Study · Leverage · July 2026

Do 2× Single-Stock ETFs Actually Deliver 2×?

By Aydin Ali · July 18, 2026 · ~8 min read · Verified data · Massive Market Data

I hold NBIL — a 2× leveraged ETF on Nebius Group (NBIS) — inside AEA's leveraged sleeve. I wanted to know, with real numbers instead of a textbook diagram, exactly how far a daily-reset fund can drift from its stated multiple over a real holding period. I pulled twelve months of real daily closes for NBIL and six other well-known leveraged/underlying pairs and did the arithmetic myself.

1. How these instruments actually work

A 2× single-stock ETF does not promise 2× your return over any period you choose — it promises 2× the daily return, reset every single trading day. If NBIS rises 5% today, NBIL is engineered to rise roughly 10% today. Tomorrow, that new, larger position gets leveraged again, from a new base. Over a straight, one-directional move this compounds in the holder's favor. Over a choppy, range-bound stretch — up one day, down the next, net roughly flat — the compounding math works against the holder every single time, because the fund is always leveraging from a slightly smaller base than a buy-and-hold position would have.

This is not a flaw in the product; it is the explicit, disclosed mechanic of daily-reset leverage. It just means the fund's prospectus promise (2× daily) and the number retail investors actually care about (2× over my holding period) are two different claims, and only the first one is guaranteed.

2. The data

Tier bands (Severe < −60pp, Moderate −60 to −20pp, Mild −20 to 0pp, Positive > 0pp) are this author's own descriptive buckets for the measured decay gap, not an industry-standard classification.

3. Why it happens

The math is symmetric compounding applied to an asymmetric problem. Say a stock goes up 10% one day and down 10% the next. A buy-and-hold position ends at 1.10 × 0.90 = 0.99 of where it started — down 1%. A 2× daily fund goes up 20% then down 20%: 1.20 × 0.80 = 0.96 — down 4%, not 2%. The higher the daily swings, the worse this arithmetic drag compounds, entirely independent of where the underlying actually ends up. NBIS's roughly 100% annualized volatility — nearly triple QQQ's — is precisely why NBIL's gap is the worst in the basket even though NBIS itself had a genuinely good year.

SOXL's positive gap is the honest counter-case: when a trend is strong and directional rather than choppy, that same compounding mechanism works in the holder's favor. Leverage decay is not a guaranteed loss — it is a bet on path, not just on direction, and the path is the part almost nobody prices in before buying.

The drag doesn't scale linearly with leverage — it scales closer to the square of it, since each extra turn of leverage both amplifies the move and shrinks the base it compounds from on the down day. This is a clean, hypothetical two-day illustration of the mechanism, not a ticker-specific measurement; the real, measured numbers are the basket in Figure 2 and the table above.

4. What this means for a long-term holder

Every fund in this basket discloses, in its own prospectus, that it targets a daily return multiple and is not designed to be held for extended periods without active monitoring. That disclosure is accurate and I had read it before buying NBIL — what this study changed is the size of the number I actually attach to that warning. A theoretical "decay happens" is a different decision input than a measured −106.5 percentage points on a position I hold.

This is also, concretely, why AEA's Investment Policy Statement caps the leveraged sleeve at 5% of the book, separate from and tighter than every other bucket cap. NBIL is currently under that cap. This study is the quantitative case for why the cap exists at all, not just a rule I inherited from a template.

5. What I'd actually change, given these numbers

I'm not selling NBIL over this. The position is small (2.87% of the book, well under the 5% leveraged-sleeve cap), I understood the daily-reset mechanic before I bought it, and NBIS itself remains a name I want exposure to. But this study changed three specific things about how I hold it. First, I now treat NBIL as a position with an active shelf life, not a buy-and-forget holding — daily-reset funds are a different instrument category from the underlying stock, not just a leveraged wrapper around the same bet. Second, I check the decay gap periodically rather than assuming the October entry math still holds; a −106.5pp gap measured today says nothing about what the gap looks like in another six months of different volatility. Third, I now run every leveraged position I'm considering through the Risk X-Ray before adding it, specifically for the hidden double-exposure check — because NBIL sitting next to a direct NBIS position is exactly the kind of overlap this study exists to catch.

The uncomfortable, honest version of the finding: SOXL's +60.7pp gap and NBIL's −106.5pp gap are the same mechanism, pointed in opposite directions by nothing more than which way the underlying happened to trend. Betting on daily-reset leverage isn't betting on a company being right. It's betting on that company being right smoothly — and smooth is a much harder thing to predict than direction.

Methodology