In the third quarter of 2026, two hedge fund indices built by the same firm, from the same list of funds, disagreed on whether the industry had made money. HFR's HFRI Fund Weighted Composite fell 0.6% and its HFRI Asset Weighted Composite gained 1.2%. The first gives every fund the same share of the average. The second weights each fund by its assets, and in that quarter the largest funds did better than the small ones.
A hedge fund index is an average of the monthly returns, after fees, that hedge funds report voluntarily to a commercial database. Each provider decides which funds qualify, how much weight each one carries and when a fund leaves. Those decisions open gaps of several percentage points a year between indices that all claim to measure the same industry.
Over the ten years to December 2025, the HFRX Global Hedge Fund Index returned 3.08% a year, against 14.82% for the S&P 500. Ten thousand dollars grew to about 13,500 dollars in the first and 39,800 dollars in the second. Over the ten years to January 2026, HFR's other flagship, the HFRI Fund Weighted Composite, returned 7.17% a year.
The main hedge fund index families
A handful of data providers publish the indices that fund prospectuses and investors quote. Each runs a broad composite and dozens of strategy and regional sub-indices.
| Index | Provider | Weighting | Funds included |
|---|---|---|---|
| HFRI Fund Weighted Composite | HFR | Equal | More than 1,000 single-manager funds, monthly since 1990 |
| HFRI Asset Weighted Composite | HFR | Each fund's assets in the prior month | The same funds as the Fund Weighted Composite |
| HFRX Global Hedge Fund Index | HFR | Strategies weighted by the industry's asset mix | Funds open to investment through managed accounts |
| Barclay Hedge Fund Index | BarclayHedge | Equal | Every fund in the database that reports a net return for the month |
| Institutional 200 | With Intelligence | Equal, reset each January | 200 institutional funds picked from the 300 largest in the database |
| BTOP50 | BarclayHedge | Equal, reset each January | The largest investable trading advisor programs, 20 of them in 2026 |
HFR runs three main families. The HFRI indices are its benchmark series: they leave out funds of hedge funds and report every return net of all fees. The HFRX indices are designed to be investable, and the HFRU indices cover funds run under the European UCITS rules. HFR also launched an investable HFRI 500 series in 2005.
BarclayHedge calculates the Barclay Hedge Fund Index as the simple arithmetic average of the net returns of all funds in its database that reported that month, funds of funds excluded.
With Intelligence now publishes the indices previously available on Eurekahedge alongside its own. Its Institutional 200 starts from a base of 100 at the end of 2007, admits only funds that charge a performance fee and treats the latest three months as provisional until every constituent has reported.
Trend followers have indices of their own. The BTOP50 holds the largest trading advisor programs that are open to investment, report returns daily and together hold at least half the investable assets in Barclay's CTA universe. Through September 2026 the HFRI Macro Systematic Diversified/CTA Index was up 15.7% for the year. The SG CTA and SG Trend indices, and how trend-following funds did from 2022 to 2025, are covered in our guide to managed futures.
How the HFRI composite chooses its funds
To enter the HFRI indices, a fund must be open to new investment and report monthly returns, net of all fees, in US dollars. It also needs either 50 million dollars under management or twelve months of trading, with at least 10 million dollars at its last report before the annual rebalance. Where a manager runs several funds with the same profile, such as onshore and offshore versions of one portfolio, only the most representative one is counted. HFR can remove a fund that closes to new money or limits its liquidity.
Each month's value is published before every fund has reported. A first estimate appears on the fifth US business day of the following month, an update on the 15th, and the third release fixes the figure. Barclay recalculates its index each time a fund reports: the August 2026 estimate rested on 2,471 funds, and on October 30, 2026 the September estimate rested on 2,294. How far an HFRI quarter moved between two releases is shown in our comparison of hedge funds and index funds.
Most HFRI indices are equal-weighted, which keeps a few large funds from dominating a strategy index that holds only a small number of funds.
Equal weighting and asset weighting, worked through
Take a database of four funds over one quarter:
- A multi-strategy fund with 20 billion dollars of assets gains 2%.
- Three smaller funds with 200 million dollars each lose 1%.
An equal-weighted index adds the four returns and divides by four: 2 minus 3, divided by 4, a loss of 0.25%. An asset-weighted index weights each return by assets: 400 million dollars of gains against 6 million dollars of losses, on 20.6 billion dollars of capital, is a gain of 1.91%.
The equal-weighted figure is the return of the typical fund. The asset-weighted figure is the return on the typical dollar, which is what the industry earned in total.
The third quarter of 2026 followed this pattern, with a 1.8-point gap between the two HFRI composites. Money has been flowing to the largest firms: those managing more than 5 billion dollars received an estimated 113.7 billion dollars of the 124.1 billion dollars of net new investment in the first nine months of 2026, when total industry capital reached 5.64 trillion dollars. Their results move the asset-weighted composite and count for one fund each in the equal-weighted one.
Multi-strategy funds held a record 900 billion dollars at the end of September 2026. How the largest of them split capital among hundreds of trading teams is covered in our guide to multi-manager hedge funds.
Investable and non-investable hedge fund indices
The HFRI composite cannot be bought. Holding it would mean subscribing to more than 1,000 funds, each with its own minimum, notice period and redemption calendar, and redoing the allocation every year.
The HFRX indices were built to be replicated. A fund qualifies only if it meets all of the following conditions:
- it reports net returns in US dollars and accepts new investment,
- it has a 24-month track record and at least 50 million dollars of assets,
- it takes subscriptions monthly and pays redemptions at least quarterly, on 90 days' notice or less,
- it accepts money through a fully transparent managed account.
HFR then sorts the eligible funds into strategy clusters by running a cluster analysis on 24 months of returns, discards the funds least like their cluster, and keeps the smallest set that tracks each cluster closely. The HFRX Global index weights strategies by their share of industry assets and is rebalanced every quarter. Constituents that give daily transparency are priced on the next business day.
Each condition narrows the field. Funds younger than two years, funds that redeem less often than quarterly or ask for more than 90 days' notice, and managers who will not run a separate account showing every position to the investor all stay outside the HFRX. Over ten years the HFRX Global returned 3.08% a year to December 2025, and the HFRI Fund Weighted Composite 7.17% a year to January 2026, a gap of about 4 percentage points a year.
A third approach leaves hedge funds out entirely. The Merrill Lynch Factor Model, which the ProShares Hedge Replication ETF (HDG) tracks, holds six liquid exposures: the S&P 500, the Russell 2000, the MSCI EAFE and MSCI Emerging Markets indices, three-month Treasury bills and an ETF that moves inversely to the euro. Each month it reweights them to the mix that would have matched the HFRI Fund Weighted Composite most closely over the previous 24 months. Over the ten years to December 2025 the model returned 4.25% a year and the ETF 3.27%, with the ETF's expenses capped at 0.95% a year until September 2027. Other replication ETFs, and the registered funds of hedge funds open to small investors, are compared in our guide to investing in hedge funds with little money.
Survivorship, backfill and revised returns in hedge fund databases
Funds choose when to start reporting, when to stop and whether to correct what they sent. The classic estimates of survivorship and backfill bias, and a correction of both on a sample of about 8,400 funds, are set out in our comparison of hedge funds and index funds. Other measurements of the same distortions vary widely:
- Survivorship, about 2% a year
- From 1994 to 2001, an average of surviving funds overstated the return of all funds by around 2% a year. For small, young and leveraged funds the overstatement reached 4% to 6%. Funds of funds closed less often, and their bias was low.
- Hidden survivorship, more than 5% a year
- When the TASS database merged with Tremont's, about 60% of the funds added between April 1999 and November 2001 had been selected from funds still alive on March 31, 1999, and their earlier histories came with them. The bias from those histories averaged more than 5% a year, inside a period that would normally have been labelled backfill.
- Revised returns
- Across database snapshots taken from 2007 to 2011, 44.9% of 12,128 funds changed at least one past monthly return, and 21.3% changed one by a full percentage point or more, when the average monthly return was 0.62%. Almost half of the 87,461 revisions concerned returns more than a year old. Downward revisions were more frequent and larger than upward ones, and funds that never revised went on to beat those that did by about 0.3 percentage points a month.
Revisions change a fund's record in the database, which is what researchers and peer-group comparisons use. Once its third release has fixed a month, the published HFRI value stays as it was.
Self-selection: what funds outside the databases earned
Funds that never report leave no returns to average, so researchers have rebuilt them from regulatory filings. The three main attempts reach different conclusions:
- Aiken, Clifford and Ellis reconstructed the returns of funds that never reported to any commercial database from the holdings disclosed by registered funds of hedge funds. Reporting funds significantly outperformed the funds that stayed out.
- Agarwal, Fos and Jiang matched the 13F stock holdings of hedge fund firms from 1980 to 2008 against five databases. Funds began reporting after abnormal gains that did not last, and funds that stopped reporting saw their returns deteriorate and their investors leave. Reporting funds did better against some benchmarks, and the difference was not significant against others.
- Edelman, Fung and Hsieh studied the largest management firms, which together ran more than half of the industry's assets, including firms that report to no database. Reporting and non-reporting giants performed alike, and the widest gaps between them came from different exposure to credit markets.
The samples differ: the first study covers funds held by registered funds of hedge funds, and the third covers the largest firms in the industry.
One way to step around the problem was proposed in 2000: measure the industry through funds of hedge funds. A fund of funds reports what its investors earned across the funds it held, including those that later closed. HFR publishes this measure as the HFRI Fund of Funds Composite, whose returns are net of a second layer of fees.
Hedge fund index returns against the S&P 500
| Index | Period ending | 1 year | 5 years, a year | 10 years, a year |
|---|---|---|---|---|
| HFRX Global Hedge Fund Index | December 2025 | 7.13% | 2.87% | 3.08% |
| Merrill Lynch Factor Model, HFRI replication | December 2025 | 7.77% | 3.58% | 4.25% |
| S&P 500 | December 2025 | 17.88% | 14.42% | 14.82% |
| HFRI Fund Weighted Composite | January 2026 | 13.85% | 7.39% | 7.17% |
| MSCI All Country World Index | January 2026 | 21.87% | 11.95% | 12.75% |
All index returns are before any product fee and, for the hedge fund indices, after the fees of the underlying funds. The HFRI figures run one month later than the others, because they come from a fund report with a January year end.
Over ten years the HFRX Global trailed the S&P 500 by 11.7 percentage points a year, and the HFRI composite trailed world stocks by 5.6 points a year. In calendar 2025 the HFRI Fund Weighted Composite returned 12.5% and the Barclay Hedge Fund Index 12.59%, against 17.88% for the S&P 500.
The indices lost less in the falling year. In 2022 the S&P 500 lost 18.11%. The equal-weighted Barclay Hedge Fund Index lost 8.22%, and HFR's first full-year estimate for its HFRI Asset Weighted Composite was a gain of 0.97%. The two come from different databases and finished 9.2 points apart, a wider gap than the 5.3 points between the Barclay index and the S&P 500 in 2025.
Part of the smoother record comes from stale prices on illiquid holdings, which lower measured volatility and correlation with stocks. Our comparison of hedge funds and index funds shows how much market exposure reappears once that is corrected, and our guide to the Sharpe ratio covers why smoothed returns flatter risk-adjusted measures.
Which hedge fund index fits which comparison
- One fund against its peers
- The strategy sub-index of the database the fund reports to, such as the HFRI Equity Hedge (Total) Index for a long/short equity fund. Equal weights match the comparison with a typical peer.
- What the industry's capital earned
- The HFRI Asset Weighted Composite, which weights funds by assets and moves with the largest firms.
- What an investor could have bought
- The HFRX Global Hedge Fund Index, or the returns of an actual fund of funds after its own fees.
- Trend followers
- The BTOP50, the SG Trend Index or the HFRI Macro Systematic Diversified/CTA Index.
- Hedge funds against an index fund
- A net-of-fee composite against the S&P 500 with dividends reinvested, over the same months. The layers of fees inside those net returns are explained in our guide to hedge fund fees.
This article is general information, not investment advice.
Questions about hedge fund indices
What is the best hedge fund index?
It depends on the question being asked. The HFRI Fund Weighted Composite is the most widely quoted broad benchmark: it averages more than 1,000 single-manager funds with equal weights, net of fees, back to 1990. The HFRI Asset Weighted Composite shows what the industry's capital earned, because large funds count in proportion to their assets. The HFRX Global Hedge Fund Index is built from funds that can be bought through managed accounts. For trend followers, the BTOP50 and the SG Trend Index track the largest programs.
Do hedge funds beat the S&P 500?
The broad indices trailed it over the last five and ten years. Over the ten years to December 2025, the HFRX Global Hedge Fund Index returned 3.08% a year and the S&P 500 returned 14.82% a year. In calendar 2025 the HFRI Fund Weighted Composite returned 12.5% against 17.88% for the S&P 500. In 2022, when the S&P 500 lost 18.11%, the Barclay Hedge Fund Index lost 8.22% and HFR's first estimate for its asset-weighted composite was a gain of 0.97%.
Can you invest in a hedge fund index?
Only indirectly. The HFRI composites are measurement tools made of more than 1,000 funds with their own minimums and redemption terms. The HFRX indices select funds that accept money through managed accounts so that the index can be replicated. Exchange-traded funds such as the ProShares Hedge Replication ETF hold liquid market exposures chosen to move like the HFRI: over the ten years to December 2025 it returned 3.27% a year, and its model returned 4.25% before costs.
Why do hedge fund index returns change after they are published?
Because funds report late. HFR publishes an estimate on the fifth US business day of the following month, updates it on the 15th and fixes the value with the third release. Barclay recalculates its index each time another fund reports. Funds also revise their own past returns in the databases: in snapshots taken from 2007 to 2011, 44.9% of 12,128 funds changed at least one previously reported monthly return.