 Hi, my name is Michael Damqua, I'm a research fellow here at UNU-Wide, and I would want to welcome all of you to this session. So yes, we have 45, what do you call them, minutes, and we have four presentations as well. What goes straight to the presentations. What we would do is that, yes, you would do all the presentations, and then after that we would take what you call the questions after that. But if you have any questions, please put them in the Q&A chat, and I would ask you to do that live, one is time for the Q&A. So please do that, keep your questions coming as the presentations go on. We would take the first presentation, and it's by, you know, Beatrice. This is from the Erasmus University, so Beatrice, the floor is yours. So welcome to this presentation on the effects of firm resilience and policy responses on employment in Central America during the COVID-19 crisis. So in a recent World Bank survey of the private sector across Central America, it was found that more than half of the survey firms had already reduced the number of employees. Supporting firms thus is essential to reduce layoffs, yet might be a hard task given the low level of public revenues in the Central American region, which is about 18% of GDP, which is still low even by Latin American standards. Among these firms in the survey, it was reported that about 11% have received some kind of government support. In this context then we explore the question, how firm level resilience capabilities interact with government support in the reduction of layoffs among formal firms in El Salvador, Guatemala, Honduras, and Nicaragua. Economic resilience is typically categorized as either static or dynamic. For our research, we defined static capabilities as a general category of resources and abilities a firm accumulated prior to the shock, in this case the pandemic onset in 2019, whereas dynamic capabilities refer to the specific responses after it. Our data is from the World Bank Enterprise Survey and our sample covers approximately 660 formal firms. Our method is as follows. First we estimate two Latin variables that capture static and dynamic resilience. For static resilience, we consider whether the firm introduced an innovation to the market three years prior to the baseline and whether the firm invested in R&D also at the time of the baseline and this can be seen in equation one. For dynamic resilience in equation two, we consider whether the firm introduced a new product or service, whether it invested in a digital solution and the share of workers that work from home. All these in response to the COVID-19 crisis. We then incorporate these two Latin variables into our main equation, which estimates the probability of a layoff at the firm level, which can be seen in equation three. For these, we include a number of standard controls like past sales, firm size, type of industry, type of ownership, etc., and also we include the weeks that the establishment had to close due to COVID-19. We employ a Bayesian framework in which the model estimations rely on the specification of a likelihood function, which in turn is solved by using the Markov chain Monte Carlo algorithm. We then perform a counterfactual analysis of four different scenarios to assess the fact of firm level resilience and government support, which we call the treatment. This can be seen in table one, and as we can see, the first group includes both forms of resilience and treatment, the second group both forms of resilience, but not treatment, and then the last two groups, either static or dynamic resilience plus treatment. We finally compared the empirical cumulative distribution function of these groups using the concept of first order stochastic dominance to assess in which group government support layoffs are less likely to occur, and these are our results. We find that the only group where we see a reduction in the probability of a predicted layoff is the group where there's only dynamic capabilities and treatment. To illustrate this small but noticeable reduction, the predicted probability of a layoff being equal to or less than 0.5 in this group I just mentioned is around 0.6, whereas in the other groups it stands at around 0.4. To visually assess this reduction, we can also look at the proportion of times in which one group exhibits a higher layoff probability than other counterfactual groups. In this graph it can be seen that this curve over here in the left is the comparison of the group where there's dynamic capabilities and the treatment versus the group that has both forms of resilience. We can also see other groups where the treatment was included and in all cases this other groups pretty much exhibit the same predicted probabilities of layoff. This is the only one where we see a difference. And to further our analysis, we also do a breakdown by sector and our results hold. In particular, we see that for the services sector, which can be seen on the right side of the slide, the fact is somehow more noticeable. To conclude, our results suggest that government support measures do play a role in reducing the probability of layoffs among firms with dynamic capabilities alone. As I already mentioned, this is proxied, for instance, by firms that arrange some kind of remote work. We also find that the effect of government support does not seem to be statistically different from the effect of a statistic resilience alone with respect to layoff probabilities. The above does not imply at all that COVID-19 support measures should be disregarded, but it really raises the question about the effective allocation of these resources. Is support going to more robust firms where it is likely to have an effect on employment or to more dynamic and yet less resource abundant firms where it is more likely to have an effect because it is more likely that people will be in the latter group? It also underlines the necessity of policies to enhance resilience more broadly and to go for more continuous government support to develop this type of solutions rather than than talk measures. And that would be all. Thank you very much. Thank you, you know, Beatrice, for the presentation. Many thanks. OK, so if if if anyone has any questions on this, please put them in the in the accurate chat and at the end of the presentations, we would ask you to do to do that live for us. So we would move to we will we would move to this second presentation and it's by Fiona and Fiona is from the University of Joe in Johannesburg. So Fiona, over to you. Thanks very much. I hope the slides are visible now. Yes, great. I'm going to turn with this presentation quickly in the in the six minutes. This is based on joint work with myself and common I do as it fits very well with Beatrice's presentation. And it's great to see this work and we should we should definitely talk more. So in our work here, we were analyzing the impact of the pandemic on enterprise performance in developing and emerging economies. We know that the effects of the pandemic have been very uneven across firms within sectors within countries, between countries. And here we we analyzing three different aspects of firm level outcomes and closures, employment and sales and how these have been affected by various sets of determinants. Firstly, looking at prior conditions at both the firm level in terms of firms, productive capabilities and other firm level characteristics and also prior country level characteristics, including GDP per capita, degree of industrialization and so on, which will come to further. Then the country level severity of the pandemic, the stringency of containment measures at the country level and the extent of economic support measures at the country level. And then the firm level, the extent to which firms have been able to to adapt their production or services in the face of the pandemic and where the firms have received assistance from government. So broadly, what we're interested in here is what is the relevance of these firm level and country level structural features as well as responses for the firm robustness and resilience in the pandemic. And here, although we we analyzing a range of factors, but we particularly interested in firm level productive capabilities, which I won't go into in detail here, and kind of sector and macro level characteristics at the country level in terms of degree of industrialization and industrial competitiveness. So our econometric approach is really translating what I've just been talking about into an equation. So I'll read it very briefly. Basically, modeling firm level outcomes, survival, employment growth and sales growth as a function of these various sets of determinants, country level variables, measuring the severity of the pandemic, stringency of containment measures and the extent of economic support and then firm level prior characteristics, firm level production response and prior country characteristics. So we first control for firm survival. We estimate a two step HECMAN and controlling for selection bias of firms that remain fully operational during the pandemic. And we use additional selection variables, such as the net profit and a dummy as to whether the exception is part of a larger firm. Let me just go on to the data. Like Beatrice, we also using the World Bank Enterprise Survey and we're combining this. I'm not going to go through it in detail, but we're combining it with other country level variables. I'm just going to skip this because of time and just go straight to our results. So I'm just going to show two sets of results here, although we have more in the paper. So firstly, in terms of the determinants of firm survival, we've split this same table into two slides just because it was a bit long to include on one slide. So what we've done, we have in terms of the productive capabilities in the results, which I'm showing here, we have calculated indices of technological capabilities and production capabilities. We also have separate regressions, which I won't show here because of time where we individually model the various components of both of these indices. So just to highlight a couple of features of interest here that for manufacturing firms, both our linear probability model and the probate, we find production capabilities being significant all the weekly and technological capabilities in the LPM, but not in the probate. And in the regressions where we look at the individual components of these indices, in particular significance of having international quality certification. Let me just move on. So this is the second part of the same table, which is determinants of firm survival. And here we're looking more at the country-level variables. The CRP score, which is the Industrial Competitiveness score from the UNIDO index. Interestingly, it's highly significant for manufacturing firms. So this at the country-level is a strong determinant of outcomes for manufacturing firms. We also see a negative effect of the stringency of containment measures and of the extent of the pandemic. In this set of regressions, we're not seeing any overall significance of the country-level economic response. And then I'll just show the employment growth regressions. So this is on the frames that remained open. So we can see some significance from both the technological capabilities index and the production capabilities index, as well as in this case from the firm-level production response. So this is the extent to which firms were able to adapt their production in response to the pandemic. And in the case of manufacturing firms, a positive impact from a firm-level receipt of government support. In terms of the country-level determinants of firms' employment outcomes, again, we see a very strong impact of the CRP score and again of the stringency index and the severity of the pandemic. Just very briefly, we also disaggregate firms by take intensity, so comparing low-tech and medium and high-tech firms. And that's just one result of significance here. For low-tech firms, it's the production capabilities index, which shows up as most significant. And for medium and high-tech firms, it's the technological capabilities index, which I think is intuitive. We're not going to show the sales results here because of time. I'm just going to my final slide. As you would have seen from the regressions, they were quite noisily estimated. So it seems that there's really a strongly idiosyncratic element to firm-level outcomes. And I'm not sure how to say that we tried a range of different specifications beyond what we've shown here. And we really found them consistently noisily estimated. So I think the pandemic just, it seems very difficult to kind of comprehensively model firm-level outcomes during the pandemic. But broadly, although the results are uneven, but we do find importance of the technological and production capabilities indices as well as of specific components within those index. And what's important to mention about that is that those kind of capabilities are not things which can be built up over that in the face of a pandemic. They are the outcome of years of investments with financial and otherwise in those capabilities, which proved to be important in a firm's resilience and robustness. And then at the country level, I think it's interesting to see that the degree of industrial competitiveness is a positive determinant of firm's employment growth. And for me, one of the interesting results from this is also for services firms. So even for services firms, the competitiveness of the manufacturing sector is an important determinant of firm's employment outcomes. Thank you. All right. Many thanks, Fiona, for the interesting presentation. So again, if you have any questions for Fiona, please put them in the Q&A chat. We would come back to that after the presentations. We would move on to the third presentation, and it's by Kenneth. Kenneth is from ETH Zurich. Kenneth, please go ahead with your presentation. Hello. My name is Kenneth and the paper I'm going to present is called Physical Distancing Becomes Impossible, a physical distancing index based on access to essential infrastructure, which is joint work with Isabel Günther, Johannes Seiler and Jürg Utziger. To give you a little bit of a background, we know that limited infrastructure remains a big challenge for many low and middle-income countries, particularly in Sub-Saharan Africa, both at the macro as well as at the micro level. And lack of such infrastructure can also affect the ability of countries to prevent outbreaks and contain the spread of infectious diseases, such as the ongoing COVID-19 pandemic. We know that most countries have responded to the crisis with a series of public health measures, including encouraging of physical distancing. And these measures, for example, school closings, work-from-home arrangements, even curfews and suspension of some public services. And we also know that the lockdowns of public life in Sub-Saharan African countries come with immense economic and social costs. For example, large shares of population are employed in the informal sector where people cannot go to work. It will result in an instant income loss for most of these people. And as a result, many African countries have quickly started to lift lockdown measures, again, despite rising daily COVID-19 cases. However, this was not the case for school closings. In this paper, we propose a very simple physical distancing index composed of five indicators, households with lack of private toilet facilities, lack of private drinking water source, lack of ICT infrastructure to be connected to the outside world, lack of private transportation means, and lack of space measured as people per room used for sleeping. To do so, we use data from the demographic and health survey database for the most reasonable survey year. In doing so, we have a sample of 34 countries between 2005 and 2018 with more than 700,000 households. What we do is we calculate the physical distancing index based on a principal component analysis and then the PDI is calculated for each household. One means lowest access to private infrastructure, PDI of zero indicates no lack of essential infrastructure. We also take into account that the lack of private infrastructure leads to more social interaction in highly or more densely populated areas. So we adjust the index by population density and in addition, use a geostatistical approach based on a Bayesian regression to provide high resolution maps in particular 10 by 10 kilometers. To show you some results, the graph shows the geospatial estimates of the PDI at the country level, regional level and pixel level. We see that high risk areas of disease transmission are particularly concentrated in the western part of Africa, such as Ghana, Togo, Iberia and Senegal, countries with lower population densities and relatively better infrastructure, such as Namibia, Gabon, Mozambique and South Africa show a lower PDI. However, the interpretation of the estimates need to be made in relation to other southern countries. For example, although South Africa shows a much brighter color, this does not mean that the country has all the infrastructure in place for people to keep distance. It shows that compared to, for example, Burkina Faso, South Africa has on average better domestic infrastructure, but still might play key elements of domestic infrastructure. Moreover, we also know nothing about the actual behavior of people, whether they actually follow a physical distancing regulation even if this is in principle possible. And the right panel of the figures will reveal some considerable spatial heterogeneity of high risk areas within countries and to zoom in a little bit, here the results are shown for Ghana and South Africa and for both countries, sub-national heterogeneity is high and high risk areas exist in both countries, particularly in highly dense urban areas. Okay, to summarize, we calculate a physical distancing index based on a multi-dimensional index approach that indicates the potential effectiveness of physical distancing regulation at the national and sub-national level. The spatial analysis shows that many households in Southern Africa lack essential private infrastructure that undermine governmental regulation to foster physical distancing. We also find large within-country heterogeneity and our results highlight the fact that different countries face different infrastructure challenges. So if not addressed in the long-term COVID-19 or other cases, we continue to spread despite drastic and costly national measures such as closure of schools and businesses and the limitations of our study include its descriptive nature lack of information beyond the indicators we are using and the limited general ability beyond the countries analyzed. Thank you. Many thanks Kenneth for the interesting presentation and again if you have any questions, please keep it coming. We would look at all of them after the presentations. So we would go to the last presenter and this Carl right? Carl is from what do you call it the IFRI and the UNU Do you have a recorded EPT or you would like We have a video. Okay, good. Hello everyone I'm presenting some work on food prices marketing margins and shocks in the context of vegetables and the COVID-19 pandemic in Ethiopia. This is joint work with IFRI colleagues from International Center in Ethiopia. So the background is basically widespread concern at the beginning of this pandemic how food value chains are coping with this pandemic especially in low and middle income countries. In Ethiopia the relevant information here is the plan borders were closed with implications to cross border trade in the first week there was a lot of confusion and disruption to domestic trade as well and transportation but importantly the country never went into a full lockdown that severely restricted movements like we've seen in some other neighboring countries such as Kenya. At the policy level there was always this emphasis on protecting food security. So we've been doing stacked value change surveys in Ethiopia to try to understand how agriculture value change function. In February 2020 we conducted one in-person survey focusing on the vegetable value chain the most important vegetable value chain that connects farmers in the central risk valley to consumers in Addis Ababa. This is a quite important value chain for the country because it supplies approximately 200 million USD worth vegetables to Addis Ababa every year. We did quantitative surveys at different levels of the value chain going from rural producers to urban retailers and then when the pandemic declared we saw the opportunity to answer and kind of look into this question of how the pandemic is shaping these value chains we randomly selected half of the original respondents that is more than 400 farmers 30 whole sale outlets and more than 200 retail outlets for follow up surveys that we conducted in May 2020 and March 2021. We also did one survey just recently last month but we don't have the dates yet to analyze here. So in this this presentation I'll be focusing the price date that we have from each round more than 10,000 price observations for most important vegetables consumed in Addis tomatoes, onions, green pepper and cabbage. So in each round we recorded price per unit, quality attributes and origin particularly at the wholesale and retail levels. So we collected this price data at different levels of the value chain. So we can use this data to analyze farm gate prices and cross margins at the wholesale and retail levels. So just diving into the results so here we see the retail prices in beer per kilogram for different vegetables and for different survey rounds so I'm going to highlight just the two cases here the case of onion and green pepper so for onion we see that the prices increased from at the beginning of the pandemic between February and May and the reason for this was that imports from Sudan that play a important role in stabilizing supply particularly during off seasons came to halt and suddenly there was an under supply of onions coming into Addis and this resulted a price increase but then in March 2021 we see that the prices kind of collapsed and the reason for this is that encouraged by this price increases farmers increasingly started cultivating onion in the previous season and suddenly there was an over supply of onions to Addis the green pepper story is basically the opposite the neighboring region Amhara forms an important market for the green pepper produced in these zones in Oramiya and in May that market came to halt partly there was less demand for onions for green pepper from Amhara there were also some trade disruptions within the country but then in March we see that the prices increase in March 2021 the prices increase quite rapidly and again this is the kind of same bounce back story in reverse many farmers switched away from green pepper and suddenly there was not sufficient supply of green pepper we can look at this more carefully by looking at the composition of these prices looking at the farm gate prices the amount that goes to the wholesalers and the amount that goes to the retailers so couple of observations here that the farm gate prices seem to take explain large portion of the final retail prices for almost all vegetable except for green pepper the other thing is that it seems that a lot of the variation we see in the final prices originates at the farm gate as well so in way of conclusion changes in consumer prices are often claimed to be linked to predatory behavior among traders motivating governments intervention to curb trading activity our findings here are kind of against this narrative they indicate that the price changes during the pandemic have not been driven by large increases in marketing margins for most part it seems that especially farmers are exposed to large price volatility because of domestic and international trade disruptions so if anything we should try to find solutions to support farmers and solve the inefficiencies at the farm level the obvious limitation here is that these are cross margins and we've seen in our data that the input prices particularly fertilizers are rising extremely rapidly since the beginning of the pandemic and they suggest that the cost of the farm level are also under rise not just the prices thank you very much many thanks Kao for the presentation let me check the chat the Q&A if I have any questions out there no I don't have any but let me let me ask this Fiona right I was able to follow your presentation I mean can you tell us more about how you were able to put together the data set I see you've got the the World Bank Enterprise Survey which is at the farm level right and then you have some then you have the what do you call the the constituency what do you call the index as well which is at the you know country level I am thinking that within a country the level of constituency it's a bit different so if you have farms which are what do you call it across a country you know how do you rate consul using one index which is at the national level for all these farms which are located across the country you know how to face the same constituency should I go ahead with that now sure yeah okay great thanks for the question so briefly in terms of how we compile the data set it's a multi level data set so the farm level data as we mentioned comes from the World Bank Enterprise Survey for countries with matched data sets over the period from 2016 and then those ones which had the COVID period pandemic COVID period surveys and then we combine that for the country level prior conditions with data from UNIDO the World Bank World Development indicators and so on and then the country level pandemic data and the constituency data we got from different sources but University one the Johns Hopkins and so on those variables are national they measured nationally in terms of the both the constituency and the effects of the pandemic so really they measure the national level responses and even the economic support measures that they measure national level responses so as far as I know there's no data available for countries internationally which will give you subnational constituency levels so where for example it's different from one state or province to another so while the constituency will vary as well as the severity of the pandemic varies those are country level variables and I guess perhaps be part of the noise in the estimation where there are strong differences we did try and include a covariate of whether the firm is located in the capital city to try and pick up some of that but we really just didn't find it significant in any of the specifications all right many things we have one question from Tic Tac to Ka Lina can you put her on the stage yes Tic, can you click the audio and video button at the top third Tic, please go go I thank you I have questions to Kale and Fiona sorry Beatrice I didn't hear your presentation earlier so I might miss your presentation the question is that for Kale for those Facebook are they produced mainly domestically and what caused the increase in input prices I think you mentioned the increase in prices due to the increase in input prices for Fiona I like to know what is the explanation for the negative sign for your export intensity in your services equation I think for the services sector you get negative sign for export intensity if I'm not mistaken thank you all right many things yes Kale all right thanks good question so we did this surveys in a particular zone has access to irrigated agriculture and this zone produces a large portion of the vegetables consumed in Addis so these are indeed domestically produced vegetables that we are focusing on here and the second question was about the input prices so this is a very dramatic story that these input prices are almost like doubling in a space of one year we are still trying to investigate the exact reasons for it but we see also that internationally fertiliser prices are going up quite rapidly as documented by the World Bank for example recently so that's one aspect to it then there is the exchange rates at the moment in Ethiopia with respect to the USD it's kind of plummeting and that's one part of the story driving prices of basically imported goods up so these are the kind of lines of research that we will be looking at now and we have new data that we've collected and we can look into this a little bit more diving into the story thank you all right Fiona you can thank you thanks for the question so the export intensity variable we included as a component of the production capabilities index but then as well in separate regressions which I didn't show here because of time we also, as I said, modeled those various components separately so we found really weak results on export intensity when we modeled it separately usually close to zero and usually insignificant and in my mind I think it's because of kind of countervailing effects of export intensity under normal circumstances we would expect this to be a strong or both an outcome of film productive capabilities can be thought about like that but as well a strong determinant of films robustness and resilience and performance and so on I think during the unique circumstances of the pandemic films which were more export intensive were also more vulnerable and more exposed to the downturn in international demand the pandemic and particularly in the early stages of the pandemic I think external demand was perhaps more strongly hit than domestic demand also due to import restrictions in certain countries disruption of value chains and so on so I think under these particular circumstances we found that some films which relied more on the domestic market were able to manage better I think it's probably these countervailing effects which ended up leading to the verbal showing up sometimes positive sometimes slightly negative but mostly insignificant all right many things let me ask Kenneth I cannot like the idea I mean the things you have done and the loads of work that has gone into but I was just wondering I mean how would this help a policy maker and say Ghana or South Africa what would be the main input as to what you call it and then another one have you said this what any of the of the policy makers in these countries thanks Michael for your question so to simply respond to the letter one no we did not yet and well we argue that since many countries tried to impose measures of social distancing they realize very quickly that this did not work especially with people a lot of people work in the informal sector it's simply not going to work but they cannot go to work so there's actually then there's not much left in order to reduce the spread of a contagious disease and one thing is physical distancing and the other one was actually school closings which obviously will have some severe long term negative effects and school closings are much longer in many African countries and for example in European countries or in the US so that's why we think it's more important for the next pandemic to really concentrate on these high risk areas because when we compare our index with actual COVID cases then this physical distancing index is actually quite a good predictor or vice versa there's a high correlation between the actual cases and our index so that's why we think that this type of research is pretty important all right I don't have any more questions over here and I think we are right on time actually it's 10.45 so if there are no questions I would want to close this you know they didn't have any feedback what do you call them I would read your paper and then send you some comments I know it was too early we were still waking up excellent I think it was a bit early and then also one thing I couldn't see the full screen later that I saw that I haven't but I would read it up and then I would drop you some comments thank you so much so many thanks Beatrice, Fiona Kenneth and to Kyle for the excellent presentations and many thanks to all of you I can see 18 people who are what you call them sticking with us as well so many thanks