Taiwan
The frames disagree, which is the interesting case: the economics read well and the danger does not, which is the split that matters most for sizing.
Rise — early debt cycle
Debt at 27% of GDP, internal. Rates stable.
Deflationary Boom — growth, no inflation
Hold Equities — efficiency and technology Eliminate Gold and cash — both cost you while capitalism compounds.
Fragile — moves 50, danger 29
Loses value when volatility rises. Needs the world to stay as it is.
Heading towardDeflationary Boom(low confidence)
TSMC monopoly — ~30% of global AI capex flows through Taiwan. Very low debt, strong fiscal position. Cheap valuations despite dominant semiconductor position. China invasion risk is the singular tail risk — peak window 2027–2028.
Asia · Updated 2026-06-05
Where in the cycle
Countries follow a long arc — building strength, peaking, declining under their own debts, then resetting and starting again. It takes centuries, and the same investment behaves completely differently depending on which part of the arc a country is standing on.
Early in the cycle. Low obligations, room to invest, nobody forcing the hand. Debt at 27% of GDP, owed internally, which means it can be arranged with its own population. Long-term debt cycle still has room.
Which quadrant
Two questions describe the whole environment: is the economy growing, and are prices rising? Answer both and you land in one of four boxes, and each box rewards owning something different.
The answers come from market prices rather than government statistics, because statistics get revised and massaged. Growth is measured by the country’s stock index divided by the oil price — if companies are outrunning the energy they burn, the economy is working. Inflation is measured by gold divided by that country’s own government bonds — when savers move from the state’s money into the metal nobody can print, they are telling you something. Both are averaged over seven years, so the reading turns only a few times a decade.
TAIEX / oil
TAIEX / oil is the strongest reading on this list, and also the most concentrated — one company dominates the index.
Gold / Taiwan 10Y
Gold / Taiwan bonds flat. Very low debt at 27% of GDP.
TWD stable but geopolitical risk premium is significant and rising
Low debt but war risk makes duration exposure unattractive
AI hardware monopoly (TSMC), cheap valuations, ~30% of global AI capex
Geopolitical risk makes illiquid assets dangerous — invasion scenario = total loss
Gave manages by exclusion, not inclusion — it is far easier to know what will fall than what will rise, so the work is throwing horses out of the race rather than picking the winner.
Danger, not risk
Two different things get called risk. One is how much the price jumps around, which is noise you can wait out if you are not forced to sell. The other is the chance the money never comes back at all. They are separate, and they often point in opposite directions — an Argentine government bond barely moved in price and went to zero nine times.
This question gets asked twice: once of the country, and again of each instrument you might buy inside it. A sound economy can still contain a lethal holding, and a country in trouble can contain perfectly durable ones.
Middling on both axes. Size it on conviction, not on comfort.
The weather above narrowed it to a class. This narrows it to a holding. Same two axes, asked of the instrument instead of the economy — how much does it move, and can it go to zero? Everything below is what deflationary boom rewards; where each one sits is how much danger you would be carrying to own it.
Broad share index
moves 46 · danger 12
A fund holding a slice of every large company in a market.
The default way to own companies. Needs the world to stay roughly calm — its returns improve as volatility falls, which is what makes it fragile in the precise sense.
Efficiency and technology shares
moves 58 · danger 18
Companies that do more with less — software, automation, semiconductors.
The specific thing that works when the economy grows without inflation. Growth is priced far into the future, so it is also the first casualty when inflation shortens everyone's horizon.
Property
moves 28 · danger 6
Land and buildings, in places people with choices want to live.
Real collateral: something physical stands behind the claim. Slow to sell, which conventional measures mistake for safety — the actual safety is that it does not vanish because a government changed its mind.
Nuclear and uranium
moves 68 · danger 20
The fuel and the companies behind always-on electricity.
Electricity demand from computing is growing faster than anything can supply it, and wind and sun cannot run at three in the morning. Volatile, but the shortage is physical rather than a story.
These are long-horizon judgements on the same 0–100 scales used for the countries, not figures computed from a return series. And they describe the instrument, not this country’s version of it — a share index is a different proposition in a state late in its cycle than in one early in it, which is what the reading above was for.
Deflationary Boom
The economics are close to flawless and the danger is binary, which is exactly the case Détente's framework is built to handle. Low volatility with a small probability of total loss is the profile that kills portfolios — the same shape as a government bond before a default, not the same shape as Bitcoin. Position size, not conviction, is the only correct answer here.
- →Cross-strait military posture
- →AI capex cycle sustaining
- →Fab capacity relocating abroad, which lowers both the reward and the danger
New to this? The three readings above are explained from scratch, with a picture for each.
Start with the principles →