Government- and donor-sponsored startup programs keep multiplying across East Africa — JICA’s Project NINJA alone has supported 824 startups across developing countries, including Uganda’s ministry-run NINJA Acceleration Program, while Entrepreneur Support Organizations in Africa have absorbed over $500 million in donor funding in a decade (1)(2)(3). Yet the coverage remains relentlessly announcement-driven — new cohorts, new calls for applications — with almost no published evidence on whether graduates actually grow revenue, survive, or raise follow-on capital. The uncomfortable truth: accelerators are measured on inputs (cohorts trained) because nobody contractually owns the output (firms scaled). Until government programs are paid on graduation outcomes — contracts won, capital raised, jobs at month 24 — they will optimize for ceremonies. And as ministries become direct program owners, program-design quality is becoming industrial policy by stealth.
Key Takeaways
- Government and donor startup programs are proliferating: JICA’s Project NINJA has supported 824 startups across developing countries, including a ministry-run program in Uganda (1)(2).
- Entrepreneur Support Organizations in Africa have absorbed over $500 million in donor funding over the past decade, yet published post-program outcome evidence remains scarce (3).
- Coverage is overwhelmingly announcement-driven — new cohorts and calls for applications — with almost no public data on graduate revenue growth, survival, or follow-on capital (4).
- The structural flaw: accelerators are measured on inputs (cohorts trained) because no one contractually owns the output (firms scaled), so programs optimize for ceremonies over results.
- As ministries become direct program owners across Kampala, Nairobi, and Kigali, program-design quality is becoming industrial policy by stealth — a high-stakes, under-scrutinized lever.
- The fix: pay government programs on graduation outcomes — contracts won, capital raised, jobs at month 24 — and require public outcome reporting, shifting the incentive from ceremony to impact.
Why do accelerators optimize for ceremonies?
Because what gets contracted gets delivered — and government and donor programs are contracted on inputs they can guarantee, not on outcomes they cannot, so the entire incentive structure rewards activity over impact.
Consider how a typical government or donor accelerator is funded and evaluated. A ministry or development agency commissions a program with a budget and a set of deliverables — but the deliverables it can write into a contract and verify are inputs and activities: run a cohort, train this many founders, hold a demo day, disburse these grants. These are things the program operator can control and guarantee. What the funder actually wants — firms that scale, jobs that endure, capital that follows — are outcomes the operator cannot guarantee, because they depend on the founders, the market, and forces outside the program. So the contract specifies what can be guaranteed (cohorts trained) rather than what is wanted (firms scaled), and the program, rationally, optimizes for the contracted inputs: it runs the cohorts, holds the ceremonies, produces the reports of founders trained and demo days held, and is paid. Whether any of those founders built a durable business 24 months later is, contractually, not the program’s problem.
This is why coverage of the sector is so relentlessly announcement-driven — new cohorts, new calls for applications, new demo days — and so empty of outcome data (4). The programs report what they are paid to deliver (activities) and there is no contractual reason to track, let alone publish, what happened to graduates afterward. The result is an “accelerator-industrial complex” that is enormously busy and almost entirely unaccountable for results: over $500 million in donor funding absorbed by African ESOs in a decade (3), thousands of founders trained, countless demo days held — and remarkably little public evidence on whether the businesses are better off. This is not because the operators are cynical; most are committed people doing real work. It is because the incentive structure rewards ceremony over outcome, and people respond to incentives. The deeper diagnosis is the same one that afflicts accelerators that report attendance instead of survival and revenue: the binding problem is that no one is measured on, or paid for, the outcome that matters.
Why does this matter more now?
Because the stakes have risen sharply: ministries are becoming direct program owners, which means accelerator design is quietly becoming industrial policy — and a poorly designed program is now a poorly designed industrial intervention.
For years, startup accelerators were largely the province of NGOs, donors, and private operators — important, but peripheral to national economic strategy. That has changed. Governments across East Africa are now direct owners of entrepreneur-support programs: ministries run accelerators, design national startup policies, and deploy public budgets into entrepreneurship programming. JICA’s NINJA program in Uganda runs with the Ministry of Trade (2); ministries elsewhere own their own programs. As donor funding recedes (the USAID collapse accelerated this), government becomes the largest remaining buyer and owner of entrepreneur support in the region. This is a profound shift in what these programs are: when a ministry runs an accelerator, that accelerator is no longer just a training program — it is an instrument of industrial policy, a public intervention meant to build the firms and jobs that drive national development. Program design has become industrial-policy design, deployed with public money toward national economic goals.
And here is the concern: this industrial policy is being conducted largely by stealth, without the scrutiny that industrial policy of this consequence deserves. If a government were building a $50 million factory, there would be feasibility studies, outcome targets, and public accountability. But governments are deploying comparable sums into accelerator programs — programs that, the evidence suggests, add value only in a small right tail while the median may destroy value — with little of that rigor, measured on cohorts trained rather than firms scaled. The quality of program design is now determining whether significant public industrial-policy spending builds durable enterprises or funds ceremonies, and almost no one is scrutinizing it as the industrial policy it has become. The region is making consequential industrial-policy bets through accelerator design, and making them in the dark. This raises the stakes on getting the accountability right: it is no longer just about wasted program budgets, but about whether a major lever of national development works.
What do government programs owe their founders?
Honest accountability for outcomes — which means being measured on, and ideally paid for, what happens to graduates, and publishing the results so the programs can be judged and improved.
The reframe is to ask what a government accelerator owes — to the founders it graduates, to the taxpayers who fund it, and to the national development goals it claims to serve. The answer is accountability for outcomes, not just delivery of activities. A program that takes public money to build enterprises owes evidence that it builds enterprises — and currently, most cannot provide it, because they neither track nor are measured on outcomes. Three obligations follow:
Owe outcome measurement. A government program should track what happens to its graduates — survival, revenue growth, capital raised, jobs created — at meaningful intervals (12, 24 months), against a counterfactual where possible. Not to punish, but to know: a program that does not measure its outcomes cannot improve them and cannot justify its public funding. This is the measurement discipline that distinguishes programs that build firms from those that run event calendars.
Owe outcome-linked payment. The structural fix is to change what programs are paid for: tie a meaningful portion of payment to verified graduate outcomes — contracts won, capital raised, jobs at month 24 — rather than only to activities delivered. This is precisely how governments should buy acceleration: as a service paid partly on results. When payment follows outcomes, programs optimize for outcomes, and the ceremony incentive dissolves.
Owe public reporting. Government programs, funded by public money and serving public goals, owe public outcome reporting — published evidence of what happened to graduates, so taxpayers, founders, and future funders can judge whether the program works. Publishing outcomes is both an accountability obligation and a discipline: a program that knows its results will be public designs differently than one that reports only ceremonies.
These obligations are not punitive; they are what converts a busy, unaccountable accelerator-industrial complex into an effective instrument of industrial policy. A program that measures, is paid for, and publishes its outcomes is one that actually serves the founders and the nation rather than the demo-day calendar.
The Accountability Ledger: making programs answer for outcomes
Here is the framework I would put at the center of any government accelerator’s design — and it is, in effect, what such a program owes. Call it the Accountability Ledger — four entries that shift a program from input-accountability to outcome-accountability.
Entry 1 — Define the outcomes that matter. Specify, before the program runs, the outcomes it is accountable for: graduate survival at 24 months, revenue growth, follow-on capital raised, jobs created. These are the real objectives of industrial policy, and naming them is the first act of accountability.
Entry 2 — Measure against a counterfactual. Track those outcomes and, where possible, compare graduates against a comparable non-participant group — because “our alumni grew 40%” is meaningless without knowing how similar non-participants did. The counterfactual is what separates real impact from selection effects, the standard the serious evidence on acceleration applies.
Entry 3 — Tie payment to outcomes. Structure program funding so a meaningful share follows verified outcomes, not just delivered activities. This aligns the operator’s incentive with the result the funder actually wants and dissolves the ceremony-optimization trap.
Entry 4 — Publish the results. Make the outcome data public. Public reporting is the accountability that public money and public goals demand, and it is the discipline that drives program improvement and lets good programs be distinguished from busy ones.
The Accountability Ledger reframes what a government accelerator is for. The input-accountable program asks “did we run the cohorts?” The outcome-accountable program asks “did the firms scale, and can we prove it?” The difference is the difference between industrial policy that works and industrial policy conducted in the dark — and it is what government programs owe the founders, the taxpayers, and the national goals they serve.
What should governments and program designers do?
Treat accelerator programs as the industrial policy they have become — with the outcome accountability that consequential public spending demands.
The practical agenda for governments and ministries now owning these programs is to design them around the Accountability Ledger: define the outcomes that matter, measure them against counterfactuals, tie payment to verified results, and publish the evidence. This means procuring acceleration as a service paid partly on outcomes rather than funding activities — buying acceleration the way a demanding customer buys any service — and selecting operators who can deliver the right-tail results, drawing on the design lessons of what actually works in acceleration. It also means honesty about scale: a program measured on outcomes will likely graduate fewer, better-supported founders than one measured on cohort size, and that is the correct trade. The region’s relationships with outcome-minded partners — including JICA and the Japan-Africa industrial pipeline that runs through programs like NINJA — are natural venues to model this outcome accountability.
The conclusion frames the stakes plainly. East Africa is building an accelerator-industrial complex — hundreds of government and donor programs, over half a billion dollars of ESO funding in a decade, thousands of founders trained — and it is doing so largely without knowing whether any of it builds durable enterprises, because the programs are measured on ceremonies, not outcomes. As ministries become the direct owners of these programs, this is no longer a question of wasted training budgets; it is a question of whether a major instrument of national industrial policy works. What government programs owe the founders they graduate — and the taxpayers who fund them — is accountability for outcomes: measurement, outcome-linked payment, and public reporting. Until programs are paid on graduation outcomes rather than cohorts trained, they will rationally optimize for ceremonies, and the industrial policy conducted through them will remain a bet placed in the dark. The accountability is not punitive; it is the discipline that turns a busy complex into an effective policy. Pay for outcomes, measure against counterfactuals, publish the results — and the accelerator finally serves the founders and the nation, not the demo-day calendar.
FAQ
How big is the government and donor accelerator sector in Africa?
Substantial and growing: JICA’s Project NINJA alone has supported 824 startups across developing countries, including a ministry-run program in Uganda, and Entrepreneur Support Organizations in Africa have absorbed over $500 million in donor funding over the past decade — with governments now becoming direct program owners (1)(2)(3).
Why are accelerators measured on inputs instead of outcomes?
Because what gets contracted gets delivered, and programs are contracted on inputs they can guarantee (cohorts trained, demo days held) rather than outcomes they cannot (firms scaled, jobs created). So programs rationally optimize for the contracted activities, and there’s no contractual reason to track or publish what happens to graduates afterward.
Why does accelerator accountability matter more now?
Because ministries are becoming direct owners of these programs, which makes accelerator design a form of industrial policy — deployed with public money toward national development goals. A poorly designed, unaccountable program is now a poorly designed industrial intervention, raising the stakes from wasted training budgets to whether a major development lever works.
What should a government accelerator be accountable for?
For outcomes, not just activities: graduate survival at 24 months, revenue growth, follow-on capital raised, and jobs created — measured against a counterfactual where possible, with a meaningful share of program payment tied to verified results, and the outcomes published. This converts a busy program into an effective instrument of industrial policy.
How should governments fund accelerator programs?
By procuring acceleration as a service paid partly on verified outcomes rather than only on delivered activities — competitive selection of proven operators, payment partly contingent on graduate results, and public outcome reporting. This aligns the operator’s incentive with the result the government actually wants and dissolves the ceremony-optimization trap.
Related Reading
- If You Can’t Report Survival and Revenue, You’re Running an Event Calendar
- How Governments Should Buy Acceleration
- Does Startup Acceleration Actually Work? The Evidence and the Right Tail
- The ESO Funding Crisis: What the USAID Collapse Revealed
Sources and Evidence
- JICA — “Project NINJA: Empowering and Connecting Africa’s Entrepreneurs and Tech Start-ups” — Source for Project NINJA supporting 824 startups across developing countries.
- WeeTracker — “Uganda NINJA Acceleration Program startups” — Source for the ministry-run NINJA program in Uganda (10 startups from 130+ applicants).
- Pollinate Impact — “Incubators and Accelerators: Impact & Innovation” — Source for African ESOs absorbing over $500 million in donor funding over the past decade.
- StartupMap Africa — “Accelerators / Funding” — Source for the announcement-driven coverage and continental VC context, with scarce public post-program outcome evidence.
- MSME Africa — “Call for Applications: Accelerate Africa Startup Program 2026” — Illustrative of the relentlessly announcement-driven sector coverage.