Understanding the impact of nutrition-sensitive social protection interventions on the use of multiple-micronutrient powders and iron supplements in rural pre-school Bangladeshi children

Conditional cash transfers (CCT), or cash plus behaviour change interventions, have become highly popular in recent years with supra-state actors such as the World Bank supporting and financing many CCT programmes in developing countries^{( 1 )}. Just about all the child cash transfers that have been implemented in Latin America are conditional. In the absence of or limited evidence on the effectiveness of unconditional cash transfers (UCT) in poverty alleviation and human capital development, there has been a significant move towards replicating the Latin American model of CCT in developing countries across the world, including sub-Saharan Africa, South Asia and India^{( 2 , 3 )}.

Nevertheless, conditionality is one of the most contested aspects of the design features of child cash transfers. Debates on conditional v. unconditional cash transfers centre on two main contending arguments. Proponents of CCT argue that conditioning transfers on behaviour change helps to address not only immediate but long-term poverty and child development outcomes through insistence on human capital investments. According to this view, while UCT may improve take up of nutrition-related interventions, an imposition of conditions will have additional effects because they will highlight to beneficiaries the importance of investing in child nutrition.

On the opposite end of the spectrum, advocates of UCT believe that human capital investments in beneficiaries can be enhanced without enforcing conditionality^{( 4 , 5 )}. UCT proponents assert that conditionality is a violation of social protection as a human right since it interferes with beneficiaries’ right to choose what and how the transfer should be spent^{( 5 )}. Ideas about needing to enforce conditions to make sure that poor people do what is necessary for themselves provokes one to ask the question, as one scholar did^{( 6 )}, ‘should disadvantaged people be paid to take care of their [own] health?’ Do they need to be incentivised to take care of their health and invest in the human development of their children? Those in support of conditionality would say yes to each of these questions. Indeed, the author who posed the first question above agreed with the use of conditions to force the poor to ‘do the right thing’. The author argued that this is not ‘excessive paternalism’ as poor people’s failure to take care of their health, if left unchecked, would have adverse effects on the whole population and that conditionality was thus the application of the principle that ‘power can be rightfully exercised over any member of a civilized community, against his will, [if the intention] is to prevent harm to others’ (p. 140)^{( 6 )}. For those who are ideologically opposed to the idea of attaching conditions to income transfers, this line of thinking adds power to the conviction that conditionality is driven by paternalism and stems from the idea that poor people need to be forced and policed to do what is right and cannot be trusted to take care of themselves^{( 5 )}. In this sense, conditionality is ‘pernicious’ in its underlying assumptions about human nature.

In the case of the trials^{( 7 )} conducted in the study published in a recent issue of Public Health Nutrition (PHN), the results showed that cash conditional on nutrition behaviour change communication (‘Cash + BCC’) yielded higher impacts in terms of knowledge of Fe deficiency, awareness of multiple-micronutrient powders (MMP) and reported consumption thereof for children aged 6–59 months, than did cash only (‘Cash’) or cash and food (‘Cash & Food’) or food only (‘Food’). While these are important findings, as they show that in contexts where MMP are widely available, nutrition BCC interventions can help increase knowledge and awareness of the importance of micronutrients and that this can result in higher consumption of said MMP, the study exhibits the same pitfall that many other attempts to measure the impact of CCT on child health outcomes fall into: the inability to categorically attribute difference purely to conditionality. In particular, in the published article, baseline measurement of MMP consumption was not conducted and thus we have no way of knowing the precise magnitude of the effect of the interventions: if consumption of MMP was much lower at baseline compared with endline in the ‘Cash’ treatment arms and higher in the ‘Cash + BCC’ treatment arms at baseline compared with endline, then the magnitude of the effect seen at endline could be much higher than we think in the ‘Cash’ arms and exaggerated in the ‘Cash + BCC’ arms. It is noteworthy that the effect size seen in the ‘Cash’ treatment arms (11·6 percentage points) was much higher than in the ‘Control’ arms, even though it was still lower than that seen for the ‘Cash + BCC’ arms. This is important because it shows that giving cash to women from poor households does result in improvements in child nutrition outcomes.

Unlike other studies that have attempted to measure the impact of conditionality on child health outcomes, the PHN study^{( 7 )} looked at knowledge and reported consumption of MMP only and did not measure any objective nutritional clinical outcomes such as anaemia. While it could be argued that once the efficacy of MMP in reducing anaemia has been proven, strategies to increase uptake, such as the ones discussed in the article, will be useful, the lack of objective nutritional clinical outcomes is an important limitation because it leaves us no nearer to answering the question of what impact CCT have on outcomes such as micronutrient deficiency.

The lack of information on participants’ nutritional status and infant feeding practices also precludes the extent to which one can assess the effect of these key factors on the impact of the interventions. Compliance with conditionality is expensive for poor people, especially in terms of time and the penalties meted out for non-compliance. The trials reported in the PHN article^{( 7 )} had intense BCC interventions that required women to allocate an hour each week for nutrition education. Monitoring attendance at the BCC sessions would have been equally laborious for community health workers who had to follow up on each person who failed to attend. Thus, it is important that we are certain that the gains promised by conditional programmes are much higher than the costs. Researchers who conducted a large cluster-randomised trial in Zimbabwe assessing the effect of conditionality v. non-conditionality on child outcomes, including completion of all vaccination visits, concluded that there was not enough evidence to suggest that attaching conditions to child cash transfers in sub-Saharan Africa has an impact on child outcomes. When compared with an unconditional grant, the Zimbabwe CCT did not achieve significantly higher effects; in fact, on many outcomes, the effect size was similar^{( 8 )}. The authors concluded that the results supported the importance of cash transfer programmes, regardless of conditionality, in developing countries; but stated that ‘further evidence is needed for the comparative effectiveness of UCT and CCT programmes in this region … further work is needed to assess whether the increased costs associated with monitoring compliance with conditions is compensated by greater improvements in child welfare outcomes’^{( 8 )}.

An important and laudable difference in the ‘Cash + BCC’ intervention implemented in the trial^{( 7 )} reported in the recent PHN issue is that the conditions were soft: the authors report that failure to comply with conditions did not result in caregivers not receiving the cash transfer, instead community health workers employed in the study followed up on each participant who failed to comply to understand their reasons for non-attendance of the BCC sessions. From the results of the trial it is clear that nutrition education regarding MMP and Fe deficiency was an important intervention in rural Bangladesh; what is also equally clear is that receipt of the cash was not strictly tied to attendance of the education sessions and yet caregivers still complied. This supports the view often proffered by those against the idea of conditionality: caregivers who are poor will do what is good for their children without needing to be forced or threatened with withdrawal of cash support. Taken together, this suggests that a ‘cash plus’ model to improving child nutrition, without enforcing conditions, is needed and yields results. There is enough evidence to suggest that cash alone cannot tackle child malnutrition, but that a myriad of interventions including cash, nutrition education, housing, access to services such as water and sanitation, and infrastructural improvements are needed. However, additional empirical data are required on the impact of these ‘cash plus’ interventions on child health outcomes, as the findings reported in the PHN article^{( 7 )} have notable weaknesses such as the lack of a BCC only comparison arm and baseline data for some of the outcomes. Furthermore, it is important to assess whether these findings can be replicated in other low- and middle-income settings.