Risk Parity Explained: How Bridgewater’s All Weather Works

A traditional 60/40 portfolio allocates 60% of capital to stocks and 40% to bonds — but roughly 90% of its risk comes from the stock allocation. Risk parity flips the question: instead of allocating capital equally, allocate risk equally across asset classes, then use leverage to scale to your target return. It is one of the most influential portfolio construction ideas of the last three decades.

1. The Problem with 60/40

The 60/40 stock-bond portfolio has been the default institutional allocation for decades. The logic is straightforward: stocks provide growth and bonds provide stability. By mixing the two, you get a return that is higher than bonds alone and less volatile than stocks alone.

But there is a hidden concentration problem. Stocks are approximately 3 to 4 times more volatile than investment-grade bonds. The S&P 500 has historically exhibited annualized volatility of roughly 15–20%, while the Bloomberg US Aggregate Bond Index has run around 4–6%. When you combine assets with very different volatilities, the more volatile asset dominates the portfolio’s risk profile.

In a 60/40 portfolio, stocks contribute approximately 90% of the total portfolio risk, while bonds contribute only about 10%. This means the portfolio’s returns are overwhelmingly driven by what happens in the stock market. In a year when stocks fall 30%, the 40% bond allocation provides some cushion, but not nearly enough to offset the equity drawdown. The portfolio is effectively a stock portfolio with a bond garnish.

The Math Behind the Concentration

Portfolio variance for a two-asset portfolio is:

σ²p = w²sσ²s + w²bσ²b + 2·ws·wb·ρ·σs·σb

With stocks at 16% vol, bonds at 5% vol, weights of 60/40, and a stock-bond correlation of roughly 0 (which held for much of the 2000–2020 period), the stock contribution to portfolio variance is:

Stock variance contribution: (0.60)² × (0.16)² = 0.36 × 0.0256 = 0.00922
Bond variance contribution:  (0.40)² × (0.05)² = 0.16 × 0.0025 = 0.00040
Total variance:              0.00962

Stock share of risk: 0.00922 / 0.00962 = 95.8%

Stocks account for nearly 96% of the portfolio’s variance despite being only 60% of the capital. This is the fundamental insight that motivated risk parity.

2. The Risk Parity Solution

Risk parity inverts the allocation problem. Instead of asking “how much capital should I put in each asset class?”, it asks “how much risk should each asset class contribute?” The answer: equal risk contribution.

In a simple two-asset risk parity portfolio with stocks and bonds, you want each asset to contribute 50% of total portfolio risk. Since bonds are much less volatile, you need to hold more bonds (in capital terms) to make their risk contribution equal to stocks. A rough calculation: if stocks have 16% vol and bonds have 5% vol, the risk parity weights are inversely proportional to volatility:

w_stocks ∝ 1/σ_stocks = 1/16 = 0.0625
w_bonds  ∝ 1/σ_bonds  = 1/5  = 0.2000

Normalize: w_stocks = 0.0625 / (0.0625 + 0.2) = 23.8%
           w_bonds  = 0.2000 / (0.0625 + 0.2) = 76.2%

This is the inverse-volatility weighting approach, which is an approximation of equal risk contribution (it is exact when correlations between all assets are equal). The result is a portfolio that is roughly 24% stocks and 76% bonds by capital. This portfolio has balanced risk, but its expected return is lower than 60/40 because it holds so much more in lower-returning bonds.

Enter Leverage

The risk parity solution to the lower expected return is leverage. You borrow at the short-term interest rate and scale up the entire balanced portfolio to achieve the return target. If the unlevered risk parity portfolio has 6% expected return and 6% volatility, and you want 10% expected return, you lever it approximately 1.7x, resulting in roughly 10% expected return and 10% volatility.

The key insight is that a levered balanced portfolio should have a better Sharpe ratio than an unlevered concentrated portfolio. The Sharpe ratio measures return per unit of risk. By equalizing risk contributions, you maximize diversification, which should improve the Sharpe ratio. Leverage then scales the portfolio up or down to any desired risk level without changing the Sharpe ratio (in theory, before accounting for leverage costs).

3. The Academic Foundation

The formal mathematical framework for equal risk contribution portfolios was laid out by Maillard, Roncalli, and Teiletche in “The Properties of Equally Weighted Risk Contribution Portfolios”, published in the Journal of Portfolio Management, Vol. 36, No. 4, pp. 60–70 (2010). They defined the risk contribution of asset i as:

RC_i = w_i × (Σw)_i / σ_p

where (Σw)_i is the i-th element of the vector obtained by multiplying the covariance matrix by the weight vector, and σ_p is the portfolio standard deviation. The equal risk contribution (ERC) portfolio is the set of weights where RC_i = RC_j for all asset pairs.

Maillard, Roncalli, and Teiletche proved that the ERC portfolio lies between the minimum-variance portfolio and the equal-weight portfolio on the efficient frontier. They also showed it has no closed-form solution in general (when the covariance matrix is not diagonal) and must be found numerically. The optimization can be expressed as minimizing the variance of risk contributions, subject to the constraint that weights sum to one and are non-negative.

4. Bridgewater’s All Weather Portfolio

Ray Dalio and Bridgewater Associates created the All Weather portfolio in 1996. It is the most famous implementation of risk parity principles, though Dalio framed it not as “equal risk contribution” in the mathematical sense but as balancing risk across economic environments.

Dalio’s framework divides the economic landscape into four quadrants based on two dimensions: growth (rising or falling relative to expectations) and inflation (rising or falling relative to expectations). Each quadrant favors different asset classes:

All Weather allocates equal risk budget to each quadrant. Since you don’t know which environment will prevail at any given time, the portfolio is designed to perform acceptably in all four rather than spectacularly in one.

Approximate All Weather Allocation

While Bridgewater has never published the exact All Weather portfolio, Tony Robbins disclosed an approximate version (attributed to Dalio) in his 2014 book Money: Master the Game. The approximate allocations are:

• 30% stocks (US equities)
• 40% long-term bonds (20+ year Treasuries)
• 15% intermediate-term bonds (7–10 year Treasuries)
• 7.5% gold
• 7.5% commodities (broad commodity index)

The heavy bond weighting (55% in total) reflects the need to balance the higher volatility of stocks. The gold and commodity allocations provide inflation protection. In practice, Bridgewater’s actual All Weather fund uses leverage to scale this balanced allocation to a target risk level.

5. A Simpler Predecessor: The Permanent Portfolio

The idea of balancing across economic environments predates Bridgewater. Harry Browne proposed the Permanent Portfolio in his 1981 book Fail-Safe Investing. The allocation is strikingly simple:

• 25% stocks — for prosperity
• 25% long-term bonds — for deflation
• 25% gold — for inflation
• 25% cash (T-bills) — for recession

The Permanent Portfolio does not use leverage and does not explicitly target equal risk contribution. It uses equal capital weights across four assets chosen to cover the four economic states. Despite its simplicity, it has delivered remarkably stable returns over long periods, with much lower drawdowns than equity-heavy portfolios. It can be viewed as a conceptual precursor to risk parity: the intuition (balance across economic regimes) is the same, even if the mathematical implementation is less sophisticated.

6. Performance: The Historical Track Record

Risk parity strategies and All Weather-style portfolios have historically delivered competitive returns with significantly reduced drawdowns compared to traditional stock-heavy allocations. The performance case rests on several periods:

2008 Global Financial Crisis

This was risk parity’s defining moment. During 2008, the S&P 500 fell approximately 37%. A 60/40 portfolio lost roughly 22%. The All Weather portfolio, by contrast, is estimated to have lost approximately 3–4%. The reason: while stocks collapsed, long-term Treasury bonds rallied sharply as the Federal Reserve cut interest rates and investors fled to safety. Because risk parity portfolios held large bond positions, the bond gains substantially offset the equity losses.

Long-Run Returns

Backtests of risk parity strategies over multi-decade periods (subject to the usual caveats about backtesting) suggest annualized returns in the range of 8–10% with annualized volatility of approximately 10–12%. This compares to roughly 10% return and 15–16% volatility for the S&P 500, and roughly 8–9% return and 10% volatility for a 60/40 portfolio. The Sharpe ratio of risk parity strategies has historically exceeded that of both the S&P 500 and 60/40.

The COVID Crash (March 2020)

Risk parity strategies performed well during the initial COVID-19 market shock in March 2020. Stocks fell sharply, but Treasury bonds rallied as the Fed slashed rates to zero and launched massive quantitative easing. The negative stock-bond correlation that risk parity depends on held during this episode.

7. The Criticisms and Risks

Risk parity is not without significant criticisms. Several structural risks can undermine the strategy.

Reliance on Negative Stock-Bond Correlation

The core diversification benefit of risk parity depends on stocks and bonds moving in opposite directions during stress periods. This relationship held reliably from approximately 2000 through 2021: when stocks fell, bonds rallied as investors sought safety and central banks cut rates. However, this negative correlation is not a law of nature. It is historically unusual.

From the 1960s through the late 1990s, stock-bond correlations were often positive. Inflation was the dominant macro driver, and rising inflation hurt both stocks (through higher discount rates) and bonds (through higher yields) simultaneously.

2022 was the critical test case. The Federal Reserve aggressively raised interest rates to combat inflation. Both stocks and bonds fell simultaneously. The S&P 500 declined approximately 18% and long-term Treasury bonds (as measured by the iShares 20+ Year Treasury Bond ETF, TLT) fell approximately 31%. Risk parity strategies that held levered long positions in bonds suffered their worst year in decades. The very asset (bonds) that was supposed to hedge stock losses amplified them.

Leverage Cost

Risk parity requires leverage to scale a balanced (low-return) portfolio to competitive return levels. Leverage has a cost: you borrow at the short-term interest rate and invest the proceeds in longer-duration assets. When short-term rates are near zero (as they were from 2009–2021), leverage is nearly free. When short-term rates rise to 4–5% (as they did in 2023–2024), the cost of leverage can substantially erode returns.

A risk parity portfolio that is 1.5x levered with a 5% borrowing cost faces a 2.5% annual drag (0.5 × 5%) on returns relative to an unlevered portfolio. This drag must be earned back through the diversification benefit of risk balancing. In high-rate environments, the math becomes much more challenging.

Duration Risk

Risk parity portfolios are typically overweight bonds relative to traditional allocations. This means they carry significant duration risk — sensitivity to changes in interest rates. In a rising rate environment, the large bond allocation can produce sustained losses. The combination of duration risk and leverage can create path-dependent outcomes where a risk parity portfolio that would have recovered eventually is forced to deleverage at precisely the wrong time.

Model Risk

The equal risk contribution calculation requires estimates of asset volatilities and correlations. These are inherently backward-looking. If volatility regimes shift (a quiet bond market suddenly becomes volatile) or correlations change (stocks and bonds move from negatively to positively correlated), the risk parity weights become stale and the portfolio is no longer truly balanced. Most implementations use rolling windows (60–252 days) to estimate these parameters, but regime changes can happen faster than the lookback window can detect.

8. Implementing Risk Parity in Practice

For practitioners building risk parity portfolios, the implementation requires several decisions:

Asset Universe

The classic risk parity universe includes: equities (US, international developed, emerging), nominal government bonds (short, intermediate, long), inflation-linked bonds (TIPS), commodities (broad index, gold), and sometimes real estate (REITs) and credit. More asset classes provide more diversification but increase estimation error in the covariance matrix.

Covariance Estimation

The standard approach uses an exponentially weighted moving average (EWMA) of returns with a halflife of 30–60 days for volatilities and 90–180 days for correlations. Ledoit-Wolf shrinkage (Ledoit and Wolf, 2004, in the Journal of Multivariate Analysis) is commonly applied to stabilize the covariance matrix and reduce estimation error, particularly when the number of assets is large relative to the estimation window.

Rebalancing

Risk parity portfolios require regular rebalancing as asset volatilities and correlations change. Monthly rebalancing is standard. Some implementations use threshold-based rebalancing: rebalance only when an asset’s risk contribution deviates from the target by more than a specified tolerance (e.g., 2 percentage points).

Leverage Implementation

Institutional implementations use futures contracts on equity indices, government bonds, and commodities to achieve leverage efficiently. Futures provide embedded leverage (you post margin, not the full notional amount) and avoid the explicit borrowing costs of margin loans. The implied financing rate in futures is closely tied to the short-term interest rate.

9. Risk Parity vs. Alternatives

10. The Bottom Line

Risk parity is a principled approach to portfolio construction that addresses a genuine flaw in traditional allocation methods: the unbalanced risk concentration in stock-heavy portfolios. By equalizing risk contributions and using leverage to scale to target returns, risk parity has historically delivered competitive returns with lower drawdowns and higher Sharpe ratios than 60/40.

But it is not a free lunch. The strategy depends on structural conditions — notably negative stock-bond correlation and low leverage costs — that are not guaranteed to persist. The 2022 experience demonstrated that when inflation forces rates higher, both sides of the stock-bond allocation can lose simultaneously, and leverage amplifies the pain.

For investors considering risk parity, the key questions are: (1) what is your view on future stock-bond correlation? (2) what is the cost of leverage in the current rate environment? and (3) can you tolerate the path-dependent risk of leveraged bond positions in a rising rate regime? Risk parity works best when you don’t know which economic environment is coming — which, to be fair, is most of the time.

Alpha Suite implements risk parity as one of the portfolio optimization methods in its quant_plugins/portfolio_opt.py module. The system computes equal risk contribution weights across position holdings, applies correlation penalties to reduce concentration in correlated names, and uses sector caps and turnover budgets to maintain practical portfolio constraints.