We apply Principal Component Analysis for zero-coupon Treasury bonds to get level, slope, and curvature series. We model these as autoregressions of order 1, and analyze their innovations. For slope, but not for level and curvature, dividing these innovations by the Volatility Index VIX made for Standard \& Poor 500 makes them closer to independent identically distributed normal. We state and prove stability results for bond returns based on this observation. We chose zero-coupon as opposed to classic coupon Treasury bonds because it is much easier to compute returns for these.

翻译：我们采用主成分分析法对零息国债进行处理，得到水平、斜率和曲率序列。将这些序列建模为一阶自回归过程，并分析其新息项。对于斜率序列（水平与曲率序列则不然），将其新息项除以标普500波动率指数VIX后，可使其更接近独立同分布的正态分布。基于这一观测结果，我们提出并证明了债券收益率的稳定性定理。本研究选择零息国债而非传统附息国债，是因为前者的收益率计算更为简便。