 Well the paper that I will be presenting is a young work with one of my colleagues at Oxford University So I need focusing on the case of Mexico So I'm very glad that our colleagues from Senegal were talking about some things that we also Reach upon in our studies. So it was nice that I didn't have to talk too much about it but Well the paper is on the website, so I will Skip many things related to the literature, which is very slim but Just by beginning one of the motivations of this paper was That's There being some interesting innovations in the context of Mexico Which tried to link public and private service institutions Potentially this can have important Effects on the on the efficiency of the provision of these words So the idea of this paper or our hope is that we can contribute to the literature on conditional cash transfers and bikes mining at the recent introduction of the electronic payment system of Progressive or now is called Prospera and We are looking at a number of dimensions which are related as well with remittances reception so I will refer to Opportunity that is Prospera or progress as we was initially referred to as pop so Mr.. Gio a background for the program pop was introduced in in August 1997 Initially to cover about 300,000 households in basically rural areas By the end of 2015 the program supported about 6.1 million households Which overall represent 25% of the population of Mexico? So these are this is a massive program, right? so pop initially paid cash Distributional points located in towns and this was very costly for recipients because they had to travel you have to also absorb some opportunity calls from often have to leave their businesses and they have implications about Looking up after kids and and also there were some risks related to those kind of approaches because as you most likely are aware crime is an issue in Mexico. So Often they were robbed by gang separating in this location. So there were a number of concerns about this approach. So in 2001 there was an important legal reform Mexico that tried to put some more coherent a Structure to the financial system and that included non banking institutions So the idea of this study is try to exploit the fact that The electronic payment system or pop was introduced. There was a pilot And it was implemented by these Bansefis called is a state-owned Development Bank in Mexico together in partnership with a very extensive rate of non banking institutions creates association Savings and create cooperatives microfinance organizations and so forth so the advantage of this strategy was that Non banking institutions affiliated to la redela gente, which is pop's network people's network, sorry Focus on localities were Progressive operators. So they basically focus on rural areas on peri urban localities and that opened the possibility to introduce this pilot So this map in a way reflects distribution of those Branches that belong to pansefi and non banking institutions and you can see the distribution is quite Spread across the country with more density in the more populated areas as you can find so well in a way what we are trying to do is We Studied the effect of these electronic payments by taking the advantage of the exonium of variation of the introduction of the program So so and the way the introduction was decided by the managers of progress and bansefi together with the red we la redela gente and that decisions were Captured in a survey that was implemented in 2004 It was a panel data set that the introduction of the pilot began in 2003 and the collection of these data panel Like I said was began in 2004. So we don't have a baseline, but nevertheless we capture households recipients of opportunities with Their cash benefiting cash or receiving progress and cash and houses receiving progress in a electronic banking account so of course Even if we can Control for sample sex election because of the way the pilot was implemented Still we cannot rule out problems of then intergenerated because of non random Selection of the non banking institutions that were participating in this pilot simply because These organizations Were likely to be selected when they were operating in localities with better infrastructure No, so most likely localities which were poorer or more remote they had let's say weaker Systems to provide these cash transfers and therefore most likely were excluded so the idea of Whereas we don't have problems with cell selection. We still have a selection problem by by by these By the fact of the rollout of the program So therefore we compute some average treatment in fact using some matching estimators So so the concerns about matching as you know is basically related to unobservable So we believe the main concerns about unobservables are somehow addressed through the way the pilot was for that But nevertheless we cannot roll out other factors. We are quite confident that We are controlling for a very important variation in this observables so Well deal the literature is very very thin so what we are trying to do here is to look the four-year impact of pops electronic payments on a number of outcomes savings decisions remittance the reception and company strategies again, I just incorrect shocks and and And this is very important because in our sample as well as you may know in Mexico one in every five or fifth household Has a member who is a migrant? so the probability that members of those households That had a migrant work is very high and we will see later that does the case So we also try to unpack some on the nap on the line mechanism through which we observe those effects So we are trying to establish some causal mechanisms so Well because I'm not very good at keeping the time So I just give you the main results and then let's see how it goes But overall what we find is that the electronic payment in a way Decrease the participation in informal savings arrangements like Roscoe's which makes sense of people when they are Faced with an opportunity to have savings in a bank account They most likely choose to use those services Although we don't find any significant effect on savings at home So, you know people keep keeping some money under the mattress for for some eventualities But the core of the savings we find a significant shift The other one is that households were less constrained on remittance reception And a result they were less likely to reduce consumption or contract loans to deal with shocks Which is in a way more or less what you were finding as well So we also find an important degree of heterogeneity in terms of the environments that characterize those households in particular between rural and urban localities so We also find that the nature of the financial institutions play an important role The opening of the bank account was free of cost so they didn't charge any fees And also the fact that they have a much more friendly environment towards clients Contrary to the traditional banking institutions in Mexico in a way help the relationship with those households Right so Here we are the other way Okay, so As I said pop is the largest cash transfer in Mexico and it's one of the largest in the world The eligibility is based on a very rigorous methodology into stages the first stage Identified localities which are poor based on a census based Marginality index and in a second stage once you identify the localities you run a census to identify Houses based on proxies proxie means test and categorical criteria houses with children and so forth So the criteria of selectivity is very very rigorous. So the households who receive the cash or the Progressive in cash or in electronic transfers are very similar on average, you know Nevertheless, there are some heterogeneity in the sample and we will see later, but Pop's income support is distributed every two months and is given to women The amount of the transfer is quite significant It's about 20% of the average of household income and more the type of population And virus because for example, you have more children you get more money and you get if you have more girls you get more so the idea is to compensate the opportunity cost and also Norms that are quite persistent in certain communities in Mexico So Well the pilot in a way involved As I said a number of institutions and The as I said the accounts were free of opening and maintenance costs so there is a previous study that just focuses on administrative data and He finds a very significant reduction in in financial and transaction costs for the household So which is something in a way positive what we do is to look at second-order effects here and and Looking at a very representative sample the other sample is a very small one so During the phase of the pilot phase More than 90% of all recipients continue to receive the transferring cash and just a small Share was shipped to the electronic trans and this can be seen in this graphic So as you can see most of the houses receive progressing cash in 2003 Percentage about and 5% began to receive the transfer in savings account. So our study covers The period in which the survey was collected and then over time Progress was moved towards a much more electronic approach and by 2011 or every single house will receive the cash in either An a savings account or in prepared cards. So nobody receives cash anymore So what we are trying to do is in a way to exploit this phase to find out the extent to which has an effect on different dimensions Oops So the data well the the survey covers 25 Out of 32 states in Mexico The sampling frame was assigned to be represented at the three renal areas in Mexico the sampling follow Frame which was random, but with a probability proportional to the number of clients, which means that those localities with most likely better infrastructure We're more likely to be selected in the sample and this is what we try to control later with these matching estimators so overall the survey cover about 17,000 observations Houses in this case and for the purpose of this study because not every single Client of these institutions receive progress so the sub sample Cover about 3000 households, which is quite a significant number And that what we try to do is to look at households who were always compliers, which means that they always receive either the cash Transferring cash or an electronic miss over the period of analysis As you saw before there is not there is no much variation in this period and therefore Difference in difference approach was not possible because there was no much variation unfortunately, right so The the decision about treatment was basically decided based on Geographical criteria so houses who live in a radius of 10 kilometers around the branch Selected this house of some members no so ideally the sample has Members borrowing from sorry members of these institutions and also houses living in the same localities Which didn't have access to this Services So the the issue of the covariate as I said it was a source of concern for us So and then when we run a very simple details We actually find that where the atoms don't exhibit significant differences the covariate does which in a way confirms our concerns about the heterogeneity in terms of the in the identification of the of the institutions which Have to deal very often with the characteristics of the localities whether urban or rural whether they have certain infrastructure and Therefore the idea of using the machine it makes sense. So so in order to Do this empirically we obviously having a linear model of this kind will be biased For the reasons that I just explained So what we do is we will follow two approaches the first one is to estimate a fully interactive linear model which as you can Imagine just the interaction allows to find out whether there is a significant Heterogeneity model. So if you find significant interactions, basically you have to deal with this heterogeneity through a different approach So I don't present this but it's in the papers of the results actually confirm our concerns that there is an issue of heterogeneity So therefore we focus on a propellant score matching To win away construct a synthetic was experimental approach So I'm sure all of you know about all these so I don't want to go into this But we essentially find another mahalo novice this and metric matching of this kind For for the empirics once we estimate the mahalo novice metric We submit an average treatment effect on the treated after computing the matching algorithms, which we essentially follow three different Approaches which in a way report very similar results and Everything is explained in the paper so so for in the matching covariates the The variables that exhibit some problems and also we're on some like As it's like a fixed-effects estimates using matching by controlling rule and an urban Localities and you can see after we run the the matching the Thunderized percentage bias across different for various Produced Before The heterogeneity in observables was basically removed and issues about unobservability unobservable heterogeneity We believe at least related to self-selection is also not a concern Right. So the results is that overall the electronic transfers as I said decrease the propensity to participate in Roscoe's We find the transaction and opportunities are the main drivers behind this We cannot rule out other interhouse on dynamics because we don't observe that but we cannot rule out potential Dynamics in that respect. So the propensity to save at home again was not affected by this intervention What is interesting is that? We find that Households were much more likely to use savings to cope with with shocks And the transmission mechanism was remittances. So remittances reception increased substantially by 90% so So because houses were suddenly receiving more frequent remittances They were able to save more and through savings is the mechanism through which households reduce for example Savings or other sorry they use savings or To cope with we shock. So I think that this is something that is very important So in a way remittances is the underlying mechanism through which savings can be accumulated and reduce different Aspects that can be detrimental when Houses are faced with shops So we find an important heterogeneity in terms of the house of composition and also the way they live Which I don't go How many minutes we have? Okay, but nevertheless Overall there is a significant level of heterogeneity which means that It's important to look at the conditions in which these problems are implemented to to in a way fine-tune the effectiveness of those Interventions and well in a way what we we argue is that and Since migration is a very consistent or persistent phenomenon in the concept of Mexico progress in this case facilitated Together with these financial institutions an additional mechanism to deal with shops For houses which are poor or very vulnerable. So Obviously the what matters is how these Alliances between public and private institutions can maximize the effectiveness of these policies. So So well, there's a lot of things in the paper, but I think that's it