Big study looks at impact of child poverty in Peru, India, Vietnam and Ethiopia


The Young Lives Project is a unique, long term study of childhood in 4 countries – Peru, Vietnam, India and Ethiopia.  It is exploring the impact of poverty on children’s development, and identifying the most effective policy interventions that could enable poor children to reach their full potential.

In this podcast, Amanda Barnes talks to Alan Sanchez, Principal Investigator in the Young Lives Project which is based at Oxford University, about the research he has been doing on child development in Peru. She started by asking him to explain the Young Lives study and what it is trying to achieve.

Alan Sanchez

Alan Sanchez

Alan Sanchez:  The Young Lives study  is an international longitudinal study of childhood poverty.  We started in 2002 and the intention is to track the livelihood of families for basically 15 years.  And we are working in four different countries: in Peru, in Ethiopia, Vietnam, and Andhra Pradesh in India. The study is led by Oxford University and has research partners in each of the countries.

We use a variety of methods: quantitative methods and also qualitative methods. The main objective is to have a better understanding of the causes and consequences of childhood poverty.

Amanda Barnes: So how many children altogether are included in the study?

AS: In total, twelve thousand.  And then we have eight thousand kids that we have been tracking since they were aged between one and two years old ( in the year 2002). We call them the younger cohort.  And then there are another four thousand kids that we are tracking, also since 2002, and they were between seven and eight years old the year when we started with the survey. So by now they are 18 years old.

AB: That’s a lot of kids. So generally, a sample of that kind of size: is that something that’s statistically representative of the population as a whole more or less?

AS:  Well, we do not intend to be nationally representative.  We want our samples to be very informative of each of the countries and to observe children growing up in communities with a variety of living standards and variation in terms of geographical location and ethnic groups. So that even though our intention was not to claim national representativeness, we can certainly say many things about the characteristics in which these children are growing up in each of the countries.

AB: Why were those particular countries chosen?

AS: One important thing was to have diversity. So even within each country we were looking to have diversity in terms of the communities in our sample.  But it was also important to have diversity in terms of the countries that we selected.  So on the one hand we were looking for regularities. So we wanted to be able to say whether certain processes are very similar regardless of the country, which we have found.  But at the same time some policies are very country-specific and if you only focus on one country then it means you are only able to say things about that specific country and the specific policies that take place in that place. But if you have other countries it means you observe a variety of social policies and see their effects on children.

AB:  Yeah, well that makes pretty good sense.

AB: So your research, some of it looked at what the data could tell you about what factors in early childhood had an impact on outcomes for them at later stages – even up to their mid-teens.  Can you tell us what specifically your study was trying to find out?

AS: Yes, well to let you know a bit more about the background.  This study is based on two studies that I did previously.  In one of the studies, using data for Peru for instance, we found that there was a food price crisis a few years ago and we were able to measure that this crisis had consequences on the early nutrition of those kids that were very young at the time that the crisis took place.  But not only that, also that these kids later on had a lower cognitive development because of this crisis. So what it suggests is basically a linkage between nutrition and cognition.  And then in another study, in which we used data from four countries, from the four countries from which we are collecting data, we also found a strong association between early nutrition and the development of social-emotional skills such as aspirations and self-esteem. So what this suggests basically is that early nutrition is then an important determinant of the formation of cognitive as well as social-emotional skills.

So in the study on which I’m currently working, what we want to do is build a structural model in which early nutrition can have an impact on these two dimensions of skills.

AB:  Presumably you can’t just sit there counting the calories that a child takes in every day.  How have you been able to measure their nutritional status, or how much they had to eat, when they were certain ages?

AS:  No it’s true.  So what we do is we use anthropometric information. For children observed early in life, the physical height of the child can basically give good information about the nutritional history of that child. So what we use are standards of the WHO,The World Health Organisation.

This standard will tell you that a child of a given age should have a given height. And then you compare how far away is the child that you observe relative to the standard. When the child is too far away, being too small compared to the standard, it means that this kid is likely to be under-nourished. So that’s what happens basically.  You know, in the countries in which we are collecting data about a third of the kids in our samples are under-nourished by the WHO standards.

AB: And is the height that a child should be at a given age, is that standard across all different ethnic groups when they’re young, or is there any variation that you need to take into account?

AS: Children have the same growth potential in the first years of life: up to the age of seven actually.  It is true that after, you know, especially beyond puberty, it’s not only nutritional status but other things that become part of the determinants of your height.  So what we’re using is only information up to the age of seven.  Because only during that period can you actually claim that height is a good indicator of nutrition.  And the standard of the World Health Organisation is based on observing healthy children in many different countries, not in only developed countries, but countries in South America and Africa.

AB: it’s pretty standard up to the age of seven?

AS:  Exactly.

AB: What about cognitive skills?  How do you test those without having any cultural bias creeping in?

AS:  In each of our countries – what we observe is multiple indicators of cognitive achievement, such as the scores of a child in a reading comprehension test or in a vocabulary knowledge test or in a maths test.  We have this information so every time we visit families we collect this information.  We use statistical techniques that allow you to extract the implicitability that a child has, based on how well he’s doing in each of the cognitive tests.  And we use a similar procedure to measure social-emotional skills.  So we have information about the aspirations of a child, and the self-esteem, the agency of the child. And this information allows us to extract an indicator of social-emotional skills, which is basically what drives the results of the kids in these tests.

AB: When you say ‘agency’, what does that mean to a lay-person?

AS: Basically it’s the level of mastery you have of your own life.

AB: So something like the capacity to decide what you want to do and then take action to pursue your own goals.

AS: Exactly.  Because it is agency as well as self-esteem that have been found to be very important predictors of labour market outcomes, which is why actually we’re interested in cognitive skills and social-emotional skills. Because they’re going to have an impact on the labour market outcomes of these children eventually.

AB: I’m just interested to know how you go about measuring those sort of things like self-esteem and agency because they sound sort of more subjective.

AS: Yes. Some years ago actually, we started looking for international scales that had been used to estimate these dimensions.  And even though there’s a lot of work behind,  they have been validated for other countries, we thought it was important to just make sure these questions were going to work in the context of our countries. So we did some validations of these instruments in our countries. And so based on that we re-phrased some of the items and we dropped  some of the items and we included other ones.  So that we are sure that these questions that we’re using now can be understood by the children given their context.

AB: Your analysis highlights evidence that height development is important for later height development, but also for later cognitive achievement too. And there’s a link between nutrition and different cognitive skills and qualities like self-esteem and agency.  Can you tell us about what seems to affect what?

AS:  One of the main findings is that early nutrition, at the age of one to two years old can have long-lasting implications for cognitive skills up to the age of 14 to 15.  And even though early nutrition also has an impact on social-emotional skills, the effect on social-emotional skills is not as persistent as it is for cognitive skills. So by the age of 14 to 15, we no longer find the effects of early nutrition on social-emotional skills. But another insight of my research is that even though early nutrition is quite important, it’s not the only aspect that matters.   So for instance nutritional status at later stages also matters. And this implies for instance that if a given child was under-nourished very early in life, that’s going to have consequences.  But, if the child can recover from a nutritional point of view then some of the negative effects can be reverted. So what this means is that, you know, the early childhood stage is important.  That’s what we find.  But it’s not the only period of childhood that matters and you can make a difference even at later stages.

AB: That it is somehow possible to have some recompense later for some children…

AS: Exactly.

AB:  All children?

AS  What we find is that in most of the cases you’re still able to do something.

AB: Can you be certain that the nutrition deficit is the cause of the cognitive problems later, or is it possible that both of them could be caused by the same thing  like perhaps just being raised in a family that’s struggling a lot to make ends meet?

AS:  Yes. Well that’s actually a very important question.  In one of our previous studies for instance we used information on the food price crisis and also on other household level shocks that affected the family, such as job loss or climatic events, as sources of variation in the economy of the family that in turn have effects on the nutritional status of the children. So what this means is that you need information about shocks that effect the household.  The household cannot do anything about it and this has consequences for the children at a nutritional level.

So in this example we have estimated the nutritional status of the kid, which is measured by the height of the kid.  We observed that household shocks affect the nutritional status and this in turn affects the development of skills.   So in such a way, you can be sure it is not just that there are some families who are better at raising kids with better health and better educational outcomes, but something else that the family cannot do anything about is affecting them. And this has nutritional affects and in the long term affects on skills.

AB: So really the existence of the food crisis helped you be sure that you eliminate the possibility that families who are struggling, and cause children to be both under-nourished and cognitively behind?

AS: Exactly.  In our statistical model we also control for the correct statistics for the household.  But eventually what you do in these cases is you use information on natural experiments that take place in each country.  And so the shocks that we have found that are important are country-specific.  For instance in the case of Peru, the most important shocks that takes place and can affect the household economy – therefore the nutritional status of the child – are low-temperature shocks in the highlands when the temperature goes below historical levels. That actually creates a problem for the households in terms of agricultural outcomes.  It can kill the crop basically –  and it also kills the livestock. So we use these natural experiments to make sure that the variation in nutritional status that we’re observing is not driven by household characteristics, but by something else.

AB: What seemed to be the most important links that you discovered between nutrition and later development?

AS: So I think the most important findings of ours is that early nutrition is a predictor of cognitive and social-emotional skills.  But on the other hand, the long-term implications are only observed for cognitive outcomes. Whereas in the case of social-emotional skills the effect is not as persistent, which is probably because there are other factors affecting social-emotional skills besides nutrition.  So that’s one important finding. But at the same time the result of the model suggest that even though early nutrition matters, nutritional status at later stages also matters as well.

AB: What did your analysis suggest about what ages are particularly important for maybe policies aiming to improve children’s later physical development?

AS: To a certain extent we’re constrained by the characteristics of our sample, so we observe the children at the age of one to two years and then at the age of four to five and then at the age of seven to eight and then later at the age of eleven to twelve and fourteen to fifteen.  So we only have tools to measure nutritional status properly up to the age of seven to eight. So within this period we find that you know between the age of one and eight it is possible to make a difference.   At least within my research I haven’t been able to obtain a result saying that whatever you do at the age of one is not going to be able to be fixed later within this period.

AB: So you’ve got another round of data-gathering coming up and that will go up to children that are a bit older?  Eighteen year-olds or something? Are you planning to do any further research?

AS: Yes, one of the objectives of the study is to understand the consequences of childhood poverty. And for that it will be important to see what’s going to happen with these families and these children that are going to become adults very soon: in the labour market or making a decision about higher education. Because that’s going to be the result of a process that took place at the household level and at the community level. So we are keen to be able to analyse the data that’s coming in the next round.

AB:  So by that time you’ll be able to find out how 18 year-olds might have been affected by their earlier experiences as small children?

AS: Exactly. We have two cohorts. So the young cohort is going to be between 14 and 15 years old in the next round of data collection.  But we also have the older cohort and the older cohort is going to be between 21 and 22 by the time we visit them again. So we’re using data from both cohorts  to look at the consequences of childhood poverty at the age of 18 to 19 and in the next round up to the age of 21 to 22.

AB: What kind of things do you think you’ll be looking for in that sort of age-group, when they’re in their 20s?

AS:  Well, it is country-specific. For instance, anecdotally, we don’t have the official numbers but in the older cohort there are many girls (or women by now) that have children.  A large proportion of the sample have children by now. Something similar, although of less magnitude, is observed in Peru. This gives us the opportunity to study the intergenerational transmission of poverty, because we were tracking them since they were very very small  (in the case of the older cohort, since the age of seven). So we have all the history of these families. We can really have a very clear understanding of their decision-making. Why is it that they are becoming parents so young? And then in relation to that, it is the decision between going to higher education or starting to work, or doing a combination of both work and study. So the decision of time allocation during this period of life is going to be the consequences of processes that took place before.

AB In your recent study there seems to be a difference between some of the country data. Vietnam stood out a little bit. Can you say  why you think that might have been?

AS:  Two of our countries: Peru and Vietnam are doing a bit better than the other two countries in terms of the economic growth that’s happened during the last few years.  And also the income per capita is higher, especially in Peru.  And also, the Peruvian sample, Peru in general has higher levels of organisation than the other countries. So what this means, for instance, is when you try to look for relationships there are what we call complementary assets that matter. So for instance, if you think of the impact of a nutritional programme, the impact of a nutritional programme on cognitive outcomes, because there is a link between the two, can be amplified.

AB: What do you mean by nutritional programme?

AS: In Peru we have the Juntos programme, which is a cash conditional transfer programme.  Families receive money on the condition that they will send their children to the health post and then they will receive plenty of information about what type of nutrition that these children should have. So you know, for example you can implement the same programme in two different countries, but then the effects of the programme are going to be affected by the characteristics of the families.  So, if in Peru the average level of education of the mother is higher compared to another country, then potentially the effects of these programme on the nutritional level and on the cognitive level of the child are a little bit higher. And conversely, although I don’t have a specific instance to show you about this, it could be the case that you have a very good programme, but the level of education of the mothers is so small that it actually makes it very difficult for the programme to have a real effect on children.

AB: So really it’s a combination of what sort of government interventions are available, but also the capacity of the population to benefit from those programmes.

AS: And also the characteristics of the communities. If there is no proper sanitation for instance, then a programme such as Juntos (this cash conditional programme that I was mentioning), might not have an effect because everything you try to do is then negatively affected by the fact that there is not enough infrastructure in a community.

AB: there have already been some studies about the link between nutrition and cognitive skills.What do you think your longitudinal data collection has allowed you to find out that previous studies haven’t been able to?

AS: One thing that we’re allowed to see is not only the short-term but the medium-term and the long-term impact of different specific programmes, for instance, in each of the countries.  But then, one of the strengths of the study is that we’re looking at four countries.  And these countries are all different, you know. But if you are able to observe that a certain input in the production of cognitive skills – and also social-emotional skills – is important in all four countries then it means that the relationship you have observed is a ‘regularity’. Regardless of the context this input matters.  On the other hand, you might be able to observe that some programmes, or some aspects of the family, are more or less important, depending on the context – then we can make a claim of external validity, which is that whenever you observe something in one country, it might not work in the other country.

We have the chance to do the same type of analysis in four very different countries. So if it works in one country only, then you need to understand why it’s working in this country and not in the other countries.  And if it works in all four countries, then what you are probably finding is a regularity.

AB: So you have found some really interesting things haven’t you.  What do you think are the implications practitioners and policy-makers can take on board from your findings?

AS: A particular implication is that the early childhood period is a very important period – probably critical for investing in the child.

But that doesn’t mean that you should put all your money there. The process of accumulation of skills takes place during early childhood, mid-childhood, late childhood and at adolescence, and even up to early adulthood. So the implication is you should also be looking at what happens in later periods because there is scope for intervention later on. And there is scope to reverse some of the bad things that might have happened in earlier life.

AB:  And there’s often concern that policy delivery is very siloed, with education specialists looking at schooling issues and nutrition being seen as to do with health policy.  What do your findings suggest for those debates?

AS:  Well actually there is an obvious connection and you can think of nutritional interventions as educational interventions.

The more you invest in nutrition then the more likely it is that the kid is going to do well at school. For instance, just to give you an example, in one of my studies we found that early nutrition was affecting cognitive development during the pre-school period. So these kids – most of them at least – were not attending school yet.  But still the undernourished kids were already disadvantaged and it was before being at school. So you know, if you can make sure that these kids are having adequate nutrition then you are giving them a better start at school. Of course it’s not the only thing that matters. After that is when you really have to start focusing on specific educational policies.  But you can see nutritional policies as part of the solution, let’s say.

AB: People at Young Lives have been starting to talk about something called ‘a life course approach’.  What does that actually mean in terms of policy priorities?

AS: One concept that we use is the life cycle approach.  It is basically the idea that there are different stages in the formation of human capital.

When I speak about ‘skills’, I am also speaking about the human capital that these children are going to be able to bring to the labour market. But you know, you need to understand that there are different stages and each stage has its own characteristics. So you should be designing specific policies for each of these stages. I think that’s one important implication. So for instance, right now we have good data for children observed during adolescence. And during adolescence an important dimension of development is risky behavior.  So you know,  you were not interested in this in earlier periods but now they are in adolescence you have to be interested in risky behaviors.  And the policies that have to be designed are specific to that age-group therefore.

AB: So is that going to be one of the next stages of some of the research?

AS: Yes, we have already collected data on risky behaviors in the older cohort: at the age of between 14 and 15 and now at the age of 18 to 19.  The young cohort they’re still too young to be assessed on that, but for the older cohort we already have data collecting in four countries and we are looking forward to start doing research.

AB: Well that’s another story. Alan Sanchez, thank you very much.

Pod Academy has a number of podcasts based on the Young Lives Research – looking at some of the studies ongoing findings: