Teacher Attendance & Incentives 


Showing up is the first step

You can read about our methodology for estimating cost-effectiveness here

Impersonally administered and direct incentives for attendance are the most effective at increasing service provider attendance. However, in settings where supervisors were given discretion over administering incentives, these programs became ineffective. Attempts to increase accountability to the poor were successful in cases where the monitors had credible authority which the service providers respected, but untrained committees who had no actual power over service providers had little effect on attendance.

On an average day, 27 percent of teachers in Uganda do not come to class, and primary health care centers in Indonesia and India experience 40 percent absence rates (Chaudhury, et al., 2006). High absence rates are not driven by a small proportion of permanently absent service providers, they are widespread and unpredictable, which makes it hard for the poor to simply plan around them.

If a teacher is absent, children often lose days at school, particularly in rural areas with single-teacher schools. Thus for adolescent girls enrolled in school in rural Bangladesh, the most common reason reported for missing school in the previous week was teacher absenteeism (Field and Glennerster, 2008). When NGO Seva Mandir increased teacher attendance at their single-teacher schools in India, this led to more teaching, more days in schools and higher test scores: the same number of students were present on days where the school was open, but since it was open more often, students were taught about 2.7 more days each month 3 .

In the case of health services, lack of predictability is the main problem. If a nurse is not there one day a patient may simply come back the next to receive treatment. If they come for an immunization, waiting a day may not have major health effects, but if there is a high probability that the clinic will be closed, patients may not even bother trying to get services from a public clinic.

There is mixed evidence on whether people change their behavior in response to a higher probability of provider attendance, by using public facilities more often or getting better healthcare. A temporary reduction in nurse absenteeism at clinics in India did not lead to more patients attending these clinics 9. However, when nurses started to come to work more frequently in response to community monitoring in Uganda, the usage rate of the facility increased 4. Similar positive results were seen in patient attendance in Cambodia when the government outsourced health care provisions to non-governmental organizations. Of course, since many other changes took place in these facilities, the increase in patient presence may not be only due to the decrease in absence. 

High rates of provider absence are enabled by administration systems which offer no real incentives for attending work, or providing services while at work. A study in India found that 41 percent of healthcare providers engage in private practices outside of their civil service jobs (Chaudhury, et al., 2006). Similar observations have been made on the prevalence of private tutoring amongst teachers. Firing teachers is almost unheard of and there are very few consequences for poor performance. In a survey of 3,000 Indian government schools, only one principal reported a teacher ever having been fired for poor attendance. Widespread absenteeism suggests supervisors are already unable or unwilling to combat absenteeism. Consistently across J-PAL studies, supervisors undermined the effectiveness of incentive programs, and tended to excuse any and all absences and reward providers whether or not they turned up. It may be possible to circumvent the perverse effects of supervisor discretion by mobilizing and empowering the local community to monitor service providers, but such programs are hindered by coordination and free-rider problems.

As these studies have shown, the institutions that deliver public service are complex and the details matter. If you are interested in knowing more about our research on service provider absenteeism, click here to download our bulletin.

Read more

Cost-Effectiveness of Interventions to Improve Attendance

Additional Days of Attendance per $100 Spent

You can read about our methodology for estimating cost-effectiveness here
Evaluations Cost Effectiveness Academic Papers
additional provider days per $100 spent
Community Monitoring
1 Contract vs. Civil Service Teachers in Kenya

Hiring extra contract teachers actually saves money relative to the use of civil service teachers.

infinite Duflo, Dupas & Kremer, 2011Duflo et al., 2015 Academic green
4 Community Monitoring in Uganda

Communities and healthcare providers created a specific action plan to improve service, whose progress was monitored by a local committee.

4 days Björkman Nyqvist et al., 2014Bjorkman & Svensson, 2009Björkman & Svensson, 2010 Academic green
5 Community Monitoring in Madagascar

Local school committees had no effect on teacher absenteeism in Madagascar.

(no impact) Lassibille et al., 2010 Academic green
6 Community Information in India

Informing villagers of poor teaching in their villages and raising awareness of accountability mechanisms had no impact on teacher attendance.

(no impact) Banerjee et al., 2010 Academic green
2 Girls' Merit Scholarships in Kenya

High-scoring girls in Kenya were offered a merit scholarship, motivating teachers to attend school.

50 days Friedman et. al., 2011 Academic blue
3 Incentives & Camera Monitoring in India

Teachers paid based on attendance and monitored with cameras attended school more often.

45 days Duflo, Hanna, & Ryan, 2012 Academic blue
7 Incentives & Supervision in Kenya

Supervisor discretion undermined attendance incentives for teachers.

(no impact) blue
8 Incentives for Test Scores in Kenya

Teacher pay linked to student test scores did not improve teacher attendance.

(no impact) Glewwe, Ilias & Kremer, 2010 Academic blue
9 Attendance Incentives for Nurses in India

Supervisor discretion undermined a time-clock program for nurses in rural India.

(no impact) Banerjee, Duflo & Glennerster 2008 Academic blue

One can build schools and clinics and stock them with books, drugs, and equipment, but if the teachers, nurses, and other providers are chronically absent, these investments will be wasted. Motivating better service delivery is complex and not all the programs rigorously tested by J-PAL researchers worked.

Impersonally administered and direct incentives for attendance are the most effective at increasing service provider attendance. When teachers in India were provided with cameras, and their salaries linked to showing daily photos with the pupils as proof of presence, teacher absenteeism halved and test scores went up 3.  However, in settings when supervisors were given discretion over administering incentives, these programs became entirely ineffective. In Kenya, school principals rewarded providers whether they showed up or not, with no improvements in provider attendance 7. In India, the doctors responsible for monitoring nurses colluded with them to grant them exemptions when they were absent, and the monitoring program eventually had the perverse effect of increasing absence 9.  

Intrinsic motivation matters, too. Student incentives for good test performance in the Girls’ Scholarship program in Kenya motivated students to study harder 2. When students and families became more motivated, teachers came to school more often, and though only girls were eligible to win the scholarship, boys and academically weaker students also benefited. However, when teachers’ incentives were made conditional on student test scores in India and Kenya, absenteeism did not change 8. Test scores rose, but mainly due to test preparation.

Attempts to increase accountability to the poor had varying results depending on the context. In India, local school committees had no impact on teacher attendance 6, even when they were given information about the quality of education in their area. However, a program in Uganda which created “report cards” for the health or education services in the area and then mobilized the community with specific “action plans” in coordination with an NGO was effective at increasing the attendance of teachers and health care providers 4. These programs need to have credible authority which service providers will respect. A program in Kenya gave community education committees funds to hire an extra teacher whom they had the power of hiring and firing, and these locally accountable teachers had significantly lower absence rates than the government civil service teachers 1. Moreover, the effects of the program were even stronger when school committees were trained to monitor the teachers.

Of the programs that worked, most interventions were surprisingly cheap. Monitoring with cameras in single teacher schools in India 3 and the Girls’ scholarship program in Kenya 2 both cost about $2.00 per additional day of teacher attendance. The Citizen Report Card in Uganda 4 had broader objectives than improving service provider attendance and achieved remarkable success, including significantly higher vaccination rates and a 33 percent reduction in under-5 mortality. If the program is considered solely as a means to improve attendance and the total cost is attributed to this goal, it achieved better service provider attendance at a cost of $25 per worker per additional day.
Contract teachers in Kenya 1 were paid a fraction of the cost of the civil service teachers, and when they were monitored by trained school committees, they ended up saving the program money compared to the cost of buying the same amount of teaching time with additional civil service teachers.  

Community monitoring where the community could not hire and fire service providers at will, such as the VECs in India 6 had no impact on teacher attendance. Incentives which were not strictly conditional, including the teacher bicycle program in Kenya 7, were also a waste of resources.  

Return to top