Lant Pritchett once said to me “Thanks for the comments. As usual they are all very smart and well-informed and I disagree with most.” I feel similarly regarding his very popular piece  posted here last week (already one of the top 10 most popular posts in our blog's short history) on how CCTs are forcing children in developing countries into terrible schools. So, here goes a reply…
I read your short piece on Uttar Pradesh, which does paint a dismal picture. I don't think anyone can argue with the point that making safety nets conditional on services (public or private) that don't exist, are too far, or are dismal (or, even worse, dangerous) makes no sense. Perhaps others more familiar with supply side problems will chime in, but I offer a few points for rebuttal.
1. In our Malawi experiment, our worry was very much like yours – i.e. because of low school quality, the marginal girls induced to go to school by the conditionality would not learn anything. However, we were proven wrong: we saw significant improvements in math, language, and cognitive skills in the CCT group. No similar improvements were detected for UCTs (because they did not attend school as much). Furthermore, what is really interesting is that we saw similar size improvements among girls who had already dropped out of school before the program came along. Deon Filmer and Norbert Schady have a paper  that suggests that such marginal children brought back to school are the reason why we don't see any learning effects under CCT programs. If, even these girls, who might have dropped out of school "rationally" due to low ability, did improve these skills, then the schools cannot be completely useless. At least, not in Malawi in any case...
2. You focus on prevalence of the students’ knowledge in your piece, but the question may be its incidence. Indeed, had I written your piece about Zomba, Malawi, I might have written the same thing: achievement levels in Malawi, apparently among the worst of SACMEQ countries, are dismal – especially among girls. But, the question is whether the increased school attendance, grade attainment, and test scores are worth more than the returns to the time spent in the counterfactual activity to schooling – even if the final level of learning is still very poor by international standards. We're going back to the field to examine this question this year.
3. Are you also not underestimating the benefits from sitting your butt in school even if formal learning is minimal? Schooling increases non-cognitive skills as well, leading to improvements in mental well-being, reduction in risky behaviors, etc.
4. I think that some of my colleagues would argue that the supply side is not fixed and might get better as a result of changes on the demand side. In fact, this paper  by Maluccio et al. (published in the Journal of Development Effectiveness in 2010) reports that Nicaragua’s CCT was not only more effective in supply-constrained areas but also improved school supply “as measured by grade availability, number of sessions per day and number of teachers.” Similar changes could occur to quality as well. Many countries increasingly focus on both sides of the coin...
5. The language of "threatening the mother with conditions” brings more readers to our blog, but is perhaps slightly unfair language. I see both sides of this argument: on the one hand, in my recent paper  with Baird and McIntosh, we argue that many people, who don't comply with the condition and hence do not benefit from CCTs, live in vulnerable households who deserve the support of the government. By foregoing such support, CCTs significantly undercut the social protection dimension of these programs, and so they should only be deployed after careful consideration. However, as we also point out in the same work, whether in Malawi or Mexico, many households (many more than the marginal ones who comply) do not comply with the condition, i.e. don’t respond to the “threat.” So, maybe (just maybe) people who won't benefit from additional schooling become non-compliers under a CCT offer while the marginal children/households who stand to benefit take it up.
6. Finally, a question regarding the issue of beating and abuse at schools in Uttar Pradesh. Is it a schools problem, or do these children suffer a similar plight at home and elsewhere, i.e. the factory, the fields, or public life? Is the 21st century equivalent of “intense heat and turpentine” really better for these children? Can you really get a good answer by asking them their choice? I agree that CCTs are likely a second-best or worse solution to whatever market failure exists, but, as you argued earlier, they may be the politically feasible one.
It’s good to attract people’s attention to the fact that supply side issues are as important in many settings when it comes to education. It’s also nice to point out that some (or perhaps many) may be opting out of schools “rationally.” But, as one of the readers who commented  on your piece said last week: “CCTs are no panacea. But they are no villain either.”
Let’s do this again sometime. Cheers,