Ideas for growth session 6: Human capital
Chaired by Imran Rasul (UCL; IGC), the Human capital session began with Orzaio Attanasio (UCL) presenting on a randomized early-childhood home-visiting intervention in Colombia. The study found that while provision of micronutrient supplements had no impact on measures of child development, a curriculum of stimulation designed to promote child development did have a significant impact on cognitive development. The results are suggestive that the mechanism of impact may be increased parental investment. It is also important to note that reliance in this intervention on local resources and existing infrastructure (including on women from the communities to conduct home visits) kept intervention costs low and contributed to community ownership of the program. Key for policy is evidence that if interventions improve human capital accumulation early in life this will not only raise human capital levels but also better enable individuals to exploit opportunities for further human capital accumulation later in life. Thus investments in human capital over the lifecycle are complementary and investments early on can have very high returns. However, when children accumulate early human capital deficits, these lags may be very difficult to reverse.
Tessa Bold (Goethe University Frankfurt; IGC) presented on experimental evidence of scaling up education reforms in Kenya. Specifically, Bold noted that while Randomized Control Trials from Western Kenya suggest that contract teachers may be a viable strategy for improving education outcomes, in a scaled up contract teacher intervention spanning the eight regions in Kenya, the Ministry of Education was unable to produce similar success. The intervention in fact compared the performance of student outcomes in schools where contract teachers were managed by an NGO with performance in schools where the Ministry of Education was charged with implementing the contract teacher program. While students in schools under the stead of the NGO did see an improvement in test scores, the impact of contract teachers administered by the Ministry of Education was essentially zero. These results are consistent with a line of criticism that questions the external validity of randomized control trials and suggests that large scale policy reforms may provoke hostile reactions from groups whose rents are threatened. A key implication is that institutions matter and that there needs to be more attention paid to experimental settings.
Finally, Karthik Muralidharan (UCSD) presented work on performance pay for teachers. Muralidharan discussed results from a 5 year long experimental evaluation of both group and individual teacher performance pay in a large representative sample of schools in the Indian state of Andhra Pradesh. The study found that students in schools under the individual teacher incentive program did significantly better than students in control schools throughout the duration of the program, and even in subjects for which there were no teacher incentives. These improvements in student outcomes were broad based and the intervention is cost effective when compared to reducing class size. However, when incentives were taken away, the effect in the year of discontinuation on student performance is basically zero, indicating a return to business as usual. The study suggests that compensation reforms have the potential to drive improvements in public sector productivity.