摘要
基于切米雷泽(Khmaladze)鞅转化技术,利用标记的参数经验过程构造两种对利率模型漂移函数参数形式的设定检验方法,并给出该检验方法在原假设条件下的渐近性质与检验统计量的计算方法.所提出的利率模型漂移函数设定检验方法不依赖于利率模型的波动函数形式,即使对于复合假设检验问题也具有渐近分布无关性.蒙特卡洛模拟结果表明这些检验方法具有合理的检验水平(size)和检验功效(power).利用这些检验方法对我国的短期利率动态特征进行实证分析也能得到较好的检验效果.
Based on Khmaladze martingale transformation technology, two new tests for drift functions of inter- est rate models are constructed by the marked empirical processes. The asymptotic property of the tests and their calculation methods are also provided. The test statistics do not need to know the information about the diffusion function of interest models, and they do not depend on an asymptotic distribution even for the compos- ite null hypothesis. Monte Carlo simulations show that the tests have a reasonable size and power performance. The tests are also in effect when applied to analyzing the characteristics of short-term interest rates in China.
出处
《管理科学学报》
CSSCI
北大核心
2014年第11期43-56,共14页
Journal of Management Sciences in China
基金
国家自然科学基金资助项目(70971087)
关键词
鞅转化
利率模型
设定检验
蒙特卡洛模拟
实证分析
martingale transformation
interest rate model
specification test
Monte Carlo simulation
empir-ical analysis