July 2026 · Dr John Mary Matovu & Dr Sebastian Krantz
Africa enters the second quarter of this century as the youngest region on Earth. By 2050 one in four people alive will be African, the continent will hold the world’s largest labour force, and its cities will have absorbed several hundred million new residents. It is also the region least responsible for climate change yet among the most exposed to its effects. Whether that demographic wave becomes a dividend or a strain depends on choices made over the next two decades: how Africa produces, how it trades, how it builds, and how it adapts to a warming planet.
The African Institute for Climate, Trade and Economic Modelling (AICTEM) was founded on the conviction that those questions need to be answered together. We therefore publish cross-country comparable diagnostics along four themes (Climate, Trade, Growth and Infrastructure), and continental syntheses that analyze the results across countries, to help policymakers form a more holistic view of African economies and of the continent’s development. We view this as a first step toward more integrated assessments. This first post explains the motivation behind these diagnostics, and behind the AI-powered tools we are building to produce them.
Structural problems, seen from four angles
For half a century, Africa’s development debate has run in silos. Growth economists studied structural transformation and the persistence of low-productivity agriculture. Trade economists asked why fifty-four economies trade so little with one another. Climate and finance specialists measured vulnerability, adaptation and the gap between what governments have pledged and what they can pay for. Infrastructure was often studied apart from all three. Each community built its own models, its own datasets and its own blind spots.
The cost of that fragmentation is practical. A finance ministry commissions a growth diagnostic from one team, a trade-corridor study from another, and a climate-vulnerability assessment from a third. Each uses different country groupings, base years and sectoral definitions, so the results cannot be added up. A minister therefore cannot ask the obvious cross-cutting question: if I invest one dollar, where does it do the most for growth, trade and resilience at once?
Yet the transformations Africa must achieve are really one problem viewed from different angles. The factory that adds value to a cocoa bean is the factory that needs reliable, increasingly renewable power. The corridor that lets a Zambian firm sell into the Democratic Republic of the Congo is the corridor a flood can cut. Take agro-processing, the addition of manufacturing value to agricultural output. It is a growth question, because it is a bridge from a low-productivity sector to a higher-productivity one; a trade question, because it needs regional markets large enough to justify the plant; an infrastructure question, because it depends on power, roads and cold chains; and increasingly a climate question, because food systems both suffer from and contribute to warming. One intervention, four dimensions.
We do not pretend that AICTEM has cracked this. Building a single model that genuinely spans climate, trade, growth and infrastructure is a hard, open problem, and we have not yet found a convincing way to do it. What we can do is bring the four analyses closer together than they usually are. A multi-region input–output (MRIO) foundation maps how every sector in every economy buys from and sells to every other; it helps calibrate a dynamic computable general equilibrium (CGE) engine that turns policies into growth, welfare and distributional outcomes; structural gravity and quantitative-trade models diagnose actual and missing trade; and a climate module feeds its damages into the same CGE that prices growth interventions. Because these tools share data conventions, sectors and country coverage, the thematic diagnostics they produce can be read side by side, and compared across countries, rather than pieced together from incompatible studies. You can read more about the modelling stack on our Technology page.
Putting Africa’s infrastructure on the map
Infrastructure deserves a word of its own, because it is the hardest to see clearly. The state of an African country’s infrastructure system is hard to establish: the reach and condition of the transport network, the coverage and reliability of the power network, the spread of digital infrastructure and access. The relevant information sits scattered across ministries, utilities, operators and a patchwork of online sources, in incompatible formats and vintages. A planner who wants a clear picture of even a single country’s system, let alone a comparison across countries, has little to go on.
An integrated diagnosis, one that reads transport, power and digital together rather than as three unrelated sectors, is rarer still. Yet these systems reinforce one another: a firm needs a road to reach the market, power to run the plant, and connectivity to coordinate the two, and a gap in any one can strand investment in the others. The returns are often highest exactly where the gaps coincide.
The AICTEM infrastructure diagnosis adds a simple yet rigorous framework, built on open geospatial data and open modelling tools, that gives an integrated picture of a country’s infrastructure system: transport accessibility, power supply and access, and digital coverage, measured the same way everywhere. At this stage it runs largely separately from the MRIO-based macroeconomic and climate assessment. Even so, it is a distinctive addition to AICTEM’s surveillance, and coupling the two more tightly is a natural next step.
Continental syntheses
Because every country runs through the same framework, the diagnostic reports are broadly comparable across countries; that consistency is what fragmented, one-off studies cannot offer. The continental syntheses build on it. They are meta-analyses of the country diagnostics across Climate, Growth, Trade and Infrastructure, reading modelling results and narratives across countries to draw out patterns and shared challenges.
We have also built interactive web applications on the same engines and data that produce the reports, so you can run custom simulations and explore the data. For infrastructure, you can upload and analyze your own data in the Transport Network Optimizer or the Infrastructure Risk Analyzer. The Kiel Institute Africa Monitor adds macroeconomic and development data for the whole continent, covering economic, social and environmental dimensions, with fast API access.
A bet on powerful AI, guided by human expertise
To produce these diagnostics at quality and scale, AICTEM is building a team of AI economists, statisticians, modellers and geospatial experts. We are stunned by the ability of frontier AI models to simulate economic models, process data and write compelling narratives around the results. Recent IMF research using models up to GPT-5 assessed how well LLMs analyze the macrofinancial issues covered in IMF Article IV staff reports, and found 70–80% accuracy relative to IMF experts; the current generation of models (GPT-5.6, Opus 4.8, Fable 5) is up to 30% better than GPT-5. But this capability needs to be guided by human expertise. We are therefore developing a set of skills that steers AI agents through simulations and report generation, and ensures the numbers are grounded in reality rather than being the product of a model alone.
Where to start
This blog will share notes from building the models, guides to reading a diagnostic, and findings as new country and continental work is published. The best place to begin is the evidence itself:
— Dr John Mary Matovu & Dr Sebastian Krantz, AICTEM