State-dependent asset allocation

The traditional mean-variance solution to the asset allocation decision requires estimates of long-run risk and expected return, and produces an efficient frontier of risk/return choices. The authors demonstrate that if the same long-run estimates are generated by combining separate states of the world with differing estimates of risk and return, the investor can use this information to alter the benchmark portfolio allocation in an efficient way in each state. When this state-dependent information is used to alter the investor's benchmark allocation, an expanded opportunity set of risk/return possibilities can be created. Capturing this expanded set requires the investor to shift the portfolio allocation in response to the state-dependent risk/return estimates, however. This investment opportunity set cannot be captured by holding a static mix in the portfolio across all states.