Testing the hypothesis that international equity market correlation increases in volatile times is a difficult exercise and misleading results have often been reported in the past because of a spurious relationship between correlation and volatility. Using "extreme value theory" to model the multivariate distribution tails, we derive the distribution of extreme correlation for a wide class of return distributions. Empirically, we reject the null hypothesis of multivariate normality for the negative tail, but not for the positive tail. We also find that correlation is not related to market volatility per se but to the market trend. Correlation increases in bear markets, but not in bull markets.

We present a set of stylized empirical facts emerging from the statistical analysis of price variations in various types of financial markets. We first discuss some general issues common to all statistical studies of financial time series. Various statistical properties of asset returns are then described: distributional properties, tail properties and extreme fluctuations, pathwise regularity, linear and nonlinear dependence of returns in time and across stocks. Our description emphasizes properties common to a wide variety of markets and instruments. We then show how these statistical properties invalidate many of the common statistical approaches used to study financial data sets and examine some of the statistical problems encountered in each case.

This paper inv estigates the potential for extreme co-mov ements between financial assets by directly testing the underlying dependence structure. In particular, a t-dependence structure, deriv ed from the Student t distribution, is used as a proxy to test for this extremal behav#a(0 Tests in three di#erent markets (equities, currencies, and commodities) indicate that extreme co-mov ements are statistically significant. Moreov er, the "correlation-based" Gaussian dependence structure, underlying the multiv ariate Normal distribution, is rejected with negligible error probability when tested against the t-dependencealternativ e. The economic significance of these results is illustratedv ia three examples: co-mov ements across the G5 equity markets; portfoliov alue-at-risk calculations; and, pricing creditderiv ativ es. JEL Classification: C12, C15, C52, G11. Keywords: asset returns, extreme co-mov ements, copulas, dependence modeling, hypothesis testing, pseudo-likelihood, portfolio models, risk management. # The authorsw ould like to thankAndrew Ang, Mark Broadie, Loran Chollete, and Paul Glasserman for their helpful comments on an earlier version of this manuscript. Both authors arewS; the Columbia Graduate School of Business, e-mail: {rm586,assaf.zeevi}@columbia.edu, current version available at www.columbia.edu\# rm586 1 Introducti7 Specification and identification of dependencies between financial assets is a key ingredient in almost all financial applications: portfolio management, risk assessment, pricing, and hedging, to name but a few. The seminal work of Markowitz (1959) and the early introduction of the Gaussian modeling paradigm, in particular dynamic Brownian-based models, hav e both contributed greatly to making the concept of co rrelatio almost synony...

Economic problems such as large claims analysis in insurance and value-at-risk in finance, require assessment of the probability P of extreme realizations Q: This paper provides a semi-parametric method for estimation of extreme #P;Q# combinations for data with heavy tails. We solve the long standing problem of estimating the sample threshold of where the tail of the distribution starts. This is accomplished by the combination of a control variate type device and a subsample bootstrap technique. The subsample bootstrap attains convergence in probability, whereas the full sample bootstrap would only provide convergence in distribution. This permits a complete and comprehensive treatment of extreme #P;Q# estimation. Keywords: Extreme value theory, tail estimation, risk analysis # Corresponding author: C. G. de Vries, Tinbergen Institute, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands, e-mail cdevries@few.eur.nl. Danielsson's e-mail is jond@hi.is. This and related papers can be downl...

We propose a method of optimization of asset allocation in the case where the stock price variations are supposed to have “fat ” tails represented by power laws. Generalizing over previous works using stable Lévy distributions, we distinguish three distinct components of risk described by three different parts of the distributions of price variations: unexpected gains (to be kept), harmless noise inherent to financial activity, and unpleasant losses, which is the only component one would like to minimize. The independent treatment of the tails of distributions for positive and negative variations and the generalization to large events of the notion of covariance of two random variables provide explicit formulae for the optimal portfolio. The use of the probability of loss (or equivalently the Value-at-Risk), as the key quantity to study and minimize, provides a simple solution to the problem of optimization of asset allocations in the general case where the characteristic exponents are different for each asset.

We examine the dynamics of extreme values of overnight borrowing rates in an inter-bank money market before a financial crisis during which overnight borrowing rates rocketed up to (simple annual) 4000 percent. It is shown that the generalized Pareto distribution fits well to the extreme values of the interest rate distribution. We also provide predictions of extreme overnight borrowing rates before the crisis. The examination of tails (extreme values) provides answers to such issues as what are the extreme movements expected in financial markets; have we already seen the largest moves; is there a possibility for even larger movements and, are there theoretical processes that can model the type of fat tails in the observed data? The answers to such questions are essential for proper management of financial exposures and laying ground for regulations.

We present a set of stylized empirical facts emerging from the statistical analysis of price variations in various types of financial markets. We first present data sources and discuss the choice of a time scale when constructing financial time series. Various statistical properties of asset returns are then described: distributional properties, tail analysis and extreme fluctuations, linear and non-linear dependence of returns in time and across stocks. Our description emphasizes properties common to a wide variety of markets and instruments. The last part deals with interest rates: we present some issues encountered in constructing yield curves from empirical data and discuss the statistical properties of the term structure fluctuations.

This paper proposes a methodology to provide risk measures for portfolios during extreme events. The approach is based on splitting the multivariate extreme value distribution of the assets of the portfolio into two parts: the distributions of each asset and their dependence function. The estimation problem is also investigated. Then, stress-testing is applied for market indices portfolios and Monte-Carlo based risk measures – Value-at-Risk and Expected Shortfall – are provided.

The asymptotic normality of a class of estimators for extreme quantiles is established under mild structural conditions on the observed stationary β–mixing time series. Consistent estimators of the asymptotic variance are introduced, which render possible the construction of asymptotic confidence intervals for the extreme quantiles. Moreover, it is shown that many well-known time series models satisfy our conditions. Then the theory is applied to a time series of returns of a stock index. Finally, the finite sample behavior of the proposed confidence intervals is examined in a simulation study. It turns out that for most time series models under consideration the actual coverage probability is pretty close to the nominal level if the sample fraction used for estimation is chosen appropriately.

Many fields of modern science and engineering have to deal with events which are rare but have significant consequences. Extreme value theory is considered to provide the basis for the statistical modelling of such extremes. The potential of extreme value theory applied to financial problems has only been recognized recently. This paper aims at introducing the fundamentals of extreme value theory as well as practical aspects for estimating and assessing statistical models for tail-related risk measures.