Tilted Nonparametric Estimation of Volatility Functions With Empirical Applications

摘要: This article proposes a novel positive nonparametric estimator of the conditional variance function without reliance on logarithmic or other transformations. The is based an empirical likelihood modification conventional local-level regression applied to squared residuals mean regression. shown be asymptotically equivalent local linear in case unbounded support but, unlike that estimator, restricted nonnegative finite samples. It fully adaptive unknown function. Simulations are conducted evaluate finite-sample performance estimator. Two applications reported. One uses cross-sectional data and studies relationship between occupational prestige income, time series Treasury bill rates fit total volatility continuous-time jump diffusion model.