 elicitate individual preference of residents to our immigrants' attributes in the Dominican Republic. So what is the motivation of this? Well, in comparison to the South-North migration, South-South is less study, literature there is sparse. And we already heard about the importance of developing countries in this topic. Over 35% of the stock of immigrants are in developing countries and over the last decade we observed that immigration flows toward between these economies have increased at a higher pace than from poor to rich countries. So this trend is likely to increase in the near future and this plays a risk, a substantial risk for countries that have weaker institutional frameworks and so on. If you don't see the citations, it's because I didn't add citations. But along this process we observed that there is an increasing negative view toward immigrants, right? And what it seems to be a case is that these negative perceptions or opinions or views are as negative in developed as in developing countries. So this is kind of in contrast with what we already know about the development effects on immigration. We know that there are long-term net positive effects. So why such a view? Well, there are two working types of hypotheses. Now those that relate to economic factors and those with no economic factors. You will know these determinants. The point is that behind those factors there are characteristics of both the immigrant and the resident. For example, in the labor competition hypothesis, you know that for example, it might be the case that highly educated resident might be against highly educated competition entering the country. So these specific attributes of each one of the parties matter. So there are two broadly, there are two types of literature studying these hypotheses. One is for example, MAIDA, that was written before, study native attitudes toward immigration, not individual subjects. And they are based on fairly broad, fairly wide public opinion surveys that covers many countries including developing ones. There are another stream of literature that study attitudes toward specific immigrants. And they are based on more specialized surveys and typically use co-joined analysis or so. So I joined this last stream and the question I asked is which immigrant profile, which type of immigrant is support for admission into that country. And I think that my contributions might be that first, well, it's kind of the first. I am not aware of a previous study using this state preference approach to a developing country. And it matters because previous studies on this stream focus on developed countries. In our way, United States. And we don't know if those findings hold for our context, which are different, which are different in terms of not only institutions, coverage of social service, labor market competition, labor market composition, etc. So I suppose that an underlying question is there are studies, how do Americans weight the preference of foreigners. I will ask the same question, how do a resident native from a developing country weight the different characteristics of a foreigner. So more in a methodological, with a methodological flavor, there are some different with previous studies. In this study, I address a greater number of attributes of the immigrant. And further, I allow for heterogeneity in the preference toward those attributes. And I study which drivers, which socioeconomic characteristics of the native might influence those references. And in addition, I study two types of choice settings, one which is for choice in the other with an either option. So this is the methodology, well, the characteristics of the methodology I use to generate the data. But it's better with an example, sorry about that. This is an example of a card, a choice situation I use in the field to raise the data. Basically, choice experiments consist in presenting hypotheticals, choice situations to the respondent, in this case the resident. And they have to choose between those alternatives and say which profile they would prefer. So the first column are the column of attributes which I extract from the literature and from a focus group in the country. And the other columns are the candidate's profile. And all the columns or the profiles or the candidates at least have a different level in at least one attribute. The point here, I mean, the source of the identification here is that I randomize attributes and I randomize levels in order to identify causality in terms of what they prefer. So, well, this is a, of course, this is a discrete choice setting. And I model that using a typical, well, random utility model with some features, right? In this case, for example, the utility that a resident I obtain in a choice situation is from an immigrant J, depends on the characteristics of the immigrant X and the weight preference alpha. I allow the preference of each resident to be heterogeneous. It depends on, be only for that resident plus a typical noise. Further, I allow that weight preference to depend of W reflecting socioeconomic characteristics of the resident and beta reflect that association. Understand the decisions rules in assumptions. This derived in a mixed logic model, which I solved computationally. So, the data. The data is about 2,500 respondents in the main and the biggest city, seven cities in Dominican Republic. As I mentioned, I used two types of experiments. One fourth choice in which the respondents have to decide, have to choose which one of the candidates he would support for admission into the country. And another with an interoption in which the resident, the respondent was allowed to deny access to all of those candidates. And as you can see here, I randomize also these treatments, these types of choice experiments between the respondents. As you can see, both samples are well balanced. The only difference is for the number of admitted profiles into the country. Which in the case of fourth choice is around 33%, while is 25% in the one with interoptions. And that difference is significantly different. I think that's it all. The results. Sorry, a little messy. These are the main results. These are the alpha coefficients we saw before in the first column for each immigrant attribute. The main message here is kind of aligned with previous findings is that residents prefer almost as competing for a job. Younger people were educated with fluency in the local language, probably with some norm of cultural resemblance. For example, religion with high skill occupations. The second column, column PSD, test if those alphas are sufficiently heterogeneous among the population. For those that are specified as random parameters. And they mostly are. So preference are quite heterogeneous. The third column, beta, test for the case of education in this case if the preference of the respondent depends on their education level. And in most cases, they don't. I tried with different specifications, different characteristics of the native, etc. These are the results of a near-octane experiment. And basically the same story with some important difference. Some attributes become strongly significant. Some others become weakly. But the main point is that the feet of the model to the decisions that we observe in this experiment increase dramatically for 50% to 24%. Overall, the message is that in terms of the model is the feet of the model understanding preference of people. For understanding preference of people is important to account for their heterogeneity and for more realistic choice settings. So, okay, just for you to believe me that I didn't find anything from the respondent explaining what is their preferences. In the case of the coefficient for country of origin, the preference for country of origin, we observe that there is not a pattern. There is wide heterogeneity in the coefficients, but the distribution doesn't seem to change across income levels of the respondent. So, the other thing, yes, to see this better, I mean, the previous table was rather fast. Well, this is the marginal probability effects of each level of each attribute. Well, again, we observe the same as before. High speed professions as medical doctors, professors, and digits are very well-rewarded when taking the decision. The same as very correlated with higher levels of education. Then you observe, sorry, you don't answer. So, this one, okay, is the preference for country of origin, okay? They don't like Haitians, but they like people from developed countries as Italy, Spain, USA. The other countries are well around zero. And the other characteristics, oops, the other characteristics are mostly small in terms of the contribution. So, in contrast with previous studies, I was rather estranged about this result with regard to the country of origin, because previous ones show that this preference for some specific countries tend to disappear when you account for other characteristics of education, et cetera. So, let me jump that. So, I try to cluster the findings and try to do conditional probabilities of getting admitted to a country. And we observe here that, again, for every level of education, again, the preference for Haitians are well below that for immigrants or potential candidates from developed countries. So, this might be reflecting, of course, heterogeneity in quality of education. But in any case, it is the finding I have. So, again, the same information in another way. Who meets the cut? The same individual, which profile is preferred for admission? And to get admitted into this experiment, I have the probability of 40%. And the only ones who have a higher probability than that are immigrants for U.S. or developed countries, and regardless, other trade-offs in the other attributes. Education, skill, even language. So, again, just to show a kind of robust check, these are the plot, the kernel densities of the probability of admission for two types of choice experiments, for the two types I applied. Of course, the one with neither option is more concentrated to the left indicating lower probability of support. But also, the other lines are different specifications, regardless which random parameter I use, which determinant I use as explanatory variable. Different distribution assumptions, regardless of the noise, the error term, etc. And they are more or less tell the same story. So, to conclude, some results are aligned with previous literature. Education, occupation, country of origin, player role in admission, some others don't. For example, immigrant status is seen not to be a determinant in any case, but of course that is expected because this country has a weaker rule of law. Then, what is unexpected is that the premium or penalty for some countries persists even after different specifications or controlling for different aspects. Just to mention that these both previous findings imply a very kind of rough reality for future and current flows of immigrants. Because these characteristics doesn't match the typical immigrant. They're skilled, well-educated, and typically the domain of the local language doesn't meet the typical immigration flow that we observe. Well, the other conclusions are more in the methodological sense. Aetherogeneity matters, increased the fit of the model. Aetherogeneity, I found, doesn't seem to be explained by observable characteristics. It seems to be idiosyncratic. Finally, the choice settings also are important. I found that choice settings with needed options are more relevant. Increase the fit of the model probably is more a realistic setting and for sure is a more stringent choice situation. So that's it. Thank you.