Stock price discreteness adds noise to price series. The noise increases return variances and adds negative serial correlation to return series. Standard variance and serial covariance estimators therefore overestimate the variance and serial covariance of the underlying stock values. Discreteness-induced variance and serial covariance depend on underlying volatility and on the size of the bid/ask spread. Simple formulas for approximating the effects of discreteness on variance and serial correlation are derived and presented. The approximations, which are accurate in daily data, can be used to adjust the standard variance and serial covariance estimators.

ABSTRACT Exact small sample population moments of the standard serial covariance and variance estimators are derived under the assumptions of the Roll bid/ask spread model. Noise explains why serial covariance estimates are often positive in annual samples of daily and weekly returns. Small sample estimator bias partially explains why weekly estimates are more negative than daily estimates. Noise causes the Roll spread estimator to be severely biased by Jensen's inequality. The French‐Roll adjusted variance estimator is unbiased but noisy. Empirical tests confirm the major implications.

Exact small sample population moments of the standard serial covariance and variance estimators are derived under the assumptions of the Roll bid/ask spread models. Noise explains why serial covariance estimates are often positive in annual samples of daily and weekly returns. Small sample estimator bias partially explains why weekly estimates are more negative than daily estimates. Noise causes the Roll spread estimator to be severely biased by Jensen's inequality. The French-Roll adjusted variance estimator is unbiased but noisy. Empirical tests confirm the major implications.