electricsheepafrica/african-decentralization-fiscal-transfers

---
license: cc-by-4.0
language:
- en
tags:
- governance
- decentralization
- fiscal-transfers
- sub-saharan-africa
- synthetic
- lmic
- local-government
- public-finance
- intergovernmental-relations
- fiscal-federalism
pretty_name: African Decentralization and Fiscal Transfers
size_categories:
- 10K<n<100K
configs:
- config_name: baseline
data_files:
- split: train
path: "data/african_decentralization_baseline.csv"
- config_name: fiscal_devolution
data_files:
- split: train
path: "data/african_decentralization_fiscal_devolution.csv"
- config_name: recentralization
data_files:
- split: train
path: "data/african_decentralization_recentralization.csv"
---

# African Decentralization and Fiscal Transfers

Synthetic dataset modeling fiscal decentralization patterns across 12 Sub-Saharan African countries under three counterfactual scenarios. Parameters are calibrated against published literature on intergovernmental fiscal relations in the region.

## Dataset Description

### Overview

This dataset contains **360,000 synthetic records** (120,000 per scenario) covering fiscal decentralization indicators for 12 SSA countries across 2000-2024. Each record represents a country-year observation with ~20 variables capturing the structure, autonomy, capacity, and outcomes of subnational fiscal systems.

### Countries

The 12 countries span a spectrum of decentralization maturity:

| Country | Gov Tiers | Subnational Units | Decentralization Type |
|---------|-----------|-------------------|----------------------|
| Kenya | 2 | 47 counties | Devolved (2010 Constitution) |
| South Africa | 3 | 9 provinces + 278 municipalities | Federal-like (post-1996) |
| Nigeria | 3 | 36 states + 774 LGAs | Federal |
| Ethiopia | 4 | 12 regions + city admins | Federal (ethnic-based) |
| Rwanda | 3 | 30 districts | Devolved (performance contracts) |
| Uganda | 2 | 135 districts | Deconcentrated |
| Tanzania | 2 | 185 LGAs | Limited devolution |
| Ghana | 2 | 260 MMDAs | Limited devolution |
| Senegal | 2 | 170 communes/regions | Limited devolution |
| Mozambique | 2 | 159 municipalities/districts | Limited devolution |
| Côte d'Ivoire | 2 | 113 regions/departments | Limited devolution |
| DRC | 3 | 26 provinces | Limited (constitutional 40% target) |

### Scenarios

- **baseline**: Status quo fiscal decentralization patterns calibrated to published data
- **fiscal_devolution**: Enhanced devolution with increased own-source revenue, autonomy, and capacity (+8pp revenue, +12pp autonomy, +10pp capacity)
- **recentralization**: Fiscal recentralization with reduced subnational autonomy (-5pp revenue, -10pp autonomy, -5pp capacity)

### Variables (25 columns)

| Variable | Description | Range |
|----------|-------------|-------|
| `country` | Country name | 12 SSA countries |
| `year` | Observation year | 2000-2024 |
| `scenario` | Counterfactual scenario | baseline/fiscal_devolution/recentralization |
| `tiers_of_government` | Number of government tiers | 2-4 |
| `subnational_units` | Number of primary subnational units | varies |
| `subnational_revenue_pct_total` | Subnational revenue as % of total government revenue | 1-75% |
| `own_source_revenue_pct` | Own-source revenue as % of subnational revenue | 0.5-80% |
| `intergovernmental_transfers_pct_gdp` | Intergovernmental transfers as % of GDP | 0.2-15% |
| `unconditional_transfers_pct` | Unconditional (block) transfers as % of total transfers | 5-95% |
| `conditional_transfers_pct` | Conditional (earmarked) transfers as % of total transfers | 5-95% |
| `subnational_expenditure_pct_total` | Subnational expenditure as % of total government expenditure | 1-80% |
| `expenditure_autonomy_score` | Expenditure autonomy index (0=none, 100=full) | 5-95 |
| `revenue_autonomy_score` | Revenue autonomy index (0=none, 100=full) | 3-90 |
| `local_government_capacity_score` | Local government institutional capacity index (0-100) | 5-90 |
| `vertical_fiscal_imbalance_pct` | % of subnational spending financed by transfers | 10-99% |
| `transfer_dependency_ratio` | Transfer dependency index | 10-98 |
| `fiscal_decentralization_index` | Composite fiscal decentralization index | 0.05-0.85 |
| `effective_devolution_index` | Effective devolution composite index | 0.05-0.85 |
| `horizontal_equalization_index` | Horizontal fiscal equalization index (0-100) | 5-95 |
| `performance_based_transfer_share` | % of transfers allocated on performance basis | 0-80% |
| `own_source_tax_effort_index` | Tax effort index relative to capacity | 5-120 |
| `borrowing_autonomy_score` | Subnational borrowing autonomy index (0-100) | 2-80 |
| `capital_expenditure_pct` | Capital expenditure as % of subnational expenditure | 5-60% |
| `decentralization_class` | Decentralization classification | 5 classes |
| `constitutional_transfer_floor_pct` | Constitutional minimum transfer floor (%) | 5-42% |

### Decentralization Classification

- `highly_centralized`: FD index < 0.20 (e.g., DRC, Mozambique)
- `moderately_centralized`: 0.20 ≤ FD index < 0.35 (e.g., Ghana, Tanzania, Rwanda)
- `moderately_decentralized`: 0.35 ≤ FD index < 0.50 (e.g., Kenya, Nigeria)
- `highly_decentralized`: 0.50 ≤ FD index < 0.65 (e.g., South Africa, Ethiopia)
- `fully_decentralized`: FD index ≥ 0.65

## Parameter Calibration

Country parameters were calibrated against published literature:

- **Kenya**: Constitutional 15% floor, transfers ~87% of county revenue, own-source ~10% (Rutto et al. 2022; Kenya Controller of Budget Reports). Post-2013 devolution of 47 counties.
- **South Africa**: Three spheres, provinces receive ~43% of revenue, local government ~9%, provinces raise only 1.3% own revenue (National Treasury, Division of Revenue data). Local government ~79% self-financed.
- **Nigeria**: Three tiers, Federation Account allocation formula (RMAFC). States ~26% of federation account, LGAs ~20%. Oil revenue dominant.
- **Ethiopia**: Federal system with 4 tiers, ~30% own-source revenue at subnational level, one of the most decentralized in SSA (IMF 2025 Selected Issues).
- **Rwanda**: Districts are primary units, 80% of transfers earmarked, 16% own revenue (MINALOC data). Performance contracts (imihigo) system.
- **Uganda**: Deconcentrated rather than devolved, ~5% own-source revenue (IMF 2018), conditional grants dominant.
- **Tanzania**: Limited devolution, 185 LGAs, transfers can enhance local revenue collection (Masaki 2018, World Development).
- **DRC**: Constitutional 40% transfer target largely unmet, very low actual subnational revenue share.

Cross-country pattern: Most SSA countries retain <15% of public revenue at subnational level. Only Nigeria, South Africa, and Ethiopia approach 50% of general government spending at subnational level. East African countries (Kenya, Rwanda, Tanzania, Uganda) are around 15-20% (Hobdari et al. 2018, IMF).

## Diagnostic Plots

The 8-panel diagnostic figure (`data/diagnostic_plots.png`) includes:
- A: FD Index distribution by scenario
- B: Subnational revenue share by country (baseline)
- C: Own-source revenue vs capacity scatter
- D: Transfer composition (unconditional vs conditional)
- E: FD Index boxplot by scenario
- F: Vertical fiscal imbalance by country
- G: Capacity score over time
- H: Decentralization class distribution

## Usage

```python
import pandas as pd

# Load specific scenario
baseline = pd.read_csv("data/african_decentralization_baseline.csv")

# Load all scenarios combined
df = pd.read_csv("data/african_decentralization_all_scenarios.csv")

# Filter to specific country
kenya = df[df["country"] == "Kenya"]

# Compare scenarios for a variable
df.groupby("scenario")["fiscal_decentralization_index"].describe()
```

## Generation

```bash
pip install -r requirements.txt
python generate_dataset.py
python validate_dataset.py
```

## Limitations

- Synthetic data calibrated to literature means; does not capture country-specific political economy dynamics
- Scenarios are stylized counterfactuals, not predictions
- Capacity and autonomy scores are composite indices without standardized measurement across countries
- Time trends assume gradual linear capacity building; real reforms are episodic
- Correlation structure is assumed, not empirically derived from panel data

## References

1. Hobdari, N., Nguyen, V., Dell'Erba, S., & Ruggiero, E. (2018). Lessons for Effective Fiscal Decentralization in Sub-Saharan Africa. IMF Departmental Paper No. 2018/009.
2. Rutto, A., Minja, D., & Kosimbei, G. (2022). Intergovernmental Fiscal Transfers and Decentralization Initiatives in Sub-Saharan Africa: A Case Study of SNGs in Kenya. SDMIMD Journal of Management.
3. Masaki, T. (2018). The impact of intergovernmental transfers on local revenue generation in Sub-Saharan Africa. World Development, 106, 173-186.
4. IMF (2025). Ethiopia - Fiscal Federalism: Fiscal Policy Considerations for the Medium Term. Selected Issues Paper No. 2025/108.
5. World Bank Decentralization Net. Intergovernmental Fiscal Transfers Primer.
6. Fjeldstad, O.-H. & Chambas, G. Local Government Taxation in Sub-Saharan Africa. Chr. Michelsen Institute.
7. IMF (2025). Tax Expenditures in Sub-Saharan Africa. Departmental Paper.
8. South Africa National Treasury. Division of Revenue Bills (various years).
9. Nigeria Revenue Mobilization Allocation and Fiscal Commission (RMAFC).
10. Rwanda Ministry of Local Government (MINALOC). Decentralization reports.

## Citation

```bibtex
@dataset{african_decentralization_fiscal_transfers_2026,
title={African Decentralization and Fiscal Transfers},
year={2026},
note={Synthetic dataset calibrated to literature on sub-Saharan African fiscal decentralization},
license={cc-by-4.0}
}
```