Japan's job-based pay: how many companies really do it

Many Japanese companies now call themselves "job-based" (jobu-gata). Very few actually pay by the job. Depending on how you ask, the share falls from almost 80% to around 6%.

If you're hiring from, benchmarking against or selling into Japanese companies, the label tells you little. Three questions tell you a lot more.

Three numbers for the same thing

In a 2019 survey by the Japan Productivity Center, 78.5% of companies said they used role- or job-based pay for managers, and 57.8% for non-managers. Read on its own, that would mean Japan already pays by the job.

In Keidanren's 2025 survey of its member companies (337 responses), 26.4% said they had introduced, planned or were considering job-based HR. The other 73.6% had not.

Look inside that 26.4%. Three quarters have actually introduced it. Only 22.9% apply it to all employees, and only 2.4% to all non-managers. Most apply it to managers only. That 22.9% includes companies still planning or considering it, so multiplying through gives an upper bound: at most about 6% of responding companies have job-based HR for everyone in place or on the way. Count only those already running it and the share is smaller.

Where the gap comes from

I see two sources.

The first is vocabulary. Role-based pay (yakuwari-kyu), which pays a band according to the size of a role, is common in Japan, and surveys often group it with job-based pay. That makes job-based pay look widespread. But a role is usually broader than a job. Very different work can sit inside one role, and moving to another assignment often doesn't change pay. Even among Keidanren companies that have adopted job-based HR, only 55.3% use job pay as the pay item. A quarter still use role pay.

The second is maintenance. Job-based HR runs on job descriptions: what the job does, what it requires, what it's accountable for. In Keidanren's survey, 40.7% of adopting companies said they don't update job descriptions on any regular schedule. A description written once and left alone stops matching the job. After that it can't set pay.

The label moved faster than the system. Companies that haven't adopted it are blunt about why. It "doesn't fit our business" (50.4%) and jobs are "hard to define" (40.3%).

Three questions that separate real from label

When a company tells me it's job-based, I check three things.

Is there a job description, and when was it last revised? If it exists and has been touched in the past year, it's in use. If there isn't one, or "we got one when I joined," it's probably just a name.

When someone changes roles, does their pay change? If pay follows the job, moving to a bigger or smaller job moves the pay band. If the work changes and the money doesn't, pay is still attached to the person.

Does a demotion mean a pay cut? This is the one that tells you most. Japanese labor law and case law make pay cuts hard, and many companies that claim job-based pay only ever move pay up. If it can't go down, it's still personal.

Far fewer companies can answer all three than call themselves job-based.

What this means for hiring and competitor work

In interviews, these questions tell you what a candidate's grade at their last company actually means. Two people with the title "manager" can have very different scope and pay depending on whether the company really pays by the job or by tenure. Benchmark pay on titles alone and you'll get it wrong.

The same goes for competitor analysis and talent mapping. A competitor that says "job-based" on its website isn't necessarily one whose people you can recruit by the job. Where pay follows the job, prices are visible and you can talk in terms of the role. Where it's only a label, pay is buried in tenure, and you need to know how much of it a candidate loses by moving. I wrote about choosing which companies count as competitors in Picking the wrong competitors.

For the long-run picture on pay levels, see Japan's wages since 1945.

Keidanren's members are mostly large companies. Smaller firms are probably more cautious still. The Japan Productivity Center figures are from 2019, and I couldn't find a newer survey on the same basis, so the two sets can only be shown side by side.

My guess is that job-based HR in Japan never becomes the US model. It may settle as a hybrid, with managers paid by the job and everyone else by the person. If so, looking at the specific role matters even more than what the company calls itself.

How many of the three questions can your company answer?