 So around 650 million women and girls alive today were married before the age of 18. Previous research have already said that child marriage leads to lower educational attainment, early motherhood, higher fertility, and worse education and health outcomes of the children born into these child marriages. So eradicating this practice seems to be a key component in the fight of fighting against global poverty. Even though most countries in the world have lost established a minimum age of marriage at age 18, there exist a lot of exceptions that make in practice child marriage legal. Some of these exceptions, for example, are parental consent, pregnancy, court authorization, or just religious laws. So a common proposal for ending child marriage is just to eliminate all these exceptions. We're going to be looking at this question more closely in the context of Mexico, and the first thing that we want to analyze is if there's increasing the minimum age of marriage to age 18 without any kind of exception actually is able to reduce child marriage rates. Now it's important to say that child marriage rates do not only include formal marriages, but also include couples that they cohabitate at these marriage where one of the partners, normally the girl, is below the age of 18. And actually in this paper we're going to be focusing on child marriages where the girl is under 18 because child marriages where the boy is under that age are actually not that common. The second question that we actually want to investigate in this paper is the application of these laws are actually effective at reducing the negative consequences normally attached to the child marriages. In particular we're going to be analyzing what is the effect of banning child marriages on the girl's educational attainment, the probability that they become mothers before the age of 18, and the infant health at the moment of delivery of their kids if they end up having kids. Actually it's quite challenging to analyze the causal effect of banning child marriages and the reason is because these kind of policies are normally implemented at the national level at one point of time, leaving us with no contractual to analyze its causal impacts. So in this paper we're going to do is we're going to be taking advantage of a particular implementation of a child marriage ban in the context of Mexico in 2014. So before 2014 child marriage was legal in all 32 states of Mexico. Miners could just get married if they had the consent of their parents and a minimum age. This age was normally 14 or 16 depending on the state. But kids, even if they were younger than in age or they didn't have the consent of their parents, could still get married if they caught the solicitation of the court or the municipal mayor. In some states also they were allowing very young girls to get married if they were pregnant. In December of 2014 the Federal Congress passed a law increasing the minimum age of marriage at age 18, eliminating any kind of exception. However this federal law was not directly applicable and each of the states had to modify the legislation to make this change happen. How they modified this legislation was actually quite gradual. So in the paper what we do is we take advantage of this gradual implementation of the laws across states over time using a difference in difference model. Another advantage of the context of Mexico is the existence of granular data on births, marriage and school attendance. In particular we are going to be using the universe of all birth certificates and marriage certificates in Mexico and this large amount of data is going to allow us to identify if there's any effects on birth and marriages even if the effects are very small. So in our paper we have two main results. In the first result we just focus on what is the effect of banning child marriages on child marriage rates. And we actually find that child marriage bans reduce marriage formal marriages of 15 and 16 year olds by around 49%. We also find a reduction in formal marriages for 14 and 15 year olds, however this effect is a bit smaller given that child marriage was not that common at this ages. It's important to know that this effect does not tell us that these laws were not effective. In our estimation strategy what we are able to capture is what is the effect of these changes in laws beyond and above the national wide decrease in child marriage rates that was already happening at that time. We also analyze is what is the effect of this reform on the marriage rates of 18 years old. And we actually do not find any impact there. Our interpretation of this result is that these couples are even delaying longer than they should to get formally married or these couples are just dissolving before they arrive to the minimum age of marriage. So in our second result we want to see if these laws were actually effective and reducing the detrimental effects normally associated with child marriage. However even though we find that these laws were properly applied we do not see any effect on girls' educational attainment, the probability that they become teenage mothers or the infant health of their kids. The reason why we don't find any impact there could just be driven by the fact that the decrease in the amount of formal marriages is just upset by the increase in the same amount of informal unions. However this very interesting mechanism is actually very difficult to investigate and the reason is because normally girls at that age have a lot of incentives to lie about their civil status when they are surveyed. So in order to analyze these mechanisms we are going to be taking advantage of the fact that birth certificates normally have data on the civil status of the mother at the moment of giving birth. Also mothers that are having kids at such a young age have less incentives to actually lie in the birth certificate about their civil status. More importantly for us, before the reform 65% of child rights were having their first birth before the age of 18. So in this pool of very young women we are going to have a large percentage of the girls and women that are going to be affected by the reform. What we can learn about these results is that informal unions have a very low social penalty in Mexico and girls that want to get married when they don't have this possibility they just substitute these formal marriage by informal unions. Therefore in this context that is very common in the whole of Latin America, banning child marriages will not be affected at reducing some of the detrimental effects that we are very worried about and will not have any impact on girls' educational attainment on the probability that they become teenage mothers. So our results are very related to previous research that has studied what are the determinants of early unions and also studies that analyze policies that have the objective of reducing these kind of practices in particular. In fact, there's already a couple of papers out there that they have analyzed what is the effect of increasing the age of marriage. However, the contribution of our paper is to analyze these kind of questions in a context where formal marriages are not the only option for these young couples. In particular, we are the first ones to find that when informal unions is a common practice, banning child marriages are not effective at reducing child marriage rates because girls just substitute these formal marriages by these informal unions. We are also the first ones to analyze these kind of questions in the context of Latin America. Latin America has been very understudied in this topic and is particularly important because it's nowadays home of almost 10% of child rights and the drivers of child marriage in Latin America are quite different from the drivers of child marriages in other places in the world. So of analyzing if these kind of reforms are actually effective in this area will be very important. Also, even though in the rest of the world we have seen a decrease in the child marriage rates, in Latin America these child marriage rates have remained constant over the past 25 years. As the reform we are examining in this paper was really recently enacted, in the paper we just focused on the short term effects of this reform. In the future we plan to investigate also the medium and long term effects of these reforms because ex ante is not very clear what should be these effects. In particular, the substitution between formal marriages and informal unions that we find now would have negative welfare effects on the girls if these girls that are now in informal unions do not have the same legal recognition or the same legal rights as girls that end up in formal marriages. With this project we actually find that these kind of reforms are not effective at reducing child marriage rates. So what the next steps would be is to investigate what kind of policies are actually effective at reducing child marriage rates. Previous research have already pointed out that policies that give girls economic opportunities or direct incentives to delay marriage actually work quite well in reducing child marriage rates. The next steps would be to understand if these kind of reforms are also going to be effective in a context like Latin America where the drivers of child marriages are very different.