#770 2003 · Seva Mandir (with Esther Duflo, Rema Hanna, Stephen Ryan) · Education / NGO service delivery

An Indian NGO stopped asking teachers to report their own attendance and let a camera do it

问题

Rural one-teacher schools had no reliable way to know whether the teacher actually showed up

背景

Seva Mandir ran roughly 100 informal one-teacher schools for children in rural Udaipur, Rajasthan, who fell outside the government system. Each school had exactly one adult present, so there was no colleague, principal or parent on site to vouch for whether the teacher had come to work; unannounced inspector visits were rare, predictable, and easy to schedule around. Teachers were paid a flat monthly salary regardless of days worked, and self-reported attendance registers were trivial to falsify. The result was a measured absence rate near 44 percent — teachers were missing school nearly every other day, and there was no way to tell an honest absence from a fabricated one.

The obvious fixes were the ones every NGO and ministry already tried: send more inspectors, moralize about duty to students, or dock pay for absences reported by the teacher's own register. All three depended on the same broken link — verifying, at low cost, that a person was physically present in a remote village on a given morning — and none of them fixed it. More inspectors meant more cost for a marginal gain, since visits were still infrequent and predictable enough to dodge; docking self-reported pay just gave teachers a reason to under-report absences rather than reduce them.

换别人会怎么做

Send more inspectors, or moralize to teachers about their duty to the children -- the two things every NGO and ministry already tries. Both fail for the same reason: they don't change the fact that verifying a remote worker's presence is expensive, so inspections stay rare enough to dodge and self-reported registers stay easy to falsify.

他们看到了什么

The problem was never teacher motivation -- it was that nobody could cheaply prove where the teacher was. The children sitting in the classroom every day were already an unbypassable, self-generating witness; a tamper-proof camera just turned their daily presence into a verifiable record instead of an unusable one.

那一手

Researchers working with Seva Mandir gave half of a randomly assigned set of teachers a cheap camera with a tamper-proof date-and-time stamp, and had a student photograph the teacher with the class at the literal start and end of each school day. Pay was restructured from a flat monthly salary to a base plus a per-verified-day rate, calculated directly off the timestamped photos rather than any register the teacher filled in.

为什么管用

A flat salary paid regardless of verified attendance gives a teacher no cost to skipping school, so absence is limited only by conscience and the rare inspection. Tying pay directly to a timestamped photo makes every day's presence self-certifying and cheap to check, so the incentive to attend no longer depends on catching the absence after the fact -- it is built into the payment itself. Because the verification cost collapses to almost nothing, the program can afford to check every single day instead of sampling a few, which is what actually closes the gap between the true absence rate and the punished one.

值了多少

Teacher absence in camera schools fell from about 44 percent to roughly 21-22 percent, close to half the comparison group's rate, and students gained about 30 additional instructional days a year. A year in, test scores in camera schools were 0.17 standard deviations higher than in comparison schools, and children were substantially more likely to be admitted into regular government schools afterward. The program cost about $120 per teacher a year, worth roughly $0.11 per extra day of schooling delivered per child.

什么时候会失灵

The mechanism depends on the proof staying harder to fake than the job is to do -- once local staff or supervisors can bypass, disable, or collude around the verification step (as happened when similar monitoring was extended to government health-worker attendance in India), the incentive collapses back to the self-reporting problem it was built to fix. It also needs a genuinely low-cost, tamper-resistant proof to exist for the behavior in question; jobs whose output can't be reduced to a timestamped, unfakeable record don't get this fix for free.

后来呢

The study, published as "Incentives Work: Getting Teachers to Come to School" (American Economic Review, 2012), became one of development economics' most cited field experiments and helped establish tamper-proof, technology-verified attendance as a credible alternative to inspection regimes across public-service delivery research. It is not a universal fix: later attempts to extend similar photo- or biometric-linked monitoring to health-worker attendance in Indian government clinics saw local staff and supervisors find ways around the verification step itself, a reminder that the mechanism only holds while the proof stays harder to fake than the job is to skip.

资料来源

[1]Encouraging Teacher Attendance through Monitoring with Cameras in Rural Udaipur, IndiaAbdul Latif Jameel Poverty Action Lab (J-PAL), MIT, 2012povertyactionlab.org
[2]Monitoring Works: Getting Teachers to Come to School (NBER Working Paper 11880)National Bureau of Economic Research, 2005nber.org