Frequently Asked Questions

What is a narrative fiscal consolidation?

A narrative fiscal consolidation is an instance of planned discretionary fiscal tightening identified through historical narratives such as budget speeches, policy statements, and IMF reports. Unlike mechanical measures that rely purely on fiscal outturns, the narrative approach aims to isolate policy decisions that are not driven by contemporaneous economic conditions.

Which countries are included?

The initial release includes 14 Sub‑Saharan African countries. See the country coverage table on the home page for details. We plan to expand to additional countries in future releases; see the roadmap in governance/roadmap.md.

How are the sizes of consolidation episodes measured?

The tax and spend variables record the projected fiscal impact of the announced measures, expressed as a percentage of GDP. These projections are taken from official documents (budget reports, medium‑term frameworks) or IMF staff reports. The total variable is the sum of tax and spend.

Which file should I use for replication?

Use dataset.dta at the root of the repository. It is the authoritative paper replication dataset for the two IMF working papers listed on the home page. It is preserved as a fixed version; future annual updates of the narrative dataset will be released separately.

Can I download the data programmatically?

Yes. Use the raw GitHub URL for dataset.dta:

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

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

* Stata
use "https://github.com/Hanomics/fiscal_narrative_dataset/raw/main/dataset.dta", clear

How can I contribute a new country or episode?

Please read CONTRIBUTING.md and governance/submission_guide.md before submitting data. In short, you should fork the repository, add your data following the schema, document your sources, and open a pull request. We welcome contributions that adhere to the narrative identification methodology.

Who maintains this project?

The project is maintained by a small team of researchers. Contact information is available in SECURITY.md and the repository description.