国债利率期限结构的组合优化模型研究

The Combinatorial Optimization Model Research of National Term Structure of Interest Rates

摘要: 我国正在加快利率市场化进程，利率期限结构越来越重要，利率期限结构在国债结构的设计、金融产品定价和国家掌握宏观经济等方面具有重要的意义，因此选取国债利率期限结构的优化模型进行研究。本文选取上海交易所某一天的静态数据，选取三次多项式样条函数模型和Nelson-Siegel模型以市场价格和理论价格误差最小为约束条件进行模型参数估计，基于熵权法以平均绝对偏差和均方差为判断标准将两种模型赋予权重进行组合，得到组合优化后的模型，利用国债价格拟合误差最小原则进行实证分析，结果显示组合后的模型在拟合国债价格要优于两个单一模型，另外，用样本外的数据进行国债定价的预测。本文最后得出结论：组合优化模型的利率期限结构的拟合优于两个单一模型。基于这一结论，本文为我国利率期限结构模型未来的研究方向提出了建议。

Abstract: Our country is accelerating interest rate marketization process. The term structure of interest rates is more and more important. The term structure of interest rates is important for the structural design of the Treasury, financial products pricing and national macro economy and so on, so this paper selects optimal model for the Treasury term structure to study. This paper selects the Shanghai Stock Exchange one day static data, Cubic polynomial spline function model and Nelson-Siegel model using the actual price and the theoretical price as minimum error constraint to estimate the model parameters. Based on Entropy Method and in the mean absolute deviation and standard deviation for the criterion, the two models are given a combination of weights to be a combination of optimized. The principle of minimum fitting error of Treasury prices is used to do empirical analysis. The results showed that the combined model in fitting Treasury prices better than two single models. In addition, data-sample is used to predict pricing bonds. Based on this conclusion, this paper puts forward some suggestions about future research directions of interest rate term structure model.