GARCH Model With Fat-Tailed Distributions and Bitcoin Exchange Rate Returns

Liu, Ruiping; Shao, Zhichao; Wei, Guodong; Wang, Wei

doi:10.20448/2002.11.71.75

Cited by 17 publications

(9 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Later, Lahmiri et al (2018) employed the FIGARCH model and found that volatility in all the Bitcoin markets considered in their study exhibits long-range dependence, while Mensi et al (2018) found that, after accounting for structural breaks, long memory in the mean and variance decreases. Among other authors who have studied the price volatility of cryptocurrencies are Chu et al (2017), Liu et al (2017), and Takaishi (2018), all of whom also used GARCH-type models, while Phillip et al (2018) employed the stochastic volatility model.…”

Section: Literature Reviewmentioning

confidence: 99%

An empirical investigation of volatility dynamics in the cryptocurrency market

Katsiampa

2019

Research in International Business and Finance

116

View full text Add to dashboard Cite

By employing an asymmetric Diagonal BEKK model, this paper examines volatility dynamics of five major cryptocurrencies, namely Bitcoin, Ether, Ripple, Litecoin, and Stellar Lumen. It is shown that the conditional variances of all the five cryptocurrencies are significantly affected by both previous squared errors and past conditional volatility. Moreover, in the case of Bitcoin, Ether, Ripple, and Litecoin, asymmetric past shocks have a significant effect in the current conditional variance. Similar results are obtained for the cryptocurrencies' conditional covariances, which are significantly affected by cross products of previous error terms and past covariance terms while capturing asymmetric effects of past shocks accordingly.It is also shown that time-varying conditional correlations exist and are mostly positive. Finally, the cryptocurrencies' volatility dynamics are found to be responsive to major news, with Bitcoin and Litecoin exhibiting one structural breakpoint each in the conditional variance. The results improve our understanding of interdependencies between cryptocurrencies as well as of the events that affect their volatility dynamics and thus have important implications for both cryptocurrency users and investors.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

An empirical investigation of volatility dynamics in the cryptocurrency market

Katsiampa

2019

Research in International Business and Finance

116

View full text Add to dashboard Cite

show abstract

“…For adequacy and parsimonious reasons, Exo − GJR − MGARCH(1, 1) and xo − GJR − DCC − MGARCH(1, 1) (Hansen and Lunde 2005;Ashley and Patterson 2010;Lim and Sek 2013). As such, based on the empirical properties of return (Cont 2001), the distributional assumption on the error vector z t is the multivariate Gaussian distribution (Liu et al 2017). The joint density is specified as follows:…”

Section: Models Identification Strategymentioning

confidence: 99%

“…The return of Bitcoin empirically exhibits leverage effect (high volatility during low return periods and low volatility during high return periods) and has evidence of return asymmetric (Bouri et al 2017a;Liu et al 2017). Therefore, since the return of Bitcoin is asymmetric and prone to leverage effect, this paper incorporates the GJR (Glosten et al 1993) asymmetric component in out GARCH model.…”

mentioning

confidence: 99%

Could stock hedge Bitcoin risk(s) and vice versa?

Okorie

2019

Digit Finance

View full text Add to dashboard Cite

This paper is saddled with the task of investigating the Bitcoin market behavior in the presence of a government risk. This is because both the institutional and retail investors' interests in the Bitcoin market are growing rapidly. Conversely, the seemingly unregulated nature of this market is a serious concern to most economies and results to the placement of ban on Initial Coin Offering (ICO) in some economies by the government. Daily series of return and volume within the window of the ICO ban in China was used for the Bitcoin market and S&P500 stock market to examine the effect of a government risk in the Bitcoin market and possible hedging capabilities. Empirical results show that the ban dampened Bitcoin returns and the returns from each market can predict the other. The Exogenous Dynamic Conditional Correlation (Exo-DCC) model result suggests that, yes! the S&P500 stocks are capable of hedging Bitcoin risk, while Bitcoin can also hedge S&P500 stocks' risks and vice versa. The Exogenous BEKK (Exo-BEKK) model result shows evidence of bidirectional volatility spill over between the two markets studied. In practice, investors (institutions and retailers) can comfortably form a robust investment portfolio with (at least) these two assets and develop a hedging strategy, such that the impacts of risks on the portfolio's returns are safely hedged.

show abstract

“…Ma (2017) and Ma et al (2017) showed that the NRIG cannot outperform the Students t-distribution in modelling the metals spot return. Liu et al (2017) also suggested that the older Students t-distribution still performs better than the NRIG distribution for modelling the daily Bitcoin exchange rate returns. Additionally, some researchers suggest that the NIG and full GHD provide a better fit to some financial data than an NRIG.…”

Section: Generalised Hyperbolic Distribution and Its Subclasses In Var Estimationmentioning

confidence: 99%

Modelling Crude Oil Returns Using the NRIG Distribution

Chinhamu

Mabaso

Chifurira

2020

Stat., optim. inf. comput.

View full text Add to dashboard Cite

Over the past decade, crude oil prices have risen dramatically, making the oil market very volatile and risky; hence, implementing an efficient risk management tool against market risk is crucial. Value-at-risk (VaR) has become the most common tool in this context to quantify market risk. Financial data typically have certain features such as volatility clustering, asymmetry, and heavy and semi-heavy tails, making it hard, if not impossible, to model them by using a normal distribution. In this paper, we propose the subclasses of the generalised hyperbolic distributions (GHDs), as appropriate models for capturing these characteristics for the crude oil and gasoline returns. We also introduce the new subclass of GHDs, namely normal reciprocal inverse Gaussian distribution (NRIG), in evaluating the VaR for the crude oil and gasoline market. Furthermore, VaR estimation and backtesting procedures using the Kupiec likelihood ratio test are conducted to test the extreme tails of these models. The main findings from the Kupiec likelihood test statistics suggest that the best GHD model should be chosen at various VaR levels. Thus, the final results of this research allow risk managers, financial analysts, and energy market academics to be flexible in choosing a robust risk quantification model for crude oil and gasoline returns at their specific VaR levels of interest. Particularly for NRIG, the results suggest that a better VaR estimation is provided at the long positions.

show abstract

GARCH Model With Fat-Tailed Distributions and Bitcoin Exchange Rate Returns

Cited by 17 publications

References 9 publications

An empirical investigation of volatility dynamics in the cryptocurrency market

An empirical investigation of volatility dynamics in the cryptocurrency market

Could stock hedge Bitcoin risk(s) and vice versa?

Modelling Crude Oil Returns Using the NRIG Distribution

Contact Info

Product

Resources

About