The Arsenal Account

The problem

Proposal № 010 argued that Americans should be shown, on one page, every year, what the wars cost. Proposal № 019 argued that when a war ends its funding should lapse into the endowment rather than being quietly absorbed into the base budget.

Both are proposals about restraint, and this one belongs with them, because it starts from the same premise: the cheapest war is the one that does not happen, and the second cheapest is the one that ends quickly because one side was obviously going to lose.

The United States spends on the order of $850 billion a year on defence, more than the next several countries combined. It is therefore surprising to discover how thin the actual production base underneath that number has become.

Before 2022, American production of 155mm artillery shells ran at roughly 14,000 a month. When allied demand arrived, the entire Western alliance discovered that its combined output was a fraction of what a mid-sized land war consumes, and the effort to scale toward 100,000 a month has taken years and billions of dollars and is still in progress. Stinger production had effectively stopped; the line had to be restarted, with engineers recalled from retirement to explain components that were no longer manufactured.

Shipbuilding is worse. By tonnage, Chinese commercial and naval shipbuilding capacity exceeds American capacity by a factor that US Navy intelligence has put in the hundreds. Virginia-class submarine deliveries have run near one per year against a requirement closer to two, before accounting for commitments to allies.

The prime contractor base has consolidated from more than 50 firms in the early 1990s to about five.

The cause is not mysterious and it is not primarily greed. It is the appropriations cycle. Congress funds procurement one year at a time. A firm asked to add a production line, hire a workforce and qualify a supply chain is being asked to make a multi-year capital commitment against a one-year contract that may not be renewed. The rational response is not to build the capacity. Every executive who declined to expand a line in 2015 was making the correct decision on the information available.

So the country holds an enormous budget and a fragile industrial base, which is the specific combination that invites the miscalculation it is supposed to prevent.

The proposal

A separate, multi-year capital account that buys production capacity rather than units. Warm lines, deep magazines, and co-production with allies, funded and accounted for as insurance.

How it would work

1. Capacity is the product. The account contracts for a sustained minimum rate of production over 5 to 10 years for a defined list of critical munitions, paying for the line's existence rather than only for delivered units. A manufacturer that knows the floor is guaranteed for a decade will finance the expansion privately, which is the entire point.

1. Multi-year and block buys, as the default. Multi-year procurement authority for munitions, extended and made routine rather than exceptional. Historical savings from multi-year contracting run 10 to 20 percent against annual buys, so this is cheaper per unit as well as more reliable.

1. A funded magazine target. A published stockpile objective for each critical munition, expressed in days of high-intensity consumption rather than in units, with the account funded to close the gap on a schedule. Days of supply is the metric that matters and it is the one that is not published.

1. Co-production with allies. Joint production lines with treaty partners, with shared qualification standards, so that allied demand deepens a common industrial base rather than competing for the same scarce output. This is the Open Hand version of deterrence: partners who can make their own shells are partners who do not need an American expeditionary force.

1. Second sources, deliberately. No critical munition dependent on a single supplier or a single sub-tier component. Where the market will not sustain two producers, the account pays for the second, and calls that cost what it is: insurance against a single plant fire deciding a war.

1. Accounted for honestly, and it lapses. The account is reported separately in the Peace Ledger of № 010, so citizens can see what deterrence costs as a distinct line rather than buried in a topline. And per № 019, it is capacity funding: it does not become a permanent entitlement for the base budget, and unspent authority lapses to the Fund rather than being reprogrammed.

The numbers

Munitions capacity. Closing the gap on artillery, air defence interceptors, long-range precision fires and their critical sub-components is an investment on the order of $15 to $25 billion over several years, plus a sustained annual rate of $8 to $12 billion.

Submarine and shipbuilding industrial base. Workforce, supplier qualification and yard capacity is a further $5 to $8 billion a year sustained, and this is the item least amenable to money alone, because the binding constraint is welders and machinists rather than dollars. It runs directly into № 028.

Call it $20 billion a year at steady state. That is roughly 2 percent of the defence budget.

Now the comparison that makes the case. The post-9/11 wars cost approximately $8 trillion, the figure № 010 exists to put in front of people. $20 billion a year for fifty years is $1 trillion. An insurance policy costing an eighth of one war, which reduces the probability of that war, is not a close call on expected value.

And the honest counterpoint, stated in the same breath: this is precisely the argument every defence programme in history has made, and the deterrent effect is not measurable. Nobody can produce a counterfactual in which the war did not happen because the magazine was deep. We are making an argument from plausibility, not from evidence, and readers should discount it accordingly.

The honest objections

"This catalog published a War Budget Sunset and a Peace Ledger, and now proposes $20 billion a year more for weapons. Pick one." The most important objection because it goes to consistency, and it deserves a direct answer rather than a reconciliation. № 019 targets war outlays, the supplemental appropriations that fund active operations and then never lapse. This targets production capacity in peacetime. The distinction is deliberate: a country with deep magazines and no war spends less than a country with empty magazines that goes to war, and the historical pattern of American conflict is that a thin base is filled by emergency supplementals at wartime prices with no competition. Buying capacity in peace is the fiscally conservative option. That said, a reader who thinks any increase in defence industrial spending inevitably ratchets the base budget upward has history on their side, which is exactly why item 6 exists and why it should be the hardest-drafted clause in the bill.

"Guaranteed multi-year contracts to five firms with no competition is a recipe for cost-plus capture." Well founded, and consolidation makes it worse. Mitigations: item 5 mandates second sources, which is the only real discipline available in a market this concentrated; contracts should be fixed-price where the product is a known munition rather than a development programme; and capacity payments should be tied to audited, demonstrated surge rate rather than to declared readiness. None of these fully solves a five-supplier market. The alternative on offer is a one-supplier market.

"Arming allies means American weapons in wars America did not choose." True and unavoidable, and item 4 makes it more likely rather than less. The mitigation is that co-production is with treaty partners under existing transfer controls, not general export promotion, and that end-use restrictions have to be real and enforced. This is a genuine cost of the proposal and one over which reasonable people will differ on principle rather than on facts.

"Warm production lines that are never used are waste." Yes, in the same sense that an unclaimed insurance policy is waste. The relevant question is the premium against the loss, and the loss here is a war fought with insufficient ammunition, which is the scenario in which wars become long. A country that keeps a strategic petroleum reserve and a vaccine stockpile already accepts this logic.

"Twenty billion a year would do more good as twenty billion a year of anything else in this catalog." Possibly. It would fund the rural care floor of № 036 three times over. We publish both and we are not going to pretend the budget constraint is imaginary. The argument for this one is that it is the only item in the catalog whose failure mode is a war, and that no other line item survives that failure mode intact.

Sources

• US defence budget authority; Department of Defense comptroller and CBO analyses (cbo.gov)
• 155mm artillery production rates, pre-2022 baseline and expansion targets; Army acquisition statements and CRS reporting (crsreports.congress.gov)
• Stinger production line restart and component obsolescence; DoD and manufacturer testimony
• Comparative shipbuilding capacity assessments; Office of Naval Intelligence briefings on Chinese shipbuilding output
• Virginia-class submarine delivery rates against programme requirements; Government Accountability Office shipbuilding reviews (gao.gov)
• Defense prime contractor consolidation from over 50 firms in the early 1990s; DoD State of Competition within the Defense Industrial Base (2022)
• Multi-year procurement savings estimates of 10–20 percent; CRS, Multiyear Procurement and Block Buy Contracting
• Post-9/11 war costs of approximately $8 trillion; Brown University Costs of War Project (watson.brown.edu/costsofwar)
• Proposals № 010 (The Peace Ledger); № 019 (The War Budget Sunset); № 028 (The Apprentice Republic); № 030 (The Compute Reserve); № 036 (The Rural Compact)