Sammendrag
A numerical strategy for solving low-dimensional Bellman equations through the traditional backwards recursion is formulated. A simple error analysis suggests that the approach handles many multi-period portfolio selection problems, and a number of examples confirm this experimentally. Minimum downside risk procedures are studied and it is demonstrated how multi-period efficient frontiers can be calculated for such criteria. A closing example examines the impact of heavy-tailed distributions on optimal, multi-period risk.