Shine a (night)light: decentralization and economic development in Burkina Faso

Decentralization, championed by international institutions, has been one of the most prominent public sector reforms of the last decades, particularly in sub-Saharan Africa. To date, few studies propose a quasi-experimental evaluation of its capacity to contribute to local development. We exploit the phase-in of decentralization at the commune level in Burkina Faso. We use satellite information on night-time light density as a proxy for local development levels, which has the advantage of being measured and comparable over time and space. The communes that were decentralized first can be compared to the others after the reform relative to the pre-reform situation. The difference-in-difference approach includes commune fixed effects and inverse propensity score reweighting to account for time-varying differences across communes. We find a positive impact of decentralization on the night-light intensity trends of the early-decentralized communes. This is supported by alternative measures (remote sensing of built-up settlements and a welfare index), which shows the possibly broader scope of decentralization gains. We show that decentralization did not lift all boats: only the communes with the ability to generate own-source revenues benefited from effective decentralization.

Context and Objective

Decentralization has, for several decades, been one of the institutional reforms most strongly promoted by international institutions, particularly in sub-Saharan Africa. Delegating power to local authorities is presented as a means to improve governance by reducing information asymmetries between decision-makers and citizens, while introducing greater competition and checks and balances against central governments that have lost legitimacy. This reform is expected to enhance the accountability of local governments, improve the provision of public goods and services, and ultimately reduce poverty (Bardhan, 2002). However, few rigorous empirical studies have provided a causal evaluation of its capacity to foster local development.

Burkina Faso offers an ideal setting to address this gap in the literature, enabling a quasi-experimental study through the phased implementation of its decentralization reform in three stages: 33 municipalities were decentralized in 1995, 16 in 2000, and the remaining 321 in 2005. This staggered rollout provides an opportunity to establish comparison groups and evaluate the causal effects of the reform.

Methodology

We use a database covering the period 1992–2010 and rely on satellite data on night-time light intensity as the outcome indicator. The use of night-time light data as a proxy for local economic development captures local economic trends in contexts where conventional economic indicators are scarce (Henderson et al., 2012). The analysis is complemented by alternative measures, such as the degree of urbanization obtained through remote sensing of built-up areas and a household well-being index derived from Demographic and Health Surveys.

Our difference-in-differences model—including fixed effects at the municipality and time levels, as well as control variables related to climatic conditions and other local characteristics—neutralizes biases stemming from time-invariant differences between treatment and control groups while capturing common temporal variations. We also apply propensity score reweighting to address biases linked to the non-random selection of municipalities in the early phases of the reform, allowing for comparisons among more homogeneous subgroups of early and late decentralized municipalities (Abadie, 2005). The propensity score is estimated using variables related to demographics, geography, and the economic resources of the municipalities. Finally, we explore the heterogeneity of effects to understand the mechanisms at play and identify the characteristics of the municipalities that benefited the most from the reform.

Key Findings and Conclusion

Our results show that decentralization had a significant positive impact on local economic development. Municipalities decentralized in 1995 experienced an increase in night-time light intensity of 14% to 18% compared to those decentralized later. These effects are robust to various heterogeneity tests and parallel trend checks. Data on built-up areas and household well-being indices also confirm these findings.

The ability to generate and collect own-source revenues through taxation emerges as a key factor—not only in explaining which municipalities were decentralized first but also in understanding the benefits of decentralization and, among early decentralized municipalities, which ones performed better. This underscores the importance of local resources for the effective implementation of decentralization reforms (Caldeira et al., 2012).

These heterogeneous effects also suggest that decentralization did not benefit all municipalities equally. The findings align with the literature on the unequal effects of decentralization, which can contribute to a permanent increase in inequalities by favoring already advantaged populations (Prud’homme, 1995; Galiani et al., 2008).

Référence de l’article

Olivier B. Bargain, Rose Camille Vincent, Emilie Caldeira, Shine a (night)light: Decentralization and economic development in Burkina Faso, World Development, Volume 187, 2025, 106851, ISSN 0305-750X.

Références

1.    Abadie, A. (2005). Semiparametric difference-in-differences estimators. Review of Economic Studies, 72(1), 1–19. 
2.    Bardhan, P. (2002). Decentralization of Governance and Development. Journal of Economic Perspectives, 16(4), 185-205.
3.    Caldeira, E., Foucault, M., Rota-Graziosi, G. (2012). Does Decentralization Facilitate Access to Poverty-Related Services? Evidence from Benin. World Development, 40(9), 1737-1751.
4.    Callaway, B., & Sant'Anna, P. H. C. (2021). "Difference-in-differences with multiple time periods." Journal of Econometrics, 225(2), 200-230.
5.    Galiani, S., Gertler, P., Schargrodsky, E. (2008). School Decentralization: Helping the Good Get Better, but Leaving the Poor Behind. Journal of Public Economics, 92(10-11), 2106-2120.
6.    Henderson, J. V., Storeygard, A., Weil, D. N. (2012). Measuring Economic Growth from Outer Space. The American Economic Review, 102(2), 994-1028.
7.    Prud’homme, R. (1995). The dangers of decentralization. World Bank Research Observer, 10(2), 201–220.

Photo © IRD - Patrice Brehmer