摘要
以期货合约的每一交易日的对数涨跌率来反映市场风险,借助VaR风险价值法,运用加权核估计技术(WKDE)和指数加权滑动模型(EWMA),建立了基于期货组合中持有头寸不同且可以进行风险对冲的期货组合市场风险非线性叠加评价模型,解决了同种商品、不同月份期货组合每一交易日最大损失的确定问题,并通过实证研究验证了模型的实用性。该模型的特点一是借助WKDE法预测组合中单个合约每一交易日涨跌率最大日亏损值,充分体现了期货合约涨跌率的实际走势,使VaR估计更加精确。二是通过动态迁移相关系数矩阵的计算保证了模型的精确性。采用EWMA模型预测动态变化的方差-协方差矩阵,从实证的角度得到更精准的动态迁移相关系数矩阵。三是考虑了组合中多头和空头不同头寸之间的风险对冲,避免了实际中期货组合风险的线性相加而造成放大风险或减少风险的不准确性,从而能较好地保证了模型的预测精度及准确性。四是通过基于风险非线性叠加建立的期货组合风险评价模型解决了SPAN系统中期货组合风险的线性叠加问题,从而得到更合理的组合风险预测值。
This paper using every trading day logarithmic fluctuation reflecting Futures contract market risk, in virtue of value at risk method, and adopting weighted kernel density estimation technology (WKDE) and exponentially weighted moving averages (EWMA), the same commodity of Futures portfolio market risk evaluation model based on different position's risk hedging and nonlinear addition is set up in order to solve the problem of the Futures portfolio every trading day's maximum loss. At the same time, we validate the model's practicability by demonstration research. The characteristics lies on four aspects: Firstly, using WKDE to forecast the single Futures of portfolio every trading day's volatility reflects the Futures volatility's trend, and this make the evaluation more precisely. Secondly, the model's precision is guaranteed by adopting dynamic transferred matrix. Using EWMA to forecast the portfolio's dynamic transferred variance-covariance matrix, we can get more reasonable and precise dynamic transferred coefficient matrix. Thirdly, different position contracts' risk are hedged, which avoids the biggish error and not very precise in practice. And this guarantees the forecasting model's precision and accuration. Fourthly, using this model based on risk nonlinear addition solves the portfolio's risk linear addition problem in SPAN system, and this will help us to get more practical forecasting value.
出处
《管理工程学报》
CSSCI
2006年第4期82-88,共7页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(70571010)
中期协联合研究计划资助项目(GT200410和ZZ200505)
大连市科技计划项目(2004C1ZC227)