The Amagasaki Disaster

The Amagasaki Derailment

In our workshops and training, we often use real-world stories as a way to explore the dynamics of both failure and success. Stories are a powerful tool to help us reflect on our own experiences, and sometimes challenge our assumptions too.

One story we return to often is the Amagasaki Disaster. It features prominently in our Human and Organisational Performance (HOP) workshops because it offers a rich case study of systemic failure. Through it, we explore key concepts such as production pressure, the normalisation of risk, the distinction between the Sharp End and the Blunt End, the role of blame, Local Rationality, and much more.

Let’s explore what happened.

The Derailment

On the morning of the 25th April 2005, during rush hour on the Fukuchiyama Line in Amagasaki, near Osaka, Japan, a tragedy unfolded. Around 700 passengers were aboard a seven-car JR West “Rapid” train, most of them on their way to work. At 09:19, the train reached a sharp curve between Tsukaguchi and Amagasaki – a 300m radius bend with a speed limit of 70 km/h. The train, however, was going much faster – around 116 km/h. 

The consequences were devastating. Hitting the bend at this speed derailed the front four cars of the train. The first two cars were completely crushed when the third and fourth landed on top of them, slamming into a nearby apartment building. 107 people died, including the driver, 23 year-old Ryūjirō Takami (高見 隆二郎, Takami Ryūjirō), and a further 562 were injured. It was Japan’s most serious rail disaster since the Tsurumi accident in 1963.

In the lead-up to the crash, a series of small incidents contributed to growing pressure. About 25 minutes earlier, the driver, Takami had accidentally passed a red signal, triggering the automatic train stop (ATS) system. He was forced to reverse the train back to the platform, costing valuable time. Then, just four minutes before the derailment, he overshot Itami Station by more than three carriages, once again having to reverse to align with the platform. Together, these two errors caused a 90-second delay.

By the time the train reached Tsukaguchi – the station just before Amagasaki – Takami had sped up, and managed to reduce the delay to just 60 seconds. But, as we now know, that effort to recover time came at a terrible cost.

It’s also worth noting that Takami did use the service brake to try to slow the train before the bend. But at this speed, and with this short a distance in which to reduce speed, the service brake wasn’t effective enough. He didn’t use the emergency brake, which it was later suggested may have slowed the train enough to avoid disaster.

Takami’s error and unsafe behaviour – his willingness to exceed safe speed limits and refusal to use the emergency brake – undoubtedly led to the disaster. But should we stop investigations at human error?

What Was Really Going On?

At first glance, the Amagasaki derailment might seem like a tragic case of individual failure – a young driver making reckless decisions that led to disaster. But as we often ask in our Human and Organisational Performance workshops: What sense did it make to the person at the time? What was going on in the system around them?

To understand the disaster, we need to look far beyond the moment of derailment.

Production Pressure

JR West, like other Japanese rail companies, upheld very high punctuality standards, where trains were expected to run perfectly on time. Many commuters rely on trains arriving and departing with clockwork precision to make their connections, because trains share the same platform so that people can switch directly between trains. This allows seamless transfers but also makes for a very tightly coupled timetable – even a brief delay can ripple through the day’s schedule. 

Indeed, schedule adjustments over the preceding three years had reduced the buffer time between Takarazuka and Amagasaki – from 71 seconds down to just 28 seconds, over a 15-minute run – which leaves very little margin for error. And Takami knew that his earlier errors left him under immense pressure to make up the time and reduce further delays.

Punishment Over Learning: “Nikkin Kyoiku”

Takami’s earlier mistakes weren’t seen as operational slip-ups – they carried very real personal consequences. Ten months prior to the disaster, after overshooting a platform by around 100 metres, Takami had been sent to JR West’s “retraining” programme, named “Nikkin Kyoiku” or “Dayshift Education”. 

Rather than providing support or training, Nikkin Kyoiku was punitive. It came with a loss of pay, and involved tasks which were intended to be humiliating, such as cutting grass, cleaning the toilets, sitting in the middle of a shared shift room whilst copying by hand the company rule book, or standing on a platform for days whilst greeting other train drivers. And Nikkin Kyoiku was widely feared. After Masaki Hattori made an unscheduled 60-second stop in 2001, to conduct routine safety checks, he was made to sit in a tiny office whilst being yelled at continuously. Hattori was found hanged in his home the following day. 

In total through his career with JR West, Takami had already endured 13 days of Nikkin Kyoiku. And understanding this helps us answer the question: given his speed, why didn’t Takami apply the emergency brake? Well, use of the emergency brake would have been automatically recorded. It would have triggered a formal investigation, and had Takami not been able to adequately justify his actions, he would have been blamed and punished for the mistake.

Takami’s desire to avoid yet more “retraining” was recorded as a major contributing factor in the final investigation – context drives behaviour.

Prioritising Profit over Safety

The control systems in place at the time (ATS-S) merely alerted the driver, without intervening. A newer system, ATS-P, could have intervened in an overspeed incident and automatically braked the train – but its rollout was delayed by JR West, citing cost and operational complexity.

This decision to delay ATS-P rollout to save on costs and avoid operational headaches left drivers without the safeguard that might have prevented this disaster. That trade-off – cost and convenience over risk mitigation – speaks volumes about what the organisation truly valued – and it wasn’t safety.

And it’s worth noting that the driver’s actions are far from the only contributing factors in the severity of this incident. There are also physical and engineering aspects; train lines in Japan typically run on narrower gauge than the rest of the world (1067mm instead of 1435mm), and train lines are often raised above the ground, very close to surrounding buildings. This is largely due to a lack of space as well as high trust in railway safety. What this means in reality is that the consequences for a rail disaster are often higher in Japan than elsewhere. 

Normalisation of Risk

Under the weight of relentless punctuality targets and a culture of blame, unsafe practices gradually became normalised. Speeding to make up time, even when this risked passenger safety, was common – even expected. In the months before the Amagasaki derailment, the same bend had seen at least eleven recorded “overspeed” incidents.

This normalisation of risk, or normalisation of deviance, may have led Takami to believe what he was doing was acceptable. It’s analogous to other disasters, such as the Costa Concordia, where over time, unsafe practices became the acceptable norm.

The Silencing of Concerns

The incentive structure at JR West was clear: drivers who hit targets got promotions and pay rises. Safer drivers who flagged risks or refused to make up time by exceeding speed limits were seen as troublemakers, and may even receive formal reprimands or be made to undertake Nikkin Kyoiku. Mistakes were seen as personal failing, rather than weak signals of system failures. 

So unsurprisingly, concerns about unsafe behaviours or practices were never raised. Unlike with CRM in commercial aviation or other safety-critical domains, there were no anonymous or psychologically safe channels in which to raise them. Near misses, safety issues, or the consequences of excessive system pressures went unspoken, and the silence wasn’t passive – it was enforced.

Power Gradients

There was a strong adherence in JR West to formal power structures: deference to those above you in the hierarchy is typically both part of Japanese culture more widely, and was also enforced through career consequences. Rather than working to decrease power and authority gradients, JR West managers amplified them – punishing and demeaning those “below” them for mistakes, deviations and missed targets. 

This also resulted in a huge gap, really more of a chasm, between those at the sharp end (the people doing the work) and the blunt end (those setting the policies). Senior managers and those making the rules and targets rarely even rode the line, so they had little awareness of what it was really like for front-line operators. Their decisions – to cut buffer time, delay safety upgrades and enforce punishment – were made far from the realities of daily operations.

Local Rationality

What Takami did was self-evidently wrong. But it made sense to him at the time. He was under pressure. He had already made mistakes. And he was trading the certainty of humiliating punishment and penalties against the risk of an accident. 

Takami was still a relatively inexperienced driver, who, it seems, was struggling already, having already overshot two platforms. It’s likely his performance was already impaired by stress or fatigue. And just prior to entering that bend, he’d received a non-essential phone call from the control station, which may have distracted his attention at a critical point. 

In the final seconds, the report notes there was no voiced “emergency-brake” confirmation from Takami; a procedural omission that the final report flagged as an indication of cognitive overload – stress, fatigue or task-saturation. Why did it make sense for Takami to behave the way he did?

Systems shape behaviour

It’s too easy to read stories like these personal failures – as one driver’s failure to do it right. But that’s not what happened here. What we see are the clear indicators of the failure of a system:

• Punctuality was prioritised over safety.
• Unsafe behaviours were rewarded.
• Concerns were suppressed. 
• Mistakes were punished.
• Power structures and blame suppressed learning.

Behaviours are not the problem – behaviours are symptoms of the problem. At JR West there were multiple opportunities to see failures as not just opportunities for individual learning, but as windows into a failing system. Instead, what happened at Amagasaki reflects a classic case of Self Organised Criticality – seemingly small failures and concerns accumulated over time, piling up like grains of sand, the slope getting steeper and steeper, until a single grain tipped the whole system into collapse.  In hindsight it looks inevitable. At the time it was invisible.

Aftermath and Change

In the aftermath, JR West was forced to learn and to act. It was necessary to consider not “Who was to blame?” but “Where and why did our systems enable, reward or require the behaviours that led to this disaster?”

They fast-tracked ATS-P installation across the Fukuchiyama Line. The Nikkin Kyōiku programme was overhauled to remove its public-shaming rituals. And safety-focused think-and-act training grounded in human-factors principles was rolled out across the company.

It shouldn’t take tragedy to trigger change. But the lessons of Amagasaki remain vital today for any organisation navigating risk, performance, and learning. 

Sources and further reading:

• Amagasaki Railway accident investigation report (Excerpt): https://jtsb.mlit.go.jp/eng-rail_report/RA2007-3-1en.pdf 
• JR train driver faulted in final report on crash: https://web.archive.org/web/20160609172252/https://www.japantimes.co.jp/news/2007/06/29/national/jr-train-driver-faulted-in-final-report-on-crash
• Drivers win toilet battle with Japan train firm: https://web.archive.org/web/20180618230444/https://www.fm-world.co.uk/news/fm-industry-news/drivers-win-toilet-battle-with-japan-train-firm/ 
• JR West says non-punishment policy for errors improves safety: https://www.asahi.com/ajw/articles/15726247
• Stressful Schedule: The 2005 Amagasaki (Japan) Derailment: https://mx-schroeder.medium.com/stressful-schedule-the-2005-amagasaki-japan-derailment-12156ddd488
• Union chief blames ‘bullying’ rail firm for Japanese train crash: https://www.theguardian.com/world/2005/apr/29/japan.justinmccurry
• Reddit thread on The 2005 Amagasaki Derailment:  https://www.reddit.com/r/CatastrophicFailure/comments/swwvhc/the_2005_amagasaki_japan_derailment_a_train/
• Japan ex-rail chiefs cleared of fatal 2005 crash: https://www.bbc.co.uk/news/world-asia-24296363

“Human and Organisational Performance” and “Delivering Effective Feedback” Workshops

Booking is now open for “Delivering Effective Feedback” and HOP Fundamentals
The success of our inaugural HOP Fundamentals workshop meant that we had to schedule another immediately, and we’ve also opened the next Feedback workshop which is already half full before even advertising it.

The Delivering Effective Feedback workshop is designed to be applicable for those working in all industries, domains and roles. It may be most useful for those with people responsibilities, including managers, leaders, trainers, supervisors and mentors. 

In the HOP Fundamentals workshop, we use the 5 principles as a scaffold to explore HOP and how to put it into practice. We cover many safety-related topics, such as human error and Work-As-Imagined vs Work-As-Done, but it is not specifically a health and safety workshop.

Join us and build understanding as well as great practices to take away and run with your own teams or clients, and meet wonderful participants from all over the world!

“Genuinely one of the most thought-provoking and practical courses I’ve attended in a long time.”

Find out more and book your place here: https://psychsafety.com/psychological-safety-training-for-individuals/