Why Does Bureaucrat-Led Tech Policy Lag in Japan? Job Rotation and the Humanities/Science Track Split

Generalist-track personnel make technical judgment hard

Look across Japan's technology-policy ministries and you'll find that officials with a technical background are a small minority. Career-track staff hired through the national civil service exam (the old "Type I" exam) climb a ladder -- staff, deputy section chief, section chief, deputy division director, division director -- while rotating between bureaus and divisions every one to two years. The whole system is designed to produce generalists.

That design makes sense for building broad administrative judgment. But in fast-moving technology fields, it has a structural weakness: by the time an official actually starts to understand the technology, they've often already rotated to a different post. It's common for someone trained in law, economics, accounting, or tax policy to be assigned to a technology-policy role with no grounding in the underlying technical basics, and then be expected to forecast the next several years of technological change from that position. That's an extremely difficult bar to clear.

The author once dealt with a technical ministry where the point of contact happened to hold a PhD in biology from Kyushu University, and the depth of that person's expertise made a real difference in the quality of the response. In the author's experience, though, that was a fairly rare case. Technical PhD holders staying long-term at the center of the bureaucracy appear to be the exception rather than the rule.

A slow budget process: over a year from request to approval

Technology-policy work often flows through ministries like MEXT commissioning research to universities or research institutions. Here, Japan's single-year budget cycle becomes a drag. Each ministry compiles its budget request (概算要求, "gaisan yōkyū") around midsummer for the following fiscal year; after review and Diet deliberation, it's finalized near the end of the fiscal year. In other words, getting from a budget idea to actually usable funding typically takes at least a year.

In fast-moving technology fields, that one-year gap can be fatal. On top of that, research funding is often said to concentrate in a small number of institutions (the University of Tokyo and the University of Tsukuba are frequently cited examples, given their track record of running large-scale projects).

Procurement structure and the IT industry's subcontracting layers

Government system procurement in Japan connects into an industry structure topped by a handful of large systems integrators, with multiple layers of subcontracting beneath them. When the buying side (the ministry) lacks technical judgment of its own, it tends to lean on the vendor's own explanations during requirements definition and acceptance testing, which in turn makes relationships with specific vendors more likely to become entrenched. This dynamic isn't particular to any one ministry -- METI, the Ministry of Internal Affairs, and MEXT all show roughly the same pattern. It's probably more accurate to say that "bureaucrat-led tech policy runs late" is a structural feature shared across the entire bureaucracy, not a defect of any single ministry. The subcontracting structure of Japan's systems-integrator industry itself is examined in more detail, from the angle of contract types and billing units, in a separate article.

At the root: the humanities/science track split at 17 or 18

The personnel, budget, and procurement problems described above all trace back to an "entry point" issue. Japan's high school and university entrance system effectively locks most students into a humanities or science track at the early age of 17 or 18. Once a student chooses a business, law, or economics faculty, the path back to a systematic science education (mathematics, physics, computer science) is technically still open, but in practice very hard to access.

It would be heavy-handed to simply declare "the humanities are bad." That said, the role of expertise like law, accounting, and tax policy may itself be changing: these fields lean heavily on relatively formal rules and pattern-matching against precedent, which makes them comparatively easy targets for specialized AI to take over, and the need for human specialists in these areas is plausibly going to shrink over the medium-to-long term. If that's right, then the more accurate framing isn't "the humanities are bad," but rather "building a career on a single narrow specialty leaves you exposed to the risk of AI displacement." Nor is this an argument that every student needs an aptitude for math and engineering -- aptitude and interest vary between individuals and have nothing to do with gender. What's actually at issue is the design of a system that, regardless of aptitude or interest, effectively closes off access to a basic science education after a single choice made far too early, at 17 or 18. Asking someone that age to pick a career path based on a forecast of how technology will develop years down the line is an unreasonable demand.

Summary

• Tech-policy delays should be understood as the product of several combined systems -- generalist-track job rotation, the single-year budget cycle, and procurement structure -- rather than individual bureaucrats' competence
• Rotational personnel management has a structural weakness: officials often rotate away just as they start to genuinely understand a technical field
• The 1+ year lag between a budget request and its approval is a poor fit for fast-moving technology areas
• Funding concentration and ties to the multi-layer subcontracting structure are common across the bureaucracy, not specific to one ministry
• The root cause traces to the humanities/science track split at 17-18, which can be understood through the economic concept of information asymmetry
• The issue isn't "the humanities are bad" as a judgment on people -- it's an education system design that closes off access to science education after one premature choice, at a time when narrow legal/accounting/tax specialization may itself face AI-driven displacement