Portfolio Construction

We boil down the overall attractiveness of an ETF into one number, which we call “utility.” Utility is calculated as follows:

Expected return
–  Systematic risk penalty
–  Residual risk penalty
–  Implementation shortfall (fund tracking error to benchmark index including expense ratio)
–  Bid-ask spread
–  Commission cost
–  Borrow cost (for shorts)


=  Utility

After rank-ordering the attractiveness of the ETFs by utility, we target weights for the 16 highest as follows:

  • 10.0% in #1
  • 9.5% in #2
  • 9.0% in #3
  • …..and so on down to…..2.5% in #16

at which point 100% of capital has been invested.  For our (long/short) Global Macro Portfolio, short target weights are similarly set for the ETFs with the most negative utility. This methodology has the effect of concentrating our holdings in the most attractive ETFs.  In practice, we typically establish positions up to within 3% of the target weight (maximum 10% as explained above), with actual portfolio weights usually 1%-3% below the target.  Targets are updated daily, but we do not typically trade positions unless actual weights are outside of +/- 3% relative to the current target weight.  We find that this helps to reduce unnecessary turnover.