金融波动的赋权“已实现”双幂次变差及其应用

金融波动是金融研究中的热点问题。金融高频数据比低频数据包含了更丰富的日内收益波动信息,因此对金融高频时间序列的研究成为金融领域中备受关注的焦点。"已实现"波动是利用高频数据计算金融波动率的全新方法,目前在金融高频数据的研究中应用十分广泛,但它具有误差较大和不稳健的缺点,因此各种改进方法应运而生,其中"已实现"双幂次变差克服了"已实现"波动的不稳健的缺点。本文提出赋权"已实现"双幂次变差的概念,不但继承了"已实现"双幂次变差的稳健性,而且满足无偏性和最小方差性,通过理论证明和实证研究都表明其能够更准确的度量金融波动率。     

基于TrTS取样的股票收益率RV测度的改进

由于噪声的存在使得高频数据的分析过程存在着诸多困难,本文探讨了高频数据情况下的金融资产收益率已实现波动率的估计问题。在离散化的跳跃模型基础上,通过混合泊松分布而非传统的连续扩散模型来描述价格过程,并进一步提出了不同于以往文献研究的噪声假设,即在独立同分布的噪声假设基础上放松约束条件,保持噪声的独立性,但是允许噪声强度随时间变化,以此改善了传统的固定时间间隔取样模式。为了进一步改善估计效果,我们结合了TrTS(Transaction Time Sampling)以及一阶偏误修正的RV(realized variance)估计方式RVAC(1) (first-order AutoCorrelation to RV)。对来自两个交易所不同板块股票的价格数据进行的实证研究结果表明,本文的估计方式虽然对于个别股票价格数据会产生与实际背离潜在真实价格参数,但整体上对于已实现波动率的估计效果是比较稳健的。     

中国ETF基金价格“已实现”波动率、跟踪误差之间的Granger关系研究

在高频数据条件下,中国ETF基金价格"已实现"波动率与跟踪误差之间是否存在着因果关系并存在着信息的先导效应？基于"已实现"波动、跟踪误差计算方法及Granger因果检验过程、VAR模型等,本文对此进行了深入研究。研究结果认为:中国ETF基金价格"已实现"波动率与两种跟踪误差分别具有Granger因果关系,后者是前者的Granger原因;中国ETF基金价格"已实现"波动率序列与两种跟踪误差序列的同期及一、二阶滞后相关性较高,而跟踪误差滞后于"已实现"波动率;当ETF基金的跟踪误差受外部市场条件的某一冲击后,将给ETF基金价格"已实现"波动率带来同向的冲击,这一冲击具有一定的持续性和滞后性。     

Modeling and Forecasting Realized Volatility

We provide a framework for integration of high–frequency intraday data into the measurement, modeling, and forecasting of daily and lower frequency return volatilities and return distributions. Building on the theory of continuous–time arbitrage–free price processes and the theory of quadratic variation, we develop formal links between realized volatility and the conditional covariance matrix. Next, using continuously recorded observations for the Deutschemark/Dollar and Yen/Dollar spot exchange rates, we find that forecasts from a simple long–memory Gaussian vector autoregression for the logarithmic daily realized volatilities perform admirably. Moreover, the vector autoregressive volatility forecast, coupled with a parametric lognormal–normal mixture distribution produces well–calibrated density forecasts of future returns, and correspondingly accurate quantile predictions. Our results hold promise for practical modeling and forecasting of the large covariance matrices relevant in asset pricing, asset allocation, and financial risk management applications.     

考虑微观结构噪声与测量误差的波动率预测

以测量误差的分布理论为基础,本文将微观结构噪声的影响引入到测量误差的方差中,构建了包含微观结构噪声影响的HARQ-N模型。使用蒙特卡洛模拟与中国股市的高频数据对HAR、HARQ、HARQ-N模型与HAR-RV-N-CJ模型的估计和预测进行了比较,研究发现,HARQ模型和HARQ-N模型的测量误差修正项对波动率的影响系数统计显著为负,HARQ-N模型的测量误差项影响系数远大于HARQ模型,更大程度地减弱当期微观结构噪声和测量误差的影响。并且,考虑微观结构噪声和测量误差的HARQ-N模型样本内和样本外预测效果在统计上显著优于HAR模型、HARQ模型与HAR-RV-N-CJ模型。     

基于四次幂差修正HAR模型的股指期货波动率预测

股指期货波动率建模与预测是揭示其波动运行规律和市场风险是重要途径。本文基于跳跃、好坏波动率与符号跳跃建立四组HAR模型,提出单级纠偏HARQ类模型和多级纠偏HARQF类模型,实证研究揭示股指期货波动运行规律,并采用MCS检验来评估模型优劣。HAR建模考察连续与跳跃波动、好与坏波动率的两种已实现波动分解。为了降低波动率估计偏差,基于最小化MSE准则确定最优抽样频率,利用已实现核修正的ADS检测法识别跳跃,采用已实现核估计修正好坏波动率与符号跳跃。基于沪深300股指期货的实证研究表明：连续波动比跳跃波动对未来已实现波动贡献更大;好坏波动率具有不对称波动冲击,而符号跳跃对未来波动具有负向冲击;好坏波动率分解优于连续与跳跃波动分解;中位数已实现四次幂差能够显著提升HAR类模型的样本内外预测能力;与样本内预测相反,样本外预测中单级纠偏HARQ类模型优于多级纠偏HARQF类模型;MCS检验得出HARQ-RV-SJ模型表现最佳。研究结论与启示对认识股指期货波动规律和市场风险具有意义。     

Inference for Continuous Semimartingales Observed at High Frequency

The econometric literature of high frequency data often relies on moment estimators which are derived from assuming local constancy of volatility and related quantities. We here study this local‐constancy approximation as a general approach to estimation in such data. We show that the technique yields asymptotic properties (consistency, normality) that are correct subject to an ex post adjustment involving asymptotic likelihood ratios. These adjustments are derived and documented. Several examples of estimation are provided: powers of volatility, leverage effect, and integrated betas. The first order approximations based on local constancy can be over the period of one observation or over blocks of successive observations. It has the advantage of gaining in transparency in defining and analyzing estimators. The theory relies heavily on the interplay between stable convergence and measure change, and on asymptotic expansions for martingales.     

股市波动率的短期预测模型和预测精度评价

基于幂转换以及不设定扰动项的具体相关结构和分布形式,构建了半参数的短期预测模型来预测中国股市的波动率.模型采用基于极值估计量的两阶段估计法进行估计,估计方法的小样本性质表现良好.此外,还通过具有Bootstrap特性的SPA检验实证比较了新模型与其他6种预测模型的预测精度.实证结果表明,在各种损失函数下,半参数短期预测...     

我国黄金期货价格波动率预测研究：来自模型缩减方法的新证据

黄金作为重要的避险资产,对其价格波动的定量描述和预测对于各类投资者的风险管理决策意义重大。基于标准回归预测模型,采用主成分分析、组合预测和两种主流的模型缩减方法(Elastic net 和Lasso)构建新的波动率预测模型,探究哪种方法能够更有效地利用多个预测因子信息。进一步,运用模型信度集合(model confidence set,MCS)、样本外R2和方向测试(Direction-of-Change,DoC)三种评价方法检验新模型的样本外预测精度。实证结果显示：不论是基于哪一种评价方法,相比其它竞争模型,两种缩减模型的样本外预测精度均为最优,可以为我国黄金期货价格的波动率预测提供可靠保障。     

中国股票市场个股已实现波动率估计

从修正市场微观结构等干扰因素的角度出发,基于高频交易数据,同时采用一阶偏差修正方法估计了深圳证券交易市场个股的已实现波动率。研究结果表明,基于分笔交易数据计算的已实现波动率能有效提高个股真实波动率的估计精度,从而可以为波动率相关研究提供一个参照基准的波动率值。     

基于LSTM神经网络的金融时间序列预测

本文提出将小波分析与纳入时间序列依赖特征的长短期记忆(LSTM)神经网络相结合,构建金融时间序列数据预测模型,以克服现有模型对金融时间序列数据非平稳、非线性、序列相关等复杂特征以及数据间非线性交互关系无法反映的缺陷。同时,以道琼斯工业指数日收盘价为例,探究LSTM神经网络对实际金融时间序列数据的预测能力,比较其与多层感知机、支持向量机、K近邻、GARCH四种模型的预测效果。实证结果表明LSTM神经网络具有更高的预测精度,能够有效预测金融时间序列数据的长短期动态变化趋势,说明了其对金融时间序列数据预测的适用性与有效性。此外,对金融时间序列数据进行小波分解与重构,可有效提高LSTM预测模型的泛化能力,以及对长短期动态趋势的预测精度。     

Environmental Surveillance System Characteristics and Impacts on Confidence About No Undetected Serotype 1 Wild Poliovirus Circulation

Surveillance for poliovirus during the polio endgame remains uncertain. Building on prior modeling of the potential for undetected poliovirus transmission for conditions like those in Pakistan and Afghanistan, we use a hypothetical model to explore several key characteristics of the poliovirus environmental surveillance (ES) system (e.g., number and quality of sites, catchment sizes, and sampling frequency) and characterize their impacts on the time required to reach high (i.e., 95%) confidence about no circulation (CNC95%) following the last detected case of serotype 1 wild poliovirus. The nature and quality of the existing and future acute flaccid paralysis (AFP) surveillance and ES system significantly impact the estimated CNC95% for places like Pakistan and Afghanistan. The analysis illustrates the tradeoffs between number of sites, sampling frequency, and catchments sizes, and suggests diminishing returns of increasing these three factors beyond a point that depends on site quality and the location of sites. Limitations in data quality and the hypothetical nature of the model reduce the ability to assess the extent to which actual ES systems offer benefits that exceed their costs. Thus, although poliovirus ES may help to reduce the time required to reach high confidence about the absence of undetected circulation, the effect strongly depends on the ability to establish effective ES sites in high‐risk areas. The costs and benefits of ES require further analysis.     

小波包与混沌集成的股票市场预测新方法

基于小波包变换和混沌理论提出了一种股票市场建模及其预测的新方法,既能刻划时间序列的规律,又能捕捉混沌状态的特征.首先,应用小波包变换对上证综指和深证成指日收益率序列进行三层分解,分别得到第三层从低频到高频八个频率成分的时序,并在此基础上作进一步分析,结果表明中国股市存在混沌特性;然后,应用混沌理论分别建立从低频到高频八个时序的预测模型,分别对八个时序进行预测;最后,基于小波包理论对混沌模型预测的结果予以重构,实现对原始收益率序列的预测.与现有方法比较,结果表明该方法具有较高的精度,有极大的应用范围.     

全球金融危机下的股票市场波动跳跃研究——基于高频数据的中美比较分析

本文以雷曼破产日至2009年1月底这段时期内上证综指、恒生指数以及S&P500指数的日内高频数据作为研究对象,采用跳跃显著性检验方法和扩展HAR模型,对波动跳跃特征进行了实证研究.结果表明:雷曼危机导致股市波动的显著提高,但中国内地股市受到的影响最小;中国香港股市成为波动跳跃发生频率最高、跳跃幅度最大的市场,且波动跳跃主要发生在夜间休市时间内;雷曼危机使得波动率模型的预测精度大大降低,股市风险变得更加难以预测,对于新兴市场来说这一现象更加明显.     

已实现波动GAS-HEAVY模型及其实证研究

引入日内高频数据计算的已实现波动，能够提高波动模型预测能力。本文将日收益和已实现波动联合建模，提出一种新的波动模型。选取尺度调整t分布和F分布作为日收益和已实现波动的分布，更为充分和灵活地捕捉厚尾性，采用得分驱动方法设定波动模型更新项，得出广义自回归得分（GAS）波动模型，提高对实际模型的逼近效率。本文对模型遍历性和平稳性进行证明，并与同类模型进行比较。蒙特卡罗模拟实验显示，在数据生成过程误设的情况下本文提出的GAS-HEAVY模型比同类模型具有更好的数据拟合效果。基于沪综指、深成指和沪深300指数2013.1至2017.4日内1分钟高频数据实证分析表明，不同损失函数的SPA检验下GAS-HEAVY模型的波动预测能力显著强于其它同类模型。本文给出的GAS-HEAVY模型为有关理论研究和市场应用提供了新的波动计量工具。     

中国股票市场的时变杠杆效应研究——基于随机Copula模型的实证分析

构建随机Copula模型研究了中国股票市场在极端市场条件下的时变杠杆效应.为了解决金融市场中波动率不可直接观测的问题,采用已实现波动率测度作为隐波动率的代理变量,进而运用基于有效重要性抽样的极大似然(EIS-ML)方法估计了随机Copula模型的参数.基于沪深股市数据的实证研究表明:中国股票市场的杠杆效应具有非对称特征,即股市低收益率伴随高波动率,但股市高收益率不一定伴随低波动率;中国股票市场的杠杆效应存在显著的时变性,沪深股市杠杆效应表现出类似的变化趋势;随机Copula模型相比其它Copula模型(静态Copula模型和时变Copula模型)具有更好的数据拟合效果.     

跳跃的估计、股市波动率的预测以及预测精度评价

本文基于C_TMPV理论估计已实现波动率的跳跃成分,在此基础上构建考虑跳跃的AHAR-RV-CJ模型和MIDAS-RV-CJ模型来预测中国股市的已实现波动率,并评价和比较各类波动率模型的预测精度。实证结果表明:基于C_TMPV估计的波动率跳跃成分对日、周以及月波动率的预测有显著的正向影响;AHAR-RV-CJ模型和MIDAS-RV-CJ模型的样本内和样本外预测精度在不同的预测时域上都是最高的,尤其是对数形式的模型;MIDAS族模型的样本外预测精度在中长期预测时域上比HAR族模型高;AHAR-RV-CJ模型和MIDAS-RV-CJ模型的样本外预测能力在中长期预测时域上比基于低频数据的Jump-GARCH模型、SV-CJ模型和SV-IJ模型好。     

The Repeatability of Uncertainty and Sensitivity Analyses for Complex Probabilistic Risk Assessments

The performance of a probabilistic risk assessment (PRA) for a nuclear power plant is a complex undertaking, involving the assembly of an accident frequency analysis, an accident progression analysis, a source term analysis, and a consequence analysis. Each of these analyses is, in itself, quite complex. Uncertainties enter into a PRA from each of these analyses. An important focus in recent PRAs has been to incorporate these uncertainties at each stage of the analysis, propagate the subsequent uncertainties through the entire analysis, and include uncertainty in the final results. Monte Carlo procedures based on Latin hypercube sampling provide one way to perform propagations of this type. In this paper, the results of two complete and independent Monte Carlo calculations for a recently completed PRA for a nuclear power plant are compared as a means of providing empirical evidence on the repeatability of uncertainty and sensitivity analyses for large-scale PRA calculations. These calculations use the same variables and analysis structure with two independently generated Latin hypercube samples. The results of the two calculations show a high degree of repeatability for the analysis of a very complex system.     

Inventory record inaccuracy – The impact of structural complexity and lead time variability

This article presents a rigorous prospective analysis of the impact of Inventory Record Inaccuracy (IRI) on complex multi-echelon Supply Chains (SCs). Specifically, key operational factors (i.e., the magnitude of the error, frequency of the inventory audits and lead time variability) and SC structure are systematically assessed. We find that the detrimental effects of IRI are exacerbated by the structural complexity of the SC and lead time variability. Furthermore, we show how the efficacy of countermeasure strategies may vary depending on SC configuration and operational conditions. These results allow us to provide interesting managerial recommendations to guarantee investment in prevention and correction strategies.     

External Initiators in Probabilistic Reactor Accident Analysis—Earthquakes, Fires, Floods, Winds

This article discusses the methodologies presently available for analyzing the contribution of "external initiators" to overall risks in the context of PRA (probabilistic risk assessment) of large commercial nuclear power reactors. "External initiators" include earthquakes, fires and floods inside the plant, external floods, high winds, aircraft, barge, and ship collisions, noxious or explosive gases offsite, and so on. These are in contrast to "internal initiators" such as active or passive plant equipment failures, human errors, and loss of electrical power. The ability to consider external initiators within PRA has undergone major advances in recent years. In general, uncertainties associated with the calculated risks from external initiators are much larger than those associated with internal initiators. The principal uncertainties lie with development of hazard curves (such as the frequency of occurrence of an event exceeding a given size: for example, the likelihood of a hurricane with winds exceeding 125 knots). For assessment of earthquakes, internal fires and floods, and high winds, the methodology is reasonably mature for qualitative assessment but not for quantitative application. The risks from other external initiators are generally considered to be low, either because of the very long recurrence time associated with the events or because the plants are judged to be well designed to withstand them.