Measure and manage portfolio risk Lesson 8 of 14
Portfolio and risk · Measuring return and risk
Portfolio beta: how much of the market a portfolio carries
How beta is calculated from returns, how to combine holdings into a portfolio beta, betas of funds and stocks against SPY, and why the number drifts.
On this page
Beta is the amount an investment has tended to move for each 1% move in the market. A portfolio's beta is the weighted average of the betas of the things in it. It is the standard measure of market risk: the part of a portfolio's ups and downs that comes from simply being invested when the whole market moves.
Market risk matters because it is the part that diversification cannot remove. Owning forty stocks instead of four averages away company-specific surprises. It does nothing about a year like 2008, when nearly every stock fell together. Beta estimates how exposed a portfolio is to that. This lesson shows how beta is measured, how to add up a portfolio's beta from its holdings, and how far a beta can be trusted.
The idea
Take every month in a period and plot a dot: the market's return that month across, the fund's return that month up. If the fund simply copied the market, the dots would lie on a 45-degree line. If it moved twice as much, the line through them would be twice as steep.
Beta is the slope of the line that fits the dots best.
For QQQ, the Nasdaq-100 fund, against SPY, the slope is 1.13. In a typical month when SPY rose 1%, QQQ rose about 1.13%; when SPY fell 1%, QQQ fell about 1.13%. The dots do not sit exactly on the line. How tightly they hug it is a separate number, R-squared, here 0.82: the market's moves account for 82% of the variation in QQQ's monthly returns, and the other 18% is QQQ's own.
Two funds can have the same beta and very different scatter. Beta is the tilt of the cloud, not its width.
How it works
The second form is the one to remember. Beta is high when an investment is both strongly correlated with the market and more volatile than it. QQQ's correlation with SPY over the period was 0.91, its volatility 18.4% and SPY's 14.8%: 0.91 × (18.4 ÷ 14.8) = 1.13.
It also shows how a volatile asset can have a low beta. Gold's volatility was 17.4%, higher than SPY's, but its correlation with SPY was 0.09. Its beta: 0.09 × (17.4 ÷ 14.8) = 0.11.
Betas of funds, 2005 to 2026
| Fund | Beta to SPY | R-squared |
|---|---|---|
| IWM (US small companies) | 1.19 | 0.78 |
| XLK (technology sector) | 1.14 | 0.78 |
| QQQ (Nasdaq-100) | 1.13 | 0.82 |
| VNQ (real estate) | 1.06 | 0.53 |
| XLE (energy sector) | 1.06 | 0.35 |
| EFA (foreign developed stocks) | 0.97 | 0.73 |
| BRK-B (Berkshire Hathaway) | 0.65 | 0.31 |
| XLP (consumer staples) | 0.57 | 0.47 |
| GLD (gold) | 0.11 | 0.01 |
| AGG (US bonds) | 0.07 | 0.05 |
| TLT (long Treasury bonds) | −0.08 | 0.01 |
Energy and real estate both show a beta of 1.06, the same as each other and close to the market's. Their R-squared values, 0.35 and 0.53, say something the beta does not: most of what moved the energy sector was not the stock market. Where R-squared is near zero, as for gold and Treasuries, the beta is describing almost nothing.
From holdings to a portfolio beta
A worked example with four of the funds above:
| Holding | Weight | Beta | Weight × beta |
|---|---|---|---|
| QQQ | 40% | 1.13 | 0.452 |
| XLP | 20% | 0.57 | 0.114 |
| AGG | 20% | 0.07 | 0.014 |
| GLD | 20% | 0.11 | 0.022 |
| Total | 100% | 0.60 |
As a check, the returns of this portfolio, rebalanced to these weights every month, were regressed directly on SPY over the same 262 months. The slope is 0.60. The shortcut and the long way agree, because beta is linear: the slope of a weighted sum is the weighted sum of the slopes.
Read it as an exposure. A beta of 0.60 says that in a month when SPY falls 10%, the market-driven part of this portfolio's return would be about −6%. The actual result will differ by whatever its non-market parts did that month.
Three consequences follow from the formula:
- Cash dilutes beta. Moving a fifth of a portfolio with beta 1.0 into cash lowers it to 0.8.
- Borrowing raises it. A portfolio 150% invested in a beta-1.0 fund, with the extra 50% borrowed, has a beta of 1.5.
- Weights drift. A holding that rises faster than the rest becomes a larger share of the portfolio, and the portfolio's beta moves toward that holding's beta without a single trade.
Betas of single stocks
Stocks are noisier than funds. From daily closes over November 19, 2024 to November 17, 2026 (500 days), against SPY:
| Stock | Beta | R-squared | Its own annual volatility |
|---|---|---|---|
| TSLA | 2.21 | 0.44 | 56.3% |
| NVDA | 1.83 | 0.48 | 44.7% |
| AAPL | 1.04 | 0.37 | 29.0% |
| MSFT | 0.94 | 0.29 | 29.5% |
| JPM | 0.90 | 0.42 | 23.6% |
| XOM | 0.16 | 0.01 | 25.4% |
| PG | 0.10 | 0.01 | 19.8% |
| JNJ | 0.02 | 0.00 | 20.3% |
| KO | −0.04 | 0.00 | 18.4% |
Coca-Cola, Johnson & Johnson, Procter & Gamble and Exxon Mobil all had betas near zero in this window. None of them was a riskless stock: each swung by 18% to 25% a year. Their movements were simply unrelated to the index over these two years. A portfolio of only such stocks would have a beta near zero and plenty of risk.
Where it fails or misleads
Beta changes over time
A beta is an average over the window it was measured in. The full-period beta of long Treasury bonds, −0.08, hides two different regimes. Measured over the previous 36 months it was −0.85 in July 2013, when bonds rose as stocks fell, and +0.82 in December 2025, after both had fallen together in 2022. The energy sector's 36-month beta went from 1.89 in March 2021 to −0.12 in October 2026.
Anyone relying on Treasuries' negative beta to offset stocks in 2022 found that it had turned positive. A beta is a description of a period, and a hedge built on it holds only while that period's relationships do.
Betas tend to move toward 1
Stocks with very high or very low measured betas usually show less extreme betas in the next period, a pattern documented by Marshall Blume in 1971. Part of an extreme estimate is measurement noise. Some data providers therefore publish an "adjusted beta", commonly two-thirds of the measured beta plus one-third of 1.0. A measured beta of 1.83 becomes 1.55. When two sources disagree on a stock's beta, this adjustment, the window (two, three or five years), the return interval (daily, weekly or monthly) and the index used are the usual reasons.
The line is not equally steep in both directions
Over 2005 to 2026, QQQ's beta in the 89 months when SPY fell was 1.01, and in the 173 months when SPY rose it was 1.22. One number averages the two. In SPY's worst month, October 2008 (−16.5%), QQQ's beta of 1.13 predicted a loss of 18.7% and the fund lost 15.5%. Consumer staples, with a beta of 0.57, were predicted to lose 9.4% and lost 12.6%. In a panic, low-beta assets often fall more than their beta suggests.
Beta ignores everything that is not the market
A single stock's risk is mostly its own: earnings, lawsuits, one product. With an R-squared of 0.37, nearly two-thirds of Apple's daily variation had nothing to do with the index. For one stock, beta is a small part of the picture. For a diversified portfolio, where company-specific risk has largely averaged out, it is most of it.
The benchmark is a choice
Beta against the S&P 500 is the convention for US stocks. A portfolio of foreign stocks, bonds or gold has a beta against SPY, but it answers a narrow question. A bond fund's risk is interest rates, and its beta of 0.07 says nothing about that.
Compared with its neighbours
| Measure | Question it answers | Range | Says nothing about |
|---|---|---|---|
| Beta | How much does it move per 1% market move? | Any number | Risk unrelated to the market |
| Correlation | How consistently does it move with the market? | −1 to +1 | The size of the moves |
| R-squared | What share of its variation does the market explain? | 0 to 1 | Direction or size |
| Volatility | How much does it move in total? | 0% upward | Where the movement comes from |
| Alpha | What did it return beyond its market exposure? | Any number | Whether that will continue |
How Finabase does it
Finabase shows a beta for individual stocks and funds, and does not calculate one for a portfolio.
- Symbol pages and comparison tables list "Beta" among the key statistics. The value comes from the market-data provider as a finished number; Finabase does not compute it, and the page does not state the window or the index. Treat it as a conventional beta against the broad US market, and expect it to differ from the figures on this page, which use specific dates.
- The screener does not filter or sort by beta. The figure is included when screener results are exported to a spreadsheet.
- The chart's BETA study is a different calculation with the same name; see the BETA study lesson before using it as a market beta.
- Portfolios show each position's market value and an allocation bar, with no beta.
To find a portfolio's beta, take each position's market value as a share of the total, look up each symbol's beta, multiply and add, exactly as in the worked example. Count cash as a holding with a beta of zero.
Frequently asked questions
What does a portfolio beta of 1.2 mean?
That the portfolio has tended to move about 1.2% for each 1% move in the benchmark, in both directions. In a market fall of 20% its market-driven loss would be about 24%, before anything specific to its holdings.
Can beta be negative?
Yes. A negative beta means the investment has tended to rise when the market fell. Long Treasury bonds had a negative beta for most of 2008 to 2020. Inverse funds are built to have one. Negative betas are also the least stable.
Is a low-beta portfolio safer?
It is less exposed to broad market declines. It can still be concentrated, volatile or sensitive to something else, such as interest rates or oil. Beta measures one risk.
Which beta should I use: daily, weekly or monthly?
Monthly returns over three to five years are the most common choice for funds. Daily data gives more observations and is affected by stocks that trade at different hours or react to news with a lag. Whatever the choice, use the same one for every holding.
How do I lower a portfolio's beta?
Arithmetically, by raising the weight of low-beta holdings or cash and lowering the weight of high-beta ones. Whether to do so is a decision about how much market risk suits the investor, which this lesson cannot answer.
Summary
Beta is the slope of an investment's returns against the market's: its correlation with the market times its relative volatility. A portfolio's beta is the weight-times-beta sum of its holdings, and the shortcut matches a direct regression. From 2005 to 2026 the Nasdaq-100 carried 1.13 times the market's moves, consumer staples 0.57 and gold about a tenth. Beta is a good summary of a diversified portfolio's exposure to market-wide falls and a poor summary of anything else: it ignores company-specific risk, it differs in rising and falling months, and it drifts, sometimes changing sign, from one period to the next.
Check your understanding
5 questions on this lesson. Each answer comes with the reason, and nothing is scored.
A portfolio is 50% in a fund with beta 1.2, 30% in a fund with beta 0.6 and 20% in cash. What is its beta?
- 0.78
- 0.90
- 1.80
Show the answer
A. 0.78 Portfolio beta is the sum of weight × beta: 0.5 × 1.2 + 0.3 × 0.6 + 0.2 × 0 = 0.60 + 0.18 + 0 = 0.78. Cash counts with a beta of zero.
Gold was more volatile than SPY from 2005 to 2026, yet its beta was 0.11. How is that possible?
- Beta ignores volatility
- Its correlation with SPY was close to zero, and beta is correlation times relative volatility
- Gold funds are not included in the market
Show the answer
B. Its correlation with SPY was close to zero, and beta is correlation times relative volatility Beta = correlation × (asset volatility ÷ market volatility). Gold’s correlation with SPY was 0.09, so even with higher volatility its beta was 0.09 × (17.4 ÷ 14.8) = 0.11.
Two funds each have a beta of 1.06 to SPY. One has an R-squared of 0.90 and the other 0.35. What is the difference?
- The second fund is less risky
- The market explains most of the first fund’s movement and only about a third of the second’s
- The second fund’s beta is wrong
Show the answer
B. The market explains most of the first fund’s movement and only about a third of the second’s Beta is the tilt of the scatter and R-squared is how tightly the points hug the line. A fund with a low R-squared moves for many reasons other than the market, so its beta describes a small part of its risk.
A stock’s beta measured over the last two years is 0.02. What does that tell you?
- The stock barely moves
- Over those two years its moves were unrelated to the index; it may still be volatile
- It will protect a portfolio in the next market fall
Show the answer
B. Over those two years its moves were unrelated to the index; it may still be volatile A beta near zero means no relationship with the index in that window, not small moves. Johnson & Johnson had a beta of 0.02 and annual volatility of 20% over the two years tested. Betas also drift between periods.
A portfolio’s largest holding doubles while the others stand still, and nothing is traded. The holding has a beta of 1.8 and the rest average 0.8. What happens to the portfolio’s beta?
- It rises, because the high-beta holding is now a larger share
- It is unchanged, because no trades were made
- It falls, because gains reduce risk
Show the answer
A. It rises, because the high-beta holding is now a larger share Portfolio beta depends on current weights. A holding that grows becomes a bigger share of the total and pulls the weighted average toward its own beta.
Exercise
One task to do in Finabase. Nothing is graded or saved.
Open the comparison of NVDA, SPY and KO and find the Beta row in the table. Work out the beta of a made-up portfolio: 40% NVDA, 30% SPY, 20% KO and 10% cash. Multiply each weight by its beta, count cash as zero, and add the four results. If SPY fell 10%, what market-driven move does that portfolio beta imply?
Compare NVDA, SPY and KOWhat you should find
SPY’s beta is about 1, so it adds about 0.30, and cash adds nothing. The NVDA term is the largest by far, and KO’s low beta adds little. Multiply the total by −10% for the market-driven part of the loss; the actual result would differ by whatever the holdings did on their own. The Beta shown comes from the data provider with no stated window, so it differs from the lesson’s two-year figures.
Sources and further reading
- William F. Sharpe, “Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk,” Journal of Finance, vol. 19, no. 3 (1964)
- Marshall E. Blume, “On the Assessment of Risk,” Journal of Finance, vol. 26, no. 1 (1971) — the tendency of betas to move toward 1
- Zvi Bodie, Alex Kane and Alan J. Marcus, Investments, 12th ed. (McGraw-Hill, 2021) — the single-index model
- Finabase chart data — monthly bars of SPY, QQQ, IWM, EFA, VNQ, XLK, XLE, XLP, BRK-B, GLD, AGG and TLT through October 2026 and daily closes of ten symbols, November 2024 to November 2026; betas computed by Finabase
