INTERMARUM: THE FREE-WILL COALITION, PRICED
The attraction of the free-will coalition is stated the same way every time: an alliance architecture that rejects both hierarchical dependency on a distant patron and absorption into a neighbouring sphere. A third option. Membership without a master.
The proposition is coherent as politics and it is a specific, testable claim as economics, because a military alliance is a machine for producing one particular kind of good, and that kind of good has a known provision problem. The third option is not refuted by pointing at history. It is constrained by a threshold, the threshold is a fraction, and the fraction can be calculated in one line.
The mechanism
Deterrence is non-excludable within an alliance. If an integrated air and missile defence network makes an attack on the coalition unattractive, it does so for every member, including those that paid nothing. There is no technical means, and no politically survivable means, of leaving one member uncovered in order to punish it for arrears. The good arrives for everyone or it does not arrive.
That is the definition of a public good, and the standard result in collective-action theory applies without modification: where benefits are shared and costs are not, the good is provided only if some single member’s private benefit exceeds the entire cost of provision. Everyone else’s rational move is to wait. [UNVERIFIED: this result is stated here from general familiarity with the collective-action literature; no primary text was accessed for this note.]
That result is usually presented as a warning about free-riding. It is more useful read the other way round, as a definition. A hegemon is not a country with an attitude. A hegemon is whichever member’s share of the benefit crosses the threshold at which building the thing alone beats waiting. Where no member crosses it, there is no hegemon, and there is also no good.
The threshold
Model a coalition of 10 members. Assume an integrated defensive capability that costs 10.0 units per year to build and operate, and delivers 18.0 units per year of benefit to the coalition as a whole. Net value to the coalition is 18.0 - 10.0 = +8.0 units per year. It is plainly worth building.
Assume benefit accrues to each member in proportion to its weight in the coalition. A member will build it unilaterally when its own share of the benefit exceeds the full cost:
share x 18.0 > 10.0, therefore share > 10.0 / 18.0 = 55.6%
| Member weight | Private benefit (weight x 18.0) | Cost if built alone | Builds alone? |
|---|---|---|---|
| 35% | 6.30 | 10.0 | No |
| 45% | 8.10 | 10.0 | No |
| 55% | 9.90 | 10.0 | No |
| 55.6% | 10.01 | 10.0 | Threshold |
| 60% | 10.80 | 10.0 | Yes |
A coalition designed so that no member exceeds roughly 55% is a coalition designed so that nobody builds the capability. The design succeeds precisely at the thing it set out to do, and the cost of that success is a good worth +8.0 units per year that does not exist.
Note what this is not. It is not a claim that the members are selfish, disloyal, or insufficiently committed. Every member in the table above is correct to decline. A 35% member who builds alone spends 10.0 to receive 6.30 and is worse off by 3.70 for having been public-spirited. The failure is structural and it survives goodwill.
The three escapes, and where each one goes
There are exactly three ways out and each is worth pricing.
Escape one: build it nationally. Each member provides its own segment, retaining full sovereignty. Model a national segment at 1.45 units per member, so 10 x 1.45 = 14.5 units against 10.0 for the integrated system, a scale penalty of 4.5 / 10.0 = 45%. Because the segments are procured separately they interoperate imperfectly, so model realised benefit at 13.0 rather than 18.0.
Net value: 13.0 - 14.5 = -1.5 units per year.
Sovereign provision has converted a good worth +8.0 into a liability worth -1.5, a swing of 9.5 units, while allowing every member to report a defence capability. This is the most popular escape because it is the only one that requires no negotiation.
Escape two: a binding cost-share treaty. Ten members pay proportionally into a common fund totalling 10.0 units. The 35% member pays 3.50 and receives 6.30, which is strongly positive. Even the 3% member pays 0.30 and receives 0.54. On paper every member is better off and the treaty is stable.
The instability is not in the arithmetic, it is in year three. A defector saves its full contribution and loses only its marginal effect on a good it still receives, because the good is non-excludable. Deterring defection therefore requires the ability to exclude a defector from coverage, and the ability to exclude requires somebody to hold the switch. Whoever holds the switch is a hegemon in function, whatever the treaty calls them. Escape two does not remove hierarchy. It relocates it into an annexe and renames it a secretariat.
Escape three: an external guarantor. The coalition contracts the public good from a power outside it. This works, and it is the arrangement the free-will coalition was formulated to avoid. It has the additional property of reintroducing the dependency at a worse price, because an external guarantor’s benefit share of a coalition it does not belong to is zero, so its provision is a favour rather than an interest, and favours are repriced annually.
The strongest counter-argument
The public-good frame may simply be the wrong frame, and this is the serious objection.
A large fraction of what an alliance actually supplies is excludable. Intelligence products can be withheld from a specific capital. Interoperability certification can be revoked. Joint procurement contracts can be signed without a given member. Common logistics standards, shared repair depots, harmonised munition specifications, exercise access, and officer-exchange programmes are all club goods, not public goods, and clubs have no provision problem at all, because exclusion is cheap and immediate.
Price one. Ten members each procuring 100 units of a common item at a unit price of 1.00 spend 100.0 each, or 1,000.0 in total. A combined order of 1,000 units attracts a volume discount; model it at 14%. Each member now pays 86.0, saving 14.0. A member that refuses to contribute to the club’s administration is simply left off the contract and pays 100.0. Provision is stable, no hegemon is required, and nobody had to trust anybody.
This is the honest core of the free-will coalition, and it is real. Central Europe can build a genuine club without a patron, and the club can be substantial: procurement, standards, logistics, industry, intelligence exchange, rail gauge, ammunition. All of it is achievable at 10 roughly equal members.
What the club cannot supply is the non-excludable part. You cannot exclude a member from the deterrent effect of a capability that exists on the map. So the coalition converges on a stable equilibrium in which it is excellent at capability and structurally incapable of deterrence, which is a description that fits several real organisations and explains why they are simultaneously praised as useful and dismissed as unserious.
Consequence, and the test
The free-will coalition is not a fantasy. It is a club being marketed as an alliance, and the distinction is not rhetorical: it is the difference between goods you can withhold and goods you cannot.
Three things follow. A coalition of near-equals should build everything excludable and should not promise anything non-excludable, because the promise will be believed by its own populations and by nobody else. Any member that does approach the 55.6% threshold will be accused of seeking hegemony at exactly the moment it becomes capable of supplying the thing the coalition was formed to obtain. And the recurring proposal to solve this with a stronger secretariat is a proposal to create a hegemon with no population, no army, and no ability to be punished.
The test is clean. If a coalition of near-equals, none holding a majority of benefit share, ever fields a jointly funded, non-excludable strategic capability and sustains it through a full budget cycle in which at least one member had a domestic incentive to defect, the model in this note is wrong and should be discarded. The earliest observable signal would be a common fund whose contributions survive a member’s national election without renegotiation.
Nothing in the arithmetic says the third option is undesirable. It says the third option has a price, that the price is denominated in the goods you agree not to promise, and that a coalition which will not name that price has not chosen a third option. It has chosen not to decide.