分位数虚假回归与误差修正模型的理论研究及应用

发布时间：2018-06-07 21:20

本文选题：分位数误差修正模型 + hs300股指期货　；参考：《华侨大学》2013年硕士论文

【摘要】：分位数回归是用给定的解释变量X来估计被解释变量Y的条件分位数的一种基本方法。它不仅可以度量解释变量对被解释变量分布中心的影响，还可以度量分布极端情况下的该种影响，具有经典最小二乘回归所不具备的优点。经济金融领域中的时间序列多是非平稳的，直接对他们进行最小二乘估计，有可能出现虚假回归现象。通过引入均衡误差项建立误差修正模型，不仅能够避免虚假回归，而且能够同时研究变量之间的长期均衡关系和短期参数关系。通过蒙特卡罗实验发现，非平稳时间序列的分位数回归同样存在着虚假回归的问题，而目前对该问题的研究还不多。本文结合传统的误差修正模型提出了分位数误差修正模型，该模型将分位数回归法与误差修正模型结合起来，同时具有两者的优点，，能够更为全面地研究误差修正模型的调整机制。具体研究思路是：先采用最小二乘法估计具有协整关系的变量间的长期均衡关系，然后对误差修正方程进行分位数回归，并讨论在各分位点上的调整机制，特别是在极端分布情况下的调整机制。本文进一步利用分位数误差修正模型对沪深300股指期货市场的价格发现功能及其与现货市场之间的信息传导关系进行了实证研究，并将研究结果与用传统的误差修正模型得到的结果进行了比较。结果表明：随着分位的增加现货市场的信息效率逐渐提高，体现出比期货市场更大的信息优势，而传统的误差修正模型放大了市场调节机制的作用。
[Abstract]:Quantile regression is a basic method for estimating the conditional quantiles of the interpreted variable Y with a given explanatory variable X. It can not only measure the effect of explanatory variables on the distribution center of explained variables, but also measure the effects in extreme cases of distribution. The time series in the field of economics and finance are mostly non-stationary, and if they are directly estimated by least squares, false regression may occur. By introducing the equilibrium error term to establish the error correction model, not only the false regression can be avoided, but also the long-term equilibrium relationship and the short-term parameter relationship between variables can be studied simultaneously. Monte Carlo experiments show that the quartile regression of non-stationary time series also has the problem of false regression, but there is not much research on this problem at present. In this paper, a quantile error correction model is proposed based on the traditional error correction model. The model combines the quantile regression method with the error correction model and has the advantages of both. The adjustment mechanism of error correction model can be studied more comprehensively. The specific research idea is: first, using the least square method to estimate the long-term equilibrium relationship among the variables with cointegration relationship, then the quantile regression of the error correction equation is carried out, and the adjustment mechanism at each locus is discussed. Especially in the case of extreme distribution, this paper makes an empirical study on the price discovery function of the Shanghai and Shenzhen 300 stock index futures market and its information transmission relationship with the spot market by using the quantile error correction model. The results are compared with those obtained by the traditional error correction model. The results show that with the increase of quantiles, the information efficiency of the spot market improves gradually, which shows that the information advantage is greater than that of the futures market, and the traditional error correction model amplifies the role of the market regulation mechanism.
【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【相似文献】