Bilateral aid to Africa fell by nearly a quarter in 2025, the largest annual decline in the history of official development assistance. Meanwhile, sovereign debt interest payments now consume on average 27% of government revenues across the continent, up from 19% in 2019.
The pressure to fund development from within has never been greater. But meeting it requires African governments to understand their own economies with precision: which tax policies work, which incentives serve their purpose, how fiscal decisions distribute their consequences.
Administrative tax data, the anonymised filings, returns and transaction records generated through the tax system that African revenue authorities already hold, is one of the most powerful tools for answering those questions. South Africa, Uganda and Zambia have built the means to use it, and what they are finding is shaping how they govern.
Each has established a secure research data lab where researchers work with anonymised tax records under strict confidentiality protocols. All three were developed with support from the United Nations University World Institute for Development Economics Research (UNU-WIDER). It provides technical expertise and facilitates knowledge-sharing across countries, while ensuring that the data, the research agenda and the findings are retained by the institutions that use them.
We have been part of that support, from the establishment of the labs and making raw administrative data research-ready, to facilitating partnerships and ensuring findings reach the people placed to act on them.
What these data labs are producing is evidence that has fed into solutions including tax policy reform, budget decisions, labour market programmes and social protection. It can deepen how governments understand the economies they are responsible for, and the people within them.
From research findings to decision
In South Africa, the National Treasury Secure Data Facility, the cornerstone of the Southern Africa Towards Inclusive Economic Development programme, has been doing this work for over a decade. The cumulative impact reflects that longevity. The data has repeatedly illuminated how the economy works, often differently from how policy expected.
Findings have shaped a number of decisions. For example:
-
Research revealed that the corporate tax system was quietly favouring debt over equity financing. This was nudging firms to borrow more than they otherwise would, making companies and the economy more fragile in downturns. This informed corporate tax restructuring in Budget 2020.
-
Analysis of the Employment Tax Incentive, a wage subsidy for young workers in a country where nearly 60% cannot find work, revealed a more complicated picture of impact than its designers had anticipated. This informed a decision to expand the subsidy during the COVID-19 pandemic.
-
Research which describes how much economic activity increases when the government raises spending or cuts taxes emphasised the importance of growth-oriented investments such as infrastructure, education, or public health.
-
Analysis of behavioural patterns at tax thresholds provided evidence for designing fairer policies that reduce avoidance and broaden the tax base.
Similarly, in Kampala, Uganda’s commercial capital, adjustments were made to policies based on the use of the data.
-
Research conducted through the Uganda Revenue Authority’s secure data lab found that fewer than 15% of firms were paying both national corporate income tax and the local trade licence fee. The gap had existed for years. It had simply never been quantifiable before.
-
Separate research revealed that corporate tax incentives were costing approximately US$42 million in forgone revenue. More than half of the benefiting firms likely remained profitable at the full statutory rate of 30%. An incentive regime designed to attract investment was, in measurable terms, more generous than the investment required.
For policymakers, findings like these do not simply describe a problem. They reframe it. The questions have shifted from how to collect more revenue to where the system is working against itself, and what can be changed.
Some examples from Zambia:
-
Tax gap research estimated the country’s compliance gap at between 47% and 56%. This helped quantify, for the first time, where revenue was being lost and how audit resources could be better targeted. The findings fed directly into the 2026 budget. The government’s audit strategy was reshaped. And the findings informed the deliberations of the Tax Policy Review Committee.
-
Separately, research on VAT administration uncovered a structural inefficiency: large firms generating simultaneous liabilities and credits on the same accounts, a circular flow consuming administrative effort without producing revenue. This was invisible without transaction-level data. When it was identified, the problem was corrected.
From a number to a life
Statistics, even compelling ones, exist in abstraction. What brings them to reality is the chain of consequences, from research finding, to decision, to the life that decision shapes.
Zambia’s domestic revenue contribution to the national budget is rising from 55.7% in 2020 to a planned 73.1% in 2026. This reflects fiscal decisions made with increasing precision and confidence, in a country that now has the tools to interrogate its own tax system rather than rely on external assessments of it.
That stronger revenue base has also made possible a free education initiative that has brought approximately 2.3 million children who were previously out of school back into classrooms. A research lab did not put those children there. Evidence-informed fiscal policy did.
South Africa’s record adds another dimension. Microsimulation modelling showed that the Social Relief of Distress grant was significantly reducing poverty just when a decision about its future needed to be made. The evidence changed what that decision could credibly be. The distinction matters, because it clarifies what these labs are actually for: not the production of research, but the conditions that allow governments to govern better.
The case in practice
Africa’s financing challenge will not be resolved from outside. The most durable path runs through domestic revenue systems that are efficient, fair, and sharpened by what the evidence shows. That requires a particular kind of analytical honesty: the willingness to examine fiscal systems rigorously and act on what the data reveals.
Development economics has long argued that evidence-based policymaking produces better outcomes. South Africa, Uganda and Zambia are making that argument in practice. They are doing this through what they have built, what they have found, and what they have chosen to do because of it.
At a moment when external financing is contracting and debt service is rising, the quality of fiscal decisions is not an academic question. It is the difference between governments that can see their own economies clearly enough to act, and governments that cannot.
The data is already there. The model has been proven. What remains is the will to use it.
