Maximum Diversification Portfolio (MDP) Lab

Yves Choueifaty & Yves Coignard (2008) Diversification Ratio (DR) optimization & equi-correlation risk budgeting

MDP Diversification Ratio
--x
DR = Weighted Vol / Portfolio Vol
MDP Portfolio Volatility
--%
Realized portfolio standard deviation
Min-Var DR Ratio
--x
Markowitz CLA Diversification Ratio
Max Asset Concentration
--%
Largest single asset weight in MDP
Asset Class Volatilities (%)
%
%
%
%
%
Multi-Strategy Allocation & Equi-Correlation Comparison
Asset Class Vol MDP MinVar ERC HRP Inv Vol 1/N MDP Corr
Notice how the MDP maximizes the Diversification Ratio by balancing individual asset volatility and correlation, equalizing constituent correlation to the portfolio.
Diversification Ratio vs. Portfolio Volatility Space

Scatter plot comparing the Diversification Ratio (DR) and total portfolio volatility across all 6 asset allocation strategies. The MDP achieves the absolute theoretical peak of the DR dimension.

Verification of Choueifaty's Equi-Correlation Theorem

Constituent asset correlation with the portfolio: $\rho(R_i, R_p) = \frac{(\Sigma w)_i}{\sigma_i \sigma_p}$. In the MDP, all active holdings exhibit an identical correlation to the overall portfolio ($\rho = 1 / DR$).

Portfolio Weights Across Optimization Frameworks

Detailed comparison of weights assigned by the Most Diversified Portfolio vs. Markowitz Minimum Variance, Equal Risk Contribution, and Hierarchical Risk Parity.

Matrix 1: MDP Diversification Ratio (DR)
Tech-Equity Correlation (Rows) vs. Bond-Equity Correlation (Columns)
ρ(SPY, QQQ) \ ρ(SPY, TLT) -0.60 -0.30 0.00 +0.30 +0.60
Green (≥1.70x) Peak diversification benefit • Yellow Compressed DR under high cross-asset correlation.
Matrix 2: MDP Treasury Weight (%)
Cross-Asset Correlation Shift (Rows) vs. Equity Vol Shock (Columns)
Corr Shift \ Vol Shock 0.8x 1.0x 1.2x 1.5x 2.0x
Blue (≥35%) High Treasury flight-to-safety allocation • Yellow Compressed weighting.

Institutional Quantitative Foundations: Maximum Diversification

Introduced by Yves Choueifaty and Yves Coignard (2008) in The Journal of Portfolio Management ("Toward Maximum Diversification"), the Most Diversified Portfolio (MDP) formalizes a mathematically rigorous measure of portfolio diversification: the Diversification Ratio (DR).

1. The Diversification Ratio (DR)

$$DR(w) = \frac{\sum_{i=1}^N w_i \sigma_i}{\sqrt{w^T \Sigma w}} \ge 1$$ DR measures the ratio of the portfolio's weighted average constituent volatility to its net portfolio volatility. The difference represents risk eliminated purely through diversification.

2. Equi-Correlation Property

$$\rho(R_i, R_{\text{MDP}}) = \frac{(\Sigma w)_i}{\sigma_i \sigma_{\text{MDP}}} = \frac{1}{DR_{\text{MDP}}} \quad \forall i : w_i > 0$$ In the MDP, every single active constituent has the exact same correlation to the portfolio!

3. Sharpe Ratio Equivalence

If all assets have identical Sharpe ratios ($SR_i = \text{constant}$), the Most Diversified Portfolio is mathematically identical to the Markowitz Tangency (Maximum Sharpe Ratio) portfolio, achieved without return forecasts!

Why Maximize Diversification?

In practice, expected returns are notoriously noisy and difficult to forecast, whereas volatilities and correlations can be estimated with much greater statistical accuracy. By maximizing the Diversification Ratio, institutional asset allocators harvest the "only free lunch in finance" (pure diversification alpha) without suffering from the estimation errors inherent in traditional mean-variance optimization.

Institutional Mastery Assessment: Maximum Diversification

1. What does the Diversification Ratio (DR) mathematically quantify?

2. Under what condition is the Most Diversified Portfolio (MDP) identical to the Markowitz Tangency Portfolio?

3. What does Choueifaty's Equi-Correlation Theorem demonstrate about constituents in the MDP?

4. Why is the MDP advantageous over traditional Mean-Variance Optimization in real-world asset management?

Frequently Asked Questions

The Maximum Diversification Portfolio (MDP), introduced by Yves Choueifaty and Yves Coignard in 2008, is an asset allocation strategy that maximizes the Diversification Ratio (DR). It seeks to extract the maximum possible risk reduction from cross-asset non-correlation without requiring expected return forecasts.

The Diversification Ratio is the ratio of the portfolio's weighted average constituent volatility to its total portfolio volatility: DR(w) = (w^T * sigma) / sqrt(w^T * Sigma * w). By definition, DR >= 1. A higher DR indicates superior diversification and greater cancellation of idiosyncratic risk.

A fundamental mathematical theorem of the Most Diversified Portfolio (Choueifaty et al., 2013) states that for all active assets with positive weight (w_i > 0), their correlation with the overall MDP portfolio is identical and strictly equals the reciprocal of the Diversification Ratio: rho(R_i, R_MDP) = 1 / DR.

If all individual assets in the investment universe have identical Sharpe ratios (meaning each asset offers the same expected excess return per unit of volatility), the Most Diversified Portfolio is mathematically identical to the Markowitz Tangency (Maximum Sharpe Ratio) portfolio.

Minimum Variance minimizes portfolio variance regardless of individual volatilities, frequently concentrating into low-volatility assets. Risk Parity (ERC) equalizes marginal risk contributions across all assets. MDP explicitly penalizes cross-asset correlation, allocating more heavily to assets that provide genuine diversifying power even if they have higher standalone volatility.
Related Quantitative Portfolio Strategies
Hierarchical Risk Parity

Machine learning tree clustering and recursive bisection without matrix inversion.

Risk Parity & ERC Lab

Model Equal Risk Contribution (ERC) portfolios and 60/40 risk illusions.

Black-Litterman Model

Blend subjective investor views with market equilibrium priors.

Resampled Frontier Lab

Michaud Monte Carlo resampling and parameter estimation uncertainty.