Skip to content
Menu
Finabase Learn

Portfolio and risk · Portfolio theory

Risk parity: weighting a portfolio by risk instead of by money

How risk contribution is calculated, why a 60/40 portfolio is 94% stock risk, and how a stock and Treasury risk parity mix did from 2008 to 2026.

On this page
  1. The idea
  2. How it works
  3. Using it
  4. Where it fails or misleads
  5. Compared with its neighbours
  6. How Finabase does it
  7. Frequently asked questions
  8. Summary
  9. Check your understanding

Risk parity is a way of building a portfolio so that each asset class contributes the same amount of risk, instead of receiving a set share of the money. In a conventional portfolio the two are far apart. A mix of 60% stocks and 40% bonds sounds balanced, but stocks are so much more volatile than bonds that they account for nearly all of the portfolio's ups and downs.

A risk parity portfolio corrects that by holding less of the volatile asset and more of the calm one. The result is a steadier portfolio with a lower expected return, which is why the approach is usually paired with borrowing to bring the risk back up. This lesson calculates risk contributions for a 60/40 portfolio, builds a two-asset risk parity mix from SPY and a Treasury fund, and follows it from 2008 to 2026, including the year that exposed its main weakness.

The idea

Ask of a portfolio: when it has a bad month, what caused it? For a 60/40 portfolio the answer is almost always the stocks. Stocks move more than twice as much as intermediate Treasuries, and they hold the larger share of the money. The 40% in bonds moderates the swings without really sharing in them.

Where the money is and where the risk is Stocks (SPY) Treasuries (IEF) 60% 40% 60/40: money 94% 6% 60/40: risk 31% 69% Risk parity: money 52% 48% Risk parity: risk Risk share: each asset’s part of the portfolio’s variance, 2008–2026. Risk parity at its average weights.
Share of the money and share of the portfolio’s variance for SPY and IEF, January 2008 to October 2026. Risk parity is shown at its average weights over the period.

The left pair of bars is the conventional portfolio: 60% of the money, 94% of the risk. The right pair is a risk parity portfolio: less than a third of the money in stocks, and risk split close to evenly. The portfolio now depends on two things going reasonably well instead of one.

How it works

Each holding's contribution is its weight times its covariance with the whole portfolio. A holding contributes a lot of risk if it is large, volatile, or moves with everything else.

When the correlation is zero, or when there are only two assets, equal risk contribution has a simple solution: weight each asset in proportion to the inverse of its volatility.

With more than two assets and unequal correlations, the exact weights need a numerical search, but inverse-volatility weights are usually close.

A worked example: 60/40 by risk

The data: monthly total returns of SPY and IEF (7–10 year Treasuries), January 2008 to October 2026. SPY's volatility was 15.6%, IEF's 6.7%, and their correlation −0.05.

Using variances (volatility squared): 0.156² = 0.0243 for stocks and 0.067² = 0.0045 for bonds. The covariance term is 0.156 × 0.067 × −0.05 = −0.0005.

  1. Stocks: 0.6 × (0.6 × 0.0243 + 0.4 × −0.0005) = 0.6 × 0.0144 = 0.0086.
  2. Bonds: 0.4 × (0.4 × 0.0045 + 0.6 × −0.0005) = 0.4 × 0.0015 = 0.0006.
  3. Share from stocks: 0.0086 ÷ (0.0086 + 0.0006) = 94%.

Now the inverse-volatility weights: 1 ÷ 15.6 = 0.064 and 1 ÷ 6.7 = 0.149. Stocks get 0.064 ÷ (0.064 + 0.149) = 30%, and bonds 70%.

The portfolio through time

In practice the volatilities are not known in advance. The test below sets each month's weights from the volatility of the previous 36 months, then holds them for one month. The stock weight averaged 31% and ranged from 22% to 39% as conditions changed.

Growth of $1: 60/40, risk parity and risk parity with borrowing, 2008–2026 $1 $1.5 $2 $3 $4 60% SPY / 40% IEF Risk parity, no borrowing Risk parity at 1.5× (same volatility as 60/40) 2010 2015 2020 2025
Growth of $1, January 2008 to October 2026, monthly total returns: 60% SPY / 40% IEF; risk parity from trailing 36-month volatility; and the same levered 1.48 times with borrowing at the Treasury bill rate.
January 2008 to October 2026 Return a year (CAGR) Volatility Sharpe ratio Deepest fall 2008 2013 2022
SPY 11.1% 15.6% 0.67 −48.2% −36.8% +32.3% −18.2%
IEF 2.8% 6.7% 0.23 −22.8% +18.3% −6.1% −14.8%
60% SPY / 40% IEF 8.1% 9.6% 0.71 −28.4% −17.9% +15.6% −16.5%
Risk parity, no borrowing 5.4% 6.5% 0.63 −17.7% −1.8% +4.5% −15.5%
Risk parity, levered 1.48 times 7.3% 9.6% 0.63 −25.8% −3.6% +6.6% −23.2%

Returns are after fund expenses. The levered version borrows at the Treasury bill rate, which is lower than any individual can borrow at; trading costs and taxes are not included.

Without borrowing, risk parity is a conservative portfolio. Its volatility was two-thirds of 60/40's, its deepest fall 17.7% against 28.4%, and it came through 2008 almost untouched. The price was 2.7 points of return a year.

With borrowing, the portfolio is scaled up until its volatility matches 60/40's. That took leverage of 1.48: for every $100 of capital, $148 of investments and $48 borrowed. The return rose to 7.3%, still below 60/40's 8.1%. On a risk-adjusted basis the conventional portfolio was slightly ahead over this period, 0.71 to 0.63.

That result needs context. The period contains a long, strong run for US stocks (11.1% a year) and, in 2022, the worst bond market in decades. A portfolio that deliberately holds less of the first and more of the second was bound to trail. The case for risk parity is that an investor cannot know in advance which asset the next period will favour, and the case against it is the 2022 column.

What happened in 2022

Risk parity depends on stocks and bonds not falling together. For most of 2008 to 2021 they did not: IEF gained 18.3% in 2008 while stocks lost 36.8%. In 2022 inflation pushed interest rates up sharply, and both fell: SPY by 18.2% and IEF by 14.8%. A portfolio with 70% in bonds had nowhere to hide. The unlevered mix lost 15.5%, almost as much as 60/40, and the levered mix lost 23.2%, because leverage multiplies losses as well as gains and the cost of borrowing was rising at the same time.

Using it

Few individuals run a levered risk parity portfolio, and the lesson is not that they should. The useful parts are the measurement and the question it raises.

  1. Calculate where the risk is. For any portfolio of two or three asset classes, the contribution formula shows which one drives the outcome. Most balanced portfolios turn out to be stock portfolios with a cushion.
  2. Decide whether that is intended. An investor with a long horizon may be content for stocks to dominate, since they have had the higher return. The point is to choose it knowingly.
  3. Add assets that respond to different conditions. Institutional risk parity funds go beyond stocks and nominal bonds to inflation-linked bonds and commodities, so that some holding does well in each combination of rising or falling growth and inflation. Two assets, as 2022 showed, are not enough.
  4. Treat leverage as a separate decision. The unlevered portfolio is a legitimate low-risk choice. Leverage adds borrowing costs, the possibility of forced selling, and losses that can exceed what the holder planned for. See Leveraged and inverse ETFs: daily reset math and volatility drag for how leverage behaves over time.

Where it fails or misleads

  • It depends on the stock–bond correlation. When both fall together, as in 2022 and in much of the 1970s, the balance fails.
  • It concentrates in bonds by money. A portfolio that is 70% bonds is heavily exposed to rising interest rates, and with leverage that exposure exceeds 100% of capital.
  • Leverage is not free or guaranteed. The test assumes borrowing at the Treasury bill rate. Real borrowing costs are higher, and lenders can demand repayment in a crisis.
  • Volatility estimates lag. Weights based on the past three years react slowly. After a calm period the portfolio holds its most in an asset just before that asset's volatility rises.
  • Volatility is not the only risk. Equalizing volatility says nothing about the chance of a large, sudden loss, which differs between asset classes.
  • It makes an implicit forecast. Risk parity is often described as needing no return forecasts. Equal risk weights are the best choice only if every asset earns the same return per unit of risk, which is itself an assumption.

Compared with its neighbours

Approach How weights are set Needs return forecasts? Typical stock share by money
60/40 Fixed shares of money No 60%
Equal weight The same money in each asset No 50% of two assets
Minimum variance Lowest portfolio variance No Very low
Risk parity Equal contribution to variance No (but see above) 25% to 40%
Highest Sharpe Best expected return per unit of risk Yes Whatever the forecasts imply

How Finabase does it

Finabase does not calculate risk contributions or volatility-based weights.

A portfolio page shows the split by money: each position's market value and the allocation bar. That is the "60/40" view. To get the risk view, the missing inputs are the volatility of each holding and their correlation, which can be computed in a spreadsheet from exported monthly prices and put through the formula above.

The comparison chart helps with the key assumption. Overlaying SPY and a Treasury fund such as IEF over two years shows directly whether the two have been moving in opposite directions or together, and the Treasury page shows where yields stand, which sets how much room bonds have to rise in price if stocks fall.

Frequently asked questions

What is risk parity in simple terms?

Sizing each part of a portfolio so that it causes the same amount of fluctuation. Volatile assets get small positions and calm assets get large ones.

Why does risk parity usually involve leverage?

Because a portfolio dominated by bonds has low risk and a low expected return. Borrowing scales both up to a level comparable with a conventional portfolio. Without leverage it is a conservative portfolio.

Is risk parity the same as the "All Weather" portfolio?

All Weather is the name of a fund strategy launched by Bridgewater Associates in 1996 and is the best-known example of the approach. Published "all weather" recipes for individuals are unlevered approximations with fixed weights.

Did risk parity fail in 2022?

It had its worst year, because stocks and bonds fell together and many funds were levered. Whether that is a failure depends on the claim. It was never protection against rising interest rates.

Can I check my own portfolio's risk split?

Yes, with the volatility of each asset class and their correlation. The contribution formula in this lesson needs nothing more. For a 60/40 mix, or one heavier in stocks, the stock share of risk usually comes out above 85%.

Summary

Risk parity weights assets so that each contributes equally to the portfolio's variance. The calculation shows that a 60/40 portfolio of SPY and IEF took 94% of its risk from stocks over 2008 to 2026, and that balancing the two needed about 31% in stocks. That mix was much steadier and returned less; levered to the same volatility it still trailed 60/40, in a period that favoured stocks, and it fell harder in 2022 when stocks and bonds declined together. The lasting value of the idea is the measurement: knowing which holding a portfolio's results really depend on.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A portfolio holds 60% in stocks with 16% volatility and 40% in bonds with 6% volatility, with a correlation near zero. Roughly what share of its risk comes from stocks?

    1. About 60%
    2. About 75%
    3. More than 90%
    Show the answer

    C. More than 90% Contributions scale with weight squared times variance: 0.36 × 0.0256 = 0.0092 for stocks against 0.16 × 0.0036 = 0.0006 for bonds. Stocks supply about 94% of the variance.

  2. Two assets have volatilities of 15% and 5%. What inverse-volatility weights equalize their risk?

    1. 50% and 50%
    2. 25% in the first and 75% in the second
    3. 75% in the first and 25% in the second
    Show the answer

    B. 25% in the first and 75% in the second Weights are proportional to 1 ÷ volatility: 1/15 and 1/5, or 1 to 3. The volatile asset gets a quarter of the money.

  3. Why do risk parity funds usually borrow?

    1. To raise a low-risk, bond-heavy portfolio to a level of risk and expected return comparable with a conventional one
    2. Because bonds cannot be bought without leverage
    3. To avoid paying taxes
    Show the answer

    A. To raise a low-risk, bond-heavy portfolio to a level of risk and expected return comparable with a conventional one Equalizing risk puts most of the money in low-volatility assets. Leverage scales the whole portfolio up; it multiplies losses as well as gains.

  4. What made 2022 a bad year for risk parity portfolios?

    1. Stocks rose too fast
    2. Stocks and bonds fell together as interest rates rose, so the large bond position did not offset stock losses
    3. Volatility fell to record lows
    Show the answer

    B. Stocks and bonds fell together as interest rates rose, so the large bond position did not offset stock losses The approach relies on the two assets not declining at the same time. In 2022 both did, and leverage enlarged the loss.

Sources and further reading

  1. Edward Qian, “Risk Parity Portfolios: Efficient Portfolios Through True Diversification,” PanAgora Asset Management (2005)
  2. Sébastien Maillard, Thierry Roncalli and Jérôme Teïletche, “The Properties of Equally Weighted Risk Contribution Portfolios,” Journal of Portfolio Management, vol. 36, no. 4 (2010)
  3. Clifford S. Asness, Andrea Frazzini and Lasse H. Pedersen, “Leverage Aversion and Risk Parity,” Financial Analysts Journal, vol. 68, no. 1 (2012)
  4. Board of Governors of the Federal Reserve System, 3-Month Treasury Bill Secondary Market Rate (TB3MS), via FRED
  5. Finabase chart data — monthly bars and dividends of SPY and IEF, January 2005 to October 2026; statistics computed by Finabase