ETFs to Avoid: 6 Structural Red Flags

Not every ETF is built to serve you

The ETF wrapper is one of the better financial innovations of the last 30 years. A low-cost, tax-efficient, transparent way to own a diversified basket of stocks or bonds. That's the good version.

The other version is a structure where the wrapper exists to sell a story or extract fees, and the product design quietly works against the person who buys it. These funds are usually legal, well-marketed, and advertised with a striking yield or a headline return multiplier. The problem isn't that they hide anything. It's that the mathematical consequence of the structure is easy to underestimate.

Six patterns recur often enough to be worth naming. Most "worst ETFs to avoid" lists end up reaching for some version of them. Each can have a legitimate use: hedging a known event, selling volatility on purpose, capping downside before a specific expense. The argument isn't that every fund below is bad. It's that for a long-term buy-and-hold investor, they're usually the wrong default.

1. Leveraged and inverse index ETFs

What they look like: TQQQ (3x Nasdaq-100), UPRO (3x S&P 500), SOXL (3x semiconductors), SQQQ (inverse Nasdaq-100).

The pitch is simple: multiply the daily return of an index by two or three. The prospectus is explicit about the daily reset. Most buyers still assume multi-year returns will scale proportionally.

They don't. Daily rebalancing creates volatility decay. On a flat index with realized volatility around 20%, a 3x product typically bleeds 5% to 7% per year purely from the compounded daily-return path. In low-volatility trending markets, leveraged products can outperform the nominal multiple. In sideways or choppy markets they drift down. The structure is built for tactical use over days, not buy-and-hold.

Cheng and Madhavan (2009) formalized this decay. The drag grows with realized volatility and holding period. The same math applies to inverse products, where the compounding path makes multi-week holds unpredictable even when the benchmark moves in the direction you expected. That is the core of leveraged ETF risks. The product works the way the prospectus says, just not the way most buyers expect over months and years.

Legitimate use: a one-day or intraweek directional expression. A tactical overlay when you actually want to express a view on tomorrow's move. Not a long-term bet on semiconductors or Nasdaq growth.

2. Leveraged single-stock ETFs

What they look like: NVDL (2x Nvidia), TSLL (2x Tesla), AMZU (2x Amazon).

These apply the daily-leverage mechanics of TQQQ to a single stock. The compounding drag is worse, because single-stock realized volatility is typically two to three times that of a diversified index.

Bessembinder, Cooper, and Zhang (2024) studied these products directly. Their finding: 56% of leveraged single-stock ETFs posted negative absolute returns over their first year of trading, with an estimated 9.5% annual drag relative to simply holding 2x notional of the underlying stock. The stock could rise, and the leveraged wrapper could still deliver less than the unlevered holding over the same window.

Add an expense ratio of 0.75% to 1.15%, plus the financing cost of the embedded leverage, and the math is sharply pointed against any holding period longer than a few days.

Legitimate use: very short-term trading only, and even then, synthetic exposure through options is often cheaper once costs are measured honestly. For a conviction view on a specific company, own the stock.

3. Single-stock income wrappers

What they look like: TSLY (Tesla covered-call income), NVDY (Nvidia covered-call income), MSTY (Strategy covered-call income). Most are YieldMax products.

These advertise eye-catching distribution yields, often 40% to 80% annualized, occasionally over 100%. The mechanic: the fund holds or synthetically replicates a single stock and sells short-dated call options against it. The option premiums are distributed monthly. That's where the yield number comes from.

Two structural catches. First, the distribution is not income in the economic sense. It's a partial return of the upside you would have earned from holding the stock. Second, a covered-call overlay on a single stock concentrates two already-risky exposures (the stock itself, plus short convexity on it) while adding a 0.99% management fee. Israelov and Nielsen's 2015 study of covered-call strategies formalized the mechanic: the income comes from selling upside, and over full market cycles the total return of the strategy lags simply holding the underlying.

The gap is hidden by distribution-heavy return summaries. A fund that paid out 90% and appreciated 0% looks identical on a yield chart to one that returned 90% as capital gains. For tax and compounding purposes they are nothing alike.

Legitimate use: almost none for a long-term investor. If you want income from a specific stock, owning it and selling calls directly is strictly better. No 0.99% fund fee, full control of strike and tenor, optional skip in months you don't want to cap upside.

4. Covered-call index funds

What they look like: JEPI (S&P 500 income), JEPQ (Nasdaq income), QYLD (Nasdaq covered call), XYLD (S&P covered call).

Less aggressive than single-stock wrappers. The underlying is a diversified index, so one layer of concentration risk is off the table. The structural point is the same, though. The yield is financed by capping upside, and over multi-year periods these funds tend to underperform their underlying index by 100 to 300 basis points per year depending on the option-overlay design.

Israelov and Ndong (2023) examined the tradeoff explicitly. Their conclusion: investors chasing yield in covered-call funds systematically misread the distribution as free income, when it's actually a reduction in long-term total return. A monthly distribution reads as 8% annualized income; the same fund might deliver 7% total return against 10% for the plain index. That 3% annual gap compounds into major lifetime underperformance.

Legitimate use: a conscious decision to trade expected long-term return for monthly cash flow, usually in retirement decumulation where the portfolio is being drawn down anyway. Not a total-return vehicle, and not a bond substitute for a long-horizon investor.

5. Buffer and defined-outcome ETFs

What they look like: BUFR (quarterly buffer), BUFP (annual buffer), BJUL and BAPR (Innovator's dated series), the First Trust Vest ladders.

The pitch: equity exposure with a floor. If the market drops, you're protected up to some buffer, typically 9% to 30%. In exchange your upside is capped, typically 8% to 20%.

The real buffer ETF cost shows up in two places: the explicit fee, and the upside cap. The mechanic is a collar. The fund buys put options to build the floor and sells call options to finance the puts. Morningstar's 2023 analysis of defined-outcome ETFs found fees in this category averaging 0.70% or higher, against 0.03% to 0.10% for the underlying index. The cap and the floor both reset annually, so the outcome you actually get depends on when in the outcome period you enter.

For most investors worried about near-term drawdown, a simple stock-bond mix achieves similar downside behavior at a fraction of the cost, with no reset mechanics and no upside cap to track. Buffer products are usable but rarely the most efficient solution to the problem they're sold against.

Legitimate use: a specific one-year risk window, known in advance, where capping both directions is genuinely the preferred outcome. That is a narrow situation.

6. Narrow thematic ETFs

What they look like: MJ (cannabis), PSIL (psychedelics), BLOK (blockchain), single-theme clean-energy and AI-pure-play funds.

Thematic funds package a story. The index is usually designed after the theme is already popular. The fund's launch date tends to land near or after the peak of the narrative. Most investors who buy the theme through an ETF buy it close to the top.

The evidence on thematic ETF underperformance is hard to argue with. Morningstar's 2024 Global Thematic Funds Landscape found that only about one in 10 thematic funds have survived and beaten a broad global equity benchmark over the past 15 years. Roughly 60% of thematic funds launched over that period have been closed outright, often after the theme failed to deliver on its promise. This isn't a few weak products. It's the category.

The narrative is always compelling. The track record almost never is. If beating a broad market index is hard even for full-time professionals, leaning into a single theme is worse math, not better.

Legitimate use: a small, explicitly-sized tactical position on a theme you have a specific view on, sized so that going to zero doesn't meaningfully affect the portfolio. Not a core holding, not a retirement anchor.

How to spot the pattern

These six structures share a common signal. The fund's marketing leads with a number that sounds like a return (daily leverage, distribution yield, protection level) but isn't a total return. When the headline metric is not the thing that actually compounds your money, the structure is usually working for the issuer rather than for you.

Three practical filters:

1. If a fund requires a sentence explaining what the yield isn't, it's probably not free income.
2. If a product's historical total return lags a simple index of what it holds, the structure is eating the difference.
3. If the prospectus includes "not suitable for long-term holding" in a category where you're thinking long-term, believe it.

Beacon flags the six structures above on each fund's detail page with a plain-English verdict and a cited source. The screener also has a toggle to hide flagged ETFs from results entirely. It's off by default so that direct searches still return what you asked for, but turning it on produces a cleaner universe for buy-and-hold selection.

Most long-term investors don't need any of these structures. The ones who do usually have a specific, temporary reason: a one-day hedge, a defined-outcome window, a tactical theme bet. Not a permanent allocation, and almost never the default.