 Thanks for coming to the session. I'm very happy to present the paper here at this conference. So I'm going to present you this paper entitled The Geography of NGO Activism Towards Multinational Cooperation. And I will start by showing you these two pictures. I suppose you're very familiar with this picture over here you've heard about them. These are both campaigns by NGOs that focus on international production and or sales. So the one on the left is a campaign by Greenpeace focusing on the production of the input used to produce the chocolate bar. And the problem here is the way that this input is produced and the fact that it is contained in the final product. And the one on the right is also a campaign by Greenpeace that we can find, of course, similar ones by other NGOs. And this one is concerned with the sales of two firms, the Mamut and the North Face, that are problematic because they contain different elements that are dangerous for the environment. So why am I showing you these pictures? I'm showing you these pictures for this reason because academics have been recently and especially since the beginning of the year 2000, very much interested in getting to know the microeconomics of globalization. And in terms of what relates to firms, we've made a lot of progress. Now models are very well published and are able to describe with very much precision what firms do, what they produce, how they behave. And these models have been complemented with data, very disaggregated data, especially customs data, which a researcher can use. So there's another agent that is very much involved in globalization on which this availability of the data and of evidence is not at all at the same stage and this is NGOs. So NGOs are another agent that is involved in international trade just because they target firms. At some point, they claim to influence their practices. So this is what they say and we know a little bit but we don't know much about it. And the other thing that they do, besides trying to influence the practices of firms, is trying to influence the regulation that countries put into place. So why do they do this? They do this because NGOs are involved in protecting and preserving resources, mainly environmental resources or human resources that are used in producing these goods. So an example of the potential, importance of these agents in shaping trade and sales are the recent law. For example, in France here, which is called a vigilance law, which addresses to firms with more than 5,000 employees, what do we know about these advocacy NGOs who act in international trade? Not very much, especially because we don't have a lot of data. We don't have a lot of data because there's no mandatory collection of data on these NGOs, on these associations. And because also it is complicated to collect data on activities that are very different. So you would have to collect data on activities of, for example, service provision by some NGOs who provide health services, who provide education services, or you would have to collect data on the campaigns that the NGOs publish and on the impact of these campaigns. So right now what we have is not exhaustive, of course. But what we have is, for example, this paper by Kutanyi and Ad. So there's this early database on NGO campaigns, on some NGO campaigns that allows to study the communication of activists. There are also case studies, so a lot of case studies. And this Harrison and Scorsepapers at 2010 in the AR is not case studies, but it is the only large-scale study that aims at studying the impact of activism on the behavior of firms. And they do it, and they study it on the wage that firms that produce textiles do in Indonesia. So what I'm going to present you today, and what I'm presenting you today is work on, so it's a whole work program that we have on new data, new data on these activists that campaign against multinational firms on production, on production abroad, but also on sales. So I'm presenting you part of this work program. I'm presenting you today a sort of panorama of activism towards firms. And another paper that I am working on is on this second bullet point. Studies the impact of campaigning on the sourcing decision of firms between different origin countries. So I'm going to show you, not show you, but just explain to you a little bit what we're working on, what we have, because this is really new. You have almost exhaustive data on what NGOs publish about firms. So we know for the period 2010, 2015, who says what about which firms, basically, if I can summarize it this way. So we know which NGO has campaigns about which firm. We know the date. And we know the reason. The cause of the campaign is summarized by a keyword. And also about the country, which we will call here in the rest of the presentation the action country. It's a country where the damage has taken place. So in all, we have more than 3,000 NGOs. These are originated in more than 100 countries, more than 7,000 firms are targeted at least once during that period in 139 countries. So to summarize, we have three different locations that we will use specifically in this work. To summarize the campaign, we have the country where the NGO issues the reports the news. We have the country of the firm, so where the firm is headquartered. And we have the country where the action, the damage has taken place. So this is what we do in these first papers. The first part of the paper is devoted to showing regularity, stylized facts, important stylized facts, because these do not exist up now. So for example, on where are the activists? How many such activists do we have per country? How are they distributed? Do we have more uniform distribution? Or do we have something that pretty much resembles the export panorama of firms? We're a very minority of firms export. And then we have a description. I'm going to show you a description about where NGOs target. And this question of where NGO targets, of course, will lead us to studying the determinants of the campaigns. The first category is a very straightforward fact, actually. These are two facts that relate to the distribution of campaign. So the first one shows us that the distribution of NGO campaigns is actually very skewed. So this is shown, for example, in these figures. So these figures are very straightforward. They just average the number of campaigns per year, per NGO, in each country. So we see that there's a minority of NGOs who campaigns a lot, and the rest campaigns very few. And another way to show you the same stylized fact, sorry to go that quickly on these slides, is to express the fact that, in fact, a majority of the campaigns are done by a very reduced number of NGOs. So there is not a uniform distribution where each NGO would publish the same number of records. On the contrary, it is very much skewed. We now turn to the second set of stylized facts. And these stylized facts relate to the bilateral connection between the NGO and the firm. So why does an NGO target a specific firm? And here we investigate in particular the fact that an NGO would target either a home firm or a firm that is headquartered in a foreign country. And I'll show you this table. So this table summarizes this internationalization of NGOs, specifically in the third column, where I compute the percent of activists in each country that have at least one campaign that includes a foreign firm. So this number seems pretty high. So we see that there are a lot of NGOs that are, if we can say it like this, internationalized in the sense that they don't target only domestic firms. Another way of investigating this internationalization is in column four, so the percent of foreign campaigns, among the NGOs that target abroad. What is the share of campaigns that are directed at foreign firms? And again here, this is a very high number. So now we turn to the second stylized fact that is related to internationalization. And here we see that the picture is actually a little bit nuanced with respect to what I just said, because in fact we realize that for more than three quarters of the campaigns that target foreign firms, well, the action is indeed related to something that has been done at home. So I'll show you this number. This is the very last column, so the percent of foreign campaigns with home as the action country. So this table and these both of these stylized facts are the basis of the question that we will address in the second part of the paper. And indeed, we're questioning the reason why a specific NGO targets a specific firm and we're asking whether an NGO is interested in targeting a firm that the audience relates. The odd possibility would be to think that the NGO, in fact, targets the firm that is related to the largest damage, for example, completely independently of the nationality and of the firm that is targeted. So how do we do this? We estimate a gravity equation for campaigns. So gravity equations have become the workhorse in international trade to study the determinants of trade flows. And there are now backed up by theory model that describe the determinants in three different terms. If I can summarize it very briefly, there are the characteristics of the source of the flow, the characteristics of the recipient and then the characteristics that are bilateral. So here we're gonna be interested in the bilateral connections, whether the bilateral connections do count in explaining the number of campaigns between two locations. To do this, we assume, so I summarized it in just two sentences here, but we assume that the NGO faces a discrete choice between different targets and her objective, the objective of the NGO is to maximize her pay of the campaign. And we assume that this payoff is maximized whenever the audience donates and the audience donates whenever she is familiar with what is contained in the campaign. Okay, so this is what, this is the micro funded model that is behind this estimated equation that we have. So this estimated equation is aggregated at the country level right now. And so yes, I have to clarify the indices. I is the origin country of the campaign, J is the country of the firm and K is the country where the damage is taken. Okay, so with this full triadic equation, we have these three countries that are involved and we are able to estimate the importance of bilateral determinants, I, J, okay, and I, K in shaping the total, the final number of campaigns. So what the determinants were interested in this equation are the X, I, J, just behind the beta two here. We will control, so doing this, we will control for the unilateral factors. So for the elements that are specific to the country of the NGO, okay, the size of the country and all the other determinants that might explain that a country publishes a lot of campaigns from its NGO and we also on the other side control for all the characteristics of the recipient countries, okay, of the target countries, explaining that the country receives a lot of campaigns or is the object of a lot of campaigns. Just before showing you the table, the regression table, I'll show you just these two table, these two figures that can illustrate a little what we mean by connections or bilateral, not both of them are bilateral, but the one on the left is, the one on the left illustrates the distance between the country of NGO and the country of the firm. So of course it's not an estimation, so we don't control for anything else. We illustrate, we graph the number of campaign received divided by the share of, by the population, by the share of the population, and we graph this on distance for the three different countries, France, the UK and Germany. And so a sort of pattern emerges which we will find confirmed with the regression, which is that bilateral connection here showed, illustrated by distance, bilateral connections do seem to impact the flow, if we can say it like this, with a similarity with international trade flows, the flow of campaigns here, of news flowing between a country and another one. The second figure is not bilateral, but we aim at looking for data that would make it bilateral. Right now it's just the number of brands that is, that pertains to each country. So we can compute the number of evaluated brands that each country have all sectors together, okay? And we just very simply right now just to illustrate our issue graph whether the number of campaigns that is received by a country increases, and it's the case with the number of brands illustrating a sort of pattern of visibility, okay? And so this is what we will investigate in the estimation, so whether the target needs to be visible to the audience. And we will investigate different sort of measures of this visibility. The first one will be really the familiarity that the audience has with the target, okay? Okay, so this is the first regression table. We will, I just highlighted two numbers that I will comment in this reduced amount of time. So just to explain the table, the above variables are the unilateral variable, okay? You see the size variable, population GDP, population and GDP for J, and we're mostly interested in, as I said, bilateral variables, okay? Why? Just because in the last three columns we're controlling for the unilateral variables by fixed effects, which is also conformed to what is done in trade, gravity equation. So I will focus directly on this number of, that is negative, that is the coefficient on distance, on the distance i, j, and this number shows us that controlling for the unilateral determinants of campaigns, well, distance does decrease what the number of news that NGO that activists publish towards another, in other words, firms of another country, okay? So the elasticity is approximately, well, it's of 3% for a 10% decrease in distance. The second fact, the second finding that can illustrate approximately the same behavior is the coefficient on the dummy for the home campaigns. So this means, this positive number and significant number means that in priority, NGOs will target firms from their own country, okay? From the home country, this is computed, computing having an internal distance, okay? So which means that for a given distance, for a given language, for given the rest of the variables, well, NGOs will prefer to target home firms and whenever the firm is not home, well, proximate firms or firms in proximate countries will be favored when choosing the target. So there remains to me just to tell you that these estimations are not fully convincing if I don't explain to you how I control for the third country that we have in the database, okay? So as I told you at the beginning, I have the possibility to identify the three countries, the countries where the NGO is located, but also where the action has taken place, okay? So here you might tell me, yes, but what if the distance between the NGO country and the action country is very much correlated with the distance to the firm country? Of course, this might bias our results, so what we do is estimate the same gravity equation for campaigns on the triadic campaigns, okay? The same table, but now I have not only the similarity with the usual gravity equation involving two countries, but here an estimation on a set of campaigns involving the whole three countries, okay? Which means a very interesting fact, it means that looking, for example, at this coefficient above here, which is negative of 0.205, well, I'm controlling for the action country. So basically I'm asking the question of whenever two NGOs from different countries, okay? Let's say France and Germany target two firms for an action in a given country, let's say Indonesia, okay? Will they tend to target firms from different countries, okay? And the answer is yes, but we know it's not due to the distance to the action country. It is really the effect of the nationality of the firm and this effect is still significant and still negative, okay? The same for the home campaigns. So I will summarize and I will get to the end of the presentation. These are the results. So NGOs do select in priority, in priority, sorry, home firms and this result holds when controlling for the distance to the action to, and this is valid for a given distance, a given size and given income of the target country and the takeaway so is that the number of campaigns towards firms in the 10% more removed country decreases by 2% and NGOs report 30% more on firms from countries sharing their home language on the domestic firms, on the home firms, I already said it and something I didn't point to which is very interesting is that the two coefficients are negative which means that in fact the proximity to the action or to the firm country are substituted, okay? So in order to have a successful campaign so it would mean that I can either have, I need one element of proximity but I can either have it from the firm country or from the information of the place where the action has taken place. The availability of data on NGO was very important for this project and it's still important and we are very much working on it. It allows to quantify and it allows to shed light on the phenomenon of private regulation which is potentially important in international trade. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you.