Skip to main content
Track Asset allocation
25 / 26
Search the archiveLibrary
Library

2019issue C128-15

Evaluate diversification by correlation, not by the mix label

A 60/40 mix of a stock index ETF and an intermediate government-bond ETF, rebalanced annually from late July 2002, remained almost fully aligned with the stock ETF, with a reported correlation of 0.96. Changing the weights, then replacing the cash bond sleeve with a futures overlay, changed the reported equity correlation and the volatility mix.

  • A 60/40 mix of a stock index ETF and an intermediate government-bond ETF, rebalanced annually from late July 2002, remained almost fully aligned with the stock ETF, with a reported correlation of 0.96.
  • That mix showed a lower peak decline and lower annualized monthly volatility than a 100% stock-ETF holding, along with a lower compound annual growth rate.
  • Reversing the weights to 40/60 reduced drawdown, volatility, and equity correlation versus 60/40, but the reported correlation to the stock ETF remained high at 0.8.
  • A leveraged bond overlay of 40% stock ETF, 100% 10-year government-bond futures, and 50% three-month Treasury bills reported a 0.63 equity correlation and about half the stock-ETF volatility, with a higher return than the unlevered 40/60 mix.
Entries in this reading3 entries
Current trackAsset allocation55 readingsDiversification59 readingsCorrelation analysis71 readings

The mixes in the comparison

The source backtest compares a 60/40 mix of a stock index ETF and an intermediate government-bond ETF with a 100% stock-ETF holding, then reverses the weights, then replaces the cash bond sleeve with a leveraged bond overlay. A later comparison places an all-weather example next to the Wilshire 5000.

Asset allocation, Diversification, and Correlation analysis appear here as checks on weights, realized equity correlation, and how the bond instrument changes the risk mix.

What the 60/40 mix tracked

A 60/40 mix of a stock index ETF and an intermediate government-bond ETF, rebalanced annually from late July 2002 in the source backtest, remained almost fully aligned with the stock ETF. The reported correlation was 0.96.

In that same 60/40 comparison, the mix showed a lower peak decline and lower annualized monthly volatility than a 100% stock-ETF holding. It also showed a lower compound annual growth rate.

The source argues that most of a conventional 60/40 mix's volatility and risk come from equities, so the calmer bond sleeve is washed out and the whole book still tracks stocks.

Reversing the weights

Reversing the weights to 40% stocks and 60% bonds reduced drawdown, volatility, and equity correlation versus 60/40. The reported correlation to the stock ETF remained high at 0.8.

Scaling the bond sleeve with futures

A leveraged bond overlay used 40% stock ETF, 100% 10-year government-bond futures, and 50% three-month Treasury bills. In the same comparison set it produced a lower reported equity correlation of 0.63 and about half the stock-ETF volatility. It also showed a higher return than the unlevered 40/60 mix.

Bond futures plus idle cash in bills are presented as a way to scale bond exposure toward equity-like volatility without borrowing cash to buy bonds, because futures often embed a positive carry when the yield curve slopes upward.

The later all-weather comparison

In the later all-weather versus Wilshire 5000 comparison, the example mix's volatility was about two thirds that of the stock index. The mix also showed a milder worst month and worst year than the index series shown.

What the notebook actually ran

The illustrative Python notebook is hard-coded to three price series, fixed weights, and yearly reallocation. It omits commissions and slippage.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
25 of 26 in the Asset allocation track
202059-59 pp.Next on Asset allocationCost-first fund sleeves and a stock-bond mixCost-layers such as sales charges, turnover or trading costs, taxes, management fees, marketing costs, and idle cash reduce the return that reaches a mutual-fund investor.
All readings on this track · 26 readings
  1. 1990Stress-testing a futures sleeve across equity regimes
  2. 1990Staged sleeves from one equity forecast
  3. 1990Score a trade after allocation, diversification, and a drawdown limit
  4. 1993A three-sleeve regime map for equities, rates, and futures
  5. 1994Late-cycle breadth as an allocation exam
  6. 1994Ratio MACD crossover as a portfolio weight filter
  7. 1995Sizing a trend futures sleeve in a 1995 policy shift
  8. 1996A 1990s case on liquidity cycles, rank rotation, and cross-market sleeves
  9. 1997Regime-aware allocation versus binary timing
  10. 1998Hedge-fund sleeves under a hard drawdown limit
  11. 2010True diversification as portfolio construction
  12. 2012Regime-gated trend baskets and layered sizing
  13. 2013Score fund cost drag before the equity-versus-defensive mix
  14. 2013Cost-first fund mix in portfolio construction
  15. 2015Cost-first fund construction and the stock-bond mix
  16. 2016Audit the fund cost-stack, then set the sleeve-mix
  17. 2016Read fund listings through costs, then allocation, then diversification
  18. 2017A four-channel own-account book with signal-replacement
  19. 2017Listed REITs as a standalone portfolio sleeve
  20. 2017Cost-first stock and bond fund construction
  21. 2017Cost-first fund mix in a diversified portfolio
  22. 2018Mutual-fund costs come before returns in portfolio mixes
  23. 2018Cost gates before a stock-bond fund mix
  24. 2019Cost-stack and bond-equity mix around a trade
  25. 2019Evaluate diversification by correlation, not by the mix label
  26. 2020Cost-first fund sleeves and a stock-bond mix
By fieldPortfolio construction155 readingsBy year2019140 readings
All 55 readings tagged Asset allocation
Also on Asset allocation5 readings
1990Stress-testing a futures sleeve across equity regimes1990Score a trade after allocation, diversification, and a drawdown limit1993A three-sleeve regime map for equities, rates, and futures1995Sizing a trend futures sleeve in a 1995 policy shift1996A 1990s case on liquidity cycles, rank rotation, and cross-market sleeves