实证资产定价中的时间序列回归多用于估计资产的因子暴露(beta值)。此外,时间序列分析可以通过GRS检验来考察资产收益率是否存在未被风险因子(risk factor)解释的异象(anomaly)。
本文根据Bail et al.的著作Empirical Asset Pricing编写相关程序,时间序列回归的模块是EAP.time_series_regress,下面将对该模块进行详细介绍。本文的Package已发布于Github:
Github: GitHub - whyecofiliter/EAP: empirical asset pricing
time_series_regress
class TS_regress()
This class is designed for time series regression,
$$
r_{i,t} = \beta_if_t + \epsilon_{i,t}
$$
to obtain the beta for each asset.
def __init__(self, list_y, factor):
This function initializes the object.
input :
list_y : The return matrix with i rows and t columns.