Thematic ETFs

What the format does well — and what you are actually buying

A thematic fund is the only ETF category that arrives pre-loaded with a narrative you already believe: AI will reshape work; the grid must be rebuilt; cyberattacks are increasing. Usually the narrative is true. That is exactly the trap: a widely-known true story is already reflected in prices (Chapter 6's arithmetic, applied to a sector), so your edge would have to lie in the market mispricing something everyone can read in a newspaper. The realistic sources of edge are therefore time and temperament: a willingness to hold the theme longer than the product cycle, through the drawdown that shakes out the late arrivals.

Set against that, the format’s genuine advantages deserve stating plainly. The basket beats the bet. If your conviction is “AI infrastructure wins the decade,” picking the single winning stock inside that theme is the hardest version of the trade; Chapter 18’s skewness data says most single names disappoint even inside winning industries. A theme fund owns the contenders, so the theme being right is enough; you don’t also need to out-pick everyone else who believes it. Early weight in future giants. A company that may define the next decade can sit at a few hundredths of a percent of a world index, economically invisible to a core holding. The theme fund holds it at a weight where being right actually shows up, years before cap-weighting would deliver it to you. Conviction with structure. A theme in satellite form comes with a published basket, a rulebook and a daily price: a far more disciplined way to hold a view than an improvised collection of story stocks.

Add the industry's incentive: thematic funds carry 0.35–0.75% fees against a core's 0.05–0.20%, so launching them is profitable, and they launch where demand is: after a run. The result is a product cycle rather than an investment cycle: theme performs → assets flow → launches cluster → theme mean-reverts → fund closes quietly a few years later (Chapter 67 covers the closure end; thematic funds are heavily represented in it).

The construction problem, in detail

Turning a story into a rulebook is genuinely hard, and the seams show:

Definitional drift. What is a "robotics company"? Providers use revenue-share thresholds, keyword screens of filings, patent classifications, or vendor taxonomies. The result: the same theme from two providers can share as little as 20–30% of holdings. Your "AI fund" is one vendor's definition of AI, and definitions get revised as the theme evolves, meaning today's portfolio may not be what you bought.

Purity versus liquidity. The purest exposure is usually small, illiquid companies; the tradable version leans on mega-caps that happen to have a division in the theme. Many AI funds are substantially the same handful of mega-caps your world fund already holds at market weight, wearing a 0.55% fee. Check overlap with the core before buying: it is frequently the whole argument.

Concentration and capacity. Narrow eligible universes plus fund-law caps (Chapter 40) produce 30–60 stock portfolios with 40%+ in the top ten, plus a genuine capacity problem: when a niche fund grows large relative to its small constituents' float, the fund's own flows move its holdings, on both the way in and, painfully, the way out.

Turnover. Themes are re-screened as definitions evolve; turnover often runs 30–100%+, importing Chapter 41's toll at scale. Backtests, of course, pay none of it.

The flow gap: the number that should end most conversations

Morningstar's "Mind the Gap"-style analyses consistently find investor returns lag fund returns most severely in the narrowest, most volatile categories: for thematic and sector funds, gaps of several percent per year are typical, versus near-zero for broad index funds. Mechanism: money arrives after runs and leaves after drawdowns. So the thematic investor faces a stacked deduction: broad market return, minus definitional dilution, minus fees, minus turnover costs, minus the timing gap they themselves create. Each layer is small enough to dismiss and they compound into the category's reputation.

The five-question test (and permission to fail it)

Buying a theme well starts with five questions, in writing:

1. What do I know that the price doesn't? If the answer is the newspaper's narrative, stop: you are buying consensus at consensus prices.
2. Does the theme's success reach these shareholders? Transformative technologies frequently enrich customers and private investors, not listed-equity holders (Chapter 30's airline-era logic). Value capture is a separate question from technological truth.
3. What's the overlap with my core? Pull the top 20 holdings against your world fund's. If it's mostly the same mega-caps, you're paying 0.55% for a tilt you could size in one line.
4. Which definition am I buying, and can it drift? Read the index methodology's inclusion rule and revision policy (Chapter 40's habits, applied to a moving target).
5. What's my sizing, horizon and exit condition, in writing? Satellite caps (Chapter 61), a benchmark, and a pre-agreed review that isn't "when it hurts."

Letting a theme go when it fails the test is part of buying them well; the core already owns every theme at market weight, so passing is never urgent. But a theme that clears all five has earned its satellite slot honestly, and holding it with Chapter 61’s rules is a fully legitimate way to invest a conviction.

Where themes work hardest for you

Three situations where the format is at its strongest:

• Access: some exposures (uranium miners, specific supply chains, frontier biotech) are genuinely hard to assemble as an individual, and a fund is the practical route: a real service, worth a real fee, for someone with a real thesis.
• Contrarian entry: the launch-clock evidence cuts both ways. Buying a theme after a 70% drawdown, when the launches have stopped and the fund is at closure risk, is the opposite of the flow-gap behavior: a legitimate, uncomfortable strategy for the temperamentally suited.
• Expression with discipline: a capped satellite held through a full cycle with written rules is a defensible use of the format, and materially better than the stock-picking it usually replaces.

A worked example

Two investors, same theme, 2021 clean energy.

• Priya buys after a 140% two-year run at 15% of her portfolio, on the "energy transition is inevitable" thesis. Inevitable it may be; the drawdown to 2023 is 60%, she sells at −45%, and her realized loss is roughly the flow gap plus the fees plus the mean reversion. Every layer of this chapter, in one position.
• Marco runs the five questions in 2023, after the launches stopped. Overlap with core: low. Definition: read, narrow, stable. Sizing: 4%, five-year horizon, exit only on thesis change. He is buying an unloved theme with written rules: still risky, but he has removed the two mechanisms (late entry, panicked exit) that did most of Priya's damage.

Same fund, different decades of the same product cycle, and the difference is entirely procedural.