A risk-averse US investor adjusts the shares of a portfolio of short-t
erm nominal domestic and foreign assets to maximize expected utility.
The optimal strategy is to respond immediately to all new information
which arrives weekly. We develop a model to estimate the cost of optim
izing less frequently and find that it is generally very small. For ex
ample, if the investor adjusts portfolio shares every three months, an
average expected utility loss of 0.16 per cent p.a. is incurred. Henc
e, slight opportunity costs of frequent optimization may outweigh the
benefits. This result may help explain forward discount bias.