What is reusable identity? Verify once, use everywhere

The three architectures of reuse

Wallet credentials are the standards-track architecture, the one with the deepest institutional backing: a government or accredited issuer verifies the person once and provisions a cryptographically signed credential, an mDL under ISO 18013, an EUDI wallet attestation, a W3C verifiable credential, into the holder's device. Reuse is presentation: the holder shows the credential, or one fact from it, to any relying party, which verifies issuer signatures against trust lists with no call home. Selective disclosure is native, the privacy story is strongest here, and the model's trust anchor is the issuance ceremony.

Identity networks are the operational architecture: a provider, an ID.me, a bank-consortium scheme, a national login, verifies the person once into an account, and reuse is federated login plus attribute assertion, the network vouching to each relying party. Networks ship fastest, carry their provider's liability and support, and concentrate exactly what they amortize: one provider's enrollment weakness, outage, or breach touches every connected service at once.

Portable verification records are the pragmatic architecture: the result of a strong verification, not a government credential, made reusable, a verified-identity token a user carries between services inside one platform's ecosystem or a vendor's network. It is reuse without waiting for issuers, and its assurance is precisely as good as the original verification event and the binding that connects the present user to it.

The trust equation every reuse must solve

All three architectures compress into one equation: the relying party trusts tonight's presentation to the exact degree that three earlier things held, the enrollment, the binding, and the lifecycle.

The enrollment is the root: whoever verified the person originally set the assurance ceiling for every reuse after. A credential provisioned against a stolen identity kit, or a network account enrolled past a weak liveness check, is a fraudulently minted master key, cryptographically perfect forever. This is why issuer and network provisioning posture, who proofs enrollment, with what liveness, against what attack mix, is the first diligence question, and why enrollment ceremonies are where the fraud economy has already moved, from fake passkey registrations to wallet provisioning with breached documents.

The binding is the nightly question: is the person presenting the credential the person it describes? Device possession plus a PIN answers it weakly; the wallet's biometric gate answers it better but invisibly to the relying party; a session-level liveness-anchored match against the credential's signed portrait answers it directly. Risk tiering lives here: an age gate can ride possession, a payout cannot.

The lifecycle is the long tail: credentials get revoked, networks suspend accounts, people's attributes change, and a reuse architecture is only as current as its revocation and refresh machinery. Relying parties must actually check status at presentation, and the measured lag between issuer revocation and relying-party refusal is a number worth demanding from any scheme.

What reuse changes economically

The economics explain the adoption curve better than the technology does. For relying parties, verified-user acquisition cost falls toward the price of a credential check, and conversion rises as onboarding drops from minutes to seconds, with projections like the US mDL base growing from 21.7 million to 143 million by 2030 defining the addressable curve. For users, the hundredth service no longer means the hundredth document photo in the hundredth database, a data-minimization gain privacy regulators actively prefer. For fraud economics, reuse cuts both ways: strong reusable credentials raise the attacker's cost per identity dramatically, while successful attacks amortize identically, one fraudulent enrollment reused across a hundred services, which is why book-level analytics, the same face or credential recurring where it should not, remain load-bearing even in a credential world.

Consuming reusable identity well

For relying parties, five rules keep reuse an upgrade. Treat every reusable source as weighted evidence in one decision policy, never as a bypass lane: an mDL presentation, a network assertion, and a fresh document check should land in the same decision plane with weights reflecting enrollment posture and binding strength. Maintain an issuer and network posture table, and let it move weights as schemes mature or stumble. Tier the present-person bind by action: possession for low stakes, liveness-anchored portrait match for onboarding, money, and recovery. Keep equal-assurance fallback routes, because reuse coverage is a minority of users for years yet, and a reuse-only flow is an exclusion engine. And log acceptance per decision, source, issuer, assurance, binding, outcome, because the regulators who opened the door to credentials will audit how they were consumed, and the acceptance playbook is ultimately an evidence discipline.

The strategic read for 2026: reuse is no longer a prediction, it is a quarter-by-quarter rollout, and the relying parties that built consumption as a risk discipline, rather than a checkout shortcut, will be the ones for whom the wallet era compounds instead of detonates.

The liability question nobody has settled

Reuse redistributes verification work; it has not yet cleanly redistributed verification liability, and relying parties should price the open question. When a fraudulently enrolled credential passes a bank's onboarding, the regulatory obligation, forming a reasonable belief in the customer's true identity, stays with the bank, whatever the issuer's enrollment failure, which is precisely why the US guidance frames credentials as evidence within a program rather than a safe harbor. Networks negotiate the question contractually, assurance levels, indemnities, and liability caps living in the relying-party agreement, and those terms deserve the same scrutiny as the technology, because a network that warrants nothing is selling convenience, not assurance. Wallet ecosystems are building the formal answer, trust frameworks that certify issuers and define reliance rights, the EUDI regulation and the UK's register being the furthest along, but certification historically attests process, not outcomes, and the first large mDL provisioning fraud will test how reliance actually allocates loss. The practical posture until the law matures: treat reusable sources as strong evidence with documented weights, keep your own binding and risk layers where obligations are yours, and record per decision which source you relied on and how far, because whatever the eventual allocation rules say, they will be applied to the records you kept.