See the risk that lives between participants.
FromIsolated records→ToA network graph
Each participant can look healthy on its own while the risk sits between them. This shift asks whether finance could see the connections: the shared dependencies, concentrations and correlated losses that no single record shows.
Healthy one by one
Seen one by one, every participant in Meera’s transaction looks healthy illustrative. Seen as a graph, two things appear. Meera and two other creators whom Arjun stocks all depend on the same processing collective. And every one of Arjun’s deliveries runs along a single mountain road that closes in the monsoon.
Neither shows in any individual record. Both show in the connections.
From records to connections
Each participant, separately
The relationships between them
| Today | On an open network |
|---|---|
| Each participant separately | The relationships between them |
| Individual exposures | Shared dependencies and correlations |
| Separate records | An auditable trail across the network |
An open network is a graph. A good deal of its risk lives on the edges.
How it would work
A network graph, with governance. The graph connects operator, brand, product, supplier, manufacturer, stockist, geography, channel, transaction, capital and outcome. It gives an auditable trail from a decision to a capital commitment, an order, fulfilment, payment and outcome.
Governance separates six kinds of information, each with its own access rules: participant data, transaction records, derived intelligence, underwriting information, reputation information and proprietary intelligence.
Go deeper
What already exists
Banks already measure concentration and dependency. The Basel standard on large exposures limits a bank’s exposure to a single counterparty, or to a group of connected counterparties, to 25 per cent of its Tier 1 capital[1]. Open networks already connect participants through shared protocols and roles[2]. This shift applies the same thinking to the dependencies between small participants.
Scope. We haven’t built or tested a graph, and we don’t know how much earlier it would flag a concentration than watching participants one by one.
Who sees what
A graph links businesses and, through them, people. Who can see which links is the heart of governance: participants see their own, providers see what consent allows and nobody sees everything. Data-protection law applies wherever the graph holds personal data[3].
Getting it right
Start with the dependencies that matter and are cheap to see: a shared producer, a shared road, a shared node.
Use it to size and spread exposure, not to blacklist.
Explain every flag. A signal nobody can explain won’t be acted on.
Who it helps. Providers spot a concentration before it becomes a loss. Meera and Arjun can be offered a second route or a second collective instead of a refusal. The network gets steadier as it grows.
What could go wrong
- Surveillance by another name.A graph that sees everything gives too much to whoever holds it, so access follows role and consent.
- Flags without causes.A signal nobody can explain won’t be acted on, so each flag shows the connection behind it.
- Costly to build for modest gain.The pilot builds the cheapest useful links first.
- Misleading links.Two participants can share a supplier harmlessly, so the graph shows dependence, not guilt.
How we’d test it
In one corridor, build a simple graph of shared producers, routes and stockists.
- The test
- Graph against watching one by one
- What we would measure
- Test whether the graph flags concentration and correlated losses earlier than monitoring each participant separately, and early enough to act.
- What would end the idea
- The graph doesn’t flag concentration or correlated losses earlier than participant-level monitoring, or too late to act.
Sources
- [1]Bank for International Settlements, Supervisory framework for measuring and controlling large exposures (Basel Committee, 15 April 2014)Checked; the standard-setter’s own page
- [2]Ministry of Commerce & Industry, Revolutionizing Digital Commerce: The ONDC Initiative (4 January 2025)Checked; a government document
- [3]PRS Legislative Research, The Digital Personal Data Protection Bill, 2023Checked; a summary of the Bill as passed by Parliament in August 2023
Spotted a source we missed or a rule we misread? Tell us what we got wrong.
The network can now see risk, evidence it and move it. How would anyone know it’s working?