Hierarchical Archimedean Copulas Research Articles

Modeling dependence structures among international stock markets: Evidence from hierarchical Archimedean copulas

This study investigates dependence structures among international stock markets, including developed, emerging, and frontier markets, using the hierarchical Archimedean copula model. Empirical results indicate that emerging markets show the strongest dependence with European markets. Frontier markets show the weakest dependence with other market. After the global financial crisis, the lower dependence structure among the international stock markets has changed. Negative news have a larger impact on the degree of dependence than positive news. Contagion effect is observed in both the global financial crisis and the EU debt crisis.

Conditional Systemic Risk with Penalized Copula

Financial contagion and systemic risk measures are commonly derived from conditional quantiles by using imposed model assumptions such as a linear parametrization. In this paper, we provide model free measures for contagion and systemic risk which are independent of the specification of conditional quantiles and simple to interpret. The proposed systemic risk measure relies on the contagion measure, whose tail behavior is theoretically studied. To emphasize contagion from extreme events, conditional quantiles are specified via hierarchical Archimedean copula. The parameters and structure of this copula are simultaneously estimated by imposing a non-concave penalty on the structure. Asymptotic properties of this sparse estimator are derived and small sample properties illustrated using simulations. We apply the proposed framework to investigate the interconnectedness between American, European and Australasian stock market indices, providing new and interesting insights into the relationship between systemic risk and contagion. In particular, our findings suggest that the systemic risk contribution from contagion in tail areas is typically lower during times of financial turmoil, while it can be significantly higher during periods of low volatility.

MODELING DEPENDENCE BETWEEN LOSS TRIANGLES WITH HIERARCHICAL ARCHIMEDEAN COPULAS

AbstractOne of the most critical problems in property/casualty insurance is to determine an appropriate reserve for incurred but unpaid losses. These provisions generally comprise most of the liabilities of a non-life insurance company. The global provisions are often determined under an assumption of independence between the lines of business. Recently, Shi and Frees (2011) proposed to put dependence between lines of business with a copula that captures dependence between two cells of two different runoff triangles. In this paper, we propose to generalize this model in two steps. First, by using an idea proposed by Barnett and Zehnwirth (1998), we will suppose a dependence between all the observations that belong to the same calendar year (CY) for each line of business. Thereafter, we will then suppose another dependence structure that links the CYs of different lines of business. This model is done by using hierarchical Archimedean copulas. We show that the model provides more flexibility than existing models, and offers a better, more realistic and more intuitive interpretation of the dependence between the lines of business. For illustration, the model is applied to a dataset from a major US property-casualty insurer, where a bootstrap method is proposed to estimate the distribution of the reserve.

An approach to structure determination and estimation of hierarchical Archimedean Copulas and its application to Bayesian classification

Copulas are distribution functions with standard uniform univariate marginals. Copulas are widely used for studying dependence among continuously distributed random variables, with applications in finance and quantitative risk management; see, e.g., the pricing of collateralized debt obligations (Hofert and Scherer, Quantitative Finance, 11(5), 775---787, 2011). The ability to model complex dependence structures among variables has recently become increasingly popular in the realm of statistics, one example being data mining (e.g., cluster analysis, evolutionary algorithms or classification). The present work considers an estimator for both the structure and the parameters of hierarchical Archimedean copulas. Such copulas have recently become popular alternatives to the widely used Gaussian copulas. The proposed estimator is based on a pairwise inversion of Kendall's tau estimator recently considered in the literature but can be based on other estimators as well, such as likelihood-based. A simple algorithm implementing the proposed estimator is provided. Its performance is investigated in several experiments including a comparison to other available estimators. The results show that the proposed estimator can be a suitable alternative in the terms of goodness-of-fit and computational efficiency. Additionally, an application of the estimator to copula-based Bayesian classification is presented. A set of new Archimedean and hierarchical Archimedean copula-based Bayesian classifiers is compared with other commonly known classifiers in terms of accuracy on several well-known datasets. The results show that the hierarchical Archimedean copula-based Bayesian classifiers are, despite their limited applicability for high-dimensional data due to expensive time consumption, similar to highly-accurate classifiers like support vector machines or ensemble methods on low-dimensional data in terms of accuracy while keeping the produced models rather comprehensible.

Nonparametric estimation of the tree structure of a nested Archimedean copula

One of the features inherent in nested Archimedean copulas, also called hierarchical Archimedean copulas, is their rooted tree structure. A nonparametric, rank-based method to estimate this structure is presented. The idea is to represent the target structure as a set of trivariate structures, each of which can be estimated individually with ease. Indeed, for any three variables there are only four possible rooted tree structures and, based on a sample, a choice can be made by performing comparisons between the three bivariate margins of the empirical distribution of the three variables. The set of estimated trivariate structures can then be used to build an estimate of the target structure. The advantage of this estimation method is that it does not require any parametric assumptions concerning the generator functions at the nodes of the tree.

Properties of hierarchical Archimedean copulas

Abstract In this paper we analyse the properties of hierarchical Archimedean copulas. This class is a generalisation of the Archimedean copulas and allows for general non-exchangeable dependency structures. We show that the structure of the copula can be uniquely recovered from all bivariate margins. We derive the distribution of the copula values, which is particularly useful for tests and constructing confidence intervals. Furthermore, we analyse dependence orderings, multivariate dependence measures, and extreme value copulas. We pay special attention to the tail dependencies and derive several tail dependence indices for general hierarchical Archimedean copulas.

On the structure and estimation of hierarchical Archimedean copulas

In this paper we provide a method for estimating multivariate distributions defined through hierarchical Archimedean copulas. In general, the true structure of the hierarchy is unknown, but we develop a computationally efficient technique to determine it from the data. For this purpose we introduce a hierarchical estimation procedure for the parameters and provide an asymptotic analysis. We consider both parametric and nonparametric estimation of the marginal distributions. A simulation study and an empirical application show the effectiveness of the grouping procedure in the sense of structure selection.

Systemic Weather Risk and Crop Insurance: The Case of China

AbstractThis article explores the possibility of spatial diversification of weather risk for 17 agricultural production regions in China. We investigate the relation between the size of the buffer load and the size of the trading area of a hypothetical temperature‐based insurance. The analysis adopts the hierarchical Archimedean copula approach that allows for flexible modeling of the dependence structure of insured losses. We find that the spatial diversification effect depends on the type of the weather index and the strike level of the insurance. Our findings are relevant for the current discussion on the viability of private crop insurance in China.

Hierarchical structures in the aggregation of premium risk for insurance underwriting

In the valuation of the Solvency II capital requirement, the correct appraisal of risk dependencies acquires particular relevance. These dependencies refer to the recognition of risk diversification in the aggregation process and there are different levels of aggregation and hence different types of diversification. For instance, for a non-life company at the first level the risk components of each single line of business (e.g. premium, reserve, and CAT risks) need to be combined in the overall portfolio, the second level regards the aggregation of different kind of risks as, for example, market and underwriting risk, and finally various solo legal entities could be joined together in a group. Solvency II allows companies to capture these diversification effects in capital requirement assessment, but the identification of a proper methodology can represent a delicate issue. Indeed, while internal models by simulation approaches permit usually to obtain the portfolio multivariate distribution only in the independence case, generally the use of copula functions can consent to have the multivariate distribution under dependence assumptions too. However, the choice of the copula and the parameter estimation could be very problematic when only few data are available. So it could be useful to find a closed formula based on Internal Models independence results with the aim to obtain the capital requirement under dependence assumption. A simple technique, to measure the diversification effect in capital requirement assessment, is the formula, proposed by Solvency II quantitative impact studies, focused on the aggregation of capital charges, the latter equal to percentile minus average of total claims amount distribution of single line of business (LoB), using a linear correlation matrix. On the other hand, this formula produces the correct result only for a restricted class of distributions, while it may underestimate the diversification effect. In this paper we present an alternative method, based on the idea to adjust that formula with proper calibration factors (proposed by Sandström (2007)) and appropriately extended with the aim to consider very skewed distribution too. In the last part considering different non-life multi-line insurers, we compare the capital requirements obtained, for only premium risk, applying the aggregation formula to the results derived by elliptical copulas and hierarchical Archimedean copulas.

Constructing hierarchical Archimedean copulas with Lévy subordinators

A probabilistic interpretation for hierarchical Archimedean copulas based on Lévy subordinators is given. Independent exponential random variables are divided by group-specific Lévy subordinators which are evaluated at a common random time. The resulting random vector has a hierarchical Archimedean survival copula. This approach suggests an efficient sampling algorithm and allows one to easily construct several new parametric families of hierarchical Archimedean copulas.

Modeling the Dependence Structure between Bonds and Stocks: A Multivariate Copula Approach

We apply a copula-GARCH approach to modeling the joint distribution of excess returns of four major assets: one year and ten year Treasury bonds and S&P 500 and Nasdaq indices. Three approaches are attempted in building the multivariate copula for the dependence structure of multiple assets: (1) n-dimensional normal copula and n-dimensional Student's t copula, (2) hierarchical Archimedean copula and (3) mixed copula. To ensure better convergence in estimation, we adopt a two-stage estimation procedure where marginal distributions are estimated separately in the first stage and then the copula is estimated in the second. Our results show that normal copula and Student's t copula yield higher log-likelihoods while hierarchical Archimedean copula and mixed copula yield lower log-likelihoods. We plot overall and rolling correlation, exceedance correlation and tail dependence of the data and find significant time-varying features with correlations between stocks and bonds, and with asymmetric exceedance correlation among all assets. With a time-varying dependence structure, our Student's t copula yields the highest log-likelihood and generates the most consistent dependence patterns with the data.