CAJ | 학술논문

本文在概述国债利率期限结构模型的基础上，针对当前被发达国家广泛采用的NS和SV模型所存在的不足，通过扩展指数多项式的方法，构建出NSM模型。为了更好地估算利率期限结构模型中的参数，本研究针对目标函数优化求解，经分析比较多种优化算法后，确定选用GRG2非线性最优化算法。通过使用上海证券交易所2005．1．4-2007．11．30的国债每日交易数据对NS、SV、NSM三个模型的实证分析比较，表明NSM模型不仅保留了NS模型的经济含义，克服了SV模型参数估计依赖初值的缺点，能够反映出利率曲线多峰的情况；而且其在拟合精度、价格误差等多项指标上均优于NS模型和SV模型，并具有良好的适应性和稳健性，能够满足我国当前的国债市场需要。
본문재개술국채리솔기한결구모형적기출상，침대당전피발체국가엄범채용적NS화SV모형소존재적불족，통과확전지수다항식적방법，구건출NSM모형。위료경호지고산리솔기한결구모형중적삼수，본연구침대목표함수우화구해，경분석비교다충우화산법후，학정선용GRG2비선성최우화산법。통과사용상해증권교역소2005．1．4-2007．11．30적국채매일교역수거대NS、SV、NSM삼개모형적실증분석비교，표명NSM모형불부보류료NS모형적경제함의，극복료SV모형삼수고계의뢰초치적결점，능구반영출리솔곡선다봉적정황；이차기재의합정도、개격오차등다항지표상균우우NS모형화SV모형，병구유량호적괄응성화은건성，능구만족아국당전적국채시장수요。
On the basis of summarization of researches on term structure of interest rate of treasury bond, the paper constructs a new model named NSM （ Nelson-Siegel Modified Model） by extending exponential polynomial to overcome the deficiencies of the popular NS model （ Nelson-Siegel model） and SV model （ Nelson-Siegel with Svenssen extended model）. For the better estimation of the model parameters, GRG2 optimization algorithm is selected after contrasting against various optimization algorithms. By taking the data of treasury bond daily traded in Shanghai Stock Exchange from January 4^th, 2005 to November 30^th, 2007, the authors empirically compares NS, SV and NSM model, and the results indicate that NSM model inherits the NS model＇s economic significations and behaves less dependent to initial values of model parameters than SV model. Furthermore, NSM model has better fitness and less pricing errors than NS and SV model, so it is believed that NSM model is commendably applicable to the treasury bond market in China.