A Common-Sense Lesson Inside the ETF Boom: When What You Own Is No Longer What You Think You Own

A late-night office floor: a warm desk lamp over open paper reports in the foreground, a corridor receding into depth, ending at a wall of blue-lit market screens with silhouettes gathered around it

Therefore the skilled commander seeks victory from momentum, and does not demand it from his men; he selects the right people and lets momentum carry them. — Sun Tzu, The Art of War, “Energy”

Sun Tzu’s point: a good general bets on momentum, not on any individual’s heroics. His job is to put the right people in the right positions and let the situation do the pushing.

That line from twenty-five centuries ago works surprisingly well as a footnote to this episode. Taiwan’s active ETFs have posted absurdly good numbers this year — 80%, 90% annual returns are everywhere. But the researcher’s read is that this is probably momentum, not skill. And telling the two apart requires waiting for the day momentum runs the other way.

What This Episode Covers

This is episode 206 of MacroMicro’s After Meeting, recorded on 9 July 2026, and the whole thing is about ETFs. Host Roger brings in research manager Dylan to lay out what happened in the ETF world in the first half of the year: where the money went, why emerging markets outshone US equities, why the cloud giants — supposedly furthest ahead in the AI wave — lagged, and how to actually evaluate the active ETFs Taiwanese investors are currently obsessed with.

One number sets the scene: Taiwan’s total ETF assets have reached NT$7.8 trillion, and the first half alone already exceeded the entire prior year’s accumulation. Once a vehicle absorbs more capital in six months than it did in the previous twelve, it stops being merely a vehicle. It becomes a force acting on prices. Half of this episode is really about the consequences of that.

(There’s also a hiring segment in the middle where the host earnestly pitches the company’s “spotlessly clean bathrooms” and “front-row views of Taipei 101.” That’s the texture of a real finance show — funny, and genuinely human.)

Key Takeaways

1. Blowout earnings, falling stock — the explanation isn’t fundamentals, it’s positioning

The episode opens with Samsung: revenue up 120%, operating profit up 1,810%, a quarterly operating profit that not only set a company record but exceeded Nvidia’s prior quarter — more earned in one year than in the previous forty combined. The stock fell.

The market’s first explanation was “a miss”: revenue came in at KRW 171 trillion against a more optimistic estimate of 174 trillion, a gap of 3 trillion. That’s under 2%. The researcher’s read is that the 3 trillion isn’t worth over-interpreting; the real pressure came from Korea’s own leverage structure. In May, Korea approved a batch of leveraged ETFs tied directly to individual semiconductor names, and they pulled in as much as US$45 billion. These products rebalance daily: if the index falls 2%, the fund must fall 4%, and then sell to maintain its leverage ratio. The selling is mechanical. It doesn’t ask whether earnings were good.

The most striking datapoint: KOSPI has triggered market-wide circuit breakers only 12 times in its history — and 6 of those occurred in the first half of this year. The head of Korea’s financial regulator publicly said he somewhat regretted approving those single-stock leveraged ETFs in late May. Regulators admitting they got it wrong is rare enough to be worth recording.

2. Volatility decay: 2x leverage doesn’t mean 2x return

The show uses a clean piece of arithmetic. Up 5% on day one, down 5% on day two: the underlying index ends around −0.25%. Same path, but a 2x leveraged ETF — which multiplies and recalculates every single day — ends around −1%.

The key phrase is “every single day.” It makes your return depend not only on the destination but on the route. Two paths from A to B, and the bumpier one costs the leveraged product more. The longer the chop and the deeper the drawdown, the more likely long-term returns fall short of the index the fund was built to amplify. This isn’t a fee. It’s structural mathematics.

3. The label didn’t change; the contents did

Global equities rose about 11% in the first half; US equities about 10%. That’s a second consecutive year of US underperformance versus the world — genuinely unusual given the intuition that innovation happens in America.

More interesting is the divergence inside emerging markets. Korea rose 118% and Taiwan 62%, while Indonesia, China and India — carrying the same “emerging markets” label — posted negative first-half returns. Over five years, China’s weight in the EM index has fallen by more than 10 percentage points, losing its position as the largest component. Taiwan is now 27.3%; China is 19%.

Which means something rarely said out loud: buying an EM ETF used to mean buying Chinese domestic demand and India’s demographic dividend. Today it means buying Asian tech hardware and the semiconductor supply chain. Not one letter of the index’s name has changed, and yet the allocation logic has completely turned over.

4. A shift in free cash flow explains why money left the cloud giants

If AI demand is this strong, why did the cloud service giants — the ones furthest out in front — underperform this year? The episode’s answer is free cash flow.

Free cash flow is roughly cash on hand minus capital expenditure. When every giant announces data centres and joins the arms race, capex eats the cash flow. The market estimate cited on the show: the combined free cash flow of the five largest cloud players could approach zero by 2027.

That money didn’t evaporate — it went upstream, into memory, foundry, hardware. Free cash flow on the hardware side is rising sharply from here, with the increase widening over the next several quarters and the flow persisting into 2027. Cash bleeding out of one end and pouring into the other: where capital rotates is almost arithmetically determined.

This is also why investors began using factors to spread risk. Two product types outperformed the broad market this year: one selecting companies with the strongest free-cash-flow growth, the other screening large-cap index members on return on equity, accrual ratio and financial leverage to isolate balance-sheet quality. That isn’t picking a lane — it’s picking constitution. When the lane gets crowded, quality becomes the new diversification.

5. What won this year was probably sector allocation, not stock selection

Taiwan’s first active ETF only listed in May of last year; the category has already passed NT$900 billion. This year’s performance tables are full of active products at 80–90%, nearly all of them named after AI, “new momentum,” or the “new economy.”

The structural argument for active is sound. In a market-cap-weighted ETF, the largest constituent can easily exceed 50% of the fund — you think you bought a whole market, but half your money sits in one company. An active manager sets the weights, can avoid that concentration, and can position early in less mature small- and mid-cap suppliers.

But the show then does the most valuable thing in the whole episode: it looks for a control group. Over the same period, passive semiconductor-themed ETFs also returned 80–90%, some above 100%. Once active and passive produce the same order of magnitude on the same theme, “active managers pick better stocks” stops holding up. The plausible reading is that this year, getting the AI supply chain right was sufficient — active or passive.

Hence the researcher’s conclusion: you can’t yet declare active superior to passive. The real test comes when the manufacturing cycle slows and turns down — whether managers actually cut exposure in time. That’s the moment stock-picking skill becomes visible.

The episode even hands over the timing and the indicators. MacroMicro’s manufacturing cycle index has climbed to 0.55, a multi-year high, with a rare second upturn, and we’re now in active restocking. The next phase is passive restocking — manufacturers still building inventory while downstream demand starts falling, so inventory piles up. To judge whether we get there, watch three things: big-tech capex, Taiwan’s exports, and the new-orders-to-inventory ratio. Q4 is the checkpoint.

6. “We’re not prophets”

Near the end, the hosts read a listener comment along the lines of: aren’t you supposed to be prophets — why don’t you say more prophet-like things?

The reply matters more than any datapoint in the episode: we aren’t prophets; we surface views early based on how the data moves, and the decision has always been yours.

It’s genuinely hard for a financial media outfit to resist being treated as a tip service, because tips are what audiences want. Saying so plainly, on air, is worth more than another statistic.

Going Further

1. The unit of diversification is exposure, not labels

That “the label didn’t change, the contents did” observation can be pushed further than the episode takes it.

Index rebalancing is a silent change to your holdings: you place no order, and yet your exposure profile is rewritten. While China’s weight fell 10 points and Taiwan’s rose to 27.3%, nobody holding an EM ETF received a notification.

The danger is in the stacking. Suppose a portfolio holds an EM ETF, a Taiwan market-cap ETF, and a few semiconductor names. On paper that’s three different things, handsomely diversified. Look through to the underlying holdings and the same Asian semiconductor supply chain may be bet three times over. Real concentration doesn’t show up in the number of line items — it hides in overlapping constituents.

So the way to check diversification isn’t counting positions but asking: if the Asian semiconductor supply chain weakens tomorrow, what fraction of my portfolio gets hurt at once? The answer is usually higher than people assume.

The counterpoint deserves saying plainly: overlap isn’t automatically wrong. Concentrating in one industry you genuinely understand, while its cycle is still rising, is a legitimate source of return. What’s wrong is believing you’re diversified when you aren’t — taking concentrated risk without demanding the conviction and monitoring intensity that concentration requires. Misreading your own exposure is more dangerous than concentrating on purpose.

2. Leverage charges you for the path, not for the holding period

The Tao Te Ching says: one who stands on tiptoe does not stand firm; one who takes long strides does not walk far. It’s a satisfying line to attach to leveraged ETFs, but on its own it’s a correct platitude.

The precise version: volatility decay is a toll on path roughness, not on elapsed time. If the underlying grinds upward in a straight line and barely retraces, compounding can push a 2x product above twice the return — that side has to be admitted honestly. The decay happens along choppy, back-and-forth paths; the wider the swings and the more round trips, the more gets ground away.

The inference is therefore quite specific: whether a leveraged product is usable depends on whether you have a view on the path, not merely on the direction. If all you know is “I’m bullish,” without knowing whether the road is bumpy, leverage magnifies precisely the part you don’t know. Directional bets with a defined time window and an expected one-way move are its legitimate use. “Hold long term and forget about it” is not.

Korea’s US$45 billion adds one more layer: once these products get large enough, their mechanical rebalancing manufactures the chop. You think you’re bearing the market’s volatility; some of it is volatility the product created. The tool has altered the thing it was built to track — a problem shared by every financial product that outgrows its underlying.

3. To verify skill, you need a control group facing a headwind

Point 5 is the sharpest piece of reasoning in the episode, and it’s worth isolating because it demonstrates a repeatable move.

“Active ETFs returned 80–90% this year” is a fact, and standing alone it proves nothing. What the researcher did was find the same-period passive semiconductor ETFs and discover returns of the same magnitude. Only with the control group does the original claim get tested — and the excess return gets attributed to sector allocation rather than stock selection.

Generalise that into a discipline: whenever you see impressive performance, first ask “compared to what?” Performance without a control group isn’t evidence, only an outcome. And the control has to be strict enough — comparing an active semiconductor fund to the broad market guarantees a win. The right comparison is the passive product on the same theme, because that’s the one that answers “what’s left after subtracting sector luck?”

Better still is writing the test conditions down in advance. This episode already hands them over: Q4, big-tech capex, Taiwan’s exports, the new-orders-to-inventory ratio. If the manufacturing cycle really turns down, that’s a natural headwind experiment — and the thing to examine then is not the returns table but the holdings. Did the active fund cut cyclical exposure? Did it rotate defensive? Was the timing ahead of the turn or behind it?

Write the criteria first, wait for the data, then mark your own paper honestly. The value isn’t in how much you make. It’s that it lets you book luck and skill in separate ledgers. Without that separation, everyone is a genius in a bull market — and the invoice arrives in the next cycle.

Disclaimer

This article consists of personal notes and extended reflections after listening to a podcast. It is educational and commentary in nature, and does not constitute investment advice, an offer, or a recommendation. Indices, product categories and companies mentioned serve only to explain the discussion; nothing here is a buy or sell recommendation, and no price targets or entry/exit levels are provided.

Figures and views cited come from the episode, may contain transcription or interpretation error, and will become stale over time. Please defer to primary sources and the official public filings of the relevant institutions. Investing carries risk. Every investment decision should rest on your own financial situation, risk tolerance and independent judgement, and you should consult a qualified professional where appropriate. The author accepts no responsibility for any decision made or outcome experienced as a result of reading this article.

Raw Markdown

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.