Arms transfers are the alignment data nobody reads
States lie about their alignments constantly and buy weapons honestly. Procurement is a decade-long dependency, chosen deliberately, and it is published.
Alignment is usually assessed from the things states say and the rooms they sit in: communiqués, votes, summit attendance, the language of official statements. All of it is cheap. A vote costs nothing, a communiqué costs less, and both are routinely produced to obscure a position rather than to declare one.
Weapons procurement is not cheap, and that is what makes it useful.
Why a purchase is a commitment
Buying a major weapons system is not a transaction. It is the beginning of a relationship that runs for decades and is very hard to exit.
The platform needs spare parts from the manufacturer. It needs software and munitions from the same source. Crews and maintainers are trained on it, in a doctrine that comes with it. Integration with existing systems locks in a standard. Financing frequently ties the buyer to the seller's export credit apparatus for years. By the time the platform is delivered, the buyer has accepted a dependency that persists across several changes of government in both countries.
States understand this perfectly well when they sign. Which is why the choice of supplier is one of the most deliberate alignment signals a government ever sends, and why it is a far better guide to where a state actually sits than anything it says at a summit.
A vote at the UN costs nothing. A thirty-year spare parts dependency costs everything, which is exactly why it tells you more.
Reading the data honestly
The transfer data is public and has three traps in it that will produce wrong answers if ignored.
Orders are not deliveries. The gap between a signed contract and a delivered platform is routinely years, and the political meaning attaches to different moments. The order tells you what the government decided; the delivery tells you what it actually received, which may be after the government that ordered it has fallen. Both matter, and conflating them produces a picture that is confidently out of date.
The measure is not a price. The standard datasets express transfers in a constructed unit designed to capture military capability transferred, not the money paid. This is the right choice for comparing across time and across deals with wildly different financing, and it means the figures are not dollars and should never be reported as if they were. Nothing generates false authority faster than a capability index presented as a sales figure.
Diversification is a strategy, not indecision. A state buying from three suppliers is often read as hedging or incoherence. Frequently it is deliberate: distributing dependency so that no single supplier can impose terms in a crisis. India has run this approach for decades. Several Gulf and Southeast Asian states run versions of it. The pattern of who supplies which capability is more informative than the totals, because states tend to diversify the systems they could survive losing and concentrate the ones they cannot.
Twenty suppliers appear in that decade in all. Read as a single number the mix looks like dependence on Russia, and a great deal of commentary stops exactly there. Read as a distribution it is something else: a dominant historical supplier at just under half, and three others each large enough that none of them can set terms alone. That is what deliberate diversification looks like on the register: a state buying room to manoeuvre, priced in decades of spare parts.
| The trap | What the data actually records | What it gets read as |
|---|---|---|
| Orders vs deliveries | A contract signed, or a platform received, years apart | One event |
| The unit of measure | Military capability transferred, by construction | Dollars paid |
| Diversification | Dependency distributed deliberately | Hedging, or incoherence |
What the pattern shows that the statements do not
Three things become visible in transfer data and are hard to see any other way.
Switching, early. A state that has bought exclusively from one supplier for thirty years and places a significant order elsewhere has made a decision that will not be announced. The order precedes the announced realignment, often by years, because the procurement decision is made when the strategic assessment changes and the diplomacy catches up afterward.
The capability gap it is worried about. The category matters more than the volume. Air defence purchases say something different from transport aircraft. A state buying long-range strike is answering a different question from one buying coastal patrol. Procurement is a state describing, in the most expensive language available, what it thinks is going to happen to it.
Who is willing to sell. The supplier's decision is itself a signal, and often the clearer one. Export approvals are political acts. A supplier state agreeing to transfer a sensitive capability is making a statement about the relationship that it may be unwilling to make in words.
An open register
The unusual thing about this evidence is not that it is hard to obtain. The datasets are published, the categories are documented, the methodology is explained at length by the institutions that compile it, and none of it costs anything.
It is that almost nobody in finance reads it. Procurement sits in the security-studies literature, and exposure sits in the credit and market literature, and the two rarely occupy the same desk. That is precisely why a signal that precedes sanctions exposure, payment restrictions and counterparty risk by several years goes largely unpriced.
Whether that gap closes is an interesting question. Alignment data has none of the properties that usually make an edge disappear: it is not fast, not proprietary, and not automatable into a signal. It just requires somebody to read the security literature and the credit file as one problem.
← Back to all insights