Does money buy good teachers?

Why it's tough to find a clear link between teacher pay and student performance

Improving teacher quality can have a profound impact on educational and societal outcomes. Eric Hanushek of Stanford University calculates that if the United States replaced just the 5-8% least effective teachers in the country with merely average ones, the net present value of the economic benefits would total around $100 trillion.[1]

Clearly, investment in teacher quality could yield vast returns. One obvious potential direction for such spending is to increase teacher salaries as a way both to attract more qualified people to the profession and to motivate those who are currently in it. Evidence suggests, however, that this would not have a marked effect. A comparison of average remuneration at the country level – as a proportion of GDP per person – and national scores in the Global Index of Cognitive Skills and Educational Attainment (created for the Learning Curve by the Economist Intelligence Unit) indicates that salary levels and results are remarkably uncorrelated. As the accompanying chart shows, the four countries that pay their teachers the most relative to their economy’s average earnings all finish in the bottom third of the Index. Similarly, although second-place Korea has relatively high teacher salaries – 1.6 times the average gross wage – first-place Finland and fourth-place Japan pay slightly less than the average earnings across their economies.

Teacher pay vs educational performance, selected countries

Teacher pay vs educational performance, selected countries


Note: The Global Index scores are represented as z-scores. The process of normalising all values in the Index into z-scores enables a direct comparison of country performance across all the indicators. A z-score indicates how many standard deviations an observation is above or below the mean of all the countries in the Index.

Source: Economist Intelligence Unit and OECD.

These results confirm the existing conventional wisdom among education experts. Overall wage levels for the profession may have so little an impact on results, however, because payment systems are not designed to motivate better performance. In many countries, salary levels are based largely on seniority, which in turn comes from years spent teaching rather than any formal connection to student results. How might pay explicitly linked to educational outcomes change things?

The question of what is variously dubbed merit or performance-related pay is a highly contentious one. Teachers’ organisations frequently resist its introduction, arguing that measuring the impact of individual teachers is nearly impossible given the other influences on results, and invidious as it creates competition between educators who need to work as a team to provide the best possible outcomes. Advocates reply that, especially with modern data collection and analysis, it is possible to come up with fair estimates of value-added and different reward arrangements – including pay for small groups of teachers or entire schools that obtain good results. Moreover, they contend that rewarding the best is both fair and will lead to better overall results.

Both sides point to a variety of statistical studies to support their arguments. Experiments in a number of countries, in particular India and Israel, indicate that incentives can have a positive effect on student grades, but recent studies in the US have typically indicated that they yield few such results.[2] 

The problem with much of the research in the field is that it is focussed on the analysis of short-term initiatives that have specific performance measurement and reward characteristics. Generalising from them is problematic: for example, a recent study in Chicago found, unlike other American research, a highly marked improvement in student results when teachers were paid performance bonuses at the start of the year – and had to repay them if targets were not met – rather than receiving them retrospectively at year’s end.[3]

Such experiments of limited duration also cannot measure the impact of the long-term effect of performance rewards. Two relatively recent multi-country studies of OECD data shed some light on the latter and indicate that performance related-pay can have positive effects in at least certain cases.

A 2012 edition of OECD’s PISA in Focus magazine compared the results of national education systems which had some form of bonus or promotion preferment for teachers deemed to have performed better with the outcomes of systems that did not. On the surface, the results seemed to show that there was no link, but a closer examination found that for countries where average teacher earnings were below 115% of the average gross wage, performance rewards had a positive effect. For countries where the average wage exceeded this level, however, the opposite was true.[4] The study did not examine why this was the case, but presumably, if nothing else, where pay levels are higher, the rewards represent less of an incentive.

This is not the only study of OECD data, however. In 2010, a much more detailed analysis of student PISA results from 2003 that – unlike the PISA in Focus report – controlled for student, school, and country characteristics found that performance-related salary adjustments led to significant improvement everywhere in maths, reading, and – to a lesser extent – science scores.[5] This suggests strongly that, all other things being equal, performance pay does motivate teachers to do better over the long term.

It would be surprising if performance-related pay had no impact at all on performance, but the often conflicting research makes two points clear. The first is that the form of reward matters and the optimal approach may well differ based on the context of a given education system. The second is that pay is only part of what motivates teachers. Issues of social status and professional independence are typically even more important. Pay arrangements that are consistent with promoting these rewards are therefore more likely to be successful.

[1]Eric A. Hanushek, “The economic value of higher teacher quality,” Economics of Education Review, 2011.

[2]Karthik Muralidharan and Venkatesh Sundararaman, “Teacher Performance Pay: Experimental Evidence From India”, National Bureau of Economic Research Working Paper 15323, September 2009; Victor Lavy, “Performance Pay and Teachers’ Effort, Productivity, and Grading Ethics,” American Economic Review, December 2009; Matthew Springer et al., Teacher Pay for Performance: Experimental Evidence from the Project on Incentives in Teaching, 2010; Steven Glazerman, Allison Seifullah, An Evaluation of the Chicago Teacher Advancement Program (Chicago TAP) After Four Years, March 2012.

[3]Roland G. Fryer et al, “Enhancing the Efficacy of Teacher Incentives through Loss Aversion: A Field Experiment” National Bureau of Economic Research Working Paper 18237, July 2012.

[4]“Does performance-based pay improve teaching?”, PISA in Focus, May 2012.

[5]Ludger Woessmann, “Cross-Country Evidence on Teacher Performance Pay”, Forschungsinstitut zur Zukunft der Arbeit Discussion Paper 5101, July 2010.

Putting a face to career options

The case for transparency

The language of learning