At a glance
TMF charges 0.9% a year. It has $2.20B in assets, the largest of that group; it paid 4.72% of its price in distributions over the past 12 months, quarterly. Over the past year it returned −29.1% with distributions reinvested, the weakest of those 4 funds; over three years it returned −6.8% a year. At its worst it fell 43.8% in the 2020 crash (S&P 500: 33.7%) and 69.5% in the 2022 bear market (S&P 500: 24.5%), and is still below its earlier high.
The short version
Direxion Daily 20+ Year Treasury Bull 3X Shares (TMF) aims to deliver three times the daily return of an index of U.S. Treasury bonds with more than 20 years left to maturity. It gets that exposure through swaps and shares of the iShares 20+ Year Treasury Bond ETF (TLT). The leverage resets every day, so returns over months or years can drift far from 3× the index. It is a short-term trading tool.
Price history
The interactive chart needs JavaScript.
TMF -29.1% over the year to Oct 27, 2026, total return with distributions reinvested.
How it has behaved
| YTD | 1 yr | 3 yr | 5 yr | 10 yr | Since start | |
|---|---|---|---|---|---|---|
| TMF official prospectus, to Dec 31, 2025 | — | -3.01% | — | -34.72% | -13.52% | — |
| TMF price data to Oct 27, 2026 | -18.3% | -29.1% | -6.8% | -34.4% | -17.2% | -6.6% |
- Over the year to Oct 27, 2026, TMF returned -29.1%, 4th of 4 among similar funds.
- Its 5-year annualized return was -34.4% and its 10-year -17.2%.
- In the 2020 COVID crash, while stocks fell, it gained 35.7%.
- Its worst full calendar year was 2022 (-72.5%); its best was 2011 (+109.1%).
- The deepest fall in its price history was 94.2% (Mar 9, 2020 to Oct 5, 2026), not yet recovered.
Show as a table
| Year | Price data | Prospectus (official) |
|---|---|---|
| 2026 (to date) | -18.3% | — |
| 2025 | -3.0% | -3.01% |
| 2024 | -35.9% | -35.82% |
| 2023 | -13.0% | -13.07% |
| 2022 | -72.5% | -72.80% |
| 2021 | -19.8% | -19.47% |
| 2020 | +39.0% | +38.80% |
| 2019 | +34.8% | +36.27% |
| 2018 | -11.0% | -12.06% |
| 2017 | +22.7% | +23.17% |
| 2016 | -2.5% | -3.61% |
| 2015 | -13.7% | — |
| 2014 | +97.3% | — |
| 2013 | -39.0% | — |
| 2012 | +0.6% | — |
| 2011 | +109.1% | — |
| 2010 | +16.2% | — |
| 2008 financial crisis stocks fell Oct 9, 2007 – Mar 9, 2009 | Before this security existed |
|---|---|
| 2020 COVID crash stocks fell Feb 19, 2020 – Mar 23, 2020 | Gained 35.7% over that stretch · deepest drop along the way 43.8%, not recovered by the end of the data |
| 2022 bear market stocks fell Jan 3, 2022 – Oct 12, 2022 | Lost 68.8% over that stretch · deepest drop along the way 69.5%, not recovered by the end of the data |
Dividends and income
Show as a table
| Year | Per share | Change |
|---|---|---|
| 2026 (to date) | $0.9400 | |
| 2025 | $1.5200 | -11.1% |
| 2024 | $1.7100 | -5.0% |
| 2023 | $1.8000 | +28.6% |
| 2022 | $1.4000 | +150.9% |
| 2021 | $0.5580 | -92.9% |
| 2020 | $7.9000 | +216.0% |
| 2019 | $2.5000 | -13.8% |
| 2018 | $2.9000 | +222.2% |
| 2017 | $0.9000 | +44.0% |
| 2013 | $0.6250 | +108.3% |
| 2012 | $0.3000 | -76.0% |
| 2011 | $1.2500 | -71.8% |
| 2010 | $4.4250 |
What's inside
As of Jul 31, 2026, most of the fund sat in cash, Treasury bills and money-market holdings (44.91% of net assets), with its exposure coming from 8 swap and futures positions rather than from owning the stocks outright.
| Fund held | Weight | |
|---|---|---|
| ISHARES 20+ YEAR TREASURY BO | 66.48% |
By country
United States 66.48%By asset type
Stocks 66.48%What changed recently
- Over the three months to Jul 31, 2026, investors withdrew a net $241.9M (creations minus redemptions, per the N-PORT report).
- Monthly total returns reported for that quarter: +0.52%, +2.47%, -13.76%.
How it works
The fund follows an index of U.S. Treasury bonds with more than 20 years left until they mature. Each bond is weighted by its market value. To qualify, a bond must pay a fixed rate, be in US dollars and have at least $300 million outstanding, not counting what the Federal Reserve holds. The index leaves out inflation-protected bonds, Treasury bills, agency debt and stripped zero-coupon bonds. It is rebalanced on the last business day of each month and held 45 bonds as of December 31, 2025.
TMF doesn't simply buy those bonds. At least 80% of its assets go into instruments that together give 3× daily exposure, mainly swaps (contracts in which a bank pays the index's return) and futures. As of July 31, 2026, TLT made up 66.5% of net assets, cash and short-term holdings 44.9%, and the fund held eight derivative positions. Using only a few swap partners adds counterparty risk, meaning the risk that a bank fails to pay.
At each day's close, Rafferty Asset Management resets the fund's exposure. It adds exposure after the index rises and cuts it after the index falls. Portfolio turnover was 142%. Because of this daily reset, the fund loses money if the index ends flat over time, and it can lose money even when the index rises.
What it really costs
- TMF charges 0.9% a year.
- Of that, 0.75% is the management fee; the rest covers other fund expenses.
- The fee includes a 0.11% waiver from the adviser, which may end.
- Using the prospectus's standard example, $10,000 invested would cost about $92 in the first year and $1,226 over 10 years (assuming a 5% annual return).
| Cost of $10,000 (SEC example) | 1 yr | 3 yrs | 5 yrs | 10 yrs |
|---|---|---|---|---|
| TMF | $92 | $311 | $547 | $1,226 |
Structure and tax notes
- Resets its exposure every day. Over periods longer than a day, compounding means returns can differ sharply from 3× the index — usually worse in choppy markets. Built for short-term trading, not buy-and-hold.
- Invests in other funds rather than directly in stocks or bonds, so its holdings are those funds.
- Bond prices move opposite to interest rates; the longer the maturities it holds, the more its price reacts to rate changes.
In a retirement portfolio
Its 3× leverage resets daily, and over longer periods that has led to deep, lasting losses that don't match 3× the index.
TMF launched in April 2009, so it has no record from the 2008 financial crisis. During the 2020 COVID crash it gained 35.7% overall, but it fell 43.8% between March 9 and March 18, 2020. In the 2022 bear market it lost 68.8%, and 2022 was its worst calendar year, at -72.8%. Its deepest drop, 94.2% from March 9, 2020 to October 5, 2026, had not been recovered as of October 27, 2026. Anyone who had to sell shares in a stretch like that would lock in losses that are very hard to win back.
The trailing yield was 4.72%, paid quarterly, as of October 27, 2026. That income is small next to the price swings. Costs are 0.90% a year, or $92 in the first year on $10,000 under the SEC's standard example. Past results don't predict the future, and a plan should look at the whole portfolio, not one fund.
Who it may suit — and who should look elsewhere
May suit
- Active traders who want a one-day, 3× position on long-term Treasury prices and can check it every day.
- Experienced investors taking a brief position on a view that long-term interest rates will fall.
- Traders who want the largest fund in its group: $2.2 billion in net assets as of July 31, 2026, against three similar leveraged Treasury funds.
Look elsewhere if
- Retirees wanting steady bond income: its 4.72% yield came alongside a 94.2% drop that hadn't recovered as of October 27, 2026.
- Buy-and-hold investors: it lost 34.72% a year on average over the five years through December 31, 2025.
- Those wanting less leverage: ProShares Ultra 20+ Year Treasury (UBT) targets 2× and fell 17.7% over the year to October 27, 2026, against TMF's 29.1%.
Close alternatives
Matched by leveraged/inverse funds on the same index (treasury), same direction.
Net assets from each fund's latest N-PORT or annual report where available. Returns through Oct 27, 2026.Questions investors ask
Is TMF good for retirement?
TMF isn't designed for retirement portfolios. It seeks 3× the return of long-term Treasuries for a single day only, and its prospectus warns that holders could lose all of their money. Held for longer periods, it lost 72.8% in 2022 and 35.82% in 2024. Its average annual return was -13.52% over the 10 years through December 31, 2025. Its fall of 94.2% from a March 2020 peak had not been recovered as of October 27, 2026.
Why has TMF lost so much money?
There are two reasons: falling bond prices and daily compounding. Long-term Treasury prices fall when interest rates rise, and TMF triples each day's move. Because it resets every day, a choppy run of gains and losses wears down its value even when the index ends near where it started. Over the five years to October 27, 2026, its price fell 88% in total, far more than 3× the drop in long-term Treasuries would suggest.
Does TMF pay dividends?
Yes, TMF pays a distribution quarterly. Over the 12 months to October 27, 2026, it paid $1.40 per share, a trailing yield of 4.72%. The income comes from interest on the fund's holdings, which include TLT shares, cash and short-term instruments. The payouts are small next to the price swings. TMF fell 29.1% over the same one-year period, so the income did not come close to making up for the price losses.
What's the difference between TMF and TLT?
TLT holds long-term Treasury bonds without leverage. TMF aims for 3× TLT's segment for one day at a time. In fact, TLT is TMF's largest holding: 66.5% of net assets as of July 31, 2026. TMF adds swaps on top of that to reach 3× exposure. That leverage, the daily reset and a 0.90% expense ratio mean TMF's long-term results can look very different from three times TLT's, and they have often been worse.
TMF vs UBT: which is less risky?
UBT carries less leverage. ProShares Ultra 20+ Year Treasury (UBT) also tracks 20+ year Treasuries, but it targets 2× the daily return instead of 3×. Over the year to October 27, 2026, UBT fell 17.7%, while TMF fell 29.1%. Over three years, UBT averaged -0.4% a year and TMF -6.8%. TMF is much larger, with $2.2 billion in net assets against UBT's $110.7 million. Both reset daily and are built for short-term trading.