Mini-grids are key for improving electricity access in poor rural areas, but the high collection cost of field data is an obstacle to their developmental impact evaluation. This paper discusses the use of night-time light (NTL) to assess and monitor impacts of rural mini-grids. A positive impact on NTL reveals improvement in electricity access, but it does not necessarily reveal an impact on electric power consumption (EPC), which is key in any theory of change associated with impact assessment of electrification projects. Many NTL-EPC descriptive regressions have been estimated, leading the way to possibly mistaken interpretation of impacts on NTL as impacts on EPC. These equations are often heterogeneous at fine granularity levels. A major source of heterogeneity is related to the role of public streetlamps, which emit vastly more radiance at night per electricity consumption than any other appliance. We propose a framework to disentangle this issue and we apply it to data on a mini-grid project in Madagascar. We show in this case study that overlooking street lighting may bias, by up to a factor 2, the impact of this project on EPC.
This study evaluates artificial neural networks (ANN) as credit-scoring tools for microloan allocation in a public non-profit microfinance institution. Using data from vulnerable Uruguayan entrepreneurs, the model incorporates borrower, business, and loan characteristics available at application. Results show that ANN substantially outperform traditional approaches in predicting structural default, defined as repayment of half or fewer scheduled instalments, rather than minor or temporary delinquency. The findings demonstrate that low-cost, readily available information can improve default-risk assessment. ANN can complement loan officers' judgement, optimise resource allocation, and reduce structural default, thereby strengthening the institution's financial sustainability and resilience.one
Evidence on workers' willingness to pay (WTP) for location flexibility in Latin America remains scarce. Using a discrete choice experiment with 4,785 working-age participants in Argentina, Chile, Colombia, Mexico, and Peru, we estimate WTP for hybrid and fully remote formal jobs in manufacturing and information technology. Conditional logit models show that participants sacrifice 10% of wages, or USD 109 monthly, for hybrid work and 6%, or USD 60, for fully remote work. Women value both arrangements more than men, with statistically significant gender differences. Heterogeneity by age and commuting distance is limited; caregiving differences are not detected. Robustness checks confirm results.
This study evaluates a programme that improved caregiver training and infrastructure at locally managed childcare centres in high-poverty areas of Bolivia. In 2012, 170 centres were randomly assigned to treatment or control groups. The programme increased centre survival by 46 percentage points in 2018 and 23 percentage points in 2023, with larger effects among the most vulnerable centres and municipalities. It also substantially improved service quality, with surviving treatment centres scoring 1.23 standard deviations higher than controls. Mediation analysis suggests that human-capital investments, rather than infrastructure, drove these effects. The estimated benefit-cost ratio is 2.6, largely due to prevented closures.
Home garden interventions are often believed to empower women, but evidence is limited, especially in Africa. This study examined the impact of home gardens on women's empowerment in Mali using interview data from men and women across 605 households and applying a difference-in-differences estimator. The analysis distinguished between the average effect across all participants assigned to the intervention (intent-to-treat) and the effect among active participants (treatment effect on the treated). It employed the project-based Women's Empowerment in Agriculture Index. Qualitative data provided context. The intervention increased women's empowerment, gender parity, vegetable area, and growing period, but only among active participants.
Secure and transferable land rights are promoted as a key to ensure the reallocation of land to productive users through land rental markets to improve food security and reduce poverty, especially in developing countries. Using a unique nationally representative panel data from the land certification program in Ethiopia, this paper examines the effect of land certification on household's decision to rent out land and heterogeneity in the impact of the program on rental market participation across regions of the country. We also examined regional rural land regulations that include regulations on rental markets and their implication on rental market development. Using different panel data approaches, we find that improved tenure security through land certification has contributed positively to rental market development in the country. Moreover, we also find evidence of heterogenous effect of land certification on land rental market participation across regions in Ethiopia, where stronger rental market effects are observed in Amhara region and relatively, the effects are significantly lower in the Oromia and the SNNP regions.
Policies aimed at improving school governance via community-based management are common but understudied. We present the results of a cluster-randomised trial of a multi-faceted education intervention that involved support for local School Management Committees (SMC) in rural India. The intervention increased SMC meetings held by 17 per cent and completed school improvement plans by 38 per cent, gains that persist 1 year after the intervention ended. There was also a 14 per cent increase in the number of hired teachers but no other improvements in infrastructure. We provide suggestive causal mediation analysis, showing that SMC activities and teachers hired only explain a small and insignificant portion part of the bundled programme's total effect on learning.
We evaluate the impact that bicycle provision has on education and empowerment outcomes for girls in remote communities. Bicycles were provided as a sub-component of a large multidimensional girls' education programme implemented across rural Zimbabwe. The bicycles and the larger programme were designed to address barriers experienced by the most marginalised children in these communities, including girls who have to travel long distances to get to school. Evaluating the specific components of multifaceted development programmes is often challenging when randomisation is not feasible or when sub-components are prioritised for specific populations. We demonstrate how optimal full matching can be used to isolate the marginal impact of this single component within a larger, multi-dimensional education intervention. Optimal full matching leads to a more balanced match across pre-treatment characteristics than more common matching exercises that use one-to-one matching. This improved match quality has implications for the resulting treatment effects. Using this approach, we show that the bicycle programme improved attendance and girls' sense of empowerment, but unlike other bicycle programmes implemented independently, there was no evidence of improved learning.
This paper capitalises on a randomised cash-transfer intervention implemented in rural Morocco between 2008 and 2010 to shed new light on the interplay between household inequality, as driven by gender and unpaid care work (UCW) dynamics, and children's schooling. The study explores the effect of the cash transfer on school progression and UCW, including analyses of how effects vary by gender and time spent on UCW prior to intervention implementation. Results suggest that the intervention increased the likelihood of girls progressing through grades on time by approximately 6 to 10 percentage points, while it had no discernible effect for boys' timely grade progression. Nonetheless, among girls the benefit of the treatment on timely grade progression was halved for each additional hour of UCW prior to intervention implementation, and the transfer proved ineffective in lessening the care burden. Taken together, findings suggest that - as a result of the intervention - girls performing UCW were staying in school more but were less likely to progress on time relative to their counterparts not engaged in UCW. Insights from this research shed light on whether promoting gender equitable opportunities within the household might enable children to follow a more regular school path.
Obesity rates are rising around the world and so are its health and socio-economic costs. In order to effectively address the problem of obesity, it is important to understand the drivers of the disease. Education is one potential determinant and while there is a large body of literature on the causal impacts of schooling on obesity status, bulk of the existing evidence comes from developed countries. In this paper, we study the education gradients in adult female obesity in two African settings - Uganda and Zimbabwe. In the past, these countries have implemented large national school reform programmes that significantly enhanced educational attainment. We leverage age-specific exposure to these reforms to instrument for women's education and use a two stage least squares (2SLS) methodology to isolate the effects of education on later-life obesity. Despite evidence of positive and statistically significant education-obesity correlations, we cannot reject the null hypothesis that schooling has no overall causal impact on obesity in either study setting. We also find no causal effects on body mass index and being either overweight or obese. The significant cross-sectional associations between education and obesity are evidently driven by correlations between schooling and unobserved factors influencing both schooling and obesity status.
We study the effects of integrated social services on the utilization and subjective life satisfaction of women in El Salvador.The Ciudad Mujer "one-stop shop" centers integrate health, legal, employment and other services into a single secure environment for women.These integrated services could boost demand by reducing the cost of access, improving quality and exploiting complementarities in service provision.Using a randomized encouragement design, 4,062 women in the vicinity of three centers were randomly encouraged to visit Ciudad Mujer (treatment group) or a local health clinic (placebo group), or they received no encouragement (control group).Approximately 1 year later, women who were exposed to Ciudad Mujer through encouragement visited the center an additional 2.1 times, increased the use of public services by 0.47s.d. and reported an improvement of 10% in life satisfaction relative to both the placebo and control groups.
Community-based savings groups (SGs) both provide credit to the poor and incentivise repayment by allowing community members to borrow from their neighbours. However, available capital is typically limited by members' savings. We study an intervention relaxing this constraint by exogenously adding initial capital to SGs. Unlike traditional bank loans, this outside money is paid back with the same timing and formula as the other members of the SG, greatly simplifying the process for the SG. Using a randomised experiment in rural Uganda, we find this increases both loans made to SG members and payouts received by SG members. We fail to find evidence that the intervention compromises the functioning of the SGs. Our analysis fails to identify significant effects on the quantity of savings or on SG functioning as measured by disbandment, defaults or failure to repay loans. Cost-benefit calculations indicate a high social rate of return to the intervention.
Cost-Effectiveness Analysis (CEA) is widely used among policymakers due to its intuitive way of estimating the cost of an intervention per unit of outcome improvement. For interventions with multiple outcomes, however, this standard approach tends to underestimate programme efficiency. We propose a method to discount cost-effectiveness ratios (CER) based on multiple benefits and illustrate it using RCT findings from a violence-reduction programme in Liberia. The standard CER suggests a cost of $51.8 to improve economic performance by 0.1 SD among cash-transfer recipients, but our discounting approach reduces this to $5 once two additional outcomes are considered. Although these figures reflect only private benefits without considering positive externalities for society at large, accounting for multiple outcomes helps identify cross-sector synergies and compare the efficiency of interventions achieving similar results.
This paper re-examines the labour market effects of Brazil's Bolsa Familia using nine waves (2001-2009) of individual-level PNAD data and a fixed effects strategy that controls for census-tract unobservables, state-year shocks and tract-specific trends. In contrast to earlier tract-average studies, the individual specification reveals no significant change in aggregate labour force participation or weekly hours. Beneath this neutral mean, however, clear heterogeneity emerges: employment rises for women, for non-white adults and in rural areas, while a small but significant shift from formal to informal jobs is observed among urban men close to the eligibility threshold. The pattern is consistent with liquidity and child-substitution channels outweighing pure income effects, while the sharp means-test creates incentives for marginal formal workers to conceal earnings. Overall, Bolsa Familia does not discourage work; instead, it modestly expands employment among disadvantaged groups. These findings reinforce the programme's value as a poverty-reduction and labour-activation tool, thereby supporting the UN Sustainable Development Goals related to inclusive growth and effective social protection.
I investigate the impact of conditional cash payments in Kazakhstan on adult labour supply and the household financial outcomes. The identification strategy exploits the random assignment of sub-districts into treatment and control groups. The results show evidence that the benefits increased the likelihood of employment among mothers by 7.6 percentage points, but had no effect on household heads or fathers. Additionally, households in the treatment group were 8.5 percentage points more likely to borrow. This study extends the literature on developing countries by examining the effect of a conditional cash transfer programme in a new, understudied setting.
This study investigates the impact of an irrigation rehabilitation scheme on rice productivity and water management practices during wet and dry planting seasons, using data from small-scale community irrigation systems in the Philippines. It examines whether rehabilitated irrigation infrastructure has reduced the risk of climatic event, in particular, droughts. The study also explores whether the intervention led farmers to sustain the management of communal water resources, ensuring sufficient water to improve rice productivity under the extreme climatic condition. Using a sample of 2,583 parcel-household levels from 113 community irrigation system clusters managed by irrigators' associations and random assignment to treatment and control groups, estimates suggest that rehabilitated irrigation infrastructure mitigated the effects of drought, leading to increased rice productivity. Farmers also demonstrated collective action through voluntary maintenance, particularly among those who attended water management training.
Credit and savings interventions have been heralded as a promising path to empowering women. However, the evidence base is inconclusive. Thus, the objective of this review is to synthesise the evidence in relation to the impact of credit and savings interventions on women's economic empowerment and agency adopting a meta-regression approach. Eighteen experimental studies were identified and synthesised finding that effect sizes are very small and mixed in their sign and level of significance. Upon examining whether specific characteristics are associated with the effectiveness of credit and savings interventions, no statistically significant effects could be identified for credit interventions and only small effects for savings interventions, possibly because the latter have less of a risk of indebtedness. Overall, effects are small, levels of heterogeneity are high, risk of bias concerns of the included studies prevail, and publication bias exists. Credit and savings interventions by themselves are not the path to economically empower women. It may be that the non-financial features that emphasise awareness raising, skills training, education on women's rights and enhancing women's social networks are the driving force behind the empowerment of women, more so than financial features.
Do country-level factors make a difference in whether a foreign aid project achieves its intended outcomes? If so, what country characteristics are most important? This study examines the project-level evidence regarding how the criteria considered in the allocation of foreign assistance translate into better development outcomes. I exploit a novel database of project ratings for more than 20,000 projects standardised across twelve different donors. I examine which country-level factors matter for project outcomes and produce evidence in support of the allocation criteria that are drivers of project success. Using a fixed effects model, I find that higher-income countries achieve better project outcome ratings, good governance and control of corruption are critical factors in project success, and democracy is not a key driver of project performance. While it is not clear that donors should favour higher-income countries based on these findings, my results suggest that donors could potentially improve their effectiveness by shifting their allocation of resources towards better-governed countries. These findings provide support for the World Bank and MCC performance-based allocation models, though slight tweaks should be considered.
This study evaluates the impact of Malaysia's People's Housing Project (PPR) on crime outcomes using a staggered difference-in-differences design and administrative police data (2013-2020). Results show significant reductions in property crime, especially house break-ins, vehicle-related theft, and robberies, with declines of about 21%, 10%, and 16% respectively. Effects on violent crime are mixed. A supplementary analysis using satellite-based nightlights shows modest but directionally consistent effects. Overall, findings indicate that well-planned social housing can improve local safety in addition to meeting housing needs.
