Journal Design Science Quartz
African Rural Development Studies (Interdisciplinary - | 14 June 2022

Methodological Diagnostics of Municipal Water Systems in Rwanda

A Systematic Review of the Difference-in-Differences Model for Yield Optimisation
M, a, r, i, e, C, h, a, n, t, a, l, U, w, a, s, e, ,, T, h, a, r, c, i, s, s, e, N, i, y, o, n, z, i, m, a, ,, V, a, l, e, n, t, i, n, e, M, u, k, a, m, a, n, a, ,, J, e, a, n, d, e, D, i, e, u, U, w, i, m, a, n, a
difference-in-differenceswater managementimpact evaluationRwanda
Limited corpus of studies identified in systematic review
Frequent failure to test parallel trends assumption in DiD applications
Treatment effects often lack confidence intervals or robust errors
Methodological practice remains nascent despite theoretical suitability

Abstract

{ "background": "Municipal water systems are critical for agricultural productivity and rural development. In Rwanda, optimising the yield of these systems is a policy priority, yet rigorous methodological evaluations of impact assessment models, particularly quasi-experimental designs, are lacking in the literature.", "purpose and objectives": "This systematic review aims to critically evaluate the application and methodological rigour of the difference-in-differences (DiD) model for measuring yield improvements in Rwandan municipal water systems within the agricultural sector.", "methodology": "A systematic search of multiple academic databases was conducted following a pre-registered protocol. Eligible studies were those employing a DiD design to assess water system interventions on agricultural yield. Studies were appraised for methodological quality, focusing on model specification, identification assumptions, and robustness checks. The canonical two-way fixed effects model is $Y{it} = \\alpha + \\beta (Treati \\times Postt) + \\gammai + \\deltat + \\epsilon{it}$.", "findings": "The review identified a limited corpus of studies. A dominant theme was the frequent failure to test the parallel trends assumption, with fewer than 30% of studies employing placebo tests or event-study frameworks for validation. Where reported, treatment effect estimates were often presented without confidence intervals or robust standard errors, limiting inferential certainty.", "conclusion": "The application of the DiD model in this specific context is methodologically nascent. While the design is theoretically well-suited, common shortcomings in empirical practice undermine the reliability of reported yield optimisation impacts.", "recommendations": "Future research must prioritise rigorous pre-intervention trend analysis and comprehensive robustness checks. Policymakers should require such diagnostic evidence when evaluating programme efficacy. Capacity building in advanced quasi-experimental methods for local researchers is essential.", "key words": "difference-in-differences, water resources, agricultural productivity, impact evaluation, quasi-experimental design, public utilities", "contribution statement": "This review provides the first dedicated methodological audit of DiD applications in