Smarter through social protection? Evaluating the impact of Ethiopia’s safety-net on child cognitive abilities

ABSTRACT

Ethiopia’s productive safety net is the second largest Social Protection Program in sub-Saharan Africa and has been rolled out to almost 10 million beneficiaries since 2005; its effects are therefore of general interest. We provide the first estimates of its impact on children’s cognitive abilities. To identify impacts of this program, we exploit four rounds of data on a cohort of children surveyed repeatedly between 2002 and 2013. We find a small but significant positive effect of the programme on both numeracy skills and vocabulary. This is driven mainly by children in households that had graduated (left) the programme just before 2013. We argue that this is at least partially related to time allocation: graduates of the programme spent more time in school than continuing beneficiaries. We also find evidence that the maths (though not language) improvement is more pronounced for boys.

KEYWORDS:

• Ethiopia
• social protection
• children
• cognitive development

JEL CLASSIFICATION:

• I38
• O22

Acknowlegment

Thanks to the survey team at Young Lives in Oxford and Addis Ababa, and in particular to Grace Chang for excellent research assistance. Young Lives is an international study of childhood poverty, following the lives of 12,000 children in 4 countries (Ethiopia, India, Peru and Vietnam) over 15 years. www.younglives.org.uk. Young Lives is core-funded from 2001 to 2017 by UK aid from the Department for International Development (DFID), and co-funded by IrishAid from 2014 to 2015. The views expressed are those of the authors. They are not necessarily those of, or endorsed by, Young Lives, the University of Oxford, DFID or other funders. Comments welcome to [email protected]

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Vocabulary was measured by the Peabody Picture Vocabulary Test as administered by the YL survey, and memory tests from Spanish language version of the revised Woodcock-Muoz test.

2. http://www.worldbank.org/en/news/loans-credits/2014/09/30/ethiopia-productive-safety-nets-project-4, last accessed 19 December 2017.

3. Most households who received the programme in 2009 reported receiving it since 2006–7. If households received the programme in 2013 and in 2009 we assume that treatment was uninterrupted. For the fixed effects specification, we consider ‘post treatment’ for participant households as 2009 and 2013, since otherwise graduates would appear to be untreated in 2013.

4. The wealth index takes values between 0 and 1, such that a larger value reflects a wealthier household. It is the simple average of a housing-quality index, an access-to-services index and a consumer-durables index (Briones (2017)). In the analysis we use terciles of wealth index using those households in the bottom fertile of the wealth-index distribution as reference group.

5. We investigated this information to see whether it could be used to establish more nuanced treatment variables, but we found it to be too incomplete to be of use.

6. The final sample for the maths regression model is 947, 42 observations are dropped due to missing data from the control variables and the PPVT sample is 824, the difference due to those missing a PPVT score in any of the rounds being dropped.

7. Children from graduated families spend less time in school and studying than the control group children and they are less likely to be in the top tercile of the wealth index. There are still fewer mothers with greater than primary education in the treated group, who also spend less on non-food items than the control group. Finally, the maths z-score at baseline is lower for those children in households that continue to be beneficiaries of the programme in 2013.

8. Community fixed effect included for Maths models, while in PSNP we control for language dummy which cannot be combined with the community fixed effect, as language is homogeneous within most communities.

9. The test language for the PPVT in each round is the same for all but 33 children. The results presented are robust to excluding these children.

Additional information

Notes on contributors

Marta Favara

Marta Favara is a Senior Research Officer at the Oxford Department of International Development University of Oxford. She has a PhD in Economics from the University of Essex and a Masters in Economics from the University of Leuven (Belgium). Her recent work focuses on adolescent brain development, skills formation and transition to the labour market. In addition, she investigates the role of aspirations and subjective expectations as potential self-enforcing mechanisms underlying poverty and gender inequality.

Catherine Porter

Catherine Porter is an Associate Professor at Heriot-Watt University, Edinburgh, and a Research Associate at Young Lives. She has a PhD in Economics from the University of Oxford, and a Masters in Economics from the University of Bristol. Her work focuses on risk and vulnerability in Sub-Saharan Africa, the short-and long-term impacts of shocks on child and household outcomes, and the effectiveness of social protection programmes on household and child outcomes.

Tassew Woldehanna

Tassew Woldehanna is Professor of Economics at the Department of Economics, Addis Ababa University and Senior research Fellow of Ethiopian development Research Institute. Currently, he is the Principal Investigator of several projects including Young Lives, (an international study of childhood poverty), RISE Ethiopia (A six years project on Assessing the Impact of General Education Quality Improvement Program) and Early Learning Partnership Research Programs.