Funding Digital Public Infrastructure as lawful capability, not platforms
- Ott Sarv
- 2 days ago
- 5 min read
A licence is refused. The person appeals. The agency can reproduce the outcome but cannot reconstruct the authority behind it. The logic sits in configuration. The logs are operational, not evidentiary. The correction path is unclear. Downstream systems have already acted on the decision.
This is a composite scene. It is also a predictable consequence of funding digital infrastructure as technology while expecting it to behave as public authority.
The governing rule is simple. The principle set out in Law Before Code is not a slogan. It is a sequence.
Governments are also changing how they fund digital work, moving towards staged and performance-oriented models for live services. That shift creates a fork. Funding can buy lawful capability, or it can buy platforms that simulate authority until the first serious dispute, a tension visible in the Performance Review of Digital Spend.
Why platform-led DPI funding fails when decisions are contested
Platform-first funding is often rational under real constraints. It simplifies procurement, produces visible milestones, and creates a single escalation route when capability is scarce and timelines are tight. Those incentives are legitimate.
The weakness appears at contestation. Platforms execute behaviour. Public authority requires attribution, evidence, and reversal. If correction and reversal are not funded as standing capability, disputes force manual reconstruction, policy improvisation, and uneven outcomes across cases.
A system that cannot show lawful origin, competent mandate, authoritative records, and a working remedy pathway is not merely incomplete. It is institutionally unsafe.
How governments are changing the DPI funding model in 2026
The Performance Review of Digital Spend argues for a significant shift in how digital initiatives are funded, managed, and tracked, with approaches better suited to iterative delivery and continuous improvement of live services.
This matters for funding digital public infrastructure because DPI is never finished. It is operated, improved, governed, and disputed. A build-only budget produces a build-only institution, and build-only institutions fail when reliance becomes widespread.
Staged, performance-oriented funding only works for DPI when the staged proof is not a feature list, but a set of lawful capabilities that can survive dispute.
What lawful capability means in funding digital public infrastructure
Term | Working definition for funding decisions |
Shared digital systems that support delivery of and access to services across society, which means a failure mode scales across multiple services and institutions | |
Platform | A technical product or suite that provides capabilities and integration patterns, which can accelerate delivery without establishing public authority or enforceable remedy |
Lawful capability | Ability to produce legally attributable outcomes and to correct and reverse them under contestation, consistent with the sequence in Law Before Code |
Remedy | Practical power to suspend effect, correct authoritative records, and reverse outcomes across dependencies, made operational in the Friday test |
Lawful capability is the missing budget line in many DPI programmes. When it is unfunded, legality becomes narrative, and the platform becomes the de facto regulator by default.
Law before code as a funding discipline, not a design preference
International definitions frame Digital Public Infrastructure as foundational systems enabling secure and seamless interactions across people, businesses, and government. The OECD definition of DPI similarly positions DPI as shared systems that support inclusive delivery and access.
That framing becomes incomplete once DPI supports eligibility, licensing, enforcement, certification, or access decisions. At that point, the state is producing public effects, and those effects must remain lawful, attributable, and reversible.
The Seven Layer Model resolves the ambiguity by sequencing law, mandate, canonical records, governed execution, and remedy as prerequisites for enforceable delivery, with operational pressure-testing expressed in the Friday test. The procurement implication is direct: a programme cannot be funded as DPI unless it can pass the sequence under pressure.
Rules as Code that strengthens authority instead of hiding it
Rules as Code is an approach to turning policy and relevant law into maintainable machine-consumable versions that can be used to design and create compliant services.
The opportunity is precision and speed in implementation. The danger is inversion. The rules engine becomes the de facto law, and interpretive judgement becomes hidden configuration. That is the failure mode that Law Before Code is designed to prevent.
For DPI funding, the practical instruction is that rule execution is not enough. Rule lifecycle and evidence must be funded as first-class deliverables, so the state can explain decisions, correct them, and prove what changed and why.
DPI funding release gates that prevent authority drift before production reliance
Funding release gate | Minimum proof for production reliance |
Legal origin gate | A traceable mapping from legal provisions to decision types and scope constraints, reflected in receipts and decision records, then a contested-case reconstruction that shows the legal basis for each decisive rule end to end, aligned to Law Before Code |
Institutional mandate gate | A mandate statement and delegation model that identifies the competent institution and supervisory chain, including the power to compel correction and reversal, then an escalation run that produces an attributable decision to uphold, correct, or suspend, consistent with guidance enforcement |
Canonical records gate | A record authority model covering provenance, retention, correction workflow, and propagation semantics, then a correction executed and confirmed across dependent systems within defined timelines, consistent with plumbing does not grant access |
Governed rule execution gate | A rule registry with version history, regression tests, approval logs, and rollback plan, then a rule change deployed, tested against scenario coverage, and rolled back without losing evidentiary trace, aligned to Rules as Code discipline |
Public interface gate | Receipts that preserve legal effect, status timelines, and accessible contestation entry points, then a contestation triggered that yields a complete evidence pack for review without manual reconstruction, aligned to the Friday test |
Remedy gate | A remedy authority model and reversal propagation design across dependencies, then a remedy drill that suspends effect, corrects authoritative records, and confirms downstream reversal with evidence-grade logs, consistent with Law Before Code |
These gates align naturally with staged, performance-oriented funding because each produces verifiable artefacts and observable drills rather than narrative claims, which maps to the direction set by the Performance Review of Digital Spend.
What finance ministries should demand in a DPI business case under staged funding
Funding review question | What the business case must include |
What public effects are being produced and what is the fiscal and legal exposure from errors | A quantified outcome and risk model for decisions and downstream reliance, including contestation volumes and the cost of incorrect outcomes, aligned to the live-service posture implied by the Performance Review of Digital Spend |
What counts as evidence, and can it be produced without manual reconstruction | An evidence model that specifies what is evidence-grade, how it is generated automatically, retention boundaries, and case reconstruction steps, aligned to the accountability posture in Law Before Code |
How policy change becomes controlled operational change | A rule lifecycle with versioning, tests, release governance, and rollback, aligned to Rules as Code |
Who can compel correction and how correction propagates across dependencies | A mandate and remedy reach statement showing who can suspend effect, correct records, and compel reversal, with drill plans aligned to the Friday test |
How the programme stays lawful as it scales | A staged release plan explicitly tied to passing the release gates, with measurable checkpoints rather than narrative milestones, aligned to the funding shift in the Performance Review of Digital Spend |
The funding decision that stops code becoming law by default
Platforms can produce outcomes. DPI must justify them and reverse them.
Production reliance is not authorised until the remedy gate is proven by drill, including correction of authoritative records and confirmed reversal across dependencies, consistent with Law Before Code.
