2024 Problem

ON 1 APRIL 2024, the Japanese government introduced a constraint for the construction industry. A hard cap on overtime hours, 720 per year, monthly ceilings, no exceptions, came into force for construction workers. The law had been passed in 2018, triggered by karoshi – death from overwork – but construction had won a five-year grace period to prepare. That period bought time. It did not buy transformation.

The Japanese called it the 2024 Problem. Not because anyone doubted it was coming, but because an industry that had spent decades compensating for a shrinking workforce through excessive hours was about to discover what happens when you remove the hours. Available labour supply fell off a cliff. Unfilled orders surged to a record ¥15 trillion ($103 billion). Expo 2025 pavilions slipped. A Tokyo municipal building was delayed two years. The workforce had already fallen 30% from its 1997 peak; 37% of those remaining were over 55; only 12% were under 30.

The constraint was absolute. And then the interesting part began.

BIM adoption rose from 48% to 59% in two years. The Construction RX Consortium, a shared robotics platform founded by 16 contractors in 2021, grew to 312 members across 15 subcommittees developing common standards for welding, transport, and inspection robots. Remote construction expanded from disaster-recovery niche to routine operations. NIKKENREN (the Japan Federation of Construction Contractors) saw its initial productivity target achieved five years early; a new target of 25% improvement by 2035 replaced it. The construction robotics market reached $10.1 billion in 2024 and is projected to triple by 2033.

Ernest Rutherford put it best a century ago: “Gentlemen, we have run out of money. It is time to start thinking.”

Japan ran out of hours. It started thinking. The pattern holds closer to home. When Covid shut UK construction sites in 2020, the CLC produced Site Operating Procedures within days. Digital inductions replaced toolbox talks. Remote inspections moved from pilot to standard practice. Design coordination that had required physical colocation shifted to model-based collaboration overnight. The tools had existed for years. Adoption curves that the industry had projected over a decade compressed into weeks.

Organisations and the individuals within them naturally exhibit status quo bias: the known costs of current practice are familiar and distributed, whilst the costs of change are concentrated and uncertain. Rational actors choose the familiar path even when they accept the trajectory is unsustainable. Optimism bias compounds this: every decision-maker believes they can defer the reckoning one more year. But constraints eliminate the option to defer.

Encouragement moves the needle. Constraint moves the system.

Japan had tried encouragement first. i-Construction, launched by the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) in 2016, was the government’s flagship programme to modernise the sector through technology. It covered ICT-enabled earthworks, 3D surveying, drone inspection, BIM standardisation, and productivity benchmarking across all public works contracts. By FY2022, 86% of MLIT-managed earthwork projects used ICT methods. It achieved a 21% productivity improvement over eight years. Respectable. But also wholly insufficient to address a workforce projected to halve by 2045.

Three Design Principles, Not Three Policies

What Japan built in response deserves attention not as a set of policies to replicate but as an architecture to consider.

• The push-pull. In June 2024, Japan amended its Construction Business Act. The amendments simultaneously tightened constraints on traditional practices (prohibiting below-cost pricing, mandating good-faith cost renegotiation) and relaxed regulations inhibiting technology adoption (allowing site engineers to supervise two projects via remote monitoring, formally promoting ICT within the statutory framework). They made the old way progressively harder and the new way progressively easier.
• The shared platform. The Construction RX Consortium places all five of Japan’s Super General Contractors  (the equivalent of UK Tier 1’s) alongside 282 cooperating companies, jointly developing robotics standards on a common open-source software platform. Competitors collaborated on standards. They competed on delivery. This was platform thinking applied to construction technology.
• The institutional connector. Japan’s MLIT simultaneously wrote the regulations, procured the projects, funded the R&D, ran the training programmes, and published the standards. When it mandated BIM for all public works in FY2023 (two years ahead of schedule), it simultaneously allocated ¥6.5 billion (c.$43 million) in subsidies for smaller firms to adopt it. When it set a 2040 target for site automation, it immediately began defining new occupational categories, “Automated Construction Coordinators,” “Remote Construction Operators”, and developing national training programmes to fill them. One body. Joined-up action.

The Expensive Lesson of the 1990s

Japan’s experience carries a warning alongside the invitation. In the 1990s, Shimizu, Kajima, Obayashi, and Taisei each built fully automated “building factories”: robotic enclosures that assembled skyscrapers floor by floor. At least 60 buildings were erected this way. Billions invested. Then the firms quietly stopped. The ROI horizon was 20+ years. The systems were too rigid for variable project types. The technology worked. The economics didn’t. (The construction sector’s talent for expensive lessons in technology adoption is, it turns out, universal.)

What Japan learned is that construction automation succeeds at portfolio scale through modular, task-specific interventions coordinated by shared standards. Shimizu’s current Smart Site deploys a welding robot on one floor, an autonomous forklift on another, a ceiling installation unit on the next. Each handles a discrete, repeatable task across the company’s entire project portfolio. For a single 30-storey building, this approach projects 75% labour reduction in material transport and 70% in welding. The difference between the 1990s and now is not the ambition. It’s the architecture.

The UK’s Missing Connective Tissue

The UK faces every structural pressure Japan faces, on a delayed timeline. Workforce ageing. Recruitment challenges. Flat productivity. Projections that over 250,000 additional workers will be needed by 2028 just to meet existing demand. Housing targets require building at rates we have not achieved in decades. Net zero commitments require retrofitting millions of homes using trades already in acute shortage.

We have good raw materials. World-leading BIM standards. An active ConTech ecosystem. A Construction Playbook that says the right things.  Institutions that care: the CLC, Build UK, NISTA.

None of these bodies lack competence or commitment. What they lack is the connective authority that MLIT possesses. No single institution in the UK can simultaneously set the regulation, shape the procurement, fund the R&D, and mandate the training. Our functions are distributed across at least seven bodies. Each produces valuable work. But, without constraints, none can compel the others to act in concert.

The challenge is not the absence of good technology. It is the absence of the institutional architecture that makes technology deployment rational at scale.

There is a deeper problem still. Much of the UK’s response to workforce shortage focuses on training more people into the existing system rather than redesigning the system itself. If the underlying production model remains labour-intensive, fragmented, and site-dependent, training simply feeds people into the same bottleneck. Japan’s lesson is that workforce constraint must be met by changing what the workforce does, not just how many people do it.

Individual firms under-invest in automation because the payoff accrues across portfolios, not projects. Industry under-invests in shared standards because the collective benefit is diffuse whilst the individual cost is concentrated. Government under-invests in system design because the political reward goes to visible initiatives, funding announcements, strategy launches, rather than invisible infrastructure: regulatory alignment, institutional coordination, standards development. The coordination failure runs three layers deep. Japan has begun to overcome each layer not through any single intervention but through an interconnected system where constraint, collaboration, and enablement reinforce each other.

The question is not whether we face our own version of the 2024 Problem. It is whether we design the response in advance or improvise it under crisis conditions. A government that builds the coordination architecture now can manage the transition. A government that waits will be managing the crisis.

Japan’s construction sector has, under duress, built exactly the kind of institutional architecture that the UK knows it needs and has spent two decades not building. The principle is there for the taking.

Rutherford’s question stands. Have we run out of enough yet… to start thinking?