Portfolio optimizers enhance investment performance by optimally using information. However, optimizers frequently create portfolios that perform poorly compared to equal weighted portfolios. Poor estimates and bad luck are not the primary reasons for these disappointments, though they are often blamed. The essential problem is that traditional portfolio optimizers improperly handle the uncertainty inherent in all investment information. Optimizers assume that optimization inputs are accurate and 100% certain, but investors are neither 100% accurate nor 100% certain of future estimates of risk and return.
New Frontier's approach to investing focuses on dealing with investment uncertainty. Michaud optimization calculates portfolios spanning the many ways that assets and markets may perform based on the optimization inputs. Implementing this procedure requires setting a range for the multiple risk and return scenarios, which depends inversely on the manager’s confidence in the reliability of risk and return estimates.
Exposing the Forecast Confidence parameter in the Optimizer software permits investors to tailor the optimization to their confidence in their estimates. The optimizer uses the forecast confidence level to calculate the Michaud frontier most effectively. With a high forecast confidence level, the Michaud Resampled Efficient Frontier™ more closely resembles the classical efficient frontier, since the classical optimization process has total confidence in the estimates. The rebalance test also applies smaller statistical regions at higher forecast confidence levels. Conversely, with a low level of forecast confidence, the rebalance test assumes larger regions of equivalence.