Safety-first portfolio selection (Q5918317)

This paper examines the relationship between safety-first portfolio selection and mean-variance optimization. Two different investment philosophies proposed in 1952 have inspired many modern investment studies. Namely, mean-variance optimization (modern investment theory; MVT) identifies the lowest risk for a given target return, and in mean-variance utility (MVU) approach, investors use the risk-aversion coefficient to determine their portfolio choices. Then portfolio optimization with mental accounts (POMA), which connects the MVT and MVU settings to a behavioral portfolio-style problem, was introduced. In this paper the authors present a POMA closed-form solution with the threshold return relative to the orthogonal portfolio on the mean-variance frontier. They employ the \(t\)-statistics of stepwise regression coefficients to sort the suboptimal subset of assets. Finally, from the VaR (value-at-risk) perspective, they examine mean-variance efficiency testing to determine whether the frontier spanned by a complete set of assets coincides with a subset of assets.