Monetary and Non-monetary Teacher Incentives - Evidence, Concerns and Challenges: Ensuring a Learning India S5 E.004

[29th post in 'Ensuring a Learning India' series. 42 in total. One post per day. 13 more to go.]

The previous section discussed problems of public school teachers as perceived by them.  Teacher incentives is one of the most cited solution to improve the quality of teaching in public schools. This section discusses three broad categories of such incentives. While a completely literature review of various forms of incentives is beyond the scope of this section, some important studies are highlighted.

I. Monetary incentives

There are broadly four types of monetary incentives discussed in the context of school teachers.

1. Teacher pay: One of the most cited causes for the poor quality of public schools is the low teacher pay, and hence the recommendations to pay good salaries. Some people often suggest the large scale recruitment of low-salaried contract teachers, which led to overall teacher quality.

Evidence doesn’t seem to support these claims. One, regular teachers in public schools are paid as per the 6th pay commission (and 7th pay commission if it comes out) and are not low as generally perceived. Two, as discussed in the section on ‘Four not-so critical constraints’, large scale evaluation of the contract teachers in Andhra Pradesh suggests that contract teachers who receive a fraction of salary, are no worse than regular public school teachers. If salary was the main constraint, then the regular teachers would have been out-performing the low salaried teachers. While there is a risk to extrapolating this evidence to other contexts in India, this at least shows that the widely held belief needn’t necessarily be true in some contexts. Similar studies in other countries also suggest that, merely increasing teachers’ pay doesn’t lead to outcomes.

This doesn’t mean that teachers shouldn’t be paid good salaries. They can be paid as the resources may allow but citing this as the main reason, prioritizing it over everything else, and giving disproportionate space in the scarce discourse bandwidth, may not be the wise thing to do.

2. Pay for performance: Under this, teachers are paid incentives as per the improvement of test scores of the children they taught. There is good amount of literature on this aspect but the one conducted in India may be particularly relevant.

Karthik Muralidharan and Sundararaman conducted a large scale randomized evaluation of ‘pay for performance’ incentives in Andhra Pradesh. This study is an important not just because of its scale and it’s randomized evaluation but the incentives in this evaluation are carefully designed to prevent some of the widely known ways of gaming the incentive systems. One group of teachers in 100 schools across 5 districts in Andhra Pradesh were paid Rs.500/- for every percentage point of mean improvement in test scores of their students. The other group of teachers in 100 schools (control group) didn’t receive any incentives and worked in the status quo conditions. The study[i] finds that

(i) ‘students who completed their full five years of primary school under the program performed significantly better than those in control schools by 0.54 and 0.35 standard deviations in math and language tests respectively. These students also scored 0.52 and 0.3 standard deviations higher in science and social studies tests even though there were no incentives on these subjects’.

(ii) “test score gains represent genuine additions to human capital as opposed to reflecting only ‘teaching to the test’. Students in individual teacher incentive schools score significantly better on both non-repeat as well as repeat questions; on both multiple-choice and free-response questions; and on questions designed to test conceptual understanding as well as questions that could be answered through rote learning.’

(iii) “We measure changes in teacher behavior and the results suggest that the main mechanism for the improved outcomes in incentive schools is not reduced teacher absence, but increased teaching activity conditional on presence. We also measure household responses to the program – for the cohort that was exposed to five years of the program, at the end of five years – and find that there is no significant difference across treatment and control groups in either household spending on education or on time spent studying at home, suggesting that the estimated effects are unlikely to be confounded by differential household responses across treatment and control groups over time.”

3. Group incentives: It is argued that individual incentives might not be fair because a students’ development is a combined effort of all teachers in the school and that the factors at school level should also be considered. For this purpose, instead of paying each individual as per their own performance, all teachers in the school are paid as per the performance of the school. There are reports from other countries that some schools are recently considering shifting to ‘school based’ incentive model as opposed to ‘individual based incentive model’.

The large scale study in AP discussed above also evaluates this aspect of incentives. It finds that while group incentives were as effective as individual incentives in the first year of the study, at the end of five years, the difference wasn’t significant though it was positive.

4. Pay upfront but return on failing to achieve targets: In this structure of incentives, certain amount of money is paid upfront to the teachers at the beginning of the school year with an agreement that, they have to return some or all of it at the end of the year, based on their performance. The idea is that, the prospect of losing money propels individuals more than the prospects of gaining money, a phenomenon called as loss aversion.

A study[ii] conducted in Chicago by Roland Fryer, Steven Levitt, John List and Sally Sadoff finds that it increases math scores between 0.201 and 0.398 standard deviations, while the incentives designed as per standard structure had smaller but insignificant effects.

Two key aspects of monetary incentives: One, the incentives above are paid as per the improvement of students of the same teacher, and not as per ranking of students in a particular exam. Two, incentive systems are different from accountability systems. Incentive systems pay extra money conditional on performance, while accountability systems may both incentivize and also penalize teachers based on performance of student, the penalties ranging from salary cuts to termination of service. The section above only deals with incentives conditional on performance and not punitive actions.

In summary, monetary incentives can be an effective way of improving teacher performance. The design of these incentives is crucial; it must take the possible perverse incentives into account to prevent gaming of the system, and must be iterated continuously depending on the feedback and response of the system.

Concerns and challenges of monetary incentives:

1. Teaching to the test – It is argued that incentivizing teachers for the performance of students on specific assessments makes teachers to narrow down the teaching to only those concepts relevant for the assessments.

One way to address this problem is to design the assessments appropriately so that they test the desired skills of students, so that even if one focuses only on the assessment, they will still be able to acquire the desired skills.

In governance theme, we discussed that the approach to reform depends on the state of the system, transition from poor to good systems and good to best systems. If the employees in an organization aren’t working, not attending the office regularly, then by just imposing strict rules on attendance and monitoring inputs like work hours clocked etc., might show at least show outcomes initially, merely because of the fact that they are at least now putting some efforts. This approach might not work for reforming systems which are already good but trying to become the best. One must see the ‘teaching to the test’ in this context. In contexts like India, where lack of efforts is one of the constraints, then by merely getting teachers to put some efforts can result in some gains initially. The Andhra Pradesh study corroborates this.

The argument that monetary incentives for performance of students will narrow down the teaching isn’t appropriate in systems of weak capacity because, even if aims to achieve broader outcomes, effort is a necessary condition. Due to the lack of this effort, currently students aren’t learning much. The increase in efforts due to these incentives will at least result in some outcomes, something better than the current situation. It is true that this shouldn’t be seen as the only approach to incentivize teachers and the incentive structures must adapt as the system progress to ‘good state’ and while attempting to transform to ‘best’.

2. Difficulty in measuring test gains attributable to teacher – One, it is argued that the test gains can’t be solely attributed to teacher and the other socio-economic factors are also responsible. This can be true but when student performance is measured against their own baseline, this shouldn’t be an issue.

Two, all teachers don’t have students of same ability. It may happen that some teachers get students of high ability and hence showing improvements of their test scores is much difficult as compared to improving students with low test scores (increasing from 30 to 40 is easier compared to increasing from 90 to 100). This is a genuine concern and even in the large scale study in AP, researchers had to adjust for this after first two years, calculating targeted scores for each child and measuring against the percent achieved. When such systems are implemented in large scale, concerns like these may result in political resistance, especially if there is no stake holder buy in. Such changes in formula can be troublesome politically. However, one should note that in the AP study, “85% of teachers reporting a favorable opinion about the idea and 68% mentioning that the government should try to scale up programs of the sort implemented under this project.

3. Structure and design of incentives matter – Paying salary over the base pay, giving money upfront as opposed to giving it at the end of the year, and individual vs. group based incentives are broad categories of design but even the micro details of such incentives matter. At times, they can make or break the system.

a) Amount of incentive: ‘Stakes’ often determine people’s response to incentives. Amount of money in this context decides the stakes and the response of people and the extent to which they can go to game the system depends on this. Giving Rs.10,000/- per percentage gain in scores is different from giving Rs.1,000/- per percentage gain.

b) Threshold effects: If the incentive is based on achieving certain threshold and is not paid for every percentage improvement, then it is possible that teachers might just focus on the good performing students to boost the average score or they may prevent the low performing students from taking the test.

c) Feasibility to achieve the incentive: The metrics of the incentive system should be such that they are perceived as achievable by the teachers. If they are designed to make it difficult to achieve, then teachers may end up putting efforts because they wouldn’t be getting incentive anyways. This might even become a justification for not putting efforts. This is important especially in cases where the incentives are designed based on loss aversion principle.

If money is given upfront to teachers, and the incentive metrics are perceived as unachievable by teachers, and if teachers have to end up paying the money back, then teachers might view returning money as paying fine for not working, and can justify their actions.

4. Credibility of assessments - At the end of the day, teachers should perceive these assessments as credible, and are measuring the true abilities of the child. Design of these assessments is going to be crucial and the government has to invest on this.

5. Administrative issues - There are definitely administrative issues with conducting assessments like transferring question papers, preventing cheating in examination and other issues which are common to most other policies as well.

6. Monetizing non-monetary preferences[iii] - It is argued that there are certain aspects in life which resist monetization, pure friendship for example. Teaching is also one such act, where the passion to teach students is one of the important factors. Monetizing such preferences may not be effective.
It may be true that some people teach out of passion but it shouldn’t mean that they shouldn’t be paid for that. Money is a necessary condition but might not a sufficient condition. It is understandable that one should rekindle the passion to teach in teachers but monetary incentives and this approach can be complementary, and need not be perceived as a zero-sum game.

7. Monetary incentives aren’t sustainable in long term: It is argued that monetary incentives aren’t sustainable in long term as teachers may find it initially motivating but this may fade out with time. Though the evidence of AP’s study suggests that it needn’t be the case, it can be a concern. Hence, these incentives should be adapted to changing circumstances and be revised.

8. Are teachers in elite private schools getting performance pays? : Some argue that teachers in elite private schools work without performance pays and are performing better. Thus, strict monitoring and accountability is the solution and not incentives. One must note that these contexts are different; monitoring and incentives needn’t be seen as zero sum issues, both can be done.

9. Will such initiatives work in government setup? Some of the studies demonstrating positive effects of monetary incentives are cases where an NGO provided these incentives. Incentives in AP study discussed above were also given by an NGO. It is not known if same effects persist if it is implemented through government machinery. For example, a policy in Kenya empowered head masters to give out bonuses to teachers based on their attendance.  The amount was up to 85% of the teacher salary and the unspent amount would go into other accounts of the school. The evaluation results suggest that head masters ended up paying uniform bonuses to everyone, despite their absence. This underscores the need for both the design of the incentive programmes and credible independent assessments.

In summary, monetary incentives can be cost effective way of improving teacher performance. The design of these incentives is crucial; it must take the various possibilities of gaming the system and design the incentives accordingly, and must be iterated continuously depending on the feedback and response of the system.

II. Non-monetary methods to motivate teachers

We have discussed the problems of public school teachers in the previous section. Addressing these issues can be one of the best ways to motivate and incentivize teachers. It is true that some of these issues may have been exaggerated but it is the way they are being perceived and some of them can be genuine issues. At least these genuine issues should be addressed. If these issues aren’t addressed or at least the efforts aren’t made, then these can become good alibis for non-performance.

As discussed earlier, these issues can be broadly categorized into infrastructural, lack of appropriate support and training, lack of recognition, lack of autonomy, unfair systems and overburdening due to extra academic tasks. The infrastructural issues like lack of enough teachers, toilets, textbooks etc., need to be addressed to ensure conducive atmosphere for teachers to work. We already discussed the aspect of teacher training and support in detail. We discuss the issues of recognition, unfair systems and lack of autonomy here.

Recognition - The previous section quoted instances where teachers desired more inspections I craving for praise and recognition. One of the important steps should be to address this and rekindle the spirit of motivation in teachers.

One, while best teacher awards are regularly given to teachers, these are either perceived to be politicized or unachievable. It is also not possible to recognize efforts of individual innovations in classrooms with single award. Thus, a community of local teachers where these are shared and appreciated can be created.

Two, teachers who guide students to do extra-curricular tasks like participating in inter-school competitions etc., can be given extra points. These points can be used while preparing the ranking list during the teacher transfers. Andhra Pradesh state currently does this.

Three, increase levels of hierarchy in public teacher service. The unique feature of public teacher service is that there are only two to three promotions possible in a career of 30-35 years. A primary school teacher can become a teacher in high school and then the head of a high school. Some with higher degrees go onto become lecturers or go to other administrative posts but they are very few in number. A promotion once in 10 years is surely not an encouraging aspect. One can create more levels of hierarchy, not necessarily varying much in the pay scale but at least the fact that one is progressing above can act as a booster. Of course, this alone is not a panacea but it can be one of the solutions.

Unfair systems - Teacher transfers in some states are highly politicized. Most teachers prefer to get postings in urban areas as it is convenient to travel and for other evident reasons. This makes the urban postings highly pricey and these are often awarded directly to teachers based on orders from minister. This can be highly demotivating to teachers who end up working in interior areas. Hence, it is essential to build credible transfer mechanisms for teachers.

Lack of autonomy – Teachers in Mooij’s study reported lack of autonomy as one of their concerns. In some cases, it is true to an extent. Some teachers feel highly constrained by the fact that they have to deal with infrastructural issues in schools like lack of teachers, unmotivated parents and also that they don’t have much elbow space to move around. Even simple requests sent by schools take long time to get approvals and often seemingly small issues like school timings can be troublesome. For example, in some states the starting time and closing time of schools are fixed. In some villages, the first bus to the village reaches after the starting time of the school and sometimes the last bus from the village leaves before the school ends. It is a daily hassle for teachers to deal with these issues. Even if one doesn’t want to break the rules, such situations bring in unnecessary ethical conundrums. In cases like this, schools can be given autonomy to decide starting and ending times, if needed keeping the duration of school fixed. Other than this, schools can be given autonomy so that the teachers feel that they are genuinely empowered and have the resources to bring change, if they desire to.



[i] Long-Term Effects of Teacher Performance Pay: Experimental Evidence from India – Karthik Muralidharan
[ii] Enhancing the Efficacy of Teacher Incentives through Loss Aversion: A Field Experiment -  Roland G. Fryer, Steven D. Levitt, John List, Sally Sadoff
[iii] Phrase taken from the book Tort Theory by Kenneth D. Cooper-Stephenson, Elaine Gibson (page 139)

No comments:

Post a Comment