Science, Finance and Innovation

‘Labeling’ can improve cash transfer program outcomes: evidence from Kenya

4 min


Carolyn J. Heinrich and Matthew T. Knowles

Encouraging poor households to increase their take-up of healthcare and education is a key objective of development policy, leading to the growing use of conditional cash transfers. But implementing and enforcing the conditions of these programs is demanding for both administrators and families. This column reports surprising evidence from Kenya that similar outcomes can be produced via ‘labeled’ cash transfer programs, which provide guidance for family spending but do not involve monitoring or enforcement.

Cash transfers are one of the most popular aid interventions directed at reducing poverty. Conditional cash transfer (CCT) programs are operating in more than a fifth of all countries, but unconditional cash transfer programs (UCTs) are proliferating rapidly as well, and are among some of the largest cash transfer programs today. The implementation and enforcement of conditions requires substantial infrastructure and administrative capacity, and the direct costs of complying with conditions can be burdensome for families, too.

For these and related reasons, the implementation of UCTs has become more commonplace in very low-income countries. We are also seeing the introduction of ‘labeled’ cash transfer programs (LCTs), where guidance for spending the transfer is communicated but not monitored or enforced.

CCTs differ in their implementation of ‘hard’ versus ‘soft’ types of conditioning: they vary by the extent of monitoring or the likelihood of fines or other sanctions being imposed for non-compliance with the conditions. An LCT fits somewhere along the continuum between a CCT and a UCT. This raises the question of whether the ‘nudge’ or guidance that an LCT provides to encourage intended uses of the cash transfer is sufficient to achieve the desired program outcomes (in contrast to the monetary stick used with CCTs to ensure compliance).

In our research, we take advantage of a unique opportunity to compare the effects of a CCT that imposed monetary penalties for non-compliance with program conditions to an LCT arm of the same cash transfer program, which provided only guidance for cash transfer use.

We undertake this analysis in the context of the Kenya Cash Transfer Programme for Orphans and Vulnerable Children (CT-OVC), an LCT that was distinct in its random assignment of conditions with penalties within locations that were randomly selected to receive cash transfers.

The CT-OVC conditions include regular visits to health facilities for immunizations, growth monitoring and nutrition supplements, school enrollment and attendance, and attendance at caregiver ‘awareness’ sessions. Accordingly, we are interested in the program’s effects on children’s health (that is, vaccinations and Vitamin A supplements received) and absences from school, as well as on household consumption.

We find that the labeling of cash transfers results in similar beliefs and perceptions of the program rules and expectations between households in the CCT and LCT program arms. The one exception is that CCT households are more likely to believe that they would be penalized with a monetary fine for failing to comply with the program rules.

For this reason, we expect that assignment to the CCT arm would mostly affect program outcomes differentially through the increased likelihood of being fined. In fact, about one third of the households assigned to the CCT arm in the CT-OVC receive a penalty fine within two years of the program starting.

As in the one prior experimental study that we know of comparing a CCT with an LCT (in the Tayssir program in Morocco), we find no evidence that CCT households experienced different outcomes, on average, than LCT households. Even in the presence of a larger cash transfer amount in the CT-OVC than the Tayssir program, coupled with a higher probability of being penalized for noncompliance, households are not more likely to comply with the conditions than those receiving only guidance or labeling with the transfer.

But when we examine how the effects of assignment to the CCT (relative to the LCT) varies by household wealth measured at the start of the program, we find that the poorest households in the CCT arm report significantly lower non-food consumption than the poorest LCT households. It appears that very poor households, who are disproportionately affected by the penalty fines in the CCT arm, respond to this loss of cash resources by reducing consumption that is likely to be less essential.

Overall, our findings imply that in the Kenya CT-OVC program, imposing conditions on households (with a real risk of being penalized) did not lead to improvements in any of the outcomes of interest to the program developers. If anything, it appears as though receiving cash transfers with conditions instead of labeling only serves to reduce non-food consumption among the poorest households – an unintended and regressive effect of the policy.

In light of the non-trivial costs of implementing and enforcing conditions in cash transfer programs, we believe this calls into question the prevalence of conditioning, and suggests that more implementation and experimentation with labeling in cash transfer programs should be pursued.


Carolyn J. Heinrich
Patricia and Rodes Hart Professor of Public Policy, Education and Economics
Matthew T. Knowles
Ph.D. Student, Vanderbilt University