Multivariate Hyper-Rotated GARCH-BEKK

For large multivariate models of generalized autoregressive conditional heteroskedasticity (GARCH), it is important to reduce the number of parameters to cope with the ‘curse of dimensionality’. Recently, Laurent, Rombouts and Violante (2014 “Multivariate Rotated ARCH Models” Journal of Econometrics 179: 16–30) developed the rotated multivariate GARCH model, which focuses on the parameters for standardized variables. This paper extends the rotated multivariate GARCH model by considering a hyper-rotation, which uses a more flexible structure for the rotation matrix. The paper shows an alternative representation based on a random coefficient vector autoregressive and moving-average (VARMA) process, and provides the regularity conditions for the consistency and asymptotic normality of the quasi-maximum likelihood (QML) estimator for VARMA with hyper-rotated multivariate GARCH. The paper investigates the finite sample properties of the QML estimator for the new model. Empirical results for four exchange rate returns show the new specifications works satisfactory for reducing the number of parameters.

Betas in the time of corona: a conditional CAPM approach using multivariate GARCH model for India

PurposeThis paper empirically investigates the effect of the coronavirus pandemic (COVID-19) on the Indian financial market and firm betas, perhaps the first paper to do so. The results will be helpful for investors tracking betas during future the coronavirus waves.Design/methodology/approachA conditional capital asset pricing model (CAPM) and multivariate generalized autoregressive conditional heteroskedasticity (GARCH) model is used to estimate time-varying daily betas of the 50 largest Indian stocks spread across 16 industries over five years (Nov 2017 to May 2021), including the two waves of COVID-19 in India.FindingsThe results show that the betas increased during the COVID wave-1 (2020) but not during COVID wave-2 (2021). Moreover, the increase is more pronounced for consumer goods, infrastructure, insurance and information technology, unlike energy (oil and gas, power and mining) industries. Further, there are positive abnormal residual returns during the COVID waves. The results will be helpful for investors tracking betas during future COVID-19 waves.Originality/valueThis is perhaps the first paper to study the firm betas in light of the COVID-19 pandemic.

Analysis of financial contagion in influential African stock markets

Drawing from the experience of the global financial crisis that sprang forth from the US stock market, an empirical assessment of the dynamic correlation analysis of financial contagion with evidence from (5) African countries (South African, Nigeria, Egypt, Kenya, Tunisia) is presented. Monthly stock prices indices from 2004 to 2018 was analyzed using the dynamic conditional correlation multivariate GARCH model to ascertain the contagious effect of the US to the selected African markets. By analyzing the correlation coefficient series, three phases of the crisis periods were identified {pre-crisis (2004–2007); crisis (2007–2009) and post-crisis (2009–2018), respectively}. The study revealed that a significant relationship exists between the returns of the US market and the African markets. The inspection of the pre-crisis, crisis, post-crisis mean and variance estimation shows that the crisis period is characterized by substantial increases in volatility, establishing that the shock experienced in the US posed a threat to the African markets being examined. Further, evidence revealed that in the crisis period, an increase in correlation (contagion) existed, while a continued correlation (herding) existed in the post-crisis period.

Did Brexit change asset co-movements?

Purpose This paper aims to examine the impact of the Brexit referendum on the risk structure of financial asset prices. Co-movements are analysed using daily price returns of major stock and bond indices as well as commodities and exchange rates from June 2014 to June 2018. The authors used a multivariate GARCH model to study the dynamics of the conditional correlation matrix of asset returns. It was found that the conditional variances and correlations of assets spike on and after the Brexit referendum and then quickly revert to normal levels, suggesting that the effect of the referendum was transient rather than structural. The findings are of interest to investors as co-movements of financial assets can significantly impact financial portfolios and hedging strategies. Design/methodology/approach The authors used a multivariate GARCH model to study the dynamics of the conditional correlation matrix of asset returns. Findings It was found that the conditional variances and correlations of assets spike on and after the Brexit referendum and then quickly revert to normal levels, suggesting that the effect of the referendum was transient rather than structural. Research limitations/implications The findings are of interest to investors as co-movements of financial assets can significantly impact financial portfolios and hedging strategies. Originality/value To the best of the authors’ knowledge, research studying the underlying asset co-movements around Brexit does not exist.

Sectors stock indices aggregate correlations and expectations: Evidence from the Greek stock market

Based on the cyclical movements of the Athens Stock Market, the paper empirically examines the behavior of seven sectors (markets) namely: industry-services, emporium, construction, petroleum, telecommunications, food-beverages, and banks. Specifically using daily observations from January 2006 to August 2017, we estimate a dynamic equicorrelation multivariate GARCH model (DECO-MGARCH) developed by Engle and Kelly (2012), to analyze the dynamic behavior of these sectors. Furthermore, using time-dependent entropic measures we examine empirically the uncertainty (expectations) regarding the correlation behavior of these seven sectors. The empirical results are in line with previous findings (Tsai & Chen, 2010; Garnaut, 1998) and provide evidence supporting the view of high correlations during periods of crises. In addition, the dynamic entropy shows that the expectations of market participants were more concentrated (less spread out) during these periods of crises. Therefore, the empirical evidence of the paper supports the view that market participants share the same opinions (entropy exhibits low uncertainty) during crises and therefore are acting in a similar fashion (exhibiting high correlation).

Volatility spillovers and dynamic correlations among foreign exchange rates and bond markets of emerging economies

In this paper, we first examine how important historical shocks during and after the 2007-2008 global financial crisis affect the size and the persistence of volatilities among exchange rates and the ten-year bond rates of the Fragile Five countries (i.e., Brazil, India, Indonesia, South Africa and Turkey). We then investigate separately the dynamic interactions between exchange rates and the ten-year bond rates of the Fragile Five. We utilize a multivariate GARCH model (FIAPARCH-DCC model) and volatility impulse response functions to achieve these objectives. The results suggest that shocks? positive impacts on expected conditional variances of the variables are largely market-specific and different. Shocks have a more significant impact on bond markets than on foreign exchange markets. We also find that the dynamic conditional correlation series of bonds exhibit much lower correlations than those associated with exchange rate returns.