 I'm going to be talking about employment-oriented social protection, and in particular the National Employment Guarantee Scheme in India, and the targeting the ultra-poor programme in Pakistan. Now there is a very sound rationale for focusing on employment orientation, because of the fact that we are in a part of the world which has had quite strong rates of growth, but it has been described as jobless growth. It simply has not generated the level of decent jobs that would help to pull large sections of the poor out of poverty. But it isn't just the deficit in jobs that is a problem in South Asia. It is also the problem of intersecting inequalities of location, caste, gender, ethnicity and so on, and the ways in which these trap people in chronic poverty over generations. And because of the scale of the problem, we have to remember that only about a maximum of 10% of people are in formal employment in South Asia. The scale of the problem means that we cannot talk only in terms of social assistance. So the language of protection and promotion is one that has originated from work in South Asia, and it refers to finding forms of protection that contain the potential for helping people to climb out of poverty, what are called opportunity ladders. Both these programmes are relatively new, as part of the emerging, but interestingly both of them have their roots in other programmes that go back many years. So one of the first points I think I'd like to make is that policy continuity is really quite important in the field of social protection. Partly because you need to institutionalise social protection and introduce a bit of predictability in the lives of people who are living very insecure lives, but also policy commitment allows for learning from experience. And both the programmes I'm talking about are the products of quite a lot of experience. So the NREGS has its origins in the Maharashtra Employment Guarantee Scheme that was put in place in the state of Maharashtra in 1978 after some major droughts. And that introduced the idea that the government had a legal obligation to provide employment on demand if a minimum of 15 people within a locale demanded it. The employment had to be provided within eight kilometres of that locality and within 15 days or the government could be taken to court and an unemployment benefit could be paid. It required equal wages for men and women, on-site drinking water, childcare and first aid. India has had many other employment public works programmes, but it was not until 2005 when the United Progressive Platform came in that the idea was that the government could be taken to court and an unemployment benefit could be paid. It was until 2005 when the United Progressive Platform came in that the idea of an employment guarantee was revived, this time at a national level. So the National Employment Guarantee Act built on the Maharashtra Scheme but added a few features which were lessons learnt. One is that it limited the guarantee of employment to just a hundred days rather than indefinite. It required state minimum wages to be paid. It banned private contractors in implementation and put the locally elected officials in charge. And it had, because of the process deficits that have been documented, it had a number of accountability provisions such as the right to information about documents, the public payment of wages and social audits. Later provision was made for wages to be paid into a bank account and according to the website wage employment in 2013-14 was for over 73 lakhs, a lakh is 100,000 individuals of whom 48% were women, so much higher than the quota. Bangladesh has also had public employment programmes and in fact in the food crisis of 2008 the government hurriedly introduced something called the 100 days employment. But it was very hurriedly implemented and it is to another programme, the rural maintenance programme, that we might look to see how lessons are learnt over time. The rural employment programme was the product of a pilot in 1983. Bangladesh is a country where women are not supposed to work in the public domain. And the existing food for work programme, many women were turned away because officials didn't approve of them, others didn't even come forward. So they experimented and then built a rural maintenance programme which was meant only for what we call destitute women. Either women without husbands or alien husbands, women who are primary breadwinners. And these are about maintaining rural, urban works all year round and they provide work to tend astute women per union. And because demand outstrips supply, the women are chosen by lottery. There is a mandatory savings provision, about a fifth of their wages is put into a bank account, which they can't touch for the four years that they are on that programme. And regular training was added over time, first of course in road maintenance, then in health, social awareness and then in business and financial skills. So around 42,000 women were provided employment each year and in 2006 the government took it over. It was managed by Care International till then. Targeting the Outrepoor also has its origins. This time in a vulnerable group feeding programme that was put in place in 1975 after a major famine which evolved into vulnerable group development. Brach and World Food programme got together to add an income generating component to the vulnerable group development programme which had provision for giving women access to loans, microfinance, training and so on. But it was found by the end of the 1990s that the extreme poor amongst these groups were simply not benefiting. They were dropping out, they were not taking advantage of loans and so on. And so in 2002 we got the targeting the Outrepoor programme, which I think is a very good example, a paradigmatic example of a programme designed to overcome the multiple and overlapping constraints that make up the gendered nature of extreme poverty in Bangladesh. So it has a mixed methodology to identify people, it has a transfer of productive assets plus a cash stipend so that women can focus on their enterprises rather than having to do wage work, health support, mandatory savings, training and mentoring, intensive mentoring. I persuaded the brag officials not to call it handholding because that sounds a little patronising. So we call it intensive mentoring and elite support and by 2011 it had reached 400,000 women. Now when we talk about sustainability, when we talk about sustainability obviously the financial aspect is very important. Because I think this new generation of evaluations which point to the wider impacts of social protection are very important in justifying these programmes. And what we find with the NREGS is that because it's been scaled up to the national level, very mixed outcomes, however while we see evidence that there is some rationing of jobs because demand outstrips supply, it is reaching the poorest groups and it is very effective in reaching the tribal and Dalit, the lowest caste groups. And as we saw women make up far more of the 30%, it has had significant poverty reducing effects particularly for the poorest, changes in people's lives and sending children to school and so on. However one of the more negative depending on which side of the debate you are aspects of it is that it seems to be driving up agricultural wages. Particularly female wages which is not so good for farmers, small farmers, but it's very good for agricultural workers who are amongst the poorest paid in the Indian economy. But it's not clear if the wages are being driven up by market forces as someone has suggested or by actual active bargaining. Certainly the Maharashtra Employment Guarantee scheme documents many examples of workers being empowered by the scheme to bargain for their higher wages. There's been the decline in the stress migration in West Bengal and these are the state level findings. And also interestingly given what we've been hearing from the others, participation in the NREGS results in better educational outcomes for children, particularly girls. Over and above the income effect and it seems to be related to the increase in bargaining power, decision making power within the household. While far this participation is associated with negative educational outcomes, particularly in poorer households. Of course the process deficits continue, we have 40 muster rolls, long delay in payments, corruption and leakages and so on. But as one of the less positive evaluations concluded, despite this what we need is an overhaul of the governance of the programme. But that sense of entitlement that it has instilled in very poor workers is one that needs to be valued and built. Development potential of public works, there was a very interesting study a long time ago about the local economy effects of the Food for Work programme. Wage employment increasing, self-employment declining, agricultural production increasing and so on. Nothing of this scale has been done more recently, but the care evaluations of the road building projects show overall increases in commercial freight and passenger volumes. And that while men continue to make more use of the roads that are built, the rate of increases is much higher for women. And in particular women's access and children's access to school, health clinics and NGOs have been improved by the building of infrastructure. The other evaluations of the rural maintenance programme show what it does for individual women, that many of them continue to earn what they were earning on the programme after they've left and are able to provide means for their families. And a recent study by If Free says that both the IGVGD, BRAC's and the RMP resulted in longer term improvement in the income of participants which lasted 18 months for the IGVGD and 25 months for RMP. They also found, and this is interesting, that the two public works programmes they evaluated, road maintenance and food for assets led to far more positive outcomes in terms of women's empowerment and that we're talking about decision making, mobility and so on than the IGVGD which was promoting enterprise and self-employment. Looking at the development potential of TUP it has also benefited from randomized control trials built into the programme from the outset. So anybody who's anybody I think in the RCT field have come in to evaluate the TUP so we've got people and the pilots. So we're finding very positive outcomes, the first round of studies three or four years after the programme was put in place and the programme is supposed to graduate women out of poverty in two years through this combination of supports. Found higher incomes and greater livelihood asset ownership. The second round found that, which was actually carried out by the IGC at LSE, found that women were moving out of poorly paid wage employment into more regular forms of self-employment and that total earnings rose by 38% two years after they had graduated out of the programme and savings were also increasing. There was some evidence of a rise in wages for unskilled wage labour, female wage labour. However a much more recent study finds that the impact of women's ownership and role in decision making does not extend beyond the assets transferred by the programme. All other assets and all other forms of decision making are under male control. And it appears in fact to have had a negative impact on women's control over income savings and expenditure decisions. Although the qualitative data on TUP shows something more positive and that is women appreciate having work within the home because of the hostility of the external environment to women. In terms of the questions that we were asked, I think the fact that we have public works programmes across the world more or less successful suggests that they are scalable and transferable. Provided of course the administrative and logistical requirements are met but the issue is how transferable and scalable is the employment guarantee and that really is a very different kettle of fish. The move from the Maharashtra scheme to the national scheme is an example of both transferring to other states in India and scaling up but it has met as we know with problems and very uneven performance across the states. An employment guarantee needs a responsive state and active citizens and these do not always exist. But I think the question we might ask is could it act as a catalyst to building more responsive bureaucrats and active citizens. And I say this because the Maharashtra employment guarantee scheme has many examples of wage workers becoming more active, of civil society organising poor people to demand wages, of putting pressure on local officials and so on. So going back to Corshik's talk yesterday about the focal point, which I don't like that word very much, it means something else to me. But anyway this business of creating a legal obligation can it generate the kind of impacts or the kind of outcomes that we want to see in terms of states and civil society. I think there is a very interesting comparison that McMore makes between the Maharashtra employment guarantee scheme and the employment assurance scheme in India which is just like the Maharashtra one but without the guarantee. And you find no evidence of mobilisation around the employment assurance scheme but you did find a great deal of mobilisation around the Maharashtra one. And you do find it around the NRAGS. Transformity scalability, we're looking into TUP, there are pilots being carried out, ten pilots in eight countries across the world. And I said, as I said, there's Esther Duflo, Brigitte Balergy, Dean Carlin, all the people who have been leading on the RCT field are involved in these pilots, in evaluating these pilots. And the results are coming in, they can help us a little bit with the transferability question, not scalability because these pilots are very small. But as a recent conference in Paris reported on the BRAC website, four out of the five RCTs showed increased food security, increased as well as more diverse incomes, increased assets, all in all, we are setting the stage for further explorations of scalability. However, some caution is needed. The four out of five were positive, the one that wasn't was in Andra. And it found that the income that women made through being involved in the TUP programme was offset by the income they lost through withdrawing from wage employment. And this of course is at a time when wages for agricultural labourers were rising partly because of the NRAGS and women's wages were rising. So, as Modoc says, you really do have to look at the larger opportunity structure in order to decide whether providing women with these opportunities is the most feasible way of helping them. And certainly in Sydney, where we did an evaluation of TUP, we found that there was no wage labour for women, but also that women's earning capacity and mobility capacity were so restricted that all they did was basket wheeling and embroidery. In that context, it was women who had husbands who could make use of TUP support that were able to benefit from the programme. And then a very recent study of the Bangladesh TUP finds that six years, so by the era the LSE studies were four years after, this is six years, that 54% of TUP households had experienced a decline in asset stock and income. And of course the SINTH and other studies from I think Haiti tell us that those who have some kind of initial advantage are much more likely to benefit. So there is uneven impact even within, across countries and amongst participants. And then my final point I think would be that some aspects of TUP are easier to transfer than others. The intensive mentoring, how easy is that to transfer? The fact that the West Bengal study that we did was carried out by an organisation that was very responsive to the needs of the poor, had worked with extremely poor women and self-help groups for a long time, made them manage and negotiate the TUP programme much better than the SINTH one where it was being managed by a microfinance programme whose only experience was with male entrepreneurs. So to that extent I think that the skills and the human element of the programme must be taught through before you think of transferring. So my conclusion then would be, my reading of this literature echoes a debate that took place in the Indian literature around the time of the integrated rural development project etc. And that is to say at that time credit interventions seemed to work best for those who had the complementary resources you need to make use of credit. Those who were asset less were most likely to benefit from wage employment. Today we're looking at self-employment enterprise development for extremely poor women. And the literature on enterprise is always looking for the entrepreneurial personality, the person who has the traits to be an entrepreneur. And the key trait is the ability to take risks. And I don't see how you can expect extremely poor women to take the kind of risks that are necessary to make success of their enterprises without the kind of support that say a self-help group could do. So the other thing of course that troubles me is the empowerment studies. And this is very much in line with the study that we've done at IDS, which found that in Ghana, Bangladesh, Egypt, where the three countries we did a survey, that regular wage employment and employment outside the house had far stronger association with empowerment indicators than work within the home, particularly unpaid work with the home. That's not dramatically surprising, but it seems to be mirrored now in the empowerment evaluations of wage employment in Andhra Pradesh and in Bangladesh and self-employment for very poor women. Thank you.