Traditional Mean-Variance Optimisation (MVO) — the bedrock of portfolio construction since Markowitz in 1952 — has a structural flaw: it forces investors to pre-commit to either a target risk or a target return before optimisation begins. That arbitrary choice introduces bias and cannot guarantee a globally optimal outcome.
The Dual-Objective Optimisation Approach (DOOA) solves this by simultaneously minimising risk and maximising return — no pre-commitment required. It maps the entire universe of possible portfolio positions visually, identifies the true global optimum, and keeps return forecasting cleanly separated from the optimisation engine for full governance transparency.
Tested on an 11-ETF ASX portfolio against MVO, DOOA delivered lower volatility, higher Sharpe ratio, lower beta, and higher alpha — with more accurate underlying calculations.
A pragmatic, governance-ready evolution of classical portfolio construction.
In this paper, we sought to provide these perspectives:
1. MVO’s Fundamental Flaw Traditional MVO forces investors to pre-select either a target risk or return level before optimizing, introducing arbitrary bias and no guarantee of a truly global optimum.
2. DOOA Solves Both Simultaneously Rather than treating risk and return as sequential problems, DOOA optimizes both at the same time — mathematically identifying a single, globally optimal portfolio position.
3. Full Visibility of Possibilities DOOA visualizes up to 10,000 possible portfolio allocations, giving investors clear context on where their current portfolio sits relative to everything achievable — something MVO simply cannot offer.
4. Real-World Evidence In a live test with 11 ASX-listed ETFs, MVO’s “maximum Sharpe ratio” portfolio was shown by DOOA to not actually sit on the Efficient Frontier. DOOA found a genuinely better position with lower volatility and improved risk-adjusted returns.
5. Built for Professional Use DOOA separates return forecasting from the optimization engine, supports weighting constraints, and can integrate into institutional workflows — making it more governance-friendly and auditable than traditional MVO.
Beyond the core MVO vs. DOOA comparison, the paper also offers a concise primer on factor models for estimating expected returns — covering CAPM and the Fama-French-Carhart 5-Factor Model — giving readers practical context on how return forecasts can be fed into any optimizer.
It also includes a candid discussion on data quality and comparability, warning practitioners that differences in data frequency (daily vs. monthly returns) and risk-free rate benchmarks (3-month bill vs. 10-year bond yield) can significantly skew volatility and Sharpe ratio calculations between tools — a subtle but important real-world consideration.
Finally, it briefly touches on practical implementation, noting how DOOA can slot into quarterly rebalancing cycles, model portfolio construction, and risk reporting dashboards, making it not just a theoretical proposal but an operationally viable framework for institutional asset managers.
To explore the full analysis, data, and manager landscape in detail, download the complete whitepaper below.
