Abstract
In recent years skewness has become a much-discussed factor in financial research, and many studies/models involve the skewness of various financial variables. This paper (i) points out the universal neglect in the finance literature of skewness' sampling error and its significant consequences; (ii) presents a simple approach for roughly constructing a confidence interval for skewness estimated from lognormal populations; (iii) points out directions of further research for developing a comprehensive approach for estimating skewness reliably.