>> So let me start off by saying that what I would be talking about today is not just my work but is a work that I’ve been doing with many people and some of them are listed here People here at UCL but also in other places around the world You notice I’ve been– this is a topic, the one of every childhood development which is extremely fashionable and important And so there are lots of people that– that are working on it now and– so let me first start by motivating the– our interest in this topic So there’s more and more evidence that what happened to children in early years is critical in shaping the development of the child and what happens in early years has long lasting effects on a variety of outcomes in adult lives that varies from social skills and emotional well being to depression, participation in criminal activities, success in school, academic achievement, cognitive development, long run health outcomes, and earnings and economic well being Just to give you an example, I could have filled the lecture with examples of this sort of relationships This is a graph taken from a paper by Janet Currie published last year in the American Economic Review, where the top part of the graph relates an index of cognitive development to– at age 6 to earnings of individuals at age 24, 27 And you can see there is a very strong relationship in cognitive development even at age 6, as early as age 6 is a good predictor And you can go even earlier You can relate birth weight, which is well known to be infants by what type of nutrition the child gets in– during pregnancy to earnings 25 years later So what happens in early– in the first part of the life cycle, it is gonna be extremely important And it’s not just cognition that is important as well non-cognitive– what we call non-cognitive skills are also very important, have shown to the important Many of you might have seen the– might know about the marshmallow experiment that was conducted 30 years ago at Stanford You know, you find these beautiful videos on YouTube where you see these children that were given a marshmallow and if they were willing to wait for 10 minutes, they were given a second one And so it turned out that the behavior of those children age 3 to 4, in terms of their ability to self-control, was a very good predictor of their academic achievement and eventually of their earnings in adult life It’s a very impressive piece of evidence We also know that cognitive and non-cognitive skills develop at different points in time and moreover, they are affected to a very large extent by the environment where the individuals live In a way there– I think there is a growing consensus that they all debate about in nature versus nurture, is quite obsolete And even genes find different ways to express themselves depending on the environment where they can do that We also know that the– there are important interactions between cognitive and non-cognitive skills in terms of the– during their childhood And as the skills develop they interact in important ways with each other Now all this is consistent with a bunch of recent neurological research that indicates that the brain develops very early in life Moreover, different parts of the brain that control different functions develop at different points in time And the development, the development of the brain is a function of the quality and range of early experiences interactions And this is a graph that shows the– how different parts, different functions for instance, seeing and hearing develops in the very early months while receptive language and speech develops later And of course, if the early functions have developed well,

it is easier for functions that develop later to grow in an efficient way Now, if this is true for everybody this type of problems are surely extremely important in developing countries where risk factors that they might affect negatively the development of children And therefore have an influence for the rest of their lives are particularly salient According to an important set of papers published in The Lancet in 2007, there are about 200 million children around the world that are at risk of not developing– developing their full potential and here is a map where– about where these children are located and the development– that the– but the poor countries are where you can find it In these countries, there are many risk factors that might affect the development of children They range from nutrition and value factors and the nutrition during pregnancy and early childhood; iron deficiency, iodine deficiency, and in general deficits in various micronutrients In many of these countries, infectious diseases and parasites are very prevalent There are other environmental factors like the lack of clean water hygiene and general health conditions might be problematic And there are psychosocial factors and when you live in poverty, you are likely to be stimulated Parents might be thinking about other pressing issues like survivor– survival, and therefore might give less care to children There’s the violence, there is maternal depression, these are all risk factors that had been shown to have a negative impact on the development of children So poverty is associated with the lack of development and the effects of poverty on child development are visible and they are dramatic And we know that, you know, this lags that are accumulating in early year as I’ve said, they will be difficult to fill in later years This is some evidence from Ecuador, where the development in receptive language skills or receptive children is plotted over their age and the children are divided according to their– the wealth decile of the family where they live And so a score of a hundred, these are scores that are normalized by age, so a score of a hundred is the score of a normal child, what you expect a child to develop at that particular age So normally children should grow so to keep their development in around a hundred, as they age And this one happens to– to children in the forth decile and probably they’re not plotted here, the children living in wealthy families But look what happens to the poorest children, the bottom 10 percent At the beginning, at 36 months where these observations started, they are not particularly delayed There is a little bit of a delay but is no dramatic By age 6 or by age 7, they fell off the scale There is a three standard deviations gap between the– within the poorest and the relatively middle children, this is the forth decile This is equivalent to 2 and half years delay in development These are the children that enter school so you put a– in a classroom with normal 6 years olds, working on a curriculum designed for 6 years old, a child who has the development in terms of their– of his or her language skills of a 3 year old, that child is not gonna be successful And this is the beginning of a life of misery So these are really important problems It is not just– this is true of Ecuador, up in the Andes We’ve been with Sally Grantham-McGregor and [inaudible] and others We have been looking at example of Bangladeshi children, these are all poor children and we show that even among those children that are all poor, there are differences in wealth and you can find differences in the development– cognitive development of the children at very early ages We can find some differences in seven months And by the time the children are 5 years of age, those differences have increased tremendously in terms of their cognitive development Moreover, which is particularly important, a large fraction, not all of it, but a large fraction of this delay seems

to be related to stimulation or in this case the lack of stimulation >> We have measures so far of the quality of the environment and simulations that they received Something similar in Chile, so this is a middle-income country, in one work that I’ve been doing with one of our graduate students of this year, Pamela Jervis [phonetic] on a sample of Chilean children, we find that there are important difference in the cognitive development of children and they grow with age by– between 6 and 24 months of age, is there almost 0.4 standard deviations that– more than doubles by– for children aged 24 to 60 months And once again, the– a big chunk of this delay seems to disappear once you control for– the inputs that the parents may give in terms of stimulation and in terms of other books, stories, et cetera that a large chunk of this gap is associated I’m not saying that this is consult but there seems to be an association there Early years are therefore very important in the long run They are affected by the environment, they are affected by poverty Just in the same way we see that they can be affected by variety of risk factors It means that their outcomes in early years are malleable and the implication is that they might be possible to design interventions that can remedy these delays and can affect the outcomes in early years and as a consequence have long run effects And there’s mounting evidence that this type of interventions, the if well-designs, if successful, not only do they affect outcomes in early years but they have longer effects Let me just mention a few examples It was a few years ago in intervention in Guatemala where a nutritional supplement reinforced with iron and other micronutrients were– was supplied in a randomly selected set of villages for two years And these children were followed for over 30 years Now I don’t wanna go into the details of this table We don’t have enough time to go this But just look at the differences in wages between the children that received that treatment between say, zero and three years which are the first two rows and the children that did not There is a very strong and significant difference in the wage that they were able to command once they grew up in our– 35 years later Interestingly, the effect is much lower, is 0.2 rather than 0.6 for the children that received this intervention within 3 and 5 years, 3 and 6 years of age So yes, the early years interventions can be affected in the long run, but has to be early Three years, after three years it might be already too late Now this is a– this is a nutrition intervention in the Guatemalan villages We can move to Michigan in the U.S And there was a famous experiment that was done in the ’60s, where a number of high risk children age between 3 and 4, were assigned to high quality pre school program And then they were followed into adulthood These were the results of this experiment So the blue bars refer to the outcomes for children that received the intervention, the yellow ones are those that did not So you can find effects, you know, many years later on almost anything you look at Number of percentage of youth arrested more than 5 times by age 40, 36 percent versus 55 percent Percentage that earns more than 20,000 dollars at age 40 is sixty percent versus 40 percent Graduated from high school, 65 versus 45, basic achievement at 14, 49 versus 15, IQ of more than 90 at age 5, 67 versus 28 percent So this is an intervention that seems that been extremely successful An interesting question is how costly is this intervention? Is it worthwhile? Does it cost a fortune? Well the cost was in 2,000 dollars worth

about 15,000 dollars per year, per child so it seems a fair amount On the other hand if you look at the benefits, here is what you save in special educations, 7,000 dollars What you make– what the government means by taxing the increased earnings that these children will earn is 14,000 dollars The welfare savings– welfare payment savings, another 2700 dollars and especially the savings in terms of crime, the incidence of crime can make this program extremely cost-effective Now stimulation and nutrition interact and there is evidence of that In the 1980s, there was a– a famous study was started in Jamaica and Sally Grantham-McGregor is sitting here in this room And her team in Kingston, Jamaica identified 129 children, aged between 9 and 24 months, so early And they developed a structured home-visitation curriculum to promote cognitive simulation as well as mother-child bonds And the curriculum was delivered by community nurses, nurses’ aides through a one hour weekly visits to children’s homes during 24 months So this is probably intensive, intensive intervention Now, the stunted children were randomly divided into 4 groups One group receives stimulation, another group received nutrition, another group received both, and then there was a control group And the children were followed at the end of the intervention at age 7-8, at age 11-12, at age 17-18 and lastly at age 22-23 The results were pretty stunning This is from the first observation You can see here, these were the control children You can see that delay that this has accumulating relative to the now stunted children These are the children that received the supplementation, these are the children that received the stimulation, these are the children that received both They almost catch up But this is not the end of the story By age 17, while the effect of the nutrition intervention has faded out the effect of the stimulation is still there These are the non-stimulated children, these are the stimulated children They don’t quite make up with all the way to the now stunted children but you can fill half the gap This is 15 years after the intervention was finished And this is 22 years after the intervention This is a very recent study that I don’t believe is published But you can– but the results are being made available This is the exhibition of earnings in the current job and you can see that the green line are the stimulated children, the blue– the orange density are the non-stimulated children and you can see a shift in terms of earnings How large is it? Probability of being employed 22 years after the end of the intervention is 15 percentage points higher for the stimulated children The earnings can be as much as 50 percent higher So these affects are– are amazing So before I go into what I think is gonna be the center of my lecture, what I want to tell you what we do, let me summarize what we think we have learned and this first few slides that I showed you were a bit of a summary We learned that development during the first three years sets the stage for long-term development Human capital is a multidimensional [inaudible] these different dimensions are all important, both cognitive and non-cognitive skills are important Nutrition is important to develop But stimulation is probably even more important and more effective in the long run The physical and social home environment is central to development of cognitive and non-cognitive skills and moreover there are important interactions Children that develop early on non-cognitive skills like self-restrain, ability of social communication, executive functions of various types, they might be able to better develop cognitive skills later on, for instance when they go to school However, there’s lots of stuff that we still don’t know We still don’t know– we still don’t fully understand how nutrition affects the development of cognitive and non-cognitive skills and I could give you lots of examples I’ve taken two examples One is about iron and medicine We know that anemia is bad for young children because iron plays a– seems to be playing an important role

in the development of the brain during early childhood >> There’s also some evidence that zinc is protective against a number of diseases So, you think, okay, zinc should also have an effect Well, there is at least one study that shows that zinc has a negative effect on cognitive development maybe because it counteracts the ability of the body to absorb iron Now, I don’t of know this and the problem is that I’m not the only one not to know this stuff There are still lots that need to be learned Iron, so iron seems to be good for children but we are not sure whether it’s a good idea to give iron in areas where malaria is endemic There has been a famous study in Tanzania that showed that they might have very negative effects So there’re still lots that need to be learned even in terms of the Biology of these issues We still don’t know the role played by different inputs and their interactions How does nutrition interacts with stimulation for instance We still don’t know what determines investment in human capital This is where economists like myself, might have an important role to play We need to understand what parents actually do? What are the constraints that poor parents face? Our problem– simply problem of resources or there is also information issues, knowledge, beliefs, attitudes, what role do they play? And more– and importantly I think how resources are allocated within the family is still not completely clear We have some idea that the mothers are better than fathers in caring about children but we don’t fully understand what determines all this And what we don’t know, we know that it is possible to construct effective intervention What we need to know is how– cost-effective interventions, how to go from efficacy to effectiveness? And this is important because when you go to a politician you say, “Look, this is in the long run, this is gonna be very beneficial We spend 58,000 dollars now, in 25 years, we recuperate 150,000 dollars, one to ten And that sounds a really great investment.” Well, unfortunately most politicians don’t have such a long horizon and they will want to have benefits now or to have really low cost So it is crucial to build interventions that are cost-effective And we may develop in the case of the developing countries, there are budgetary issues Simply, the money is not there So, what is our research agenda? In terms of the work that we have been doing in this area, I think we have two main purposes One is to try to understand this mechanism, put a formation of [inaudible] and accounting skills and what motivates and what determines the decisions that parent make, parents make about investing in their children And then we want to identify cost of effective stimulation and nutrition intervention that can be scaled up and sustained So, over the last three years, we have developed a new intervention in Columbia in collaboration with Sally, we have adopted her Jamaican curriculum that I mentioned earlier and we are trying to deliberate using community-based resources So we want to develop a program that can be scaled up If we do show impacts, we want to be able to go to the Colombian government and say, “These are the impacts and this is how much it costs and you can afford it and that– for the– this should be scaled up.” This research has been financed by a patch work of organizations and grants So let me give you a little bit of a background to show you how we tried to do this So in Columbia like in many other Latin-American countries, they adopt this Conditional Cash Transfer Programs The Colombian one is called Familias en Accion, one of the most famous one, the Mexican, one of those called Progressa The idea is quite simple They give– they pay poor people if they comply with certain conditions that typically have to do with the– with investment in human capital in children So they have to take their children to growth and development check ups in health centers, they have to enroll in to school, they have to have them vaccinated, et cetera So these are all fashionable now and throughout Latin-American and many other countries as well So Familias en Accion in Columbia is now the largest welfare program in their country So now, the beneficiaries of this program which account for about 20 percent of the poorest families in Columbia Within a neighborhood or within a village, they elect their representative which is called the Madre Lider And so the mandate of this Madre Lider

which are elected every three years is to connect with the administration of the program and organize a number of activities Now, I worked on the evaluation of this particular program One thing that we realized at the– at a very early stage is that Madre Liders are different These are women that are very distinct, very distinguishable from the other women in the community for their leadership skills, for their education, for their entrepreneur– they are more entrepreneurial, they have more community networking abilities And so the idea we have is that these skills can be used These are smarter women, these are entrepreneurial women and they live in the communities and we might use those skills So the idea is to use these human resources available in the communities And so we train and hire the Madre Liders so they deliver the stimulation curriculum that we adopted from the Jamaican curriculum through weekly home visits Using local resources is not only cheaper but it can mobilize local communities and spread information on child development At the same time, the problem introduced in new institution and improves knowledge and understanding of child education Now, to train the Madre Liders, we also need to hire and train some psychologist who eventually become supervisors and support staff for the Madre Liders And so we have to set up an entire– an entire structure So this is a– the program, the intervention itself is made of a number of components, I’ve put some pictures here to the curriculum The curriculum is extremely well-structured Every week has got its own activity This is a page from the book, from the curriculum so it’s specific So this is month 32, first week There’s a– every week there’s a different set of activities and tells what are the games that they’re supposed to play, what are the activities they are suppose to do, there are instructions These are the some of the puzzles we do The first one on the left is the easiest puzzle The child simply has to– this comes out, the round part of the flower, has to fill it in The clown has got the square that comes out so it’s a little bit more difficult The house is quite difficult because there are all sort the pieces that they have to put together So, as the child grows, you make available harder and harder puzzles We have story cards So these, you know, a coffee growing and you know, then let’s put into sacks then it’s taken to the market and eventually have a nice cup of coffee, the butterfly story So we provide this to the mothers to– and encourage the mothers to interact with the child to form these bonds, to use words, to use language, to develop skills in the child We produce some little books This is one of the stories, you know, there is a dirty shirt that the mother washes and then dries So, simple stories that are appropriate to their culture and to their language An important and very successful enterprise is how to build toys with stuff that is around the house So we use plastic bottles, colors, rags, also the things, these are example of toys that the mothers eventually build and give to the children So we also have as in the original Jamaica study, we have a nutrition part We don’t use calories, we have micronutrients These are sprinkles so these are little satchels that contain little powder which has irons, zinc, vitamin A, and vitamin C. It’s one a day You put it on the food, there’s no taste, no color and provides all the important nutrients This is– one of the reasons why this research has been so much fun is that you have to learn all sorts of skills We bought this stuff in India and we have to import it in Columbia so we find ourselves importing some of powder in to Columbia [ Laughter ] >> They have to train– – this is a picture of the training of the Madre Liders that we do for a period of three weeks before they start with visits This is a section In each town we train three of them This is an example of a visit If you’re in a household, so this is the Madre Lider, this is the mother, she’s got her child, this is the young child on her lap, there are the siblings also being involved in the visit This is the same Madre Lider in a different environment with a different child in a house, look here, this is a story card that she’s using

>> That’s– that’s yet another one So how do we evaluate this? We are implementing– we have been implementing this intervention, how would you evaluate it? So we identify 96 small towns in Columbia in three different geographic areas and we randomly allocated them into 4 intervention groups One that receives only home visits, then home visit plus nutrition, nutrition only, and control, total we have 1400 children at the age– our baseline in January 2010, they were between 20– between 12 and 24 months and this intervention has been going on for 18 months And then we measure There’s a child being– we measure in height in this particular case but we measure an awful lot of stuff and that’s probably one of the most expensive parts of our study We will measure motor and cognitive development in children through the babies’ test, we measure the social emotional development, we measure language development, nutritional status, height, weight, hemoglobin and morbidity, food intakes, child care and arrangements and time use We have a mother interview where we collect lots of information both socioeconomic stand of variables but also aversion to inequality and to risk, depression, knowledge in parenting, parenting practices and home environment and so on and so forth So we have hours and hours of interviews that we’ve been developing And then we have information on the Madre Lider General characteristics, we have the logs of the visits, we have information collected by the supervisors on observing the quality of the visit and we have been running some focus groups to get some more qualitative information So where do we stand? We’re almost finished with the follow-up data so the intervention is now finished And we are collecting the data in the 96 surveys, we are in the middle of it, we should be done by end of the month or early December I believe So we know the results in the next few months and we already internalized the baseline survey so we know how these children were at baseline Now the– one surprising thing is, these are again the index of social development and again the normal– hundred is about normal And here I plot– so our children as I said, at the beginning of the intervention were between 12 and 24 months So at 12 months, they were not in a bad shape This is cognitive development, this is language development But then you can see that, you know, as you look at older and older children, they’re starting to accumulate delays And so this– our challenge is, can we bring this line up? Can we make up for this? Can we avoid this dropping off? And that’s the challenge and we will soon know whether this has been a success or not And now honestly, I don’t know I mean we have– I could tell you lots of anecdotes but that’s not how science is made We need to collect the data, and we need to show whether we– work we’ve done has been effective or not Now if we do have an impact, we know that this can be scaled up on a national level One temptation that we resisted was to put lots of resources, money into this thing We wanted to build something that could be replicated To us and the intervention cost about 500 dollars per child per year Now this is partly because there’s a large cost of supervision which when you go to scale, you can have economies of scale We– the moment we have in each town, three Madre Liders, about 15 children Probably a supervisor can deal with more than that, maybe 10, 15 Madre Liders rather than three and so you can economize on the supervision cost We calculated that we run this thing– a scale with may be 300-400 dollars per child per year This is affordable in a country like Columbia, in middle income country is affordable and is indeed is the cost of some of the interventions that they already are running So if we do have an impact, we can go to the government and say, “Hey, this is something you can do and this is gonna have an impact.” And we will also be able to use– the reason why we’ve been measuring so much is that we want to be able to model what’s happening Why– if we do have an impact, where does this impact come from? Is the attitude of the mother that changes? Is what she does with the child? Is the involvement of the fathers? Whatever. We want to know that So the future agenda, this is my last slide– slide We are now building new interventions in other parts of the world We are running a small pilot in the urban slums of Sambalpur which is a city in Orissa in India which has been financed by generous UCL Alumnus And we are building on that pilot

to have a really large project in rural Orissa, again in India, where we applied for a very large grant which we might be able to get In both cases, the visits will be delivered by local community organizers working with a large micro-finance institution Once again, the idea we have is that we want to involve community, people living in the community, human resources available in the community so that not only you deliver an effective intervention cheaply but also the community themselves take ownership of this type of interventions In the second case, in the case of the large pilot, we’ll also pilot group rather individual visits which is another interesting way to save money So far, we’ve been working with children up to the age of three where in the future we are thinking of working with other interventions after the age of 5 but that’s for the future And finally, let me conclude by saying all these measurement is really important And as a scientific community I think there is still an enormous amount of work that needs to be done [inaudible] building measurement tools that can be effectively used to assess the effectiveness of our interventions Thank you very much [ Applause ] [ Inaudible Remark ] [ Noise ] >> Oh yeah, thanks The randomization, was that done at the village level? >> Sorry? At the village level, yes >> You randomized at the village level >> Yeah >> How do you motivate the control villages? Those parents– the parents and the children in the control villages, how do you motivate them to go through a very extensive– you have a lot of different exams that you put them through to look at many different aspects, how do you prevent them from becoming attrition? >> We have been collaborating with the welfare program that– to which they are all associated because even in the control villages, the welfare program, the families that’s in program is operating and we’ve been working with the officials of the program so we didn’t have– we haven’t had much problems in the control village in terms of willingness to be interviewed and the like Of course they don’t know that we are doing the same thing in other– different things in other villages [ Noise ] >> How is this program different to other programs of intervention that the government is already running and how do you interact with the family welfare institutions in Columbia? >> That’s a very good question I don’t have an hour to answer that, I could take an hour to answer that question but it’s really this way, the interaction with the Familias en Accion program has been extremely nice They’ve been very collaborative, they helped us with the logistics because they give out the list of Madre Liders, they help us to contact them and they kind of kept the 96 villages alone so they are not doing anything else because they themselves are trying different things but they are doing it in different other parts We’re also starting to– now there’s a new government since last year in Columbia and we’ve been talking to the new government about what they want to do about child development but– and they’re doing a bunch of things but we haven’t been coordinating much with them [ Inaudible Remark ] >> Hi. In your photograph, you showed a mother leader with a child and it looked like the interaction was directly between the mother leader and the child and I wondered how much you empowered the mothers themselves to learn the skills and do the stimulation and the interaction themselves and how much the focus is around the session and the interaction between the mother leader and the children? >> That’s a– that’s a very good question, and probably Sally would be able to answer much better than I can but there is an emphasis in the curriculum We try to train the mother leader to put emphasis in their action with the mother than herself ’cause the mother has an important role to play After all, after that hour, it’s the mother who’s gonna be with the child So we do try to stress to the mother leader that she has to encourage the mother to participate and actually, not just the mother but possibly other family members Here, this is– this is a good evidence So this is the mother holding the child in her lap and so the mother leader is interacting with both of them and the older siblings are also participating which is nice

But it is a concern educationally, you know, the mother leader takes too much of a role but– >> Well ladies and gentlemen, on behalf of all of you, I thank Professor Attanasio for a valuable and fascinating lecture [ Applause ] >> Thank you very much