15 Years After 3/11, 2 Years After Noto — The Structural Limits of Japan's ¥41 Trillion Recovery

On March 11, 2026, Japan marked the 15th anniversary of the Great East Japan Earthquake. According to the Reconstruction Agency, approximately 570 km of roads, 30,000 units of public housing, and seawalls have been largely completed—infrastructure recovery is officially "essentially done." The total recovery budget reached over ¥41 trillion[1]—a massive public expenditure spanning 14 years.

Yet the numbers tell a different story. In the coastal municipalities, as compiled by the Reconstruction Agency from census data, Iwate's twelve coastal municipalities held 274,086 people in 2010, before the disaster, and 227,103 in 2020[2] — 46,983 people, or 17.1%, gone in a decade. Miyagi's fourteen coastal municipalities excluding Sendai fell from 662,613 to 608,637, down 53,976 people, or 8.1%.

The workforce in disaster-affected areas is shrinking at 1.6 times the national average[3]. Maintenance costs for rebuilt infrastructure have ballooned to 1.8 times pre-disaster levels[3].

One caveat matters here. The Reconstruction Agency's own document notes that while the coastal decline is steeper than the national rate, "in the sense that population has entered a phase of decline, the trend is the same." The 2025 census bears this out: 1,558 of Japan's 1,719 municipalities (90.6%) had fewer people than five years earlier, with 62.4% down by 5% or more and 27.7% down by 10% or more[4]. That the disaster area is losing people is not, in today's Japan, unusual. What is unusual is the speed.

Infrastructure was rebuilt. But livelihoods did not return. This single sentence encapsulates 15 years of recovery.

Two years after the Noto Peninsula earthquake on January 1, 2024, the same structural pattern is emerging. Suzu City's population has fallen below 10,000, and the four Oku-Noto municipalities have lost 12% of their population[5]. Have the lessons of 3/11 been applied? This analysis examines the structural challenges both disasters share.

What was rebuilt

Roads

about 570 km

Disaster public housing

about 30,000 units

Sea walls

essentially complete

Reconstruction budget

over 41 trillion yen across 14 years

What did not return

Municipalities losing population

nine in ten of the 42 affected

Of which, losing 20% or more

about half

Rate of workforce decline

1.6 times the national average

Infrastructure maintenance costs

1.8 times pre-disaster

Two years after the Noto Peninsula earthquake, Suzu has fallen below 10,000 residents and the four municipalities of Oku-Noto are down 12%.

Where Did ¥41 Trillion Go?

Breaking down the ¥41 trillion recovery budget reveals a clear structure: approximately half was allocated to "hard" infrastructure projects. Seawalls and coastal protection received roughly ¥5 trillion, recovery roads (~570 km) about ¥3 trillion, public housing (~30,000 units) about ¥2 trillion, and land readjustment and highland relocation about ¥3 trillion.

Recovery Budget Structure (FY2011–2025 / 14 years)

¥41T+

Hard infrastructure (~50%)

• Seawalls & coastal protection~¥5T
• Recovery roads (~570km)~¥3T
• Public housing (~30K units)~¥2T
• Land readjustment & relocation~¥3T
• Other infrastructure~¥7T

Soft programs (~50%)

• Nuclear damage compensation~¥8T
• Industry & job rebuilding~¥4T
• Survivor support & health~¥3T
• Decontamination & waste~¥3T
• Other recovery programs~¥3T

Infrastructure was rebuilt, but livelihoods did not return

📉Population declined in 90% of 42 affected municipalities👤Workforce shrinking 1.6x faster than national average💰Maintenance costs 1.8x pre-disaster levels

The remaining half in "soft" programs included nuclear damage compensation (¥8 trillion), industry and employment rebuilding (¥4 trillion), survivor support and healthcare (¥3 trillion), and decontamination and waste processing (¥3 trillion). However, as Mitsubishi Research Institute points out, investment in "livelihood recovery"—community rebuilding and resident return support—was structurally insufficient.

This "hardware bias" has institutional roots. Recovery grant programs tend to favor "hard" projects whose completion can be clearly measured—roads, seawalls, housing units—while "soft" initiatives like social capital restoration and local economic revitalization, whose outcomes are harder to quantify, struggle to secure funding. The result: physical reconstruction is completed while social recovery is left behind.

The Depopulating Disaster Zone

Fukushima Prefecture's population fell from approximately 2.029 million in 2010 to 1.834 million in 2020—a decline of roughly 10%[6]. Iwate dropped by about 7%, and even Miyagi saw a 2% decrease. The aging rate in the Tohoku region reached 32%[7], well above the national average of 28.4%.

The number of evacuees has decreased from a peak of approximately 470,000[8] to about 25,000[9], but 15 years on evacuees are still spread across 791[9] municipalities. In surveys, over 30% of evacuees report stalled life rebuilding, and more than 50% describe economic hardship.

Academic research confirms this disconnect. A ScienceDirect (2024) paper finds that changes to physical characteristics alone do not restore a "sense of place." Social connections, daily routines, and identity are the true keys to return—infrastructure recovery is merely a precondition.

Noto's Question: Recovery in Depopulating Areas

The Noto Peninsula earthquake disaster zone carries a different context from 3/11. The four Oku-Noto municipalities held 48,600[10] people as of 1 August 2025, down 12%[10] from 1 January 2024 (the prefecture as a whole fell 1.5%[10] over the same period). Suzu City dropped below 10,000 for the first time, to 9,952[10], a 15%[10] fall from 11,721[10] at the time of the quake. NHK's mobile location data analysis estimates the actual resident population at approximately 7,900[11].

Structural Comparison of Two Disaster Recoveries

	Great East Japan (2011)	Noto Peninsula (2024)
Scale of damage	~19,000 deaths	~470 deaths
Affected population	3 prefectures, ~5.7M	4 municipalities, ~55K
Recovery budget	¥41T+ (14 years)	~¥660B (1 year)
Population decline	90% of 42 municipalities	Oku-Noto ▲12% (1.5 yrs)
Evacuees	470K → 26K	~13K still in temp housing
Structural issue	Hard infra bias, rising costs	No recovery model for depopulated areas

Despite different scales, both share a structural problem: 'rebuilding infrastructure doesn't bring people back.' Noto further poses an unresolved question: how to rebuild in depopulating areas.

The applicability of the 3/11 recovery model—massive seawalls and land elevation for "in-place reconstruction"—to a depopulating area of 50,000 people has been widely questioned. Located approximately 150 km from Kanazawa at the tip of the peninsula, Suzu City's geographic isolation made even supply delivery extremely difficult. Limited flat land complicated temporary housing site selection, and a shortage of construction workers able to enter the disaster zone delayed restoration work.

Approximately 13,000 people[12] remain in temporary housing, with over 80% requesting extensions. No timeline exists for permanent housing transitions, and the "prolonged temporary housing" pattern from 3/11 is repeating.

The Build Back Better Gap

In March 2015, at the 3rd UN World Conference on Disaster Risk Reduction held in Sendai, Japan led the adoption of the Sendai Framework, which positioned Build Back Better as one of four priority actions. Andrew Zolli systematically examines the essence of resilience, including the Build Back Better concept, in Resilience: Why Things Bounce Back（外部サイト、新しいタブで開きます）. The concept calls for treating disaster recovery not as mere restoration but as an opportunity to overcome societal vulnerabilities.

Yet Japan's own recovery diverges from this principle. The hardware-biased budget allocation contrasts sharply with Cambridge Core's quantitative analysis showing that investment in social infrastructure improves community survival rates.

New Zealand's response to the 2011 Christchurch earthquake offers a contrast: the Canterbury Earthquake Recovery Authority (CERA) was established as a five-year temporary organization and launched the "Share an Idea" initiative, collecting 106,000 ideas[13] from citizens. While criticized for top-down decision-making, the integration of citizen participation into the recovery process represents a significant difference from Japan's approach.

Can "Pre-Disaster Recovery" Work?

The Ministry of Land, Infrastructure, Transport and Tourism's 2023 guidelines for pre-disaster recovery planning aim to spread the concept of preparing recovery frameworks before disasters strike. The guidelines put the take-up rate at about 65%[14] as of the end of July 2022. But "take-up" here counts any municipality that answered "considered" or "under consideration" on any of five items — framework, procedure, drills, baseline data, or goals — and the guidelines themselves note that most have got no further than considering the framework and procedure.

The projected damage from a Nankai Trough megaquake far exceeds that of 3/11. The concept of pre-disaster recovery is gaining traction, but large-scale residential relocation and land-use regulations requiring community consensus remain unrealized. Noto's experience concretely illustrates what happens when a region that neglected pre-disaster recovery planning is struck by disaster.

The Disconnect Between Infrastructure and Social Recovery

The clearest lesson from 15 years of recovery is a structural fact: physical infrastructure restoration and the rebuilding of residents' lives do not move in tandem. Roads may be rebuilt, but without workplaces at their destination, people will not return. Public housing may be constructed, but without community, residents will not stay.

Ryo Yamazaki demonstrated in 『コミュニティデザイン: 人がつながるしくみをつくる』 (Community Design)（外部サイト、新しいタブで開きます） (『コミュニティデザイン』) that infrastructure restoration alone cannot regenerate community life. This disconnect stems from a fundamental problem in how recovery is evaluated. "Road completion rates" and "housing units built" are measurable, but "degree of social capital restoration" and "resident life satisfaction" are not easily quantified. As long as recovery is assessed through measurable indicators alone, hardware bias will be systematically reproduced.

Measurable

Share of roads reopened

countable

Housing units completed

countable

Hard to quantify

Recovery of local social capital

hard to measure

Residents' satisfaction with daily life

hard to measure

While recovery is judged on measurable indicators, the bias toward hard infrastructure reproduces itself institutionally. A reopened road does not bring people back without work at the end of it.

The Absence of a Depopulation Recovery Model

The most fundamental question Noto raises is whether full-scale recovery in depopulating areas is possible—or even desirable. Suzu City's population was already declining before the earthquake; even complete infrastructure restoration offers no prospect of population recovery.

This demands a paradigm shift in recovery policy: from "restoring what was" to "restructuring around contraction." Compact urban development, regional cooperation for maintaining essential services, and digital technology for remote service delivery—these are rarely discussed in recovery contexts but are unavoidable considerations for sustainability in depopulating areas.

Will the Disaster Prevention Agency Be a Turning Point?

The Disaster Prevention Agency, scheduled for establishment during fiscal 2026, is expected to integrate pre-disaster preparedness with recovery. Mitsubishi Research Institute has proposed "peacetime recovery policies," and a Nikkei editorial has argued for "turning the lessons of 15 years into the foundation for pre-disaster recovery."

But establishing an agency alone will not solve structural challenges. What is needed is institutional design that shifts recovery budgets from "hardware" to "livelihoods," planning standards that assume contraction in depopulating areas, and fiscal equalization mechanisms that compensate for disparities in municipal financial capacity. The question is not about creating an organization—it is about fundamentally transforming the recovery policy paradigm.

The lesson left by ¥41 trillion and 15 years is clear: rebuilding infrastructure does not automatically restore livelihoods. Whether this lesson can be applied to the next disaster is the essential question facing the new Disaster Prevention Agency.

---

For a practical framework on assessing disaster vulnerability at the community level, see "Disaster Vulnerability Mapping — A Practical Guide to Visualizing Hidden Community Risks（このサイトの記事）."

References

復興予算41兆円は何に使った 数字と図解で見る東日本大震災15年 — Nikkei (2026)

東日本大震災、ハード偏重の復興 全国比1.6倍速で減る働き手 — Nikkei (2026)

東日本大震災からの復興の状況と取組（2026年2月版） — Reconstruction Agency (2026)

東日本大震災から15年 防災庁は「平時からの復興政策」を — Mitsubishi Research Institute (2026)

Disrupted sense of place and infrastructure reconstruction after the Great East Japan Earthquake — ScienceDirect (2024)

How social infrastructure saves lives: A quantitative analysis of Japan's 3/11 disasters — Cambridge Core / Japanese Journal of Political Science (2023)

Guidelines for Pre-Disaster Recovery Town Planning — Ministry of Land, Infrastructure, Transport and Tourism (2023)

東日本大震災復興の総合的検証 — Hyogo Earthquake Memorial 21st Century Research Institute (2018)

Reference Books