An optimal investment policy model for the long term growth of expected uti
lity of wealth is considered. The utility function is HARA with exponent -i
nfinity < gamma < 1. The problem can be reformulated as an infinite time ho
rizon, risk sensitive control problem. Then the dynamic programming equatio
ns for different HARA exponents and different policy constraints are studie
d. We obtain some estimates for the solution of each equation. This can be
used to derive an optimal policy with some interesting properties.