Bond ETF Statistics 2026
By Axis Intelligence Research
Co-author: Sarah Davis (Digital Finance & Fintech) | Last updated: September 18, 2026 | License: CC BY 4.0
US bond ETFs held $2,568.3 billion at the end of July 2026, up 28.6% in twelve months across 978 funds. Over the five weeks to September 9, they absorbed 40.5% of all ETF net issuance while holding just 16.4% of ETF assets. Axis Intelligence Research finds the six largest bond ETFs now carry $26.87 billion of value per 100 basis points of rate movement.
Quick Answer
US bond ETF assets reached $2.57 trillion in July 2026 (ICI), a rise of $571.8 billion year over year. Bond ETFs are pulling money far faster than their asset base implies: Axis Intelligence Research calculates a flow capture ratio of 2.47× — 40.5% of recent net issuance against a 16.4% share of assets. The BEDX™ reading, Axis Intelligence Research’s measure of aggregate rate exposure across the six largest bond ETFs, stood at 5.52 years on June 30, 2026.
Key Findings
- Assets of US bond ETFs were $2,568.3 billion at July 31, 2026, versus $1,996.5 billion a year earlier, according to the Investment Company Institute’s July 2026 Exchange-Traded Fund Data release.
- Axis Intelligence Research finds bond ETFs captured 40.5% of US ETF net issuance over the five weeks ended September 9, 2026, against a 16.4% share of ETF assets — a flow capture ratio of 2.47×.
- In the week ended September 9, 2026, bond ETFs took $10.88 billion of $11.33 billion in total ETF net issuance while equity ETFs recorded negative net issuance of $2.64 billion (ICI).
- Axis Intelligence Research’s BEDX™ reading was 5.52 years at June 30, 2026, equal to $26.87 billion of market value per 100 basis points across a $486.4 billion basket.
- The number of US bond ETFs rose to 978 in July 2026 from 849 a year earlier, a 15.2% increase, while bond ETFs’ share of total ETF assets slipped from 17.0% to 16.4% (ICI; Axis Intelligence Research calculation).
How Big Is the US Bond ETF Market in 2026?
$2,568.3 billion, across 978 funds, at July 31, 2026. That figure comes from the Investment Company Institute’s monthly Exchange-Traded Fund Data release — the industry’s own census, collected from the funds themselves rather than estimated from share counts.
The growth rate is the part worth pausing on. Bond ETF assets rose 28.6% over twelve months. The whole ETF complex rose 33.3%. Fixed income grew, and still lost ground.
| Category | Jul 2026 ($B) | Jun 2026 ($B) | Jul 2025 ($B) | YoY change | Source |
|---|---|---|---|---|---|
| Domestic equity | 10,090.4 | 10,152.3 | 7,690.9 | +31.2% | ICI, July 2026 |
| Global/international equity | 2,636.5 | 2,627.6 | 1,790.8 | +47.2% | ICI, July 2026 |
| Bond | 2,568.3 | 2,553.0 | 1,996.5 | +28.6% | ICI, July 2026 |
| Commodities | 315.4 | 311.8 | 231.4 | +36.3% | ICI, July 2026 |
| Hybrid | 60.8 | 57.6 | 46.2 | +31.6% | ICI, July 2026 |
| All ETFs | 15,671.3 | 15,702.3 | 11,755.7 | +33.3% | ICI, July 2026 |
Percentage changes calculated by Axis Intelligence Research from ICI values.
Sarah Davis: An asset-share decline during a year of record inflows is not a demand problem. Bond ETF NAVs do not compound the way equity NAVs did over this stretch — a portfolio yielding roughly 4.5% with a coupon-driven return profile cannot keep pace with an equity complex repricing on multiples. Share of assets measures returns as much as it measures conviction. Flow share is the cleaner read, and flow share is telling a different story entirely.
How Many Bond ETFs Are There, and How Big Is the Average Fund?
978 bond ETFs were listed in the US at July 2026, up from 849 a year earlier. Average net assets per bond ETF work out to $2.63 billion, against $3.40 billion for the average domestic equity ETF (Axis Intelligence Research calculation from ICI July 2026 fund counts and assets).
That gap understates the concentration. Two funds — BND and AGG — hold $298.7 billion between them on their June 30 fact sheets, roughly 11.7% of the entire bond ETF complex. Most of the other 976 are competing for shelf space in a market where the core allocation was decided years ago.
Are Bond ETFs Taking Flows From Equity ETFs?
Yes, and the ratio is not subtle. The ICI weekly estimated net issuance report breaks creations minus redemptions by objective. Five consecutive weeks to September 9, 2026:
| Week ended | Bond ($M) | of which taxable | of which municipal | Equity ($M) | All ETFs ($M) | Bond share | Source |
|---|---|---|---|---|---|---|---|
| Aug 12, 2026 | 14,485 | 14,008 | 477 | 16,494 | 34,217 | 42.3% | ICI weekly |
| Aug 19, 2026 | 13,785 | 12,811 | 974 | 34,216 | 49,970 | 27.6% | ICI weekly |
| Aug 26, 2026 | 13,783 | 12,054 | 1,729 | 14,488 | 32,089 | 43.0% | ICI weekly |
| Sep 2, 2026 | 12,272 | 11,567 | 705 | 18,201 | 33,240 | 36.9% | ICI weekly |
| Sep 9, 2026 | 10,882 | 9,651 | 1,231 | −2,636 | 11,325 | 96.1% | ICI weekly |
| Five-week total | 65,207 | 60,091 | 5,116 | 80,763 | 160,841 | 40.5% | ICI weekly |
Bond share column calculated by Axis Intelligence Research.
Set that against the asset weights. Bond ETFs are 16.4% of ETF assets and took 40.5% of the money. Axis Intelligence Research calls this the bond ETF flow capture ratio: 40.54 ÷ 16.39 = 2.47×. A category pulling two and a half times its own weight is not being rebalanced into. It is being bought.
The week ended September 9 is the one journalists will want. Equity ETFs went negative — $2.64 billion out — and bond ETFs supplied $10.88 billion of the $11.33 billion total. On that single week’s arithmetic, 96.1% of every net dollar entering a US ETF went into fixed income.
One caution on the weekly series: ICI labels these estimates, and actual net issuance is collected monthly. Prior weeks get revised. The five-week aggregate is the number to cite; any single week is a signal, not a settlement.
Where Inside Fixed Income Is the Money Going?
Taxable bond ETFs took $60.1 billion of the $65.2 billion five-week total. Municipal bond ETFs took $5.1 billion — 7.85% of bond ETF net issuance (Axis Intelligence Research calculation from ICI weekly data).
Sarah Davis: Munis punching at 7.85% of bond ETF flows is respectable for a segment whose buyer base is defined by a tax bracket rather than an asset allocation. The muni wrapper has a structural problem the taxable side does not: the underlying market trades thinly, in odd lots, with wide dealer spreads, and the ETF has to hold that basket together through creation units built from what authorised participants can actually source. Flow into munis is a vote of confidence in the plumbing as much as in the after-tax yield.
BEDX™ — The Bond ETF Duration Exposure Reading
Assets and flows tell you how much money is in bond ETFs. Neither tells you how much interest-rate risk that money is carrying. A dollar in SGOV and a dollar in TLT are both “a dollar in a bond ETF,” and they behave nothing alike when the curve moves.
So Axis Intelligence Research built one.
BEDX™ (Bond ETF Duration Exposure) is the assets-weighted effective duration of a fixed, disclosed basket of six flagship US-listed bond ETFs — the BEDX-6 — together with the aggregate dollar duration those funds carry. The basket spans the full rate-sensitivity range, from T-bills to the long end, and held $486.4 billion at the June 30, 2026 reading, or 19.05% of ICI’s bond ETF total for the same month.
The Q2 2026 Reading
| Ticker | Fund | Net assets ($M) | Effective duration (yrs) | Dollar duration ($M·yrs) | Basket weight | Source |
|---|---|---|---|---|---|---|
| BND | Vanguard Total Bond Market ETF | 159,816.00 | 5.80 | 926,932.8 | 32.86% | Vanguard fact sheet |
| AGG | iShares Core U.S. Aggregate Bond ETF | 138,847.46 | 5.80 | 805,315.3 | 28.55% | iShares fact sheet |
| SGOV | iShares 0-3 Month Treasury Bond ETF | 95,892.45 | 0.11 | 10,548.2 | 19.71% | iShares fact sheet |
| TLT | iShares 20+ Year Treasury Bond ETF | 41,099.93 | 15.31 | 629,239.9 | 8.45% | iShares fact sheet |
| LQD | iShares iBoxx $ Investment Grade Corporate Bond ETF | 33,113.14 | 7.92 | 262,256.1 | 6.81% | iShares fact sheet |
| HYG | iShares iBoxx $ High Yield Corporate Bond ETF | 17,626.12 | 2.99 | 52,702.1 | 3.62% | iShares fact sheet |
| BEDX-6 | Basket | 486,395.10 | 5.52 | 2,686,994.3 | 100% | Axis Intelligence Research |
All constituent figures as of June 30, 2026, taken from each issuer’s own fact sheet. Basket figures calculated by Axis Intelligence Research.
The formula, in full:
Dollar duration (fund) = net assets × effective duration
BEDX reading = Σ dollar duration ÷ Σ net assets
Value per 100bp = Σ dollar duration × 0.01
Worked: $2,686,994.3M·yrs ÷ $486,395.1M = 5.52 years. Multiply the aggregate dollar duration by one percentage point and you get $26,869.9 million — $26.87 billion of market value per 100 basis points, or $268.7 million per single basis point.
What a 17-Basis-Point Week Actually Costs
Between September 9 and September 15, 2026, the 10-year Treasury constant maturity yield moved from 4.83% to 5.00% — 17 basis points in four business days, per the Federal Reserve’s H.15 release.
Applied to the BEDX-6 dollar duration, that move is worth roughly $4.57 billion of mark-to-market across six funds. Axis Intelligence Research labels this an estimate: it is a first-order, parallel-shift calculation that ignores convexity, credit-spread movement and the fact that the belly and the long end did not move in lockstep.
| Maturity | Sep 9, 2026 | Sep 15, 2026 | Change | Source |
|---|---|---|---|---|
| 3-month | 3.95% | 4.11% | +16 bp | Federal Reserve H.15 |
| 2-year | 4.43% | 4.67% | +24 bp | Federal Reserve H.15 |
| 5-year | 4.61% | 4.83% | +22 bp | Federal Reserve H.15 |
| 10-year | 4.83% | 5.00% | +17 bp | Federal Reserve H.15 |
| 20-year | 5.28% | 5.40% | +12 bp | Federal Reserve H.15 |
| 30-year | 5.28% | 5.36% | +8 bp | Federal Reserve H.15 |
Sarah Davis: Read that curve from the bottom up. The front end moved more than the long end, which is a bear-flattening pattern — the market repricing policy, not term premium. For BEDX-6 that is the less painful version of a selloff, because the basket’s dollar duration is concentrated in BND and AGG at 5.8 years and in TLT at 15.31. SGOV, holding 19.7% of the basket’s assets, contributes 0.39% of its dollar duration. Nineteen cents of every dollar in this basket is parked somewhere that barely notices the curve at all.
Yield Is Not Free, and the Fact Sheets Say So
| Ticker | 30-day SEC yield | Effective duration | Weighted avg maturity | Yield per year of duration | Source |
|---|---|---|---|---|---|
| SGOV | 3.54% | 0.11 | 0.11 yrs | 32.2% | iShares fact sheet |
| AGG | 4.51% | 5.80 | 8.11 yrs | 0.78% | iShares fact sheet |
| TLT | 4.89% | 15.31 | 26.03 yrs | 0.32% | iShares fact sheet |
| LQD | 5.24% | 7.92 | 12.89 yrs | 0.66% | iShares fact sheet |
| HYG | 6.51% | 2.99 | 3.72 yrs | 2.18% | iShares fact sheet |
Yield per year of duration calculated by Axis Intelligence Research. Vanguard does not publish a 30-day SEC yield on the BND fact sheet, so BND is shown in the BEDX table only.
The ranking is instructive and it is not the one most allocation decks imply. TLT buys 4.89% with 15.31 years of rate sensitivity. SGOV buys 3.54% with 0.11. HYG’s 6.51% is paid for in credit risk rather than duration — 41.5% of the portfolio sits at B or below on its own fact-sheet breakdown, and its convexity reading is negative.
Sarah Davis: The spread between SGOV’s 3.54% and TLT’s 4.89% is 135 basis points. That is what the market is paying investors to accept fifteen years of duration instead of six weeks of it. Anyone who lived through 2022, when TLT returned −31.41%, should price that compensation carefully. The fee, as usual, is not in the expense ratio. It is in the risk you agreed to carry for a coupon that looked like the reward.
Methodology
Collection. Category assets, fund counts and monthly net issuance come from the ICI Exchange-Traded Fund Data release for July 2026, published August 28, 2026. Weekly net issuance by objective comes from the ICI Estimated ETF Net Issuance release published September 15, 2026, covering the five weeks ended September 9, 2026. Treasury constant maturity yields come from Federal Reserve statistical release H.15, released September 16, 2026. Fund-level net assets, effective duration, weighted average maturity and 30-day SEC yield come from each issuer’s own fact sheet dated June 30, 2026 — BlackRock for the five iShares funds, Vanguard for BND. Every source was fetched and read on September 18, 2026; every figure in this page appears as a row in the downloadable CSV with its URL and retrieval date.
BEDX™ construction. Constituents are fixed: BND, AGG, SGOV, TLT, LQD, HYG. The basket is chosen to span the full rate-sensitivity range of the US bond ETF market rather than to rank funds by size, and it is not rebalanced between quarterly readings. Inputs are the effective duration each issuer publishes on its own fact sheet at the quarter-end date, and the net assets attributable to the ETF share class. Dollar duration is net assets multiplied by effective duration. The BEDX reading is aggregate dollar duration divided by aggregate net assets. Value-per-basis-point figures are first-order and exclude convexity; BlackRock’s own convexity readings for the basket range from −0.19 (HYG) to 3.27 (TLT), so a large move in either direction will not be linear. Vanguard reports average duration to one decimal and BlackRock to two, and the two houses use different option-adjusted duration models — the reading is comparable across quarters, not across issuers to the basis point.
What the reading does not capture. Currency-hedged, leveraged, inverse and interval funds sit outside the basket, as do the roughly 972 bond ETFs not named above. BEDX measures rate exposure inside the basket, not credit exposure, and not the exposure of the bond ETF complex as a whole.
Arithmetic. Every calculation on this page was computed programmatically and re-verified against its source values before publication.
About This Dataset
bond-etf-statistics-2026.csv contains 133 observations covering US bond ETF assets, fund counts, weekly and monthly net issuance by objective, BEDX-6 constituent characteristics, BEDX readings, and the Treasury constant maturity curve for September 9–15, 2026. Every row carries source_org, source_document, source_url, retrieved_date, is_primary, axis_calculated and, where Axis Intelligence Research computed the value, a method_note naming the formula.
Licensed CC BY 4.0. Reuse, redistribute and build on it; attribute Axis Intelligence Research and link back.
Cite This Research
APA — Axis Intelligence Research. (2026). Bond ETF statistics 2026: Flows, duration and the BEDX™ reading. https://axis-intelligence.com/bond-etf-statistics/
MLA — Axis Intelligence Research. “Bond ETF Statistics 2026: Flows, Duration and the BEDX™ Reading.” Axis Intelligence, 18 Sept. 2026, axis-intelligence.com/bond-etf-statistics/.
Chicago — Axis Intelligence Research. “Bond ETF Statistics 2026: Flows, Duration and the BEDX™ Reading.” Axis Intelligence, September 18, 2026. https://axis-intelligence.com/bond-etf-statistics/.
FAQ
How much money is in US bond ETFs right now?
$2,568.3 billion at July 31, 2026, spread across 978 funds, according to ICI’s monthly Exchange-Traded Fund Data release. That is up $571.8 billion from $1,996.5 billion in July 2025.
Are bond ETFs outpacing equity ETFs on flows in 2026?
On a share-of-flow basis, yes. Bond ETFs captured 40.5% of US ETF net issuance over the five weeks ended September 9, 2026, while holding 16.4% of ETF assets — a flow capture ratio of 2.47×, per Axis Intelligence Research calculations from ICI data. In the final week of that stretch, equity ETFs recorded negative net issuance and bond ETFs supplied 96.1% of the industry total.
What is the average duration of the largest bond ETFs?
5.52 years across the BEDX-6 basket at June 30, 2026 — the assets-weighted effective duration of BND, AGG, SGOV, TLT, LQD and HYG, using each issuer’s published effective duration.
How much value moves across big bond ETFs when rates shift by one basis point?
About $268.7 million across the BEDX-6, and roughly $26.87 billion per 100 basis points. That is a first-order estimate from aggregate dollar duration; convexity and credit-spread moves are excluded.
Does buying a bond ETF mean buying duration?
Not necessarily, and that is the most commonly mispriced assumption in the category. SGOV holds 19.7% of the BEDX-6 basket’s assets and contributes 0.39% of its rate exposure, at an effective duration of 0.11 years. TLT holds 8.45% of the assets and contributes 23.4% of the dollar duration. Two funds in the same ICI category, two entirely different risk products.
Why did bond ETFs’ share of total ETF assets fall while their assets grew 28.6%?
Because the rest of the complex grew faster — 33.3% year over year, driven by equity NAV appreciation rather than by flows alone. Asset share conflates price return with allocation. Flow share separates them, which is why the 2.47× capture ratio is the more useful read.
Do municipal bond ETFs get a meaningful share of bond ETF flows?
They took $5.1 billion of the $65.2 billion in bond ETF net issuance over the five weeks to September 9, 2026 — 7.85% of the total, per ICI weekly data.
Why do ICI’s weekly and monthly bond ETF flow numbers differ?
ICI publishes the weekly series as estimates representing industry totals, with prior weeks subject to revision for data adjustments and reclassifications. Actual net issuance is collected monthly and reported in the Monthly Exchange-Traded Fund Data release. For any claim that needs to survive a fact-check, use the monthly figure.
Which is larger, BND or AGG?
BND, by a clear margin at the June 2026 quarter-end: $159.8 billion in ETF share-class net assets versus $138.8 billion for AGG. Both reported an effective duration of 5.80 years and both charge 0.03%, so the two funds are close to interchangeable on exposure.
What does the September 2026 curve move mean for bond ETF holders?
The 10-year constant maturity yield rose from 4.83% to 5.00% between September 9 and 15, 2026, with the 2-year up 24 basis points and the 30-year up 8 — a bear flattening. Short-duration funds were largely insulated; the long end took the price hit, though less than a parallel shift would have implied.