 1994 was obviously the crucial turning point in South Africa and the Constitution had rewritten what the welfare task was going to be. I had been very involved in pension reform for elderly people and for people with disabilities. There had been a maintenance grant for children, but it had gone largely not to African people in rural areas and the Constitution was clearly insisting on a massive sway of services and delivery towards the poorest people who are by and large African people in South Africa. So on the one hand it was about shifting services so that they could reach many, many more people, albeit at probably a lower rate. And that was the second thing that constrained the principles was that the new more conservative macroeconomic and economic policy came down the tubes towards us very fast, much faster than anyone could have expected. And so you were having to spread services broader but with a fiscal constraint that was being tightened all the time with the new economic policy. What was helpful was that Nelson Mandela had his first call for children as it was called. He was completely committed to the cause of child development, the nurturing of children, special places in society for children. And so the fact that we were doing in a sense a near universal reform towards poorer children got a lot of backing from with inside cabinet who were of course former president Mandela was revered and I think his authentic love for children was something that was a basis of a new kind of solidarity in South Africa. We already knew that the pension for elderly people and for people with disabilities was incredibly important in strengthening women's financial position within the households that they were in and that where pensions went to women they had a large impact on the household because so many poorer people live in extended families and we knew even by 1994 that pensions that went to grannies in households rather than grandfathers in households was spent better on good household goods. The child support grant then came in for the first time going to many, many more poorer families although the previous grant I must insist had been going largely to so called colored and Indian families who were poor and some of them were going to lose the grant. The evidence came in quickly that the child support grant was making a difference specifically on child's enrollment in schools and on nutritional status of children. I don't think we were sure early on that the general impact of the child support grant on women was so marked partly because the child support grant was so small relative to the pension for elderly people and for people with disabilities. What we are increasingly seeing is that we have evidence that the pension for elderly grannies enables younger women with young children to leave the rural home and go to town probably largely to seek work. We don't know convincingly yet how the child support grant is impacting on women. What we do know is that the so many researchers from inside the country and outside come in trying to prove that the child support grant causes younger teenagers to get pregnant or causes decreases in agricultural production. There is absolutely no evidence whatsoever for this despite the fact that some very, very, very clever people have been trying to find exactly that evidence. I think with the child support grant that is a particularly important question. The child support grant was never intended to be a standalone thing that would solve monetary problems in the family or which would have huge effects on child nutrition. It was meant to be one of a set of components of support for children and families which should work synergistically together. I think it is one of the most intractable problems of development that in fact you can make your policies as interactive as you like. But in fact policies are delivered on the ground in segments with no necessarily well coordinated overall plan and I think this is the big problem. The cash transfers undoubtedly work positively. The future trajectory for the better quality of life of children and pathways out of poverty depends critically on other things such as in South Africa the education system. We know that the child support grant gets children into school earlier and it makes them stay for longer and that's particularly the case with regard to girl children. We don't know what the outcome of that is. Education has probably been the most botched part of policy reform in South Africa and until that gets right you can put the money in it first but you're not going to get the outcomes that would come from radically improved schooling which need not cost a lot. It's got to be a change in mindset. It's got to be a more coordinated set of thoughts coming from the government departments about how the different bits, grades and education interlink with each other and how to teach them within the limits of the teacher workforce in South Africa.