We formalize a modified present value rule for portfolio decisions und
er uncertainty based on a state-dependent valuation with opportunity c
osts. This decision rule is shown to select a unique portfolio which a
lso maximizes the expected logarithmic utility of payoffs.