Leveraged and inverse ETFs are built to deliver a stated multiple of their underlying's move, but only for a single day. Hold one longer than that, and the daily results compound on each other rather than adding up in a straight line, so what a holder actually gets can drift from the multiple printed in the fund's name. Over the three months to Sep 18, 2026, that drift showed up in most of the funds ETFIQ tracked: 395 of 532 leveraged and inverse ETFs finished short of their own stated multiple, with the median fund landing 3.8 points below it.
The gap does not scale simply with holding period. Looking across windows, the median difference from stated was -1.6% at one month, -3.8% at three months, -10.9% at six months, -8.1% at one year and -9.7% at three years. The one-year and three-year numbers sit closer to the six-month number than a straight line from the shorter windows would suggest, a reminder that this gap tracks how far and how roughly the underlying travelled over the period, not simply how long the fund was held. That pattern also shows in the volatility breakdown for the three-month window: funds tracking underlyings with 0% to 30% volatility had a median difference of -1.8%, rising to -5.0% for 30% to 50% volatility, -6.8% for 50% to 80%, and -13.2% for underlyings at 80% volatility and above. The proportion finishing short of stated rose alongside this, from 106 of 165 funds in the calmest bucket to 94 of 115 in the most volatile one.
Because the underlying's daily path matters more than its net move, funds tracking the same underlying at the same stated multiple can still finish at different distances from it, even when run by different issuers over the exact same days. Among the 53 underlyings with more than one issuer offering the identical mandate, the spreads between the closest and furthest fund varied. The widest was ETHA +2x, where four issuers produced a 4.4 percentage point spread between ETHU (+10.2) and ETHT (+5.8) while the underlying moved +54.7%. CPER +2x showed a 4.1 point spread between CPXR (-1.9) and UCOP (-6.0) on a +3.5% underlying move. Narrower spreads appeared in trades such as XLF +2x, where UYG (-1.9) and XLFX (-3.4) differed by 1.5 points, and NVDA -2x, where NVD (-7.9) and NVDQ (-9.4) differed by the same margin.
None of this reflects a fund failing to do what it is built to do. The stated multiple applies to a single day's return before compounding; measuring it against a longer holding period is measuring something the fund was never designed to deliver directly. What the tables show is the size of that difference, fund by fund, window by window, and issuer by issuer, calculated the same way from the same start and end dates.
Readers can pull up any fund by name on the live desk to see its own distance from stated multiple over the window of their choice, alongside the underlying volatility that accompanied it.