The integrated economic, climate and infrastructure models, the AI report-generation pipeline and the interactive web applications that power AICTEM’s country diagnostics.
AICTEM’s four report series follow a 3 + 1 design: the three economy-wide diagnostics — Climate, Growth and Trade — run on a common core of computable general equilibrium (CGE) engines, while Infrastructure is powered by a separate spatial and power-system stack built entirely from open geodata. A shared data foundation keeps results comparable.
Two general-equilibrium engines and dedicated climate & carbon modules, sharing global input–output databases and national accounts.
The Climate-Dynamic CGE couples a multi-region model built on GTAP 11 with the Dynamic Integrated Climate–Economy (DICE) model to assess emissions, warming, climate damages and the economics of adaptation and mitigation.
The Structural-Dynamic CGE, solved with the PATH mixed-complementarity solver, projects output, productivity, sectoral reallocation and welfare — and, in its trade configuration, AfCFTA tariff scenarios, cross-border value chains and embodied emissions.
Connects economic activity, emissions pathways and climate damages, so policy scenarios compare on a consistent, physically-grounded basis.
Builds the land-use carbon accounts behind the Climate reports, adding emissions and land-use-change analysis to the CGE core.
Spatial economic and power-system models at grid-cell and settlement level, built entirely from open geodata and deep-research.
A spatial-equilibrium model on populated 0.25° grid cells, calibrated to gridded population, productivity and OSRM travel times, identifies the road upgrades with the largest national welfare and market-access gains.
A city-network general-equilibrium model (Fajgelbaum & Schaal, 2020) prices individual road works and allocates a fixed construction budget for maximum welfare — the engine behind our Transport Network Optimizer.
A nodal PyPSA dispatch and capacity-expansion model, on open PyPSA-Earth grid data, assesses supply adequacy, transmission bottlenecks and the cheapest mix of new generation to serve future load.
Settlement- and hexagon-level analytics on the Africa Infrastructure Database (Krantz, 2024) combine wealth estimates with asset distances, and causal ML estimates of power–digital complementarities.
The CGE stack builds on GTAP 11 and the EMERGING MRIO v2.2 input–output tables, integrated with African national accounts and environmentally-extended satellite accounts — globally consistent, locally relevant. The spatial stack rests on open geodata: the Africa Infrastructure Database (AID), OSRM routing, LocalGDP/population grids, PyPSA-Earth + OSM and deep-research power-system data.
AICTEM uses Claude Opus agents to turn the modelling stack into finished country diagnostics. A first set of agents runs a fixed, pre-defined suite of simulations on the models; a second set compiles the raw outputs, produces the figures and drafts the narrative. The workflow is encoded in a growing library of purpose-built skills that we standardise and continuously improve. We welcome feedback on the quality of the reports and suggestions for improvements.
Agents run a fixed, pre-defined set of scenarios and analyses on the models — the same suite for every country, so results are reproducible and comparable.
→Further agents ingest the simulation outputs, build the tables and figures, and assemble them into the structured sections of each report.
→Agents draft the narrative around those results, applying a shared, versioned library of skills that we refine with every reporting cycle.
Every figure reported is drawn directly from factual model simulation outputs based on real state-of-the-art economic and geospatial data — never generated or estimated by the large language model (LLM). The agents orchestrate the pipeline and write the prose; the numbers come from the models, keeping each report auditable and free of LLM hallucination.
Each modelling stream — Trade, Growth and Climate — is exposed as an interactive web application, so policymakers and researchers can run on-demand scenarios and watch results update live. These are built as Julia-native tools with the open-source Genie.jl framework, on the same engines and data that produce AICTEM’s reports. Alongside them, the Transport Network Optimizer, Infrastructure Risk Analyzer and Africa Monitor support the Infrastructure series with transport-network design, climate-risk screening and macroeconomic monitoring.
Trace goods, tariffs and value chains as they cross African borders under AfCFTA and trade-policy scenarios.
Project output, productivity and welfare as economies respond to shocks, investments and reform pathways.
Track warming pathways, climate damages and adaptation economics feeding AICTEM’s Climate reports.
Compute welfare-maximizing transport networks on user-supplied graphs, based on the quantitative spatial model of Fajgelbaum & Schaal (2020).
Automated climate and natural-hazard risk analysis of infrastructure assets against global hazard layers.
A macroeconomic information platform for Africa — leading indicators, data stories and custom visualizations, updated monthly.