Portability or ownership screens? Wrong question.¶
A plain-language companion to the working paper "Exit or Own? Portability Mandates and Ownership Screens" (v0.3, July 2026). The paper carries the model and the proofs; this text carries the ideas.
The short version. Europe runs two families of sovereignty instruments and argues about which to prefer. Ownership screens verify who the supplier is: control caps, immunity from foreign law, key custody. Exit mandates verify whether you can leave: portability, reversibility, switching rights. Priced in one model, they turn out to be answers to different halves of the problem, and, in a precise and policy-relevant band, strict complements: the screen only works when the mandate stands beside it.
Two instruments, two kinds of harm¶
The deep distinction is about observability. Some harms announce themselves: a price hike, a service cutoff, a deprecation. Exit-type instruments (portability, escrow, second sources, tested fallbacks) insure exactly these, because the trigger is visible and the response, leaving, is mechanical. Other harms never announce themselves: quiet legal access, influence over what you see and store. No exit clause fires on a harm you cannot observe; only knowing who controls the supplier prevents it. Screens are for the harms that stay silent; exits are for the ones that shout. Severity has nothing to do with it; observability is the whole line.
Why they are complements¶
Here is the mechanism, and it is one sentence: an exit mandate shrinks the prize a fake "European champion" would collect if it slipped through the screen, because a chunk of that prize is lock-in rent, and portability destroys lock-in rent; a smaller prize is easier to police, so the ownership screen needs less audit accuracy to deter the fake. The paper turns that sentence into an exact band: for a whole range of realistic verification capability, the ownership screen is infeasible without the exit mandate and feasible with it. Neither instrument substitutes for the other inside the band: without the mandate, no screen; without the screen, nothing touches the silent harms.
The three policy regions¶
Comparing the four policies (nothing, exit alone, screen alone, both), the map has three regions worth naming. Where the exposure is denial-dominated, the observable kind, and the screen is costly, the exit mandate alone is optimal: it treats the bulk of the exposure and the untreated remainder is small. Where the silent exposure is large and verification capability sits in the band, screen-plus-exit is optimal and neither alone works. And where capability falls below even the with-mandate bar, the honest conclusion is bleak: nothing reaches the silent exposure, and buying portability there and calling the workload protected is a category error. Portability in that region is a consolation, and the paper says so in those words.
Two subtleties worth carrying¶
Exit and price are dual instruments, and neither is free. A price premium strengthens the winner's discipline (more to forfeit if caught misbehaving) and simultaneously fattens the prize that attracts fakes. The exit mandate is the mirror image: it shrinks the fake's prize, and it also weakens whatever share of the genuine winner's discipline was lock-in rent. The model prices both faces instead of celebrating either instrument.
Exit mandates on licensed technology. An EU-owned operator running licensed foreign technology carries little legal-access exposure and keeps two silent-adjacent ones: the update stream, and license withdrawal, a denial with a delay. An exit mandate written on the licensed technology itself is the instrument the debate under-uses, and it prices the "sovereign wrapper" case directly.
The caveat, as everywhere in this programme: this is a model, its numbers verified against brute force rather than measured in the field, and its contribution is the logic: which instrument answers which harm, and when only the pair answers at all.