OK, great, thank you very much um so thanks to everyone and thanks for uh for uh joining us, um, in this in this event, um, uh, we, we're gonna be discussing today what are the impacts of school closures, um, in, um, on, on, on, on education outcomes, um, which I, I would say it's an extremely, uh. Uh, critical issue for all of us, um, in, um, uh, during this, during this pandemic, I, as, as, as, as you know, we're living probably the worst education crisis, um, of the last 100 years. We have, we have never seen something like this in which we have, um. Extended periods of school closures in in in all across the whole world and at the same time we have deep economic recessions so we have lived in school closures before in one countries have had political upheaval, so or strikes and there has been interruptions in the schooling process that has happened. Uh, we have lived economic recessions with impacts on budgets and in family disposable incomes that has happened as well. Uh, what has never happened is that we have the two things at the same time the school closures and then precisely when you need resources in order to deal right with the, uh, negative consequences, you have deep economic recessions and at the same time this is happening across the whole world, right? So we have a twin shock on education. And if we see it from a financial perspective, I, I would say that we have a triple shock, right, because we will see that public budgets are going to be under pressure, family disposable incomes are going to be under pressure, and given that this is happening across the whole planet, both for rich and poor countries, right, even international aid is going to be under pressure, right? Even countries who would say, look, we want to stick to providing international aid at the level of 0.7% of GNI. Uh, which some, some rich countries do, even if they stick to that, it will be a 0.7 of a lower GNI. So, so on the, on the, uh, on the financial sector we could potentially see a triple shock. So this is, this is really, really, um, once in a lifetime event hopefully. And uh we we really need to need to have a better assessment of what are the potential learning losses and what are the potential impacts on on on the education system in order to decide what is that we need to do in the in the in in the future. And when I say future, I say the next weeks and months, right? I mean the future is propelled to today because we need to react extremely quickly. And let me just mention that I mean countries fortunately have not standing still. Countries are doing stuff in terms to try to mitigate the negative impacts of this crisis. So there's a lot of experiments on remote learning in, in, I would say in all the countries, and we, our records show that about 130 countries have implemented some type of multi-platform. Um, remote learning combining online resources with the smartphones, with radio, with TV, with social networks in order to try to reach as much as as many kids as possible. However, we know that the effectiveness and the depth of those interventions is very terogeneous across countries and with. And and within countries so we really need to have a better understanding of what are the short term and the long term impacts of this uh of of this crisis such that we can also design better right the uh um uh school school systems that uh will have to be much more, um, uh, resilient and that it will have to address the losses that we have seen that we will be seeing, um, as a consequence of this pandemic. So to understand better, those those losses we have, um, um, um. To, uh, the presentation of two very, very interesting studies, um, um, uh, that we're gonna, um, present first, and then we'll have, we'll open for a Q&A. So first we're gonna have Jishnudas, um, who will, uh, present a recent paper of his and co-authors that studies the long-term effects of a devastating earthquake that occurred in northern Pakistan in, in 200 in 2005. Um, uh, Jishnu probably doesn't need a lot of introduction because he has been, uh, with, with the bank for a, for a long time and then now he's, he's, he's in academia but it's just in the other, uh, in the other side of the street, so, um, so he's very well known for, for, and by, by all of us, uh, and, um. And after, uh, after the presentation by Jishnu we'll have, uh, Joe Pedro Severo who is a lead economist in the education global practice in which, uh, and he will present a recent report of his and co-authors that presents the results of simulations considered diff diff different, uh, different, uh, lengths of school closure and different, different levels of mitigation effectiveness. And with those assumptions trying to stimulate then what will be the impacts in terms of in terms of, uh, foregone earnings in terms of learning in terms of uh uh uh years of schooling adjusted by quality so um. We'll go first with uh with Gisnu and the um and the uh to tell us what's what was the experience and the analysis regarding the as a consequence of this uh uh earthquake that occurred in northern Pakistan in 2005 and then we'll go to uh Joo Pedro. So Jishnu, uh, the floor is yours. Uh, thanks, Jaime, and thank you for having me here. Uh, it's always a pleasure to, uh, present to, uh, audience at the World Bank. So this is joint work with uh Tahirind Ravi, who's at Pomona College and was the inaugural Dean, um, of the School of Education at Lahore University of Management Sciences, and Ben Daniels, who is at Georgetown University right now. So, a couple of things, right? I mean, uh, one is, and I don't need to reiterate this with this audience, uh, we know from a large amount of work, uh, especially over the last two decades, that insults in childhood, but by which I mean deprivations that children suffer, and especially early childhood, so the way, uh, people have thought about this is as the 1st 1000 days of life. Can have lasting impacts on adult health, education, and earnings. So, even in disasters where the immediate mortality impacts are among elderly men, as seems to be with the case a little bit with COVID, uh, what we typically find is later on, the impacts are much larger, even in terms of mortality for those who were between 0 and 1000 days at the time of the disaster. What we know less about is, first of all, what are the impacts across a range of different outcomes. So if you look at the literature, uh, typically they'll focus on one or two outcomes in every single disaster. So we don't have a clear picture of, you know, if we look at one important disaster, what is the range of outcomes and what is the range of impacts you could have across adults, children, health, education. The second one is what are the impacts across the full age range for children. So if you think about the way these studies have been done, we typically look in a way this is looking at starlight, right? Uh, we look at adult populations now, and we say, well, let's look at where they were when there was a big earthquake, or there was a, you know, a terrible wind, or there was a locust infestation, or there was, there were floods. And we say, well, 40 years ago, you were in, you know, X district that was flooded. What's happened to your outcomes now? Now, because of various problems with trying to tease out the impacts of those floods versus you being a certain type of person who's in a flood prone area, what they'll typically do is compare. Children in the 0 to 1000 days to those who are older, that's great, and it tells you exactly how, how much larger the impact is for children on 0 to 1000 days, but it doesn't give you an idea of what are the potential impacts across the full age range for children, and we'll try to do that here. Third. When you look at these disasters and you're looking at adult populations now, you're looking at children who were exposed to these events 25 to 50 years ago, you know, at a time when social protection was not fully developed, at the time when, especially in low-income countries, uh, we didn't even believe that they had the fiscal resources to help populations at this point, right? This is very different in COVID. I mean, there is no country that's saying we should lock down, we should have people in disaster, but we will not give any compensation. So, we have been struck by the fact that, look, we need to have some data, some information on is this compensation enough, right? And what we're gonna do is follow the economist's tool of saying, let's look at a case where there was substantial compensation. And then say, even if there was that substantial financial compensation, was it enough, right? So, what we're gonna do is we're gonna focus on a particular disaster, and I should be clear that every disaster is a case study and we can discuss what's special about this one. And we're going to say, what happens 4 years after a big disaster, where, and I'll show you this, households actually received substantial compensation, right? Uh, so, what's the context? It's the Pakistan 2005 earthquake. It was a very large earthquake. It was 7.6 magnitude in the Richter scale. Uh, it was in a mountainous area in northern Pakistan. It left 73,000 to 78,000 dead, more than 80% of houses destroyed, and most physical infrastructure destroyed. It was also the first big earthquake to hit after 1935, so there's had a, there's been a period of 70 years of earthquake calm, and in a sense, it's completely unanticipated, right? I mean, this is one important piece about earthquakes is they're really, they're impossible to predict. And there is no population experience of an earthquake, a big earthquake in this region, at least for the last 70 years. So what do we do? Uh, we started working in November to March 2005 with extended field trips to the area and I worked pretty closely with the World Bank at that point to design a bunch of programs to work to help in these populations. Then we went back four years later, in June 2009. We took a sample of 126 villages in these affected areas, which was 28,297 households. And we visited every single household in these 126 villages. Uh, why did we do that? We wanted to get very detailed information on aid. We wanted to get information on, on mortality, which we knew would be a small, small fraction, and we wanted to get information on all the demographic characteristics of. Households in these areas and at that time itself in the census, we did a bigger, bigger, bigger version of the questionnaire for 8,351 households. Then between July and December, we sampled from these 28,200 households and did a detailed survey, household survey of 2,456. Uh, what are the two key components we're gonna use for children's outcomes? Well, first of all, nutritionists and, uh, economists and, uh, people working in health have been very, very interested in looking at stature or height. Uh, and one of the implications over this time has been insults in childhood that cause children to reach less than their, you know, optimal height can lead to severe deficiencies, deficits in wages, earnings, primarily, it seems, coming from a lower ability to learn during their schooling years. So we measure heights, and we measure heights for children who are older than 3 at the time of the survey. Right? Important to realize that these are constraints in the field rather than any, uh, any, uh, you know, judgment due to the efficacy of the findings. The reason why it's greater than 3 is because less than 3 have to put the children flat on a board, and that is the way funerals are carried out for children in these areas. So that was just something that was not feasible, uh, uh, uh, for our survey. The second piece we have is we have test scores that we tested children at home in English, mathematics, and Urdu for all children who were 7 to 15 at the time of the survey. Remember, 7 to 15 at the time of the survey is between 3, is between 3 and 11 at the time of the earthquake. Why not younger? It's because children in Pakistan and many low-income countries, as, as you know, uh, are fairly behind in terms of their reading abilities. And below 7, we have found you need interactions, uh, on a face to face basis with the child. You can't give them a test even with something simple written like 5 + 2 and expect them to answer it. And that requires a different degree of standardization and training. That we just couldn't achieve, right? And we were, it was taking us 3 days to walk to the villages, you know, enormous time just battling the cold, battling the, the, the, uh, mountainous terrain, and that was something that the survey, we couldn't manage with the survey. So these are the populations that we're going to work with, uh, uh, over here. OK. The main contribution, and, you know, this is precisely why I think Raul's presentation is gonna be so important. Is because we are going to try and convince you that we are estimating causal impacts, right? Why do we think that's really important? Well, for COVID, it's going to be really hard to estimate what the causal impact of COVID is, given that every area is going to be affected, right? I can't compare schools that have closed down in Pakistan to schools that are not closed somewhere else because there's nowhere else where schools are open, right? So one of the big things we did want to provide to the community was a strong sense that we have estimates that we can hang our hats on in terms of what happens in a disaster uh after compensation, right? It turns out, you know, uh, that this is not that easy to do, and let me just give you a couple of ideas why it's not that easy to do. You know, this, so here's an example of a place where this would not work. I've given you a picture of the San Andreas fault line, and this is a picture taken from space, and you can see it's a big fault line that's totally visible from far away, right? A person or a household that chooses to live here has to be different from a household that's choosing to live somewhere here. This is a household that loves risk, right? So one of the big problems is gonna be, well, if people are living or, you know, you think about floods, it's quite clear that people who are forced to live in flood prone areas, especially in low-income countries, are likely to be poorer than people who can live away from flood prone areas, right? So that's just the starting point. As we worked through this paper, we realized that getting a causal impact is quite a lot of work, even in the case of a disaster. And just to give you a sense of the various issues that are tackled in the paper, which there are links to at the end of the presentation, and I'm sure can be shared, uh, one is you have to make sure that you don't have selective exposure. By which I mean you have to make sure that the people who are exposed to your disaster can be thought of as random with respect to the important characteristics relative to people who are not exposed, right? That's not gonna be true for a flood. It is not going to be true for something like um uh uh for something like uh even heavy winds or locust infestations where different populations are going to be differentially exposed. Second, and, and this we found interesting, which is, you know, if you look at a disaster like a hurricane, it's anticipated. You start getting forecasts two days in advance saying a hurricane is coming, get out, get out, get out. So then, the people who choose to get out are going to be very different from the people who choose to stay behind. So you need a disaster that's also unanticipated, right? On top of that, you need a disaster where mortality, destruction may be high, but mortality needs to be low, right? There's a wonderful paper by Elizabeth Frankenberg and Duncan Thomas showing that the people who died in the tsunami were different from the people who survived because primarily the people who survived were stronger, tend to to be more male, and they lived in different places, right? The mortality numbers for those living along the beach where the mortality was highest in the tsunami was 30%. Now it's really, really hard to say, well, if those 30% who were selected had not died, what would have been their outcomes today, right? Finally, uh, and these are problems that can be solved with better survey survey data collection, households move after a disaster, so you need to track where they are in a case like ours where we're coming in afterwards, um, it's really hard to, to figure that out properly and in our case, actually the households didn't move a lot, reasons we can get into later, uh, and. Importantly, what I'll show you also is aid was very well targeted, so we can, we can think about the impact of a disaster followed by aid that is received by the same population, right? Uh, so we will be able to give you what are the causal impacts of a large disaster after which families receive a considerable amount of aid, and I'll show you how much on a bunch of different outcomes. OK, so here's our strategy. Um, we are going to look at this area in Pakistan. Uh, I'm showing various things in this picture. So this, this dotted line is the main fault line that was activated. It's called the Balakot Bhag fault line, right? And all of these clusters are the households that we actually surveyed, OK? One important point I'll show you is this disaster is very localized. So, this distance is about 40 kilometers. OK. Then all these other lines you see are other fault lines in the area, and I want you to note one very important thing. Yes, we have lots of households that are on the activated fault line, but in fact, we have a lot of other households that are living on one fault line or the other, all of which could have been struck. So what we always do in our econometric specifications is we say, look, two households, both of which are living on a fault line, remember these fault lines are invisible and under the ground, they're not like the San Andreas fault line, must be pretty similar, and we do a bunch of checks to show that's the case. It just turns out that by luck this fault line was struck, but this one was not, or this one was not, right? So that's the, that's the main crux of our econometric strategy, uh, in, in, in, in thinking about this, this uh uh whole disaster, and to convince you of one part of it, I'm not going to go over all parts of it because I think that would be overkill, but just to convince you of one part of it, these are all variables that are determined prior to the earthquake. Sorry. So these are all variables that are predetermined, right? So, where do we get them from? We get them either from household censuses in 1998, at the village level, so that's these ones, or we get them from our own surveys, and what are those? Well, things like adult height and female, male and female height. The earthquake can't have affected those. Those are what they were, uh, for adults by the time the earthquake hit, or the primary education of adults, or the the ages of adults, right? And what you find, so we have plotted now these variables that were predetermined against the distance to the fault line, which is our measure of the intensity of the, of the disaster. And what you find is there is absolutely no correlation between any of them, right? All of them are all over the place. Not a single one is significant except electricity in the house. And that comes from one village that is very far away. Our results don't change if we include those. So one very important piece, and we go in the paper through a large number of these checks, shows that there's absolutely on all observable characteristics that we can find, there's just no difference between the households that were affected and the households that were not, OK? Uh, now, here is the local part of the earthquake, that's the second part of our strategy. We need to try and find places that were affected and places that were not affected, something we can't do in COVID properly. So let me just emphasize for you that this is a very localized phenomenon. So it's a big earthquake, and before we started working on this, we didn't realize this, but even with such a big earthquake, all the effects are concentrated in 30 kilometers around the fault line. So what I've drawn here are two things at the back, this histogram that you see is our sample. So it's telling you these were those households that were close to the fault line. These are households that are further away. Right? And what you see, whether we look at home destroyed, whether we look at whether you died, whether we looked at whether village facilities were destroyed, all of these three, there is absolutely, you know, clear evidence that if you were close or right on the fault line, you were, you know, going to be badly hit. But within 30 kilometers, things were pretty much back to not being hit at all, right? And then if you look at people who died after the earthquake, which is not during or after, again there is no relationship with distance to the fault line, and you wouldn't expect one, right? If these two populations are identical, why you should not see. Any difference in deaths after the earthquake, uh, except maybe a little budge due to injuries or something later on, but in general you would not expect to see that and you don't see it, OK? Having established that this is the last part I want to establish before showing you the results. Which is what do we mean by aid coming in after the earthquake. So what we have plotted here is now the same thing. So this histogram that you're seeing is always going to be the same. It's our sample of villages from the fault line. Uh, this now is showing you the total aid that came in. I've broken it up by different buckets, but you should just focus on this total cash aid for now. It's, there was a lot of money that came in. Right? So if you look at places that are right next to the fault line, they received more than 150% of their annual consumption budget, right? So just thinking about it in terms of COVID, I mean, this is saying, you know, you receive 1.5 times your yearly income as cash compensation, right? Admittedly, a lot of this went, some portion of this went to rebuilding houses, but it was not all of it, right? So we need to rethink, you know, you need to calibrate that for COVID a little bit because asset destruction is not that important here. But on the other hand, over here, after the earthquake, there was massive demand for labor, right? Because everything needed to be rebuilt. In COVID, you have a problem that there is no demand for labor at all, right? So that's what I mean by a, by a compensated earthquake. At the same time, you have to realize that this compensation took some time to come, right? So this is pictures I took in December, about 4 months later. And these were the kinds of schools that had been set up. The first schools that started running back up were private schools, and you can see they're basically set up a tent that had been given by the UN or that had been given by another relief agency. Uh, the children are all sitting there, they're trying to do the best that they can, but these are not ideal circumstances by any, any measure, right? Uh, here are houses that are, you know, here's the way people were living, and, uh, you'll have to think about the results in this context. Couple of things I want to point out. Everybody put their tents up on top of their pre of the houses that existed before. So this is this person's house. The reason was they did not want to leave the area because unless they could show the government inspectors that the house had been destroyed, they would not receive the compensation. So in fact, they didn't even clear the debris from their houses until the government inspectors had certified that they had a house that had been destroyed as a result of the earthquake, right? OK, what are the results? Uh, so 4 years later, uh, first of all, and I, I'm not showing this here just for brevity, we find that all adult outcomes are back to normal. We don't find any change in adult weight. And when we look at household, we find no change in per capita consumption, so they're, you know, doing just as well as others. And when you look at housing stock, it's actually better, so more people close to the earthquake are now living in permanent houses, more of them have electricity, so in terms of adult outcomes, things are back to normal. If you look at children's outcomes, and just a quick way on how we're presenting this because I both want to show you ages and distances. In this first graph, here are children's ages during the earthquake, and we've just divided that up by whether you were far from the fault line or, or close to the fault line. You essentially see there was no difference, right? You know, yes, it's true, as you grow older, your weight for age seems to decline, but that decline is the same regardless of whether you're living close or far from the fault line, right? The other way you can see that is you can break up age categories and do distance to the fault line, right? If you had expected that people close by were doing worse, you would have expected something like that, right? That says people close by are very, you know, doing worse in terms of the way people are, are, are doing better. But that's not the case here at all, you know, there's some nonlinear patterns throughout, but basically it's flat, right? And in fact, for greater than 3 years old, it even looks like people, you know, the people close by were doing a little better than the people far away, right? That's not the case at all for height. So now you see very clearly that for height, things are totally different. So once again. Exactly what you would expect from the literature so far, after the age of 3, there's really nothing going on. Whether you're far or close to the fault line, there's no difference between these two groups. Before the age of 3, you see this gap in height developing between those who are close and far from the fault line, and that gap is largest if you were in your terror at the time of the shock, right? Let me, this is again another way of showing the same thing above the age of 3, flat with distance to the fault line, below the age of 3, it's a big impact, uh, as a big, big change difference between whether you're close or far from the fault line. OK, here's the, the, the why, the reason why we're here, we still have 10 minutes to kind of run through this, uh, or 8 minutes. First result, right? When we turn to education. We find no differences in enrollment. And we find no differences in enrollment for girls, for boys, and we find no difference in grade attainment either, right? We think that these are really important results partly because one of the places we're working, Khyber Pakhtuwa, as it is, female enrollment is much lower, but this is evidence saying, look, in this particular case, it's not as if the disaster which did lead to school closures led to children dropping out permanently. OK, neither did it lead to them being held back, and in fact, you know, I would argue that's something we need to talk about because even in COVID, children will not be held back, which means they will proceed to the next class not knowing a lot of what they should have learned in the current class, right? The place where we should see it show up very, very clearly is in test scores. So, importantly for test scores, it doesn't matter what age you were. Right? Whether you were 3 or whether you were 11, there's a little bit of tapering, but it's significant across the age range, and we can't reject that the coefficients are the same. There's a big difference in test scores between children who were far from the fault line, which is these guys, and children who were close to the fault line. In fact, you can see how big this is by just calibrating it, right? So if you look at a child who's 4 and far from the fault line, their test score is here, right? That's identical to a child who is 6.5 and whose test score is at the same level, right? And at every point, if you plot these out, it's about 1.5 to 2 year difference, right? So for every age, they fell behind in their learning by about 2 years. Let me show you a couple more things, uh, I'm not gonna show you this, but let me benchmark this total, total effect. So if you use wage height and wage years of schooling associations from Pakistan, uh, not from other countries, and you assume that these deficits that we're seeing will continue into adult life, then the numbers are quite frightening. So we estimate that at peak losses, so remember these kids will all join the labor market. And other people will exit, so peak losses happen when the youngest is 18. Then the earning losses for these cohorts are of the order of 15 to 18%. Right? And that corresponds to 5% of earnings in the area. Why is it a lot less? Because it's one cohort, you know, there are other people in the labor market as well, right? But if you think about now, these effects are big. It's 15% for an entire cohort. That drop in potential GNI down the road is orders of magnitude higher. Than what we're seeing right now, right? So if this is a problem that we can't correct right now, and interestingly, most of this is due to loss in schooling years, and there's an additional effect of 6% due to loss of height, right? That's the key, key issue that we're trying to highlight that look, even when households receive compensation, it's not going to be enough. We're looking at a disaster 10 years down the road, so this is not a question of just, you know, taking it through this one year. It's a question of thinking far ahead in terms of what's gonna come 10 years from now, right? I want to show you two more results and then stop hopefully on time. One is, um, one of the issues that people have been working really hard on is to say, what do we do during this time that schools are closed, right? Some numbers, uh, so on average schools were closed about 4 months next to the fault line here, right? Important pieces, so these are test scores by age again in the second graph that I'm showing you here, right, we have controlled for the school closures, OK? What we find is two things. One is the more the schools were closed, the more children lost. That's what you would expect. But in fact, think about this simple, simple number which is people were closed, schools were closed for 4 months, but children lost 2 years of learning. So clearly, if all of this came from school closure, it just can't be, right? And when we do this more formally through what we think about what what's usually called mediation analysis, what we show is that at most school closures account for 10% of all losses. Right? Let me just put that out there and we can come back to this in the discussion. Uh, so in other words, time out of school is leveraged 10 times by ongoing home and school factors post resumption of schooling, OK? Let me give you one last slide and then conclude. Uh, I don't want to go over my time. This is the second piece of it, which I suspect that all of us have a good inkling of, uh, just looking at what's happening with our children and others. What we have plotted here is what happened to children whose mothers had some education versus those who didn't, and we've plotted it against that distance to the fault line, right? Uh, here's the important thing, for children whose mothers were educated, you don't see any effect at all. There's nothing, this line is flat. It doesn't matter how far you were from the fault line for your test scores today, right? For children whose mothers were not educated, those are the ones who took the massive hit. Right, those are the ones who when they're close to the fault line, have taken an absolute plummet in their test scores over this time, right? So very, very importantly, this is not just about differences between those who were hit and those who are not. What's really important is this is going to even further increase inequalities that already exist in our systems, right? So let me conclude there by saying look, uh, this is, you know, our, our emphasis here and our duty has been to say what are causal impacts in these large disasters followed by aid, right? Uh, and what we're showing is look, you can get recovery in infrastructure consumption, adult anthropometrics, school enrollment, grade attainment. Right, but those hide deep scars, and those deep scars are coming in children's test scores and height for age. And the academic advantages of children with educated mothers are magnified 2 to 3 times relative to others. Very importantly, this means that look, it's not just the losses they suffer while the schools are closed, these losses are going to build after they resume. Schooling unless you know, and, and you guys will know more about this unless we take that particular measurement and experimentation and putting changes in to deal with the fact that we're gonna have to promote all these kids. While in fact they might now be if they were 1 year behind now, they'll be 2 years behind now, right? So let me conclude there, hopefully I've not gone uh over time, uh, and hand it over to Johuahua, who I think will now benchmark these losses in a much bigger sample to try and get us some idea. What I hope I've convinced you is to take very seriously this, not as a correlation, but this is a true impact. Jaime, uh, uh, uh, uh, let me hand it over. Wonderful. Great, uh, thank you very much, um, just new, um, I, I mean, these, there, there are many, many, this generates many questions and also, um, um, many implications in terms of the type of number of analysis that, that we'll have to do in terms of how will this, uh, pandemic will, will impact, um, both enrollments and learning. But, uh, before going to the questions now we will have the presentation by, uh, Joe Pedro Acevedo. Joe Pedro, the floor is yours. Uh, we, you're muted, uh, Joe Pedro. Yeah Hello. OK, uh, I believe it. OK, thank you very much, uh, Jaime, uh, Gishnu, and, and colleagues. Uh, so yes, I think, uh, this is, uh, an interesting, uh, segue, um, as I think our results, uh. Basically corroborate and and and give scale to to some of the numbers that Jusnu has presented uh for Pakistan. So this work is built on a on a report that was launched uh last June, uh, last month, uh, and is joint work with Amir Hassan, Diana Goldenberg, uh Aroup, uh, Iqbal, and Knevin. And basically, what we, what, uh what I wanna do today in this presentation is to share uh with you, uh, basically, the main aspects of the simulations, the analytical framework that we have used, as well as the methodology, and present some of the. The, the main findings and, and basically some uh concluding remarks uh as how we, we see this uh going forward, right, is, is, is this massive uh shock that our educational systems are, are enduring. Um, and you can find all the details of this work in, in a working paper in a, in a report, and we'll be circulating the slides, I'm sure. So, uh, how do we see the, the use of the simulations, right? I think to, to, to start, I think that it is important to, to highlight that this is not an effort to, to assess the impact of COVID, right? Uh, this is a work that is is completely forward-looking and, and we We see it kind of as, as a way to, to actually enable uh policymakers to have a sense of what are the consequences of inaction or potentially what would be the ex ante assessment of what might happen if uh countries respond and, and give scale to, to some of the responses that are uh basically taking place. Um, so some of the, the elements that we see, uh, this being used for is to assess the, the learning losses due to the school closure and the income shock, the economic shock that, that COVID is, is generating, uh, according to different mitigation assumptions, right? Um, we are representing or we are reporting this, this results in terms of learning adjusted years of schooling or delays measure. Uh, which is an interesting measure because it combines precisely the, the quantity and the quality aspects of education, right, the schooling and the learning. We also have a set of, of simulations in which we're presenting this in terms of PISA scores, which of course as PISA, it's only representative of, of the lower secondary uh cohort of students uh in the population, and, and we definitely want to translate this into life cycle earnings and, and, and some monetary measures and it's interesting because, uh, just to give you a little bit of a, of a, of a uh preview. Uh, our, uh, estimated learning loss as, as a percentage of the earnings, uh, of, of the workers is actually very close to, to what Jixinnu has found. We've, we're coming out at 16% of, of earnings. Lost, right, uh, Juum highlighted 15% for the Pakistan case if I, if I remember, uh, correctly, and believe me this is completely independent, right? But it's reassuring, uh, to, to actually hear that our numbers coming, uh, close to that now, uh, I think it is important to, to highlight that we don't see this work as, as a, as, as. The only uh information that, that, uh, policymakers should use to guide the decision to reopen schools, as a matter of fact, you know, uh, first and foremost, uh schools reopening should be seen as a health, uh, uh measure and policy, and, and should be treated as such, right? But definitely, I think for the debate, uh, we feel that it's a useful, uh, effort and, and a useful resource that, that. You know, uh, governments, uh, donors and other stakeholders can, can make use of in order to, to, to precisely have a sense of, of what are the orders of magnitudes. So, moving on, I think it's, uh, we start by acknowledging that, you know, the world is currently in a, in a massive learning crisis, right? Uh, our results are, are dismal, um. And, and basically, we, we have recently uh seen that the possibility for the country to basically uh reduce learning poverty by 2030 is, would require, uh, basically an unprecedented level of, of, of, of effort, right? So today, about 53% of the children in the world are in learning poverty. Meaning that they're not able to read and understand a simple age appropriate text, uh, and the historical trend will take us to 43% in 15 years, right, which is far from what we, uh, have, uh, committed as, as uh, as, as a, as the world, uh, by the SDGs, right? So, in halving this number, which is the, the recently released kind of World Bank target would imply in itself uh increasing the, the rate of, of improvement to, to close to the 80th percentile of, of the regions. So and that would be, and that's an extremely optimistic kind of effort, right? I mean, we believe it's feasible, but it's still, it's ambitious. And, and we know that COVID will definitely take us off track, right? Uh, COVID will, will definitely slow down our progress and, and it can be, um, significant, right? Um. So this is kind of one of our main motivations for this work is to really make sure that, that we understand the consequences of this, of this shock. Now, in terms of uh our analytical framework and methodologies for the stimulation, and again, I think here as well, we have a nice alignment with what was presented by Tishhinu because we start by uh our departing premise is that. Uh, COVID and, and the school closures will affect learning through two main channels, right? The one is the learning that will not take place when schools are closed, uh, and the second is the learning that will be lost due to, uh, basically forgetting and. That is gonna be uh created due to the school disengagement and the dropouts. So those are, we, we believe that those two channels are extremely important and, and those are two things, two elements that have been uh also discussed by the previous presenter, um. So with this understanding that these two elements will be affecting the learning losses, right? The, the, the learning that will not take place, and the, the forgetting of, of, of learning due to, to the disengagement, we built a simulation model, and here I'm just going to present some of the stylized channels of this, this model, right? And again, more details are available in the, in the report and you're all welcome to. To take a look and, and contact us uh if you think uh can be of interest. Um, so, basically, in terms of the simulation parameters and, and the main outcomes which we, we, we have built this system around. Um, basically, we, we start from the premise that when kids go to school, they learn, right? And, and when they do not go to school, they forget what they have learned, right? Uh, and, and that gets translated into, into some measure of learning gains, which is, uh, P, and let me just highlight the, the pointer here, um. And we have, uh, schools are going to be closed by a certain amount of, of months, so it represented in terms of uh share of the school year. Governments are deploying a mitigation, mitigation strategies, and those mitigation strategies and the effectiveness of this mitigation will be a function of three elements, uh, what the government is supplying, which is the, the G. Um, what is the propensity of the households to absorb, uh, or to access what the government is supplying, which is the access by the families to what is being supplying. As an example, uh, governments might supply online training, but not all households might have internet. Connectivity, right? So, so definitely that's another element that we take into account and the third and last element is the effectiveness of this, uh, remote learning, um, right, which, uh, can be very from 0 to 100%, uh. From a second important channel is gonna be the economic one or the family income, and, and here, um, as we've noticed, right, COVID is, is not just a health crisis, but it's also an economic crisis and household incomes are dropping and our estimates of, of those, of that economic shock is being adjusted by the day. And, and we also take into consideration a second channel which will be uh the income effect of, of uh on dropouts on students' dropouts and, and basically those two things combine the m. You know, the what happens in terms of mitigation and, and what happens to income will get, uh, will get translated into both schooling losses and learning losses, which then can be reported in terms of either learning adjusted years of schooling in the case of our first simulation, um, in the case of the 2nd and 3rd in terms of PISA scores and, and both the PISA, uh, average score as well as the minimum proficiency level of PISA. Um, now, of course, uh, this simulation is leaving certain elements out, uh, such as, uh, what. Which we believe are important, uh, but that we don't have data, we don't have a good grasp of, of those channels yet, such as school safety concerns, right? What we see happening a lot in many countries is that, uh. You know, in conversations about schools reopening, many parents are not feeling comfortable in sending their kids back to school, not because they might not have the income for that, but because they're still not, uh, they, they don't trust that the, that this, the school will be a safe environment for their kids to be at, um, so these are elements that are important. Uh, and, but we just feel that we don't have data today to actually, uh, incorporate them in the model. So we've just, um, wanna be upfront about this. Now, moving into some of the parameters that we are using, so this is a, uh, basically a summary representation of, of the school uh enrollment relationship vis a vis household income levels, and we definitely find um significant uh Between country differences and here we have the, the relationship between uh income level of households and the share of children out of school, right? By income groups and this is uh built using micro households. Survey data from 130 countries and we have here the low income countries, the lower middle income countries, upper middle and high income countries, and for two age groups, the 4 to 11, which would be the primary, and the 12 to 17, which would be our secondary. Uh, cohort and the lines are both for females and males, right? So I think the main takeaway here is that they're important between country differences as we can see for high income countries, especially in primary enrollment, uh, you know, results are completely flat, basically, there's a, you know, uh, pretty much the same share of children out of school across all the different income quintiles, um. And while in, in the low income countries, this range is extremely high and the level is significantly higher than in the high income countries, right, suggesting there that there is both uh between countries significant difference but also within country significant differences on the equality of opportunity of of accessing school, right? Um. Now, on the secondary education for the older children, the 12 to 17 cohort, um, the within-country inequality seems to be, uh, present in, in, in all country groupings, although the, the, the between is definitely different, right? So this is the information that we're taking into consideration as we produce our global estimates. So moving to the, into the mitigation, as I mentioned earlier, so we have these 3 kind of elements of, of the mitigation effectiveness, the, what the government is supplying, what is the access to distance learning, and the effectiveness of this distance learning, and the idea here is that each one of these indicators will vary from 0 to 1. Um, where 0 is, uh, basically, it, it doesn't exist or is of very bad quality. One is that it's perfect, if, if fully comparable to, to, you know, uh. To presential education and, and what it's important for us in this construction is that basically the system or the mitigation effectiveness will be as strong as almost the weakest link, right? So there is a disintegration effect here across this diff these three different elements. It's not enough to be thinking in terms of what the government is supplying. We need to be thinking about what is the ability of the households to, to absorb that, uh, and, and definitely what is the effectiveness. Now, we, we do have data on what governments are doing and, and we are getting better data uh by the day on that. Uh, on access, I think that is still an area for, for the work because, Um, the data that we have comes from household surveys, and I'll come to this in a little bit, um, but there's a lot of, of, uh, access or intra-household, uh, bargaining or inequality that we really don't know, right? Many households. Might have just one computer, uh, and, and that has to be shared across multiple children or even as a, as a, as a tool for the parents to work. So it's not obvious how this intra uh household allocation of these resources are, are, are truly available and, and definitely we need a lot more work on the effectiveness of this, right? And I think this is an area that we definitely need to be extremely deliberate and intentional in, in really pursuing. Um, um, significant, uh, advancement on, on our understanding on, on, on how this, uh, technologies and, and approaches, um, radio, TV, paper-based, that, that are being used to, to precisely to blur, you know, the, the distance, uh, between home and schools, uh, how effective they are truly being, right? Um. So this is an example uh using the Gallup World Survey of 2019 on just mobile phone ownership and internet access, by again for data for 140 countries in the world, um, and we can see very clearly that, Uh, access, uh, on, on these technologies, uh, between countries is, is certainly very different by type of, of technology, right? While in the In the mobile phone, high income countries, uh, 94% of the adult population have a, uh, uh, a mobile phone today, while in low-income countries, only 61%, but, you know, in upper middle or lower middle is close to 80%. So the range exists, uh, but it's, uh, but it's significantly lower than the range on internet access, right? So while in high income countries, 87% of the households. Have uh internet connectivity in low income countries, this number, the average number is down to 24% and as you can see by the range of the bars, there's also significant between countries inequality of those measures, uh, even within these groups of countries, right? But that's definitely not the full story, right? We, uh, that, you know, we have the between country differences on, on access to these things, but it's also important to, to remind ourselves that there are significant within country differences, and this is what the lower panel actually portrays, right? Uh, and, and that this, and this is just presenting the ratio of access to both mobile phone ownership and internet access by the top 60% versus the bottom 40% of the population of the country, right? So if they are equally distributed, this number would be 1, if the top, uh, 60 has uh larger access, this number would be greater than 1. If the bottom 40 in the unlikely scenario have greater access, this number will be lower than 1. And what it's uh remarkable is that technology like mobile phone in the high income countries and even in, in upper middle income countries seem extremely well uh distributed within country, right, even in low middle income countries and, and in low middle, low income countries, you know, there is a certain difference, but it's clearly not as nearly as much as internet connectivity. Right, uh, the ratio of the top 60 with the bottom, uh, 40 is 2 times, and the range is massive, right? Uh, in the low income countries. So even in the situation in which, and, uh, in which, you know, some of these solutions are provided by the countries, right, this is definitely something that is likely to create significant within-country inequalities, right? Give it access. Differences. So, all of this is being used to us, by us to actually inform uh our simulation scenarios, which I'll come to in a, in a bit, um, but in terms of the kind of what are the channels, just to kind of be a little bit more specific, uh, in terms of our first approach or the simulation one, we have basically the learning adjusted years of schooling. Uh, or the impact of, of the school closures on lays being transmitted through both the, the learning or the HLO, the harmonized learning outcome, and the expected years of schooling. In the case of the HLO, the, the main, uh, effects are gonna be school closure, the mitigation effectiveness that we just talked about, and of course, the, the learning gain or the expected school productivity of, of a particular country, right? Um, on the expected years of schooling, uh. Both school closures and mitigations are also going to be important, but we have the dropouts, right, and those dropouts will be a function of the expected economic shock as well as, as though that elasticity that we presented uh a bit earlier. On the case of our second simulation, we, we have that basically learning, uh, scores or learning will be uh affected through both the school closures and the mitigation effectiveness, but in this case, we're able to take into account, uh, basically the heterogeneity of this shock across socioeconomic status, right? Or, or welfare of, of the households. Um, one last, uh, approach that we've also used was precisely to To basically uh translate the average effect into a level effect, meaning the share of learners that are below a particular proficiency level, and, and this is particularly important, uh, if we think about the importance of governments when schools reopen to teach at the right level, right? and. And the minimum proficiency level, it's a, it's an important milestone that governments need to be mindful of because uh it can have significant, uh, detrimental effects on, on the ability of, of children's to, to understand what is being presented if, if the, the teaching is. Not adjusted right, as, as we do have evidence from Banerjee and, and others of, uh, precisely that when we and when students are, are taught at a, at a level that is different from their ability or, or their level uh they learn less, right? So we need to, to be mindful of that, um. Now, of course, we also translate this into life cycle earnings uh using uh basically present value uh calculation and some additional uh assumptions when, when we aggregate this number. I think uh one thing that is important to highlight when we aggregate the number is not only we're taking all the numbers are um only for this particular student cohort that is being affected. By the crisis, so children that are today between 4 and 17 years of age, but um we are also adjusting for the fact that not all of those children will live to age 65, which is the age in which uh they, they would be considered uh of working age, right? So we adjust for the expected survival rate and in addition to that we. Also adjust for the fact that not all this human capital will be fully utilized, meaning that not all these uh students uh of today will be in gainful employment throughout the entire work life, right? So, so keep that in mind when you, you hear our results in terms of monetary losses in aggregate because we are definitely taking into consideration. The fact that, uh, you know, we're not, you know, if you will, uh, overestimating this, we feel that we are, it's more of a conservative, or we're being more on the conservative side of, of our, uh, given these assumptions and these additional adjustments that we made, that we do make. So, in terms of the scenarios, we, we're working with uh 3 scenarios. One is uh optimistic, uh, an intermediate, and a pessimistic scenarios, and we have different assumptions depending on the, the level of income of the country. So here we have some of the main parameters uh in terms of low-income countries, uh, lower, middle, upper middle and high income, and, and those parameters are informed by precisely some of the, you know, the results that we've shared, uh, so far, right, in the presentation in which we, we acknowledge that there are important heterogenized important differences across this, this grouping. Of countries and we, we do the, uh, we do try to take that into account for our global and regional estimates. One thing that is extremely important is that we feel that this, this exercise, it's a useful exercise as a global and regional and, and aggregate, uh, type of estimate, but we are also making this so available to to country teams so they can actually. Replace our parameters uh by country specific parameters which we feel that it are significantly more meaningful uh for a country dialogue, right? uh, in which basically in many cases we'll have better understanding of of and better data to, to inform some of those uh choices that are being made. Um, 11 last important remark is just that the, the macro, uh, shock, this, uh, these average numbers are actually weighted by the student population, right? So this, and not, not the, uh, not the whole population of the country, right? And then, and for that reason, they might differ from, from some other numbers that, uh, some of you might have seen. So now moving to the results, right? So, we have a baseline, uh, scenario, which is, you know, the world as, as it is today, and, and basically, the results reported, uh, according to the three scenarios that I just described, and in the paper, you'll be able to find these results broken down by other subgroups such as regions and, and income levels of countries and even World Bank uh classifications. So in, in our intermediate scenario, we're expecting basically a loss of around 0.6, 0.6 lays, so about half a year of learning adjusted year of schooling, a little bit over half a year of learning. HSC or schooling, which is uh about a 7% uh decline in lays um in the most pessimistic scenario, this number uh might go down all the way to 7 lays uh points which will be an 11% reduction. Now just to put it in, in, in perspective, this is a, uh, a quite significant and, you know, change and, and, and it's also as, as mentioned um by Chinu earlier, right, you cannot. Basically all of this is with a scenario in which schools are closed by about 44 months, right? So schools are closed by 4 months, but actually students are losing more than that 4 months, than those 4 months, right? Precisely because it's not just the, the schooling that is not taking place, but also the forgetting, right? Uh, that is uh an important mechanism in, in, in this uh simulations. Now in terms of PISA scores or PISA points, um, basically we, we're seeing that, uh, countries might, might lose about 4, 4% of PISA points in the intermediate scenario, dropping from 440 to 425 in the pessimistic 413. And in terms of the pizza levels. And there we see a very different story. Uh, I think qualitatively because basically, today we have about 40% of our children, uh, falling below PISA level two, which is the minimum proficiency level of PISA, right? Um, and in our intermediate scenario, this number would go up to 50% or, and that is a 25% increase in the share of, of children falling below minimum proficiency. Now, a lot of those children will be falling just a little bit below the minimum proficiency, right? Um, some of them might be falling significantly more, but this is an important result when we think about the, the remedial, uh, strategy once schools reopen, right, because, uh. We do believe that it is critical for systems to be able to, to teach at the right level and, and the, and, and to adjust teaching material according to that, right? So, uh, a kid that is, let's say 2 or 3 levels below the minimum proficiency threshold. We need a different support than a child that is just one level below or two levels below. So being able to accommodate this heterogeneity is gonna be a critical element of the strategies or the recovery strategies as we go forward. Now, shifting gears a bit and and presenting these numbers now in terms of dollars, right, uh, basically using uh the expected earnings of these countries and, and assuming an 8% returns to education, what we do find is, is basically that in the intermediate scenario, uh. Children would expected to lose by year annually about $872 in. Now if we aggregate this to the entire population and bring it at present value, uh, what we expect to be about 35 years of, of productive work life with the, the adjustments on life expectancy and human capital utilization. This number translates to approximately $10 trillion right? Uh, and the range is a little bit higher, between 4.3 to $16 trillion depending on the, the scenario assumption that we, we do make. Now. I think we do have these results, as I mentioned, uh, broken down by, by some, uh, regions and, and income groups of the countries and, and you are welcome to revisit those. And 11 way that we feel that it's particularly useful to think about these numbers is in terms of the share of investment in education of this student cohort. So if we basically uh multiply the, the annual expenditure or investment in education that governments in the world are making. Right, um, and for this particular student cohort, right, of about 12 years or 13 years of, of time in school. This, uh, loss is equivalent to 16% of this, uh, of this investment, right? Um, so it is a very significant, uh, loss and, and it's a particularly important loss for low-income countries, right? Low-income countries, uh, this loss can represent up to almost 40% of their investment in education, right? Uh, which, uh, is extremely, extremely significant. So some of the, you know, uh, final remarks and, and kind of main takeaways of, of this work, um, and then we can definitely open for, for questions. Um, we definitely have no precedent for a, uh, a pandemic of, of this, uh, magnitude and, and, and here we're very much in uncharted territory. We've informed our simulations as much as, as we could on the base of the literature and, and the best data that we have access to. Uh, but we feel that as we go forward, we'll be able to basically replace some of the assumptions that we have made by better understanding. And, and definitely we need to, to have, I think a very strong kind of uh research and and knowledge agenda to, to learn from this experience, especially because this might not be uh the last time that we might see uh a pandemic um of this nature in, in, in our lifetimes. So systems need to be built to, to become more resilient to this kind of shocks and, and for that, um, we need to learn from, from the current experience. I think, um. It is uh critical to to recognize also that the magnitude of, of this learning losses are going to be uh proportional to how effective educational systems are currently functioning, right? So in the case of countries in which uh children are not, are going to school and not learning. Um, the, the learning losses are likely to be lower, right? Um, now, of course, there are other channels that are very important, such as nutritional and, and violence, right, and we know that those, that schools are, are, you know, serve other purposes, and, and we need to be mindful of that as well, right? There are very important other, you know, losses that are happening beyond learning with school closures, right? We know that violence and, and abuse in children is increasing. We know that uh children are losing uh uh the opportunity to, to have access to, to. To school feeding programs and, and, and those are also important losses let's say that, that are taking place because of the closure of the systems, but those go beyond the scope of this particular work um we, we also learned that the income shocks um are gonna be. Will be affecting older children more than the younger children, and this is also aligned with the idea that the opportunity costs are gonna be different, right? for younger children, uh, children in primary, the opportunity costs of them going to school, uh, tend to be lower than children in, in secondary or older children which might have, uh, other, uh. Competing demands from the families, um, we're also ignoring, um, other channels which we believe are important, but for which we have no data, right, so gender specific channels like uh teenage pregnancy, um, or teenage early marriage, uh, that we feel that are important and can be very relevant, uh, although, uh, with more. Is important regional differences, right? We've learned from the Ebola crisis, that the teenage pregnancy has increased during the, the Ebola outbreak and early marriage as well in Haiti where school systems have been closed since early uh. Earlier last year, um, again, uh, teenage pregnancy seems to be on the rise, so there are evidence that there are other important channels that are going to affect the welfare of, of our students, uh, but those go beyond the, the, the current work. Uh, I think another important element is that this crisis is not over yet, so we know that this is, uh, changing by the, by the date, and, and our, even our number of, of millions of children are out of school has gone up from 1.9 to 6.9 as we revise these numbers. And as we take into consideration the within country distribution effects of these shocks, right, these numbers are likely to go, to become even larger, right? Given that not all uh households have the same ability to, to absorb this shock, right? And, and. So in terms of, I think uh we'll need to, we need to start wrapping up uh very quickly, Joan Pedro, so we have a few 3 minutes for a, for a Q&A, OK, and this is the last slide is basically this is an opportunity for us to, I think, also reimagine the educational system. And, and really kind of uh address the, the learning inequalities that are happening, right? I think that a lot of these platforms that are being built on remote learning uh can be used, I think beyond the, the response of COVID and, and I think that those should be uh strategies to, to actually provide more personalized learning for our students once uh. Uh, schools reopened and beyond, right, and, and those are gonna be critical elements to, to make educational systems more resilient to, to future interruptions, um, and of this nature. So I'll stop here and, and looking for any questions or comments. Thank you very much, uh, Joan Pedro. So we have a few questions from the, um, um, from the, uh, the participants. So I will. Uh, ask him, I'll just go first with, uh, Gish knew a couple of questions, um, uh, for you, and then we go to, uh, Joe Pedro. So, um, so first, um, for Gish new, um. Let me see. So, uh, one question is regarding the Pakistan study. If you're saying that 90% of the effects on learning losses are not direct school closure effects, but it's, but it's regarding to what's happening later. So how much of that 90% of what's happening later is about not adapting the curriculum to the additional, uh, to the, um, initial to the initial losses. Um, if the problem is not teaching at the right level, for instance, one would one expect, wouldn't one expect differences in the impacts by age, um, um. Uh, or, or by school grade, um, um, for that matter, and then a second question just new, um, is about, um, from Theresa K, uh. Uh, I'm interested in, uh, in your finance, what your findings show about the uh mechanisms for persistent learning losses. And the question is, if can one interpret your findings as suggesting that learning losses effects were unrelated to permanent trauma or nutrition. Or nutrition effects on brain development, but we're actually we're related more to the differential ability to catch up and fill in the learning, the learning gaps. Um, and let, let me stop there. Then I have one more general question for, uh, for the two presenters, but let, let, let, what did you, uh, tackle both, um, uh, those two questions I just knew, please, uh, thank you, Jaime. Uh, Martin, with respect to Martin's question. Uh, so, I think there are a large number of things going on. I mean, remember, these were also children who might have been, um, affected in terms of their, uh, resilience, in terms of dealing with trauma, um, in terms of a whole number of issues afterwards. I think what I'm comfortable saying at the moment is, look, this is really a demonstration that even after you give a lot of compensation, these learning losses happen and they happen after coming back to school, right? Or a large fraction of them. So, at the minimum, it says, don't just test children when they come back or get, you know, assess their learning when they come back and think it's over at that point, right? What we really need to do at this point is kind of set up cohort studies that are saying, look, we're gonna track these children for 4 or 5 years. In these 8 or 10 countries or whatever it might be, uh, that, that look like other places, because otherwise, we will not see if these learning losses are compounding, and I think we have enough evidence to say that there's a real possibility there. Uh, then the question of, is it because of not adapting the, uh, curriculum to the initial losses? My suspicion is yes, uh, because all these kids were promoted, as will the children in COVID. Uh, so suddenly now you haven't done your grade 3 curriculum. You were barely at the grade 2 curriculum to start with, and now you're trying to cover a grade 5 curriculum, knowing grade 2 material. It's not gonna work. Um, on Teresa's question, which is very similar on the mechanisms for permanent persistent learning losses, I think Teresa is exactly right that look, because these are happening across the age range. And it's not happening, and remember, we didn't find height losses for children above the age of 3 at the time of the earthquake. So we know that despite not having any height losses, you can still have these. And my guess more and more is this is quite closely related to what was going on in school. So I would pitch these uh uh things as saying, look, it's a cautionary tale, saying it's not only what happens while the schools are closed, it what happens after they reopen is going to be really important and don't think that because these are children who haven't suffered perhaps or hopefully haven't suffered nutritional losses during this time, although look, you know, in India 50,000 children were born every day of the lockdown, we don't know what situations they were born in. Um, so there is a persistent concern with what's gonna happen with these kids as well, uh, but don't think that the children who escaped nutritional deficiencies will not suffer learning losses, uh, in the current crisis. So, overall, you know, it's this, uh, I don't want to say doom and gloom, but maybe it is a bit of doom and gloom to say, look. Uh, all the things that we are thinking about, can we make it through the school closure period and then we're done because the children have not suffered nutritional losses. I think this study is a persistent reminder that no, none of those assumptions need hold, uh, in this particular case. So let me stop there. No, and, and, and definitely I mean the, uh, the, the learning losses due due to nutrition, I mean, uh, most likely will, will be there. I mean, uh, some countries have been able to compensate for that, but others, for others have been much, much more difficult, and that, that's where we're gonna see a lot of inequality in terms of some areas of some countries in which the mitigation has been there on the nutritional side, but other parts of the country in which the mitigation has not been there. But then a key point of what you're saying is that a lot, I mean. It's not necessarily doom and gloom. It depends a lot on what we do, right, when kids start going to school in a, in a different way, um, but let me, let me now go to, go to, uh, uh, Joe Pedro, um, for a coup, a few minutes' response, uh, Jo Pedro, two points, um. Uh, one is, um, the question, um. By Martin as well, um, um, that I mean, well, he's, he, he's saying we have been using your simulation in South Africa and there's a lot of interest on this, uh, but in your discussion the report you referred to the likelihood of learning losses actually exceeding the actual schooling years, right? Do and actually exceeding what is due to forgetting. But in your calculations, it seems that learning losses equals time lost and might not be incorporating the forgetting part. Remember that you were saying that this is about how much you are, you're, you're not learning how much you're, you're, you're forgetting of what, what you have learned in the previous cycles. So maybe you can explain better in which way the um, the, um, simulation captures that captures the, uh, the forgetting part. And then this, the, the other question, um, for usual Pedro, and that this is from Veronica uh Diaz from the Peru office, and I think this is, this is a question to test your political skills more than your econometric skills because she's, she's asking how do you advise and communicate the effectiveness parameter right that you are using when applying this tool, the effectiveness of the mitigation actions. When the country might be very optimistic about their remote learning strategy, but actually we might not have actual certainty of that level of um um of effectiveness of that remote learning strategy, and, and I, and I think this applies to many countries because what you see in paper in terms of what they are doing in terms of the mitigation strategy and the whole multi-platform learning, learning, the multi-platform remote learning strategy. In many countries you see one thing in paper and you see a very different thing in terms of what's happening in the country. But then how do you, what, what do you put in your simulation? What the, um, what is in paper according to what the government says or, uh, what might be something closer to the truth? Let me stop there for those two questions, I'm Joel Pedro, very quickly because we're almost running out of time. Sure, thank you very much, uh, Martin, uh, for your question, and I'll be happy to follow up uh bilaterally as well, but basically, we, we do have both elements, uh, both the, the learning and the forgetting in the context of the simulation one, the first of the three simulations that I presented, which is uh reported in terms of learning adjusted years of schooling. And, and I can get more into the math and the details bilaterally or uh and with others as well as, as needed, but it is there and, and, and one way you can see that is that even if you choose uh a a scenario in which let's say schools are gonna be closed for 50% of the time, right, you're gonna lose in terms of learning adjustities of schooling more than that, right? You'll lose more than. In a single year now it is important to keep in mind that the forgetting is also gonna be a function of the length of the school closure, right? School closures are gonna affect the forgetting because, uh, and that is a very similar argument to the, to the summer learning losses, right? Um, so they, they are captured in the lace and be happy to, to elaborate further and, and, and follow up bilaterally. I think on Veronica's point, I think, uh, which I, I think a couple of, of, of thoughts on that. I think this is why we, we think in terms of scenarios, right? Uh, I think that's one of the values of scenarios, and you can have a pess an optimistic scenario, which can be very much aligned with, uh, maybe, uh, the vision of, of, uh, what the government is, is presenting. But this is a kind of uh a situation in which you wanna hope for the best, but you wanna be prepared for the worst, right? And, and it's just not enough to, to be optimistic about this, right? We need to also uh be ready when things don't, don't work as planned, because implementation in the context of COVID is particularly complex. So, building alternative scenarios can be a useful way of, of bringing this together. I think another aspect is, is to actually acknowledge that, look, it's not just about what the government is supplying, which is the G in, in our exercise, right? Remember that we also have the A, which is the ability of the household to actually absorb what the, the government. is supplying and we have the E which is the effectiveness. So even if government is perfect in terms of the supply and what the government is supplying, the effectiveness is also perfect, it's 1 if the A or if the households don't have access to, to good bandwidth or if the computers are not there or if uh the resource has to be shared among. Multiple children, right, uh, effective, the overall mitigation effectiveness might actually be hurt, right? So it's not just about what the government is, is, is providing, but this combination of, of several other elements, and this is something that the government needs to be mindful of, and, and some of those elements, to be quite frank, are not things that the government will have direct influence, right, during the crisis and. And in spite of all the good intentions and and efforts, right, they need to be accounted for, uh, in order to have a realistic uh response when schools reopen. So, uh, thanks Joe Pedro. So very quickly, and we're probably gonna run out of run 2 or 3 minutes, uh, over the time, uh, so very, very briefly. So a couple of questions for Jishnu, uh, first, and then one for, um, for both. So, so let me, let me go first to the 22 questions for Jishnu. One is, um. Uh, is there, are there any lessons from the recovery process, uh, in Pakistan on how to build resilience on the education system to ensure that children can continue with their education when, when you will have another shock like this, um, and then for Michael, uh, Crawford, um, is, if, if this is closely linked to what's happening in school, math and English should suffer greater losses. Since, or to the extent that Urdu is, uh, learned in school and partially throughout the daily daily use, um, uh, and language in everyday life, I mean that's, that's very specific to to Pakistan, do you find this pattern or is there a loss in the same three subjects, so math, math, math and, um. Um, math and English on one hand and Urdu separately, uh, and then finally one question to, to both, uh, is the implication of this, uh, of this exercise that needs to massively then increase testing when kids go get back to school, uh, with guidance to teachers about how to adjust, uh, the curriculum, uh, that's from Marianne from Marianne Fay, so to. To both of you, so you that's your your your final words, um, additional and then uh Joe Pedro. You're muted, uh, just. Uh, I'm going to take only one of those. Look, I think the biggest thing that the Pakistan earthquake has taught us is, and, and one that I'm, um, you know, quite, uh, um. That I think is important is we are having this discussion today. Right? I mean, a lot of you have worked on disasters, and I can't think there was no, nobody in the Pakistan earthquake was discussing what should we do with schooling when schooling resumes. It was not on the cards, right? So think back to, you know, the number of times we've had to deal with disasters, right? Floods, um, other stuff, all of it, you, you, you know, Jaime, this is a question to you actually. How many times is the education group involved in post-disaster reconstruction? So that's, you know, it'd be good to get your sense on that and the fact that we are now actually actively involved in this from an education front is, I think, the big learning from Pakistan. So, you know, let me, let me just stop there because I think it'd be more important. For people to hear what your thought is on, look, you know, there are, we have a full disaster. The bank has a full, the GFDDR and this and that. How involved is the education group in that, in local disasters, and, and, you know, can, is this the big change that needs to happen. Great, thanks, uh, thanks, Inu for putting me on the spot, um. So let me, let me, uh, go to Jo Pedro while I think how to answer, uh, additional's question. Joe Pedro, please, sure. So on Marianne's question, um, I think it, it's not just about testing, right, uh, um, in terms, it's about. Testing and, and, and actually linking that with uh class practices, right? And, and it has to be, I think a smart type of assessment, right? It's not about large scale assessments, it's not about, you know, doing, you know, national learning assessments at scale, it's really about what is happening in the classroom, assessing what are the levels of, of, of learning that those children are when they arrive in the system again. And also, uh, matching the, the support and, and the teaching material that is available to, to the needs of these students, right? It has to be actionable, it and it has to be very concrete, and it has to be linked to, to, to teaching training. Uh, teacher's training has to be linked to the support to the parents as well, right? But yes, definitely I think we need to be more adaptive because if we're, if we're not and we're flexible, if we're not, I think the risk is that a lot of these children that have lost a little bit might end up dropping out of the system, might, might actually, uh, you know, lose a lot more in the, in the long run. Thank you. Great, thank you very much uh Joe Pedro and Yishnu. I think as as Yu said, I mean it's it's extremely um, uh, important that we're having this conversation, um, um, unfortunately we're having it also because of that we are now experiencing. A dramatic shock and the difference is that it's not happening in one place in the world like in Egyptian's experiment, but it's happening absolutely everywhere. That's what, what we were saying at the beginning. This is the, the most dramatic education shock the education system. Will be probably living but my sense, um, and trying to answer your question is that, um, the not necessarily what is the impact on children's welfare from a perspective of education and learning is in the top of mind of societies, uh, politicians and and and and when when we're thinking of what we need to do right after the crisis. And exactly the same, I mean, as, as you were saying, I mean in this even in this crisis it is true. I mean we need to obviously we need to pay a huge I mean attention to uh to the health shock that we're living and to the social protection issues, but I'm not sure that education is in the news in the way it should be in all countries. It's, it's happening in some I would say that it's, I, I would not say that this is uniform across the um across the, uh. The world, we just had a conversation on, on Egypt. I mean 2 hours, 2 hours before this, uh, this meeting, and Egypt, for instance, for, I mean for, uh, and I know it was a chance, but it happens to be that 3 years ago they were starting a huge change in the structure of the delivery of the education system, education in in Egypt in which there was a much larger role of technology. Uh, it, which is obviously is not that has closed completely the digital divide, but there, there was already a huge discussion in the country, right, in terms of how testing has to be reformed and what's the new role of the teacher and in which way techno, I mean the, the education has to be a blend between presential and online that had happened already 3 years ago. So now they are in a much better position and there's huge discussion in terms of how to continue revamping that system now because of the pandemic. So happens to be that in that the country education is in the news right with this with this analysis here in a in a specific case of Pakistan and with the analysis that Joe Pedro has shown us for the whole for the for the whole planet. Uh, I think both are critical in order to make sure that really education is in the news and that, um, that we internalize much more the huge impact that this type of shocks might have in welfare of children, right, because of the loss on the academic side but also because of the huge loss in terms of the social interaction that the school that the school needs, uh, and in terms of mental health, etc. so. The losses are um um are very large and probably the numbers that Joa Pedro has shown are most likely an underestimation of the complete, I mean, impact. I mean if we take into account nutrition, if we take into account mental health, and also his um his assumption is that the school closure is for 5 months, uh, in his intermediate scenario and most likely in many countries that's not gonna be 5 months, so, um. So I think, I think, I mean we're here for the long haul. We need to continue, I mean, at, uh, um, improving, improving this type of analysis, but also I mean uh there's a huge homework for all of us in terms of then. How we avoid this doom, more gloom and doom and scenario that Gno was saying, and we do have the opportunity of rebuild education systems right in a way that are much more resilient and much more equitable, but actually that will be the um the um the uh that will be the um the discussion for another for for another seminar because there's a lot to do to to discuss and talk about that. So I just want to thank, I mean, uh, all the participants, and I just wanna, uh, thank, uh, Jishnu and Joan Pedro for the presentation of these, these two excellent papers, and, uh, thanks to everyone, and we'll, we'll keep on going. Thank you very much and thanks to all. Take care. Thanks, thanks, thanks, June. We'll, we'll, we'll have you back soon. Wonderful. Bye and thank you. Thank you for sharing. Bye bye. Thank you. Stay safe. Stay safe, everybody. Stay safe. Thank you so much. You too. Take care. Bye.

showAllTranscripts

duration

PT1H35M59S

scene7File

worldbank/SchoolClosures2

scene7Domain

https://worldbank.scene7.com/

scene7FileAvs

worldbank/SchoolClosures2-AVS

title

SchoolClosures2

description

SchoolClosures2

showTimestampAndTranscript

yes

sec-spacing

section-none

sec-spacing-bottom

section-none

sec-swoosh

sec-swoosh-type16

style

default, transparent-white-components, heading-center, add-style, bg-primary-blue-70-gradient, bg-dynamic-height

col-xs-12

col-sm-12

col-md-3

col-lg-3

col-xs-12

col-sm-12

col-md-7

col-lg-7

add-style
lp-body-content

School closures due to COVID-19 have left over a billion students out of school. The potential impacts are wide-ranging. Disruptions to schooling and the resulting learning losses—from shocks to schooling and incomes, from unequal access to remedial measures such as distance learning, and from subsequent inequalities once students return to school—could be large. In this webinar on July 9, 2020, Jishnu Das (Professor, Georgetown University) and João Pedro Azevedo (Lead Economist, World Bank) presented two approaches to understand these processes and model their potential magnitude.

lp-heading-top-medium

lp-heading-bottom-medium

col-xs-12

col-sm-12

col-md-2

col-lg-2

sec-spacing

section-none

sec-spacing-bottom

section-padding-between

style

default, video-transcript

everything

https://webapi.worldbank.org/aemsite/everything/search

true

{"facets":[],"orderby":"contentDate desc","select":"*","searchFields":"*","search":"*","filter":"contentType eq 'Video'","top":20,"skip":0}

See Rest of Series

yes

sec-spacing

section-none

sec-spacing-bottom

section-padding-between

style

default, bg-neutrals-15-gradient, remove-border, lp-mini-card-primary

The Long-term Impacts of School Closures

RELATED

See Rest of Series