A Narrative Fiscal Consolidation Dataset for Sub‑Saharan Africa

A transparent, narrative-based dataset of discretionary fiscal consolidation episodes in Sub-Saharan Africa, designed for policy analysis, empirical research, and replication.

What it is

Country–year fiscal consolidation episodes identified from narrative evidence in IMF staff reports.

Why it matters

Isolates discretionary fiscal actions more cleanly than mechanical balance-based measures.

Who it is for

Researchers, policy economists, central banks, IFIs, and students of fiscal policy.

⬇ Download replication dataset (.dta) Open Quotes Explorer

Reproducibility-first design

This release is the fixed paper replication dataset associated with the IMF working papers. Future annual updates of the narrative dataset will be released separately and documented through versioned releases.

Fixed replication dataset: dataset.dta
Transparent narrative evidence through the Quotes Explorer
Future annual updates kept separate from the paper replication release

Typical applications

Estimate fiscal multipliers using narrative identification.
Compare tax-based and spending-based consolidations.
Study timing and state dependence of fiscal adjustment.
Trace country–year episodes back to source text evidence.

Narrative transparency

Every coded episode can be traced back to the underlying text evidence. Each record links a country–year action to the IMF staff report that documents it, together with the dataset’s narrative motivation summary.

Kenya · 2011 · Tax-side · Included · IMF Country Report No. 11/48

“A fiscal consolidation from the tax side amounting to 1.1 percent of GDP, driven by increases in income tax and VAT, and motivated by the objective of reducing the fiscal deficit and public debt and enhancing fiscal sustainability.”

Explore narrative evidence →

Headline result

Estimated effect of a 1 percent of GDP fiscal consolidation on real GDP.

Why narrative identification matters

Narrative identification isolates discretionary fiscal policy actions from cyclically driven movements in deficits. Compared with identification based on the cyclically adjusted primary balance (CAPB) or on forecast errors, narrative shocks imply larger and more persistent output effects.

Narrative shocks vs. CAPB vs. forecast-error identification — Narrative shocks imply larger and more persistent output effects than CAPB or forecast-error approaches.

Dataset coverage

The replication dataset covers 16 Sub-Saharan African countries from 1990 onwards. The timeline below summarises the narrative fiscal consolidation episodes identified for each country.

Timeline of narrative fiscal consolidation episodes by country.

Key heterogeneity

Tax-based vs. spending-based fiscal consolidation — Spending-based consolidations are more contractionary than tax-based consolidations.

Output effects across booms and slumps — Output effects differ across booms and slumps.

Additional results — external adjustment channels and external financing conditions

Fiscal consolidation also operates through trade and exchange-rate channels, and its effects depend on the availability of external financing.

Response of imports (BoP) — Imports (BoP).

Response of exports (BoP) — Exports (BoP).

Response of the current account balance — Current account balance (share of GDP).

Response of the real effective exchange rate — Real effective exchange rate.

Output effects under different ODA conditions — Output effects of fiscal consolidation differ with external financing conditions.

Associated papers

The first paper documents the construction of the dataset. The second paper applies the dataset to estimate fiscal multipliers in Sub-Saharan Africa.

A Narrative Fiscal Consolidation Dataset for Sub-Saharan Africa — DOI: 10.5089/9798229034661.001
The Fiscal Multipliers Narrative of Sub-Saharan Africa — DOI: 10.5089/9798229037792.001

Get started in 60 seconds

Direct download: dataset.dta

# R
library(haven)
df <- read_dta("https://github.com/Hanomics/fiscal_narrative_dataset/raw/main/dataset.dta")

# Python
import pandas as pd
df = pd.read_stata("https://github.com/Hanomics/fiscal_narrative_dataset/raw/main/dataset.dta")

View country coverage table

| ISO3 | Country | First year | Last year | Episodes | | --- | --- | --- | --- | --- | | AGO | Angola | 1990 | 2024 | 0 | | CIV | Côte d'Ivoire | 1990 | 2024 | 8 | | CMR | Cameroon | 1990 | 2024 | 9 | | COD | Democratic Republic of the Congo | 1990 | 2024 | 0 | | ETH | Ethiopia | 1990 | 2024 | 5 | | GHA | Ghana | 1990 | 2024 | 8 | | KEN | Kenya | 1990 | 2024 | 6 | | MOZ | Mozambique | 1990 | 2024 | 3 | | MUS | Mauritius | 1990 | 2024 | 3 | | NGA | Nigeria | 1990 | 2024 | 1 | | RWA | Rwanda | 1990 | 2024 | 7 | | SEN | Senegal | 1990 | 2024 | 6 | | TZA | Tanzania | 1990 | 2024 | 9 | | UGA | Uganda | 1990 | 2024 | 12 | | ZAF | South Africa | 1990 | 2024 | 6 | | ZMB | Zambia | 1990 | 2024 | 2 |

How to cite

APA

Abdel-Latif, H., Bechchani, K., David, A., & Lemaire, T. (2025). A Narrative Fiscal Consolidation Dataset for Sub-Saharan Africa. International Monetary Fund Working Paper. https://doi.org/10.5089/9798229034661.001

BibTeX

@techreport{abdel_latif_2025_narrative_ssa,
  author      = {Abdel-Latif, Hany and Bechchani, Khalil and David, Antonio and Lemaire, Thibault},
  title       = {A Narrative Fiscal Consolidation Dataset for Sub-Saharan Africa},
  year        = {2025},
  institution = {International Monetary Fund},
  type        = {IMF Working Paper},
  doi         = {10.5089/9798229034661.001},
  url         = {https://github.com/Hanomics/fiscal_narrative_dataset}
}