Subtract three from 2030 and the baseline sits near 2027 [9], which is the part the deferral reading obscures: on that account the guidance rescheduled a strike rather than removing it [8], moving it to the moment when the building exists, the tools are in it, and the money is unrecoverable [7]. A dispute available before the concrete is poured costs little; the same dispute available only after it has cured costs whatever the ramp is worth per month. And the report's own logic says the reassignment question ripens late by construction, since it becomes a bargaining subject once staffing plans are specific enough that changes to work location or duties can be objectively anticipated [4].
The mechanism worth studying is bundling. A union need not oppose the plant at all; it can welcome the investment and negotiate everything downstream of it, folding housing support and relocation allowances into the same proposal as the staffing, and a company that has already spent the capex finds an outright refusal expensive [13]. The ministry's answer is procedural, and it works only if demands arrive pre-sorted: when opposition to investment, reassignment and pay sit in a single proposal, the argument moves to where the line falls and stays there [12].
Then the substitution question, which is where this actually bites. If transfers are contestable, the obvious workaround is hiring, and the Korea Enterprises Federation's position is that core workers cannot be secured through new hiring alone, so a finished plant could sit unable to run [17]. Deploying existing staff necessarily changes work location or job duties [6], meaning the workaround and the exposure are the same few hundred people. One business community official put it as timing rather than capital or technology being central to semiconductor competitiveness [10], which is a polite way of saying the start-up date is now partly a negotiated variable.
What the material does not contain is the number that settles it: no estimate of what a month of a dark fab costs, with 2030 the only forward date on offer [21]. Mediation could fix the boundary in a handful of cases and leave the residue as paperwork [11]. It could just as easily stay contested instead, with legal judgments over where personnel authority ends dragging on until delay itself becomes the bill [14]. Academics raise a third possibility, that the scope of personnel and management rights gets read down well past any single project [20].
The item worth pricing here is the option rather than the event: the credible possibility of a reassignment dispute raises a union's leverage during investment execution at zero strikes [15], and critics argue that is precisely what the guidance left unresolved [2]. It also constrains something quieter, which is the freedom to move skilled people between sites as a matter of routine [3]. What would falsify the thesis is a ministry that in practice holds the line it claims it can draw between a management decision and the working-condition changes that follow from it [5], because the report itself concedes the two are hard to separate.
