ETFIQ

ETFIQ Quarterly

Leveraged ETFs against their own promise, 2026-Q2

The second quarter of 2026, Apr 1, 2026 to Jun 30, 2026

347 funds measured · published Sep 11, 2026

287 of the 347 leveraged and inverse ETFs that traded the whole of the second quarter of 2026 returned less than their own stated multiple implied, by a median of 2.78 percentage points.

ETFIQ is an independent publisher of data about exchange-traded funds. It is not a fund issuer, broker-dealer or investment adviser and makes no recommendations. Every figure here is arithmetic on a named public source and carries the date it applies to. These figures do not change once published.
etfiq.com · the data file behind this report is at etfiq.com/data/quarterly/2026-Q2-leverage.json

A closed quarter, Apr 1, 2026 to Jun 30, 2026. These figures do not change. Published Sep 11, 2026.

Leveraged ETFs against their own promise, 2026-Q2

287 of the 347 leveraged and inverse ETFs that traded the whole of the second quarter of 2026 returned less than their own stated multiple implied, by a median of 2.78 percentage points.

The problem

You bought a fund with 2x in its name. The thing it tracks rose over the quarter. You did not make twice what it rose, and nobody told you in advance how much less you would make or why. The fund did nothing wrong and its prospectus said so, in a paragraph about daily rebalancing that reads like boilerplate until it costs you money.

The reason it is hard to check is that no one publishes the comparison. A leveraged ETF promises a multiple of the underlying’s move over a single day, then resets. Over a quarter the honest benchmark is that promise compounded day by day, which is not the multiple times the quarter’s move and is not on any fact sheet. Working it out needs every fund’s own stated multiple, its own underlying, and the daily path of both.

What happened in the second quarter of 2026

Of the 347 leveraged and inverse ETFs that traded the whole quarter, 287 finished behind their own stated multiple and 60 finished ahead. The median fund was 2.78 percentage points behind.

The direction of the fund decides the sign. Long funds ran a median 3.45 points behind; inverse funds a median 1.75 points ahead. That is not inverse funds being better run. A short position sits on collateral and the collateral earns interest, which lands in the fund and offsets the cost of running it. It is why ETFIQ ranks long and inverse funds in separate sets and never against each other.

Why the gap is there

Three things sit inside it and none of them is the market. The fee, which is charged daily on the leveraged exposure rather than on what you put in. The cost of financing that exposure, which is interest paid on a long fund and interest earned on an inverse one, and is the whole of why the two signs differ above. And the slippage of rebalancing every single day, which is the fund buying after a rise and selling after a fall, for ever, because that is what holding a constant multiple requires.

What is not in it is compounding. Compounding is in the promise: it is already inside the figure each fund is measured against, because that figure is the multiple applied to each day and carried forward. This is the residual after the mechanic has been accounted for, which is why it can be read as what the fund cost its holder to run.

Closest and furthest, inside each multiple

The table of 347 funds below is sorted on the gap, which puts every +2x on a stock that doubled at the top of it. That ranks the underlyings. Inside one stated multiple the funds are answering the same question, so this is the comparison that is about the funds: the closest to its own promise and the furthest from it, in each multiple with at least three funds in it.

Not a ranking of how well these funds are run beyond the quarter measured, and not a recommendation. A fund at the top of one multiple can sit at the bottom of the next quarter, and the underlying it happens to track moves the figure further than anything the manager does.

Every fund, 2026-Q2

All 347 funds, furthest behind its own promise first. The underlying’s return and the fund’s are total returns with distributions reinvested, on the same days. Figures in percentage points.

How this was computed

For each fund, the product of (1 + L × r) over every daily return r of its own underlying across the quarter, where L is the multiple stated in the fund’s own name and prospectus. That is the exact path its mechanic would have produced. No volatility term, no approximation, nothing that breaks at the edges. The fund’s own total return over the same days is subtracted from it.

Prices are exchange end-of-day, split adjusted, distributions reinvested, from Tiingo. Each fund’s stated multiple and its underlying come from its SEC registration. A fund had to trade the whole quarter to be in this study: one that listed inside it has no quarter, and measuring it over the part it traded and printing that beside funds that traded all of it would be the most common way this arithmetic is got wrong.

What was left out, and why

This is the part of a study worth reading first. Two tests run before any fund is measured, and both of them throw out funds that would otherwise have been the headline. Neither of those gaps was a cost anybody paid.

The price test. A fund whose price is unchanged on more than half its days is a quote rather than a trade, and its return is what somebody paid on the days anyone traded rather than what the fund did. Measured naively, a price that never moves reads as the largest gap on the desk. The flattest this quarter was AIRL, unchanged on 100 percent of its days, at 1 distinct closing price in the whole quarter.

The pairing test. Each fund’s daily returns are regressed on its underlying’s. If the realized slope is not near the multiple it is filed as, the two are not related and no difference between them means anything. 9 funds failed it this quarter, the furthest out being MARU, filed as +2 times MARA and moving like +0.08.

Quoted rather than traded: AIRL, AMDS, FRDU, GIGI, XPM, SNDY, PXIU, XRPK.

Pairings refused

Check it yourself

Every figure above is in the file below, fund by fund, with the excluded funds and the reason for each. It is free to use, including commercially, with attribution to ETFIQ and a link.

Where to next

Cite this page. ETFIQ, Leveraged ETFs against their own promise, 2026-Q2, data as of Sep 11, 2026. https://etfiq.com/research/quarterly/2026-q2-leveraged/ Free to use with attribution; the underlying files are at Open data.