Portfolio and risk · Measuring return and risk
Upside and downside capture: a fund in rising and falling markets
How capture ratios are calculated, what they show about ten funds measured against SPY since 2011, and why equal capture on both sides is only leverage.
On this page
Capture ratios split a fund's history in two: the months its benchmark rose and the months it fell. Upside capture is the share of the benchmark's gains the fund earned in the rising months. Downside capture is the share of its losses the fund suffered in the falling months. A fund with an upside capture of 90 and a downside capture of 70 kept nine-tenths of the good months and took seven-tenths of the bad ones.
The pair is popular on fund fact sheets because it describes behaviour in words an investor already uses: how does it do when the market is up, and how does it hold when the market is down? This lesson calculates both for ten funds against SPY and shows the one comparison that makes the numbers meaningful: whether the two differ from each other.
The idea
Sort the months into two piles according to what the benchmark did. In one pile, average the fund's return and the benchmark's; do the same in the other. The comparison of the two averages in each pile is the whole measure.
The low-volatility fund USMV is shorter than SPY on both sides: smaller gains in good months, smaller losses in bad ones. QQQ is taller on both sides. Neither picture alone says whether the trade was a good one. For that, the two sides have to be compared: USMV gave up about 27% of the average up month to avoid about 35% of the average down month.
How it works
This is the compound version used on most fund fact sheets. A simpler version divides the average monthly returns instead. The two give similar but not identical numbers; when comparing published figures, check that they use the same method, period and benchmark.
A worked example on USMV
The data: 179 monthly total returns of USMV and SPY, December 2011 to October 2026. SPY rose in 125 of those months and fell or was flat in 54.
- Up months. Compounding only SPY's 125 up months and scaling to a year gives 46.3%. USMV's return in the same months, treated the same way, is 32.0%.
- Upside capture: 32.0 ÷ 46.3 × 100 = 69.
- Down months. SPY's 54 down months compound to −34.5% a year. USMV's are −24.1%.
- Downside capture: −24.1 ÷ −34.5 × 100 = 70.
- Capture ratio: 69 ÷ 70 = 0.99.
So over these fifteen years USMV captured about 70% of the market in both directions. Its capture ratio is almost exactly 1: a smaller version of the market, with no extra protection beyond what its smaller size gave.
The annualized figures in steps 1 and 3 are not returns anyone earned. No one can hold a fund only in the months the index will rise. They are an intermediate quantity, and only their ratio is used.
Ten funds against SPY
| Fund | Upside capture | Downside capture | Ratio | Return a year (CAGR) | Deepest fall |
|---|---|---|---|---|---|
| SPY (the benchmark) | 100 | 100 | 1.00 | 14.8% | −23.9% |
| QQQ (Nasdaq-100) | 122 | 106 | 1.16 | 19.3% | −32.5% |
| BRK-B (Berkshire Hathaway) | 78 | 72 | 1.08 | 13.8% | −24.3% |
| USMV (low volatility) | 69 | 70 | 0.99 | 11.7% | −19.1% |
| XLP (consumer staples) | 57 | 58 | 0.98 | 10.1% | −13.6% |
| VTV (large value) | 88 | 90 | 0.97 | 13.3% | −25.1% |
| SPLV (low volatility) | 59 | 62 | 0.96 | 10.2% | −21.4% |
| RSP (S&P 500, equal weight) | 95 | 102 | 0.93 | 13.1% | −26.8% |
| IWM (small companies) | 100 | 121 | 0.83 | 10.8% | −32.3% |
| AOR (60/40 allocation fund) | 56 | 71 | 0.79 | 8.0% | −20.8% |
| EFA (developed ex-US) | 76 | 103 | 0.73 | 8.0% | −27.5% |
Monthly total returns after fund expenses; taxes and trading costs are not included. BRK-B is a single company, shown as a familiar example of a "defensive" holding.
Three groups appear.
- On the diagonal (ratio near 1): the low-volatility, staples and value funds. Each is, in effect, a diluted S&P 500 over this period. They lost less because they were exposed less, and they gained less by the same proportion.
- Above it: QQQ gained 22% more than SPY in up months and lost only 6% more in down months. Berkshire kept 78% of the upside for 72% of the downside.
- Below it: small companies took the full upside and 121% of the downside. International stocks took three-quarters of the upside and all of the downside. Both trailed SPY by a wide margin, and the capture ratio shows how: the damage was done in the falling months.
The allocation fund AOR looks poor on this measure, and the reason is a limit of the measure. Its bonds fell alongside stocks in 2022, which raised its downside capture against a stock index, while in rising stock markets its bonds contributed little. A stock index is the wrong yardstick for a fund that is 40% bonds.
Using it
The test: compare with a mix of the index and cash
Any capture pair on the diagonal can be copied with two holdings. A portfolio that keeps 70% in SPY and 30% in Treasury bills has an upside and downside capture of about 70 each (a little better, since the bills earn interest). So the question to ask of a "defensive" fund with a downside capture of 70 is whether its upside capture is clearly above 70. If it is not, the fund has delivered what a smaller stock position would have, usually at a higher fee.
This is the same idea as Alpha: the return that market exposure does not explain, seen from a different angle. A fund with a positive alpha captures more on the way up than its beta alone would explain, or less on the way down, or both.
One year can drive the result
| Fund | 2022 (SPY −18.2%) | January 2023 to October 2026 (SPY +107.1%) |
|---|---|---|
| XLP | −0.8% | +32.5% |
| VTV | −2.1% | +77.0% |
| SPLV | −4.9% | +30.2% |
| USMV | −9.4% | +51.8% |
| QQQ | −32.4% | +164.9% |
In 2022 the defensive funds did exactly what their holders hoped. Over the following period they earned between about a quarter and three-quarters of the market's gain. Whether the trade was worth it depends on which of those two periods an investor weighs more, and a single capture number over fifteen years blends them. Looking at the same ratios over each half of the period, or year by year, shows how stable they are. Often they are not.
What a capture pair is good for
- Describing a fund's character in plain terms before buying it, so that a year like 2023 to 2026 (for a defensive fund) or 2022 (for an aggressive one) is not a surprise.
- Checking a manager's claim of "protection in down markets" against the price paid for it in up markets.
- Comparing two funds in the same category against the same index over the same period.
Where it fails or misleads
- Equal capture is leverage, not skill. 70 and 70, or 130 and 130, can be had by mixing an index fund with cash or borrowing. Only the difference between the two sides carries information.
- The period decides the answer. From December 2011 to October 2026, 125 of 179 months were up months and the down months were mostly brief. A period containing 2000 to 2002 or 2008 would give defensive funds more to show.
- The benchmark must fit. Against a stock index, a balanced fund or a foreign fund produces capture numbers that mostly describe the difference between asset classes.
- Monthly piles hide what happens inside the month and across months. A fund that falls less in the first month of a decline and catches up in the third looks better than it was.
- It weights months, not money. A month in which the index rose 0.1% counts as an up month the same as one in which it rose 10%, though the compound method lets the large months dominate the result.
- Small denominators distort. Over a short period with few down months, the downside figure rests on a handful of observations. With three years of data it can swing by tens of points from one more month.
- Negative and odd values occur. A fund that rose in the benchmark's down months has a negative downside capture, which is good, but the ratio of the two then stops meaning anything.
Compared with its neighbours
| Measure | Question it answers | Number of figures | Separates up from down? |
|---|---|---|---|
| Beta | How much does the fund move per 1% move in the market? | One | No |
| Capture ratios | What share of the market's gains and of its losses did it take? | Two | Yes |
| Alpha | What return is left after accounting for beta? | One | No |
| Sortino ratio | How much return per unit of the fund's own downside? | One | Downside only |
| Maximum drawdown | How far did it fall at worst? | One | Downside only |
Beta is a single slope fitted through all months. Capture ratios are, loosely, two slopes: one for good months and one for bad. When the two captures are equal, they carry no more information than beta.
How Finabase does it
Finabase does not publish upside or downside capture ratios for funds or portfolios.
The nearest tools are the comparison views. A symbol's comparison tab with SPY added draws both as percent changes from the start of the range, so a period in which the market fell shows directly whether the fund fell more or less. Setting the range to cover a known decline and then a known rally, one after the other, gives a rough reading of both sides. The table under the chart lists price returns over one week to two years for each symbol and the provider-supplied beta. These returns leave out dividends, which understates high-yield defensive funds such as XLP by two to three points a year.
For exact figures, export month-end prices of the fund and the index, mark each month as up or down by the index's return, and compound each pile separately as in the formula above.
Frequently asked questions
What is a good downside capture ratio?
Below 100 means the fund lost less than the benchmark in falling months. On its own that is only a statement about exposure. It becomes a good result when the upside capture is meaningfully higher than the downside capture.
Can upside capture be above 100?
Yes. It means the fund gained more than the benchmark in rising months, as QQQ did against SPY (122). Funds with more market exposure, such as growth and small-company funds, usually show this, along with a downside capture above 100.
How is the capture ratio different from beta?
Beta fits one line through all months. Capture ratios measure rising and falling months separately. A fund that behaves the same way in both has captures roughly equal to its beta times 100; a fund that behaves differently in each shows a gap that beta averages away.
Over what period should capture ratios be measured?
Long enough to include at least one sustained decline, and the same period for every fund compared. Three-year figures from a rising market contain too few down months to say much.
Why do two websites show different capture ratios for the same fund?
They may use different benchmarks, periods, return frequency, or the average-return method in place of the compound one. The numbers are comparable only when all four match.
Summary
Capture ratios report the share of a benchmark's gains a fund earned in rising months and the share of its losses it took in falling months. Against SPY from December 2011 to October 2026, low-volatility, staples and value funds captured the same share on both sides, between 57% and 90%, which is what a smaller position in the index would have done. QQQ and Berkshire Hathaway kept more of the upside than of the downside; small-company and international funds did the reverse. Read the two numbers together, against a benchmark that fits, over a period with a real decline in it.
Check your understanding
4 questions on this lesson. Each answer comes with the reason, and nothing is scored.
A fund has an upside capture of 80 and a downside capture of 80 against the S&P 500. What does that show?
- The manager protected investors in falling markets
- It behaved like a portfolio 80% in the index and 20% in cash
- It beat the index
Show the answer
B. It behaved like a portfolio 80% in the index and 20% in cash Equal capture on both sides is what a smaller position in the index delivers. Protection worth paying for shows up as an upside capture clearly above the downside capture.
In the months the index rose, it compounded to 40% a year and a fund to 36%. In the months the index fell, it compounded to −30% and the fund to −21%. What are the fund’s captures?
- Upside 90, downside 70
- Upside 111, downside 143
- Upside 36, downside 21
Show the answer
A. Upside 90, downside 70 Upside capture = 36 ÷ 40 × 100 = 90. Downside capture = −21 ÷ −30 × 100 = 70. The fund kept 90% of the gains and took 70% of the losses.
A balanced fund holding 60% stocks and 40% bonds shows a downside capture of 71 and an upside capture of 56 against a stock index. What is the main caution?
- The fund is badly managed
- A stock index is the wrong benchmark for a fund that is 40% bonds, so the numbers mostly reflect the difference between asset classes
- Capture ratios cannot be below 100
Show the answer
B. A stock index is the wrong benchmark for a fund that is 40% bonds, so the numbers mostly reflect the difference between asset classes Capture ratios only describe behaviour against the chosen benchmark. Measured against an index of different assets, they show the gap between the asset classes, not the quality of the fund.
Why can a three-year capture ratio taken from a steadily rising market be unreliable?
- It contains too few down months, so the downside figure rests on a handful of observations
- Capture ratios need daily data
- Rising markets have no upside capture
Show the answer
A. It contains too few down months, so the downside figure rests on a handful of observations The downside capture is calculated only from the months the benchmark fell. With few of them, one more month can move the figure by tens of points.
Sources and further reading
- Morningstar, Investing terms — Downside Capture Ratio (a down market is a month or quarter with a negative market return)
- CFA Institute, Global Investment Performance Standards (GIPS) Handbook — recommended risk measures for composite reports
- Andrea Frazzini and Lasse Heje Pedersen, “Betting Against Beta,” Journal of Financial Economics, vol. 111, no. 1 (2014)
- Finabase chart data — monthly bars and dividends of SPY, QQQ, IWM, RSP, EFA, VTV, USMV, SPLV, XLP and AOR, and monthly bars of BRK-B, December 2011 to October 2026; statistics computed by Finabase
