Developing a stress testing framework based on market risk models

Annals of Financial Economics, 2018

In this paper, we address how to evaluate tail risk forecasts for systemic risk (SRISK) measurement. We propose two loss functions, the Tail Tick Loss and the Tail Mean Square Error, to evaluate, respectively, Conditional Value-at-Risk (CoVaR) and MES forecasts. We then analyse CoVaR and MES forecasts for a panel of top US financial institutions between 2000 and 2012 constructed using a set of bivariate DCC-GARCH-type models. The empirical results highlight the importance of using an appropriate loss function for the evaluation of such forecasts. Among other findings, the analysis confirms that the DCC-GJR specification provides accurate predictions for both CoVaR and MES, in particular for the riskiest group of institutions in the panel (Broker-Dealers).

Journal of Emerging Economies and Islamic Research

The purpose of this paper is to estimate the market risk exposure within the Tiger Cub Economies regions in calm and stormy market conditions. The secondary objective of the empirical research is to determine the reliability and accuracy of the market risk model used by most banking sectors within the region as the primary tool for mitigating potential systematic risk. The precision of the market risk model was assessed using the 250-day trading data of major indices from five emerging ASEAN countries or known as the Tiger Cub Economies stretching from January 2018 until December 2021. It consists of two sub-samples which are known as before and during the pandemic COVID-19. The current study contributes to the existing literature on the ability of the VaR-HS model in estimating accurate market risk exposure in light of the recent pandemic COVID19 within the Tiger Cub Economies region. Interestingly, it is also evident that inaccurate VaR-HS tend to overestimate the risk and VaR-GAR...

CBU International Conference Proceedings 2013 - Integration and Innovation in Science and Education, 2013

The purpose of this paper is to evaluate performance of value-at-risk (VaR) produced by two risk models: historical simulation and Risk Metrics. We perform three backtest: unconditional coverage, independence and conditional coverage. We present results on both VaR 1% and VaR 5% on a one-day horizon for the following indices: S&P 500, DAX, SAX, PX and Belex 15. Our results show that Historical simulation 500 days rolling window approach satisfies unconditional coverage for all tested indices, while Risk Metrics has many rejection cases. On the other hand Risk Metrics model satisfies independence backtest for three indices, while Historical simulation has rejected more times. Based on our strong criteria to accept accuracy of VaR models only if both unconditional coverage and independence properties are satisfied, results indicate that during the crisis period all tested VaR models underestimate the true level of market risk exposure.

SSRN Electronic Journal

This paper proposes an original approach for backtesting systemic risk measures. This backtesting approach makes it possible to assess the systemic risk measure forecasts used to identify the financial institutions that contribute the most to the overall risk in the financial system. Our procedure is based on simple tests similar to those generally used to backtest the standard market risk measures such as value-at-risk or expected shortfall. We introduce a concept of violation associated with the marginal expected shortfall (MES), and we define unconditional coverage and independence tests for these violations. We can generalize these tests to any MES-based systemic risk measures such as SES, SRISK, or ∆CoVaR. We study their asymptotic properties in the presence of estimation risk and investigate their finite sample performance via Monte Carlo simulations. An empirical application to a panel of U.S. financial institutions is conducted to assess the validity of MES, SRISK, and ∆CoVaR forecasts issued from a GARCH-DCC model. Our results show that this model provides valid forecasts for MES and SRISK when considering a medium-term horizon. Finally, we propose an early warning system indicator for future systemic crises deduced from these backtests. Our indicator quantifies how much is the measurement error issued by a systemic risk forecast at a given point in time which can serve for the early detection of global market reversals.

Economics Bulletin, 2020

The importance of extreme events in finance is increasingly recognized but the existing methods for estimating the tail risk are not very satisfactory. The Hill estimator which is the most commonly used in the literature, suffers severe deficiencies. As an alternative, we propose to estimate the relative tail risk and not the absolute tail risk. This allows us to revisit a larger pool of tail index estimators, including stable law estimators, which have been discarded because of their bias when estimating the absolute tail risk. From a database of 74 international stock market indices, and using both simulated and empirical data, we find that the McCulloch estimator gives an estimate of the relative tail risk up to 50% more precise than the Hill estimator. These results are confirmed by the White (2000) and Hansen (2005) superior predictive ability tests.

SSRN Electronic Journal, 2015

Conditional value at risk (CoVaR) and marginal expected shortfall (MES) have been proposed as stock return based measures of the systemic risk created by individual financial institutions even though the literature provides no formal hypothesis test for detecting systemic risk. We address this shortcoming by constructing hypothesis test statistics for CoVaR and MES that can be used to detect systemic risk at the institution level. We apply our tests to daily stock returns data for over 3500 firms during 2006-2007. CoVaR (MES) tests identify almost 500 (1000) firms as systemically important. Both tests identify many more real-side firms than financial firms, and they often disagree about which firms are systemic. Analysis of the hypothesis tests' performance for plausible alternative hypotheses finds that return skewness can cause test rejections and, even when systemic risk imparts a strong signal in stock return distributions, hypothesis tests based on CoVaR and MES may fail to detect it. Our overall conclusion is that CoVaR and MES are not reliable measures of systemic risk.

2023

Since their emergence in the late 1980s, the Value at Risk models have quickly become the global standard in risk management and forecasting. However, their role during global financial crises has raised the question about the effectiveness of such tools in managing risk during extreme events while some researchers even attribute global financial contagion to Value at Risk practices. We propose a new measure called Value at Risk elasticity to measure the responsiveness of a portfolio's Value at Risk to a percentage change in the weight of each investment asset. Using BIS data on consolidated, international claims, we test our new measure with five historical financial crises. Specifically, we estimate Value at Risk elasticities for the portfolios of major banking centers prior to each crisis as an indicator for an increased risk factor in assets and contrast our predictions of high-risk assets with the actual losses in the aftermath of the crises using stock market data. The results show that our proposed measure performs well in the prediction of asset losses during several global crises. Based on our findings, we recommend including Value at Risk elasticity as an index to better manage risks in portfolios.

SSRN Electronic Journal, 2014

An emerging literature proposes using conditional value at risk (CoVaR) and marginal expected shortfall (MES) to measure financial institution systemic risk. We identify two weaknesses in this literature: (1) it lacks formal statistical hypothesis tests; and, (2) it confounds systemic and systematic risk. We address these weaknesses by introducing a null hypothesis that stock returns are normally distributed. This allows us to separate systemic from systematic risk and construct hypothesis tests for the presence of systemic risk. We calculate the sampling distribution of these new test statistics and apply our tests to daily stock returns data over the period 2006-2007. The null hypothesis is rejected in many instances, consistent with tail dependence and systemic risk but the CoVaR and MES tests often disagree about which firms are potentially "systemic." The highly restrictive nature of the null hypothesis and the wide range of firms identified as systemic makes us reluctant to interpret rejections as clear evidence of systemic risk. The introduction of hypothesis testing is our primary contribution, and the results highlight the importance of generalizing the approach to less restrictive stock return processes and to other systemic risk measures derived from return data.

This paper develops a fi nancial stress index for the United States, the Cleveland Financial Stress Index (CFSI), which provides a continuous signal of fi nancial stress and broad coverage of the areas that could indicate it. The index is based on daily public-market data collected from four sectors of the fi nancial markets-the credit, foreign exchange, equity, and interbank markets. A dynamic weighting method is employed to capture changes in the relative importance of these four sectors as they occur. In addition, the design of the index allows the origin of the stress to be identifi ed. We compare the CFSI to alternative indexes, using a detailed benchmarking methodology, and show how the CFSI can be applied to systemic stress monitoring and early warning system design. To that end, we investigate alternative stress-signaling thresholds and frequency regimes and then establish optimal frequencies for fi ltering out market noise and idiosyncratic episodes. Finally, we quantify a powerful CFSI-based rating system that assigns a probability of systemic stress to ranges of CFSI outcomes.