Box–Cox realized asymmetric stochastic volatility models with generalized Student's<i>t</i>-error distributions

Nugroho, Didit Budi; Morimoto, Takayuki

doi:10.1080/02664763.2015.1125862

Cited by 22 publications

(16 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among the well-known ones are Student-t distribution and generalized Student-t distributions (e.g., see Nugroho et al (2014Nugroho et al ( , 2016). …”

Section: Future Workmentioning

confidence: 99%

Estimation of Exchange Rate Volatility using APARCH-type Models: A Case Study of Indonesia (2010–2015)

Nugroho

Susanto

Pratama

2017

JESP

View full text Add to dashboard Cite

Volatiliy measurement and modeling is an important aspect in many areas of finance. The main purpose of this study is to apply seven APARCH-type models with (1,1) lags to investigate the behavior of exchange rate volatility for the EUR, JPY, and USD selling exchange rates to IDR for the duration from January 2010 to December 2015. The competing models include ARCH, GARCH, TARCH, TS-ARCH, GJR-GARCH, NARCH, and APARCH used with Gaussian normal distribution. In order to estimate the model parameters, this study applies the Bayesian inference using the adaptive random walk Metropolis method in the MCMC algorithm. Empirical results based on the deviance information criterion indicate that the GARCH (1,1), APARCH (1,1), and TARCH (1,1) models provide the best fit for the EUR, JPY, and USD data, respectively. In those models, both the JPY and USD data have significant negative leverage effect at the 99% credible level. Moreover, the JPY returns also have significant Taylor effect in return volatility at the 99% credible level.

show abstract

“…Among the well-known ones are Student-t distribution and generalized Student-t distributions (e.g., see Nugroho et al (2014Nugroho et al ( , 2016). …”

Section: Future Workmentioning

confidence: 99%

Estimation of Exchange Rate Volatility using APARCH-type Models: A Case Study of Indonesia (2010–2015)

Nugroho

Susanto

Pratama

2017

JESP

View full text Add to dashboard Cite

show abstract

“…For example, the basic SV‐ t model (KASTNER, ; ABANTO‐VALLE et al , ) is specified as

y_{t} = e^{h_{t} / 2} \cdot λ_{t}^{1 / 2} z_{t}, λ_{t} \sim IG (k / 2, k / 2), z_{t} \overset{i i d}{\sim} N (0, 1),

where λ t is the scaling factor and its prior distribution can be equivalently expressed as a scaled inverse χ 2 distribution with df = k and the scaling parameter =1. Many other variants also adopt this construction, such as the Box–Cox realized asymmetric non‐central SV‐ t (BCR‐ASV‐NCT) model proposed by NUGROHO and MORIMOTO (). For these models, given the typical diffuse prior k ∼Gamma( α , β ) or when α =1, k ∼Exp( β ), the second‐order derivative of the log full conditional is derived as

\frac{d^{2}}{d k^{2}} normal l o g π_{1} (k | bold-italic \cdot) = \frac{T}{4} ([], \frac{2}{k} - ψ_{1} ((), \frac{k}{2})) - \frac{α - 1}{k^{2}},

which is negative by using the result of Equation , implying log‐concavity of π 1 ( k | · ).…”

Section: Lf‐dcmsv Modelmentioning

confidence: 99%

“…NUGROHO and MORIMOTO () obtains log‐concavity of π 1 ( k | · ) using the aforementioned argument based on KU et al () in their BCR‐ASV‐NCT model (which includes the BCR‐ASV‐T variant). They further compared the performance of ARS and the Riemannian manifold Hamiltonian Monte Carlo (RMHMC) methods on the posterior simulation.…”

Section: Lf‐dcmsv Modelmentioning

confidence: 99%

“…where t is the scaling factor and its prior distribution can be equivalently expressed as a scaled inverse 2 distribution with df = k and the scaling parameter = 1. Many other variants also adopt this construction, such as the Box-Cox realized asymmetric non-central SV-t (BCR-ASV-NCT) model proposed by Nugroho and Morimoto (2016). For these models, given the typical diffuse prior k ∼ Gamma( , ) or when = 1, k ∼ Exp( ), the second-order derivative of the log full conditional is derived as…”

Section: Log-concavity Of the Full Conditional Distribution For Dfmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic correlation multivariate stochastic volatility with latent factors

Ghosh

et al. 2017

Statistica Neerlandica

View full text Add to dashboard Cite

Modeling the correlation structure of returns is essential in many financial applications. Considerable evidence from empirical studies has shown that the correlation among asset returns is not stable over time. A recent development in the multivariate stochastic volatility literature is the application of inverse Wishart processes to characterize the evolution of return correlation matrices. Within the inverse Wishart multivariate stochastic volatility framework, we propose a flexible correlated latent factor model to achieve dimension reduction and capture the stylized fact of 'correlation breakdown' simultaneously. The parameter estimation is based on existing Markov chain Monte Carlo methods. We illustrate the proposed model with several empirical studies. In particular, we use high-dimensional stock return data to compare our model with competing models based on multiple performance metrics and tests. The results show that the proposed model not only describes historic stylized facts reasonably but also provides the best overall performance.

show abstract

“…These considered distributions are the noncentral Student-t and skewed Student-t as suggested by [26,27] in the context of stochastic volatility. …”

Section: Conclusion and Extensionsmentioning

confidence: 99%

Volatility modeling for IDR exchange rate through APARCH model with student-t distribution

Nugroho¹,

Susanto²

2017

AIP Conference Proceedings

View full text Add to dashboard Cite

Abstract. The aim of this study is to empirically investigate the performance of APARCH(1,1) volatility model with the Student-t error distribution on five foreign currency selling rates to Indonesian rupiah (IDR), including the Swiss franc (CHF), the Euro (EUR), the British pound (GBP), Japanese yen (JPY), and the US dollar (USD). Six years daily closing rates over the period of January 2010 to December 2016 for a total number of 1722 observations have analysed. The Bayesian inference using the efficient independence chain Metropolis-Hastings and adaptive random walk Metropolis methods in the Markov chain Monte Carlo (MCMC) scheme has been applied to estimate the parameters of model. According to the DIC criterion, this study has found that the APARCH(1,1) model under Student-t distribution is a better fit than the model under normal distribution for any observed rate return series. The 95% highest posterior density interval suggested the APARCH models to model the IDR/JPY and IDR/USD volatilities. In particular, the IDR/JPY and IDR/USD data, respectively, have significant negative and positive leverage effect in the rate returns. Meanwhile, the optimal power coefficient of volatility has been found to be statistically different from 2 in adopting all rate return series, save the IDR/EUR rate return series.

show abstract

Box–Cox realized asymmetric stochastic volatility models with generalized Student'st-error distributions

Cited by 22 publications

References 24 publications

Estimation of Exchange Rate Volatility using APARCH-type Models: A Case Study of Indonesia (2010–2015)

Estimation of Exchange Rate Volatility using APARCH-type Models: A Case Study of Indonesia (2010–2015)

Dynamic correlation multivariate stochastic volatility with latent factors

Volatility modeling for IDR exchange rate through APARCH model with student-t distribution

Contact Info

Product

Resources

About