hirschdynamics.net ↗
Special Edition · Global Risks
DE EN
Special Edition · September 23, 2026

Global Risks
with Carmen Reinhart

An analysis of the Reichmuth Lecture at the University of Zurich: Prof. Carmen Reinhart (Harvard Kennedy School, former World Bank Chief Economist) on public debt, the AI bubble and the dollar — and why Switzerland is the counter-model.

📅 Wednesday, September 23, 2026⏱ ~10 min analysis🎓 Reichmuth Lecture · Uni Zurich
01 / Public Debt · Deep Dive

Wartime debt — and interest rates that return

Dataset
21 economies
1900–2025 · trend rising, US leading
Historical parallel
end of WWII
then: young population, few transfers
Warning
CBO
adverse debt dynamics in the US
Ratings
downgraded
Japan · US · France · Switzerland AAA

Reinhart: public debt last this high at the end of WWII — but the starting position is harder today. Source: lecture, Uni Zurich

01 / Causal Chain

Why the debt bites only now

Trigger
Central banks buy
government debt in record size · BoJ > 100% GDP
→
Step 1
Real rates negative
belief: "low forever"
→
Step 2
Incentive to borrow
debt explodes · public + private
→
Step 3
Rates normalize
debt servicing up · budget squeezed
→
Result
Ratings slide
the risk feeds on itself

It's not high debt alone that is the problem, but the return to normal interest rates — it turns a bearable burden into a growing budget line.

01 / Counter-Check & Consequence

When high debt is not a crisis

✅
Bearable at the end of WWII: a young, growing population and small social spending — you could grow into it. Or a country with a debt brake (Switzerland) never has the problem.
⚠️
The menu for reduction is unpleasant: austerity (hard, see France) · "through growth" (a wish, not a policy) · inflation · financial repression · restructuring/default.
🤫
The quietest way: simply providing fewer services — a silent retreat of the state.
🧭
For planning: don't count on permanently cheap money — plan for a state that delivers less and taxes more.
02 / AI Bubble · Deep Dive

The AI bubble — Reinhart's greatest concern

Bubble symptoms
all present
liquidity · credit boom · narrow spreads · valuations
RoW in US equities
all-time high
Fed data since 1945 · "all AI"
Fed above target
> 5 years
ECB too · little room left
Long-run GDP trend
invisible
internet, electricity — no miracle jump visible

Reinhart: the rest of the world's investment in US equities is at an all-time high — and the surge is practically all AI. Source: lecture, Uni Zurich

02 / Causal Chain

Why a real technology becomes a bubble

Trigger
Real technology
railroads then · AI today
→
Step 1
"This time is different"
euphoria · everyone wants in
→
Step 2
Overbuilt on credit
winners — and leveraged losers
→
Step 3
Rates rise
defaults · contagion via non-banks
→
Precedent
Panic of 1873
opaque networks · hidden debt

Reinhart is not an AI pessimist — it's enough that not everyone is a winner. The risk sits in leverage and opacity, not in the technology itself.

02 / Counter-Check & Consequence

Not AI pessimism — but not like 1987

✅
It takes no total failure of AI. It's enough that it works for some, but not everyone is a winner — just like the railroad.
⚠️
This is not 1987: back then wealth effects were limited. Today pension funds and sovereign wealth funds make a correction far heavier — close to a direct hit on households.
⏳
No one can predict the timing — Reinhart urges humility.
🧭
To apply: treat high valuations and hidden leverage as the real risk, not AI adoption itself.
03 / Growth Headwinds

Demographics, China — and weaker world trade

Demographics
aging
a drag everywhere — including China
China
no engine
overleveraged · housing bust · "Japanese phase"
World trade
slowing
since the financial crisis · Greece −15% CA
Trade wars
outlast
every US administration

Reinhart: in 2008 China was the world's growth engine, propping things up with stimulus. That buffer is gone today — the world economy starts from a shakier base.

04 / Doom Loop

The vicious loop of bailout and debt

Trigger
bailout
too big to fail → government debt rises
Holders
captive
domestic banks + central bank hold the debt
Rates rise
capital losses
on bank balance sheets → the loop closes

Especially relevant for Japan and Italy — a large part of the ECB's bond buying aimed precisely at defusing this loop at Italian banks. And it touches central-bank independence.

05 / The Dollar

Dominance "fraying" — but no quick replacement

Debasement trade
gold & oil
oil ~5th-largest annual rise in 126 years
Euro
no substitute
one currency, but no unified bond market
Renminbi
blocked
capital controls
Lesson
crash ≠ end
dollar can fall without losing reserve status

All the historical markers show the dollar still playing its anchor role. Even China lends abroad in dollars, not yuan. Dornbusch: you only go to a party when you know you can leave.

06 / Central-Bank Independence

Not lost de jure — but under pressure de facto

In law
not lost
but under pressure in practice (Trump vs Fed)
Fed holdings
record
US Treasuries above the WWII level
Relationship
more debt
= harder to raise rates

The higher the public debt, the greater the pressure on the central bank to keep rates low — that is what Reinhart means by independence threatened "de facto."

Conclusion · The real lesson

Two time horizons —
and a counter-model

🐢
Slow-moving: the risk of a fiscal or dollar crisis. The dollar can crash without losing reserve status — as after Bretton Woods and in the 1970s.
⚡
Her greatest concern: the "this time is different" syndrome around AI — more immediate and harder to gauge.
🇨🇭
The counter-model: the Swiss debt brake — in this lecture not a footnote, but the answer to almost everything Reinhart warns about.

Stay informed.
HDH Macro Intelligence – next week.