Portfolio Construction: When Optimization Helps and When It Fails
Summary
The document challenges claims that sophisticated diversification and portfolio optimization methods reliably outperform simple weighting schemes. It describes how optimization can be sensitive to uncertain Sharpe ratio and correlation estimates, while volatility is comparatively easier to estimate. Under those assumptions, minimum variance or maximum diversification may be reasonable when expected returns are unknown; inverse volatility may suit cases where correlations are also unknown, and equal weights may be a sensible fallback when little can be predicted.
It compares equal weighting, market capitalization weighting, and equal sector risk, noting that each can have concentration at the sector or firm level. It also warns that low correlations may not protect a portfolio in a crisis and that diversification methods can create unintended small-cap or value tilts. The document offers assertions and a few portfolio allocation figures, but no detailed methodology or broad empirical evidence. Its conclusions are therefore best read as arguments about estimation risk and portfolio assumptions, not universal proof that one weighting method is superior.
Key ideas
- Optimization outcomes can change sharply when estimated Sharpe ratios or high correlations change slightly.
- Volatility is presented as more predictable than Sharpe ratios, with correlations falling between them.
- The appropriate weighting method depends on which return and risk inputs an investor believes can be estimated.
- Equal weighting, market capitalization weighting, and sector risk weighting can all produce concentration.
- Low historical correlations may fail to provide protection during a crisis.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.