Teachers determine the quality of student learning; however, teacher effectiveness is difficult to predict and does not necessarily correlate to wages
Personal experience and extensive research both tell us that a good teacher can change a student’s life. Research from the US shows that good teachers significantly increase test scores (Chetty et al. 2014), but not all teachers are good and predicting teacher quality is difficult. There is a lack of data on teacher effectiveness in low-income countries, where student learning is poor (Pritchett 2013) and teacher salaries comprise 80% of recurring educational expenditures (UNESCO Islamabad 2013). Given the importance of teachers for learning, this represents a serious gap in the education literature.
Filling this gap is critical, since questions of how to recruit and retain high quality teachers are the subject of a growing policy debate. One prominent view is that higher teacher salaries are needed to attract high ability teachers to the profession. For example, a McKinsey report on the problem of low teacher quality argues, “[m]oney was the most powerful lever in attracting and retaining top one-third students” (Auguste et al. 2010). Determining optimal salaries requires us to both understand labour-market returns in relation to the teacher characteristics that predict student learning, and whether higher teacher salaries would induce individuals with these characteristics to enter the teaching profession in low-income countries.
The study: Primary school students and their teachers in Punjab
In a recent paper, we use a unique dataset that follows primary school students and their teachers in both public and private schools from 2004-2007 in Punjab, Pakistan to shed light on these issues (Bau and Das 2017). These data track students as they went through primary school, matching them to the teachers who taught them in each year. One novel aspect of this dataset is that both students and teachers in this data were administered primary school-level tests in Urdu (the vernacular), English, and math.
Our findings are remarkably like those from the US, with a few differences:
1. Teachers matter!
Using student test score data, we estimate a value-added (a proxy for teacher productivity) for each teacher. A value-added is an estimate of the test score gains (measured in test score standard deviations) a random student would receive if she was assigned to a given teacher (Chetty et al. 2014). These estimates indicate that moving a student from a teacher at the 5th to the 95th percentile would increase her test scores by 0.5 standard deviations (the equivalent of more than one extra year of school).
2. Teacher effectiveness is hard to predict
Using these measures of productivity, we find that only teacher content knowledge (as measured by teachers’ test scores on the primary school exams) and the first two years of teacher experience predict teacher productivity. In line with the findings in the US (Staiger and Rockoff 2010), having a bachelor’s degree and teacher training are not associated with greater teacher productivity.
3. Teacher effectiveness and wages reveal varying correlations within the public sector and private sectors
We then examine what teacher characteristics predict higher salaries in Pakistan in both the public and private sectors. Figure 1 graphs a selection of the coefficients. In both sectors, having a bachelor’s degree positively predicts wages. In the private sector, females – who have lower outside options – are paid less, but not in the public sector. However, in the private sector, teachers with higher value-added are paid more, while in the public sector, there is no relationship between teacher value-added and teacher wages. These findings suggest that teacher productivity is at least somewhat observable and that the private sector identifies and rewards more productive teachers with higher wages. In contrast, higher wages in the public sector are not associated with teacher productivity and are highly associated with characteristics that do not predict productivity.
Figure 1 Association between teacher characteristics and teacher pay
The lack of an association between teacher productivity and wages suggests that the public sector does not reward better teachers with higher wages. However, this does not mean that increasing teachers’ wages wouldn’t attract higher quality applicants to teaching positions. To measure the relationship between teacher wages and the quality of individuals entering the teaching profession, we estimate the effects of a hiring regime change in Punjab.
In 1998, Pakistan initiated unanticipated nuclear tests, which led to international sanctions and a worsening budgetary position for the province. As a result, teacher hiring was largely frozen for three years. When hiring resumed, almost all new teachers were hired as temporary contract teachers with 35% lower salaries. To see if lower teacher salaries attracted worse teachers, we compare the productivity of teachers hired right before the nuclear tests to those hired right after hiring resumed with the caveat that the new teachers’ contract status may have also increased accountability.
4. Moving to contract teachers with 35% lower salaries did not reduce the effectiveness of new hires
We never find a negative effect of being hired at a lower salary on teacher productivity, and when we account for the fact that teachers are observed at different experience levels, contract teachers performed better than the permanent teachers. This is consistent with findings in Duflo et al. (2011) and Muralidharan and Sundaraman (2013) for Kenya and India, respectively.
The fact that reducing teacher salaries by 35% did not negatively affect students’ outcomes initially seems surprising. However, comparing teacher salaries in the private sector and the public sector helps illustrate why this is the case (Figure 2). There is virtually no overlap between public and private sector salaries, with the average public sector teacher making approximately five times as much as the average private sector teacher.
Figure 2 Distribution of teacher salaries in the public and private sectors
This large public sector premium is not unique to Pakistan. Finan et al. (forthcoming) use data from 52 countries to show that public sector premia are larger in low-income countries relative to high-income countries. Altogether, these results suggest that calls to increase student learning by raising teacher wages, while keeping the same hiring and pay structure in place, are misguided. Redesigning compensation systems for government employees is complex enough as it is; one issue that policymakers need not worry about is that teacher salaries are insufficient to attract talented teachers to the pool.
Photo credit: Hashoo Foundation.
Andrabi, T, D Jishnu and A Khwaja (2008), “A dime a day: The possibilities and limits of private schooling in Pakistan”, Comparative Education Review 52(3): 329–355.
Auguste, B, P Kihn and M Miller (2010), Closing the Talent Gap: Attracting and Retaining Top One-Third Graduates to Careers in Teaching: An International and Market Research-Based Perspective, McKinsey & Company.
Bau, N and J Das (2017), “The Misallocation of Pay and Productivity in the Public Sector: Evidence from the Labor Market for Teachers”, World Bank Policy Research Working Paper No. 8050.
Chetty, R, J Friedman and J Rockoff (2014), “Measuring the Impacts of Teachers I: Evaluating Bias in Teacher Value-Added Estimates”, American Economic Review 104(9): 2593–2632.
Duflo, E, P Dupas and M Kremer (2011), “Peer effects, teacher incentives, and the impact of tracking: Evidence from a randomized evaluation in Kenya”, American Economic Review 101(5): 1739–74.
Finan, F, B Olken and R Pande (forthcoming), “The personnel economics of the state”, in Handbook of Field Experiments.
Muralidharan, K and V Sundararaman (2013), “Contract teachers: Experimental evidence from India”, NBER Working Paper No. 19440.
Pritchett, L (2013), The Rebirth of Education: Schooling Ain’t Learning, Brookings Institution Press.
Staiger, D and J Rockoff (2010), “Searching for effective teachers with imperfect information”, The Journal of Economic Perspectives 24(3): 97–117.
UNESCO Islamabad (2013), Education Budgets: A Study of Selected Districts of Pakistan.