 Hello everyone, thank you for to be here in this session from the wider conference. So we will start the presentation from Horace and each one of you have approximately six minutes to do the presentation. And then if someone have any question please write in the Q&A part of the, you can find that in the, your right side that is chat, people in the Q&A you can write there your questions. And in the end we can select some questions to ask the participants and they also can answer in the chat. So please Horace, you can start your presentation. Thank you very much for the opportunity to present this paper entitled COVID-19 and children's school resilience evidence from Nigeria. This paper is a joint work with Sylvan Desi, Luca Tiberti and Marco Tiberti. So basically in this paper we are interested to see how COVID-19 affects school dropouts in Nigeria. Briefly talk about the motivation. Over the next month there has been an increasing amount of research on the impact of COVID-19 on many socio-economic outcomes including education, health, food insecurity and so on. Addressing this issue is not surprising in the sense that since the beginning of the pandemic until today the World Health Organization count more than 200 million of cases of COVID-19 and this includes around five million of deaths. So to contain the spread of the COVID-19 government around the world put in place many restrictive measures and this includes school closure, water closure, any lockdown measures. This lockdown measure in Nigeria for example has a negative impact on household income. So for example in Nigeria 2,000 of children in our analysis sample are from household whose total income declined after the COVID-19 outbreak. So in developing countries aggregate income shock increase children's vulnerability to child labour or child marriage so to cultural practice not to undermine school attendance. Giving the negative impact of COVID-19 on household income is likely that the opportunity cost of households to send their children to school could increase as Bandera and Kota have shown in Sierra Leone in the context of Ebola. So we are contributing to this literature by showing how COVID-19 affects school dropout. Our second research question is that does COVID-19 increase gender inequality in schooling? So talking about the data, this research mainly relies on the Nigeria COVID-19 national longitudinal full survey RANSIS. This data is actually a sub-sample of households that have been interviewed as part of a Nigeria general household survey in 2019. This survey includes information at the individual level before COVID-19 and during COVID-19 after school reopened in Nigeria in October 2020. And this data includes many information like say respondents school attendance, their age, their gender, their place of residence and so on. So to identify the impact of COVID-19 on school attendance, we estimate the following individual fixed effect linear probability model. So school attendance is a dummy variable acoustic value of one if the response is going to school and zero otherwise. We are measuring this variable before COVID-19 and during COVID-19 after school reopened. So our main coefficient of interest is alpha zero to capture the impact of COVID-19 on school attendance. So talking about the main results. So we find that COVID-19 lockdown measures reduce children probability of attending school after the school system reopened. And the impact that we are getting increase with age. What does it mean? This results highlight that age is a significant factor in how households school attendance decision respond to COVID-19 lockdown measure. So we can see that households are more likely to pull older children out of school than very young women. So the second result that we found is that we did not get any gender difference. However, the evidence showed that in the child marriage from Northwest part of Nigeria, COVID-19 increased gender inequality in education among children aged 12 to 18. So what does it mean? So given that Northwest Nigeria is one of the region where child marriage is more prevalent. These results provide suggested evidence that child marriage in this region may increase due to COVID-19. So policy implications are prepared so just that public policy to mitigate the adverse socioeconomic impact of COVID-19 should target the school residents of adolescents girl in setting where child marriage is relatively common. Thank you very much for your attention and your question comments are very welcome. Thank you, Otis. Just to let who arriving this session after Otis is at the presentation. So if you have any questions you can ask in the Q&A and he can answer there or we can have a session in the end. So we will continue with, sorry, with Kibrom, please Kibrom. You can start your presentation. Thank you. Can you see my slides? Yes, we can. Okay. Thank you. Thank you for joining us for this session. So I'm going to talk about response fatigue in phone surveys and this is a better on experimental data on diversity from Ethiopia. So this is a joint work with my colleagues, Gush Burhani from Ifri, John Haudenay from Cornell and Kibrom Taffer at the World Bank. So obviously the outbreak of a pandemic like the COVID-19 necessities, the need for further monitoring of welfare outcomes including food security. However, we also know that this type of pandemic creates substantial obstacles for survey operations and mainly for running face-to-face surveys and interviews. Because of the disruption in survey operations and face-to-face surveys, we have seen a lot of increasing interest in remote data collection tools after the outbreak of the pandemic. However, we know quite little about the implication of this in terms of effects on data quality. So in this paper, we evaluate the overall and differential impact or implication of fatigue in phone surveys. There are several reasons and justifications to argue that fatigue is likely to be more pronounced in phone surveys than face-to-face surveys. I'm going to cover a lot of this but I think it's arguably that phone surveys are more prone to fatigue compared to face-to-face surveys. So here we have the data and the experimental design. We're using two rounds of phone survey data. The first one round was collected in June 2020 and the second round was collected in December 2020. And this data builds on a large household survey and mainly a large mother's survey that preceded the pandemic where the primary respondent was the mother or the caregiver of the young child. So in the first phone survey, we reached out to about 1,500 households and mothers. And in our December survey, the second round, we introduced a random assignment of respondents to one of the two questionnaire types that we prepared. That is 50% of the respondents were assigned to receive the instrument on women's diversity 15 minutes earlier in the interview. And that's immediately after the introduction of the survey. And the remaining 50% were assigned to the control group and they received the module 15% in the interview. And that's similar to what we had in the baseline survey. So here we have some baseline characteristics. I can quickly skip. I only want to mention a couple of points here. You can see that we have a fairly balanced baseline characteristics and you can see that before the intervention or before the random assignment we did, you can see that mothers in the treatment and control group reported similar diversity and also similar food groups. Here we have an empirical strategy, which is a straightforward fixed effect specification where we have Y at the outcome of interest, which is diverse score. And then we have individual fixed effect and time dummy and then the treatment indicator. Here we have some results. So you can see in the first two columns, you can see that mothers in the treatment group report consumption of 0.25 more food groups. And this translates to 8.5% reduction in diet diversity because of the delay in the arrival of the module. And the last two columns show that delay in the arrival of the women's module reduces women's meeting minimum of four food groups by about 28%. Here in this table we taught that fatigue may entail differential impact by food groups depending on how frequently these food groups appear in the household menu. And so we split the diet diversity questions into several disaggregations. And here we have in the first two columns, we have the mostly commonly consumed food items and then the last four columns are for food items that we think are less frequently consumed. Here you can see that delay in the module or in the arrival of the women's dietary diversity module leads to about 8.6% reduction in the probability of reporting consumption of animal source foods. And this translates to about 40% reduction in the share of matters reporting consuming animal source food. This is substantial reduction. Here we have, we report some heterogeneous impacts by respondent characteristics and what we find is relatively older matters and those with lower level of education and also those matters coming from larger households are more vulnerable to fatigue. And there are several intuitive reasons to justify these heterogeneous results. So to conclude, we find that delaying the timing of matters food consumption module by about 15 minutes leads to 8 to 17% reduction in diet diverse score or about 40% reduction in the number of matters reporting consumption of animal source foods. This is substantial, especially given the time difference that we are interested in this. So there are several implications to this finding. I'll just mention three of them. The first one is if 15 minutes delay in the arrival of a specific module leads to this much of difference then this implies that compiling of statistics across different surveys or across countries or across time may be confounded by the position of the position of modules in this different surveys. And secondly, you can see that the heterogeneous results that we document imply that fatigue exhibits system or response fatigue exhibits systematic pattern and this will introduce what we call the name classical measurement error, which had a significant inferential implication on our econometric analysis. And finally, our findings highlight important tradeoff between the volume of information to be collected and the quality of data. That means researchers would have to carefully trade off between the two dimensions to data quality and also to the side of information. Thank you. Let me stop here. Thank you, Kibron. So we will follow now with Ingrid. So welcome Ingrid, my colleague from the Wider Summit School. Thank you, Rodrigo. At the outset, I would like to thank UNU Wider for giving me the opportunity to present our paper and the paper has been co-authored with my colleagues, Dr. Mansi and Russian Thomas. See this COVID-19 raging across the globe has affected the education sector as well. And as part of the effort to contain the virus, the countries temporarily shut down educational institutions. And as a consequence, the right to education has come to be stalled. And as per the statistics provided by UNESCO, early of this pandemic that made April 2020, 188 countries opted for nationwide closure, which has affected the estimate of over 90% of the total world's enrolled student. India as a country alone accounts for 320 million students have been affected due to the nationwide closure of school on March 25, 2020. And out of these 320 million students, about 10 million students are studying in primary pre-primary level and more than 143 million are studying in primary levels. The most worrying fact is that as the lockdown period increases, academic skill are likely to be affected negatively. And this impact, this elongated, the impact of this elongated interruptions are unequal depending upon the socioeconomic statuses of the students. Having said that, there are several innovative responses to address the unprecedented pandemic that has caused a catastrophe in the education sector. Online platforms are being used as a substitute to continue with the education. However, given this much importance of this technology, this technology has created a new era of this digital divide and the education is divided between those who own technology and those who do not have access to the technology. So with the online classes being concentrated in the urban private school, the gap between urban private school and rest of the rest is expected to increase drastically. Therefore, the objective of this study is to assess the impact of COVID-19 pandemic on primary education in India, explicitly focusing on the differential impact between public and private school going children. So in order to accomplish the objective, we have conducted a primary survey comprising both online and offline mode. Thus our survey covers the 377 sample of parents of both private and public school going children in Karnataka state in India. And the first descriptive result shows that the comparison between the mean study hours between public and private school going children and there is a marginal difference in the study hours at home. When study hours are school is added with the study hours at home, this difference becomes much significant. And the question arises if the students are allowed to go for online teaching, online, this student learning process, then this private public school going children, whether they are in a position to access online classes or not. And therefore we have made a comparison of access to different resources required for attending online education between public and private school going children. And our results this figure too clearly shows that in terms of the availability of different online means such as smartphone, iPad, computer, laptop, broadband connection, each and every indicator that private school going children account for much higher value as compared to the public school going children. And even the total the inequality in terms of this access to different resources and the study hours are decomposed between and decomposed by between and within inequality among public and private school going children. And it has been seen that the between public and private school going children's inequality is much higher as compared to the within inequality. But in addition to this descriptive analysis to identify the robustness in the in the in the in the result we have used the econometric analysis, keeping this study hours at home study hours at home and school and access to resources as dependent variable and this choice of school between public and private as the main independent variable keeping other socio economic variables as control variable. But here the problem is this choice of school may suffer from this endogeneity problem. And therefore following but this at this choice of school is a binary variable following old rich and angriest peace. We have first estimated a binary response model by keeping this choice of school as a dependent variable and this nighttime luminosity in the district in our case as the instrument variable and keeping all other constant remains same. And then we have used the fitted probabilities from the first stage, the from this binary equation in our ultimate to SS less model. So the main findings based on this econometric analysis is that first stage regression shows that the lifetime luminosity in a district is negatively and significantly related with the choice of government school, implying that the choice of school is very much dependent on the region's level of development from the less developed region mostly students come from come for this this public schools and the second stage regression our ultimate regression shows that the public school going children study fewer hours at home as compared to the private school going children but that the magnitude of the coefficient increases substantially when school hours are added during the pandemic. So this this has intensified the gap between private and public school drastically. And the second specification shows that the public school going children have lower access to resources that are required to attend online classes. So which is a matter of great concern if online mode of teaching learning process are implemented for government school all relationships are robust because we have tested with different specification of the variables and the regressions are controlled with the gender social groups delige on location parents occupations and educational level so again. Please I would like to ask you to move to the your final thoughts because we are. Yes, yes, final slide. Okay, so the main conclusion is that this pandemic has intensified the gap between public and private school. And this crisis has made us aware of underlying issue which needs to be addressed starting from the access to resources to implementation of better educational policies and in this implementation active involvement of all stakeholders are very important. So with the pandemic there have been several innovative ways to that with the inner in which children were reached out to address the challenges in the education sector. However, it has to be inclusive of the students belonging every section of the society. Thank you. Thank you and sorry for it is because we have. Lisa, please you can put your presentation. Hello, my name is Lisa Palette and I will present this joint work to work with you to get quick men this on my child regarding the impacts of the COVID-19 on higher education for you. So the aim of this paper is to provide empirical evidence for developing countries such as your way on the effects of COVID-19 during 2020 on educational outcomes of first year university students. We think that it is relevant to understand the effect of the pandemic on tertiary education since the importance of education for human development and the development of countries. So in order to do this, we will exploit a rich data set coming from different administrative records from the public university in your way to analyze not only the effects of the whole population, but also the Russian effects. So to give some context, the law is the only public university in your way and it is a university where you can study your undergraduate career for free. The university offers around 100 undergraduate degrees and cover 86% of university students in the country. And regarding the situation with COVID-19, it is important to mention that in your way the academic year starts on March. Therefore, the pandemic arrived at the beginning of the academic year and imply the suspension of in-person classes. But after one month without classes by mid-April, most of the courses were changed to online courses. What we are going to use is two sources of information on one hand administrative records regarding educational events of students. And on the other hand, we will use information of the enrollment form completed by students when they enroll to a new career. So we will end with a database that recreates the individual academic trajectory combined with socioeconomic and socio-demographic information at the individual level. What we are going to do is to compare the first-year students from previous years from 2017 to 2019 with the first-year students from 2020. The outcomes that we are going to analyze are the probability of being enrolled but not doing any academic activity, the number of courses that the student takes, the number of courses that the student approves, and the mean grade. So here I present the table with the results and what we find is that COVID-19 reduces in approximately 3 percentage points the student activity versus previous cohorts. And although it's less significant, COVID-19 also reduced slightly the annual number of courses taken by students. However, the class of 2020 on average obtained better scores in comparison to previous generations. When we analyze the results by subgroups, we can see different results. Panel A showed disparities across different parental educational backgrounds comparing those students with parents without a university degree and separately those students with parents with university education. Independently of the parental education background, students in 2020 reduced their activity in comparison to previous cohorts of students with similar characteristics. Less significant is the effect of COVID-19 on the number of enrolled courses, specifically while students with less educated parents reduce their enrollment in 2020 in comparison to previous generations with similar characteristics. No statistical effect is found for students from better parental backgrounds. However, for the students with more educated parents, we observe an increase of approved courses in 2020, and regarding the average score obtained during the first year, what we observe is that students from relatively worse parental educational background do better in 2020 than before. In panel B, what we can see is differences by gender. First, note that boys do relatively worse in 2020 than in previous years, regarding the probability of not doing any activity. For girls, their result is less significant and also half in magnitude. And in addition, considering the taken courses, boys take less courses in 2020 than in previous years, while no effects are found for girls for both genders. We do not find significantly effects of COVID on the total number of approved courses in the first year. And on average girls in 2020 do better than girls of previous cohorts, while boys do slightly similar to previous generations of boys. Finally, we analyze students coming from a public secondary institutions and those from private ones. We first know that both groups of students decreased their activity in 2020. And in addition, students from private secondary institutions reduced the number of courses taken during 2020, while no statistically significant effects are found for those students from public high schools. In addition, for students from public schools, they increased the number of approved courses and regarding the scores, students from private high schools in 2020 do better than students from the same background in previous years. We don't find statistically significant differences for students from public secondary institutions in 2020 in comparison to previous years, regarding the mean scores. Overall, this study wants to provide relevant evidence of the potential impacts of COVID-19 in university education outcomes. Our results are, to the best of our knowledge, the first ones to exploit administrative data at the national level for a developing country. The results show an increase in students that are involved, but do not do any activity and a decrease of the number of courses, but also an increase of the number of approved courses and an increase of the mean grade for some group of students. And we must stress that in order to disentangle the mechanisms behind these effects, we need more data sources that we would like to exploit in the future. And also, we would like to add the results for the first semester of 2021 and also to add more traditional effects to understand what is happening in your way regarding the effects of the pandemic on tertiary education. Thank you, Elisa. So let's go to the last presentation. Please, Christopher, you can share your presentation for us. Thank you for the opportunity to participate in the wider conference today and to outline our paper on rebuilding the capital. Over the last year, there's been an enormous volume of work on economics of the COVID-19 pandemic. On the one hand, there's a body of macroeconomic work that's focused primarily on the short-term, medium-term impact of the pandemic and very often feeding into debates about adjustment programs and their financing, whether that be domestic or external. On the other hand, there's a rich body of work that has drawn attention to the impact of the crisis on human capital, on health, skills and education. And this literature explores the effects of the crisis in disrupting labor markets, on skills acquisition, and in particular on the effects of loss of learning in the education system. In addition, it also looks at the effects of the diversion of health care spending away from primary health care and non-commodable diseases, for example, towards addressing the acute effects of the COVID-19 crisis. What we seek to do in this paper is bring together insights from these two literatures in order to assess and quantify the medium-term macroeconomic and welfare impacts of COVID-19 in low-income countries, and in particular, focusing on the operation through the transmission of the deterioration of human capital formation onto productivity and growth in the medium term. Ultimately, the aim, of course, is to assess the macroeconomic and public finance implications of public policy measures that are aimed at reconstructing skills, health and education capital post-crisis. Our approach is to deploy a dynamic macroeconomic model that we have developed elsewhere. And we calibrate this to a representative low-income country in this paper with a relatively conventional representation of the short-run economic shock associated with COVID, and then extend that to focus in particular on human capital dynamics and the links to growth and productivity in the medium term. On the short-term side, this is an example of the approach that we might take. This is taken from an earlier paper where I and others have looked at the shock on Uganda, where against a background of trend per capita consumption growth, we decompose the shock, that's the vertical bars, into direct effects of lockdown measures, the general equilibrium effects from the disruption of the domestic supply side of the economy, and then add in the spillover effects on the demand side operating through the current capital accounts, the balance of payments, and then using this framework to assess the outcomes under different public policy responses. The key point here is that whilst we embed this basic framework into our current model, this focuses only on the short-term. And what we want to do is build on this to incorporate a more plausible representation of the human capital dynamics associated with the crisis. Now there's an extensive literature on the returns to public investment in education and health, and our core calibration pays very close attention to trying to get those rates of return and productivity effects right. But the important addition is looking at the new literature to think about the depth and duration of loss of learning effects and health diversion effects as a result of the crisis. And there's some really exciting work out there that we're able to draw and work from the World Bank. It's some very interesting work from Michel Kaffenberger and Lan Pritchett on loss of learning. And coming out of the same RISE program, some really useful work from Andrew Abbey, Daniels and Das looking at duration effects of loss of learning shocks, for in their case looking at the Pakistan earthquake of 2005 and the duration of the effects of that shock through the education system. Let me move very directly to the key policy conclusions given the shortage of time. We're always going to face a serious challenge of calibration with models of this type, and that requires us to do a lot of sensitivity analysis. We've no space here to describe this in detail, but we believe that the results that are coming out of our analysis are plausible and compelling. Number one, the direct short run economic effects of the COVID crisis have been and will continue to be brutal in many low income countries. This is true for any calibration, any plausible calibration. And importantly, these are almost always out of proportion to the health costs of the pandemic itself, at least for the moment. Recovery is likely to be very slow and the distributional effects disturbing. This is a shock that bounce back is unlikely to terribly rapid. The aging term effects may well be severely exacerbated through these loss of learning effects and diversion of health spending effects, unless countries can commit to significant reconstruction process. But given the highly constrained fiscal environments in which many low income countries were operating in just prior to the crisis, fiscal adjustment burdens may be very severe and may not be politically feasible. Another indication being that recovery is delayed even further. In such circumstances the returns to effective external assistance are likely to be high, but will require extended engagement and tolerance of high debt levels well into the future. My final slide just illustrates this, showing that in the short run the shock to growth, to private investment, to incomes and unemployment is likely to be very severe, but more importantly recovery is extremely slow, recovery back to trend. Only if we are able to think about public investment programs that help return productivity and hence private investment back to trend do we have a hope of reversing some of the loss of development impacts that have occurred over the last few years. Let me stop there, thank you very much. Thank you Christopher. So we still have approximately eight minutes for the end of the session. If someone have any questions please type on the Q&A. So there is a question for Oras. Oras if you can read in the Q&A and prepare your answer. I can start first saying that thank you for everyone. I think this session was really important discussing the impact of the pandemic on human capital is fundamental. And there was one thing that all the papers here were discussing that is the lack of data to analyze the impact of COVID on human capital. And most of the papers here tried to create some data to collect some data and also discuss the problem with that. So I would like to start the discussion with a question to Kibron about how do you think is the external validity of your results? Because it's really important I think your work. Because now everyone is working with fund service. If you look for example to the World Bank, the World Bank basically conducted fund service for almost all countries in Africa, in Latin America. And we have this trade off about collecting more information and having some bias in the answer. So I would like to listen a little bit for you. Thanks. Yeah, we have been thinking about the external validity of our findings and I think this is a valid question. And I think the best I can tell is we are trying to apply this in a first to first setting and also in another fund survey. But I agree that I mean, because also the context and the environment that you operate this type of surveys would also matter. So the only thing I can tell is like we're trying to get another round off survey with a similar design, both in face to face survey and in a fund survey setting. By then I would be able to tell more about whether our findings are or take or assume some level of external validity. But for now I think I mean what I can tell is I think internally our findings seem to sound valid and intuitive. Thank you Kira. So we can move to Ores. Ores if you can answer the question in the Q&A. I don't know if you are listening to me or because you are with your phone. Yeah, perfect. Yes, I'm going to answer yes, I'm writing my sources. Okay, so you are writing. Okay, thank you. So I can finalize with another question. I have a question for Elisa too. What do you think about teachers behavior? So do you try to address this question in your survey because we know that it's really difficult to talk about causality in this type of setting. But I also think that there is some teacher behavior affecting grades and also dropouts etc. during the pandemic. Do you have anything to talk about it? Regarding I mean the first result it was about no activity that it's more similar to the dropout. And that it's more like students that enrolled to the career but then they didn't do anything. So I mean that doesn't have to do with teachers I think in the sense that they didn't go to any class at any moment. They just enrolled but then they dropped out. So in that sense teacher may not have an important role but yes in grade for sure. I mean in this research that I present we couldn't control for that. But what we are trying to work in is like there are some I mean regarding the way that the classes were taught. We cannot do much but we can do because we were like worried about the online test compared to in-person test. So what we are trying to do is to recap all the courses that were done like the evaluation test was done in the same way than the previous years. The ones that the test was done in person compared to the courses where they changed the modality of evaluation and then see there we have a research. But now we are like trying to analyze course by course which were the ones that changes and which were the ones that stay in the same modality. So we are trying to get that information too much with our administrative records of the results of the courses at the university. But yes I mean we think that that would be a way of thinking why grades are now higher. Maybe it's because I mean all the pandemics and also the way of getting the results are doing an online evaluation test. Thank you Aliza and thank you for everyone that stayed with us until now. I think it was a great session and now I need to finalize because we need to move to the first sessions. So thank you everyone. Bye-bye. Thank you. Thank you.