 I'm Susanna Sanstern from Old Bucketley University. I have a question for the second and the third speaker. So in this round of study, I was wondering why is the public works actually, I understood it's offered for the whole year. So why is that? I mean, wouldn't it be more sensible to... I mean, it will intervene with agricultural work and it's probably also impacting on the selection into the public works. So is that really sensible? And the second is when it comes to poor households and their sort of tendency to select into the program, is that a targeting problem? Because they are not sort of allowed to select into, or are they targeted, but they select out. And if it is the second, my suspicion would be that these households probably are, and as the results also suggested, they are too labor constrained. So actually then public works is not a good way to help these households because they cannot participate because they have to get a labor in the household. And then to the third, I mean, I'm a bit concerned there that you have like a problem with common support there in the sample so that the control group is too different from the actual... So for those receiving transfers, they have a different profile. They are less in agriculture, I think, or at least the extent to which they do agricultural work is different between the control group and those receiving transfers. And so that might impact a bit on the results. Then I was also, you know, I'm not sure if actually if sort of negative labor effects is often a negative thing, as you suggest, maybe it's just that kids go to school. And I was wondering if you have more sort of information about the time use of the households in the data so that you could look a little bit more carefully at... So if the households don't work, what do they do instead? Thank you. Thank you very much. I'll take a few questions. There is one behind. Hi, I'm Erin Lentz from Cornell University. I had a similar set of questions, so I'll try and keep mine brief. The mine is also directed towards Edward. I enjoyed your talk very much. And what came to my mind is both that decreased labor isn't a bad thing, especially in labor constrained households, especially when there's a possibility that you don't have a measure. My understanding is for within household work. So time spent taking care of kids, time spent potentially just recuperating. Like it's quite interesting to me that you have these unhealthy people or working less, maybe they're getting a chance to get better. You know, this is potentially quite a good thing. So I guess I was wondering if you could comment a bit more on where the interest in the decrease of labor supply is coming from. Is that a policy concern of LEAP? Or is that something else? And if so, it seems to me you have quite a good opportunity to say, maybe this isn't the right thing to be worried about. Thank you. I can't in the front. So thank you. My name is Hayford. I'm from Ghana and Lund University. My first question goes to the second speaker. With the presentation, I was wondering, is there an exit strategy at the point where once someone gets enrolled onto the program, we expect that, you know, for example, the Ghana program, I'm aware of it, the LEAP comes okay. At the point, we expect you to LEAP out of poverty to be able to be self-sustainable. But with regards to the Ethiopian case that you just reported to us, do we have anything like an exit strategy? Then also to the first speaker, I was wondering, I was looking at your sample size and I got a bit confused about the unit of measurement, where the targets that you interviewed and whether or not the sample is good enough for the desegregation you were doing, whether that really brings out the issue you want to consider. Thank you. Thank you. And I'll take another question from here. And then we'll do a second round. Thank you very much. A couple of questions. One to Frank. Could you please explain how you think the mechanism of better property rights leads to economic dynamism if it's not linked with industrialization? Just to spill that out a bit more for us, how that might work in the settler area. And then a question to Renata. You said that the impact of the program was temporary. I wonder if that is linked to the fact that the number of days that people participate is so few in 40, 40 to 50 days and whether or not a longer-term program might have a bigger, more sustained impact. Some related question, why is that impact? Why is the duration so limited? Is it because people opt out of the program or is it because there are rules that restrict it? And if it's the latter, do you think it would be sensible to make it a longer-term program if not a permanent program, which is what some of these kinds of programs are? They're not very generous in the payments. It's part-time and people can stay on for it as long as they require that supplement to their income. And you would expect a more, obviously, the longer people are on a program if there are awareness and other training elements that that impact would kick in the longer they stay on it. Thank you. Perhaps if I could start with Eric and then move this way? All right, so, with the control group, probably being different from the beneficiary group, well, what actually was done through this protest disco marching was actually to see to it that the various characteristics of these households are very similar. So in terms of the work they do, in terms of the composition, and a lot of factors went into that. So, well, I can't really want to say they are quite different from each other in terms of what they do. And then with their time use, the data actually had the use of time for the various households. So we will try to look at that and see what they actually do with the rest of the time if they are not working. Now, decreasing labor supply, no, is not a negative thing. As I said, once we had children as part of the one supply in the labor and now they are increasingly going to school, it's a positive thing. And that is one of the aims of the program to get people, our children, to go to school, to attend school, be in school, and then get out and get a better job. So if labor supply is decreasing and in respect to some of these things, it's not a bad thing. Well, the argument before the program started was that why are you going to give these guys money? Once you give them the money, they are not going to work. Okay, so we start to find out actually, is this argument holding? Is it not? So that's the way, why we went for this study. With the targeting issues, well, this is hearsays. You know, sometimes people bring in some political connotations to wait because they are given to probably those who are in their party and all that and all that. But, well, we don't have hard finders for this. So probably we just have to look more and see if that is really the case. Otherwise, I think it's a good thing that's happening. Well, thanks for the comments and questions. And the first question on why the public works was intended to be offered for a full year. That is how it is stated in the policy document when they set up the program. So initially it was conceptualized that you would provide households with one year full-time employment and that would give them sufficient income to actually make them jump. So that was actually the exit strategy. So graduation should be after one year pretty much because then they would have had sufficient income to accumulate a certain level of assets that allows them to cope in terms of shocks and maybe start up their household or off-farm enterprises already. I mean, run and government is pretty ambitious. So that was the way they thought it out and how it is phrased in the document. Obviously, with the implementation of these public works programs, it's hard to actually ensure that you have programs or projects running over a full one-year period and once one closes, the next one starts up straight away that absorbs the labor force. And that obviously is not happening because at the sector level, you don't have the capacity to actually manage those programs and ensure that there is this continuation. So you do have breaks in between one public works project and the second one and the third and the fourth. Households do select out. I mean, they are targeted. The poorest are intended to be targeted through the community-based process, but they do select out for the reasons that I showed and yes, some of the reasons are that, yeah, maybe they are more labor constrained than the others because they have to take care of members or they are still weaker in terms of their capacity anyways because they are poorer. For the exit strategy, I think, yeah, that's what I mentioned. So initially, the exit was you're participating for 12 months, then you'll be reassessed, but the assumption was that within the 12 months you would have acquired a sufficient level of assets and wealth and then that would sort of make another set of participants eligible. So you kind of have a quite pushing people through the program. In terms of the impact of the program is temporary. That's also related to this initial 12-month targeting period. So yeah, 12 months eligibility and then it's reassessed and given that, I mean, politically they have to show that the program is making progress, obviously, to show that what you do is you graduate people off. I mean, that's a quite strong signal for policy, I would say. So if you can say, yeah, those many households graduated, if it's reasonable and justifiable is a different story then, but I think it's a good indicator. I mean, it's an indicator, obviously. But I have to say there has something changed in terms of the targeting frame. So for public works participation recently, they now have a two-year eligibility period, whereas some people kind of select in and out. So they participate in one project, then they do some other work, and then they go back or something. They don't stay continuously on doing just public works. So yeah, I think that's all from my side. Thank you. Thanks. On the sampling or the unit of analysis, so the unit of analysis is the district usually. There are 110 districts in Madagascar. I lose three or four due to missing data on some of the controls, but most of the analysis is at that level. Some of the estimations on school supply and institutional outcomes are at the municipal level, but then I control for clustering of standard errors at the district level. But I agree. I mean, this is also one point that I found or that I... One discovery that I made in this study is that perhaps, yeah, it is better to look at the household level than, of course, it gets difficult because then you have to take into account mobility. I am going to look at that in the next paper and then in a larger number of countries. The causal mechanisms that link colonial land-right institutions to current outcomes, very good question. I'm also a little puzzled by that. What explains the exact processes in Madagascar? I think, really, it is the usual story that better or stronger land-rights lead to more investments, and this then has spilled over into other areas. For example, when I want to explain the stronger presence of manufacturing firms, this could be informal workshops. Unfortunately, I don't have information on what these firms are, but it could be informal workshops that help with farm equipment. It could also be that there are more businesses to process agricultural products. But it's a very good question. I've also tried to look into the data that I have so far. Don't really let me illuminate that story very much. Thank you very much. There is time for another round of questions, so if you have any further questions, could you? You cannot? This is Vene from the University of Paris-Dauphine. My question is for the second speaker. Was there an explicit goal to achieve in terms of financial inclusion in that program? That is to force people to open a bank account or they must have a bank account. They have to pay for it. Was there an explicit goal to achieve in terms of financial inclusion? Thank you. Are there any other questions? Yes, Lucy. Hi, my name is Lucy Scott from the Overseas Development Institute. My question is to the second speaker. You mentioned the double dividend of public works and obviously you've investigated the impacts in terms of beneficiaries, but I was wondering if anybody or whether you're looking at the impacts in terms of the public works outputs, if you like. Any other questions? Okay, that gives me an opportunity to ask the question myself. I was interested in this split between the direct support councils and the rest. I have two questions on this. In the Ethiopia PSNP, there was a prolonged discussion at the beginning of the program in which the government was very reluctant to expand that particular sector of the program and it had a cap of 25% of participant households. Then it was really difficult to reach that figure, partly because of the way the program was implemented. The first question is, what is the proportion, what is the relative weight of the direct support households and the rest? Secondly, how do you account for that in your estimation because clearly the direct support households have no capacity to contribute work and so clearly the outcomes must be kind of different. It's almost like having two separate anti-poverty transfers in one, in the sense that you have two different groups of beneficiaries. Okay, so if I could go again in the same order and given that we got a few minutes, if there is anything that you were burning to say, you can take a couple, two or three minutes, thank you. In trying to get the amount of money that's good and reasonable for these beneficiaries, what went into it was to look at the consumption patterns and also what they spent their incomes on and so the amount of money that were given to these guys seemed to be adequate and enough for whatever consumption pattern they were involved in. But then with increasing inflation and all that, even with a vision from eight-garner cities to twelve-garner cities at the moment, we really wonder what actually they have really combined with this amount of money. So it's my wish that they continue, not just with the reaching of this amount to really go up, so it will actually put them on a good pedestal to actually be out of poverty. Thank you. With respect to the question on if there was an explicit goal to achieve financial inclusion the bank accounts, there was definitely a goal to expand the financial sector to the rural households. And that's not only driven by actually opening the bank accounts but also by this financial services component which does provide credit and that financial services component has actually now become larger than any of the other components in the last two years, well in the last year pretty much, in terms of size of amount. But it was, I mean, the goal that they wanted to expand financial inclusion was not specifically mentioned in the policy document as such, but definitely an underlying factor. What is explicitly stated is that they want to encourage formal savings so that households have a basis to start investing. And that's why within all these trainings that they give they do strongly encourage that a part of the transfer that you receive goes to your bank account. And in some sectors they have formalized it to the extent that you can only go and take out 80% maximum of your transfer and at least 20% have to stay on your bank account or something. But that's different in certain sectors so there's a bit of variation in organization. With respect to the actual public works outputs and sort of the MISU level effects of the program, yeah, definitely something I would love to look at. We do have sector level data on the infrastructure that has been generated, the transfers that went there, but at this stage it's not really enough to make a robust assessment of it. The rest it's descriptive, that's where we stand at. But we're collecting data hopefully soon again and we hope to tackle these sort of MISU level, community level effects a bit more than. On the question of the proportion of the direct support households to the rest, the direct support households are really the smallest proportion in terms of beneficiaries. In the early years it was about less than 10% actually of households that did get direct support and the direct support households actually do not graduate so they stay pretty much on the program. So once you're eligible for direct support obviously it's like a pension almost. How do we account for it in the estimation? Basically we took out the direct support households of the sample. But the sample size that we have, we only have about 150 actual beneficiary households in the sample in one cohort. And then we have 15 direct support households so it's rather small. Our sample size is really a big problem. But at the early stage, in the short term we didn't really expect much spillovers. However, hopefully also in the 2013 survey we hopefully managed to investigate what's going on there and how much they can contribute. Okay, do you have a final comment? No, sorry. The last word, you can have the last word. No, it's okay. Okay, perhaps one day I would like to do a meta-analysis of how effects of RCTs or all sorts of evaluations vary across countries that differ in historical origins and maybe to draw a link between your work and my work. But I think we shouldn't keep you too long from lunch or from coffee. Thanks to our speakers. We had a really interesting presentation. Perhaps the models in terms of social policy and social protection for Africa probably works which are very common in all countries which some international organizations are very keen on and sort of conditional cash transfers or other forms of cash transfers that are also a different modality. So it was interesting in having the comparison there and of course the question that was posed by the first paper which is what are the likely long-term effects of this intervention? So far we just focus on the short-term effects but we don't have much information on the likely long-term effects. So many thanks to our speakers for their papers and to you for your participation in the session and we call this session to a close.