Font Size: a A A

Resesrch Of Constructing Yield Curves By Bayesian Spline Smoothing

Posted on:2016-01-18Degree:MasterType:Thesis
Country:ChinaCandidate:X R ChenFull Text:PDF
GTID:2180330461975838Subject:Statistics
Abstract/Summary:PDF Full Text Request
Constructing the bond yield acre has many important applications. Such as in the bond valuation, in providing significant reference for the research and the policy’s draft-ing of macro-economy. In recent years, the bond market in our country has an unprece-dented development, but compared with the market of Europe and America, it is still not perfect in many aspects, especially in the construction of the yield acre. The third plenum of 18th CPC central committee pointed out that it is of great importance to carry forward market-boseh reform of interest rates and sound the Treasury yield acre reflecting the relationship between supply and demand. The main work of the article is using the Bayesian Smooth spline model to fit the yield acre and comparing it with other constructing methods of the bond yield acre.First, we will introduce the development and scale of china’s bond market, discuss the importance of the construction of the bond yield acre, and point out the question we study.Second, we will instruct kinds of constructing methods of the yield acre, such as polynomial spline estimation, B-spline method, smooth spline method and NSS method.Polynomial spline estimation is the earliest spline method to construct the bond yield acre. On the basis of Polynomial spline estimation, scientists got the B-spline method. Based on the normal spline estimation which uses the least square method, the smooth spline method adds a penalty term describing the roughness of the functions to balance the fitting precision and smoothness of the estimation. NSS method has economic meaning, whose characteristic is concise.The paper also introduces the Bayesian smooth spline model.It regards the smooth parameters as a random variabale.Finally, based on the national debt and the bond of policy banks in 2014, we use Bayesian method, smooth spline method and B-spline method for fitting. With the help of different indexs for fitting superiority, we get that Bayesian method is prior to smooth spline method and B-spline method. Another main result of this paper is that Bayesian method needs further improvement as the bond whose remaining period is more than 10 years and the lower rated credit bond are both inactive.
Keywords/Search Tags:Bayesian method, smoothing spline estimation, yield curve, B-spline, polynomial spline
PDF Full Text Request
Related items