Washington's NATO Loyalty Test Only Wrote Down What Trump Already Practices

The document carries a bland bureaucratic title, "Questions for NATO Allies & Other Key Stakeholders," and a purpose that is anything but. According to an Al Jazeera investigation published this week¹, the Pentagon sent 31 NATO allies a questionnaire probing whether they have publicly supported American foreign-policy priorities, whether they align with Defense Secretary Pete Hegseth's stated approach to the use of force, what restrictions they place on US access to their bases, and whether they endorse the administration's "NATO 3.0" concept of a Europe that defends itself with less American help. Bloomberg reviewed a version of the document² and reported that it also asks whether allies have "been publicly supportive" of US policy, whether they buy from American defense contractors, and whether they oppose regulations that inhibit contracts with US firms.

The timing is what turns an awkward survey into a strategic question. The questionnaire circulated during a six-month Pentagon review of US forces in Europe, due to wrap up around December, which Hegseth announced in Brussels on June 18 with the words, "It's a review that some countries will fail, and others will pass with flying colors"⁴. He paired the announcement with fury at allies that had denied US forces basing and overflight for strikes on Iran. Weeks earlier, Reuters reported³ that Washington planned to shrink its contribution to the NATO Force Model, the pool of forces each ally pre-identifies for activation in a war or major crisis, alongside cuts of roughly 5,000 troops from Europe and a canceled Army brigade deployment to Poland.

So the question worth answering is whether the United States is quietly converting NATO into a two-tier alliance, with military protection flowing to the politically aligned. On the legal layer, the answer is no. On the operational layer, where deterrence actually lives, I think the answer is already yes, and the decisive evidence is not the questionnaire at all. It is the president's own public record. The questionnaire did not build a loyalty system. It is the paperwork for one that was already running.

Start with the case for calm, because it is genuinely strong. American pressure on Europe to spend more and do more is as old as the alliance; Robert Gates delivered his famous warning about free-riding in 2011, and the 2 percent spending target dates to 2014. The current force review was publicly announced and publicly justified as a strategic shift toward Asia under the NATO 3.0 framework introduced by Undersecretary of Defense Elbridge Colby in February⁵, not sprung as a covert punishment scheme. NATO Secretary-General Mark Rutte called the reduction in US crisis commitments something "to be expected"⁶ as the alliance weans itself off one ally. And on July 8, all 32 members, Trump included, signed the Ankara Summit Declaration⁷ reaffirming an "ironclad commitment to our collective defence under Article 5." Even the senior European official who leaked concerns to Al Jazeera conceded the questionnaire "doesn't explicitly condition US military presence on political loyalty"¹.

Here is why that account, reasonable as it sounds, no longer fits the facts. In May, the administration announced it was pulling at least 5,000 troops from Germany after Chancellor Friedrich Merz said the US was being "humiliated"⁹ in its war with Iran. Days later, Trump announced on Truth Social that Poland would receive 5,000 additional troops, "Based on the successful Election of the now President of Poland, Karol Nawrocki, who I was proud to Endorse"⁸. Defenders of the routine-management thesis note that shifting weight from Germany to Poland tracks geography: Poland borders the war, Kaliningrad, and Belarus, and every EUCOM realignment since 2004 has drifted east. True. But Trump did not cite Kaliningrad. He cited an election result he had personally endorsed, in his own words, as the basis for a troop deployment. When the commander in chief states publicly that force allocation rewards political alignment, a questionnaire asking whether allies have praised US policy stops reading like a capabilities audit and starts reading like an answer key.

The mechanism matters here, because it explains how an alliance can be tiered without anyone touching the treaty. Article 5 obliges each member to take the action it deems necessary when an ally is attacked; it has never specified which squadrons show up, or when. Deterrence is delivered through force posture, the actual arrangement of troops, bases, pre-committed units, and access rights, and through the Force Model commitments that tell NATO planners what they can count on. Those are exactly the things being cut: NOTUS reported¹⁰ that Washington planned to pull back roughly a third of the fighter jets and strategic bombers it would activate for Europe in a crisis. The Ankara declaration's ironclad language cost the administration nothing to sign. The Force Model reductions cost Europe real combat power, and they were decided in the same building that drafted the loyalty questionnaire.

I will grant the skeptics their best remaining point: nobody has yet documented a case where a specific ally's questionnaire answer traceably changed a specific allocation, and the Pentagon has confirmed no scoring mechanism. My honest reading is that the questionnaire functions less as a scorecard than as case-building. The same European official told Al Jazeera it reflects a Pentagon effort to assemble justifications for targeted, almost punitive force reductions. That is arguably worse than an explicit scoring matrix, because a matrix could at least be audited. Discretion dressed up as data cannot.

What actually constrains this system is Congress, not the treaty. The FY2026 defense bill prohibits reducing US forces in Europe below 76,000¹¹, against roughly 85,000 there now, blocks divestment of EUCOM installations, and requires that NATO's Supreme Allied Commander remain an American. A Shaheen-led provision forces the defense secretary to certify consultation with allies¹² before dipping below the floor for more than 45 days. That floor is real, and it applies to every ally equally. But everything above the floor, and everything qualitative below it (which enablers stay, which crisis commitments hold, whose airfields get priority), remains executive discretion. The loyalty logic operates precisely in that discretionary band.

Note what the test is no longer about. Europeans are, for the first time, largely passing the money exam: allies committed at The Hague to reach 5 percent of GDP on defense and security by 2035¹³, and European members and Canada added more than $139 billion in core defense investment in 2025 alone. Gates complained about budgets. This questionnaire asks about praise, base permissions for a non-NATO war with Iran, and purchases from American contractors. The conditionality has migrated from burden-sharing, where it was legitimate and bipartisan, to political fealty, where it is neither. Europe's accelerating rearmament push is best understood as rational insurance against exactly that migration.

To conclude that this is all routine alliance management, you would have to believe that Trump's Poland announcement meant something other than what it plainly said, that Germany's drawdown just happened to follow Merz's criticism, and that a Pentagon survey asking allies whether they have publicly supported the president's foreign policy is gathering the data for a neutral capabilities review. Each belief is individually possible; together they require ignoring the pattern the administration keeps announcing in plain sight. The two-tier alliance needs no founding document, only discretion, and discretion is exactly what the December review will formalize.