Socioeconomic Drivers of Household Electricity Access in Rwanda

Evidence from EICV 7

---

Authors

Name	Role	Institution
Jean Pierre NIYOMUGABO	First Author	University of Rwanda, CBE
Dr. Jules NGANGO	Supervisor & Co-author	University of Rwanda, CBE

Institution: College of Business and Economics (CBE), University of Rwanda, Huye Campus, Rwanda

Target Journal: Scientific African (Elsevier)

Status: Under preparation — May 2026

---

Abstract

Rwanda's household electricity access expanded from 34.4% in 2016/17 to 72.0% in 2023/24. Despite this progress, significant socioeconomic and spatial disparities persist. This study examines the household-level socioeconomic drivers of electricity access using EICV 7 (2023/24) microdata (n = 15,054 households) across all 30 districts of Rwanda. Binary logistic regression with average marginal effects is estimated as the baseline model, with probit and linear probability model (LPM) robustness checks. GIS spatial mapping is employed to visualize geographic disparities across provinces and districts.

---

Key Statistics

Metric	Value
Dataset	EICV 7 (2023/24)
Sample size	15,054 households
National electricity access	72.0%
Urban access	88.5%
Rural access	66.1%
Highest province	City of Kigali (92.3%)
Lowest province	Southern Province (65.0%)
Richest quintile (Q5)	92.2%
Poorest quintile (Q1)	54.1%
AUC-ROC	0.811
McFadden R²	0.2271
Variables in model	14 (all VIF < 3)
Consistent results	14/14 ✅

---

Key Findings

Top Determinants (Logit AME)

Variable	Odds Ratio	Marginal Effect
University education	9.969	+34.8 pp
Secondary education	2.911	+16.2 pp
Floor material	1.800	+8.9 pp
Urban location	1.821	+9.1 pp
Ownership status	1.505	+6.2 pp
Poverty quintile	1.394	+5.0 pp
Primary education	1.390	+5.0 pp
Number of dependents	0.881	-1.9 pp

Spatial Findings

District	Access Rate
Nyarugenge (highest)	94.6%
Kicukiro	94.0%
Gasabo	88.3%
Gisagara (lowest)	51.9%
Gap	42.7 pp

---

Project Structure

ScientificAfrican/
├── main.tex                        ← Master LaTeX file
├── README.md                       ← This file
├── setup.bat                       ← Project setup script
├── .gitignore                      ← LaTeX aux files ignored
│
├── Analysis/
│   └── regression_analysis.ipynb  ← Complete Python analysis
│
├── sections/
│   ├── abstract.tex                ← ~250 words
│   ├── introduction.tex            ← Section 1
│   ├── literature_review.tex       ← Section 2
│   ├── methodology.tex             ← Section 3
│   ├── results.tex                 ← Section 4
│   └── conclusion.tex              ← Section 5
│
├── figures/
│   ├── fig1_descriptive.png        ← 2x2 descriptive panel
│   ├── fig2_marginal_effects.png   ← Forest plot AME
│   ├── fig3_diagnostics.png        ← ROC + Confusion matrix
│   └── fig4_gis_maps.png           ← Province + District maps
│
├── tables/
│   ├── table_regression.tex        ← Logit + Probit + LPM
│   ├── table_ame.tex               ← Marginal effects
│   └── table_diagnostics.tex       ← Model fit statistics
│
└── references/
    └── references.bib              ← 28 BibTeX entries

---

Methods

Econometric Models

Model	Purpose	Status
Binary Logistic Regression	Baseline model	✅
Probit	Robustness check	✅
Linear Probability Model (LPM)	Robustness check	✅
Average Marginal Effects (AME)	Interpretation	✅
VIF Diagnostics	Multicollinearity check	✅
Hosmer-Lemeshow Test	Goodness of fit	✅
ROC-AUC	Discrimination ability	✅

Spatial Analysis

Tool	Purpose
geopandas	GIS mapping
matplotlib	Visualization
RWA_adm1.shp	Province boundaries
RWA_adm2.shp	District boundaries

---

How to Compile

# Full compile sequence (XeLaTeX + BibTeX)
xelatex main
bibtex main
xelatex main
xelatex main

VS Code Settings (settings.json)

{
  "latex-workshop.latex.recipe.default":
      "xelatex -> bibtex -> xelatex x2"
}

---

Python Environment

# Required packages
pip install pandas numpy matplotlib seaborn
pip install statsmodels scikit-learn
pip install geopandas folium scipy

Python Version

• Python 3.13.6

---

Data Sources

Dataset	Source	Year
EICV 7 microdata	National Institute of Statistics of Rwanda (NISR)	2025
Rwanda shapefiles	GADM administrative boundaries	2022

Note: The EICV 7 microdata is not publicly available. Access can be requested from NISR Rwanda: https://www.statistics.gov.rw

---

References

Key references used in this study:

• Blimpo & Cosgrove-Davies (2020). World Development, 133, 105002.
• Peters et al. (2025). Nature Communications, 16, 10438.
• Mvondo et al. (2023). Energy Policy, 176, 113499.
• Nzabarinda et al. (2021). IJERPH, 18(24), 13207.
• NISR (2025). EICV 7 Main Indicators Report. Kigali, Rwanda.
• Wooldridge (2010). Econometric Analysis of Cross Section and Panel Data. MIT Press.

Full bibliography: references/references.bib (28 entries)

---

License

This repository contains academic research files. All rights reserved © Jean Pierre NIYOMUGABO & Dr. Jules NGANGO, University of Rwanda, 2026.

---

Contact

Jean Pierre NIYOMUGABO Registration: 222008736 College of Business and Economics University of Rwanda, Huye Campus Rwanda