Worlds of Education

Education International
Education International

A crisis of cost: Economic constraints to hiring more and better teachers

published 5 October 2012 updated 5 October 2012
written by:

My previous post (“Teacher supply gaps and the looming quality crisis in developing countries”) looked at how supply gaps in quality teachers and teacher training exacerbate the well-known global teacher shortfall in both primary and post-primary phases.  We are quickly finding out, unfortunately though not surprisingly, that there are no quick fixes to simultaneously expand the teacher workforce and upgrade their pedagogical skills in low- and lower-middle income countries.  Simply spending more money is neither feasible nor necessarily desirable in many contexts: the economic crisis of late 2008 left a fiscal hole in the budgets of low-income countries, many of which have no space to increase domestic spending and are burdened by critical levels of debt. [1] And despite recent promises to double aid to education (which is welcome news), the international track record on external finance and aid grants for education is not great. [2]

In short, there is a crisis of cost.  Many national education systems need to hire more teachers and provide better training opportunities for them, but cannot afford to do so within current financing models.  Investment in personnel, teacher professional development and meaningful in-service support is inherently subject to the ‘fiscal space’, not only within total government spending, but also within governmental education budgets.  There are at least four cost constraints at work here.

For Figure 1 (and 2) see attachments

Notes: Dark bars represent low-income countries (n=20); medium bars represent lower middle-income countries (n=19); light bars represent upper middle-income and high-income countries (n=69). Countries categorized according to World Bank classifications. Data are from 2011 or most recent available.

Constraint one: Personnel remuneration as the lion's share of education spending

In the vast majority of countries, salary expenditure constitutes more than 50 percent of total education spending, but can swell to more than 80 percent in some contexts (see Figure 1). [3] As seen in Figure 1, salary expenditure consumes the majority of education spending in many countries: salary costs comprise more than 75 percent of education budgets in nearly half of the sample (50 countries), more than 80 percent of spending in one-third of countries and more than 90 percent in 12 countries.  Salary allotments over 80 or 90 percent indicate marginal fiscal space within education budgets and a ‘crowding out’ of other items necessary to enhance the teaching workforce, such as professional development and school up-keep, particularly as school infrastructure in low-income countries (LICs) is also often weak and in need of improvement.  On the other hand, low allotments to salaries can also be problematic in that they do little to encourage high-quality candidates into the teaching profession in the first place. [4]

Constraint two: Increased unit costs of post-primary teachers

In low-income countries, secondary teachers tend to earn significantly higher wages than pre-primary or primary teachers.  This wage structure constricts non-salary expenditure where the expansion of the education system is predominantly in post-basic phases (Sub-Saharan Africa, South and West Asia and the Arab States).  Matching teacher supply to rising demand for secondary and further schooling in these contexts implies an even more inflated salary bill, further crowding out non-salary investments.  Figure 2 highlights this in 25 Sub-Saharan African countries and compares average earnings for primary and secondary teachers (as a percent of GDP per capita). A reference line indicates pay parity (primary and upper-secondary teachers receive the same wage).  In most countries shown here, secondary teachers earn much more than primary, particularly in Congo, Madagascar, Rwanda, Central African Republic, Benin and Lesotho.  Only Ghana and Liberia are close to pay parity while the Democratic Republic of the Congo actually pays primary educators more than upper secondary. [5]

From these data emerges another point: teacher salaries in many LICs are more expensive (relative to GDP per capita) than in most middle- and high-income countries: in general, high-income countries pay their teachers between 1.0 and 1.5 times GDP per capita. [6] This is not to suggest that educator wages should be pegged to country wealth, but rather that a high wage bill (relative to GDP per capita) represents an expensive and significant investment for LIC governments that are simultaneously attempting to expand access to primary and post-primary phases while enhancing the overall quality of the teaching workforce.

Constraint 3: The qualification-salary connection and the performance-salary disconnect

Teacher wages typically reflect the assumption that classroom effectiveness and instructor quality increase progressively with years of schooling, teaching certificates and experience. [7] There is evidence, however, that this assumption may be specious.  In fact, these common measures of teacher quality have been found to be poor predictors of the classroom effectiveness (though it should be noted that, in these studies, ‘effectiveness’ is often extremely narrowly defined as student test achievement). [8]

This carries substantial implications for both education budgets and the quality of the teaching workforce.  The United States spent approximately $24.4 billion per year (17 percent of instruction expenditure) in the early 2000s on seniority pay for teachers despite the lack of evidence that skills increase linearly with experience. [9] This is a lot to spend on an evidently blunt instrument for increasing learning outcomes.  Other human resource effects could stem from qualification-based salary structures.  First, they may cause educational inequalities by unwittingly discouraging the best teachers from plying their trade in more challenging settings (rural, low-achieving or low-income schools).  Second, pay that rewards high- and low-performing teachers equally may erode intrinsic motivation. Third, the absence of formal recognition (monetary or otherwise) of excellence in teaching will tend to discourage potential candidates who have the most to gain from meritocratic schemes from entering the teaching profession in the first place.

Constraint 4: The necessity to undertake structural salary reforms

The current policy debates often pit qualification- and performance-based salary structures for teachers against each other.  This is understandable given the heavily politicized nature of the debate, but it is ultimately a misnomer.  All salary structures, whether traditional qualifications-based or performance-based, are in actuality incentive schemes in that they incentivize and reward particular behaviors.

Qualification-based systems evidently value and reward formal teacher credentials and experience.  Performance-based schemes may also value these characteristics, but may additionally reward, for example, achievement gains in student test scores, completed professional development or school leadership roles.  So the question posed to any given remuneration structure is not whether the introduction of incentives for teachers might be an appropriate policy intervention to elicit certain behaviors, but rather whether the behaviors desired are effectively incentivized (and undesired negative side-effects limited) by the current structure.  If the answer is ‘yes’, then all is well.  If, however, the answer is no, something needs to change.

Unfortunately, the answer is generally ‘no’.  Teacher pay scales that focus only on qualifications and years of experience do not systematically engender desired learning outcomes among students or incentivize desired pedagogical and professional behavior among teachers. [10] As such, something needs to change.  National education systems literally cannot afford to assume that incremental and linear pay raises for their teacher workforces, allocated according to seniority and educational level, will inherently reinforce the desired behaviors present in teachers of ‘good quality’.  While education level, certification and experience should not be completely eschewed as indicators of teacher quality, it is clear that a much more robust definition of teacher quality is needed.


At this point, it may seem as though I have painted myself into a corner from which to advocate simplistic policy solutions to address the aforementioned challenges.  For example, capping teacher salaries, either in real terms or as a percentage of GDP per capita (as some have advocated) might at first glance free up revenue for reforms in other necessary, non-personnel areas.  But this also has negative implications for attracting the best candidates into the teaching profession to begin with and retaining them, and does not take into consideration the extremely low and largely unlivable wages that teachers earn in many low-income countries. Secondly, hiring contract teachers with lower qualifications may appear less expensive in the short term, but this strategy (also in favor in some circles) suffers from quality control problems, and may not be less expensive in the long term when considering personnel training needs and overhead costs.  It is also unpalatable for many to fathom a human resources strategy that centers on paying low wages and offering only precarious short-term contracts to candidates who are supposedly performing a high-skill function (i.e. teaching).  Finally, merit or performance pay for teachers might, in theory, circumvent some of the problems associated with basing salaries on qualifications and experience, but it does not have an attractive track record.  Many studies have found that merit pay induces teachers to game the system, teach to the test, or to stop collaborating with colleagues (in the case of individual incentives).  Moreover, most merit pay schemes are short-lived; as such, there is scant evidence that higher quality candidates would opt into teaching if pay was based on merit or performance.

In short, I am arguing that there are cost constraints that severely limit what can be done vis-à-vis expanding teacher numbers and upgrading their skills in many low-income contexts.  However, there are no panaceas and no silver bullets to fix these; we should greet assertions to the contrary with a healthy amount of skepticism.  Human resource policy interventions do not exist in a social vacuum, but rather have implications throughout the education system.  These implications and policy trade-offs should be openly assessed and debated by all education stakeholders, and finally evaluated by their consequences (positive or negative) on desired education outcomes, which of course means more than test scores.

[1] Kyrili, K., & Martin, M. 2010. The impact of the global economic crisis on the budget of low-income countries (Research Report). London: Oxfam.

[2] OECD-DAC. 2010. International Development Statistics: DAC Annual Aggregates. Paris: OECD Development Assistance Committee.

[3] UNESCO Institute for Statistics. 2011. Global Education Digest 2011: Comparing education statistics around the world. Montreal: UNESCO-UIS.

[4] Lazear, E. 2003. Teacher incentives. Swedish Economic Policy Review, 10, 197–213.

UNICEF. 2010. Protecting salaries of frontline teachers and health workers. Social and Economic Policy Working Briefs. New York: UNICEF.

[5] Primary teachers receive more than secondary teachers in the DRC as the latter phase relies heavily on user fees which are employed to supplement government funds.

[6] Korea is an outlier among OECD nations and pays teachers 2.11 times GDP per capita. See: OECD. 2011. Education at a Glance 2011. Paris: OECD.

[7] Hoxby, C.M. 2002. Would school choice change the teaching profession? Journal of Human Resources, 37, 846–891.

UNESCO. 2005. Education for All Global Monitoring Report 2005: The quality imperative. Paris: UNESCO.

[8] Rivkin, S., Hanushek, E., & Kain, J. F. (2005). Teachers, schools and academic achievement. Econometrica, 73, 417–458.

Boyd, D., Grossman, P., Lankford, H., & Loeb, S. (2006). How changes in entry requirements alter the teacher workforce and affect student achievement. Education Finance and Policy, 1, 176–216.

Podgursky, M., & Springer, M. 2007. Credentials versus performance: Review of the teacher performance pay research. Peabody Journal of Education, 82(4), 551–573.

[9] Ballou, D., & Podgursky, M. 1997. Teacher pay and teacher quality. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.

[10] Glewwe, P., Ilias, N., & Kremer, M. 2003. Teacher incentives (NBER Working Paper No. 9671). Cambridge, MA: National Bureau of Economic Research.

McEwan, P.J., & Santibáñez, L. 2005. Teacher and principal incentives in Mexico. In E. Vegas (Ed.), Incentives to Improve Teaching: Lessons from Latin America, Directions in Development (pp. 213–253). Washington, D.C.: The World Bank.

The opinions expressed in this blog are those of the author and do not necessarily reflect any official policies or positions of Education International.