摘要
文章指出在对利率模型进行适当的离散化后,运用极大似然估计方法进行参数估计优于GMM方法。通过选择7天银行间拆借利率作为模型中短期利率的近似替代,我们第一次将极大似然估计法运用于中国市场,对一系列单因子利率模型的参数进行了估计,并对这些模型进行了似然比检验。我们发现,在中国市场中CKLS模型中γ的值约为1.5,与美国市场中的γ值相近,与英国市场的γ值相差很大,与美英两国都不同的是中国的利率变化有明显的均值回复效应。中国利率均值回复效应显著是中国人民银行对央行目标利率的调整没有美联储对联邦基金利率的调整频繁所致。
In this paper we point out that after proper discretization of interest rate models, the maximum likelihood estimation method (MLE) is superior to GMM for estimating parameters of these models. By using the 7-day interbank offer rate as the proxy for the short term rate in the models, we apply the MLE method for the first time to the Chinese market and estimate the parameters for several single factor interest rate models. We also perform ML ratio tests for these models. We find that the value of γ in the CKLS model is around 1.5 in China, comparable to its value in the US but very different from its value in the UK. In addition, we find that unlike that in the US and the UK, the movement of interest rates in China shows pronounced mean-reverting tendency. This can be explained by the fact that the People's Bank of China has made less frequent adjustments to the central bank's benchmark rate than the Federal Reserve in the US.
出处
《财经研究》
CSSCI
北大核心
2004年第10期62-69,95,共9页
Journal of Finance and Economics
关键词
单因子利率模型
极大似然估计
CKLS模型
single-factor interest rate models
maximum likelihood estimation
CKLS model
