Take age out of who runs a site, not out of what experience earns

Japanese employers have been promising to move away from seniority-based pay for at least twenty-five years, and construction has moved more slowly than most. When contractors do reform, the usual model is borrowed from manufacturers and trading houses: job grades, market reference points, a performance element and the removal of age from the pay table. I led a reform of that kind at Asakage between 2021 and 2023, and I now think the usual model gets one important thing wrong for our industry. The fault in the old ladder was not that it paid people for experience. It was that it rationed authority by age. A contractor should remove age entirely from the question of who leads a site, and keep, in some roles strengthen, the reward for years of practice where skill genuinely accumulates with time.

What the seniority ladder was actually paying for

Before 2023, an Asakage site engineer's pay rose each year with age and tenure until fifty-five, with modest differences for performance. Promotion to the grade that allowed an engineer to lead a site came, on average, at forty-four. The system had real virtues. It made pay predictable for families, reduced rivalry within site teams that depend on one another, and kept people for decades in a business where much of the knowledge is never written down. Our union defended it for those reasons, and they were not wrong.

The ladder was paying for three different things at once: loyalty, accumulated skill and the right to authority. In the decades when engineers rarely left, nobody needed to separate them. By 2021 they had come apart. Our engineers aged 25 to 34 were leaving at 9.2% a year, and exit interviews named developers, consultancies and design firms that would let them lead projects in their early thirties. Several of those leaving had in practice run large sites as deputies to managers nearing retirement. They were not complaining that they were underpaid for their age. They were complaining that the title, the authority and the pay of the role they were already doing were reserved for someone older.

Separate the right to lead a site from the years served

The core of our reform was a set of five role grades defined by the scale and complexity of the site an engineer has led or can lead. Evidence comes from projects actually run, as deputy or manager; from the first-class qualification in construction management; from safety records; and from a review by two site managers from another branch, so that nobody is assessed only by their own superior. Age appears nowhere in the criteria. A thirty-three-year-old who has run a ¥3 billion warehouse with no lost-time accidents can be graded to lead one.

The union's objections improved the design. Its chair argued that older engineers would watch younger colleagues overtake them while their own pay froze, and that peer assessment would invite favouritism. We responded with a five-year protection of pay for every employee over fifty, an appeal panel with a union member on it and a published annual review of assessment outcomes by age band. We also agreed that no engineer would be graded below their existing level during the first two years, so that the assessments could be tested before anyone lost by them. The protection costs about ¥1.1 billion a year. I regard it as the price of a reform people could believe, not as a concession.

Since April 2023, 41 engineers under forty have been appointed to lead sites, against six in the previous five years. Turnover among engineers aged 25 to 34 has fallen to 5.6%. Average pay for site engineers rose 6.8%, and total personnel expense rose by ¥2.6 billion a year. A reform presented as a cost saving would not have survived the negotiation, and I am glad we never tried to present it that way.

Where years should still count

The mistake I made, and see repeated elsewhere, was to assume that once age left the pay table, experience should leave with it. Our first grade definitions measured site leadership well. They measured very badly the estimators, structural checkers, safety officers and civil supervisors whose value lies in having seen many projects rather than in leading one. A senior estimator who can tell from a drawing set where a foundation design will cost more than the client expects has acquired that judgement over twenty years. In the first year after the reform, eleven people in those roles left, several of them to competitors who still paid for tenure.

In 2024 we added a specialist track with grades that rise with recorded breadth of experience: the number and type of projects reviewed, inspections led and defects caught before handover. It does pay for years, but for years that leave evidence, not for years alone. The same principle applies at our craftsman-employing subsidiary, where a formwork carpenter's pay rises with certified skill levels that take years to reach. In the trades, time on the tools is a fair proxy for skill, and pretending otherwise would discourage exactly the patience a craft requires.

There is a version of this reform that becomes a performance-pay system borrowed wholesale from a trading house, with annual ratings deciding most of a site engineer's pay. I would resist it. Site teams succeed or fail together, and much of an engineer's best work, a hazard noticed or a struggling partner crew kept on the job, never shows in an individual rating. Our grades reward the responsibility people can carry and the evidence they have accumulated. The annual performance element remains small, and I intend to keep it so.

I would put the argument to another contractor's board in this form. Ask what your current pay structure rations, and to whom. If it rations authority by age, change that without hesitation, because your most capable young engineers will not wait. Then ask which roles in your company become more valuable with each additional year, and make sure the new structure still pays for them. A contractor that removes the reward for experience along with its age-based ladder will find, a few years later, that nobody left can tell it what its drawings will really cost to build.