 So, this paper, Food and Security Gender and International Migration, is with my dissertation advisor, Maria Floreau, at American University, and for the sake of bureaucracy, I have to say that the views expressed by me today do not reflect the economic research service of the USDA and the United States government at all. Now that we're past that, okay, so the takeaway message of this paper is that I find that food and security is an important determinant of migration behavior for both migration intentions and migration decisions. I find that migration intentions increases monotonically with the severity of food and security. So the more food and security you are, the more likely you are to intend to migrate. I find that the association between food and security and migration intentions is larger than for any other covariate using the model, and this includes household income, social networks, or education, which are typically the biggest influences on migration intentions. And I find that migration decisions also decreases with the severity of food and security. Last, I find that these relationships vary significantly with gender as well, and I attribute these to socially inscribed gender norms and differential access to resources and employment opportunities. And I conclude that not accounting for these relationships may yield biased findings and may yield unreliable policy prescriptions. So to tell you why these results are important, migration is obviously an increasing important strategy in the developing world, but there's a notable lack of research on the relationship between food and security and migration. Further, there seems to be an institutional disconnect between migration and food security development agendas, where each has their own international agency, each has their own international gathering, and their own body of scholarly research, but often disregard each other completely or treat each other tangentially. One of these reasons for this gap in the literature was there was previously lacking a common measure of individual level food and security that could be applied around the world. To address this, FAO instituted the Voices of the Hungry Project and developed the FI scale, the Food and Security Experience Scale, and in 2014 implemented into the Gallup World Poll over 150 countries. Further, while some research has examined the overall trends of migration by gender, a few have modeled explicitly the differences in international determinants or migration determinants between men and women, despite the fact that the global share of international migrants are now increasingly women, where the common assumption in the economics literature is usually that the costs and potential returns to international migration are gender neutral, so the same between men and women, and more and more we're learning this is not true. So the objective of this paper was to examine food and security as a determinant of migration behavior, both migration intentions and migration decisions from potential movements in developing countries. We first developed a theoretical framework to demonstrate this relationship from a gendered perspective, and then we empirically examined the set of hypotheses derived from the theoretical framework. So hypothesis one is that we expected to find food and security to have a positive relationship with migration intentions, but we expected ambiguous relationship between food and security and migration decisions. For example, being food and secure may lead you to want to smooth your consumption by using migration, but you might not be able to afford it since you can't afford food. Hypothesis two is that we expected the magnitude of these relationships to increase with the severity of food and security, and given socially described gender norms and unequal gender relations in developing countries that we expected these relationships to differ by gender. Okay, so the contribution of the paper, we first provide a better understanding of the gendered relationship between international migration and food and security through our theoretical framework. We provide empirical evidence for the first time on this relationship for individuals among 90 developing countries, and we empirically demonstrate that this relationship is gender-specific. So in the paper, I drive each one of these terms, but for time we'll just jump right in. So we model differently the individual's migration intention versus the individual's migration decision. So starting here, the probability of a migration intention from country K to country J is just the probability that their expected utility in their destination is greater than the utility of staying where they are, given that they have enough resources, A, to meet the migration cost threshold to gamma C. And therefore the probability of not migrating of staying is the probability that they cannot afford to migrate, or if they can, that their expected utility is in the destination country is lower than their utility of staying where they are, Sylvester. The gender effect on the migration intentions is ambiguous. So on the one hand, higher psychic costs and fewer resources can constrain women from conserving migration, but on the other hand, higher probability of employment in certain female labor-intensive sectors in destination countries can increase their intentions to migrate. For example, there's recently an increase in the demand for care industry migrants who are typically female. In the second stage, we model the migration decision, but we follow the new economic labor migration literature, and we model the migration decision as a function of the household, so the individual and the household members are maximized as joint utility function for the benefit of the entire family. So this is displayed by N here, and so the first component of N is the individual's expected utility in the destination minus the utility of staying where they are, and the second component here is the household's utility having sent a migrant minus the household's utility of not sending a migrant. WI is the weight for the individual migrant's intention, and WH is the weight of the rest of the household, which sum to one. So the probability of migration taking place, migration decision, is that N is positive, given that the household has enough resources, A, to cover the migration cost threshold. So gender influences the household decision in three ways. So the psychic costs of migration endured by the other household members is also influenced by gender. So if the intending migrant is female, her migration may violate certain gender norms or social expectations, which could increase social costs in the household, which could increase the overall migration cost. The bargaining position of the prospective migrant is gender influenced by gender. If the female members are likely to have lower bargaining power in the household, then their weight, the WI, will go towards zero, and the weight for the household will go towards one. So the closer you are to one, the more bargaining power you have in this migration decision. And the probability of employment in destination countries, and therefore the expected remittances, is also gender. So data for the paper come from the 2014 to 15 waves of the Gallup Role poll, which is an annual survey of adults in over 150 countries. We include only low income and middle income countries and restrict the ages to 1864. Our dependent variable is the one you just saw, actually, for migration intentions. Ideally, if you had the opportunity, would you like to move permanently to another country in the next 12 months? And the migration decision variable is have you done any preparation for this move? For example, have you applied to residency or visa? So we assume implicitly that if the individual has purchased an airline ticket or applied for a visa successfully, that the household has collectively decided to send that migrant. Food and security information comes from the FAO's Food and Security Scale, which is a series of eight questions about the individual's experiences, meaning the basic food needs of the past year. So we use item response theory to assess and combine individual's responses to the survey, and a raw scale is then estimated for each country. And each country scale is then normalized to a global reference scale. And this makes that the severity of food and security equivalent across different countries. So the food security in Ghana, after this procedure, is now the same as the food security in Mexico, for example, which is a pretty big innovation for this data. And then, so how I code the person's food and security, so an individual can be food secure, mildly food and secure, moderately food and secure, severely food and secure, depending on how many items of the survey they affirm. So we model separately migration intentions, and separately we model migration decisions. So starting with migration intentions, we use a hierarchical linear model, which controls for the clustering of standard errors, the individual, the region, and the country level. X here is a demographic characteristics and socioeconomic characteristics. F is a function of the individual's ordinal food security status, W. And X is a vector of country level characteristics, GDP unemployment. For the migration decision, we use a hierarchical binary choice for the sample selection. So in our data, we only observe the migration decision data for those who selected one or said they had a migration intention. So for example, Y here is the migration decision, and M is the migration intention. So we only observe Y if M equals 1. So to control for the sample selection, we have to use the sample selection. So here's some example results that we control, obviously, a lot more covariance to this. So hypothesis one is that we expected a positive relationship. So this table is for migration intentions. Hypothesis one was that we expected to see a positive significant relationship between food insecurity and migration intentions, and we see this as the case. Hypothesis two is we expected migration intentions to increase with the severity of food insecurity. And we see that the more food insecurity you are, the more likely you are to intend to migrate. Hypothesis three is that we expected gender differences in migration intentions, and we see that women intend to migrate 3.8 percentage points less than men. In columns two and three, we decompose a sample by gender, so female individuals in column two and male individuals in column three. We see that for each level of food insecurity, the coefficients are slightly larger for men. And interestingly, given the different care responsibilities for women in developing countries, even the most mildest level of food insecurity acts as a driver of potential migration. And some other determinants, so we find that those living in rural areas intend to migrate less, and those that live in urban areas intend to migrate more compared to those that live in small towns or villages. The more educated intend to migrate more, those who are married intend to migrate less than those who are not. Those who are socially connected in the country of origin intend to migrate less, and those that are more socially connected outside of their country of origin intend to migrate more, which corresponds to the previous literature in the diaspora's migration. So for the migration decisions, this is a sample selection model, so the top panel is the outcome equation, which is migration decisions, and then the bottom panel is the selection equation, which is migration intentions. So for the first hypothesis, we see that food insecurity significantly decreases the probability of making a migration decision. The more food insecurity you are, the less likely you are to decide to migrate. We see the gender plays no significant role in the migration decision, although the gender effect could be captured by other co-bearants in the model. When we look at decomposed sample by gender, we see that the coefficients are slightly larger for women now, and we see that the mild food insecurity category is no longer correlated with migration decisions. So we find that individuals in rural areas intend to migrate more, but this is almost completely being driven by women in developing countries. And in the selection equation, we again find that migration intentions increase monotonically. So to conclude, I find a significant relationship between food insecurity and migration behavior. This implies that this vast literature, the determinants of migration, may regularly be ignoring an important component of migration decision making process. I find that significant gender heterogeneity, so this implies that gender-related migration policies are crucial, and that blanket approaches to development assistance may not be effective. And the results also imply a crucial need for an increased coordination between the international food security and migration policy agendas.