An Empirical Study on the Variation of the Treasury Bond Yield Curve Based on the Principal Component Analysis in China

According to study the yield curve of treasury bond based on principal component analysis, this thesis attempts to describe the behavior of correlated random variables in terms of a small number of uncorrelated principal components. After it is used to study the variation characteristic of the yield curve, we find that there are several factors effecting on the variation of the yield curve. Three principal components can explain about 90% of the variation of the curve, and the explanation in China is low comparing with the abroad research. We find the principal components´ change with sample period and frequency. Because the main factors are more than one parallel movement, the traditional immunization method is inefficient to manage the interest rate risk. So we should use different method to manage the different risk sources.