HOMEINVESTMENT SERVICESRESEARCHCLIENT SERVICESABOUT NFA

Newsletters|Abstracts|Articles| History of NFA's Research |Resampling Tested

New Frontier Advisors began when Richard Michaud's research led him to solve a problem concerning the practical implications of the efficient frontier.  Pioneered by Markowitz in 1959, the efficient frontier represents all optimal portfolios--the portfolios that have the maximum return or least risk for a set risk or return. The efficient frontier has been the cornerstone of modern portfolio theory for half a century and is the basis of commercially available optimizers. It is also theoretically correct, but there are numerous problems with putting the efficient frontier into practice.

As Richard Michaud pointed out in Efficient Asset Management (Harvard Business School Press, 1998), the problem with the efficient frontier is that the risk and return numbers that are used to calculate it are estimates. Classic optimization ignores the uncertainty of these inputs, resulting in unintuitive, unstable, poor performing portfolios.

Richard and Robert Michaud, two of the founders of New Frontier Advisors, developed a variant of the classic efficient frontier--the Resampled Efficient Frontier™. NFA’s resampling process acknowledges that estimates are not facts. It produces simulated returns and optimization inputs that are statistically consistent with the original estimates. The results of this process are used to compute simulated efficient frontiers. Each simulated efficient frontier may be very different from the original efficient frontier. NFA’s technique averages these simulated efficient frontier portfolios to obtain the resampled frontier. As a result, the resampled allocations are less extreme, more intuitive, and the risk estimations are more reliable. After testing, even Markowitz agreed that resampling worked better than the classical optimization process (Pensions & Investments, December 22, 2003).

The breakthrough to the Resampled Efficient Frontier was just the beginning of NFA's contributions. Since that time, New Frontier has continued to develop practical investment tools and make recommendations regarding investment practice, building on the work of those around them. Some highlights include the Rebalancing Rule and post-optimization tools, such as improved estimates using advanced Bayesian statistics. Recently, NFA has presented research on common equity optimization errors.

For details on resampling, rebalancing, and other NFA innovations, please explore our publications.