Incomplete risk-preference information in portfolio decision analysis

Portfolio decision analysis models support decisions on the allocation of resources among assets with uncertain outcomes (e.g., investments, projects or stocks). These require information maker’s risk-preferences which can be difficult to obtain in practice. Stochastic dominance criteria show promise this regard as they compare portfolios without exact specification risk-preferences, but current literature lacks practical approaches for generating efficient frontier, i.e., set those that are not stochastically dominated by any other portfolio. We address gap developing identify sets sense second- third-order stochastic (SSD, TSD). provide novel insights into composition belonging frontier by, e.g., identifying included all portfolios. Moreover, identification makes it possible utilize additional further reduce admissible portfolio alternatives, and analyze implications has amount capital should allocated each individual asset. illustrate usefulness these applications project selection financial diversification.