Fukushima Daiichi: A Manmade Disaster: How Japan's Nuclear Establishment Designed Out the Warning Signs

Q: Was the Fukushima nuclear disaster preventable?

Yes, according to the National Diet's independent investigation (NAIIC), which called it 'a profoundly manmade disaster'. TEPCO's own 2008 simulations showed tsunami waves up to 15.7 metres against a 5.7-metre design basis; countermeasures such as raised backup power and seawalls were considered and postponed, and the regulator was briefed only four days before the tsunami struck.

Q: Why didn't TEPCO act on the 2008 tsunami warning?

Internal documents and subsequent litigation show executives deferred countermeasures, instead asking the industry-funded Japan Society of Civil Engineers to re-examine the calculation methodology. Prosecutors and a 2022 civil court characterised this as delaying safety spending; TEPCO maintained the estimates were not definitive. The episode is the textbook example of regulatory capture blunting internal warning signals.

Q: How much will the Fukushima cleanup cost?

The Japanese government's December 2016 estimate is ¥21.5 trillion (about US$187–190 billion) — roughly double its previous figure — covering compensation, decontamination, interim storage and decommissioning. Japan's Board of Audit found ¥12.1 trillion already spent by FY2021, and the private Japan Center for Economic Research has estimated ¥35–81 trillion depending on assumptions.

Q: What did the NAIIC report conclude?

The Diet-appointed commission concluded in July 2012 that the accident resulted from collusion between the government, regulators and TEPCO, and from a failure of safety culture and governance — not simply from an unforeseeable tsunami. It identified regulatory capture: the watchdog NISA sat inside METI, the ministry promoting nuclear power. NISA was abolished and replaced by the independent Nuclear Regulation Authority.

Q: Were TEPCO executives held criminally responsible?

No. Three former executives were acquitted in the only criminal trial (2019, upheld on appeal in 2023). In a separate civil shareholder suit, a Tokyo court in July 2022 ordered four former executives to pay ¥13.32 trillion — reported as Japan's largest civil award — finding the disaster preventable; that ruling has been subject to appeal.

Q: How is the Fukushima decommissioning progressing?

Slowly and behind early plans. Spent fuel has been removed from the Unit 3 and 4 pools, but molten fuel debris in Units 1-3 remains the central challenge: the first trial retrieval — a gram-scale sample from Unit 2 — occurred only in November 2024. Full decommissioning is officially framed as a 30-40 year programme, with treated water discharges continuing alongside.

Ramesh Dixit

Investigation PIA-INV-017Billion Dollar FailuresCost: ¥21.5tn (US$187–190bn) official 2016 estimate; JCER private estimate ¥35–81tn; ¥12.1tn spent by FY2021 (Board of Audit)8 min read

Fukushima Daiichi

A Manmade Disaster: How Japan's Nuclear Establishment Designed Out the Warning Signs

Filed under: Asia Infrastructure · Engineering Failures · Project Risk Management · Tokyo Electric Power Company (TEPCO) · Japan · Energy & Utilities

Written and edited by Ramesh Dixit·Published 2026-08-17·Last updated 2026-08-17·Last fact-checked 17 August 2026·Editorial Standards · Editorial Policy · Corrections Policy · Methodology · Source Standards · AI Disclosure

Download Executive Brief (PDF)

The National Diet of Japan's independent investigation called Fukushima Daiichi 'a profoundly manmade disaster' — the product of regulatory capture, engineering warnings that were documented but not acted upon, and a utility that shelved its own tsunami analysis. In 2008 TEPCO's engineers calculated waves of up to 15.7 metres could strike a plant protected to 5.7 metres; countermeasures were postponed, and regulators were briefed only days before the March 2011 tsunami. The cleanup bill has since doubled and redoubled: a ¥11 trillion official estimate became ¥21.5 trillion in 2016, with private studies reaching ¥81 trillion. This investigation focuses on the governance and project failures — not the tsunami — that made the accident preventable.

Fukushima Daiichi investigation — the damaged reactor buildings behind the seawall

Key Facts

Cost

¥21.5tn (US$187–190bn) official 2016 estimate; JCER private estimate ¥35–81tn; ¥12.1tn spent by FY2021 (Board of Audit)

Date

2026-08-17

What Happened

Fukushima Daiichi's vulnerability was engineered in, then documented, then not acted upon. The plant was designed in the 1960s to a tsunami design basis of 3.1 metres, later raised to 5.7 metres after 2002 assessments; critically, emergency diesel generators and seawater pumps sat in basements and low-lying yards. Japan's own Headquarters for Earthquake Research Promotion warned in 2002 that the Japan Trench off Fukushima could produce giant tsunamis, citing the Jogan event of 869 AD. The utility's regulator, NISA, sat inside METI — the ministry charged with promoting nuclear power — and, as the National Diet's independent commission (NAIIC) later found, the relationship had curdled into regulatory capture: oversight bodies staffed and steered by the industry they policed. The decisive failure was a project decision. In 2008, TEPCO's own simulations, run against the government's long-term assessment, produced an upper-bound case of up to 15.7 metres — nearly three times the design basis and sufficient to flood the site. Executives considered seawalls and raised backup power, then postponed the work, commissioning the industry-funded Japan Society of Civil Engineers to re-examine the methodology instead. TEPCO did not formally brief NISA on the 15.7-metre figure until 7 March 2011, four days before the tsunami. On 11 March 2011, waves of roughly 14–15 metres overwhelmed the site, drowned the backup generators and produced station blackout; three reactors melted down and hydrogen explosions followed. The response phase repeated the pattern: improvised venting, confused chains of command between TEPCO, NISA and the Prime Minister's office, and evacuations that displaced some 150,000–160,000 people at peak. Fukushima Prefecture has since attributed more than 2,000 deaths to the stresses of evacuation and displacement — more than the disaster's direct radiation toll, for which a single occupational cancer death has been officially recognised. The financial reckoning is a case study in estimate escalation. Early official response costs were put around ¥11 trillion; in December 2016 the government nearly doubled the figure to ¥21.5 trillion (about US$187–190 billion), comprising roughly ¥7.9tn compensation, ¥8tn decommissioning and contaminated water, and ¥5.6tn decontamination and interim storage. Japan's Board of Audit found ¥12.1 trillion already spent by fiscal 2021 and asked the government to re-examine the total. The private Japan Center for Economic Research has estimated ¥35–81 trillion depending on assumptions. Meanwhile decommissioning milestones have slipped repeatedly: full fuel-debris removal, once targeted within a decade, achieved only its first trial retrieval — a sample of grams — in November 2024. In July 2022 a Tokyo court ordered four former TEPCO executives to pay ¥13.32 trillion in a shareholder suit, finding the disaster preventable; the ruling has since moved through appeals, and criminal trials ended in acquittal.

Why It Matters

Fukushima is a stark demonstration that a low-probability, high-consequence risk, once identified, must be dispositioned by engineering action — not deferred by committee. Every element of the failure is a governance artefact rather than a natural one: a design basis frozen in the 1960s, a regulator captured by its ministry, an operator whose internal analysis of its own plant's fatal vulnerability was softened on request and then shelved to protect asset values and restart schedules. The cost trajectory — ¥11tn to ¥21.5tn officially, up to ¥81tn privately — shows how suppressed front-end safety spend converts into unbounded back-end liability borne by the public. For any operator of hazardous infrastructure, the case defines the test of safety culture: whether inconvenient analysis produced inside the organisation can change the plan.

Timeline

design 1967-71
Fukushima Daiichi designed and Unit 1 commissioned with a 3.1m tsunami design basis; emergency generators placed in low-lying buildings.
warning 2002-07
Government's Headquarters for Earthquake Research Promotion warns a giant Jogan-type (869 AD) tsunami is possible off Fukushima; design basis later raised only to 5.7m.
warning 2008-03
TEPCO simulations based on the 2002 assessment calculate tsunami heights up to 15.7m — enough to flood the plant.
failure 2008-06/07
The 15.7m result is briefed to TEPCO vice-president Muto; countermeasures are postponed and the industry-funded JSCE is asked to re-examine the methodology.
warning 2011-03-07
TEPCO formally briefs regulator NISA on the 2008 tsunami calculations — four days before the disaster.
failure 2011-03-11
M9.0 earthquake and ~14-15m tsunami overwhelm the 5.7m design basis; station blackout and loss of cooling follow at Units 1-3.
failure 2011-03-12/15
Core meltdowns in Units 1-3 and hydrogen explosions; mass evacuation displaces around 150,000-160,000 people at peak.
inquiry 2011-04-12
The accident is rated Level 7 on the INES scale, the maximum category.
inquiry 2012-07-05
The National Diet's NAIIC report concludes the disaster was 'profoundly manmade', citing regulatory capture and collusion between government, regulators and TEPCO.
policy 2012-09
NISA is abolished and replaced by the structurally independent Nuclear Regulation Authority.
delay 2016-12
The government nearly doubles its estimated total response cost to ¥21.5tn, from ¥11tn.
settlement 2022-07-13
Tokyo District Court orders four former TEPCO executives to pay ¥13.32tn to the company in a shareholder suit, finding the disaster preventable; appeals follow.
inquiry 2022-11
Japan's Board of Audit reports ¥12.1tn already spent and calls on the government to re-examine the ¥21.5tn projection.
milestone 2024-11
First trial retrieval of a gram-scale fuel-debris sample from Unit 2 — years behind early decommissioning plans.

Root Cause Analysis

Root cause through the Leadership Blind Spot Matrix™ lens

Pattern Overconfidence

Over-reliance on past experience in novel situations

Dissent Filtering

Systematic exclusion of challenging perspectives

Risk Discounting

Underestimating emerging threats due to present-focus

Attribution Errors

Misattributing success to skill rather than context

Frameworks Applied

Framework Applied

Leadership Blind Spot Matrix™

Identifying What Leaders Cannot See — explore the framework →

Framework Applied

Project Failure Pyramid™

The 4-Level Diagnostic for Cascading Failures — explore the framework →

Framework Applied

Decision Quality Model™

How Good Decisions Produce Good Outcomes — explore the framework →

Executive Lessons

Lessons for Leaders

Regulatory Capture Is a Design Flaw, Not a Scandal

NISA sat inside METI, the ministry promoting nuclear power, and rotated staff with the industry. NAIIC named the resulting capture as a root cause. Safety regulators of hazardous assets need structural independence — budget, appointments and statutory powers — not just good intentions.

Internal Warnings Must Have a Forcing Function

TEPCO's 2008 tsunami analysis reached its executives and died there. Safety-significant internal analysis should carry a mandatory disposition process: regulator notification, independent review, and a documented accept-or-mitigate decision at board level, within a fixed period.

Postponement Is a Decision With a Price

Countermeasures were considered in 2008 and deferred pending further study; a Tokyo court in 2022 priced that deferral at ¥13.32 trillion in executive liability. 'More analysis' is not a neutral option when the hazard is time-invariant and the exposure is catastrophic.

Design Basis Must Track the Science

The plant's 5.7-metre tsunami basis survived unchanged through a decade of new palaeotsunami science (the 869 Jogan event evidence, the 2002 government long-term assessment). Design bases for long-lived assets need scheduled re-validation against current hazard science, with upgrade triggers.

Single-Point Failures Cluster in Basements

Emergency power and seawater pumps — the last line of defence — were placed in low-lying areas vulnerable to the exact hazard under analysis. Defence-in-depth reviews must walk the physical layout against the hazard scenario, not the org chart.

Evacuation Harm Is Project Harm

More than 2,000 disaster-related deaths in Fukushima Prefecture are attributed to evacuation and displacement stress, against a single officially recognised radiation death. Emergency planning is part of the project's safety case, and its human costs belong in the risk model.

Estimate the Tail Before You Build the Head

Official cleanup cost estimates nearly doubled in five years (¥11tn to ¥21.5tn) and private estimates run to ¥81tn, with fuel-debris removal decades behind early plans. Whole-life liability pricing, including accident externality, should be part of any hazardous-asset investment case.

Continue Investigating

Burj Khalifa rising 828 metres above Downtown Dubai at dusk, the world's tallest building, delivered by Emaar, SOM and Samsung C&T between 2004 and 2010

BuildingsWidely reported at approx. US$1.5 billion (tower only); never officially disclosed

Burj Khalifa: The Supertall Megaproject That Finished on Time in a Financial Storm

9 min read

HS2 high-speed rail construction works in the English Midlands, with tunnelling and viaduct structures under an overcast sky

Megaprojects£87.7-102.7B (2026 estimate, London-Birmingham only)

HS2: How Cost, Scope and Schedule Changed Britain's High-Speed Rail Programme

9 min read

The Weekly Brief — one investigation, one executive lesson and one project warning sign each week.

Subscribe free

Executive Recommendations

Guarantee structural independence of safety regulators — Separate nuclear and hazardous-infrastructure regulators from promoting ministries in law, budget and appointment; prohibit revolving-door staffing for defined cooling-off periods.

Mandate disposition deadlines for internal safety analyses — Any operator analysis indicating exceedance of design basis must be notified to the regulator and dispositioned (mitigate or formally accept risk at board level) within a fixed statutory window.

Re-validate design bases on a fixed cycle against current science — Require periodic hazard re-assessment for long-lived hazardous assets, with automatic upgrade triggers when the assessed hazard exceeds the design basis by defined margins.

Price whole-life accident liability into project economics — Require investment cases for hazardous assets to include externally reviewed accident-externality costings and mandatory insurance or liability pools sized to credible worst cases.

Independent post-facto audit of emergency preparedness — Commission periodic red-team exercises and public reporting on evacuation logistics, power-supply redundancy and flood pathways, conducted by bodies independent of operator and promoter.

PMOS Intelligence

Preview — illustrative assessment; PMOS is in development

Governance WeaknessNAIIC concluded the accident was 'profoundly manmade' — the product of collusion between government, regulators and TEPCO: NISA sat inside the industry-promoting ministry METI, with personnel rotating between regulator and utilities.

Escalation FailureTEPCO's 2008 simulation showing tsunami heights up to 15.7m — against a 5.7m design basis — was not acted on; countermeasures were postponed in 2008-09 and NISA was briefed only on 7 March 2011, four days before the event.

Decision DelayA 1960s design basis was never re-validated against modern tsunami science across four decades of operation; each review cycle deferred re-assessment rather than pricing the upgrade.

Leadership Blind SpotLeadership treated the tsunami hazard as a public-relations risk to be managed rather than an engineering fact to be engineered against — protecting the plant's economics and the nuclear programme's image over its safety case.

Risk VisibilityEmergency diesel generators and seawater pumps were sited in flood-exposed locations, creating a single-point total-loss-of-power scenario that was visible in the plant layout and documented in internal analysis, yet never escalated to board-level risk acceptance.

Evidence QualityThe suppressed 2008 simulation was high-quality internal evidence; it was shelved precisely because acting on it implied costly backfits and admission that the design basis was wrong — evidence quality was not the constraint, evidence use was.

Assurance MaturityAssurance was captured: the regulator shared the ministry's promotion mandate, stress tests were voluntary, and no independent body could compel backfits against operator resistance.

Suggested InterventionAn independent safety assurance function — statutorily separate from promoting ministries, with disposition deadlines for any analysis indicating design-basis exceedance — would have forced the seawall and waterproofing backfits between 2008 and 2011.

“TEPCO's executives did not lack information; they lacked a governance structure in which information could win. The 15.7-metre figure was calculated in-house, briefed upwards, and then subjected to the one review process guaranteed to bury it — a request to an industry-funded learned society to reconsider the methodology. That is the signature move of captured safety systems: not suppressing the data, but endlessly deferring its implications. The tsunami was the trigger; the disaster was designed in Tokyo boardrooms and ministry corridors over the preceding decade, exactly as the Diet's commission concluded.”Ramesh's Insider Take — opinion

Evidence

Documentary Evidence

official-investigation-report

The official report of the National Diet of Japan Fukushima Nuclear Accident Independent Investigation Commission (NAIIC), Executive Summary, July 2012

The official report of the National Diet of Japan Fukushima Nuclear Accident Independent Investigation Commission (NAIIC) · 2012-07

The accident was 'profoundly manmade': the result of collusion between government, regulators and TEPCO, and of regulatory capture — and it was preventable.

View document →

press-report

Associated Press reporting on the TEPCO executives' criminal trial, 2019

Associated Press reporting on the TEPCO executives' criminal trial · 2019

In March 2008 a TEPCO subsidiary projected a tsunami up to 15.7m could strike Fukushima; executives delayed seawall countermeasures to avoid additional spending, and NISA was briefed only in March 2011.

View document →

press-report

3.11 Mieruka Project fact sheet on accident costs, citing METI 2016 and JCER

3.11 Mieruka Project fact sheet on accident costs · 2016

The government's December 2016 estimate of total accident costs is ¥21.5tn (compensation ¥7.9tn, decommissioning/contaminated water ¥8tn, decontamination and interim storage ¥5.6tn), double the prior ¥11tn; the Japan Center for Economic Research estimates ¥35–81tn.

View document →

official-audit-report

Asahi Shimbun reporting on the Board of Audit of Japan, republished by the University of Chicago Atomic Age blog, 2022

Asahi Shimbun reporting on the Board of Audit of Japan · 2022

¥12.1 trillion had been spent on the disaster response by fiscal 2021, and the Board of Audit asked the government to review its ¥21.5tn projection.

View document →

press-report

Nippon.com analysis of Fukushima-related litigation and foreseeability

TEPCO ran its 15.7m simulation in 2008, took no mitigation action, and did not report the findings to NISA until early 2011, weeks before the disaster.

View document →

Asset Hub