论文部分内容阅读
Madhavan等给出了一个著名的价差分解模型:MRR模型.MRR模型利用广义矩方法估计参数,这使得该模型在扩展上受到矩条件限制.考虑到价格演变存在条件异方差现象基础上,本文给出一类极大似然估计方法,使MRR模型在模型扩展中增加的变量不受矩条件限制.以上证180成分股数据作为样本,实证结果表明带限定条件的极大似然估计法准确地捕捉了交易中的信息参数及流动成本,同时在原模型估计及扩展上有很好的稳定性.模型扩展上,进一步检验了交易强度对信息参数的影响.
Madhavan et al. Give a famous model of price difference decomposition: MRR model. MRR model uses the generalized moment method to estimate the parameters, which makes the model limited by the moment condition on the extension. Considering the existence of conditional heteroscedasticity in price evolution, A kind of maximum likelihood estimation method is proposed to make the variable added by MRR model to the model expansion is not restricted by the moment condition.The empirical results show that the maximum likelihood estimation with limited conditions is accurate It captures the information parameters and the flow cost in the transaction, and also has good stability in the original model estimation and expansion.Furthermore, the influence of transaction intensity on the information parameters is further tested on the model expansion.