What do we know about cash and in-kind transfers in humanitarian settings? Not enough

This blog is co-published with the World Food Programme. 

The numbers are stark: this year, 274 million people will need humanitarian assistance, 39 million more than last year (UNOCHA 2022). Conflict, major flooding, and droughts have increased globally over the past decade, causing widespread displacement. The COVID-19 pandemic has intensified such problems, presenting an unprecedented challenge to the humanitarian system. Humanitarian assistance and social protection responses, amounting to $30.9 billion globally in 2020, are increasingly important sources of support for crisis-affected populations, yet we know little about their effectiveness and efficiency .  

The World Food Programme Humanitarian Impact Evaluation Workstream, delivered by the World Food Programme’s Office of Evaluation (OEV), the World Bank’s Development Impact Evaluation (DIME) department, and partners including USAID’s Bureau for Humanitarian Assistance (BHA), aims to generate evidence to support humanitarian operations globally. 

There is much evidence about cash transfers in development contexts (see here for a summary from our World Bank colleagues). However, the lessons learned from non-humanitarian settings do not necessarily apply to complex humanitarian contexts characterized by greater economic, social, institutional, and security challenges. Households in these contexts often have little ability to turn humanitarian assistance into sustainable income-generating activities. For example, finding employment or growing a business may be difficult. Markets in these contexts are also often failing; therefore, cash or voucher transfers may work less effectively coupled with the limited availability of food in local markets.  

A new paper, completed for the impact evaluation workstream, reviews the existing literature focusing on the most widely used forms of social assistance programs: unconditional and conditional cash transfers, food transfers, vouchers, and public works. The review includes outcomes that have been generally overlooked: (i) basic needs; (ii) financial outcomes; (iii) women’s empowerment and gender-based violence; (iv) human development (education, health, and labor force participation); and (v) social cohesion.  

Our review reveals five findings: 

1. Despite the growing use of social assistance programs in humanitarian settings, there is relatively little rigorous research on what works, for whom, and why . The small number of studies indicates the difficulty of randomly assigning assistance in emergency settings. We find that only twenty-one studies use experimental or quasi-experimental methods ––out of which nine studies use RCTs, but only four of them have a control group–– to rigorously assess the impact of humanitarian assistance programs.  

2. The evidence base for humanitarian assistance programs varies significantly across sectors. Most evidence concentrates on basic needs outcomes, such as food security, food and non-food expenditure, and coping strategies. This is followed by studies on financial outcomes such as assets, income, credit, and savings, where the evidence is emerging. What lacks, are studies examining human development outcomes, such as health, education, and labor, as well as gender and social cohesion. Women’s empowerment and gender-based violence are the least explored outcomes. This imbalance in the evidence base suggests that policy decision, especially in terms of human development should be made with caution due to the limited generalizability of the findings. 

3. Cash and in-kind transfers are effective at improving food security, but effects on other outcomes are inconclusive. We find that most humanitarian assistance programs can improve basic needs like food security and consumption, but it is difficult to draw general conclusions about other outcomes . The small number of studies show mixed evidence on income generation, credit and savings, education, labor, gender-based outcomes, and social cohesion. Context-specificity is critical––programs should be tailored to the specific type of crisis and the broader context in which the crisis occurs, as these factors may influence the implementation feasibility and uptake by targeted beneficiaries. 

4. Decisions about the most appropriate transfer modality cannot be pre-determined or generalized to many contexts. Only six studies in our review assess the relative effectiveness of cash versus in-kind transfers. We find that the effectiveness of different transfer modalities depends on the nature of the humanitarian crisis (for example, sudden onset vs. slow onset), the objective of the program or the main outcome of interest, the profile of the targeted population, implementation costs, and local market capacity, among others. More research is needed to better understand the conditions under which cash transfers are more (or less) effective than in-kind transfers.  

5. Consistent with the broader research, our review suggests that cash transfers are cheaper to deliver than in-kind transfers, even in humanitarian settings . In particular, studies suggest that mobile money cash transfer can be the most cost-efficient transfer method provided that the mobile network infrastructure is available and easy to use. The second most efficient transfer method is manual cash delivery, followed by vouchers, with food transfer being the most expensive way to deliver assistance. Four studies in our review report the implementation costs of different transfer modalities with equivalent monetary values, but only two studies discuss the cost-benefit or cost-effectiveness of such programs. We urge future studies to incorporate costing data from budgets and spending reports that are disaggregated, intervention-specific, and captured in real time over the course of an intervention (see here for more information on costing data). 

Given the lack of rigorous causal evidence on humanitarian assistance programs, conducting more impact evaluations in humanitarian settings would be beneficial. But we also understand that rigorous studies may be challenging when the safety and well-being of affected populations are at stake. To better understand implementation design choices, such as which population to target, what type of transfer modality to use, and the duration and frequency of transfers, substantial evidence gaps need to be filled.   

Based on the findings from this review, the World Food Programme Humanitarian Impact Evaluation Workstream provides technical support to countries interested in conducting impact evaluations of their humanitarian programming. Four initial focus areas following the timeline of an emergency response have been identified in this humanitarian impact evaluation workstream: (1) forecast-based financing & climate adaptation, (2) targeting, (3) cash-based transfers, and (4) peace and social cohesion.