 I welcome you all to this second day of the conference on development challenges in Africa in the wake of COVID-19 pandemic. Yesterday, for those of you who are able to attend yesterday's sessions, we had insightful and fruitful discussions, largely focusing on the economic impact of the pandemic. So today we have two main sessions. The first session will focus on the impact of the pandemic on livelihoods and welfare, and the second session will be on the new normal and the future development of Africa. In this first session, we have three presentations. We are privileged to have three speakers. Because the first presentation is going to be made by two speakers will be making these presentations. The presentation is on the effects of the pandemic on employment and earnings. And this will be made by Michael Danqua and Simona Shota, who are both UNU-wide researchers. The second presentation is on the mitigating role of tax and benefit rescue packages for poverty and inequality in the wake of the COVID-19 pandemic. The third presentation will be presented by Jesse Lastunan, who is also a UN-wide researcher. And in total, these two first presentations are actually based on the part of the ongoing research work at UN-wide on COVID-19 pandemic. The presentation will be on gender perspectives of COVID-19 pandemic. And this will be made by Isabella Smith, who is from UN Women. Ladies and gentlemen, at this juncture, I would like to introduce to you all the three, the four speakers that I've just mentioned. I'll start by Michael. Michael is a development economist and a research fellow at UNU-Wider, currently serving as a focal point for the project on transforming informal work and livelihoods. Michael is also a visiting research fellow at the Transfer Project and a researcher for the International Growth Center, IGC, Ghana. Previously, he has worked at the Department of Economics at the University of Ghana. His research work has been published in several journals such as Empirical Economics, Economic Modelling, Review of Development Economics, African Development Review, Journal of International Development, Technological Focusing and Social Change, among others. Michael has also been recognized as the most promising young scholar and the best researcher in the School of Social Sciences, University of Ghana. Welcome Michael. Michael will be making this presentation with Simona. Simona is an applied micro-economist with research interests in development and labor economics, working specifically at the interface of poverty, inequality and employment dynamics research. Before joining UN-Wider, she worked at the German Institute for Global and Area Studies and was a consultant to the workbench. Simona's main research focus is on understanding the drivers and livelihood implications of transitions between different types of formal and informal work, as well as the interlingual between occupational change and earnings in equality dynamics in developing countries. She has published in journals such as the World Development, Journal of Economic Inequality, Journal of Development Studies, among others. Welcome Simona. Our third presenter this morning will be just a last name, who is a research associate at UN-Wider, where his work focuses on tax benefits micro simulation models. He has previously worked for organizations such as OECD, RAND Corporation, Internet Association and Technopolies Group, conducting policy oriented research on labor markets and emerging technologies. He received his PhD in policy analysis from RAND Graduate School with analytical focus in economics and quantitative methods. Welcome Jason. Not least is Isabella Schmidt, who is from UN-Women. She has a PhD in international development, MFIL degree in urban and regional science, and an MSc in human nutrition from Northwest University in South Africa. Prior to joining UN-Women, she was the chief director, social statistics, South Africa for 10 years. Her primary responsibility was to oversee the production and promote the use of social statistics in South Africa. Prior to joining statistics South Africa in 2008, she worked as an international consultant in Africa. And in various parts of the Middle East and also South Africa. Her specialization at that time was mainly on monitoring and evaluation of development projects. She has also worked in other areas, in other countries like Sudan, Mozambique, Kenya, Uganda, Angola, Democratic Republic of Congo, Peru and many others. And I was being involved in food security and rural development projects. Welcome Isabella. Now with that introduction. I'd now like to remind you about some basic housekeeping. So we are going to have 20 minutes per presentation. And then after the three presentations we shall have a Q&A session at the end of all the presentations. Over to you Michael. Thank you. All right. Many thanks, Marine, for the kind introduction. Many thanks. So let me first say that this paper falls under the UNU-wide transforming informal work and livelihoods project. And the full working paper is actually available. So for those who would want to have access to that and then engage with us, you can find that on our webpage. The authors of this paper is Simone. Simone is here with us. So I would do the first two or three introductions and then Simone would take it up with the methodology and then the discussion of the results. And then Robert is at the, you know, Issa is at the University of Ghana and then Kunal is at UNU-wide. Next one please. Right. So let me say that the COVID pandemic has posed a significant risk, not only for people's health, but also to the economic well-being as well. One thing that has been common across many countries is workplace closures and then restrictions on the movement of people as well. But these have actually affected the livelihoods of workers in particular across the globe. So what is paper the, you know, six to do is to provide a causal evidence on the impact of on the impact of stringent lockdown policies on labor market outcomes. And, you know, here we use Ghana as a case study, and we look at how the COVID pandemic and its related government measures, i.e. lockdowns and other stringent measures have affected workers in Ghana. Then we also look at to what extent of workers being affected differently based on the type of government measures in place. So here we are looking at the government stringency measure in place, whether, you know, a lockdown or not a lockdown or a partial lockdown. And then we also look at the type of activities that workers are engaged in. And here we paid much attention to the, to the, to the, to the formality status. So we were looking to formal informal and then informal self-employed and then the informal wage, you know, employed as well as well. Before, before we even go on, let me give you a bit of a background on how things unfolded in Ghana. The first cases was reported on the 12th of March, 2020. And then from the 15th of March, the government began to roll out some of these stringency measures. So yes, public gatherings and were banned and then schools and universities were also closed. The borders were also closed as well. And then from the 30th of March, what actually happened was that there was a partial lockdown in what the government called the hot spots. So these were spots that were reporting very high numbers of what you call the daily infection rates. And these were greater across and then the, the greater commercial area. So there was a partial lockdown from the 20th of March to the 19th of April, where these particular locations were locked down. You know, to what you call the restrict the movement of workers. So if you look at the figure that we have over there, one can see that yes, right from the 12th of March, these stringency measures were rolled up. And then from the 20th to the 19th of April, there, you know, one can see that the, the, the, the, the stringency measure was very high. And then right after that, the measures were begin to go down here. And so one thing that I can say is that yes, although now the, the, the number of COVID cases have gone up, Ghana is really important. We have a fairly rate of around 700, you know, but the stringency measures are not up to the, the state where we have the lockdown as well. So one thing that is actually important to this work that we would present today is the data. So what, what we did with this paper was to work with, with the, you know, ESA, ESA is at the University of Ghana. ESA working with North Western University has what is called the Ghana's socio-economic panels survey, and they have three rounds of that. Okay, the first round is in 2010, and then there's the 2014, then the latest wave was in 2019. So we do draw from the latest wave, and then we conduct, you know, rapid phones survey between the 19th of, you know, August up to the 17th of September, 2020. In all, we did have 670 respondents in this case. One thing that I would have to highlight here is that we were able to ask retrospective questions regarding the COVID situation in February, you know, the, the whole thing started, and then in April where there was the lockdown and then in September as well. So we were able to get all these retrospective questions, which is, which was very good for the work that we, you know, wanted to do. So based on this, we were able to identify the effect of the lockdowns, or the, what do you call it, the partial lockdowns by comparing the labor market outcomes in areas with different policy measures. So let me, I would want India and then Simone would look at the methodology, and then the discussion of the results and then we can come back and speak more to the paper. Thank you Michael for the first part of the presentation and for setting the scene of this paper. So as you already explained, what we're doing is we compare the labor market outcomes of respondents who were located in districts that were subject to this partial lockdown, which we call the treated and respondents located in districts where no lockdown had been implemented, which is our control group so we're using what's called a difference and difference design here. As I already explained, we can draw on information from three time periods essentially. So one is this recall from February 2020, which is the base period before COVID-19 had reached Ghana. We have April, which is the time that the partial lockdown was in place, which is our first post treatment period. And then we have the time of the interview so respondents were asked about the last seven days prior to the interviews that were conducted in August and September 2020, which is when most of the various most stringent policy measures had been relaxed. So this is our second post treatment period. As for completeness, you see our, the regression that we are estimating so we control for location, lockdown and unlockdown. We have at the end this theta t is the time fix effects for the different periods, and then we interact these two post treatment periods with our lockdown measures so this will be our main outcome variables of interest, where we see how respondents are treated in lockdown districts are treated fair compared to the control group. Given that we have information for the same individual for this three time periods, we also estimate an alternative specification where we in addition control for worker fix effects so there. The location as it as we focus on people who didn't move. The location is fixed lock so this lockdown drops and then we still look at the treatment effects that we are interested in. Okay. So this will be an idea of the study site that we're talking to as Michael explained the lockdown, or the partial lockdown was of the two major metropolitan areas in Ghana so we have greater upper and greater commassive metropolitan districts and the continuous districts that were in the lockdown so this is left panel marked in red here. And then, of course we had to take into account that those are the major urban centers so they may be different in terms of infrastructure and so on so we looked at the population density across districts which is a second panel taking from the district census information. And what we did is to have a more comparable control group was to set a population density cut off at the lowest population density, observed in the treated districts. So here in the other districts we surveyed and dark blue is a control group and red is a treatment group. And we can actually show that like time trends were more similar and comparable among this special control groups that we defined and I'm going to show you the results for the broader group and for this limited control group. Just to give an idea of like the average demographic characteristics of the sample and to what extent these are similar across treated and control respondents. You see that about half of the respondents in our sample were female the largest majority are heads of households which was the target group of our sample selection. Something important that we noticed is that the household size tends to be smaller in treated districts or in lockdown districts compared to no lockdown district so this may imply that actually people moved out of the lockdown districts to those that were not under lockdown and this is also what we see like there's a larger share of people who moved since the last interview in those districts that were not under lockdown. And we had a robustness check to kind of control for this to make sure this is not biasing any of our results. And then as Michael explained to focus on people who were actually working so they were working on the last panel survey that was done in 2018 2019. And most of those people were still working, at least in February 2020 so above 90% of our example we're working, but 25% on average we're in formal work, and about 40% were weight weight employment, either formal or informal. So this is the type of person that we're talking about. This is a general background to one of the first questions we asked people was what aspect of the corona virus pandemic actually had the greatest impact on them personally. And about two out of three sets of unemployment or loss of income was actually the largest concern and this is surprisingly similar in both lockdown and more lockdown areas. And we find some differences that people in lockdown areas were somewhat more concerned with restrictions on movements, and this being sick of getting sick but they're not really any significant differences here. However, despite this general concern expressed by respondents in both locations, what we see is that the actual employment effect that materialized in April is very different between the two areas. And here on the left side, you see the share of respondents who were employed or reported to be still actively working in April 2020. And we see that in no lockdown districts above 70% we're still working whereas in lockdown district, it had dropped from close to 90% to 34% of the respondents. So there's a sizable and statistically significant effect on the employment probability ability in April, and this is also confirmed by our regression analysis. So yeah, don't show you the full table just the main result of interest here and as I said the interesting part is this interaction term between lockdown and April and what we find is that those located in lockdown districts had about the 35 percentage point chance of still being working in April 2020, in addition to the general drop that was observed, which is around 28 percentage points. So we can say that workers in lockdown areas were more than twice as likely to stop work in April, when the lockdown was in place. Another thing we were wondering about is what were the reasons that people had to stop work and actually most people said, or the very largest majority said that workplace or business closures due to government regulations was the main reason and again, this is true both in lockdown and no lockdown areas. And as you may remember, also in the no lockdown areas, still confinement measures had been implemented just to a different degree. However, what we observe is that the type of workers who stopped work are significantly different in no lockdown with those lockdown districts. So, in those where lockdown policies had been implemented, those workers who were able to continue work were largely informal wage employment, which is kind of the most stable and secure type of employment that we identify. Whereas those who were most likely to stop work are those in informal self employment, which has also by other studies been identified as the most vulnerable form of employment. So these are workers that are, for example, street vendors, who rely on making a living on a daily basis, and who are often in very contact intensive jobs that were heavily affected by the lockdown. On the other hand, as you can see in the no lockdown district, this group was the one who was most likely to continue working, given the need to make a living and to earn a living. So in lockdown areas, more than or close to 80% of all informally self employed had stopped work in April 2020, compared to only 28.3% in the no lockdown areas. And we also try to find a more fine grained distinction within the informal sector, but we couldn't find any statistically significant differences there. Looking at the recovery, so this is now the time of the interview, the last seven days prior to the interview that was conducted in August and September, we observe that most of the workers were back to be actually working. About 86% of those in no lockdown areas, and 84% of those in lockdown areas reported to be back in employment, to be in employment at the time of the interview. And here we observe at this point, no longer any statistically significant difference and the probability to be working between the two areas was who had different, which had different policies in place. So this time the fact that we had identified for the April period has vanished in August compared to April. However, what it also implies that remember before we had like close to full employment since the employment rates were clearly still lower compared to the February level. And we observed that 18% of all men and actually 29% of all women who had stopped stopped working in April were still out of work in August and September. So this is why we do see a recovery, you also still observe an effect or a lower level of employment compared to February. So this difference between lockdown and no lockdown areas had disappeared at this point. We try to see like what kind of factors were facilitating this recovery or which workers were more likely to return to work in time. And what we observe is that generally, those who were on paid leave were quite likely to who said they had been on paid leave in April were quite likely to be again actively working at the time of the interview. So those who have said they've had this layoff was only temporary, we're still out of employment but those who said they had permanently stopped was the most likely to still be out of work at the time of the interview. So the perception of whether the layoff had been temporary or permanent actually also was visible in the recovery rates. It's interesting that we observed the set the interaction between lockdown areas and temporary stopping was actually positive. So those in the lockdown areas who said it was temporary were more likely to again return to work than those in the no lockdown areas. So this kind of recovery of those who had temporarily stopped because of the restrictions was stronger in those areas that had been subject to the lockdown policies. As discussed before we still observe that employment was lower. So here, we just look at the more near term effect between February 2020 and the time of the interview. And what we see is that at the time of the interview still employment was about 11.6 percentage points lower than it had been in February, but also we observe a statistically significant effect on earnings and on working hours. Earnings were still about 36% below what they had been in February and working hours were about 14.3% lower than what they had been in February 2020. So there's a persistent long term effect on labor market outcomes of the pandemic, which though doesn't seem to vary by the policies that had been implemented on the stringency level of the policies. And just as a last note, this long term or this near term effect was more crimes for those in self employment and for women actually so those remain heavily affected in the medium term. In fact, we do a number of robustness checks. So we look, we first check that there's some different identity. No, I'm supporting conference. That's a comment. Confirmed. Then we check for self selection as I mentioned at the beginning so we distinguish, we take out people who moved also finding the same patterns and we also exclude the major metropolitan districts. So the very city center of Acre and Kumasi, just make sure that those are not writing our results. Let me quickly summarize. So, what we do is we provide causal evidence of the impact of string and lockdown policies on labor market outcomes, making use of the specific policy settings that actually allows us to identify the causal effect. What we find is that the lockdown measures implemented heavily affected economic activity and that the shock was felt the most by workers in the most vulnerable forms of employment. So specifically workers in informal self employment to are the most in need to earn a living on a daily basis where most often forced to stop the activities during the lockdown, where those in more stable formal employment were most likely to continue. Overall, there has been a strong recovery and employment up to August or September 2020, even though employment levels, working hours and earnings remain significantly below the pre COVID level so the assistant effect. And this effect has been more felt by those in self employment and by women. And from this we conclude that the COVID-19 pandemic and the related government response measures tend to have accentuated existing vulnerabilities in the Ghana labor market. Thank you. Thank you very much Simone and Michael. Let's have the next presentation please. So you have 20 minutes. Thank you Maureen. I hope everybody can see this. Simone and Michael too for the previous presentation, very interesting work. So my name is Yesel Astonen. I'm a research associate that you are new wider. This presentation is about our paper to the rescue where we are looking at the effects of COVID in a few African countries using micro simulation. One thing I want to note first is that the work I'm presenting actually comes from a collaboration of more than 30 people around the world. So we are a pretty large group. Another thing is that the work I'm showing is currently in progress and so some of the results I'll be showing are kind of preliminary and there's a lot more analysis to come. So I'll talk a little bit about our next steps of our work at the end of the presentation. So first some quick background. So this study is being conducted as part of the Southmont research project at UNU wider. In the Southmont research project, our main focus is on developing tax benefit micro simulation models for developing countries. For those not familiar with these models, I've listed here a few of their uses. So micro simulation can be used to get information on who pays taxes, how much do they pay in taxes. And also who receives benefits, how different taxes and benefits affect the government budget. And we can also estimate the distributional impacts of changes in taxes and benefits. For instance, different social protection reforms. Another thing to note is that our work in the Southmont group is very collaborative. As I mentioned, we are a big group. We work with research teams from different partner countries as well as organizations like SASPRI and the University of Essex listed here. We also have a big focus on capacity building, basically organizing workshops and training sessions for people to learn to use our models. And one heavy focus since this summer has been the COVID study that I will be talking about in this presentation. So we are studying the effects of the pandemic on inequality and poverty in the six African countries listed here. There are also two non-African countries, Ecuador and Vietnam that are part of the study. In the case of Ecuador, we have actually just published a working paper on the short term effects of COVID-19 in Ecuador and I'll be sharing the link with you later. So here's more background on the study itself, specifically the part that focuses on the African countries. So our ongoing work covers the six countries mentioned in a previous slide. Of course, these are all developing countries that are quite vulnerable to the pandemic, for instance, compared to many Western countries. Quite clearly, there are pretty high risks to livelihoods and potential for increases in poverty, especially if the pandemic is not managed effectively. And so for all of these countries, we use the dedicated micro simulation models I mentioned earlier. And each of these countries has a single national model that we are constantly updating, improving in collaboration with national teams in each country. And that's actually one of the very useful dimensions of this project that we can actually leverage the knowledge of these local research teams we have in different countries, including in Africa. And overall, when it comes to the COVID research, for instance, these teams have been quite heavily involved in the research process. In terms of the work itself, we have essentially two main study objectives. So for each country, we estimate first the effects of the pandemic and related lockdown measures on poverty and inequality. And second, we also analyze the contribution of this new COVID related tax and benefit policies in mitigating these negative impacts. The micro simulation models allow us to look at outcomes for different groups in the population, basically different parts of the income distribution, different demographics, informal and formal workers, and so on. So next here are our methods or kind of the main steps in the analysis. Firstly, we developed so-called pre-crisis data sets for 2020, which we use as a counterfactual. So these data sets allow us to obtain scenarios in the models that show how economic outcomes would have looked like in the absence of COVID in each country. And we developed these data sets using the newest survey data available, usually from 2014 to 2018, but weighting this data so that it matches with population estimates for 2020. Secondly, we developed so-called crisis data sets that do account for the effects of COVID last year. This means taking account the effects of lockdown measures and restrictions, which mostly reduce the economic activity in these countries. I won't go into the methods here in detail, but at the moment we are using World Bank's pre-crisis and crisis demand predictions to estimate, firstly, how different industries were affected by COVID, and we then reduce individual incomes within industries kind of accordingly. Thirdly, we compile information on different tax benefit policies that were introduced in response to COVID and then include these changes into the models when possible. And finally, we use the pre-crisis and crisis data sets and also the new policies in the micro simulation models to ultimately estimate the distributional effects of the pandemic and of the different policies. So if you look at the final step in more detail, what we do in a nutshell is to run the modeling scenarios with and without the shock from the pandemic and with and without different policies in place. And by doing this, we can answer a range of different questions. For instance, how much incomes and poverty would have changed without any government intervention? To what extent do these normal tax and benefit policies help mitigate the shock from COVID and also how much additional relief is offered by the new tax and benefit measures that were implemented specifically in response to COVID? So here's the list of methods again. Next, I'll be talking about essentially what we have found so far across these different steps. So some early results. If we start with the crisis data sets, the big question here is, you know, how the lockdown measures and restrictions actually affect different industries and individual incomes in these countries we are looking at. So as I briefly mentioned, we started by using World Bank's estimates to get some information on how industry level GDP or output was affected by COVID. Here are the preliminary estimates from each country. You can see that, for instance, South Africa that's on the left had pretty large impacts across different sectors ranging from 10 to 20%, while countries like Tanzania and Zambia on the right were a little bit less affected. Also, you can see that there's quite a bit of variation in the effects between different sectors. But again, these are just kind of preliminary estimates using aggregate level data and we are working on kind of improving these estimates going forward. Anyway, given these industry level reductions in GDP, next we want to translate them to changes at the individual level. And this is kind of the step required to actually compile the so-called crisis data sets. And I'll show you quickly how we do this using as an example this 14% reduction in GDP in the construction sector in Zambia, just as an example. So first, we assume that labor income in the construction sector is reduced in proportion to this GDP shock. And this means that overall the sum of wages will be 14% lower compared to the pre-crisis situation. Secondly, we assign randomly selected workers in that sector to unemployment with zero income, so that the total labor income in the sector is reduced by 14%. Then we basically do this for randomly selected workers within all industries and all countries that experience a negative GDP shock. And finally, based on the income reductions, we also adjust household level expenditures or consumption of the affected workers. And this involves some extra steps I won't go through here. So, going back to the list of steps in the analysis, the next stage is actually gathering some information on the tax and benefit policies in each country. So, firstly, in addition to kind of updating any existing tax benefit policies in the models for 2020, we also identified a models policies specifically related to COVID-19 in each country with the help of our national teams in the different countries. And here are just a few examples. For instance, in Mozambique and Ghana, utility fees were reduced or wait for consumers for the rest of 2020, soon after the pandemic started. Tanzania instead has provided informal support, for instance, to hospitals and orphanage centers, along with related tax exemptions. And finally, one policy for which we actually have some results that I'll share with you later is the emergency social gas transfer in Zambia. And this was provided to specific poor households that are already receiving social gas transfers. It's around 20 US dollars per month per household for a period of six months total last year. So, yeah, these are just a few examples of the COVID related policies enacted in the different countries in our study. Then, if we move back to the list of steps, I'll move on to some early results from the actual micro simulation stage. But before that, just a few reservations I want to mention here. One is that poverty results we have are based on national poverty lines and national equivalent scales. And this means that currently the findings related to poverty are not comparable between countries. The final comparative paper that we are working on now, however, will use a harmonized poverty line and harmonized equivalent scale. You also saw the industry level shock estimates we have had across countries and how we obtain individual income changes based on these shocks. There are some limitations with these methods that affect and basically limitations with the data that affect the accuracy of the results, possibly underestimating the shock a little bit. And obviously we are working with pretty limited data at least since last summer. However, especially over the past few months, we've been kind of lucky enough to get some new more detailed data from many of these African countries we are looking at. For instance, World Bank has collected individual level survey data on the effects of COVID specifically on employment and incomes in at least Uganda and Tanzania for now. And we are currently in the process of using this data to kind of better understand what types of workers are affected and also how they are affected. So for instance, how many people lose their employment and incomes in full as opposed to losing only a part of their income as a result of the pandemic. But okay, finally some preliminary results. I'll show you two types of outcomes. First, increases in consumption based poverty and inequality due to COVID and then also the composition of changes in household incomes into effects in different parts of the income distribution and also for different types of workers for instance informal and formal workers similarly to the last presentation actually. So here are the results for Tanzania first. This is showing changes in inequality and poverty in 2020 due to COVID kind of based on our modeling and the assumptions we've made so far. As you can see, there seems to be a very small impact on inequality, both the genicoefficient and the ratio of income in the 18 percentile of households compared to the 20th percentile impacts on poverty were a bit larger based on our estimates. The first row under the poverty title for instance shows the poverty rate, FGT0 and you can see that the poverty rate has actually increased by 2.2 percentage points due to COVID based on our analysis. Here are some decomposition results. First, this one shows the reduction of disposable household income in different portals of households. The gray bar on the left for instance shows the average impact on the poorest 25 percent of households. Quite clearly however incomes have decreased the most in the middle of the income distribution by around 4.5 percent essentially. Another finding is that automatic stabilizers, that's the blue small bar you can see on the right, they've had a very limited impact in terms of kind of alleviating these income losses due to COVID. Basically only in the top quartile households have kind of paid slightly less taxes, less social insurance contributions because of these earnings losses due to COVID. Here's this figure again but it now shows the income losses separately for different types of workers. For instance you can see that much of the losses are concentrated on informal employees, especially farmers which is shown in the light blue bars at the bottom and also for other informal workers shown in the kind of lighter gray bars there. So to save time for Mozambique and Zambia I will only show these tables with consumption based results, consumption based changes in inequality and poverty. I'm showing Mozambique simply because it's essentially the most affected country in our study so far. So because of COVID the genicoefficient for instance increased last year from a bit over 0.5 to almost 0.8 which is pretty huge and similarly the poverty rate was substantially larger than it would have been without the pandemic. Finally here is Zambia for which we also modeled the emergency gas transfer I mentioned earlier. This one shows the total change that also includes the kind of compensating positive effects of this new gas transfer. And the changes in inequality were not huge but still noticeable but especially if you look at poverty you can see that poverty rate has increased from 41.5% to 48.8% which is pretty huge. However these increases in inequality and poverty would have been slightly larger without the gas transfer. For instance you can see that the poverty rate was reduced by 0.3% its points because of this transfer. And while it's not shown here the effect of the policy was particularly large at the very bottom of the income distribution. The poorest 10% of households the negative impact of the pandemic was basically fully compensated by this gas transfer. Okay just to kind of summarize we found modest increases in consumption based inequality and a little bit larger growth in poverty. It was a big especially the impact of the pandemic was quite large and the economic burden fell quite substantially on farmers in the informal sector. Then again countries like Zambia seem to be doing a decent job at least in protecting some of the poorest households from the economic shock. Unfortunately in many other countries we are looking at these types of policies either were not implemented or were of very small scale. And finally you know automatic stabilizers in Africa at least the countries we are looking at are quite limited in reducing the economic burden from the pandemic. And this brings us to the next steps in the study. My last slide I believe the first one is that we are really moving towards using micro data to improve our estimates of the shocks from covid and also to improve the estimates of kind of the specific labor market transitions that took place in some of these countries. Secondly, we are still in the process of modeling some policies similar to the emergency gas transfer in the remaining countries other than Zambia. Thirdly, one big benefit of these micro simulation models is that we are able to model alternative policy basically experiment with social protection reforms and tax reforms that kind of could have been enacted and possibly could have been more cost effective than policies already in place. And kind of this relates to the last point, which is, you know, the one of the bigger bigger goals in our studies to eventually also communicate these results to other researchers but also policy makers and public officials. So that's the presentation I'll leave you with some links here for instance there's the working paper on the effects of the pandemic and Ecuador. And also our recent blog post discussing the early stages of this work. But the link at the bottom might be helpful if you are interested in actually using these models yourself. So, yeah, that's it you can also contact me at yes at wider.unu.edu. I'm happy to answer any questions there or at the end of the seminar here. So thanks. Thank you very much. So next presentation is by Isabella Isabella please. So I'm the regional gender statistics specialist for you in women. And we are going to share with you findings of a recent catty survey that we did across the region to establish what the impacts of COVID-19 has been on women and men. So I'm basically going to share the findings of this study. And as you all know that due to lockdown restrictions etc. We did not use conventional data collection methods. And what we tried to do between September and November of last year was to reach out through to service providers to your poll and it source to a representative sample of women and men, demographically representative and of course limited to those who had mobile phone call initiatives you can go to the next slide. Okay, so our target population with generally 18 years and plus adult women and men. And we covered not only Ethiopia but South Africa, Kenya, Mozambique, Malawi. And we are currently busy with is what teeny but that data is not available yet. We basically use two server instruments. Because of the time limitations on an interview longer than 15 minutes on a phone call leads to respond and take. And our first questionnaire covered demographics economic activities agriculture and education. And our second questionnaire included again some of the key demographic variables, but then focused primarily on questions dealing with half human rights safety security and GBB. We tried to do a demographic panel in that in the respondents who were selected for the first interview were encouraged to continue to the second interview and in most countries we succeeded between six to to retain between 60 and 70% of the original, originally sampled individuals. We were not willing to do the second interview with then replaced with someone who had similar demographic characteristics. As per the quote us that we determined prior to this to the survey sample size was 2400 women and men and because generally the sex distribution in countries is about 50%. So the first interview of 1200 women and 1200 men next sample or next slide please. Okay, so most of the, all the service provider use random direct darling, and that is really our main claim to random selection. But in some countries, certain quotas, which we determined previously using national statistical office data could not was difficult to fulfill for example older women living in rural areas, generally do not have phones. And so to follow squatters we we sometimes had problems and then we reverted back to existing databases that may include women of that particular age group for it's just an example. I would say that 95% of our respondents were obtained via random direct dialing. We can go to the next slide. So one thing that we did specifically in terms of methodology was in Ethiopia we tested specific methodological aspects around gender based violence, as you know, survivors of gender based violence are quite often traumatized by questions related to gender based violence. We also know that sometimes the perpetrators may be present in the household, and if for example speaker phones are used. It may have implications for the women being interviewed. And so we test a few things around that. I can just say that the results of our testing are available in a technical report on the web. We basically found that it will be important to have a question that looks at whether speaker phone is being used during the interview or not. And the interview should be terminated if it's clear that there's a speaker phone being used. A second aspect that we tested was whether it makes a difference whether the interview is being conducted by a man or a woman. So gender matching or sex matching during the interview and we found that for most gender based violence questions that didn't make a difference. But there were a few where there was a marginal but not statistically significant difference. We can go to the next slide. I think the technical reports available on our website and anybody who is interested in more detailed information about these specific tests we did are welcome to approach it. So, in most of this presentation I'm actually going to present the data of the rapid gender assessment that we did using Cathy. But in some instances, data will be sourced from other sources and there it will be indicated. So what we do know related to gender and COVID-19 is that there is a sex dimension. And even though it varies between countries generally men are more likely to die than women. We also know that there's an age dimension older people more likely to be infected and die than younger people but when you look at this distribution on the screen. You'll notice that in South Africa, for example, there's nearly 50% ratio between women and men. And that is primarily because of the presence of comorbidities such as obesity and also diabetes, HIV and AIDS that influence the mortality rates in South Africa in terms of the distribution by sex. So when we look at the demographic consequences, we would see that yes clearly these mortality and morbidity consequences. In most countries, the number of deaths are relatively small and it's unlikely to influence the demographic profiles of those countries. However, we may see a surge in fertility rates due to out of school pregnancies and restricted access to family planning services during the pandemic. So that will have a demographic impact. In terms of the age structure, I already mentioned that for most countries, it's unlikely to have a big impact in South Africa. In South Africa, it may have an impact because it has a relatively older population than the rest of the countries in the region. And we do know that some of the theories around why COVID rates are lower in our region is because of our relatively young population structures. I think the two previous speakers spoke about the economic impacts of the pandemic and we expect that there may be an impact on migration. We know that women in refugee camps are more vulnerable than elsewhere. We also know that when men migrate, women left behind suffer particularly challenges in many of the countries in the region. Next slide. Looking at governance issues around gender and women economic empowerment, etc. There are several regional and global rankings that are done and through indices, but now we're specifically looking at the World Economic Forum 2020 global gender gap report. And that's particularly in relation to a gender sense of legal and regulatory frameworks. And even before COVID, you will see that these are substantial difference between countries in the region. We are rankings globally, which is the blue bars and our rankings regionally, which is sub-Saharan Africa is the black bar. You'll notice that in terms of those frameworks, they were big differences between countries with Mauritius being the lowest ranked in the region and Rwanda the highest ranked both globally and regionally. And South Africa also did well, but Kenya and Mauritius clearly need some work. And so within that context, we see the pandemic happening. And we know that these frameworks are not accommodated for women even before COVID so we expect more impact in countries that did not have good frameworks. Next one. Okay, so during the pandemic, we did some analysis with UNDP in our COVID-19 global gender response tracker. And we found for the countries for which information was available that there was also really a varied response. And when you look at the lightest blue bar and the second lightest blue bar on the far right for every country. We noticed that the areas that were most or the area that was most likely to have interventions put in place as violence against women. And then the total gender response measures, which is the far right column. We noticed that Uganda and South Africa was the most likely of all the countries in Eastern Southern Africa to implement significant numbers of specific gender interventions or normative interventions during the COVID-19 pandemic. Let's do the next one, please. So in terms of socioeconomic consideration, we also did some modeling about the percentage of women and men living in extreme poverty. Now the slide is unfortunately very, very busy. But I think if you can focus at the darkest blue bar, that gives you the percentage of women that have been projected to be living in poverty. In 2020 as a result of the COVID pandemic. And the second darkest bar represents the men. And I think what's really important in this graph is to see that there's been disparities even before COVID between women and men living in extreme poverty. And so moving into COVID, those gaps increased between women and men. And we can also see that, yes, they will definitely be a significant impact on both women and men, but more so for women as a result of the COVID pandemic. Let's go to the next slide. So here we are looking at the findings of our catty server that was conducted in several countries in the region. And it suggests that most women and men have reduced individual as well as household income, various between countries, but generally men were more affected than women. Workers in the formal sector were more likely to change the economic activities due to COVID-19 than any other sector. And the percentage of employed and not economically active individuals increased more so for women than for men. So in general, we also find that the Social Security Assistance Network provided by governments and other agencies reached very few individuals. The only country where it reached significant percentages of people was in Africa where they rolled out the $500 billion support package. We also found that the remittances that were received prior to the pandemic decreased during the pandemic. So all across the available money for social support decreased. Then the biggest source of worries for women have been becoming infected with the virus, but for men concerns about economic activity and income were the most important leads go to the next one. So changes in personal incomes during the pandemic here are the percentages. And we see that individuals who lost all personal income is the dark blue. Those who indicated that the incomes have decreased or downsized are the black ones. Generally, a country on the left will be the percentage of women affected and the second bar will be the percentage of men affected. So for example, you'll see in Ethiopia 61.4% of women had decreased downsized incomes and 64.1% of men. And so you can just go through all the countries indicated here. And you'll see that in most countries, the percentages who have downsized or lost the incomes are sort of higher in all countries except or then in South Africa. But in South Africa, a much bigger portion say that they lost all their personal incomes. Go to the next slide please. Okay, so financial difficulties and decreases in combined household incomes. Here we see that women and men reported generally whether the combined incomes of their households increased or decreased and these are actually decreases. So on the left dark bar for women would be percentage who experienced financial difficulties during the pandemic. The light blue bar for women and men would mean it's a percentage where the combined income of all household members has decreased. And so you see the impact has been big across the region. Kenya quite big Mozambique high. I think the previous speaker also said that they felt that there was a bigger impact in Mozambique than elsewhere. Let's go to the next slide. Okay, so here we just look at changes in the sector economic sector in which you were involved in dark blue work for someone else organization or a government. Next color is own account worker. Next one is worker in agriculture. Next one is not employed unemployed and then we have other activities and then you'll see everything registered as negative means that the percentage of individuals who said they worked for someone else decreased by so many percentage points. Between March 2020 which was our base before the pandemic was declared. Until the period we were collecting data which was between September and November of 2020. So basically, we are looking at a period of about six months. So the intervention took place immediately after we had significant lockdowns and movement restrictions being lifted. Let's go to the next one. Okay, and so when we look at time is I'm not going to do my presentation because I think everybody took a bit longer than we needed to. And what's important is that we also engage with the audience and have some discussions. So I'm going to quickly go through this. We regard unpaid domestic and key work extremely important and within the pandemic context we know that women and men have been affected by additional domestic and key work within households so the burden of care increased. Especially during lockdown and it's important from a women economics empowerment and sustainable development perspective because it basically prevents women from fully participating in the economy. Let's go to the next slide. So what we found and I'm not going to go through all the details in reading all the data but before the pandemic, they were already a significant burden on women and this represents our findings where women, the percentage of women who say they were mainly responsible for these duties before the pandemic. Let's go to the next slide. We also now see we unpaid care activities. So here we also see exactly that there's been a spread of increases or basically a load on women for unpaid care work prior to the pandemic if we go to the next slide. And you'll see that the time spent on unpaid domestic work has increased so here we see women and men compared. And actually in most countries, more than 50% of women and men all say that the unpaid care duties increased during the pandemic. Let's go to the next slide. And this is care work the previous honest domestic work let's go to the next slide. So what's important for us is that we highlight the fact that women are prevented from full entry and participation in the economy, because of the unpaid domestic and care activities. What we learned is during the pandemic both men and women have seen increased activity levels, but the fact that men increasingly supported women during the pandemic is something that we can leverage in our continued advocacy efforts to promote the sharing of groups between women and men as well as making sure that there are policy measures that address this such as the provision of childcare mechanisms that would enable women to be more economically active in the post covert recovery phase let's go to the next one. So I'm just going to briefly highlight what we think the impacts of the pandemic is on goals, especially the school closures, and if you then focus on the second set of flows around school closure. You'll see that learning deteriorates goals to additional housework and can work our data supports that this is true. But then you see disruptions in the school system, you see early pregnancy CC loss of household income and then goals being delegated with income earning tasks, but you see also increased risk of exposure to sexual exposure, exploitation and GBV. So goals are more at risk than boys with in the context of of covered 19 and the school closures. And if you go to the next slide. A post covered we believe that it will be very, very important to leverage some of these and pay attention specifically to to the vulnerabilities that calls face during closure. And I'll go on to the second last slide. And just so that I can finish up and we can have some time to discussions, if you can please move the slides. So here we have gender based violence, and we definitely found that most women and men in the areas where we did the survey for the gender based violence increased. We need continued advocacy work around GBV prevention and services. So safe places mechanisms and services for victims and survivors, all of those are important aspects that need attention. We need more data that's conducted with larger samples and that look at measuring the incidents of gender based violence. And of course there are all the issues around human rights training of the police and all the other agencies dealing with victims and survivors of rape and gender based violence. So I thank you and we will have a formal launch of a wider study than just the data that I've just shared with you on the regional impacts of COVID-19 on gender equality in eastern Southern Africa. I'm going to launch this during International Women's Day and I will share the program as well as the links with you in wider. So for those of you who are interested are welcome to attend the launch at that time. So now I'm handing over for a session that will be discussing the findings of the free papers that were presented. Thank you. Thank you very much. I'd like to take this time now to welcome you for your training. So what I can see we have some questions already from the participants directed to the presenters. So what I'll do is that I'll pick questions for each one of you. And then maybe we could have two questions. And then you have time to respond. And if time allows you could take another set of questions. So the first question goes to Jesse and the question is the participants. This is a no man. I would like to know how monetary policy can work coherently with the physical policy to mitigate the impacts of COVID-19 and how you you accounted this in your analysis. Right. Yeah. Thanks to Nomena for this question. I'm definitely not an expert on monetary policy on the monetary policy response to go with especially in developing countries. The simple answer I guess is that most likely in cases and in countries where there is a clear negative impact on aggregate demand and financial stability monetary policy should become a loser at least momentarily. You know lower interest rates may be more relaxed capital and liquidity requirements for banks for instance might be warranted. You know ideally these monetary policy changes would still be consistent with any mandate the country has of you know regarding price stability financial stability. Something that complicates this kind of basic prescription a little bit is that COVID has been lastly not just a demand shock but also a supply side shock which kind of makes monetary policy less effective in these cases. And also there's been a relatively quick recovery in many countries as we saw from the first paper in Ghana. But again you know the circumstances differ between countries a lot and I'm definitely not the most qualified person to comment on this. But that would be my thought in our analysis we basically you know tax benefit micro simulation models are about you know kind of static tax and benefit changes and also income changes within the country. So monetary policy wasn't something that we directly took into account. So yeah that's what I would say thanks. Thank you very much. The next question is a I hope that that answers the question to the participants that is no manner. The next question goes to Michael and Simone. Do the type of lockdown policy explain the changes in poverty and inequality across countries. And the question is the additional part of the question is that because some countries had minimum lockdown compared to others. So how would you account for that or how can you be able to explain whether the type of lockdown policy measures undertaken. Partly explains the changes in the poverty and inequality measures. Many, many, many things. In our paper, although we don't look at issues of poverty and what you call the inequality, but we can still speak to this to some extent, because what we actually find is that the areas where we have this partial lockdown, there was a significant negative effect on employment, just as Simone showed, and this and this was mainly on the informal self employed. Definitely, if you have such a shock that leads to such a labor market outcome, you know, then definitely it will have some effect on the livelihoods of the people as well. So if you take most of the countries, let's say Ghana, you have more than 80% of the people working in the informal sector. And these are the folks that we have, you know, kids, you know, looking at the, the, the, the work that we, you know, so, so yes, we don't work exactly on that, but one can look at it, looking at the very negative impact on the labor market outcomes. Thank you, Michael. And since you are still just on stage, could I also ask a question. I was wondering whether you took into consideration the, the gender differences in the, the effect of the lockdown measures on employment and earnings. I, I, I saw some gender variable in your regression but I'm not quite sure to what extent you took this into consideration. Yes, so Simone. Yeah, thanks Maureen for the question. So as we did check for differences by gender by adding an interaction with the variable capturing capturing the gender of the respondent and as I was showing we saw that especially like this more near to midterm, midterm effect, we had a significant difference between males and females so females were more likely to be still out of work, even like five months after the lockdown had been implemented, and the earnings were more negatively affected than those by men so we do look at this and do find stronger effects for females. Thank you. So back to you, I, I, I think the, the person who was asked this question about the, whether the impact of, or the type of lockdown measures explains changes in poverty and inequality. You to respond because I think your presentation touched more on the, on the changes in the poverty and inequality across countries, and we know that some countries like maybe Tanzania didn't at all have any lockdown measures, and how would that explain your inequality and poverty results. Yeah, that's actually a very good question. You know, our current approach, basically relied on world banks demand data and it doesn't this, this, this thing with between different explanations with behind the economic shock. People, person who answer, ask, ask this is definitely right in that, you know, lockdown measures do not necessarily explain the adverse economic effects, including, you know, poverty increases. There are other reasons. For instance, you know, voluntary reduction of travel and consumption within the country, maybe reduce tourism, lower demand for experts, exports in that country. So there are many, many possible reasons, but our current method doesn't kind of distinguish between these causes of the pandemic. So that's my simple answer to that. Thanks. I have some questions for Isabella. I mean, this is just like more of a comment. And I think it's, it's related to the same issue of whether the types of lockdown or maybe the measures that have been taken. For example, I noticed that the countries like Kenya and Ethiopia were more affected in your presentation in terms of the some of the outcomes. Because maybe these countries had most strict lockdown measures and how, whether this has taken into consideration some of the social protection measures that have been taken and whether this have been effective in elevating some of those gender differences. Yes, we actually definitely considered the lockdown measures in each country that were different. I think what all the countries had in common was that the survey data was collected when all international borders were open and movement restrictions internally were reduced. I think what was in place in some countries like South African Kenya was curfew was a curfew. Unfortunately, we like the first study that was presented we did not specifically collect data during different phases or compare areas with lockdown without lockdown. And so because the economic circumstances were different even prior to the study, we cannot attribute the differences that we see to specific types of lockdown measures because the starting points were different. But I do think that the country where the lockdowns were the most severe, the countries were actually Kenya and South Africa in the region. And economic activities were severely affected, especially in the first two to three months of lockdown. So I think some of these impacts that we see and differences between countries could, you know, inferentially be linked but it's not scientifically associated. Thank you very much. We still have more questions actually there are quite a number of them and just pick them selectively so please don't feel offended if I'm not able to pick on your question. I'll have a few minutes and I'll try to pick just at randomly. This is directed to Michael and Simone. On the analysis of the impact of COVID unemployment. Were you able to take into account the extent of all the linkage with the sectors for example to what extent like maybe some sectors were able to either have to act as fall back position in pushing the impact or whether there were variations across sectors like natural resources, agriculture, etc. Okay, so I mean, we have, we have enough data on the sectors to be able to this what do you call it aggregate, right, and then look at some of these things that that it was reason, but but I'll let Simone. And but we don't go to that. What do you call it a stand that that that's that was not the, the, the, what do you call it a focus of this study and so we don't do the this aggregation that he wants us but definitely we should be able to look at that even the data that we Yeah, maybe just thanks for the comment and I think it's a suggestion to explore. Just on a note like our sample focuses on workers in urban areas only because we look like those that were on the lockdown was a major urban centers. There are other like control districts where we did the survey all urban so we may not be able to speak to the sectors mentioned by Edward here which are mainly agricultural kind of so we have little information on the effect on agriculture. And this may be there's a comment here or a question that I think is directed because it applies to I think all of all of the presentations and the fact that this The results of the analysis is based on kind of rapid survey that has been undertaken just after the COVID is there a possibility of embarking on another rapid survey or longer period that perhaps capture the effects on a prolonged period, which would also cover the second wave of COVID-19 that we are witnessing. Yes, yes, yes, of course, but it takes money to do that so said if you would give us funding, we would hopefully So I mean what I want to say is that say that the, what do you call it the basis are there we've done a lot of work to build this, what do you call it the partnerships and to get some of these work going so so the base is there, we can always, you know, expand on that. That's what I want to say. So in my case, you in women, we spent about $500,000 on the study with partners like UNFPA etc contributed as well and fortunately we won't have in the next six months resources to repeat the study. We will do a global study on gender based violence in three countries in East Africa and free in West Africa, which will have a slight economic dimension because that's part of the demographic data collected. But yes, unfortunately, we can't on the scale that we've done the previous one repeated. And it may be also add that the one survey that we will be using in the future will be the World Bank's high frequency phone surveys on COVID-19 in, for instance, Uganda, Tanzania and also they have the same same types of surveys for other countries as well. And I think that survey will go on. You know, it's still going on in 2021 and what's going on at the end of last year so I think that's that's one survey that actually might give for many people many research or interesting information on the effects of COVID, including also the second wave. Thank you. So we it's already 1030 and I'm afraid I have to to end the session so that we can have adequate time for the next session. And I'd like to thank you all for wonderful presentation and also for taking time to be with us. I'd like to thank the participants that in this session. I would like to call the session to end the session and we have 15 minutes break before the next session on the new normal and the future development of Africa. Thank you all.