基于小波变换的LMSV模型对人民币汇率波动的研究

将小波引入到LMSV模型波动长记忆性的估计与检验中,提出了基于小波变换的LMSV模型波动长记忆性的伪极大似然估计法和波动长记忆性的检验方法,并对各汇率波动序列长记忆效应的大小程度进行了验证．结果表明各汇率波动序列存在长记忆效应．人民币对美元的汇率波动序列受历史信息的影响程度最高．     

基于两步法带宽选择的金融时间序列长记忆参数估计

金融时间序列长记忆参数的半参数估计方法以频域分析为主,带宽选择是其中必不可少的关键环节。不同的带宽可能给出差异明显的长记忆参数估计值,甚至产生矛盾的结论,进而影响时间序列平稳性的判断。本文提出一种两步法,用于金融时间序列长记忆估计的半参数方法的带宽选择,并进一步对长记忆参数进行估计:首先,为了克服半参数方法忽略短期结构的不足,通过信息准则判断ARFIMA(p,d,q)过程的短记忆结构;其次,用短记忆模型拟合差分后的序列,根据拟合效果确定选择带宽及长记忆参数估计值。数值模拟显示以长记忆参数估计值均方根误差最小为标准,两步法优于其他方法。经上证50指数已实现波动率日数据的实证检验,两步法在长记忆模型中的预测误差最小;与短记忆模型相比,两步法在中期提前预测步长上具有优势。     

长记忆模型的Bayes估计及其在汇率波动中的应用

本文介绍了长记忆模型及其检验方法,根据Bayes原理,提出了记忆参数的一种新的估计方法.在运用Teverovsky/Taqqu(1997)年提出的一种基于样本方差的直观方法的初步检验基础上,运用新的检验方法,以美元对人民币汇率为研究对象,说明了我国汇率波动的长记忆性.然后,将经典的GPH-估计与新方法所得出的Bayes-估计相比较,可以看出这种新的估计方法较之经典的GPH-估计要稳定.     

基于一类长记忆过程的经济时间序列建模研究

本文介绍了长记忆的概念 ,首次提出了一种与 ADF检验相结合的长记忆性判断方法。给出了将参数 d的初估计与近似极大似然估计相结合 ,将时间序列长记忆分析与短记忆分析相结合的建立ARFIMA模型的方法。论文进行了实证研究 ,并证明了该模型与其它模型相比的优效性     

基于半参数EGARCH模型的VaR和CVaR度量与实证研究

本文主要是运用半参数EGARCH模型研究中国股票市场风险.首先,运用半参数EGARCH模型估计波动率.其次,研究中国股票市场风险,即估计VaR和CVaR值.最后,利用半参数EG;ARCH模型对上证指数进行实证分析。结果表明,半参数EGARCH模型比一般的EGARCH模型能更准确地度量中国股市风险。     

半参数再生散度非线性模型中参数的投影核和刀切估计的相合性与渐近正态性

半参数再生散度非线性模型(SRDNM)是再生散度非线性模型和半参数回归模型的自然推广和发展,它包括半参数非线性模型和半参数广义线性模型等特殊模型. 基于非参数部分的局部核估计, 给出了SRDNM模型中参数的投影核估计与刀切估计, 并对其进行了理论比较. 在一定的正则条件下,得到了这两类估计的强相合性与渐近正态性. 相比之下, 刀切估计比投影核估计具有更大的渐近方差. 最后, 模拟研究和实例分析被用来说明所给方法的有效性.     

基于股市高频数据的半参数估计方法

针对非参数方法研究国内股市长记忆性时结论参差不齐的现状,本文研究了更为稳健的半参数估计方法,即局部W h ittle(LW)估计和对数周期图(LP)回归。通过对不同频率高频数据的分析,证实了LW估计方法尽管需要数值最优化,但仍然要优于LP回归。进而将LW估计首次应用于中国股市,结果表明不同频率绝对收益序列的长记忆强度基本一致;同时发现,重大突发事件发生时的长记忆性表现得最为强烈,且事件后比事件前表现的要强烈,这说明股票市场的溢出效应在事件后增强,此项结论对我国证券市场有一定的借鉴意义。     

带有不可忽略缺失数据的半参数再生散度模型的贝叶斯分析

半参数再生散度模型是再生散度模型和半参数回归模型的推广,包括了半参数广义线性模型和广义部分线性模型等特殊类型.讨论的是该模型在响应变量和协变量均存在非随机缺失数据情形下参数的Bayes估计和基于Bayes因子的模型选择问题.在分析中,采用了惩罚样条来估计模型中的非参数成分,并建立了Bayes层次模型;为了解决Gibbs抽样过程中因参数高度相关带来的混合性差以及因维数增加导致出现不稳定性的问题,引入了潜变量做为添加数据并应用了压缩Gibbs抽样方法,改进了收敛性;同时,为了避免计算多重积分,利用了M-H算法估计边缘密度函数后计算Bayes因子,为模型的选择比较提供了一种准则.最后,通过模拟和实例验证了所给方法的有效性.     

基于高频数据的时间序列长记忆系数比较研究

在高频数据中,交易持续时间序列,交易量序列和收益率平方序列都存在长记忆现象。本文采用半参数方法,对同一只股票的上述三种序列的长记忆参数进分析比较,结果显示,三种序列都存在长记忆现象,而且同一只股票的上述三种序列具有相同的长记忆参数。这些结果为微观市场的相关理论提供了实证。     

沪深300股指期货的风险测度模型研究

以沪深300股指期货指数的30分钟交易数据为例,首先对其价格变化的动力学特征及波动模式进行了全面深入的考察,然后运用严谨系统的后验分析(Backtesting analysis)方法,分别在多头和空头两种头寸状况以及5种不同分位数水平下,实证对比了8种风险测度模型对VaR(Value at Risk)和ES(Excepted shortfall)两种不同风险指标估计的精度差异。研究结果表明:我国股指期货市场的价格波动具有较为明显的有偏和尖峰厚尾分布、聚集特征和长记忆性;采用有偏学生t分布和长记忆模型有助于提高对沪深300股指期货的风险测度精度,而在波动模型中包含杠杆效应项对提高风险估计精度并无太多帮助;在综合考虑了模型对沪深300股指期货价格变化动力学的刻画效果以及对不同风险指标的测度精度等因素后,基于有偏学生t分布的GARCH模型是一个相对合理的风险测度模型选择。     

The stochastic unit root model and fractional integration: An extension to the seasonal case

In a recent paper, Yoon (Working Paper, Department of Economics and Related Studies, University of York, 2003. Presented at the ESF-EMM Second Annual Meeting, Rome, Italy, 2003) asserts that the stochastic unit root (STUR) model is closely related to long memory processes, and, in particular, that it is a special case of an I(d) process, with d = 1.5. In this paper we question this claim by using parametric and semiparametric techniques for modelling long memory. We extend the analysis by considering both non-normality and seasonality, and shed light, theoretically and by means of Monte Carlo methods, on the relationship between the seasonal STUR and the seasonal I(d) models. The results show that methods that are specifically designed for testing I(d) statistical models are not appropriate for testing the STUR model. Moreover, they have in some cases very low power against STUR alternatives. Copyright © 2007 John Wiley & Sons, Ltd.     

关于两总体均值差的一种半参数假设检验方法

近年来, 已有一些在半参数密度函数比模型下建立半参数统计分析方法的报道, 这些方法往往比参数方法稳健, 比非参数方法有效. 在本文里, 我们提出一种半参数的假设检验方法用于对两总体均值差进行假设检验. 该方法主要建立在对两总体均值差进行半参数估计的基础上. 我们报告了一些理论和统计模拟的结果, 得出该方法在数据符合正态性假设时, 比常用的参数和非参数方法略好; 而在数据不符合正态性假设时, 它的优势就非常明显. 我们还将提出的方法用到了两组真实数据的分析上.     

Bootstrap techniques in semiparametric estimation methods for ARFIMA models: A comparison study

Summary  This paper considers different bootstrap procedures for investigating the estimation of the fractional parameter d in a particular case of long memory processes, i.e. for ARFIMA models withd in (0.0, 0.5). We propose two bootstrap techniques to deal with semiparametric estimation methods of d. One approach consists of the local bootstrap method for time frequency initially suggested for the ARMA case by Paparoditis and Politis (1999), and the other consists of the bootstrapping in the residuals of the frequency-domain regression equation. Through Monte Carlo simulation, these alternative bootstrap methods are compared, based on the mean and the mean square error of the estimators, with the well-known parametric and nonparametric bootstrap techniques for time series models.     

Inference for the treatment effects in two sample problems with right-censored and length-biased data

In the study of comparing treatment effects, the data structures of two samples may be different. In this paper, we develop a unified semiparametric estimating equation approach to estimate various types of treatment effects with right-censored and length-biased data based on a semiparametric two-sample model. The large sample properties of the proposed estimators are derived and numerical studies are conducted to illustrate the proposed methods.     

基于半参数可加自回归模型的黄金价格实证分析

大量实证研究表明,半参数自回归模型较传统的线性回归而言,能更好的拟合实际数据。本文构造了一类半参数可加自回归模型,基于条件最小二乘方法及核估计方法给出了估计模型参数和未知函数的迭代算法,讨论了估计量的渐近性质。通过数值模拟验证了估计的效果。并将模型应用于黄金价格数据的实证分析之中。实证分析结果表明,我们对现有模型的改进是必要的。     

Bayesian Semiparametric Regression Analysis of Multicategorical Time-Space Data

We present a unified semiparametric Bayesian approach based on Markov random field priors for analyzing the dependence of multicategorical response variables on time, space and further covariates. The general model extends dynamic, or state space, models for categorical time series and longitudinal data by including spatial effects as well as nonlinear effects of metrical covariates in flexible semiparametric form. Trend and seasonal components, different types of covariates and spatial effects are all treated within the same general framework by assigning appropriate priors with different forms and degrees of smoothness. Inference is fully Bayesian and uses MCMC techniques for posterior analysis. The approach in this paper is based on latent semiparametric utility models and is particularly useful for probit models. The methods are illustrated by applications to unemployment data and a forest damage survey.     

基于Gibbs抽样的随机前沿面半参数模型的估计

本文对文献[1]提出的随机前沿面半参数模型的参数估计作了进一步研究,利用基于Gibbs抽样的技术,成功构造出了随机前沿面半参数模型的参数估计,并给出了生产技术有效性测算的方法及其证明,提高了随机前沿面半参数模型的应用价值.     

Strong dependence in the nominal exchange rates of the Polish zloty

We examine the nominal exchange rates of six currencies (Canadian, Australian and U.S. dollars, euro, Japanese yen and U.K. pound) against the Polish zloty by means of statistical techniques based on unit roots and other long memory processes. We use both parametric and semiparametric methods for estimating and testing integer and fractional orders of integration at the long run or zero frequency. The results show that unit roots are likely to occur in relation with the U.S. and the Canadian dollars, the Japanese yen and the U.K. pound. However, for the Australian dollar and the euro, this hypothesis is rejected in favour of smaller degrees of integration, implying mean reversion in their behaviour. Thus, for the former currencies, in the event of an exogenous shock affecting the exchange rates, strong policy actions must be required to bring the variables back to their original levels. On the other hand, for the Australian dollar and the euro, there exists less need of action since the series will return to their levels sometime in the future. Copyright © 2006 John Wiley & Sons, Ltd.