Synthos Research · Frameworks · Thesis Snapshot

Eurodollar University: how Jeff Snider actually thinks

This is not a profile. It is a working model of the Eurodollar University worldview — its causal beliefs, its own stated gates, what it changed its mind about (dated), and where it is silent — reconstructed from dated claims and tested against a holdout window the model never saw. Snider is this council's deflation-first dissenter: where most tracked voices debate how much inflation the fiscal era produces, his model says the deep current runs the other way — a global dollar shortage, transmitted through the offshore eurodollar system the Fed does not control, turning every energy shock into a deflationary tax. On its first out-of-sample test this Framework scored 90.9% direction fidelity — the highest first-eval score of any macro Framework we have published — on a small, fully disclosed 11-test sample.

A Framework may only believe what the claim record can prove · what a Framework is · how voices earn tracking · methodology

direction fidelity: 90.9%
claim record: Apr → Jul 2026
causal graph: 22 edges
recorded mind-changes: 4

Model fidelity: 90.9% direction on 11 out-of-sample tests — the strongest first eval yet, on the youngest corpus

What that number means: we froze this model on claims through June 22, 2026, asked it to predict how Eurodollar University would react to real events from the following four weeks, and graded those predictions against the claims actually made in that window — claims the model had never seen. Two disclosures up front: the corpus is young (months, not the years behind Visser or Alden), and because it is so recent the holdout window sits close behind the training window. Both make this a strong first eval, not a long track record.

10 of 11 gradeable predictions got the direction right. The pass bar, set before grading, was 60%.
The best pair was same-day and near-verbatim: on the June CPI print the model called energy a relative-price shock, core flat, breakevens pricing disinflation — and his actual same-day claims said core fell fractionally, TIPS breakevens plummeting, demand destruction rather than second-round inflation — direction, mechanism, vocabulary and conviction all matched at once.
Correct silences: 100% this round — equity single-names, an AI product launch and a labor print all correctly drew no answer rather than an invented one.
The one miss, named: when Brent fell to $72, the model fired “dollar-shock easing” at high conviction; his actual claims that week said global dollar conditions were still tightening even as the energy shock faded. The model eased and escalated the same structural variable depending on oil's direction — an internal inconsistency his actual worldview does not have. Logged as the miss it is.
Zero invented mechanisms: no killed edge ever fired in prediction.
Data integrity: Eurodollar University is a two-voice show. Claims voiced by the co-host rather than Snider were excluded from grading rather than scored as his view.

1 · How the world works, according to Snider

Six formal invariants anchor the graph — and every one of them inverts a mainstream reflex.

Energy shocks are a deflationary tax, not inflation

Formal invariant · held across the full record, never reversed

The signature contrarian belief, held through a live oil crisis: an oil spike does not start an inflation regime — it squeezes income, which cuts spending, which kills the expansion.

raise prices briefly then choke activity into recession — 2026

energy squeezes income, income slows spending, hitting revenue then margins then labor — 2026

The dollar rises on shortage, not strength — and the Fed cannot fix it

Formal invariant · the eurodollar spine everything else hangs from

Dollar strength is not a vote of confidence in America; it is stress in the offshore dollar system. And the Fed's tools do not reach that system.

An energy shock becoming a dollar shock is squeezing the global economy — 2026

Fed bank reserves aren't globally usable eurodollar liquidity — 2026

Rate moves are reaction, not stimulus

Formal invariant

Central banks follow the economy; they do not steer it. Hikes don't stop the credit cycle, cuts don't start one — policy is confirmation of what already happened.

rate hikes don't stop the cycle and cuts don't create it — 2026

officials admitting the economy is weakening and chasing the bond market — 2026

The market's own gauges beat the narratives

Formal invariant · also his named attention channels

He reads TIPS breakevens, SOFR futures curve shape and the 2s10s — not CPI headlines or Fed speeches.

TIPS breakevens price a classic energy shock, not a sustained inflation breakout — 2026

SOFR futures show a 'frown': modest near-term rate hikes — 2026

Private credit is a self-reinforcing bust already underway

Formal invariant

Not a prediction of a crash day — a staged process: hidden repo leverage and mark-to-myth pricing force recognition in waves.

damage moves from private funds to public markets to the real economy — 2026

spreads to the real economy not as a Lehman-style collapse — 2026

2 · His strongest causal chains

The energy→dollar doom loop

Pricier oil forces dollar demand just as offshore lenders pull back: pricier, scarcer oil forces more dollar demand while eurodollar providers stop recirculating
Asia gets squeezed first: these economies need dollars that are becoming more expensive and scarcer
The squeeze feeds itself: it creates more dollar demand, which weakens the currency further
Intervention fails on a timer: the yen drifts right back below 160 within weeks — central-bank intervention has a shrinking half-life and the pressure returns

The private-credit recognition cascade

Yield-starved insurers built the bubble: annuity obligations pushed them down the credit spectrum into riskier — insurers enabled and funded the private credit bubble
Hidden leverage amplifies it: repo amplifies already-risky assets
Recognition arrives in stages: as price discovery forces recognition, markdowns, non-accruals and PIK will keep rising
The run gate: Once redemptions exceed new subscriptions, gates confirm limited liquidity
Then it reaches your portfolio: forced selling and risk-off spill into equities and stock prices go lower

The Trichet replay: hike into oil, then reverse

If the Fed reads oil as inflation and hikes, it repeats the classic 2008 error: it will very likely have to turn around and cut
The curve is already saying so: the classic policy-mistake signal heading toward reinversion
Destination: as recession and deflation reassert, Treasury yields will collapse again

3 · What would change his mind — his own stated gates

If this happens

The shock spreads into wages

His own test for when oil would become real inflation: for oil to become inflation the shock must spread: wages chasing prices. Absent that spiral, it stays a deflationary tax.

If this happens

A melt-up he already priced

If the Iran conflict resolves and central banks signal no hikes, he allows an S&P blow-off — and says in the same breath it would not last.

If this happens

Energy lingers

the longer elevated energy prices linger, the higher the odds of recession or crisis — duration, not level, is his trigger.

The deep falsifier

A price level that never comes back

He concedes shocks can leave scars: the price level permanently 'phase shifts' higher and never returns — but the rate of inflation still dies. Sustained re-acceleration without a wage spiral would break the framework.

4 · What he changed his mind about — dated

Four recorded reversals — every one a tactical price call. The doctrine underneath did not move once. That split is the temperament: doctrinal core frozen, trading calls fluid.

Then · May 2026Oil over $100: gasoline toward $5 and headline CPI in the threes by July or August

Now · June 2026Strip the war premium and oil is worth ~$50 — then a partial re-reversal: prices bounce and rally in an 'aftershock'

Trigger: Iran-conflict de-escalation. Note what did not change: the deflationary framing held through both flips.

Then · April 2026Base-case chance of an ECB hike by summer

Now · May 2026the probability is diminishing as growth weakness becomes visible

An attribution nuance, disclosed The record also shows a bearish-to-bullish S&P flip — from equities ignoring the private credit crisis, which keeps escalating to a short-covering melt-up call. But the bullish near-term calls in this record are co-host-voiced; the disconnect doctrine — Manufacturing PMI hitting a 4-year high (US 54.0) is artificial front-loading — is the Snider-voiced constant.

5 · Where he is silent

Single stocks and equity fundamentals — the model has no valuation machinery, and in eval it correctly refused to answer where he has no view. He dismisses market-level strength with a basic rule of finance reflex rather than analyzing companies.
Crypto — effectively absent as its own axis; in the holdout, risk-off gold-only-haven framing carried the weight. There is no Bitcoin mechanism in this graph.
AI as technology — he engages AI only as a financing story: AI erodes recurring-revenue assumptions and lenders now market 'less software exposure'. Capability, products, chips: silent.
China appears once, as an export deflation valve: Weak domestic demand forces China to dump excess industrial output — no broader geopolitics machinery.