Long Memory Dynamics for Multivariate Dependence Under Heavy Tails

Janus, P.; Koopman, Siem Jan

doi:10.2139/ssrn.1971255

Cited by 13 publications

(20 citation statements)

References 39 publications

(42 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The simulated bivariate sample has

T = 5000

observations. The DGP in (16) is similar to the designs in Engle (), Janus et al () and Hafner and Manner ().

\begin{matrix} left \\ true{\begin{matrix} left \\ y_{1 t} = \sqrt{h_{1 t}} z_{1 t}, h_{1 t} = 0.01 + 0.05 y_{1 t - 1}^{2} + 0.94 h_{1 t - 1}, \\ y_{2 t} = \sqrt{h_{2 t}} z_{2 t}, h_{2 t} = 0.5 + 0.2 y_{2 t - 1}^{2} + 0.5 h_{2 t - 1}, \end{matrix} \\ u_{1 t} = F_{1} true(z_{1 t} true), u_{2 t} = F_{2} true(z_{2 t} true), \\ true(u_{1 t}, u_{2 t} true) \sim C true(u_{1 t}, u_{2 t}; θ_{t}, η true), θ_{t} = Ψ true(λ_{t} true) . \end{matrix}

…”

Section: Monte Carlo Simulationssupporting

confidence: 58%

Long Memory in Asymmetric Dependence Between LME and Chinese Aluminum Futures

Gong

2015

Journal of Futures Markets

View full text Add to dashboard Cite

The co-dependence of many asset returns has been shown to be asymmetric and to follow long memory dynamics in recent studies. To capture the two features simultaneously, based on Hafner and Manner (2012), we propose the new model of fractionally integrated stochastic copula allowing for long memory in the evolution of co-dependence. In the empirical analysis, the aluminum futures markets of London Metal Exchange and Shanghai Futures Exchange were shown to have stronger co-dependence in market downturns than in upturns, and long memory was present in the dynamics of both upper and lower-tail dependence.

show abstract

“…The simulated bivariate sample has

T = 5000

observations. The DGP in (16) is similar to the designs in Engle (), Janus et al () and Hafner and Manner ().

\begin{matrix} left \\ true{\begin{matrix} left \\ y_{1 t} = \sqrt{h_{1 t}} z_{1 t}, h_{1 t} = 0.01 + 0.05 y_{1 t - 1}^{2} + 0.94 h_{1 t - 1}, \\ y_{2 t} = \sqrt{h_{2 t}} z_{2 t}, h_{2 t} = 0.5 + 0.2 y_{2 t - 1}^{2} + 0.5 h_{2 t - 1}, \end{matrix} \\ u_{1 t} = F_{1} true(z_{1 t} true), u_{2 t} = F_{2} true(z_{2 t} true), \\ true(u_{1 t}, u_{2 t} true) \sim C true(u_{1 t}, u_{2 t}; θ_{t}, η true), θ_{t} = Ψ true(λ_{t} true) . \end{matrix}

…”

Section: Monte Carlo Simulationssupporting

confidence: 58%

Long Memory in Asymmetric Dependence Between LME and Chinese Aluminum Futures

Gong

2015

Journal of Futures Markets

View full text Add to dashboard Cite

show abstract

“…It is easy to generalize this specification to include more lags of f t and s t (see Creal et al, 2013), non-linearity, structural time series dynamics (see Harvey and Luati, 2014), or fractional integration (see Janus et al, 2014).…”

Section: A Motivating Examplementioning

confidence: 99%

Generalized Autoregressive Method of Moments

et al. 2015

Self Cite

View full text Add to dashboard Cite

We introduce a new estimation framework which extends the Generalized Method of Moments (GMM) to settings where a subset of the parameters vary over time with unknown dynamics. To filter out the dynamic path of the time-varying parameter, we approximate the dynamics by an autoregressive process driven by the score of the local GMM criterion function. Our approach is completely observation driven, rendering estimation and inference straightforward. It provides a unified framework for modeling parameter instability in a context where the model and its parameters are only specified through (conditional) moment conditions, thus generalizing approaches built on fully specified parametric models. We provide examples of increasing complexity to highlight the advantages of our method.

show abstract

“…The time-varying parameter f t as specified in (3) is thus observation driven in the classification of Cox [1981]. More complicated dynamics than the ones specified in equation (3) can be added to the specification; see for example Janus et al [2011] for fractionally integrated dynamics, Creal et al [2013] for higher-order dynamics, and Harvey and Luati [2014] for higher order dynamics as well as structural time series dynamics. For our current purposes, however, the GAS(1,1) dynamics suffice.…”

Section: Score Driven Ewmamentioning

confidence: 99%