Estimation Bias Induced by Discrete Security Prices
建模了证券价格取整机制，分析取整对估计方差和峰度的影响，给出最大似然估计和简单修正方法，对使用高频价格数据的实证研究者有参考价值。
ABSTRACT Commonly, equilibrium security prices are modeled by continuous‐state stochastic processes, while observed prices are rounded into discrete units. This paper models the rounding mechanism and examines the probabilistic structure of the resultant rounded process. We provide accurate and simple estimates of the inflation in estimated variance and kurtosis induced by ignoring rounding. In particular, the maximum‐likelihood estimate of security price volatility using rounded prices is developed, and a simulation analysis is performed to examine the small‐sample properties of this estimator. For many practical applications, a simple correction for rounding becomes available.