 Hi everyone, thank you very much for the introduction and for being here as it has been said this paper was a collaborative work between me Bruno Martorano who was initially supposed to present and could not attend unfortunately today and Patricia Giustino from UNU wider. So do pandemics lead to rebellion the title we choose it because it sounds cooler I would say but we actually focus a lot more on the role of inequality itself and the way inequality and pandemics interact between themselves. So the first point is actually the literature trying to understand that the causes of protests and organized collective dissent is very broad and one of the prominent theories started by Gore in 1970 with seminal paper is actually that civil unrest is motivated by inequality. However this theory is not always supported by evidence there is more recent evidence for example by Salt that says that inequality is only determining protests among top quintile earners. So there is some room here for improving the evidence and what we thought was that COVID-19 could offer the pandemic and of course the related restrictions could offer a good opportunity to study the relationship between inequality and protests. So as the virus spread as we all know in the past two years social disconsent increased. Protest increased by almost 30% between January 2020 and January 2021 and in the USA in particular with our case study angry protesters have taken to the streets since mid-April 2020 so basically the start of the introduction of the restrictions to voice their anger. Now what I why and how have these protests come about? Initially the reason behind the protest seemed to be more of political nature but later on we realized that was actually due to well at least there is a total evidence that it was due to also economic problems generated by the restriction themselves. Think about lockdowns for example. So we thought that maybe the highest risk areas which for us are represented by those with higher inequality, the highest inequality in the country have faced a perfect storm and this perfect storm has magnified long-standing social and economic grievances. Now we want to study with this paper if pre-existing as I said in the beginning pre-existing inequality is explaining the relationship between policy restrictions and the incidence of protests co-related protests in the US. Brief introduction on the data. We study a total of over 3,140 US counties in 50 states and the District of Columbia and for each county we compiled time variant information on weekly aggregated figures on COVID related protest events, weekly changes in COVID related policies and a measure of income inequality from 2019. So again this is pre-existing inequality. The period of analysis is January 2020 to December 2021. We have a working paper in which we analyze only the first year, only January 2020 to December 2021 and here we present a let's say a lengthened analysis. Our initial strategy, it's very simple actually. It's simply analyzing the role of stringency. So this is a measure of how strict the restrictions are and an interaction between stringency and inequality to give the time variance to inequality and a series of controls which I'll speak about in a minute. However, in this equation we imply that there is a linear correlation that the effect of level of policy stringency on protest incidents varies linearly with the level of inequality. So what we thought was to also analyze non-linear relationship between these elements and to do this we divide the stringency inequality into quintiles and we select the different counties based on which quintile of inequality they are. As I said here just a brief mention of the controls that we use. We use log total number of COVID-19 cases in previous week. This is to account for the fact that the level in COVID cases might influence the stringency of the measures the week after. Log average rainfall and temperature for each county. These are also factors that have been shown to be associated with protests and then a series of social and economic controls for each of the county. These are all pre-pandemic controls. These are the last ones that I mentioned. I'm not going to go through the entire table. Let me just highlight a few relevant facts. So this is our initial strategy from equation one. As you can see stringency per se has a negative effect on protest. So the higher the level of restrictions, the lower the number of protests. However, the interaction between stringency and inequality actually is positive and it's more than let's say it's higher than so that the overall effect is actually a positive effect of inequality and stringency together. Showing that in counties with higher inequality and with higher stringency there is a higher participation in protests or a higher probability of having protests in our case. Here we show the analysis by quintiles and in fact we identify non-linear effects so that only the most unequal counties, those in Q4 and Q5, the fourth and the fifth quintile of inequality are actually the ones with more positive effects on the probability of protests. We then interrogated ourselves on which are the mechanisms that can cause these effects. And as I said before, anecdotal evidence points towards economic damage caused by COVID-19 and by the restrictions. So what we thought was that we could first investigate these, unemployment for example, and then also focus on what political science scholars have highlighted as a series of characteristics that influence protest participation, which are political preferences and group identity. I show you here first the impact of our economic mechanisms. We focus on two measures, consumer expenditure and unemployment level. And we show that indeed the level, the stringency interacted with the quintile of inequality is reducing consumer expenditure only in the two highest quintiles, among the two highest quintiles, although this is only significant as you can see for the last quintile. And the same is for unemployment level. So higher unemployment levels for the fourth and the fifth quintile. So our findings here are suggestive of the fact that indeed economic downturn might affect participation in protests. Additionally here I'm just presenting briefly the social and political context here. We analyze it in a different way. We don't see this as an outcome, but we simply rerun our initial models from a question to one to three about dividing our sample between areas with above median and below median trust in the president, values for trust in president, satisfaction with democracy, and the civic engagement. And then we divide our sample between counties that have voted Republican in the 2016 election and those who have voted Democratic candidate in 2016 election. And we find that actually the results of the protests are more likely to take place in most unical counties, so these results are again significant for the most unical counties, for counties in which Republican won in 2016, where trusted in president is below the median, where citizens are less satisfied with democracy, but also where there are higher levels of social trust and civic engagement. And we proxy civic engagement by number of religion organization, civic organization, political and labor organizations. Now you might wonder, one question that was we also asked ourselves is, but what do we, what we can do is actually only identify protests if they are related to COVID or not, but protests are not necessarily a bad outcome in one might wonder why is it bad that most unical counties protest more? So also what we wanted to investigate is if this relationship between the stringency of restrictions, anti-COVID restrictions and inequality was also affecting other outcomes. And in this case we focus for example in the vaccination rate, and we actually find, as you can see here, that most unical counties, in this case 3rd, 4th and even more for the 5th quintile, have a higher, well actually a lower vaccination rate than the other countries. So the impact of stringency with inequality causes a lower vaccination rate or is related to a lower vaccination rate in these counties, the most unical ones. There are, we also tested a series of other outcomes that we can talk more if there is time later on. And we, together with our initial analysis, we also run an IV strategy, an IV strategy for which I have extra slides in case we have time, because one might raise issues of endogeneity of the results. But let's for now focus on the main conclusions. So main results shows that the implementation of policy restrictions to contain the virus led to increases in intents of protests, COVID related protests, only in counties with highest level of inequalities. These results validates the theory that civil unrest might be motivated by inequality or is somehow connected to the levels of inequality. And the mechanism that we identify is that of the drastically adverse economic effects caused by the policies. Again, the example of Dock Town I guess is what comes to mind to all of us. However, the political and social consequences of such severe economic shock are yet to be completely understood and may take decades, of course, to grasp this point. Thank you very much.