JR: past, present, and future

The last post was a recommendation: if you privatize, be like JR. This one is not. It is the record — past, present, future — so the recommendation can be checked against what actually happened.

The short version, before the numbers: the 1987 reform produced three listed, self-standing passenger companies on Honshu, a fourth listing on Kyushu 29 years later, and three companies that are still not listed and still on the public books. Both halves are the JR story.

Past: what JNR was, and what 1 April 1987 actually did

Japanese National Railways was a public corporation. By the mid-1980s it was also a fiscal event. The International High-Speed Rail Association’s reform brief puts JNR employment at more than 460,000 in 1965 and 277,000 in 1986, the year before the split. That is already a shrinking railway, not a sudden overnight cut.

On 1 April 1987 JNR was replaced by six regional passenger companies and one nationwide freight company. The passenger firms kept operations and infrastructure together (vertical integration) and were cut by region (horizontal split). Freight was separated from passenger and given almost no track of its own.

The opening balance sheet is the part most slide decks skip. JR East’s 2004 annual report puts JNR’s long-term liabilities at restructuring — including future costs — at ¥37.1 trillion. The IHRA brief and the same JR East note give the same split, in round figures:

Who took the liability	About
JR East, JR Central, JR West, JR Freight (and related entities)	¥5.9 trillion
Shinkansen Holding Corporation	¥5.7 trillion
JNR Settlement Corporation	¥25.5 trillion
JR Hokkaido, JR Shikoku, JR Kyushu	none

The island passenger companies were instead given a Management Stabilization Fund whose interest was meant to cover operating losses: about ¥682 billion for Hokkaido, ¥208 billion for Shikoku, ¥388 billion for Kyushu (about ¥1.28 trillion in total). That is not a footnote. It is how the reform treated density: Honshu companies inherited some debt and a market; the islands inherited a fund and no debt.

Workforce: of the 277,000 still at JNR in 1986, about 201,000 went to the JRs. The Settlement Corporation took on the order of 23,000 people who had not been placed; World Bank accounts of the reform put final dismissals around 1,000, for people who refused offered jobs. The cut was real. It was not a mass firing on day one.

Two design choices sit under those numbers and should not be blurred:

Debt isolation. Most of the hole did not go onto the new operators. A settlement body was supposed to pay it down from land sales and later share sales.
Vertical keep, horizontal cut. Track, trains, and stations stayed in the same company. The national monolith did not.

Present: four listed companies, three that are not, and a debt that did not vanish

Traffic and productivity in the first decade

Hiroyuki Fukui, writing in the Japan Railway & Transport Review, reports that total passenger traffic on the new JR operators rose 24% between 1986 (the last JNR year) and 1991, passing JNR’s 1974 peak, and that JR Freight volume rose 30% over the same stretch. He also says the late-1980s bubble did a lot of that lifting: volumes and revenue rose without the fare increase that had been planned. That caveat belongs in the same paragraph as the success.

On productivity, the numbers that get quoted are from Fumitoshi Mizutani and co-authors, using Ministry of Transport and JNR/JR statistics:

Mizutani and Kiyoshi Nakamura, Papers in Regional Science (1996): the estimated effect of privatization on labour productivity growth was about 29%. They also found JRs still had about 20% more employees than Japan’s large private railways, with the gap concentrated in stations and maintenance, and that employment reduction in the transition did most of the work. Serious accidents were not associated with the productivity gain.
Mizutani and Shuji Uranishi, International Journal of Transport Economics (2007): after privatization, annual TFP growth was 2.97%, of which privatization itself accounted for 1.62 percentage points. Over-capitalisation at JNR was largely corrected.
Mizutani and Nakamura (1997), in the same journal: in the first ten years, financial results, service quality, labour productivity, and operating costs improved; accident rates did not rise; fares, which had been rising almost yearly before 1987, did not increase for nine years after.

Those are not “JR is magic.” They are: the Honshu operators, in a dense market, with debt parked elsewhere, got more output per worker and did not buy it with a worse safety record in the period studied.

Who actually left the state

“Privatized in 1987” is the wrong sentence. The JRs started as special companies whose shares sat with the Settlement Corporation. Listing is the later event. The Japan Railway Construction, Transport and Technology Agency (JRTT), which inherited the settlement function, records the share sales:

Company	First sale	Fully sold	Listed
JR East	October 1993	June 2002	yes
JR West	October 1996	March 2004	yes
JR Central	October 1997	April 2006	yes
JR Kyushu	October 2016 (one tranche)	October 2016	yes
JR Hokkaido	—	still with JRTT	no
JR Shikoku	—	still with JRTT	no
JR Freight	—	still with JRTT	no

JRTT puts total proceeds from the four sold companies at ¥4.4503 trillion. JR Kyushu listed 29 years after the split, after building property and other non-rail businesses — the same station-area model the Honshu companies used, and the same model Railway Gazette noted ahead of the 2016 IPO.

JR Hokkaido, JR Shikoku, and JR Freight were never in that column. That is not a secret. It is the present tense of the reform.

What the unlisted companies look like now

For the year ending March 2025, reporting compiled from the companies’ results put JR Kyushu at about ¥58.9 billion operating profit, JR Shikoku at about ¥13.0 billion operating loss, and JR Hokkaido at about ¥48.2 billion operating loss. JR Hokkaido’s transport business remains hundreds of billions of yen in the red; Asahi (June 2025) reported group net profit only because of national support (about ¥21.7 billion that year) plus fund income, while hotels — including reuse of railway housing land — were a bright spot (about ¥11.3 billion in hotel sales).

The state is still in the room. MLIT’s support packages for “the two islands and freight” (JR Hokkaido, JR Shikoku, JR Freight) extend grants toward FY2030, with a stated aim of management self-reliance around FY2031. For Hokkaido, the FY2024–FY2026 slice is about ¥109.2 billion. The original stabilization funds were not a one-time goodbye.

The settlement debt

Land and share sales were supposed to shrink the Settlement Corporation’s pile. They did not keep up with interest. JR East’s 2004 report already describes liabilities at the settlement body rising after 1987 because land was hard to sell and listings slipped. Government-guaranteed JNR-related debt was still about ¥15.6 trillion at the end of FY2021 — lower than ¥37 trillion, not gone. The operators were isolated from most of that hole. The public was not.

That fact does not cancel the operating success of JR East, Central, and West. It means “JR paid off JNR” is false. The reform moved the corpse.

Future: two Japans, not one JR

The forward-looking evidence is already visible. It does not look like 1987 copied onto the whole map.

The listed companies will keep doing what the large private railways did before them: run dense corridors, sell the station, and treat non-rail as a real P&L line. JR Kyushu’s listing is the existence proof that this can work off Honshu if the city-region is thick enough and the property is allowed to work.

The unlisted companies are on a different path. JR Freight, as of April 2024, is a Type-1 railway on only 29.1 km; almost all of its trains run on passenger companies’ tracks, with a freight adjustment-charge system still written through FY2030. JR Hokkaido is shrinking the map, putting hotels on former railway housing sites, and remaining on MLIT support through the decade. Thin passenger lines that left JR after Shinkansen parallel-conventional splits are already experimenting with deemed vertical separation — local governments taking infrastructure risk while an operator runs trains. Hisatsu Orange Railway, a third-sector line in Kyushu, saw ridership fall from about 1.88 million in its opening year (FY2004) to about 1.03 million in FY2024. That is the demography the island JRs are running into.

So the future inside Japan is a fork, not a brand:

Vertically integrated, listed passenger companies on dense corridors.
Publicly supported, sometimes quasi-separated, operations on thin lines, with freight still a special-track user.

Anyone who treats “JR” as one object in 2026 is describing 1991. The evidence since then is that the reform completed for four companies and stalled, with ongoing subsidy, for three.

What the record licenses (and what it does not)

Licensed by the numbers:

Isolating legacy debt from the operating companies was a condition of the Honshu success, not an accounting trick on the side.
Vertical integration plus regional companies plus station business is the package that listed. Unbundled track-plus-franchise is not in this dataset.
Labour productivity and TFP moved. Safety in the Mizutani papers did not move the wrong way with them.
Density is not a residual. The same legal form produced JR East and JR Hokkaido.

Not licensed:

That share sales retired the JNR hole. They did not.
That 1987 finished privatization. Four of seven group railway companies are listed; three are not.
That the late-1980s traffic boom was privatization alone. Fukui’s own account puts the bubble in the room.
That the JR template extends unchanged onto a sparse network. Japan’s own thin-line future is already drifting toward public support and, in places, vertical-ish splits — the thing the 1987 reformers declined to do on the main island.

I am still willing to say be like JR as a policy sentence. The evidence says: be like JR East in 2002, with your eyes open about JR Hokkaido in 2025, and do not call the settlement debt someone else’s problem. That is the past, the present, and the part of the future that is already on the timetable.

Sources

East Japan Railway Company, Annual Report 2004 (JNR liability split).
IHRA, “JNR Reform” (employment, debt allocation).
JRTT, Disposal of JR shares (listing and sale proceeds).
Hiroyuki Fukui, “JNR Privatization”, Japan Railway & Transport Review.
Fumitoshi Mizutani and Kiyoshi Nakamura, “Effects of Japan National Railways’ Privatization on Labor Productivity”, Papers in Regional Science 75(2), 1996.
Fumitoshi Mizutani and Shuji Uranishi, “The Effects of Privatization on Productivity and Capital Adjustment,” International Journal of Transport Economics 34(2), 2007.
MLIT, support materials for JR Hokkaido, JR Shikoku, and JR Freight (FY2024– packages; self-reliance target around FY2031).