David McKenzie's blog
· If a program is “obviously working”, does it need an impact evaluation? Markus Olapade reflects on this on the 3ie blog.
Many important policies are implemented at the national level. Monetary policy, fiscal policy, and many regulations are key examples. Pure time series or before-after analysis of the impacts of changes in these policies on the economy of a country will be contaminated by other changes going on in the economy. Simply trying to difference out global trends will not account for systematic differences in the growth path of the country where the reform took place from the average global growth path. This makes evaluation of the impacts of such policies difficult.
· A new From Evidence to Policy note looks at the impact of a community grant program in Indonesia which gave grants to communities for health and education services. The program lowered malnutrition, and finds performance-based incentives lead to improved performance.
· On the FAI blog, Jonathan Morduch discusses what’s next in microfinance, in terms of how experiments can move towards allowing greater external validity.
On the World Bank’s today page today I saw the following:
This seemed really high to me, and a strange way of presenting statistics. Following the link, it directs you to this World Bank Data Viz Tumblir which has a bunch of statistics all presented in the form, if the World had only 100 people, then…
Since I’ve had three emails in one week asking me about this issue, I figured I might as well blog about it and have something to refer people to instead. The questions have all been variants of:
· Are women better remitters than men?
· Does having mothers migrate result in worse outcomes for kids than having their fathers migrate?
· Early results from a skills training program for young women in Liberia show massive increases in employment and earnings – although a randomized pipeline design whereby the control group get the treatment about one year after the treatment group raise concerns about strategic delay by the control group and how long-term impacts could be measured.
Many people are aware of the concept of a “placebo effect” in medicine and of the idea of a Hawthorne effect – in both cases the concern is that merely being treated can cause the treatment group in an intervention to change their behavior, regardless of what the treatment actually is.