A Comparative Modelling of Essential Characteristics of Volatility: Simulation and Empirical Study

Richard T. A. Samuel *, Charles Chimedza, Caston Sigauke · 2025

Read A Comparative Modelling of Essential Characteristics of Volatility: Simulation and Empirical Study by Richard T. A. Samuel *, Charles Chimedza, Caston Sigauke (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-02031.txt.

Introduction

Richard Samuel, Charles Chimedza, and Caston Sigauke present a 2025 study of stock market volatility. Volatility clusters, drifts back to a mean, and reacts unevenly to bad news. Tails are fat, and shocks echo at two speeds. Risk managers need models that capture all six of these traits. The team runs five time-varying models against the trait list. Both simulated data and a real index, the Indian S&P index, serve as tests.

Market returns in the tests show all six traits. Long-memory analysis splits volatility into long and short parts. Shocks hit the short part harder but fade fast there. The long part persists strongly yet moves slowly. Traders and regulators get a trait-by-trait grading of common models. Picking a model by one trait alone can leave a risk desk blind to the others.

Catalog record

CREATORRichard T. A. Samuel *, Charles Chimedza, Caston Sigauke
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE107 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-02031

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Comparative Modelling of Essential Characteristics of Volatility: Simulation and Empirical Study Richard T. A. Samuel *, Charles Chimedza, Caston Sigauke Abstract This study utilised the dynamics of five time-varying models to estimate six essential features of financial return volatility that are relevant for robust risk management. These features include pronounced persistence, mean reversion, leverage effect or volatility asymmetry, conditional skewness, conditional fat-tailedness, and the long memory behaviour of volatility decomposition into long-term and short-term components. Both simulation and empirical evidence are provided. Through the applications of these models using the S&P Indian index, the study shows that the market returns are characterised by these volatility features. Our findings from the long-memory behaviour revealed that although the response to shocks is greater in the short-term component, it is however short-lived. On the contrary, despite a high degree of persistence in the long-term component, market information or unexpected news arrival only has a low long-run impact on the market. Based on this, the long-run investment risks within the Indian stock market seem to be under control. Hence, our findings suggest that rational investors should try to stay calm with the arrival of unexpected news in the market because the long-run effect of such news will not be severe, and the market will eventually return to its normal state. keywords leverage, persistence, score-driven model, simulation, time-variation. † † firstpage: 1 † † volume: 1 † † issue: 1 † † articlenumber: 0 † † year: 2023 † † copyright-year: 2023 † † hreflink: https://doi.org/ † † titlecitation: Title † † authornames: Richard T. A. Samuel, Charles Chimedza and Caston Sigauke † † authorcitation: Samuel, R. T. A.; Chimedza, C.; Sigauke, C. † † address: 1 School of Statistics and Actuarial Science, University of the Witwatersrand, Johannesburg, South Africa; [email protected] 1 School of Statistics and Actuarial Science, University of the Witwatersrand, Johannesburg, South Africa; [email protected] 2 Department of Mathematical and Computational Sciences, University of Venda, Thohoyandou, South Africa; [email protected] † † corresponding: Correspondence: [email protected] † † firstnote: Current address: School of Statistics and Actuarial Science, University of the Witwatersrand, Johannesburg, South Africa. \secondnote These authors contributed equally to this work. 1 Introduction Stock market indices are prone to be characterised by features of volatility such as pronounced persistence, mean reversion, leverage effect (i.e., volatility asymmetry), conditional skewness, conditional fat-tailedness, and the long memory behaviour of volatility decomposition Harvey and Sucarrat 2014 . Hence, these characteristics are usually exhibited by financial returns (see Sucarrat 2013 ; Engle and Patton 2001 ). Volatility persistence is a process where the return of today affects the future’s forecast variance Engle and Patton 2001 . The economic implications of the degree of persistence of a shock include its influence on dynamic hedging policies, the valuation of options and the price of securities Porterba and Summers 1986 . Mean reversion implies that there is a normal volatility level to which volatility will eventually return. Asymmetry is a process in which positive and negative shocks have different impacts on volatility. Asymmetry in equity returns is sometimes known as leverage effect Engle and Patton 2001 . Leverage effect implies that volatility tends to be higher following negative returns Sucarrat 2013 . Conditional fat-tailedness implies that the standardised conditional return is more fat-tailed than the Normal distribution, while conditional skewness denotes that the standardised return is not symmetric Sucarrat 2013 . The long-memory behaviour of volatility decomposition occurs when volatility is decomposed into one short-term component and one long-term component Sucarrat 2013 . Volatility on its own is not directly observable, hence its measurement and the modeling of its evolution rely on some measurement methods. Two well documented methods in the literature are the Generalised Autoregressive Conditional Heteroscedasticity (GARCH) model and the Generalised Autoregressive Score (GAS) model. These two models use the conditional variance for measuring volatility (see Ardia et al. 2019 ; Chou 1988 ; Takaishi 2018 ). These models can reliably capture some features of asset returns. However, it is well reported that asymmetric GARCH models perform better at modelling volatility than the simple GARCH model (see Takaishi 2018 ), since the simple GARCH model is incapable of capturing asymmetries in the volatility process Feng and Shi 2017 . Hence, this study applies a robust extension of the GARCH model called the family GARCH (fGARCH) model Hentschel 1995 that incorporates asymmetric volatility process. Based on this, the fGARCH and GAS models, with their extensions, are used to model the stated essential features of volatility using the daily equity returns. To be specific, the study comparatively applies the fGARCH and GAS models to estimate the magnitude and dynamics of the persistence (with mean reversion) in the conditional volatility of the returns. The main difference between these models is that the fGARCH model uses the dynamics of the residuals to drive the conditional variance, while the GAS model uses the dynamics of the conditional score to drive the time-varying conditional variance. Furthermore, the study uses the extensions of these two models, i.e., the extensions of the (f)GARCH model that consist of the Threshold GARCH (TGARCH) and Component GARCH (CGARCH) models, and the extension of the GAS model that consists of the Beta-Skew- t -EGARCH models to estimate the remaining stated features of returns. Specifically, the study compares the applications of the TGARCH and CGARCH models on one side, with the multi-functions of the Beta-Skew- t -EGARCH model on the other side to model the asymmetry (leverage effect), skewness, fat-tails and long-memory behaviour of volatility. Engle and Patton Engle and Patton 2001 used the GARCH(1,1) model to estimate various stylised facts of volatility in the Dow Jones Industrial Index returns from 1988 to 2000, and found a highly persistence volatility estimate with a half-life of about 73 days. Moreover, the tendency of volatility to persist over time is well documented in the literature (see Chou 1988 ; Pandey and Kumar 2017 among others). Tsay Tsay 2005 used the TGARCH model to study asymmetry in the IBM daily returns. Oh and Patton Oh and Patton 2016 used the GAS model with a factor copula model to study systemic risk. Other GAS modelling applications are in credit risk analysis Creal et al. 2014 , spatial econometrics Blasques et al. 2014 ; Catania and Bill ´ e 2017 , high-frequency data Opschoor et al. 2018 ; Gorgi et al. 2019 , among others. See Ardia et al. 2019 ; Haddad et al. 2023 for more dynamic applications of the GAS model. Studies associated with the decomposition of return volatility to its permanent-transitory components through the CGARCH model are increasingly found in the literature (see Chen and Shen 2000 ; Chiang et al. 2006 ; Ane 2006 ; Engle and Lee 1999 ). The applications of the Beta-Skew- t -EGARCH model for modelling relevant features of volatility are well studied in Sucarrat Sucarrat 2013 , and Harvey and Sucarrat Harvey and Sucarrat 2014 . Our study should result in responses to the following questions. One, which assumed innovations are the most adequate from the fGARCH and GAS modelling to estimate the persistence of the return volatility? Two, how persistent is the market returns volatility? Three, how are the models compared in terms of performances? Four, what volatility

Read the complete text