Value-at-Risk models and Basel capital charges (original) (raw)
Related papers
Economic Research-Ekonomska Istraživanja, 2015
The aim of this paper is to investigate the performance of Value at Risk (VaR) models in selected Central and Eastern European (CEE) emerging capital markets. Daily returns of Croatian (CROBEX), Czech (PX50), Hungarian (BUX) and Romanian (BET) stock exchange indices are analysed for the period January, 2000 -February, 2012, while daily returns of the Serbian (BELEX15) index is examined for the period September, 2005 -February, 2012. In recent years there has been much research conducted into VaR in developed markets, while papers dealing with VaR calculation in CEE are rare. Furthermore, VaR models created and suited for liquid and welldeveloped markets that assume normal distribution are less reliable for capital markets in emerging economies, such as Central and Eastern European Union member and candidate states. Since capital markets in European emerging economies are highly volatile, less liquid and strongly dependent on the unexpected external shocks, market risk estimation based on normality assumption in CEE countries is more problematic. This motivates us to implement GARCH-type methods that involve time varying volatility and heavy tails of the empirical distribution of returns. We test the hypothesis that using the assumption of heavy tailed distribution it is possible to forecast market risk more precisely, especially in times of crisis, than under the assumption of normal distribution or using historical simulations method. Our backtesting results for the last 500 observations are based on the Kupiec POF and Christoffersen independence test. They show that GARCH-type models with t error distribution in most analysed cases give better VaR estimation than GARCH type models with normal errors in the case of a 99% confidence level, while in the case of a 95% confidence level it is the opposite. The results of backtesting analysis for the crisis period (after the collapse of Lehman Brothers) show that GARCH-type models with t-distribution of residuals provide better VaR estimates compared with GARCH-type models with normal distribution, historical simulations and RiskMetrics methods. The RiskMetrics method in the most cases underestimates market risk.
International Business & Economics Research Journal (IBER), 2014
This paper uses closing prices of the BRICS (Brazil, Russia, India, China, and South Africa) financial markets to implement a risk model that generates point estimates of both Value at Risk (VaR); and Expected Shortfall (ES). The risk model is thereafter backtested using three techniques namely the Basel II green zone, the unconditional test, and the conditional test. We first filter the log-return data using an Autoregressive Regression model (AR) of order one for the conditional mean and an Exponential Generalised Autoregressive Conditional Heteroscedasticity of order one (EGARCH 1,1) for the conditional variance. We thereafter fit the filtered returns by using the Generalised Pareto Distribution (GPD) model before we compute both VaR and ES estimates. We find that the use of the GPD is well suited to financial markets that are highly exposed to global financial risks. Our results show that both VaR and ES estimates for South Africa are very low when compared with those of other B...
Extreme Value Theory as a financial risk measure of the South African stock market
North West University, 2020
The incidence of rare but extreme events appears to be greater in worldwide nancial markets. This implies the need for good risk modelling systems that can envisage the likelihood of risky events in daily market fundamentals, which can help in assessing the likelihood of extreme events such as the 2007-2008 global nancial crisis. A conspicuous candidate theory when dealing with excessive events is extreme value theory (EVT). EVT naturally became the only statistical modelling approach that pledges rm models that quantify extreme risk measures such as value at risk (VaR), expected shortfall (ES) and other related risk measures. In order to overcome issues that are related with reliance prompted by volatility clustering in nancial markets, this study develops a novel approach by applying EVT models to model extreme losses of the ve South African nancial time series exchange/Johannesburg Stock Exchange (FTSE/JSE) closing banking indices, and explores the e ectiveness of risk measures for measuring risk of investment. Fiveday time series for the period of 02 January 2008 to 20 April 2018 is used and this consists of 2575 observations for each bank. The unsteadiness of this stock market invigorated an interest in assessing the underlying three models; asymmetry regime switching generalised autoregressive conditional heteroscedasticity (GARCH). To be speci c, Markov-switching threshold generalised autoregressive conditional heteroscedasticity (MS-TGARCH), Markov-switching exponential generalised autoregressive conditional heteroscedasticity (MS-EGARCH) and Markov-switching GJR generalised autoregressive conditional heteroscedasticity (MS-GJR-GARCH) models, tted using a skewed student-t distribution via maximum likelihood estimation (MLE) method. To capture extreme quantiles and estimate return levels, the generalised extreme value (GEV) with block minima method (BMM) and the generalised Pareto distribution (GPD) with peaks over threshold (POT) are also tted using MLE algorithm. In addition, the hybrid models for the four risk measures are estimated in order to express the tail risk related to extreme quantiles and return levels. The results of regime switching GARCH models suggest that all the three-asymmetry models provide good estimates of volatility clustering, with MS (2)-EGARCH (1,1) outperforming all the models because this model recorded a frequency of one ve times more than MS (2)-TGARCH (1,1) and MS (2)-GJR-GARCH (1,1). Out of the seven statistical loss functions used, this model gives a predictive accuracy of 71.45%. Moreover, results obtained by GEV and GPD showed positive shape parameters indicating a Frechet type distribution with GPD giving a type II Frechet distribution that are appropriate for the data. While estimating extreme quantiles, the 95% and 99% quantiles for GPD model did not change signi cantly, and the same was observed in the GEV distribution. Therefore, the two distributions performed similarly at both intervals. However, GEV shows a better performance as compared to the GPD, since the bias estimation is less for the GEV contrasted to the GPD, giving 6.48% less bias of the GEV than the GPD. In addition, once in 3 years, a daily loss of approximately 7.022% would be observed across the ve banks while using the GEV model. But with the GPD model only 0.9604% would be observed across all the banks. This made a GPD model a better model than a GEV in estimating extreme loses. Finally, the results of risk measures indicated that the Glue VaR risk measure has less risk estimates as contrasted with the other three measures from GEV model and GPD model. To take into account market liquidity constraints and Basel regulations, 5-day risk horizons in addition to the more typical 1-day horizon were being considered. This implies that the computation of economic capital using Glue VaR risk is more conservative than using other risk measures under the GEV model. Therefore, the conclusion that can be made is that GEV and GPD estimates of Glue VaR risk under di erent con dence e levels exhibit analogous characteristics as observed from VaR, ES and conditional tail expectation (CTE). This study extends to stationary features of EVT models. Extension of these models in the literature is quite complicated since it requires speci cation not only on how the usual threshold and MLE parameters change over time, but also those with the bulk distribution component of the models. The study also shows, in the risk management of nancial capital or portfolios, that evaluating the probability of rare extreme events is an important question. This means that EVT provides the robust basis necessary for the statistical modelling of such events and for the computation of extreme risk measures. Furthermore, a base for future researchers for conducting studies on emerging markets, more speci cally in the South African context has also been contributed
GFC-robust risk management under the Basel Accord using extreme value methodologies
Mathematics and Computers in Simulation, 2013
In , a robust risk management strategy to the Global Financial Crisis (GFC) was proposed under the Basel II Accord by selecting a Value-at-Risk (VaR) forecast that combines the forecasts of different VaR models. The robust forecast was based on the median of the point VaR forecasts of a set of conditional volatility models. In this paper we provide further evidence on the suitability of the median as a GFC-robust strategy by using an additional set of new extreme value forecasting models and by extending the sample period for comparison. These extreme value models include DPOT and Conditional EVT. Such models might be expected to be useful in explaining financial data, especially in the presence of extreme shocks that arise during a GFC. Our empirical results confirm that the median remains GFC-robust even in the presence of these new extreme value models. This is illustrated by using the S&P500 index before, during and after the 2008-09 GFC. We investigate the performance of a variety of single and combined VaR forecasts in terms of daily capital requirements and violation penalties under the Basel II Accord, as well as other criteria, including several tests for independence of the violations. The strategy based on the median, or more generally, on combined forecasts of single models, is straightforward to incorporate into existing computer software packages that are used by banks and other financial institutions.
Frontiers in Finance and Economics, 2011
Conventional Value-at-risk (VaR) models tend to underestimate stock market losses, as they assume normality and fail to capture the frequency and severity of extreme fluctuations, Extreme value theory (EVT) overcomes this limitation by providing a framework in which to analyze the extreme behavior of stock-markets returns and by quantifying possible losses during financial turbulences. This study uses the c-quantile of a fat-tailed distribution for VaR analysis. An innovation in the present work is the application of EVT not only to the left tail of the returns distribution but also to its right tail, while assessing long and short positions. A generalized extreme value distribution (GEVD) is used to analyze the two largest stock markets from Latin America, Brazil and Mexico; a conditional VaR (CVaR) model is applied to determine risk exposure from investing in those markets, with daily index data for the period 1970-2004. The results confirm the presence of fat tails in both markets as a result of the excess of kurtosis; the empirical evidence shows that VaR and CVaR based on EVT yield more precise and robust information about financial risk than conventional parametric estimations.
Evaluation of Value at Risk in Emerging Markets
International Journal of Financial Management, 2017
Financial institutions have witnessed numerous episodes of financial crises all over the world during the last four decades. The researchers, academicians and policy makers in the field of finance studied these episodes extensively and to mitigate the risk involved in these crises have proposed several measures in the financial literature, but Value at Risk (VaR) has emerged as a more popular risk measurement technique. Although a number of studies have been undertaken in this area of research for developed markets but very few studies have been conducted in developing and emerging market economies. This study makes an attempt to evaluate the performance of VaR in emerging markets namely Brazil, Russia, India and China by considering Historical, Monte Carlo and GARCH Simulations to calculate VaR for the period 1998 to 2015. The study found that GJRGARCH Simulation is more suitable for Brazil and China while Historical Simulation for Russian and Indian Stock Markets based on the backtesting experiment..in market behaviour, it becomes vital to measure the level of risk for potential investors and agents even after knowing its presence, in order to survive in the global competitive market in a dynamic manner. Unlike the matured financial markets, the emerging financial markets are characterized with insufficient liquidity, the small scale of trading and asymmetrical and low number of trading days with certain securities (Andjelić, Djaković and Radišić, 2010). In the recent times, the emerging markets have been playing a crucial role due to greater potential in terms of economic growth and investment opportunities. However, the emerging stock markets are relatively young markets and have not developed sufficiently so as to identify all information that affects the stock prices and therefore, do not respond quickly to the publicly disclosed information (Benaković and Posedel, 2010). After the financial instabilities during 70's and advent of derivative markets, floating exchange rates led to development of several risk measurement methods. Among these Value-at-Risk (VaR) has emerged as a popular measure for assessing the market risk of the portfolio among the trading community. It can be defined as the maximum potential loss of a specific portfolio for a given time horizon. Increasing availability of the financial data and rapid advances in computer technology led to the development of various VaR models that can be applied for the risk management profession. The application of VaR models and comparing their relative performance gained a momentum in the field of financial economics. However, there is no common model that can give best forecasts of these models see for example,
2007
Stock markets, particularly those from the developing countries, are characterized by high volatility which conventional models fail to capture fully, potentially leading to high losses. Value at Risk (VaR) models signified an important step to estimate losses of financial assets and portfolios. However, the stylized fact that financial returns exhibit fat tails, implies that conventional VaR models (parametrics and non-parametrics models) show important limitations because they fail to take into account the right statistical distributions to capture the frequency and severity of extreme values; the normal distribution is insufficient for this purpose. Extreme Value Theory (EVT) overcomes this limitation because it provides a framework to formally study the extreme behavior of stock markets returns and quantifies the possible losses experienced during financial instabilities and turbulences without making any assumptions about the underlying distribution of returns. This study uses ...
Value-at-risk in times of crisis: An analysis in the Brazilian market
African Journal of Business Management, 2015
The present study aimed at evaluating the predictive ability of the models of market risk estimation in times of financial crises. To this end, models were tested to estimate the financial indicator Value-at-Risk (VaR) applied to the daily returns of the BM&FBovespa, the Ibovespa index. Traditional models and those based on the Extreme Value Theory (EVT), considered as two types of distribution, the Generalized Extreme Value (GEV) and generalized Pareto distribution (GPD) were tested. The data relating to two periods of international financial crises termed the 1997 Asian Financial Crisis and the U.S. Subprime Meltdown in 2008 were explored in the study. The results indicated the inefficiency of most statistical models for VaR estimation in moments of high volatility for both periods of crisis. In contrast, the exception refers to the model based on EVT, GPD distribution that proved satisfactory in the estimates in both periods of crisis. The results are in agreement with other studies in the field.