 Hello everybody. Thank you very much for joining this parallel session on inequality. My name is Carlos Gradín and I'm a research fellow at UniWilder in Helsinki. So for this session we have three distinguished scholars who will review various dimensions of the effects of the pandemic on inequalities. The session will focus on the labor market and policies as well as domestic work and leisure. We will see evidence from South Africa, from India and from Latin America. So the first speaker will be Mara Lebrant, he's director of the Southern Africa Labor and Development Research Unit, Saldru at the University of Cape Town. He's also an resident senior research fellow at UniWilder and he works on many topics related to economic development, particularly poverty and equality of the labor market. The title of this first presentation is pre-pandemic labor market status and labor market restructuring under COVID persistent or change. Then the second presenter will be Ashwini Deshpande, who is professor of economics at Ashoka University in India. She works on economics of discrimination and affirmative action with a focus on caste and gender in India. She's also a non-resident senior research fellow at UniWilder and the title of her presentation will be the COVID-19 pandemic and gender division of paid work, domestic chores and leisure, early evidence from India. And finally we have Isordala Pogh, a research fellow at the School of Geography and Sustainable Development at the University of St. Andrews in the UK. She's a member of the population health research group. Her research interests include migration, labor economics, demography and development. The title of her presentation is the distributional consequences of social distancing on poverty and labor income inequality in Latin America and the Caribbean. We will play all three presentations in a row, 10 minutes, about 10 minutes each, while we, the presenters, we all turn our videos and audio off and later after the playing this presentation we will have some time left for discussion with the presenters here. Please feel free to address your questions or comments using the Q&A tool and don't be shy and you can request to us the question yourselves, otherwise I will read them on your behalf. Thank you very much and we can go with the first presentation. Greetings everybody, I'm Murray Lebrunt and I lead a UniWilder work stream on the SA Tide project in South Africa with the National Treasury on turning the tide on inequality and it's a pleasure for me to be participating in the session on inequality and COVID at the wider development conference. I'm bringing to this session some evidence from South Africa on social grants and employment over the pandemic. This is the course of the pandemic in South Africa. One can see COVID arrived in early 2020, it peaks, it starts to rise and peaks in the middle of the year then it wanes and falls again, builds up for the second wave around December, January 2021, falls again and then actually rose and had, we had our third wave in June, July of this year. So associated with this trend, government imposed a very strict lockdown in March and April of 2020 right here and then another strict lockdown and strict policy measures here in June, July of 2020 and again in December and actually again recently in our country. We are going to follow the impact of this pandemic and the impact of policy, emergency policy through something called the NISCRAM survey that followed 7,000 South Africans through the pandemic off the basis of a pre-existing national panel study called the National Income Dynamic Study and this graph also shows you the waves of NISCRAM data collection. Here's the first wave right straight off to the and there's part of the strict lockdown. Here's the second wave of data collection, third towards the end of the year when it's fighting again, 4th in February and 5th here in April and May. So let's see what this NISCRAM story tells us. It was a remarkable exercise of the whole of South African social science and able to use this national panel study and had to be done telephonically of course. South Africa has a very extensive grant system with a number of prongs, a child support grant, an old age pension and others and over the 10 years between 2009 and 2019 it was rolled out even further from 14 million in 2009 to 18 million in 2019. However, despite that or using that as the base of our emergency response in the four months of first four months of the pandemic an additional 4.2 million people had received COVID SRD grants, a social relief of distress grants that government rolled out as one of the emergency relief measures, a very minimal amount of money, 350 rounds per month, but rolled out to 4.2 million people plus top ups of the child support grant and the old age pension and then some labor market interventions as well for the formerly employed who had lost their jobs and then the social relief of distress grant was supposed to be targeted at more informally employed. So we had a decades worth of growth in grants over four months. What was the impact of that? The NITSCRAM survey served as the basis for a micro simulation model that was done by one of the projects in SA Tide to assess the impact of the COVID-19 of the benefit changes that have taken place, the top up of the grants, the social relief of distress and then the employment interventions. You can see in the gray here on the graph that earnings are negative and earnings growth is negative. It's mildly negative for the bottom desk, that's because the links to the land market are so poor for people there, but you can see the earnings go across the distribution in South Africa as the pandemic in the first six months really devastated the labor market. On the other hand, you can see the COVID-related benefits made a substantial contribution to changes in household disposable income, very well targeted at the poorest of the poor in the bottom desiles. So that's an considerable achievement of these grants and this work illustrates the importance of the grants. That said, these bottom desiles were very, very poor to begin with and so that doesn't mean that this is sufficient money to help households entirely cope with the impacts of the grants of the pandemic. I'm sorry and I give you one table to illustrate this point, table of hunger. One can see that in April 2020, the hunger levels that's under the strict lockdown, the pandemic had just arrived, the hunger levels were higher than that being historically for the preceding 15 years in our country way higher, substantially higher. 47% of people in the NITS Cram Survey ran out of money for food and across the board different measures of hunger you can see the spike in April. This does fall as the grants kicking and that's an important point to note, but it's also important to note that the hunger seems to stabilize at a new much higher level than historically. South Africans found it hard to cope with COVID, the grants helped, but they were only assistance, they weren't completely able to save people from hunger. Focusing on the employed then, on the labor market side, we had a 2.4 million extra jobs created in a decade leading up to the end of 2019. With the pandemic arriving, employment plummeted. According to the NITS Cram Survey, we lost 2.8 million jobs between February and June 2020. According to the Statistics South Africa Survey, we lost 2.2 million jobs between quarter one and quarter two of 2020. The NITS Cram Survey, so a decade's worth of job loss in six months equivalent to all that we'd gained in the preceding decade. The NITS Cram Survey shows this quite graphically. I'm sorry, got too excited. The NITS Cram Survey shows this quite graphically. One can see in February, the survey asked people what they were doing in February as a way of benchmarking the impact in the move from February to April with a pandemic starting in the middle. You can see the NITS employed growing dramatically from 13.7 million to 16.5 million and then stabilizing at that level of 16.5 million. The working falls from 17 million to 12.5 million and the furloughed, in other words those who have a job that aren't working and aren't being paid, it's not a category that we thought about in our country before COVID arrived. The furloughed grows by 2.4 million. That's quite substantial and then shrinks a little bit as we move out of the harsh circumstances of the April-May period in June. The longer run picture from NITS Cram following that on, we can see that then as we go to October the furloughed fall again, but notice also the actively employed rise much more than the fall in the furloughed and you can see that those who were out of the labor market or not employed also make up a portion of the jobs gained. Under the second wave, employment suffers again and then recovers almost back to its February 2020 level, but again that gain is split between those who had had positions before this pandemic struck, who had been employed and those who had not. So it's a textured story and it's worth it pushing a bit. I'll end off my presentation with two aspects of that textured story that is both a bounce back if you like of existing employment situations and existing economic activity and a restructuring of the labor market. The gender story is particularly important and I'm sure Shwini will have much to say about that in a different context. In the South African context, whilst males did recover to their February 2020 levels, that is certainly not the case with females. They did not recover to the same level and so some of the burden and the restructuring was borne by women in the labor market. It's a signal that there was this recomposition happening as well and it is unfortunate that the most disadvantaged were further disadvantaged. In addition if you add to the fact that women then were not the recipients of the formal sector employment jobs because they don't bear most of those jobs, the relief for those who had lost jobs from the formal sector or the social relief of distress grant which was given to people who weren't grant recipients or went receiving grants for their children even. So women bore double burden but my point main point here is that the labor market didn't just bounce back in a neutral way, it was restructuring. My final point about that was that to give you some statistics about those who moved into employment who were not employed in February, about 23 percent of the February employed were no longer employed a year later. They hadn't come back into the labor market and weren't part of the labor market yet 30 percent of those who were who were not employed in February 2020 were employed in March 2021. In particular the youth took up a large chunk of the new jobs or the new created jobs if you like over the pandemic and particularly youth with complete secondary schooling or complete or post-secondary training whereas older adults experienced the largest decrease in employment. So there was restructuring happening in the labor market and understanding that texture of that restructuring is crucial to exploring the poverty and inequality impacts of the pandemic in the labor market. Thank you, I'll stop there. Thank you very much for inviting me to present in this very important conference organized by UNU wider. I'll be presenting the results from a recent paper that looks at the impact of the COVID-19 pandemic and on gender division of paid work domestic chores and leisure based upon evidence from India in the first wave and its aftermath up to December 2020. Female labor force participation in India has been stubbornly low and has registered a decline over the last two decades despite favorable preconditions of declining fertility levels and increasing female education which elsewhere in the world would lead to an increase in female labor force participation rate but not in India. That issue has been and continues to be extensively analyzed. In this paper I am specifically looking at the impact that COVID-19 had and why would this be of particular interest is because during the Spanish flu epidemic there is evidence that as a result of the mortality of men in working age groups there was an increase in female employment in India and so it's interesting to see whether the COVID-19 pandemic and its aftermath had any similar impact. Now when you look at the literature on the economic impact of the COVID-19 induced lockdowns you find that more women in absolute numbers lost jobs than men in several countries of the world including in the United States so this is not difficult to understand because COVID-19 imposed social distancing norms as a result of which several sectors where which are you know female intensive sectors or that employ a large proportion of women closed down hospitality entertainment retail spaces and domestic workers such as caregivers nannies etc. or pairs could not go to their places or work. So that that was a direct effect on a women's employment. The indirect effect was caused by increased child care burdens and home schooling burdens as a result of the fact that schools and childcare centers were also closed for prolonged periods. As a result of this women who continue to have jobs and who could work from home found it difficult to balance the multiple pressures or work working from home as well as childcare and home schooling and dropped out of the labor force that led the New York Times to run a story that said in the COVID-19 economy you can have a kid or a job you can't have both. What was the picture from India like? So I looked at a panel data set and I choose a period of two years starting in January 2019 and ending in December 2020 that allows you know because it's a panel data set is it allows us to look at the same individuals pre-pandemic which I define till March 2020 and post-pandemic which means after the onset of the pandemic till December 2020. Now overall in India the result in employment is that more men lost jobs in the first month of the very strict lockdown that is April 2020 than a women and this simply reflects the historical and significant gender gaps in employment and labor force participation rates. Looking at panel estimates if you just do a difference in differences by constructing a panel of individuals that were observed in you know that was pre and post two periods you find you see that there's a decline in the probability of male employment which is greater than a decline in the probability of female employment. So you see a closure of gender gaps in the post pandemic period but this is driven by a probability a decline in the probability of male employment rather than an increase in the probability of female employment and coming to panel B we see that this result is driven by men with low levels of education that is with 10 years of less of schooling and this is not surprising because as a result of the lockdown it's really the informal sector in India which is very large as well as factories etc that closed down and these are predominantly employers of men with lower levels of education so you see that getting reflected here. Instead of just doing a pre-post panel one could also do a kind of an event study analysis where you look at each wave and I'm you know I ran a couple of models here but I'm presenting the effects the results of the dynamic panel data model because lagged employment is the strongest predictor of current period employment and you see that accounting for lagged employment effects you know women in august 2020 were not 9.5 percentage points less likely to be employed compared to men so this is you know comparing august 2020 to august 2019 so it's a pre-pandemic month however by december 2020 the gender gap in december 2020 in employment was the same as august 2019 so same as the pre-pandemic so by december 2020 the gender gap had gone back to the same level women were not more likely but the gender gap had gone back to the same level as far as employment is concerned if you look at hours spent on housework you see that again a well-known fact about south asia india pakistan in particular which is that there's great inequality in the hours spent on domestic chores so women spend far greater amounts of time on domestic chores than do men and you see that gap here in the december 2019 figures by in april 2020 which was the first month of the strict lockdown where domestic helpers could not get to their places of work you find an increase in the male hours spent on housework this seemed like a pleasant change uh green shoots of gender equality in the making except that by december 2020 the gender gap in domestic chores had actually worsened relative to december 2019 when we look at leisure which is here measured by time spent with friends the picture shows you the estimates for rural and urban areas separately you see that in both rural and urban areas in december 2019 women would spend more time with friends than did men and in april 2020 the first month of the strict lockdown you see a sharp decline in both male and female hours of time spent with friends by the time they come to there's a slight slight increase in august 2020 but again in december 2020 you see a decline uh and so which is still not as low as the worst period of april 2020 but it's still not nowhere near the time spent in uh december 2019 which is a pre-pandemic period and this has adverse implications for emotional well-being and could you know well contribute to stress anxiety and feeling of isolation that anyway engulfs all of us as a result of the pandemic so in conclusion what we find is that uh the covid 19 pandemic did not uh reverse the established gender gaps in employment uh pre in the pre-pandemic period so the gender gaps in employment in december 2020 were back to the pre-pandemic levels um looking at domestic chores the gender gaps had clearly worsened as the economy unlocked uh in terms of leisure the time spent with friends uh the gender gap gender gap had declined uh but uh both men and women were spending far less time on uh with friends compared to uh their time that they spent in the pre-pandemic period so uh you know overall the picture is fairly grim and it certainly doesn't point towards any reversal of the of the you know the deep rooted gender inequalities in the indian labour market and in indian society uh you know just casually looking at the recovery in employment since december 2020 you know just look at the numbers we see that the recovery has been uneven and not robust so it doesn't seem to be probable that there will be a you know in the any any any uh time in in the near future a reversal of the historical gender inequalities that have plagued the indian labour market so uh i think policy needs to very clearly focus on what can be done to remedy this very stark picture of inequality in labour markets and in uh unpaid work at home thank you so hi everyone thank you for attending this presentation hopefully you can see my screen uh before starting i would like to thank you no wider for organizing this conference and for giving us the opportunity to present this giant paper so this paper looks at the effect of social distancing measures on poverty and labour income inequality in black countries so as we know the the COVID-19 pandemic has had a highly important consequences in terms of the economy costs and due to this pandemic governments have taken measures such as implementing lockdowns and this has had a tremendous impact on the lives of many individuals so at the aggregate level for instance we have seen a decline in GDP it also has had a different effect across workers and this is what we're going to look at so we're going to try to measure here the effect that the pandemic has had in the labour market and in the in the latest step how this will have an effect on the potential increase in poverty and inequality so there has been a growing literature um during the pandemic trying to estimate the asymmetric effect of the social distancing measures i'm going to be brief about this but basically um there was one main paper from Diegel and Neyman to trying to assess the the different today working ability across occupations in the US many applications have been have been proposed since then in other country settings there are two papers that are mostly uh that are the main papers that are related to our paper the one from Palo Mino Eyal and Diegman they both try to estimate the effects on poverty and inequality of this differentiated impact on the labour market so in this paper we evaluate the potential distributional consequences of social distancing on poverty and labour inequality in 20 lakh countries we're going to follow Palo Mino Eyal and construct the lockdown working ability index so this is an index that represents the the capacity of individuals to remain active under the first phase of the lockdown so i will briefly describe how this measure has been constructed in our paper one thing that we that we do is that we're going to compare the formal de jure lockdown policies when we assume that perfect compliance um has taken place with the factor lockdowns so when there is some degree of non-compliance so we're going to change a little bit the expression of the lockdown working ability index and once we have identified individuals that have been able to remain active during the lockdown we examine changes in poverty and labour income inequality so we first start by computing the teleworkability index so we use information about the task content of occupations from the step surveys because they are collecting information on two lakh countries Bolivia and Colombia and we construct our teleworkability index so once we've done this we're able to to see that it varies across countries on average for the lack region it's around 12 percent so 12 percent of people are able to to perform their job from home and it varies from 7.5 percent to 16 percent roughly um so teleworking is not the only determinant of being able to remain active during the lockdown so we also have to take into account workplace closures and mobility restrictions um that have taken place during the lockdown so um workplace closures what do we mean by this basically essential workers were able to continue to work regardless of their ability to work from home in the opposite they were closed activities so the workers in closed activities were not able to perform their job so we're going to classify these activities into open closed and not open not closed categories for each country so this is depending on the laws the decrease and the press release that we have seen in each country and so to give you a rough idea of how the lockdown policies have differed across countries here you have the lockdown intensity and duration for each country in our in the lack region so we can see that for instance Nicaragua hasn't implemented the lockdown um on the other extreme of the spectrum we have Argentina for instance that has 29 percent of their workers that are in closed activities so they were not able to work in terms of the duration for instance Guatemala has implemented the longest lockdown here we're talking about the first phase of the lockdown so we construct a lockdown working ability index and the perfect compliance so we assume that everyone is respecting the lockdown rules so here for open activities we assume that everyone is able to work for closed activities no one is able to work and for those for the activities that are not open and not closed it's going to depend on the ability to perform the job from home so when we compute this lockdown working ability index we can see that the potential to to remain active varies across countries and it's around around one workout of two that are able to work from that are able to remain active under the lockdown in the lack region so this is if we assume that if we're assuming that there is perfect compliance however we know that there are some differences especially in developing countries between the facto and de jure lockdowns so this is something that we're quite interested in in this paper if you look at the how very how compliance varies over time over the first phase of the lockdown you can see that it's steadily decreased in all countries in the region so this is something that we have to take into account so we modify the expression of the lockdown working ability we're going to assume that in open activities everyone is able to remain active this doesn't change however for the closed activities we assume that it's going to depend additionally on the level of non-compliance in each across regions in each country and for those that are working from home it's also going to depend on the level of non-compliance so once we calculate this lockdown working ability correcting for imperfect compliance we can see that all the the proportions of individuals that were able to remain active has increased in every country every country so here we have for instance a larger increase in Venezuela or Brazil countries where you had a higher level of non-compliance so once we have identified the individuals that were able to remain active we're going to calculate the the potential labor income loss so this is going to be calculated by taking the individual annual labor income in t minus one so before the lockdown and we're going to correct this we're going to remove the the fraction of their annual labor income that they lose that they lost due to their inability to to to stay to remain active so this is one minus the lockdown working ability and this is multiplied by the duration of the lockdown so for how long they were unable to remain active we compute this potential labor income loss under perfect compliance and under imperfect compliance to compare the two scenarios here you can see that we were examining so we're computing lockdown incidence curves this enables us to see which part of the labor income distribution has suffered the most so in Argentina for instance you can see that the bottom parts of the labor income distribution has suffered much more than the top part of the labor income distribution and they have suffered even more in the scenario in which you assume that there is perfect compliance this is due to the fact that in imperfect compliance mostly the the most vulnerable workers are not compliant we calculate the impact on we we estimate the the impact on poverty and labor income inequality and we can see that in most countries we see an increase in poverty and inequality and inequality and the changes are higher in the perfect compliance for the same reason the most vulnerable workers are the ones that are usually the non-compliance we also decompose overall inequality into a between and a within country component and we find that social distancing has led to an increase in both inequality between and within countries yet the between country inequality components has increased significantly more than the within country component and we observe that the changes are greater in the perfect compliance so to summarize our results show a sizable potential increase in poverty in almost all our countries we also find that labor income inequality has increased in most of the countries the changes in poverty and labor income inequality are larger and the de jure compared to the fact of lockdowns so this is de jure this is assuming that you have perfect compliance so this basically highlights the fact that non-compliance acts as a mechanism to smooth labor income losses during the pandemic we have seen that social distancing measures are also likely towards an inequality in latin america and the caribbean both between and within countries and so all these findings in our paper we argue that it has important policy implications most notably to the need to assist the most vulnerable workers in lat countries in the future thank you very much thank you very much our presenter for this very nice presentation so we have a few questions in our q&a and i have also a request for for questions i will go through this and then you have a few minutes to answer them so for marae we have a specific question from test vanish it's like how do you measure child hunger in raw research then for ashrini we have a few questions if from richard frone is there any evidence that women have been relatively less likely to move into new jobs due to gender norms in jobs for example also from richard frone we have a question about if you have the paper and you could share the link and from nikita we have thank you for our representation ashrini a surprise to see men increase time in domestic shores with no effect on women during the initial lockdown for is so we have a very interesting word have you been have there being a forelaw skins in any of the countries and if yes do you consider them in your paper and i will give the word to rash katula who's requesting for intervention and then you can answer all these questions rush okay then maybe we wait and you can start with your responses anyone i'm i'm happy to go first where do i post the link to the paper i can't do i just click or ask the question sorry i just sort of asked a question i would recommend i would recommend just posting it in the general chat in the chat okay so instead of the q and a okay okay so this is my the link to the paper which is in the chat box in the question on the women's time not changing in the first month and nikita the point is that women already do a huge amount of domestic chores yeah it didn't go up significantly but it could be just that they were just anyway doing everything earlier and now you know they continued that so there wasn't that much of a change for the women but it was some change for the men but it didn't last at all actually went back and it's actually you know worse in december by december it had become worse than the pre-pandemic period also remember that this is not a time use survey this was an employment unemployment survey which basically just asked one question roughly how much time do you spend on domestic chores and the men's understanding of domestic chores might be very difficult from the women's understanding of domestic chores and women always underestimate the time that they spent spend on domestic chores everywhere and men always overestimate the time that they spend on domestic chores everywhere you know even in the u.s after the pandemic in the new york times had a survey and asked partners about how much time they were spending on homeschooling and the divergence in them the the men's and the women's answers were staggering uh you know in the in the same family and so you know take your pick about who was who was right about the time they were spending so it's a rough measure but it indicates something on the question of did they move to were there gender norm related constraints in other jobs i don't think so i don't think there's any evidence in fact employment creation has itself been an issue you know um in india ever i mean from before the pandemic but that's worsened during the pandemic so it's really no jobs you know rather than uh jobs being there which women i mean that's that does happen on the margin uh particularly transportation constraints are critical uh for rural women they can't get to the place of also even if hypothetically there was a job that they could do in the district center they can't get there so those things do matter but overall there is a huge excess supply of labor you know they're just unemployed people uh men and women both so it doesn't seem to be a gender norm story and i actually have another paper which i didn't present here where which strongly questions the gender norm story i can talk about that later but i don't want to monopolize on the time thank you very much uh moray maybe you can answer your question yeah thank you very much so a really good question about how we measure child hunger and in this instance one needs to understand that this panel survey was being implemented even under lockdown conditions where we would have been breaking the law to do in-person fieldwork so it was a telephonic questionnaire and it and it bears that uh those weaknesses in questions like that so the measuring child hunger would be related to a question that says in the past month or in the past two weeks have any uh of the children in your household experienced hunger rather than saying has anyone in your household experienced hunger um and uh so so that's how it would be measured with all the weaknesses associated with that in the longer run panel studies one can do anthropometrics and one actually measures uh the evidence of sustained hunger um a quick answer then to a very interesting question that was also put in the chat about furloughed schemes schemes for furloughed workers so obviously if you're going to do lockdowns you suspend employment and employers respond to that in different ways some of them suspend earnings at the same time um and uh and so i think it's a very interesting question because this issue of furloughed work is well well discussed in developing in developed country labor markets but completely not very well discussed i don't think in developing country labor markets and um so in south africa there was a specific scheme called the temporary employment relief scheme which which allowed but it was mediated through employers employers applied for relief for their workers under the lockdown and uh so that was a a scheme thought of on on the go as it were in the first weeks of the lockdown um but maybe others have some interesting examples from around the world about that thank you so yes um so to to answer shortly uh we haven't really included uh furloughed schemes and social assistance programs uh this is something that we discussed in the paper how this would affect our estimations the the only reason we didn't include them is uh is the fact that you can't really identify from surveys who has been receiving uh uh i mean who has been uh benefiting from a social assistance program or has been a um part of the furlough scheme uh so this is something that we we were not able to do because we were using surveys uh not uh surveys that have been conducted during the pandemic but surveys that uh we conducted one or two years before in order to have uh the first estimations uh potential estimation on uh poverty and inequality so so of course uh based on the thinking about taking this into account uh it's it would uh it might differ in the estimations that's a very quick last question there are other questions that we cannot address unfortunately but also uh for isor the holder Xavier Hara he asked about uh how the poverty inequality figures um so how does poverty inequality based on official survey data collected during the pandemic in Latin American countries compared to the estimates in the presentation how do you check that actually and that's a good that's a good question uh we we haven't uh we haven't really looked at the the figures because i guess now they are coming out and we we started working on this paper uh a year ago so that was like a really initial uh work but uh yeah it would be interesting to look at it and uh i i mean i think in terms of the incidence curves the way it would affect uh groups uh over the labor and condition distribution uh on the bottom part since they would be receiving social assistance uh i mean benefits uh so i mean yeah there they would be some uh some slight differences with the initial uh the the the estimates that have been produced now okay thank you very much clearly your paper was about the the initial effects probably there will be a lot of opportunities for following up with the effects during the next year so there are other questions that we cannot address unfortunately i included all of you who have interested in this paper to contact our great scholars that i'm sure they will follow up with you uh whatever is your question uh about your research so thank you very much all of you i don't want to take time from for the next session that is the poster session from some early career researchers mainly so thank you very much and see you soon i hope you enjoy the rest of the comments