Teacher Pay and Educational Outcomes: Evidence from the Rural Hardship Allowance in Zambia

• Date: 01 September 2019
• Published date: 01 September 2019
• DOI: http://doi.org/10.1111/saje.12227
• Author: Grieve Chelwa,Mashekwa Maboshe,Miquel Pellicer

© 2019 Economic Society of Sout h Africa.

South African Journal of Economics Vol. 87:3 September 2019

doi : 10.1111/ saje .122 27

255

TEACHER PAY AND EDUCATIONAL OUTCOMES:

EVIDENCE FROM THE RURAL HARDSHIP ALLOWANCE IN

ZA MBIA

GRI EV E CHE LWA*, MIQUEL PELL ICER†,‡ AND M ASHEKWA MA BOSHE‡,§

Abstract

This paper studies the effect of unconditional teacher salary increases on teacher and student

outcomes. To study the issue, we evaluate the rural hardship allowance in Zambia, which

corresponds to a salary increase of 20%. This allowance is allocated to schools on the basis of a

distance criterion allowing us to use a regression discontinuity design. We use administrative data

from 2004 to 2015 on school, teacher characteristics and test scores. The administrative data are

complemented with a telephone survey of schools close to the eligibility threshold. We find that

crossing the threshold increases the share of teachers obtaining the allowance by 40%. Because

of some non-compliance with the allocation rule, our estimates are fairly imprecise. Focusing on

provinces with better compliance we find some, albeit weak, evidence that the allowance increases

the stock of teachers. We, however, find no effects on teacher characteristics or on student test

scores.

JEL Classification: I2

Keywords: Educat ion, policy evaluation, regression dis continuity, Zambia

1. INTRODUC TION

There has been an explosion of research in recent years on the factors that can im-

prove student learning in developing countries (see recent reviews in Glewwe et al.,

2011; Snilstveit et al., 2016). Teachers are one type of input that has received attention.

Evidence from high-income countries shows that teachers can have a larg e role in student

learni ng and long-term outcomes (Chet ty et al., 2018, 2014; Jackson et al., 2014), and re-

cent evidence from Pakistan sug gests a high impact of teachers on grades in a developing

country setting (Bau and Das, 2017).

There have been several studies eva luating interventions aimed at increasing the pro-

ductivity of teachers in developing countries. Most of these interventions have explicitly

provided “hard” incentives to af fect specific ty pes of teacher behaviours. For instance,

* Corresponding author: Senior L ecturer, Graduate School of Business, University of Cape

Town, Cape Town, South Africa a nd Southern African Institute for Policy and Re search

(SAIPAR), Lusaka, Zambia. E-mail: grievechelwa@gmail.com

† University of Duisburg-Essen, Dui sburg, Germany

‡ Southern Africa L abour and Development Research Unit (SALDRU), University of Cape

Tow n

§ Public and Environmental Ec onomics Resea rch Centre (PEERC), University of Johannesburg

South African Journal

of Economics

256 South African Journal of Economics Vol. 87:3 September 2019

© 2019 Economic Society of Sout h Africa.

teacher presence has been incentivised by monitoring or providing financial incentives

conditional on presence (Duflo et al ., 2012), and teacher contribution to student per-

formance has been incentivised by providing f inancial rewa rds to teachers conditional

on positive student test scores (Kremer et al., 2010; Muralidharan and Sundararaman,

2011). While some of these interventions have proven successful, a recent review com-

paring the effect of different types of interventions on student learning concludes that

the effects of teacher incentive interventions have been generally sma ll (Snilstveit et al.,

2016). Moreover, some studies find that such interventions have generated undesirable

outcomes such as, for example, teaching-to-the-test (Kremer et al ., 2010).

There has been growing interest in the potential role of “soft” approaches to increase

teacher productivity that tap into behavioural re sponses such as reciprocity or intrinsic

motivation (Besley and Ghatak, 2014). In particular, higher unconditional wages may

improve the productivity of public servants via selection of more motivated teachers (Bo

et al., 2013), or via reciprocity.

This paper studies the ef fect of unconditional increases in salaries on teacher and student

outcomes. We do so by evaluating the effects of the rural hardship allowance in Za mbia,

which provides a 20% increase in salaries paid to teachers in rural schools. The rural hard-

ship allowance is al located to schools outside a given radius from district centres, and this al-

lows us to estimate the effect s of salaries using a regression discontinuity approach. We study

the effect of the allowance on teacher attrition, teacher characteristics and student grades.

Several recent papers study the effec t of unconditional wage increase s on teacher and stu-

dent outcomes in developing countries: in Brazil, Uruguay, Pakistan and Indonesia (Filho

and Pinto, 2014; Bau and Das, 2017; Cabrera and Webbink, 2018; de Ree et al., 2018). All

these studies tend to find either no effec t or a small effect on student performance. However,

more evidence is necessary to bring to bear on the question. The studies on Brazil and

Pakistan (Filho and Pinto, 2014; Bau and Das, 2017, respectively) employ either a simple

difference or a differences-in-differences estimator which may not be able to fully account

for confounding factors. The studies on Uruguay a nd Indonesia (Cabrera and Webbink,

2018; de Ree et al., 2018, respectively) address this problem using experimental and qua-

si-experimental methods. However, the setting of both studie s are middle-income countries.

Unconditional salary increase s may well have different effects in low-income countries where

teacher salaries are ver y low and unconditional salar y expansions may be most relevant.

Our paper contributes to this nascent literature using a qua si-experimental approach

to study the effect of unconditional sala ry increments in a low-income setting. Our eval-

uation of the rural hards hip allowance not only concerns a relatively low income country

such as Zambia but also schools in rural areas with little infrastructure and amenities.

The increase in pay provided by the allowance is 20%, subst antially lower than that in

de Ree et al. (2018). Nevertheless, in a challenging environment such as rural Zambia,

an increase of this magnitude can be important.

Our paper also contributes to the literature on improving state capacity in developing

countries. And here we concern ourselves with a smal l literature on whether pecuniary

incentives can be effect ive in attracting a nd retaining public servants working under dif-

ficult conditions. Bo et al. (2013) study this issue in Mexico where it is diff icult to attract

workers to work in remote or challenging municipalities. Exploiting exogenous variation

in wages generated by a field experiment, they find that prospective workers are more

likely to accept job offers in challenging municipalities if offered higher wages. Antwi