Nonparametric Method And Its Application In T Model

In this paper, we basically intend to examine the validity of a particular valuation model, T model (Tony Step,1987), which claimed to explain90%of total return of a stock (with retrospective data). We applied data of1600companies in China’s stock market in the year2008, to traditional linear models as well as nonparametric models to examine the validity of T model in China’s stock market.We have found that two of the three variables(growth factor,value factor and cash flow factor) introduced in T model are significant in explaining stock returns, while one variable(value factor) of these two(value factor and growth factor) exhibits significant linear relationship with stock returns, as predicted by the T model. We also find that local linear LSCV could not fully detect all the linear components. Local constant LSCV founded that in the T model cash flow factor is of none linear correlations between variables, i.e. irrelevent variables.Finally, in the sixth chapter, we proceed from the inherent economic logic of the T model in order to get some market-seeking evidence. Although it is just a simple and single perspective, this attempt confirms the conclusions of our empirical analysis. And it is a useful complement for the full text of the study.