[Paper Review] Cash versus Kind: Benchmarking a Child Nutrition Program against Unconditional Cash Transfers in Rwanda

Craig McIntosh, Andrew Zeitlin|arXiv (Cornell University)|Jun 1, 2021

Child Nutrition and Water Access13 citations

TL;DR

This study compares Rwanda's multi-component in-kind child nutrition program (Gikuriro) to cost-equivalent and larger unconditional cash transfers via a cluster-randomized trial. While a $124 cash transfer had no significant impact on core child health outcomes, a $517 cash transfer significantly improved consumption, dietary diversity, and child growth, suggesting cash can outperform in-kind programs when scaled appropriately.

ABSTRACT

We benchmark a multi-dimensional child nutrition intervention against an unconditional cash transfer of equal cost. Randomized variation in transfer amounts allows us to estimate impacts of cash transfers at expenditure levels equivalent to the in-kind program, as well as to estimate the return to increasing cash transfer values. While neither the in-kind program nor a cost-equivalent transfer costing \$124 per household moves core child outcomes within a year, cash transfers create significantly greater consumption than the in-kind alternative. A larger cash transfer costing \$517 substantially improves consumption and investment outcomes and drives modest improvements in dietary diversity and child growth.

Motivation & Objective

• To evaluate whether unconditional cash transfers outperform in-kind child nutrition programs in improving child health outcomes.
• To assess the cost-effectiveness of cash transfers relative to multi-dimensional in-kind interventions like Gikuriro.
• To identify the minimum cash transfer level needed to surpass the impact of the in-kind program on key child health indicators.
• To examine whether impacts vary across subgroups or depend on transfer size, informing optimal targeting and program design.

Proposed method

• Conducted a cluster-randomized trial across 248 villages in Rwanda’s Kayonza and Nyabihu districts.
• Randomized transfer amounts around the estimated cost of the Gikuriro program ($124/household), with a subset of villages receiving a larger transfer ($517).
• Used regression adjustment to enable precise cost-equivalent comparisons between in-kind and cash transfer programs.
• Measured outcomes across multiple dimensions: child anthropometrics (HAZ), dietary diversity, consumption, and investment.
• Applied machine learning to identify optimal targeting rules across multiple outcomes.
• Pre-registered the study with the AEA (AEARCTR-0002559) and obtained ethical approval from Rwanda’s National Ethics Committee and IPA.

Experimental results

Research questions

• RQ1 Does a cost-equivalent cash transfer improve child nutrition outcomes more effectively than the in-kind Gikuriro program?
• RQ2 What is the minimum cash transfer level required for cash to outperform the in-kind program on key child health indicators?
• RQ3 How do impacts scale with increasing transfer size, and is there evidence of non-linear effects?
• RQ4 Are there significant differences in program effectiveness across subgroups defined by household characteristics or child age?
• RQ5 To what extent do beneficiary preferences—revealed through choice—align with optimal program targeting?

Key findings

• A cost-equivalent cash transfer of $124 per household did not significantly improve core child health outcomes (e.g., HAZ, dietary diversity) compared to the in-kind Gikuriro program.
• A larger cash transfer of $517 significantly improved household consumption and dietary diversity, and led to modest but statistically significant improvements in child growth (HAZ).
• The smallest cash transfer at which benefits exceeded those of Gikuriro was estimated at $694 for savings outcomes and $277.37 for HAZ, indicating a threshold effect for impact.
• No evidence of non-linear effects was found for primary outcomes, suggesting a constant benefit-cost ratio across transfer sizes, though intermediate effects were observed in a sister study.
• Results showed minimal heterogeneity in impacts across subgroups, and no evidence that cash transfers were superior in any specific dimension, supporting the use of simple lump-sum transfers with preference-based targeting.
• The study highlights that beneficiary choice—revealed through cash transfers—provides a powerful benchmark for evaluating paternalistic development programs, especially when trade-offs between multiple objectives are involved.