Prediction — EU mid-cap defense primes outperform US large-cap primes by ≥10pp over 12 months

4 May 2026 · prediction · resolves on hard date

Thesis (one line)

The marginal-buyer geography for the Western defense industrial mobilisation is European mid-cap and South Korean, not US large-cap; the equity repricing has not yet caught up, and over a 12-month forward window the EU mid-cap basket will outperform the US large-cap basket by at least 10 percentage points on total return.

Origin

Filed issue-03 — The defense industrial mobilisation has a ceiling on 2026-05-04. The prediction is the staked falsifiable claim that closes the issue.

Falsification criteria (the hard line)

Resolution date: 2027-05-04 (12 months forward).

EU mid-cap defense basket (equal-weighted):

BAE Systems plc (LSE: BA)
Rheinmetall AG (XETRA: RHM)
Saab AB (STO: SAAB-B)
Leonardo S.p.A (BIT: LDO)
Thales SA (EPA: HO)

US large-cap prime basket (equal-weighted):

Lockheed Martin (NYSE: LMT)
RTX Corporation (NYSE: RTX)
Northrop Grumman (NYSE: NOC)
General Dynamics (NYSE: GD)

Pass criterion (prediction RESOLVES TRUE): EU basket 12-month total return (price + dividends, denominated in EUR) ≥ US basket 12-month total return (denominated in USD, then converted to EUR at 2026-05-04 spot for fair comparison) + 10 percentage points.

Fail criterion (prediction RESOLVES FALSE): EU basket total return < US basket total return + 10 percentage points.

Edge cases:

If a basket constituent is acquired or delisted mid-window, replace with the largest market-cap defense prime of the same geography on the day of delisting (e.g. if Saab is acquired, replace with Babcock International). Document the replacement in the resolution note.
Currency conversion uses ECB EUR/USD reference rate at 2026-05-04 (close) for the US basket; intermediate rate movements are part of the EU-vs-US relative trade and not adjusted out.

Why this is the right falsifier

The thesis claims the marginal-buyer geography is European mid-cap and Korean, not US large-cap. The EU mid-cap basket isolates the European geography exactly. The US large-cap basket isolates the legacy US-prime trade the consensus is overweight. A 10-percentage-point gap over 12 months is a material relative-return spread — well above noise on equal-weighted baskets — and is consistent with the claimed magnitude of the marginal-buyer dynamic.

A 10pp spread is not maximalist. The thesis would survive a smaller spread as "weakly correct." Setting the bar at 10pp means the claim is genuinely contestable: if the EU basket only outperforms by 4-8pp, the prediction resolves FALSE even though the direction of the thesis was right. That is the right asymmetry — the thesis is not just a directional bet, it is a magnitude bet.

(Korean exporters Hanwha / KAI / LIG Nex1 are mentioned in issue-03 as a complementary leg; they are NOT part of this prediction's basket because their currency + state-control dynamics are different enough to deserve a separate falsifier in a future issue.)

Anti-cherry-picking notes

Basket constituents were chosen before observing 12-month forward returns. They are the 5 largest EU mid-cap defense primes by 2026-05-04 market cap and the 4 largest US large-cap primes (the "big four" defense primes per S&P GICS classification).
The 10pp threshold was chosen before observing the resolution window.
Equal-weight chosen instead of market-cap-weight because the thesis is about the geography of the marginal buyer, not the size of any single firm.
No survivorship bias filter applied — if a constituent zeros, that's part of the basket result.

Resolution note (filled in 2027-05-04)

pending — will be appended at resolution date with the actual numbers, the resolved outcome, and a "what I got wrong / what I got right" post-mortem.