Portfolio theory proposes various strategies that use available information and forecasting techniques to seek better performance than a portfolio that is simply diversified broadly. Most strategies present a variety of applicable mean-variance statistics, but few resources exist to determine the appropriateness of specific measures, the potential differences in outcomes or decisions resulting from the use of particular measures, or the aggregate affect the choice might have on portfolio strategies that likely will be used. This paper provides a survey of mean-variance statistics used to summarize periodic stock returns, for cases in which the practitioner can choose between similar, but different mean-variance statistics. The paper lends itself to demonstrations of real-world problems in applied finance and the appendices include both an exemplary spreadsheet application as well as an instructional guide for teaching the material.
Available at: http://works.bepress.com/ronald-mackinnon/4/