TradFi vs Onchain Finance

Editorial status: PUBLISHED. AI-assisted educational material; user batch approval recorded on 2026-10-19. Workflow measurements are invented; no production saving or deployment is claimed.

Learning objectives

• Compare financial functions under the same completion definition.
• Separate programmable records from external rights and settlement.
• Produce a source-backed TradFi-to-onchain infrastructure map.

Compare functions rather than identities

The useful question is which financial task changes: admission, issuance, trading, custody, registration, clearing, settlement or servicing. A token can reorganize records or permissions while the issuer, custodian, payment system and law remain relevant. BIS's arrangement framework combines underlying rights, representations and governance rather than assuming technology replaces every institution. BIS

PFMI separately identifies legal basis, finality, settlement money and linked obligations. A technical demonstration must not be described as complete financial settlement unless its actual boundary and evidence support that claim. BIS: PFMI

Worked example: a small saving after exceptions

Imagine an administrator handles 100 completed eligible transfers per month. Under the old teaching process, each requires ten minutes of staff work, totaling 1,000 minutes. At 30 currency units per hour, that narrowly defined cost is 500.

A proposed shared-record process takes two minutes each, or 100 of staff cost, plus 250 monthly support. Before exceptions it costs 350, a narrow saving of 150. Now assume 20 exceptions require ten extra minutes each, adding 100. Total becomes 450 and the saving is only 50, before migration and other omitted costs.

If migration costs an assumed 2,000 and the 50 monthly saving persists unchanged, simple payback is 40 months. That is not a forecast: volume, incidents, law, service scope and support costs can change. The point is to compare the same completed eligible transfer rather than an incomplete token instruction with a full old process.

Function → old owner/record → proposed owner/record
    → external dependency → normal completion → failure/recovery
Measurement: same task, quantity, cost boundary and evidence standard

Deliver the infrastructure map

For each function, identify the responsible legal entity, authoritative record, asset or money controlled, permission to change it, evidence consumed and output delivered. Show whether the function is preserved, reorganized or proposed. A planned component must be labeled planned rather than displayed as a working service.

Connect the edges precisely. A ledger event, order record, bank instruction and legal entitlement are different data. A token transfer can complete while the bank leg or external register remains unresolved. A smaller token unit can improve technical divisibility without proving investor access or market depth.

Include failure scenarios: compromised administrator, stale eligibility, mismatched registers, failed conversion, lost key and unavailable provider. Identify who can correct the state and on what authority. A recovery button without a defined legal and operational process is not enough.

The final report should state one measurable proposed improvement, its assumptions, retained dependencies, introduced risks and validation limits. It should also explain when the existing arrangement may remain simpler. The task is analysis, not a requirement to tokenize every asset.

Common misconceptions

“Onchain means no intermediaries.” Functions can move or be shared while obligations remain.

“TradFi means every record is manual.” Inspect the actual baseline rather than invent an inefficient one.

“A demo result proves production economics.” Workload, exceptions, recovery and full scope need evidence.

Risks and limits

No real administrator, technology deployment, legal finality or saving was measured. The map is an educational deliverable, not production architecture approval. The website's existing Build and Labs pages state their available scope; this lesson does not imply a connected bank, wallet or live securities system.

Exercise and answer

Exercise: New routine cost is 100, support 250 and exception cost 100, versus old 500. What is the narrow saving and simple payback on 2,000 migration if unchanged?

Answer: New cost is 450, saving 50 monthly, and simple payback is 40 months. Omitted costs and changing assumptions can alter or remove that result.

Related learning

Connect the TradFi track map to the equities, commodities and payments reports. Use Research to preserve evidence boundaries.

Source map

bis-tokenisation and bis-pfmi were actually read on 2026-10-18. They are international research and infrastructure principles, not certification of the proposed workflow. All volumes, times, rates and costs are original teaching assumptions; no production system was tested.