A prior blog post shows what happens when no constraints are placed on an optimized portfolio. To follow up on the Swensen Six post, assume we begin with the asset allocation plan recommended by Swensen and then optimize the six ETFs using reasonable constraints. The following efficient frontier graph show the Swensen allocation (diamond dot) falling well short of the optimized portfolio (red dot).
Efficient Frontier: While the Swensen Six, with the basic allocation, comes close to the efficient frontier, we can improve on the potential return. Those changes will show up in the second screen shot.
Weights: The following slide shows the weights assigned to the optimized Swensen Six. A token 1% is allocated to TLT and that is related to the constraint I placed on that ETF. VTI receives a full allocation based on the constraint.
Rankings: U.S. Equities (VTI) shows up as the top ranked ETF from this Swensen Six portfolio. While TLT had a great run not long ago, it has not performed well over the last three months and is relegated to position number six. Keep in mind these are relative rankings and there are other ETFs that have performed even better than VTI. VFH is such an example.
Buy-Hold-Sell Recommendations: Based on the initial holdings assigned to each ETF, the following buy-hold-sell table indicates potential changes. No sales are called for. I would only add shares of VTI as VEU and TIP are not significant additions.
With this optimization analysis, one can increase the performance probability by making a few changes from the original Swensen asset allocation plan.