Nonparametric estimation of diffusion process: a closer look

A Monte Carlo simulation is performed to investigate the finite sample properties of a nonparametric estimator, based on discretely sampled observations of continuous-time Ito diffusion process. Chapman and Pearson (2000) studies finite-sample properties of the nonparametric estimator of Aýt-Sahalia (1996) and Stanton (1997) and they find that nonlinearity of the short rate drift is not a robust stylized fact but it’s an artifacts of the estimation procedure. This paper examine the finite sample properties of a different nonparametric estimator within the Stanton (1997)’s framework.