Maximum Diversification Portfolio
A Maximum Diversification Portfolio is an optimization-based construction methodology that allocates weights to maximize the diversification ratio — the ratio of the weighted-average asset volatility to overall portfolio volatility — resulting in a portfolio that exploits diversification benefits most fully relative to the sum of individual asset risks.
The Maximum Diversification approach was formalized by Choueifeng and Coignard in 2008 as an alternative to both cap-weighting and minimum variance construction. Its central insight is that traditional diversification strategies often fail to fully exploit the risk-reduction potential of low correlations between assets because they do not explicitly optimize for diversification as an objective. The diversification ratio captures how much the portfolio's realized volatility is reduced relative to a naive equal-weighting of each asset's standalone risk, and maximizing this ratio pushes capital toward assets and asset combinations that provide the greatest correlation-based risk reduction.
Mathematically, the diversification ratio is defined as the sum of each asset's weight multiplied by its individual volatility, divided by the portfolio's total standard deviation. If all assets were perfectly correlated, the diversification ratio would equal 1 — diversification provides no benefit. As correlations decrease, the denominator falls relative to the numerator and the ratio rises above 1. The maximum diversification portfolio finds the weights that push this ratio to its highest achievable value for a given asset universe.
The maximum diversification objective tends to favor assets with high individual volatility and low correlations with other portfolio components over assets that are simply low-volatility on their own. This contrasts with the minimum variance portfolio, which mechanically gravitates toward lower-volatility assets. In practice, the maximum diversification approach often produces more heterogeneous portfolios that span different risk factors and sectors more evenly, particularly compared to minimum variance portfolios that can become heavily concentrated in defensive sectors.
The strategy performs best when correlation structures are stable and when the diversification benefits it is designed to capture actually persist through time. Like all covariance-based optimization approaches, it is sensitive to correlation estimation error — correlations that appear low in normal market conditions often spike dramatically during market dislocations, temporarily destroying the diversification benefits the portfolio was constructed to exploit.
The maximum diversification concept has influenced factor-based portfolio construction more broadly. The underlying logic — that returns to capital should be measured relative to the risk actually contributed to the portfolio rather than on a standalone basis — informs risk-parity approaches, where each asset class contributes an equal amount to total portfolio risk rather than receiving an equal capital allocation. These related approaches all share the goal of building portfolios that are more genuinely diversified than market-capitalization-weighted benchmarks.