Paper
Banerjee, Abhijit V., Esther Duflo, and Rachel Glennerster (2008). Journal of the European Economic Association, 6(2–3): 487–500
Read the original paper → Opens at the publisher; use the DU library / JSTOR login if it asks for access.
In one paragraph
India’s public health system suffers from high staff absence and low effort, so patients go private. The paper reports an experiment run with the Udaipur district administration and the NGO Seva Mandir: nurse attendance at government sub-centres and aid-posts was recorded, and the government agreed to penalise the worst absentees. At first the scheme worked — nurses responded to the financial incentives. Within a few months, though, the local health administration undermined it by letting nurses claim a growing number of “exempt days”, and eighteen months in the programme had no effect at all.
Setting and question
Rural Rajasthan, government sub-centres staffed by auxiliary nurse-midwives. The question is not only “do incentives work?” but “what happens to an incentive scheme once the people administering it have discretion?”
Data and identification
Randomised assignment of facilities to the monitoring-plus-incentives treatment, with attendance measured by the NGO’s own unannounced visits rather than by the administrative records the scheme itself generated — a design choice worth discussing in class, because the two sources diverge exactly when the scheme breaks down.
Main results
A large initial rise in attendance, then convergence back to control levels as exemptions proliferate; no measurable effect after eighteen months.
Why it is on the reading list
A compact example of an RCT whose most important finding is about implementation and political economy rather than the headline treatment effect. Read it alongside Björkman & Svensson (2009): same problem (provider effort), different mechanism (community pressure versus top-down incentives), different outcome.
No interactive explainer for this paper yet. If you build one, see Contribute to get it added.