Revisiting the finite mixture of Gaussian distributions with application to futures markets

We present a new estimation method for Gaussian mixture modeling, namely, the kurtosis‐controlled expectation‐maximization (EM) algorithm, which overcomes the limitations of the usual estimation techniques via kurtosis control and kernel splitting. Our simulation study shows that the dynamic allocation of kernels according to the value of the total kurtosis measure makes the proposed kurtosis‐controlled EM algorithm an efficient method for Gaussian mixture density estimation. This algorithm yielded considerable improvements over the classical EM algorithm. We then used the discrete Gaussian mixture framework to account for the observed thick‐tailed distributions of futures returns and applied the kurtosis‐controlled EM algorithm to estimate the distributions of real (agricultural, metal, and energy) and financial (stock index and currency) futures returns. We proved that this framework is perfectly adapted to capturing the departures from normality of the observed return distributions. Unlike in previous studies, we found that a two‐component Gaussian mixture is too poor a model to accurately capture the distributional properties of returns. Similar results have been obtained for stocks, indexes, currencies, interest rates, and commodities. This has important implications in many financial studies using Gaussian mixtures to incorporate the thickness of the tails of the distributions in the computation of the value at risk or to infer implied risk‐neutral densities from option prices, to name but a few. © 2001 John Wiley & Sons, Inc. Jrl Fut Mark 21: 347–376, 2001     

Time variation in the tail behavior of Bund future returns

The literature on the tail behavior of asset prices focuses mainly on the foreign exchange and stock markets, with only a few articles dealing with bonds or bond futures. The present article addresses this omission. It focuses on three questions using extreme value analysis: (a) Does the distribution of Bund future returns have heavy tails? (b) Do the tails change over time? (c) Does the tail index provide information that is not captured by a standard VaR approach? The results are as follows: (a) The distribution of high‐frequency returns of the Bund future is indeed characterized by heavy tails. The tails are thinner for lower frequencies, but remain significantly heavy even for daily data. (b) There are statistically significant breaks in the tails of the return distribution. (c) The likelihood of extreme price movements suggested by extreme value theory differs from that obtained by standard risk measures. This suggests that the tail index does indeed provide information not contained in volatility measures. © 2004 Wiley Periodicals, Inc. Jrl Fut Mark 24:387–398, 2004     

NEWS-GENERATED DEPENDENCE AND OPTIMAL PORTFOLIOS FOR n STOCKS IN A MARKET OF BARNDORFF-NIELSEN AND SHEPHARD TYPE

We consider Merton's portfolio optimization problem in a Black and Scholes market with non-Gaussian stochastic volatility of Ornstein–Uhlenbeck type. The investor can trade in n stocks and a risk-free bond. We assume that the dependence between stocks lies in that they partly share the Ornstein–Uhlenbeck processes of the volatility. We refer to these as news processes, and interpret this as that dependence between stocks lies solely in their reactions to the same news. The model is primarily intended for assets that are dependent, but not too dependent, such as stocks from different branches of industry. We show that this dependence generates covariance, and give statistical methods for both the fitting and verification of the model to data. Using dynamic programming, we derive and verify explicit trading strategies and Feynman–Kac representations of the value function for power utility.     

分形分布在股票市场的应用研究

本文认为,对金融资产尖峰厚尾特性的描述,分形分布比正态分布更合适。我国股票市场收益率分布的特征指数不是正态分布,具有明显的尖峰态,我国股票市场具有较大的波动性;上海和深圳的两个股票市场收益率呈负偏,股票收益率分布右偏,呈现右厚尾特征;当置信水平为99%时,用分形分布拟合经验分布得到的风险系数高估了风险,用正态分布低估了风险,且分形分布的绝对偏差大于正态分布的绝对偏差,而当置信水平为90%时,分形分布对经验分布的风险系数拟合得非常好,正态分布对经验分布高估了风险,且绝对偏差比较大。     

THE GARCH OPTION PRICING MODEL

This article develops an option pricing model and its corresponding delta formula in the context of the generalized autoregressive conditional heteroskedastic (GARCH) asset return process. the development utilizes the locally risk-neutral valuation relationship (LRNVR). the LRNVR is shown to hold under certain combinations of preference and distribution assumptions. the GARCH option pricing model is capable of reflecting the changes in the conditional volatility of the underlying asset in a parsimonious manner. Numerical analyses suggest that the GARCH model may be able to explain some well-documented systematic biases associated with the Black-Scholes model.     

The characteristic function of rough Heston models

It has been recently shown that rough volatility models, where the volatility is driven by a fractional Brownian motion with small Hurst parameter, provide very relevant dynamics in order to reproduce the behavior of both historical and implied volatilities. However, due to the non‐Markovian nature of the fractional Brownian motion, they raise new issues when it comes to derivatives pricing. Using an original link between nearly unstable Hawkes processes and fractional volatility models, we compute the characteristic function of the log‐price in rough Heston models. In the classical Heston model, the characteristic function is expressed in terms of the solution of a Riccati equation. Here, we show that rough Heston models exhibit quite a similar structure, the Riccati equation being replaced by a fractional Riccati equation.     

Dependence between the global gold market and emerging stock markets (E7+1): Evidence from Granger causality using quantile and quantile‐on‐quantile regression methods

This study examined the dependence between gold and stocks during 2002–18 in seven emerging countries. The study combined the bivariate cross‐quantilogram introduced recently with quantile‐on‐quantile regression (QQR) approaches to conduct comprehensive and complementary analyses. The QQR results for the full sample revealed a weak positive dependence in all the quantiles of gold and stock returns across all the countries selected during mild market conditions. The results for pre and post‐crisis periods largely were consistent with those obtained for the full sample, except for Turkey (pre‐crisis), and China and Indonesia (post‐crisis). The results of the causality test‐in‐mean (return) and that of the causality test‐in‐variance revealed no causal relation between stock and gold in the pre‐crisis period, while causality ran only from gold to some stocks in the post‐crisis period. Further, while there was volatility causality running only from gold to stocks during the pre‐crisis period, the volatility causality between the two markets was very high during the post‐crisis period. Therefore, we suggest that gold may have been a hedge for stocks during the pre‐crisis compared to the post‐crisis period. Further, international risk factors should be considered in optimal investment decisions between domestic and global markets' assets (stocks and gold).     

Efficiency of the IBEX spot–futures basis: The impact of the mini‐futures

This paper considers whether the introduction of the mini‐futures contract for the Spanish Ibex index affects overall market efficiency. Using linear, non‐linear, and fractional integration modeling techniques for the basis term, results of this study suggest the following salient points. First, the equilibrium speed of adjustment is reduced after the introduction of the mini‐futures contract. This effect is particularly pronounced in the mini‐futures second year when its contracts are more heavily traded. Second, fractional integration tests support longer memory in the basis term after the contract introduction, again particularly in the second year. Third, the relationship between the full‐size and mini‐futures contracts appears highly efficient, with a quick speed of adjustment and short memory. Finally, an examination of the volatility dynamics suggests that in the second year of the mini‐futures contract shocks to spot return volatility exhibit longer memory. The results reported here suggest that the increased use of the mini‐futures contract after its introduction has had a detrimental impact on price discovery. © 2008 Wiley Periodicals, Inc. Jrl Fut Mark 28: 398–415, 2008     

Term Structure Models Driven by General Lévy Processes

As a generalization of the Gaussian Heath–Jarrow–Morton term structure model, we present a new class of bond price models that can be driven by a wide range of Lévy processes. We deduce the forward and short rate processes implied by this model and prove that, under certain assumptions, the short rate is Markovian if and only if the volatility structure has either the Vasicek or the Ho–Lee form. Finally, we compare numerically forward rates and European call option prices in a model driven by a hyperbolic Lévy motion with those in the Gaussian model.     

Return and price volatility of bonds with different credit risk

This article analyzes the relative price and monthly holding period return volatility of bonds with differing credit risk. The research proceeds by decomposing the causes of price volatility into that due to duration and yield volatility. Lower duration, due to higher coupon rate, and relatively stable yields, due to the behavior of credit risk and risk aversion over the business cycle, may well make lower quality issues such as A and Baa industrials exhibit less volatility than high-quality issues such as Aaa industrials and United States Treasury issues.     

Why Doesn't the Black‐‐Scholes Model Fit Japanese Warrants and Convertible Bonds?

In this paper, we investigate the systematic departures of traded prices of Japanese equity warrants and convertible bonds from their theoretical Black–Scholes values. We briefly consider transactions costs and the dilution adjustment as potential explanations of the discrepancy. However, our major focus is on shifts in volatility of the prices of the underlying stocks as a function of the stock price changes; such shifts are not taken into account in the Black–Scholes values. We assume that the pseudo‐probability distributions of prices of stocks of cross‐sections of companies which are roughly similar in size are identical. This simple assumption, which can be generalized, enables us to infer the implied probability distribution and binomial tree for stock price changes using the Derman and Kani (1994), Rubinstein (1994) and Shimko (1993) approach. The cross‐section of warrant prices implies an inverse volatility smile and a positively skewed probability density for stock prices. Rubinstein's identifying assumptions generate an implied binomial tree in which the relative size of up‐steps and down‐steps, and thus volatility, changes systematically as stock prices change. We briefly consider potential explanations for the implied behaviour, and for the difference in the smile pattern between index options and the warrants and convertibles.     

Improved estimation of portfolio value‐at‐risk under copula models with mixed marginals

Portfolio value‐at‐risk (PVAR) is widely used in practice, but recent criticisms have focused on risks arising from biased PVAR estimates due to model specification errors and other problems. The PVAR estimation method proposed in this article combines generalized Pareto distribution tails with the empirical density function to model the marginal distributions for each asset in the portfolio, and a copula model is used to form a joint distribution from the fitted marginals. The copula–mixed distribution (CMX) approach converges in probability to the true marginal return distribution but is based on weaker assumptions that may be appropriate for the returns data found in practice. CMX is used to estimate the joint distribution of log returns for the Taiwan Stock Exchange (TSE) index and the associated futures contracts on SGX and TAIFEX. The PVAR estimates for various hedge portfolios are computed from the fitted CMX model, and backtesting diagnostics indicate that CMX outperforms the alternative PVAR estimators. © 2006 Wiley Periodicals, Inc. Jrl Fut Mark 26:997–1018, 2006     

Multivariate realized volatility forecasts of agricultural commodity futures

We forecast the multivariate realized volatility of agricultural commodity futures by constructing multivariate heterogeneous autoregressive (MHAR) models with flexible heteroscedastic error structures that allow for non-Gaussian distribution, stochastic volatility, and heteroscedastic and serial dependence. We evaluate the forecast performances of various models based on both statistical and economic criteria. The in-sample and out-of-sample results suggest that the proposed MHAR models allowing for flexible heteroscedastic covariance structures outperform the benchmark MHAR models. In addition, the proposed Bayesian MHAR models allowing for t innovations improve both in-sample and out-of-sample forecast performance of the corresponding MHAR models with Gaussian innovations.     

基于GPD的稳健贝叶斯压缩感知ISAR成像方法

现有贝叶斯压缩感知（Bayesian Compressed Sensing,BCS）-逆合成孔径雷达(Inverse Synthetic Aperture Radar,ISAR)成像算法中先验分布模型不能很好地满足可压缩性,导致成像精度随脉冲数目的减小、高斯噪声的增强而急剧下降。为此,提出了一种基于广义Pareto分布改进BCS成像方法(Improving BCS imaging based on GPD,IGPCS)。该方法主要在BCS框架下利用广义Pareto先验分布替代传统的广义Gaussian先验分布,以增强模拟信号的稀疏先验和可压缩性。进一步地,为了克服后验概率模型计算困难等问题,采用最大后验（Maximum A Posteriori,MAP）方法对超参数进行估计。通过对Mig-25小型飞机的ISAR模拟实验表明,与传统方法相比,IGPCS方法能够获取极高的成像精度,并且对低脉冲数、强高斯噪声环境具有较强的鲁棒性。     

MOMENT EXPLOSIONS AND STATIONARY DISTRIBUTIONS IN AFFINE DIFFUSION MODELS

Many of the most widely used models in finance fall within the affine family of diffusion processes. The affine family combines modeling flexibility with substantial tractability, particularly through transform analysis; these models are used both for econometric modeling and for pricing and hedging of derivative securities. We analyze the tail behavior, the range of finite exponential moments, and the convergence to stationarity in affine models, focusing on the class of canonical models defined by Dai and Singleton (2000) . We show that these models have limiting stationary distributions and characterize these limits. We show that the tails of both the transient and stationary distributions of these models are necessarily exponential or Gaussian; in the non-Gaussian case, we characterize the tail decay rate for any linear combination of factors. We also give necessary and sufficient conditions for a linear combination of factors to be Gaussian. Our results follow from an investigation into the stability properties of the systems of ordinary differential equations associated with affine diffusions.     

VOLATILITY AND COVARIATION ESTIMATION WHEN MICROSTRUCTURE NOISE AND TRADING TIMES ARE ENDOGENOUS

This paper considers practically appealing procedures for estimating intraday volatility measures of financial assets. The underlying microstructure model accommodates the inherent properties of ultra high‐frequency data with the assumption of continuous efficient price processes. In this model, microstructure noise and trading times are endogenous but do not only depend on the prices. Using the (observed) last traded prices of the assets, we develop a new approach that enables to approximate the values of the efficient prices at some random times. Based on these approximated values, we build an estimator of the integrated volatility and give its asymptotic theory. We also give a consistent estimator of the integrated covariation when two assets (asynchronous by construction of the model) are observed.     

脉冲噪声下跳频信号时频图修正

为了在α稳定分布噪声的环境下获得清晰的跳频信号时频图,提出一种基于分数低阶SPWVD(Smoothed Pseudo Wigner-Vile Distribution)与形态学滤波相结合的跳频信号时频图修正算法。首先,根据接收到的多跳频信号建立跳频信号的模型和α稳定分布噪声模型;然后,采用低阶SPWVD变换抑制时频图中脉冲噪声;最后,根据形态学滤波处理方法对残留噪声进一步抑制进而得到清晰时频图。理论分析和仿真结果表明,所提算法在广义信噪比为-5 dB时仍可以得到清晰可靠的跳频信号时频图,并且基于时频图的参数估计性能优良。     

THE MOMENT FORMULA FOR IMPLIED VOLATILITY AT EXTREME STRIKES

Consider options on a nonnegative underlying random variable with arbitrary distribution. In the absence of arbitrage, we show that at any maturity T , the large-strike tail of the Black-Scholes implied volatility skew is bounded by the square root of  2| x |/ T   , where x is log-moneyness. The smallest coefficient that can replace the 2 depends only on the number of finite moments in the underlying distribution. We prove the moment formula , which expresses explicitly this model-independent relationship. We prove also the reciprocal moment formula for the small-strike tail, and we exhibit the symmetry between the formulas. The moment formula, which evaluates readily in many cases of practical interest, has applications to skew extrapolation and model calibration.     

Advances in Drama Theory for Managing Global Hazards and Disasters. Part I: Theoretical Foundation

Global risk and disaster management challenges are complex and ill-structured group decision processes characterized by time-sensitive, multi-faceted, and self-organizing negotiations, high decision stakes, extreme uncertainty, and dynamic, value-laden tradeoffs. Drama theory asserts that conflict resolution requires players to engage in a rational-emotional process of re-defining both the game and their “positions” in it until agreement on a satisfactory resolution is reached. While game theory has been widely applied to problems dealing with hazards, risk, and disasters, it assumes fixed players, options, and preferences, and hence does not allow for the re-definition of the conflict under consideration. Results show that drama theory constitutes a flexible and powerful tool for modeling group decision and negotiation processes involving natural, man-made, and health-related hazards, risk, and catastrophes in the post-911 security environment by modeling emotional responses that, throughout the course of a game, can lead to unanticipated reactions and change basic assumptions. This is achieved through the use of option boards to construct and analyze emergency, disaster, or crisis models that are structurally similar to game models. Finally, drama theory is compared and contrasted to conflict analysis, which developed from common roots in metagame analysis. The strengths and weaknesses of drama theory are critically evaluated in the context of global climate change and the mounting risk of a worldwide influenza pandemic.     

Portfolio Value-at-Risk with Heavy-Tailed Risk Factors

This paper develops efficient methods for computing portfolio value-at-risk (VAR) when the underlying risk factors have a heavy-tailed distribution. In modeling heavy tails, we focus on multivariate t distributions and some extensions thereof. We develop two methods for VAR calculation that exploit a quadratic approximation to the portfolio loss, such as the delta-gamma approximation. In the first method, we derive the characteristic function of the quadratic approximation and then use numerical transform inversion to approximate the portfolio loss distribution. Because the quadratic approximation may not always yield accurate VAR estimates, we also develop a low variance Monte Carlo method. This method uses the quadratic approximation to guide the selection of an effective importance sampling distribution that samples risk factors so that large losses occur more often. Variance is further reduced by combining the importance sampling with stratified sampling. Numerical results on a variety of test portfolios indicate that large variance reductions are typically obtained. Both methods developed in this paper overcome difficulties associated with VAR calculation with heavy-tailed risk factors. The Monte Carlo method also extends to the problem of estimating the conditional excess, sometimes known as the conditional VAR.