Expanding Lessons From A Randomised Impact Evaluation Of Cash And Food Transfers in Ecuador and Uganda

1. Introduction

There is now substantial evidence that periodic cash transfers to poor households as a form of social protection, particularly when conditional on complementary investments in child schooling and health, can lead to substantial and sustained improvements in household welfare, household food security and child schooling (Adato and Hoddinott 2010; Fiszbein and Schady et al. 2009). Similarly, food transfers can lead to substantial improvements in household food security (e.g., Barrett and Maxwell 2005) and may have persistent effects on household expenditure and food consumption (Gilligan and Hoddinott 2007). However, there is very limited evidence directly comparing impacts of the two modalities in the same setting (Hidrobo et al, 2012; Ahmed et al, 2009; Gentilini, 2007; Webb and Kumar, 1995).

This study draws from a unique set of integrated social protection experiments conducted in two countries to compare the relative impacts of cash and food transfers on household behavior in side by side comparisons in starkly different contexts: Ecuador and Uganda. The transfer modality experiments were conducted by IFPRI in collaboration with the World Food Program in order to better understand the conditions under which cash transfers may be more effective than food transfers at improving household welfare and reducing food insecurity for the poor. The original WFP-funded study included transfer modality experiments in four countries:

Ecuador, Uganda, Timor-Leste and Yemen.

The Timor-Leste study was halted after the baseline household survey because changes in the implementation of the program compromised the study design. A paper from the Yemen study on the effect of transfers on marriage markets that was part of the original 3ie proposal is not included in this final report due to problems with the data on marriage markets that made the analysis infeasible. Another research paper from this study compares how impacts of the transfer schemes on household food security and food consumption differ by transfer modality. That paper is not included in this report.

In Ecuador and Uganda, the transfer programs were provided conditional on household participation in complementary programs designed to foster additional investments to improve household welfare. In Ecuador, transfers were provided to Columbian refugees and poor Ecuadoreans conditional on household participation in training programs designed to improve employment prospects. In Uganda, cash and food transfers were provided to households with children participating in early childhood development centers in order to learn about impacts on child language and cognitive development.

Here, we report findings from three studies based on these transfer modality experiments in Ecuador and Uganda. The studies address the following research questions:

1. How does social protection affect intrahousehold conflict and intimate partner violence (IPV) in Northern Ecuador?

2. What lessons can we learn from a mixed methods qualitative and quantitative investigation on the impact of cash and in-kind tranfers on intimate partner violence in Northern Ecuador?

3. How do cash and food transfers linked to early childhood development (ECD) centers affect child cognitive development in Uganda?

In the first paper, in Section 2, we investigate the effect of transfers to women on IPV and consider whether these effects differ by transfer modality (cash, food rations and food vouchers) in Northern Ecuador. Transfers to women could decrease IPV by improving female bargaining power or reducing poverty related stress, or it could increase domestic violence if a husband feels threatened or forcefully extracts transfers. Results indicate that, overall, transfers reduce controlling behaviors and multiple forms of IPV including moderate physical and any physical or sexual violence by 6-7 percentage points. Impacts do not vary by transfer modality, which combined with the overall negative impacts on IPV, suggests that violence is not being used to forcefully extract resources.

Although there is a growing literature on the impacts of transfers on household dynamics and IPV, few evaluations have examined the pathways through which transfers impact IPV. In Section 3 we address this gap, using a mixed methods research design, to understand the potential mechanism(s) and pathways through which transfers targeted to women affected household dynamics and IPV. We find mutually reinforcing qualitative and quantitative evidence that the intervention led to reductions in IPV through the following three pathways: 1) Reduced day to day conflict and stress in the couple; 2) Improved household wellbeing and happiness; and 3) Increased women’s decision making, self-confidence and autonomy. We find little evidence that physical, sexual violence or controlling behaviours increased as a result of the transfers.

In Section 4, we study impacts of randomized food and cash transfers on cognitive development among children attending early childhood development (ECD) centers in Karamoja, Uganda. The years leading up to primary school (for children age 3-5) are now emphasized as a period of high returns to investment in child education. We estimate the impact of cash and food transfers to households with children age 36-53 months attending ECD centers on child language and cognitive development. We find that food transfers have no significant impacts, but cash transfers cause significant increases in cognitive measures, by about 9 percentage points relative to the control group. Next, using results from the cash treatment arm, we explore the relative role of improved nutrition and stimulation in contributing to the positive cognitive effects. We find evidence supporting both impact pathways and cannot conclude that either pathway dominates. Along the nutrition pathway, cash transfers reduce the prevalence of moderate-to-severe anemia in these children by 9.6 percentage points, reduce the prevalence of diarrhea by 3.5 percentage points, and increase the frequency of child consumption of starches, meat, eggs and dairy. Through transfers made to the ECD centers by cash beneficiary households, the transfers significantly improved the quality of infrastructure and sanitation at the ECD centers, including whether the center had access to a latrine and hand-washing facilities.

Along the child stimulation pathway, cash transfers increased the number of days that ECD centers were open (by 2.4 days/week on average) and the number of days children attended the center (by 1.9 days/week on average).