Asset Allocation: Working With Optimizer

by: Lowell Herr

(Click to enlarge) Engines waiting to asset ships through the Panama Canal.

How does one combine the power of the Dashboard and the Hoadley optimizer when it comes to setting up an asset allocation plan for a portfolio? ITA readers are familiar with the Dashboard as it is shown every time a portfolio is updated. In this post I will explain how one might combine the Dashboard and Hoadley optimizer. Since the Hoadley optimizer is confined to nine (9) asset classes, I’ve chosen to break the market down into the following asset classes: Large-Cap, Mid-Cap, Small-Cap, Bonds, Developed International, REITs, Emerging Markets, Commodities, and Precious Metals. This is a major reduction from the 17 or 18 asset classes we use with the Dashboard. Here is an example of the percentages one might assign to each of these nine asset classes.

  1. Large – Cap (NYSE:LC) – 12%
  2. Mid-Cap (MC) – 10%
  3. Small-Cap (SC) – 8%
  4. Bonds – 15%
  5. Developed International – 15%
  6. REITs – 20%
  7. Emerging Markets – 10%
  8. Commodities – 5%
  9. Precious Metals – 5%

This allocation plan gives us the 30% in U.S. Equities and 20% in REITs recommended by David Swensen. We are a little light on bonds and a little heavy with our international markets allocation. Also, Swensen did not include commodities and precious metals in his “Swensen Six” portfolio.

When using an optimizer such as the Hoadley, it makes good sense to include constraints or collars around each asset class. We do the same when we assign a threshold percentage to each asset class in the Dashboard. I’ve been experimenting with 40% constraints around each of the nine asset classes. For example, Large-Cap stocks are targeted at 12%. A 40% constraints works out to be 0.40 x 0.12 = 4.8% or rounded to 5%. We can set constraints for LC to be 17% on the high side and 7% on the low side. If those thresholds seem too large, then round down to 4%. I prefer to provide a little more room for the optimizer to do its work so long as the end portfolio is not completely irrational.

Commodities and Precious Metals work out to 0.05 x 0.40 = 0.02 or 2%. These two asset classes can be as high as 7% each or as low as 3%. I’ll make an exception with these two asset classes as there are times when one does not wish to hold any shares. Therefore, I’ll most likely set my limits or constraints to range from 0% to as high as 7% and let the optimizer do its calculations.