Food Transfers, Cash Transfers, Behavior Change Communication and Child Nutrition: Evidence from Bangladesh

"[T]hese results demonstrate that social protection programs that combine cash transfers with well-implemented nutrition BCC [behaviour change communication] can represent a promising approach to reducing chronic undernutrition in low-income settings."



While there are a range of interventions that appear to effectively reduce forms of micronutrient deficiency as well as severe acute malnutrition, there are fewer promising approaches to reducing chronic undernutrition, and those that exist are challenging to scale. This paper reports the results of two 2-year randomised control trials (RCTs) in two economically poor rural areas of Bangladesh. Treatment arms included monthly cash transfers, monthly food rations of equivalent value to the cash transfers, and mixed monthly cash and food transfers. Treatment arms also varied the inclusion of complementary programming that aimed to improve knowledge and practices around infant and young child nutrition through behaviour change communication (BCC); one arm combined food with BCC, while another combined cash with BCC. This design enables a direct comparison of the impacts on child nutrition of multiple modalities (such as cash transfers and food transfers) within the same setting, for comparable study populations, for equivalent transfer value, and with comparable implementation duration and frequency.



The Transfer Modality Research Initiative (TMRI) study operated from May 2012 to April 2014 in Rangpur division (province) in the northwest (hereafter "North") and Barisal and Khulna divisions in the south (hereafter "South"). The TMRI sample consists of 4,992 households at baseline: 2,498 in the North and 2,494 in the South. The sample is restricted to biological children of the household head, which leads to an estimation sample of 4,399 children: 2,218 in the North and 2,181 in the South. In each region, 10 households were selected in 250 different villages (clusters), with 50 clusters allocated to each treatment and control arm. Participating households lived in districts (Upazilas) with high levels of poverty relative to the region in which they were located.



In both the North and the South, the RCTs were designed to ensure that across many dimensions, treatments were identical. All arms were identical in terms of the value of the payments (1500 taka), the identity of the recipients (mothers of children under age 2), the duration (24 months), frequency (monthly) and timing (second week of each month), as well as the receipt of a basic mobile phone. They differed only in terms of the transfer modality (Cash, Food, or a Cash & Food combination) and whether the beneficiaries received nutrition BCC.



TMRI used the same approach to the BCC across the two regions. The BCC component consisted of a suite of activities led by a Community Nutrition Worker (CNW). The CNWs, all women, were from local communities and came from the same villages as TMRI participants. The core BCC activity was a weekly, one-hour group session in each village with a CNW. The BCC session materials were derived from material developed for Alive & Thrive (A&T) in Bangladesh, a comprehensive programme aimed at improving breastfeeding and complementary feeding practices and reducing stunting among young children. TMRI also followed A&T in terms of the content of the BCC sessions and an approach that included community engagement, group BCC sessions, and home visits. A variety of methods were used to deliver information, including presentations, question and answer, interactive call-and-answer songs and chants, practical demonstrations, and role-playing.



One of these sessions per month, with only beneficiaries participating, occurred on the day of the transfer distribution. For the remaining group BCC trainings each month, other household members - particularly mothers-in-law, husbands, and other pregnant or lactating women - were invited to attend along with beneficiaries, with the intention of creating a supportive household atmosphere and behaviour change at the household level. These combined sessions facilitated women's ability to participate in the BCC and had the potential to increase uptake of BCC messages, as husbands and mothers-in-law are also key decision makers on food purchases and child-rearing.



Quantitative and qualitative data collected throughout the intervention indicate that implementation fidelity was high.



Intent-to-treat estimators show that cash transfers and nutrition BCC had a large impact on nutritional status, a 0.25 standard deviation increase in height-for-age z-scores, and a 7.8 percentage point decrease in stunting prevalence. Cash + BCC was more effective at improving child nutrition than cash alone, and that Cash + BCC was more effective than Food + BCC. No other treatment arm affected anthropometric outcomes: Cash or food transfers by themselves have little impact on children's nutritional status. There are no differential impacts when the sample is disaggregated by child age, maternal age, or maternal education.



Furthermore, treatment arms that included BCC led to increased use of latrines (by 23.8 percentage points in the North and 10.0 percentage points in the South), as did Cash transfers in the North. Treatment arms that included BCC led to improved garbage disposal, by 5.6 percentage points in the North and 13.3 percentage points in the South. In both the North and the South, treatment arms that included BCC led to improvements in the cleanliness of the home environment and in purchases of hygiene-related goods; for some but not all of these impacts, the differences were statistically significant.



Both Cash + BCC and Food + BCC led to large increases in maternal knowledge of infant and young child nutrition (relative to arms without BCC), and effects were similar across Cash + BCC and Food + BCC. For instance, in the North, Cash + BCC increases scores on breastfeeding knowledge by 11.8 percentage points, complementary feeding knowledge by 22.6 percentage points, and all infant and young child nutrition (IYCN) knowledge by 19.4 percentage points. Other treatment arms have either no significant effect, or effects that, while statistically significant, are tiny in magnitude. Put another way, at endline, in the North, mothers knew much more about IYCN if they were exposed to the BCC activities compared to mothers in the Cash-only treatment. Similar results were found in the South.



The paper explores mechanisms at the household level and the child level to understand why these findings emerge. For example, nearly all treatment arms increased household caloric acquisition, but the increase was largest and most diversified toward animal-source foods for Cash + BCC. These large impacts of the Cash + BCC arm on household diets were driven by a combination of (a) increased income - both from the transfer itself and from the fact that Cash + BCC led to investment in new income-generating activities; (b) positive expenditure-caloric elasticities, which mean that this higher income translates into increased caloric acquisition, particularly of animal-source foods; and (c) Cash + BCC shifting household caloric acquisition towards a more diverse diet that includes some quantities of pulses and animal-source foods.



All these results on diets are consistent with the finding that only Cash + BCC resulted in improved child nutritional outcomes, with analysis indicating that both increased knowledge and income play a role.



Reflecting on the findings, the researchers conclude: "These results on the role of BCC and the mechanisms through which it operates suggests that work on interventions seeking to improve children's nutritional status may well benefit from designs that attempt to address multiple constraints - energy, diet quality, maternal knowledge, among others - rather than focusing on only one of these. In this setting, food transfers alone did not improve child nutrition outcomes, possibly because they did not lead to increases in children's consumption of foods most important for growth. Cash transfers also had limited impacts on child nutritional status."