The Constitution Nobody Reads: Why National ICT Governance Fails Before the First Project Launches

Every Caribbean government has, at some point in the last fifteen years, produced a National ICT Policy. Many have also produced National ICT Plans. Some have produced both, then produced updated versions of both. Several have established statutory bodies specifically mandated to implement those plans.

And yet the question asked repeatedly — at every regional technology forum, in every CARICOM working group, in every post-implementation review — remains some version of: why isn’t the digital transformation working?

The answer is not that the policies were wrong. Many were genuinely sophisticated. It’s that they were treated as endpoints rather than instruments. A policy that exists but isn’t operationalised has no governing power.

Governance as Constitutional Instrument

Think of national ICT governance the way you’d think about a country’s constitution. A constitution matters not because it exists, but because it’s operationalised — through institutions, through courts, through a culture of reference and enforcement.

A constitution that sits in an archive and gets cited only in speeches governs nothing.

ICT policies have the same problem at a shorter timescale. They’re launched, sometimes celebrated, occasionally referenced in budget speeches, and then quietly superseded by whatever the most urgent procurement decision happens to be. When a ministry needs a new payroll system, it rarely starts with the national data architecture policy. It starts with the vendor who gave the best presentation last quarter.

The result is a national technology ecosystem built from hundreds of individually rational decisions that collectively produce incoherence: duplicate systems holding the same citizen data in incompatible formats; agencies unable to share information with each other; digital services that require citizens to provide the same information to three different ministries because none can see what the others hold. The UK’s National Audit Office has documented this pattern across multiple decades of government IT investment — departments that procured independently, built systems that couldn’t interoperate, and then faced remediation costs that exceeded what a coordinated approach would have cost.

Three Structural Gaps

The first gap is authority without enforcement. Most national ICT bodies in the Caribbean are advisory to line ministries rather than authoritative over them. They can publish standards; they cannot mandate compliance. They can develop enterprise architecture frameworks; they cannot require ministries to follow them. A governance body with advisory powers and no enforcement mechanism is, functionally, a think tank with a government letterhead.

Singapore’s Government Technology Agency (GovTech), established in 2016, operates differently. With both advisory and implementation authority, GovTech sets whole-of-government technology standards that stick. The Singapore Government Tech Stack — a suite of shared platforms available to all government agencies — exists because an agency had the authority to build it centrally and the institutional credibility to drive adoption across ministries that didn’t commission it.

The second gap is technical capacity in the wrong place. The technical expertise in most small-island governments sits in operational roles — managing existing infrastructure, responding to helpdesk tickets, keeping the lights on. The policy function — which requires different and arguably more sophisticated skills: enterprise architecture thinking, standards design, technology risk assessment, vendor governance — is chronically understaffed. Governments end up technically dependent on the vendors they’re supposed to be governing. That’s a structural conflict of interest no procurement policy fully resolves.

The third gap is measurement without consequence. Governance performance frameworks for ICT typically measure activity: how many projects are registered, how many policy reviews were completed. Activity metrics confirm that work is happening. They don’t confirm that the work is producing value. Estonia’s Information System Authority (RIA) publishes annual cybersecurity reports and system performance data that are publicly accessible and cited in budget discussions. That’s not transparency as performance. It’s measurement designed to drive decisions and hold institutions accountable across administrations.

What the iGovTT Model Points Toward

iGovTT, Trinidad and Tobago’s national ICT company, sits at an interesting structural position: a state enterprise responsible for providing innovative and impactful ICT solutions to enhance the efficiency and effectiveness of public sector services, occupying the space between governance and implementation. The tension in that structure — is it a regulator, an advisor, a delivery body, or a shared services provider? — is not unique to iGovTT. It’s a tension most Caribbean national ICT bodies navigate.

The resolution matters because it shapes incentives. A body primarily funded through implementation contracts has incentives that can diverge from its governance mandate. A body primarily funded by government grant or subvention has dependency risks of a different kind. The most functional models — GovTech Singapore, the UK’s Central Digital and Data Office — have clarity about what they are and institutional authority that matches their mandate. Ambiguity about the first question usually produces underperformance on both.

What Functional Governance Architecture Requires

The frameworks that work share three characteristics Caribbean implementations have largely missed.

They have mandate, not just mission: the governance body has statutory authority over cross-government technology decisions, including the power to delay or block expenditure that doesn’t meet architectural standards.

They have institutional memory: knowledge of what was built, why, how it works, and what it connects to lives in the institution — not in vendor documentation. When contracts end, capability remains.

And they use the citizen as the unit of analysis: every architectural decision is evaluated against the question of what it means for the person trying to complete a transaction with government, not what it means for the ministry requesting the budget.

The Caribbean has produced some genuinely sophisticated digital government policy thinking. The gap between policy quality and implementation reality is not primarily a resources problem. It’s a governance architecture problem — and specifically, the failure to build governance bodies with the authority, capacity, and measurement discipline to make policy documents into operational realities that persist across election cycles.