 Hi, so I'm Sarhan. I'm at the University of Göttingen and this is also part of the women's work project. My paper, it ties really well with the keynote from this morning. So I'm looking at the effects of wealth shocks on early marriages and the role of marital customs like Dari and Bright Price. So we know that following from Sudeepa's presentation that child marriage is pretty common worldwide. Right now women aged 18 Well, about seven main women who are alive today are married before the age of 18 And it's quite common in South Asia and Sub-Saharan Africa. So about 50% of women are married for the turning 18. In Pakistan, it's the same thing. So despite increasing schooling, we see early marriages, especially in the rural areas and the recent DHS data shows that one in three girls in rural areas is married before the age of 18. And also in the literature there's an established relationship between early marriage and bad outcomes for the mother and the offspring. So there's lower education attainment, more domestic violence, lower bargaining power, worse health outcomes for the mother and the child. So I wouldn't spend time on that because I was already discussed. So I'll just briefly explain what I do in this study. So I try to investigate the effects of financial shocks on the probability of school dropouts and also early marriages of school-going girls in rural Pakistan. So the main question I address here is are rural households more likely to withdraw their girls from schools and marry them off if they're hit by some adverse wealth shocks? And secondly, how do marital customs affect early marriages of households which are experiencing some sort of wealth shocks? And then there's another section in the paper which looks at the returns of education in the marriage market. But I wouldn't discuss that today because I don't have enough time. So basically I estimate the returns of education in marriage market and then also look at intergenerational effects of that on the kids of mothers who receive marital assets. Yes, so why are girls married off at a young age? So the first thing as we discussed today in the keynote is that cultural norms and social norms they are the main culprit for early marriages. So there's a paper by Field and Ambrose and they show that there's a preference for younger brides because they are perceived to be more fertile. And secondly also, there's I think is a report by for many countries that shows that low-income households are more likely to practice child marriages. So yeah, they're twice as likely to do that because they perceive the child as the girl child as a economic burden and they want to reduce the number of dependents in the household. And then there's also some literature which looks at traditions of marriage payments as economic drivers of early marriages. So there are two types of marriage payments, dowry and bride price. In case of Pakistan, we have them both. So there's also payments of dowry and bride price. In India, we mostly have dowry and then in Africa we would have bride price. So I think it's a paper by Ashraf which shows that it's economically lucrative to marry off girls at an early age because that's where you get more assets. So dowry is increasing in bride price in bride's age and bride price is decreasing in bride's age. So it's lucrative to marry them early. Right. And then lastly there's some papers which look at wealth shocks. And I think Kornu and Buena have a couple of, I think right now they have three papers which look at wealth shocks and early marriages. So the one, that's my main paper of reference here. So they showed that in Tanzania, places where they've had rainfall shocks, girls are more likely to marry off early. And the idea is that if they are credit constrained, and if they marry them off early, they receive a higher bride price. And this is used as a consumption smoothing insurance if a household experience is shocked. So that's my main idea here. I want to exploit the same idea that they have in this paper. So I'll just give you a brief summary of what are marital assets, dowry and bride price. So in Pakistan's case, we have both. So on Islamic law, you have the practice of bride price, which is that when the marriage contract is drawn out, the bride gets a fixed amount from the groom. And this is a transfer given to the bride directly. And then we have dowry, which is the opposite. So dowry would be a payment or a transfer made from the girl's family to the groom's family. So the assets basically goes to the groom's household then. And Becker proposes that dowry is a pre-mortem bequest. So it's a transfer from the parents to the girl. And this is to compensate for the lack of property rights. And then in literature, if you look at this, they use assets as a proxy for bargaining power. There's also some that show that it's a negative effect, but there's also some that show that there's a bargaining effect over here. So especially in the non-corporative bargaining literature, they show that it's a non-neighbour income and it helps in the empowerment of women. And yeah, so there's these papers which kind of go in that direction. Right, so I'll come to my data here. I'm using the Pakistan Panel Rural Household Survey. This is done by IFPRI. So they have three panels, but I'm using just one panel from 2010, 2011-2010. And because this panel has information on marital assets, and I have 4,000 households in 16 districts in three provinces. And again, in these provinces, they have both customs of dowry and bride price. And in the survey, it's quite detailed. Besides the regular household questions, they have a very detailed section on marriages where they have a decent formation on bride price and dowry, where they list every item received and the monetary value of every item received. So what I use is that I use the marital assets as a sum for each woman here. And besides that, we also have a module on shocks. So you have positive and negative shocks going back five years. And again, you have value, the monetary value for each shock in rupees besides the type of shock. So I have natural disasters, health shocks, income shocks, and so on. I mostly, I think 70% of the households experience floods. And that's my main shock that's in the data here. Right, how much time do I have? Right, okay. So I'll just quickly go over my descriptives. So I have information on marriage. So basically the rupee value of each asset dowry and bride price, information on the woman's family, her parents information, how they work and their education. And then also within the marriage, her empowerment if she can, if she has rights to divorce and so on. And then also the AHS marriage and so on. And then we have information on schooling besides enrollment, there's information on the village level school characteristics. And then lastly, there's the monetary shocks, which go back five years. And I just use a sum for each year of all shocks combined. But again, most of them are coming from the last 12 months. Right, so my estimation, I'm using, I'm following Kornow and Wenna. And so their model predicts that if households experience shock at a certain age, before the age of legal marriage, they might marry off their daughters at an earlier age. So if households experience shocks, they would get married them off. So here my shock variable is shock at age 1870 and lower. And that's the monetary value of each shock in each year. And then I have the effects of that on early marriages and also school dropouts. And then I have controls, which I discussed earlier, the standard controls that you have for the household, for the parents, for the village and for the schools in the village. And then we have district fixed effects in there. So and then looking at the effects of custom marital customs, I augment this regression with a measure of average value of assets in the village. I'll come to that in a bit. I'll explain more what I do here. And then the second section in the paper, it discusses the intergenerational effects, which where I estimate the returns of education in the marriage market, but I won't, I don't have time for that. But if you look at the paper, the draft has that section so you can read it if you want. Right, so I'll just discuss the results for the early marriage and also for school dropouts. So these are my first results for school dropouts. And I have here the shock at each year, the rupee value of shock. This is 12 months before the survey, and this is 24 months and so on. So if I look at my results without the controls, I do see a small decrease of enrollment for girls in the first year, just 12 months before after the shock, but there's no effect in the long run. So any shock which is more than 12 years, 12 months longer, there's no effect there. But once I add the controls, there's no effect for gender. So both genders experience a drop in school, schooling. So the idea is that there's no disproportionate removal of girls from schools. So girls are not affected disproportionately by shocks. Boys and girls are equilateral to be removed from schools. But again, the effect is really small. So if I look at this, if the shock value increases by 1000 rupees, school enrollment would fall by 1 percentage points, both for girls and boys. And again, this doesn't last for over 12 months. So basically there's no strong effect. So households are not using school dropouts as a coping mechanism for any kind of well shocks. And then I do an analogous regression. We're looking at early marriage question. And I have the same thing, but looking at the marriage question here. And I do not see any effects of early marriages for girls. I can't estimate this for boys because the sample is too small for boys. But for girls, there's no effect of well shocks on early marriages. There's no increase in early marriages. In fact, for girls younger than 13, there's a decrease in early marriages, which goes with the idea of puberty. So you wouldn't have a marriage before puberty according to Islamic law. And therefore there's no risk of early marriage for girls that are younger than 13. But again, the theoretical prediction was that if there's a shock, there will be an increase in marriages, but we do not see that. And I could speculate on the mechanisms here. So it's possible that these households, because of climate change, we have a lot of shocks happening all the time. So it's likely that they have a larger toolbox of coping mechanisms. So they wouldn't use early marriage as a coping mechanism for these shocks. Secondly, because of increased schooling, there's an incarceration effect and girls are staying in school longer. So they wouldn't be married off because of shocks if they have any. The third thing could be that there could be an interaction with marital customs. So because these households are experiencing shocks, they can't afford to pay for dowries and bride price. So instead they would delay marriage for a time when they have more money so they can pay for the marriages. So I want to explore that question where I look at the effects of marital customs on early marriages. So what I do here is that I augment this regression with an extra variable, which is the average value of assets in the village. So I look, yeah, but this is the average value of assets in the village in the year before. So the idea is that parents, they have an indication of how much asset they would need to pay for the daughter to marry her off. So it's like expectation of how much they need to pay to marry the daughter. So if the assets, the value in the village is quite high, that means that they expect to pay a lot of money for the bride price or dowry. And if they've experienced a shock in the last 12 months or longer, then they are quite constrained and they wouldn't be able to afford to have the marriage. And that would decrease child marriages then. And that's what I do here. So if I interact the value of the shock with the average value of the asset, I see that if there's a one sign deviation increase in the interaction of these two, the early marriage is defaulted by one percentage points. Same as it's for all assets, also for dowry and bride price. The effects are still there for shocks that have been long term. So the idea is that households which do not have access to credit right now, because they've been hit by shocks. And if they expect to pay a higher dowry or a higher bride price, they would reduce early marriages. It's a small effect but still there. But if I had a bigger sample, I would say this is the lower bound of the effect here. Yeah, so that's the main analysis on early marriages. Right, so I'll just recap everything. I don't have time for the insurance education. So basically in this paper, I look at the driving forces behind early marriages. And also we contribute to the gender dynamics of coping mechanism in rural households, which I experienced the final shocks. And lastly, I look at the marriage markets and how the returns play an important role in empowerment. Right, so again, so if I look at the wealth shocks, there's no effect of them on school drop off of girls disproportionately. In fact, girls and boys, they're equally likely to be removed from schools if they have a wealth shock. Similarly, there's no effect of these wealth shocks on early marriages. In fact, in households which that expect to pay a higher dowry and higher bride price, there's a decrease in early marriages if the households have experienced a shock in the last 12 months or longer. Yeah, and then I wouldn't go through the next one. But yeah, that's my main finding that if households expect more dowry and bride price, then they would not marry their girls.