Delegated Economic Authority

The argument

Digital commerce runs on account-based trust. Authority is tied to an account and verified inside one platform, and losses are pooled and absorbed by institutions. The model works because each platform controls identity, authorisation and liability within its own boundary, and delegation already works well inside those boundaries through OAuth scopes, role-based access control and tokenisation.

Agents that discover, negotiate and pay across platforms break that assumption. A procurement agent with a mandate of office supplies only, $500 per transaction and $10,000 a month can operate on three platforms, each with its own controls, and spend $4,000 on each in a week. No local rule is broken. The $12,000 total surfaces days later in reconciliation. The essay calls this an authority coordination failure: there is no portable, verifiable representation of what the agent was permitted to do.

Payment execution cannot fill the gap. It scales by leaving authority unresolved: a processor verifies the payment method and fraud signals and has no view of the buyer’s internal spending policy. Card network guarantees do not transfer cleanly either, because agent transactions are heterogeneous and liability terms do not travel with them.

The proposed answer is delegated economic authority as a boundary mechanism: a minimal, interoperable way to express, verify and revoke what an agent may spend, on whose behalf, within what limits and until when. Its requirements are specific. Identity, authority and execution stay separable. Verification works without constant calls to a central party. Revocation tolerates real network conditions. Semantics are extensible. Audit is possible without universal observability. Explicit delegation makes liability assignable, though not absent, and turns the trade-off between traceability and privacy into a choice made by context.

Adoption should be staged: B2B procurement and enterprise automation first, regulated industries next, and consumer retail last, if at all.

Key points

• Account-based trust works where one platform controls identity, authorisation and liability.
• An agent acting across platforms can stay within every local limit while breaching its overall mandate.
• Payment processing and card guarantees leave open whether a purchase fits the buyer’s mandate.
• A portable record of delegated authority needs separable identity, local verification, tolerant revocation, extensible semantics and audit without universal observability.
• Coordination should start in bounded contexts such as B2B procurement.

Evidence and sources

• An argument from system design; the essay cites no external data.
• Delegation mechanisms discussed: OAuth scopes, role- and attribute-based access control, Verifiable Credentials
• Payment-side controls discussed: processor checks on payment methods and fraud signals, card network guarantee models

Limits

• Centralised, account-based systems remain a stable equilibrium and are likely to dominate most consumer commerce.
• In most consumer contexts, absorbing fraud losses stays cheaper than explicit delegation.
• The proposal is for coordination in bounded contexts, without a single universal standard.