantiles of stationary processes, under different dependence
assumptions. See Dominicy, Hörmann, Ogata and Veredas (2013) and the references
therein. We cannot apply their results because (η̂t ) is not a stationary process.
The next result gives the joint asymptotic distributions of (θ n , ξ n,α ). Let Dt (θ ) =
−1
σt (θ )∂ σt (θ )/∂ θ .
Theorem 1. Assume ξαi < 0, for i = 1, . . . , m, E ηt2 = 1 and κ4 := E ηt4 < ∞. Suppose
that η1 admits a density f which is continuous and strictly positive in a neighbor-
hood of ξαi , for i = 1, . . . , m. Let A1-A3 and A4(θ 0 )-A6(θ 0 ) hold. Then
√
n θ n − θ 0 L
κ4 −1 −1
4 J λ α ⊗ J −1 Ω
√ → N (0, Σ α ), Σα = −1 ,
n(ξ n,α − ξ α ) λα ⊗Ω J ζα
8 C. Francq and J.-M. Zakoı̈an
κ4 − 1 p αi
λ α i = ξα i + ,
4 2 f (ξαi )
κ4 − 1 ξα i p α j ξα j p α i (αi ∧ α j )(1 − αi ∨ α j )
ζi j = ξαi ξα j + + + ,
4 2 f (ξα j ) 2 f (ξαi ) f (ξαi ) f (ξα j )
with pα = E η12 1{η1 <ξα } − α .
Proof. In view of Francq and Zakoı̈an (Proof of Theorem 4, 2012), we have, for
i = 1, . . . , m,
√ √ 1 1 n
n(ξαi − ξn,αi ) = ξαi Ω n(θ n − θ 0 ) + √ ∑ (1{ηt <ξα } − αi ) + oP(1),
f (ξαi ) n t=1 i
and
√ −JJ −1 n
n(θ n − θ 0 ) = √ ∑ (1 − ηt2)D
Dt + oP(1).
2 n t=1
Hence
√ 1 n 1
Covas n(θ n − θ 0 ), √ ∑ (1{ηt <ξα } − αi ) = pα J −1 Ω .
n t=1 i 2 i
We have Ω J −1 Ω = 1 (see Remark 3.1 in Francq and Zakoı̈an, 2013) and thus we
obtain
√ √
Covas { n(ξαi − ξn,αi ), n(ξα j − ξn,α j )} = ζi j .
t (α ) = −σ̃t (θ n )ξ α .
VaR n,
(see Francq and Zakoı̈an (2013)), and the asymptotic variance in Theorem 1 takes
the more explicit form
κ4 −1 −1
4 J 2λ α ⊗ θ 0
Σα = .
2λ α ⊗ θ 0 ζα
where
∂ σ̃t (θ n,α )
Δ α = ξ n,α , σ̃t (θ n )II m ,
∂θ
Φα−1
0
denotes the α0 -quantile of the standard Gaussian distribution, and I m denotes
the m × m identity matrix. Note that the choice of α0 (the risk estimation level) is
independent from that of the αi ’s (the financial risk levels). Drawing such CI allows
to underline the importance of the estimation risk for VaR evaluation.
10 C. Francq and J.-M. Zakoı̈an
n
t = −σ (εt−1 , εt−2 , . . . ; θ n ) ∑ G i i−1
DRM −G η̂n,i , (10)
i=1 n n
where (η̂n,n−i ) denotes the order statistics, obtained by ranking the η̂t in ascending
order: η̂n,1 < · · · < η̂n,n .
However, deriving the asymptotic distribution of this estimator, might be a
formidable task. To our knowledge such results do not exist in the literature. In
this section, we use VaR portfolios to obtain lower and upper bounds for a class of
DRM, leading to (approximate) asymptotic CI’s for such DRM.
It is not restrictive to assume α1 < α2 < . . . < αm . Suppose that the support of
the distortion cdf G is [α1 , αm ], that is
αm
DRMt = VaRt (u)dG(u). (11)
α1
u − α1 r
G(u) = 1u∈(α1 ,αm ) + 1u∈(αm,1) , (12)
αm − α1
where
Niinpä runokin siitä sanoo: »ei ollut suuri, eikä pieni, olihan oikea
vasikka». — Hirveä oli se aitakin, joka ympäröi Pohjolan pihaa, kun
Lemminkäinen sinne matkasi häitten viettoon:
*****