Many thanks to Michael Trucano for pointing out the important J-Pal study on cameras and absenteeism. Perhaps I am getting this wrong, but I thought that the program tied teachers' wages to their attendance, and that attendance was monitored through cameras. I then thought that the paper said that it was only tying the wages to the attendance that led to an increase in attendance, and not the use of cameras per se. That is, if you pay teachers only when they come to work, they tend to come to work more, but monitoring alone does not work. The issue here is not very different: given that service providers attendance is not sensitive to Hawthorne effects in the long-run, the political problem is always salient--does it make sense to move to a system where payment is contingent on presence, and is this politically feasible? Am I getting this completely wrong?