Issue:
September 2026
The opening of Tokyo’s arms export market this year has been framed as a story about pacifism. The business story is going largely unreported, and it started five months earlier.
On 20 November 2025, Chief Cabinet Secretary Minoru Kihara confirmed that Japan had completed delivery of PAC-3 interceptors to the United States. The interceptors, built by Mitsubishi Heavy Industries under license from Lockheed Martin, came out of Air Self-Defense Force (SDF) stocks, under an export contract signed in July 2024. Kihara would not say how many Japan sold, or when, citing military operations. The purpose was to top up American inventories drained by support to other countries, principally Ukraine. Jiji Press headlined it as indirect support for Kyiv.
That was five months before the Takaichi cabinet's decision of 21 April 2026, which abolished the five mission categories that had confined exports of Japanese-built finished equipment and was reported everywhere as the moment Japan became an arms exporter. The transaction that best illustrates what the new policy is for had already happened, under the old rules, and almost nobody connected the two.
Japan is not entering the global arms trade. It is entering a narrow market: replenishing the magazines of allies who have emptied theirs arming somebody else. Tokyo cannot sell into a war: transfers of lethal equipment to countries judged to be currently engaged in combat are barred unless special circumstances arise from Japan’s own security needs, which the four-minister National Security Council meeting finds. But it can restock the arsenal of the country doing the arming. That is a derivative market, and Japan has already traded in it once. Whether Tokyo becomes an exporter or a subcontractor inside an American supply chain is open; the evidence points to the second.
The Military Order Book
The American demand is not a forecast. Over the past decade the United States bought an average of eighty-six Tomahawks from Raytheon a year. In the thirty-nine days of the air and missile campaign against Iran, to the ceasefire of 7 April 2026, it fired more than a thousand. Operation Epic Fury was declared over in early May. Analysts put THAAD expenditure at up to 290 interceptors, with reserves restored only by late 2029, and current holdings at fewer than 1,000 Patriot interceptors and about 250 THAAD.
Washington has responded with seven-year framework agreements: up to $35 billion to Lockheed Martin for THAAD, quadrupling output from 96 interceptors a year to about 400; more than $58 billion for PAC-3, lifting production from 600 a year to 2,000; five separate agreements with Raytheon covering Tomahawk, AMRAAM and the SM series. In June, President Donald Trump invoked the Defense Production Act, finding systemic constraints in the munitions industrial base without itemising them; Reuters supplied the specifics: solid rocket motors, igniters, guidance systems.
A government that commits to buy into the 2030s has concluded it cannot get the goods sooner. That is the demand signal Tokyo is looking at, and it has a seven-year floor under it.
But What Can Japan Supply?
Not much, for reasons that are not only industrial. In fiscal 2023 the Self-Defense Forces filled just 51 percent of their recruitment target, the lowest rate on record. A 2022 survey by the SDF’s own promotional magazine found 71.8 percent of young Japanese would not fight to defend the country if invaded, on a question offering no middle option. Recruitment has since recovered: hires were up 14.9 percent in the year to March 2026. But a defence industrial expansion and a rearmament of public sentiment are being asked of the same shrinking cohort.
Mitsubishi Heavy Industries builds roughly thirty PAC-3 interceptors a year. It cannot build more because the Ka-band seekers (advanced, high-frequency radar guidance systems mounted in the nose of missiles) are made by Boeing and are in short supply; industry sources expect no relief before 2027. Thirty a year, against an American production objective of two thousand.
The wider export record is thinner still. The International Institute for Strategic Studies reported in April that since the first liberalisation under the administration of Shinzo Abe in 2014, Japan’s exports of finished defence products amounted to three fixed air surveillance radars and one mobile radar, all to the Philippines. Twelve years, four radars — and the one showpiece deal since, the A$10 billion Australian frigate programme signed 18 April, gives Japan three of eleven hulls, the rest promised to Western Australia. A design licence with a starter order attached.
Japan’s own FY2026 budget request, ¥8.8 trillion in total, allocates roughly ¥313 billion to unmanned systems: three times the prior year, and the one line item pointed at a tier where Japan’s civilian base might compete without an American or Chinese chokepoint. Against the missile figures it is a fraction, but not a trivial one, and the constraint there is not a Boeing seeker Tokyo cannot get.
The loop closes in an awkward place. The US Commercial Service is now advising American firms that Japan’s export opening creates opportunities to supply components and subsystems to Japanese primes: Japanese industry has long depended on US inputs, and export-oriented production will increase that need. Washington reads Tokyo’s arms opening as a market for American parts. Given the seeker constraint, it may be right.
Media framing of the arms sale
The April decision was not ignored. It was covered heavily and argued fiercely, but through a single frame.
Every major daily editorialised. The Sankei Shimbun welcomed a decision that would strengthen allies and grow the defence industry; The Yomiuri Shimbun wanted the rules used to enlarge the roster of like-minded countries. The Asahi Shimbun called it the dismantling of a pillar of the peace state, The Mainichi and The Tokyo Shimbun warned that exports at scale invite an arms race, The Nikkei took the deterrence case as read and called strict monitoring indispensable. The Asahi, Mainichi and Tokyo Shimbun pressed one point in near-identical terms: the Diet had been left without a check. On organisers’ figures roughly 36,000 rallied outside the Diet on 19 April, about 30,000 on the 8th — both organised around constitutional revision, arms exports one grievance inside a larger one.
The polling is consistent and negative. NHK in March found 53 percent opposed against 32 percent in favour; a Kyodo poll in mid-May, weeks after the policy had already taken effect, still found 57 percent opposed against 37 percent in favour. A Defence Ministry official, speaking anonymously the day after the decision, conceded what the government would not: selling weapons to an America that had attacked Iran does make Japan complicit in that war, and the warning that “exception” becomes a route to supplying the US military deserves to be faced squarely.
So the constitutional argument was aired. As Corey Wallace observed in East Asia Forum, headlines focused, as they usually do with Japanese security shifts, on a conservative government abandoning the postwar commitment to pacifism.
What went unexamined is whether Japanese industry can supply at scale or is bound by an American component; whether the customer is, overwhelmingly, one customer; or whether Japanese firms add the missile or only its assembly. Almost nobody asked whether the November sale was the template for the new policy rather than a footnote to the old one.
The FCCJ came closer than most. On 2 April, nineteen days before the cabinet decision, the FCCJ hosted Ogi Hirohito and Inoue Rintarō from the Institute of Geoeconomics on “Building Up Japan’s Defense Industry”. Their framing was that Japan’s defence industry, which had for decades served one customer only (the SDF) was entering an “era of excess demand”. The instinct is right, and it arrived here first. But excess demand for what, and whose? Ogi and Inoue were describing Japan’s own procurement surge and China’s shadow over it. The sharper version: the excess demand Japanese industry can reach is American, runs through a single component supplier, and is capped there. Nobody appears to have filed the follow-up.
There is a structural reason for the media’s misinterpretation. The pacifism frame is available to any correspondent on any deadline: it has protest photographs, opposition quotes, a constitutional principle and a poll. The industrial question requires reading budget documents in two languages, and it yields a conclusion: a historic policy change may produce little for some years, harder to headline than 36,000 people outside the Diet.
It has an oversight dimension that ought to interest this readership. Japan’s 1976 ban on the export of military equipment was never legislated. It was a unified government view, stated to the Diet by Miki Takeo, extending restraint beyond the earlier principles. A cabinet decision could undo it. Under the 2026 rules the Diet is notified after the fact. Three opposition parties, Komeito among them, pressed for mandatory prior notification above a monetary threshold. They did not get it; the government said only that it would consider the idea. It has promised to monitor custody, security and loss-reporting procedures after transfer. Whether Japanese-built weapons can in practice be tracked to their end use is a different question, one no one has had to answer.
The Precedent Nobody Invoked
We have been here before. Between 1950 and 1953, American special procurement (tokuju) placed orders with Japanese suppliers to sustain the Korean War. Counted in dollars alone, figures calculated by Japanese economic historian *Yoshio Asai show exports bringing in about $400 million of dollar receipts in 1952 against $800 million from procurement; in every year from 1951 to 1954 procurement earned more dollars than exports did.
[was: amazon.co.jp/IMF8条国移行―貿易・為替自由化の政治経済史-浅井-良夫/dp/4818823899
Weapons manufacture was prohibited in occupied Japan until 1952, so through the boom’s first phase tokuju bought blankets, gunny sacks, barbed wire, trucks and repair work: an export-control regime determined the composition of what Japan sold to a war it was not party to. The boom disappointed while it was running, too: in January 1953 the US embassy in Tokyo cabled Washington that against a planned $350 million for the year, only $135 million had been contracted in six months. Both features - a control regime that shaped the product mix and a forecast that outran the contracts - carry into the present.
Two senior research fellows with the Economic and Social Research Institute, under Japan’s Cabinet Office, put Korean War procurement at one to two percent of GNP a year. Tokyo's August white paper argues that defence production and civilian industry now strengthen each other, a virtuous cycle. Across six hundred and ten pages it never once mentions tokuju. The one precedent available to it was decisive for foreign exchange and marginal for output.
The industry that delivered that boom was forbidden to make weapons and was asked for blankets and barbed wire instead. Now, Mitsubishi Heavy Industries is being asked to build missile components and cannot. At least not yet.
Christopher Gerteis (www.christophergerteis.net) is Senior Lecturer (Associate Professor) in Modern and Contemporary Japanese History at SOAS University of London. He writes Japanese Modernity (www.japanesemodernity.substack.com), a Substack on the history behind Japan's current affairs.
