论文部分内容阅读
针对期货最优套期保值策略估计中可能存在的估计风险问题,本文对单变量线性回归模型(OLS模型)和多变量线性回归模型(VAR模型和EC-VAR模型)进行贝叶斯分析,并采用Gibbs抽样方法对中国铜期货市场的最优套期保值策略进行了实证分析。本文还同时估计了基于频率统计方法的最优套期保值策略,并对贝叶斯统计下和频率统计下的最优套期保值策略进行了分析比较。实证结果清楚表明,估计风险对模型结果有重要影响。在处理估计风险方面,贝叶斯统计较频率统计方法有明显优势。
In view of the possible estimation risk in the futures hedging strategy estimation, this paper carries out Bayesian analysis on the univariate linear regression model (OLS model) and the multivariate linear regression model (VAR model and EC-VAR model) The Gibbs sampling method is used to analyze the optimal hedging strategy of Chinese copper futures market. In this paper, we also estimate the optimal hedging strategy based on frequency statistics, and analyze and compare the optimal hedging strategies under Bayesian statistics and frequency statistics. The empirical results clearly show that the estimated risk has an important impact on the model results. Bayesian statistics have a clear advantage over frequency statistics in dealing with the estimation of risks.