‘Too hungry to learn? Inequalities in adolescent learning achievements by timing and persistence of household food insecurity in India’



By Elisabetta Aurino, Imperial College London



Jasmine Fledderjohann, University of Lancaster




Sukumar Vellakkal, BITS Pilani Goa



For Citation: SITE4Society Brief No.12-2018
Related to SDG Goals and Indian National Programmes: #SDG2 (Zero hunger) #SDG4 (Quality education) #SDG10 (Reduced inequalities) #NFSM (National Food Security Mission)
Economics sub-disciplines: Development economics, Economics of education, Health economics.
SITE focus:  Need for better Engagement
Country focus: India and, more broadly, low- and middle-income countries
Based on: Elisabetta Aurino, Jasmine Fledderjohann, Sukumar Vellakkal (2018). Inequalities in adolescent learning: Does the timing and persistence of food insecurity at home matters? Department of Economics, Ca’ Foscari, University of Venice Working Papers No. 9.


Food insecurity at home—that is, difficulties consistently accessing safe and nutritious food–can shape children’s cognitive, health and general development trajectories in multiple ways. For instance, being food insecure at home can influence children’s choice between continuing school or leaving school for employment, their focus in school, their psychological well-being, and their ability to participate in family and community life without worry.

India is a particularly relevant setting in which to study food insecurity, in part because it is home to one out of three people globally suffering from hunger, and to the largest share of chronically malnourished pre-schoolers globally (about 50 million). This reality has long-term consequences in terms of impaired human development. Given the sheer magnitude of food insecurity in India, it is hardly surprising that it has been ranked at the “top of India’s policy agenda”.

At the same time, despite the striking improvements in access to basic primary and secondary education of the past few decades, learning achievements have not followed: for instance, only 27% of a sample of about half a million rural Indian students were able to perform a 2-digit subtraction in Grade III.

Could food insecurity in childhood and early adolescence be an explanation for child learning inequalities? This query paved the way for the present work.

figure 1
 Research Questions
 Our recently published working paper addresses the following questions:

  1. Is there an association between household food insecurity in childhood and educational outcomes in early adolescence?
  2. (How) do timing and persistence of food insecurity in childhood matter for educational outcomes? And does this vary by type of skill considered?


Motivation for Research Questions

Food insecurity may shape adolescent[1] learning through multiple pathways: first, faced with food insecurity, households may prioritise the purchase of basic foodstuffs as compared to non-food items, and consequently may invest less in educational inputs (e.g. school fees, private tuition, educational materials, uniforms). Second, children may be more likely to work within or outside the household as part of a household’s response to food insecurity, which may lead to increased absenteeism, less time to study, and earlier dropout. A further channel is health-related: hunger and morning fasts have adverse effects on cognition, particularly through slower working memory, fatigue and distraction.

Importantly, the timing and persistence of food insecurity may matter, too: for instance, early childhood food insecurity may most strongly affect vocabulary development, as this skill is usually built in the early years, while continuous spells of food insecurity may have larger repercussions on maths learning due to the cumulative nature of the curriculum.

These aspects have been seldom investigated, particularly in low- and middle-income countries and with a focus on adolescents.


© Young Lives/Sarika Gulati

Methodology Used

We used data from the Young Lives study in India (in the states of Andhra Pradesh and Telangana), which is a longitudinal study of child poverty. A longitudinal sample of about 2000 children and their households were followed over time in 2002, 2006, 2009, and 2013[2], when the children were aged about 1, 5, 8 and 12 years. In this paper, we focused on 2006-2013 data for whom data on child learning achievements and household food insecurity were collected. The longitudinal nature of the data allowed us to examine how timing and persistence of food insecurity at different stages of childhood (at 5, 8 and 12 years) are associated with children’s learning achievements at 12 years.

How do we measure household food insecurity trajectories?

Household food insecurity is a multidimensional phenomenon, which includes limited food availability and/or access, nutrient loss through food preparation techniques, and inconsistency of these dimensions over time. In this study we focused on the food access dimension, measured through a specific experience-based scale of food insecurity. The scale captures the extent to which households cope with situations of food insecurity, including cutting food quality and quantity, skipping meals and going hungry. Similar scales are currently used by the UN Food and Agriculture Organisation as one of the measures to assess global food insecurity.

Based on the household food insecurity scale methodology, we categorised a household as food secure or food insecure in each survey round. We were then able to track household food insecurity dynamics over time–that is, whether the household was never food insecure, whether it was always food insecure, or whether it transitioned in or out of food insecurity at specific periods in children’s lives. Based on the possible combinations of moving into and out of food insecurity, we ended up with six categories capturing these pathways over time. These categories are show in the table below.


Table 1. Categories of food insecurity variable capturing pathways into and out of food insecurity over time


The outcomes we were interested in were adolescent learning outcomes at age 12 across four domains: reading, vocabulary, maths, and English. These were measured by scores on separate learning tests related to each of these domains. Tests were developed by educational specialists to be linked with the local curriculum at different ages.

Our analysis comprised:

  1. A descriptive look at which households were at risk of food insecurity, and whether there were bivariate associations (that is, associations not controlling for any potential confounders) between food insecurity and our childhood and pre-adolescent learning outcomes of children at 12 in reading, vocabulary, maths, and English.
  2. Multivariate regression models were estimated to identify the associations of household food insecurity trajectories and child learning after controlling for other potential drivers of achievements. For example, it could be hypothesized that household food insecurity doesn’t matter, but rather it is poverty that affects both food insecurity and children’s learning. Thus, a variety of possible drivers of learning  such as socio-economic status and child, caregiver, and household attributes (e.g. child age, caregiver age and education, household size, household wealth, living in an urban area and so on) were included in the estimates. All models also included controls for early childhood test scores in vocabulary and grasp of quantitative concepts to account for unobserved differences between children and families such as child ability, early investments in child education and health, and parental preferences for education.
  3. As a further step, we explored in more refined models additional potential explanations of adolescence learning such as: (i) educational inputs (e.g. school enrolment and type of school (private/Government); (ii) health inputs (nutritional status as measured by height-for-age z-scores); (iii) and adolescent and parental psychosocial skills (e.g. educational aspirations, self-efficacy and self-esteem scores), which were all measured at 12 years old.


Main Findings
Descriptively, the data show:

  • 12-year-olds who experience household food insecurity have poorer learning outcomes across the board (see Figure 3).
  • Those experiencing food insecurity at age 5 years, and experiencing it for a longer period of time over childhood have particularly negative learning outcomes at 12 years.
  • Even some of the wealthiest household experienced food insecurity, but poorer households faced a much higher risk.

Our multivariate models indicate:

  • Food insecurity in childhood is associated with worse learning outcomes in adolescence; early and chronic food insecurity are the most consistent predictors of impaired cognitive skills at age 12.
  • Temporary spells of food insecurity (i.e. transitioning into and back out of food insecurity) are not associated with negative vocabulary and reading outcomes, suggesting children may bounce back from short-term food insecurity in some learning domains.
  • However, transitory food insecurity does matter for maths. This may be due to the cumulative nature of maths learning, which may make catch-up learning more difficult.
figure 3.jpg

Figure 3. Food insecurity and adolescent learning outcomes, Young Lives India, 2006-2013. Mean cognitive achievements by household food insecurity trajectory

Some concluding ‘food for thought’

Our findings suggest that pursuing the Sustainable Development Goal 2 on ending hunger and achieving food security for all is not only critical to tackle the burden of hunger and food insecurity, but also because it may have ripple effects and foster progress on other SDGs, such as reducing inequalities (SDG #10) and ensuring inclusive and quality education (SDG #4).

figure 4.jpg

© Young Lives/Sarika Gulati

Food insecurity at home during childhood matters for learning achievements in early adolescence. Moreover, when food insecurity occurs and how long it lasts for can also have important implications for learning outcomes.

We do not claim any causality in this research, but we do provide rigorous findings on the relation between food insecurity and learning. Based on our results, and complementing existing research and policy recommendations, the following points provide some “food for thought” to policymakers:

  1. Intervene early. Because some curriculum is cumulative, it may be difficult for children who experience childhood food insecurity to catch up on learning later. Early intervention is therefore essential. Increased budget allocation and effective implementation of India’s early childhood health program (ICDS), especially for revising and reforming the ‘supplementary nutrition provision’ to ‘full feeding’, could be beneficial too.
  2. Offer remedial learning opportunities for groups at higher risk of food insecurity (e.g. tribal populations), particularly for those skills for which catch-up is more challenging (e.g. maths). Children who are in situation of chronic household food insecurity are at particular risk of falling behind their peers, with a cumulative effect over the lifecourse in terms of lower learning outcomes and increased probability of drop-out. Teaching at the right level and remedial education programmes can help children who have fallen behind to catch-up with peers.
  3. To improve learning, (also) address food insecurity. Children in food insecure households fare worse across a range of learning domains. This is true even after more commonly-cited culprits such as household socioeconomic status or earlier investments in child cognition are taken into account. Efforts to raise learning levels may be more effective when paired with a focus to address food insecurity.
  4. Improve targeting, monitoring and nutritional contents of food-based social protection (e.g. Mid-day Meal Scheme (MDMS), public distribution system (PDS), etc). Studies have shown that the MDMS can improve learning and classroom effort. Expanding this programme to include breakfast for children at higher risk of food insecurity or in the lean season may prevent hunger and its negative repercussions on child cognition. Also, the nutritional value of the PDS could be enhanced by provisioning eggs or traditional cereals such as millet.
  5. Make children’s work work for children. While it is tempting to advocate strict child labour laws in order to ensure that food insecure children stay in school, this is only efficacious where adequate social protection exists. Findings based on interviews with more than 28,000 rural adolescents aged 14-18 years across India revealed that the vast majority of them work. Emphasis on creating decent and safe employment opportunities during school breaks might minimise the risk of disruptions to schooling. Also, improving the quality of vocational training can incentivise them to stay at school for longer and transition to better employment later on.



EA would like to acknowledge the “Guido Cazzavillan” and the Imperial College London research fellowships that supported this work. Also thanks to Shyama Ramani and Nanditha Mathew for their encouraging support and feedback. Thanks to Anastasia Bow-Bertrand and the Young Lives team to share their pictures with us and a very special thanks to the children and the families of the Young Lives India sample for their time and their insights.

[1] We use the WHO definition of adolescence, which refers to the period between 10 and 19 years.
[2] A further round was collected in 2016 but are not yet publicly released.

Leave a Reply