norc at the university of chicago the university of...

NORC at the University of Chicago

The University of Chicago

The Design of Teacher Incentive Pay and Educational Outcomes: Evidence from the New YorkCity Bonus ProgramAuthor(s): Sarena F. Goodman and Lesley J. TurnerSource: Journal of Labor Economics, Vol. 31, No. 2 (April 2013), pp. 409-420Published by: The University of Chicago Press on behalf of the Society of Labor Economists and theNORC at the University of ChicagoStable URL: http://www.jstor.org/stable/10.1086/668676 .

Accessed: 03/05/2013 10:11

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press, Society of Labor Economists, NORC at the University of Chicago, TheUniversity of Chicago are collaborating with JSTOR to digitize, preserve and extend access to Journal ofLabor Economics.

http://www.jstor.org

This content downloaded from 129.119.32.27 on Fri, 3 May 2013 10:11:19 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/action/showPublisher?publisherCode=sle

http://www.jstor.org/action/showPublisher?publisherCode=norc

http://www.jstor.org/stable/10.1086/668676?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


The Design of Teacher Incentive Pay
and Educational Outcomes:
Evidence from the New York CityBonus Program

Sarena F. Goodman, Columbia University

Lesley J. Turner, University of Maryland, College Park

Teacher compensation schemes are often criticized for lacking aperformance-based component. Proponents argue that teacher in-
centive pay can raise student achievement and stimulate system-wide innovation. We examine a group-based teacher incentivescheme implemented in New York City and investigate whetherspecific features of the program contributed to its ineffectiveness.Although overall the program had little effect on student achieve-ment, we show that in schools where incentives to free ride wereweakest, the program led to small increases in math achievement.Our results underscore the importance of carefully considering thedesign of teacher incentive pay programs.
I. Introduction

Teacher compensation schemes are often criticized for their lack of per-formance pay. In other sectors, incentive pay increases worker effort andoutput by aligning the interests of workers and employers, providing in-formation about the most valued aspects of an employee’s job, and moti-

We are especially grateful to Jonah Rockoff for his thoughtful comments andadvice.We also thank ToddKumler, BentleyMacLeod, BenMarx, DerekNeal, Petra
Persson, Maya Rossin, Jesse Rothstein, Miguel Urquiola, Till Von Wachter, ReedWalker, and seminar participants at Columbia’s appliedmicroeconomics colloquium,the AEFA annual meeting, Teacher’s College’s economics of education workshop,
[ Journal of Labor Economics, 2013, vol. 31, no. 2, pt. 1]© 2013 by The University of Chicago. All rights reserved.0734-306X/2013/3102-0005$10.00

409



vating workers to provide costly effort ðGibbons 1998; Lazear and Oyer2012Þ. In this article, we examine a group-based teacher incentive scheme

410 Goodman/Turner

implemented by theNewYorkCityDepartment of Education ðDOEÞ andinvestigate whether specific features of the program contributed to its in-effectiveness.In 2007, close to 200 schools were randomly selected from a group of

high-poverty schools.1 These schools could earn school-wide bonuses bysurpassing goals primarily based on student achievement. Successful schoolswould earn lump sum payments equal to $3,000 per union teacher ð3%–7%of annual teacher payÞ. Several independent studies show that the bonusprogram had little overall effect on either math or reading achievementðSpringer and Winters 2009; Goodman and Turner 2010; Fryer 2011Þ. Weshow that in schools where smaller groups of teachers were responsiblefor instructing tested students, the program led to small but significant in-creases in student achievement. Our finding is consistent with predictionsthat group-based incentives are diluted by the potential for free riding whenpayments depend on actions of a large number of workers ðHolmstrom1982Þ.Several features of the educational sector complicate the design of

teacher performance pay. First, performance pay is most effective whenemployers can measure worker output or when observable effort and pro-ductivity are closely aligned. Monitoring teachers is costly and measuringindividual teachers’ contributions to student achievement is difficult. Sec-ond, although education is a complex good and teachers must allocatetheir effort across several activities, teacher incentive pay is often linkedto a single performance measure ðe.g., student test scoresÞ, which maylead teachers to direct effort away from other beneficial classroom ac-tivities ðHolmstrom and Milgrom 1991Þ.2 Despite these issues, studiesfrom outside the United States demonstrate that teacher incentive pay canincrease student achievement ðe.g., Lavy 2002, 2009; Muralidharan andSundararaman 2011Þ.Specific features of the New York City bonus program may have limited

its effectiveness. First, the program linked incentive pay to school-wideperformance goals. In theory, group incentive pay is most effective in the

and Harvard’s Kennedy School’s Program on Education Policy and Governance’s

1 This experiment was designed and implemented by the New York City De-partment of Education and the teachers’ union. Random assignment was con-ducted by Roland Fryer, and RAND performed the official evaluation.

2 Teachers may also be induced to focus on narrow, exam-related basic skillsmanipulate test scores, or focus on students whose performance contributes moretoward goals ðe.g., Jacob and Levitt 2003; Jacob 2005; Cullen and Reback 2006Neal and Schanzenbach 2010Þ.

Merit Pay conference for useful feedback. We are grateful to the New York CityDepartment of Education for the data used in this article. Contact the correspondingauthor, Lesley J. Turner, at [email protected].


,

;


context of a joint production technology ðItoh 1991Þ. For instance, if anindividual teacher’s effort has positive impacts on the effort exerted by her

Design of Teacher Incentive Pay 411

peers ðe.g., Jackson and Bruegmann 2009Þ, group incentives may outperformindividual incentives. Otherwise, relative to individual incentives, group in-centives decrease individual returns to effort and will lead to free riding un-less workers monitor each other’s effort.We test for free riding by allowing the bonus program’s impacts to vary

by the number of teachers with students who are tested ðand thereforecontribute to the probability that a school qualifies for the bonus awardÞ.To test for the importance of joint production and monitoring, we exam-ine whether program impacts vary by the degree to which teachers reportcollaborating in lesson planning and instruction using a survey adminis-tered prior to program implementation. We show that the bonus programraised math achievement in schools with a small number of teachers withtested students, although these impacts are small ð0.08 student-level stan-dard deviationsÞ and only marginally significant in the program’s sec-ond year. We present suggestive evidence of positive program impacts inschools with a high degree of collaboration.Second, teachers already faced negative incentives when the bonus

programwas implemented. In fall 2007, the DOE instituted a district-wideaccountability system that imposed sanctions on schools that did not meetthe same goals used in determining bonus receipt. Thus, estimated impactsof the bonus program represent the effect of teacher performance pay inschools already under accountability pressure. However, this may be themost appropriate context to examine since many states have implementedaccountability systems and all public school districts face pressure fromNo Child Left Behind provisions. Finally, we find no differences in theimpacts of the bonus program when we compare schools under differentdegrees of accountability pressure. This suggests that our results are notsolely driven by the dilution of incentives due to the accountability systemðGoodman and Turner 2010Þ.Third, teachers’ lack of understanding of the bonus program’s complex

goals may have limited its efficacy. Alternatively, since bonus awards wereprovided if a school’s performance reached a set threshold, if thresholdswere set too high or too low, a large number of teachers may have opti-mally responded by not changing their behavior ðNeal 2011Þ. However,the metrics used to determine bonus payments were the same goals usedby the district-wide accountability system, and Rockoff and Turner ð2010Þshow that negative incentives provided through this system increasedstudent achievement.3

On a related note, a committee within each school had some discretion overbonuses would be distributed. However, the distribution scheme was set ex, and most schools chose equal or close to equal distributions.

3

howante



II. Data and Empirical Framework

412 Goodman/Turner

Our analyses focus on schools classified as elementary, middle, and kin-dergarten through grade 8 ðK–8Þ, schools eligible for selection into thebonus program. A total of 181 schools were chosen to participate in thebonus program; 128 schools were placed in the control group.4 We usepublicly available DOE data and measure academic achievement usingaverage math and reading test scores in the 2006–7, 2007–8, and 2008–9school years.We estimate the main effect of the bonus program using the following

model:

yjt 5 dDjt 1Xjtb1 εjt; ð1Þ

where yjt is the outcome of interest for school j in year t, Djt is an indi-cator selection into the bonus program’s treatment group ðregardless ofwhether the school ultimately participatedÞ, Xjt is a vector of schoolcharacteristics, and εjt is an idiosyncratic error term.5 School observationsare weighted by the number of tested students. With successful randomassignment, Djt is independent of omitted variables and d̂ represents thecausal impact of the bonus program.

III. Results

A. Group Bonuses and the Free Rider Problem

Teachers should respond to the bonus program by increasing their ef-fort until the expected marginal benefit is equal to the expected marginalcost. However, the probability that a treated school reaches its goal andreceives a bonus primarily depends on its students’ performance on math

4

5 Covariates include the outcome measured in 2007; school type indicators ði.e.elementary, middle, or K–8Þ; the percentage of students who are English LanguageLearners, in special education, Title I free lunch recipients, and minorities; andperformance under the New York City accountability system ðschool account-ability scores and peer indicesÞ.

A small number of experimental sample schools were excluded prior to ran-dom assignment. Moreover, two of the 181 schools originally assigned to thetreatment group were moved to the control group prior to notification of theirassignment; we classify these as treatment group schools. Treatment schools wereeligible to earn bonuses if 55% of full-time United Federal of Teachers staff votedin favor of participation. Twenty-five schools voted not to participate or withdrewfrom the program after voting. Finally, four schools that were originally assignedto the control group were allowed to vote and to participate in the bonus programwe consider these control schools. Ultimately, 158 schools were eligible to earnbonus payments.


,

;


and reading exams. Thus, the impact of an individual’s teacher’s effort onher expected bonus is decreasing as the number of teachers with tested


students increases.6 The diffusion of responsibility for test score gainsacross many teachers may dilute the incentives of the bonus scheme. More-over, monitoring may be more difficult in schools with more teachers.We test for evidence of free riding by allowing treatment effects onmath

and reading scores to vary by the number of math and reading teachers,respectively. We only focus on teachers whose students take these exams,rather than the full set of teachers in a school, since only teachers withtested students contribute to the probability that a school earns its bonus.7

The first set of regressions in table 1 show the main effect of the bonusprogram on math and reading achievement.8 We first add an interactionbetween the number of math/reading teachers ðrelative to the mean num-ber of such teachers in the sampleÞ and the treatment indicator ðcols. 2and 5Þ and then interact treatment status with an indicator for schoolsin the bottom quartile of the number of teachers with tested studentsðapproximately 10 or fewer teachers in elementary and K–8 schools andfive or fewer teachers in middle schoolsÞ. We only present results fromspecifications that include covariates; however, results are similar whenwe exclude covariates or instrument for actual treatment with initial as-signment.We find evidence of free riding. For schools at the bottom of the dis-

tribution of the number of teachers with tested students, we estimate apositive effect of the bonus program on math achievement in the first yearof the program and a positive but insignificant effect in the second year,although we cannot reject a test of equality of effects across years. In 2008,the bonus program resulted in a 3.2-point ð0.08 student-level standarddeviationÞ increase in math achievement.9

6 Consider two extremes, a school with only one teacher with tested studentsand a school with an infinite number of these teachers. In the first case, the teacherwill either respond to the program by increasing her effort to the expected levelnecessary to achieve the school’s goal or not respond ðif the size of the bonus is lessthan the cost of exerting this level of effortÞ. In the second case, changes in a giventeacher’s effort do not affect the probability that the school receives the bonus, andit will be optimal for teachers not to respond to the program.

7 On average, treatment and control group schools have 55 teachers in total, butonly 16 teach tested students.

8 The small number of middle and K–8 schools that are missing information onthe number of teachers with tested subjects are excluded.

9 Another implication of this finding is that, in schools with a large number ofteachers with tested students, the bonus program had a negative impact on studentachievement. One explanation is that the bonus program crowded out teachers’intrinsic motivation and only in schools where incentives were not diluted by freeriding did the potential monetary rewards lead to increased teacher effort.



Table 1Free Riding and the Impact of Teacher Incentives on Student Mathand Reading Achievement

Reading Math

ð1Þ ð2Þ ð3Þ ð4Þ ð5Þ ð6ÞYear 1 ð2007–8Þ:Treatment 2.372 .046 2.667 2.871 2.536 21.445

ð.490Þ ð.499Þ ð.519Þ ð.530Þ ð.568Þ ð.561Þ*Treatment � numberof teachersðmean 5 0Þ 2.233 2.176

ð.089Þ** ð.097Þ1Treatment � firstquartile of numberof teachers 2.044 4.670

ð1.575Þ ð1.483Þ**Treatment effect:schools in first quartile 1.377 3.225

ð1.481Þ ð1.395Þ*Observations 300 300 300 301 301 301

Year 2 ð2008–9Þ:Treatment 2.579 2.395 2.909 21.297 2.979 21.893

ð.539Þ ð.572Þ ð.556Þ ð.668Þ1 ð.726Þ ð.689Þ**Treatment � numberof teachersðmean 5 0Þ 2.126 2.171

ð.099Þ ð.144ÞTreatment � firstquartile of numberof teachers 2.122 4.826

ð2.067Þ ð2.579Þ1Treatment effect:schools in first quartile 1.213 2.933

ð1.968Þ ð2.461ÞObservations 294 294 294 294 294 294

NOTE.—The dependent variable is average math or reading test scores. For each year, each columndenotes a separate regression. For each year, the first row displays the estimated impact of treatmentgroup assignment; in cols. 2 and 5, treatment group assignment is interacted with the ðdemeanedÞ numberof teachers with tested students; in cols. 3 and 6, treatment group assignment is interacted with anindicator for being a school in the lowest quartile of teachers with tested students. The number of mathteachers for schools in the first quartile is less than or equal to 10 for elementary and K–8 schools and fivefor middle schools; the number of reading teachers for schools in the first quartile is less than or equal to10 for elementary and K–8 schools and six for middle schools. The regressions are weighted by thenumber of tested students in each school. Schools with no information on teachers with tested students aredropped. Robust standard errors are in parentheses. See the text for a description of additional controlsincluded in regressions.

1 p < .10.* p < .05.** p < .01.



Group-based incentive pay may outperform individual incentives in thecase of joint production. If the degree to which teachers work together


varies across schools, the bonus program may have been effective inschools with a high level of cooperation among teachers. To proxy for theextent of joint production in a school, we construct a measure of schoolcohesiveness using teachers’ answers to a set of five survey questions priorto the announcement of the bonus program.10 This measure may alsoincorporate the degree to which teachers are able to monitor their col-leagues. We sum responses across survey questions and standardize theindex so it has a mean of zero and standard deviation equal to one. Schoolswith high levels of cohesion are distinct from those with a small number ofteachers with tested students.11

Table 2 tests for heterogeneity in the impact of the bonus program byschool cohesion. We first interact treatment with the linear index ðcols. 2and 5Þ and then interact treatment with an indicator for schools withabove-average cohesion ðcols. 3 and 6Þ. The point estimates for schoolswith below-average cohesion are marginally significant and negative inboth subjects and both years, while the interaction of treatment and theindicator for above-average cohesion is significant, positive, and of greatermagnitude. Results suggest that the bonus program may have had detri-mental effects in schools with low levels of cohesion and small positiveeffects on achievement in cohesive schools.

B. Teacher Effort

A primary motivation for performance-based pay is to provide teacherswith incentives to increase effort devoted to raising student achievement.Althoughwe do not directly observe teacher effort, we canmeasure teacherattendance, which may be correlated with effort decisions and which con-tributes to student achievement ðe.g., Miller, Murnane, and Willett 2008;Herrmann and Rockoff 2012Þ. We measure teacher absences using aggre-gate statistics from individual teacher data and estimate models where thedependent variable is the average number of absences taken duringthe months when schools first learned of their eligibility for the bonusprogram andwhen the last examswere taken.12 If teachers believe that their

10 These surveys were administered in spring 2007. Questions include: ð1Þ the
extent to which teachers report feeling supported by fellow teachers, ð2Þ whethercurriculum and instruction is aligned within and across school grades, ð3Þwhetherthe principal involves teachers in decision making, ð4Þ whether school leadersencourage collaboration, and ð5Þ whether teachers collaborate to improve in-struction. We exclude schools with a survey response rate under 10%.
11 This index has a small, negative, and statistically insignificant correlation withthe number of math and reading teachers in a school.

12 We thank Jonah Rockoff for constructing these aggregate statistics for thepurpose of this research.



Table 2School Cohesion and the Impact of Teacher Incentives on Student Mathand Reading Achievement

Reading Math

ð1Þ ð2Þ ð3Þ ð4Þ ð5Þ ð6ÞYear 1 ð2007–8Þ:Treatment 2.328 2.091 2.908 2.628 2.270 21.264

ð.494Þ ð.515Þ ð.591Þ ð.530Þ ð.551Þ ð.674Þ1Treatment �cohesion index .283 .283 .766

ð.547Þ ð.547Þ ð.619ÞTreatment �above-averagecohesion index 1.840 1.962

ð.975Þ1 ð1.131Þ1Treatment effect,schools with aboveaverage cohesion .932 .698

ð.789Þ ð.891ÞObservations 300 300 300 301 301 301

Year 2 ð2008–9Þ:Treatment 2.544 2.328 21.139 21.118 2.669 22.266

ð.540Þ ð.562Þ ð.666Þ1 ð.666Þ1 ð.679Þ ð.869Þ**Treatment �cohesion index .361 .361 1.105

ð.626Þ ð.626Þ ð.839ÞTreatment �above-averagecohesion index 1.850 3.347

ð1.093Þ1 ð1.446Þ*Treatment effect:schools with aboveaverage cohesion .710 1.081

ð.868Þ ð1.105ÞObservations 296 296 296 297 297 297

NOTE.—The dependent variable is average math or reading test scores. For each year, each columndenotes a separate regression. For each year, the first row displays the estimated impact of treatmentgroup assignment; in cols. 2 and 5, treatment group assignment is interacted with the teacher cohesionindex ðmean5 0, SD5 1, across all New York City schoolsÞ; in cols. 3 and 6, treatment group assignmentis interacted with an indicator for having a cohesion index greater than zero. The regressions are weightedby the number of tested students in each school. Schools with a teacher survey response rate below 10%are dropped. Robust standard errors are in parentheses. See the text for a description of additional controlsincluded in regressions.

1 p < .10.* p < .05.** p < .01.



attendance can affect the probability of bonus receipt by raising studentachievement, the program’s impacts on absenteeism should be largest over

Table 3The Impact of Teacher Incentives on Teacher Absences Due to Personaland Sick Leave

All TeachersTeachers of Tested

Students

ð1Þ ð2Þ ð3Þ ð4ÞYear 1 ð2007–8Þ:Treatment .001 2.158 2.217 2.156

ð.091Þ ð.146Þ ð.148Þ ð.163ÞTreatment � number of teachers ðmean 5 0Þ .013

ð.022ÞTreatment � first quartile of numberof teachers 2.236

ð.390ÞTreatment effect, schools in first quartile 2.391

ð.352Þ

Year 2 ð2008–9Þ:Treatment .045 .151 .203 .161

ð.119Þ ð.175Þ ð.192Þ ð.200ÞTreatment � number of teachers ðmean 5 0Þ .005

ð.032ÞTreatment � first quartile of numberof teachers .158

ð.621ÞTreatment effect, schools in first quartile .319

ð.576ÞNOTE.—The dependent variable is average absences/teacher has taken for personal or sick leave be-

tween November and March ðyear 1Þ or September and March ðyear 2Þ. Observations 5 301 for year 1and 294 for year 2. For each year, each column denotes a separate regression. The first row for each yeardisplays the estimated impact of treatment group assignment on absences for all teachers. The second rowfor each year displays the estimated impact of treatment group assignment on absences for teachers withtested students. In col. 3, treatment group assignment is interacted with the ðdemeanedÞ number ofteachers with tested students; in col. 4, treatment group assignment is interacted with an indicator forbeing a school in the lowest quartile of teachers with tested students ðsee the note to table 1Þ. Schools withno information on teachers with tested students are dropped. Robust standard errors are in parentheses.See the text for a description of additional controls included in regressions.


this period.13 We only examine absences that teachers likely have somecontrol over—those taken for illness and personal reasons.Table 3 presents these results; for the given years, each column contains

the estimates from separate regressions. The first column examines theeffect of the bonus program on absences across all teachers within a schooland shows no measurable impact on overall attendance. Column 2 focuses

13 In the first year of the program, schools learned of their eligibility in November,
whereas in the second year, eligibility was known in September. In both years, the lastexams occurred inMarch. Results are robust to alternate definitions of the time periodðe.g., November toMarch in the second year or September toMarch in the first yearÞ.


on teachers with tested students, while the third and fourth columnsfollow the same approach as table 2 and interact the treatment indicator

418 Goodman/Turner

with the number of teachers with tested students ðcol. 3Þ or an indicatorfor whether a school falls in the bottom quartile of the number of suchteachers ðcol. 4Þ.Program impacts on attendance are not consistent across years. In the

program’s first year, for schools with a small number of teachers withtested students, attendance increased.14 Conversely, in the second year ofthe program, we find positive but insignificant impacts on absenteeism.Finally, we test whether the bonus program had heterogeneous impactsaccording to initial teacher effort. For instance, initially low-effort ðhigh-absenceÞ teachers may be the only group with the ability to respondthrough increasing attendance. Conversely, if, ex ante, high-effort teachersbelieved that achieving the bonus program goals was a high probabilityevent, theymay have responded by reducing their effort.However, we findno evidence that teacher absenteeism varies along this dimension ðresultsavailable upon requestÞ. In the United States, attendance may not be thedimension along which teachers respond to incentive pay.

IV. Conclusions

In many sectors, performance-based pay enhances effort, output, andother desirable outcomes. Evidence from Israel and India suggests thatproperly structured teacher incentive pay programs can benefit students.However, despite substantial expenditures—over $40 million in the pro-gram’s first 2 years—the New York City bonus program did not raise stu-dent achievement. This article discusses several features of the New YorkCity bonus program that may have contributed to its ineffectiveness. Weprovide suggestive evidence that the group-based structure of the programmay have been detrimental in the majority of schools where the number ofteachers responsible for tested students is large. Conversely, the programimproved math achievement in schools with fewer teachers responsible fortested students or that had a more cohesive group of teachers. A lack ofmonitoring as well as the diffusion of responsibility for test score gainsamong many teachers may have diluted the incentives of the opportunity toearn bonuses. Our results are consistent with the long-standing literature ineconomics on the importance of taking into consideration free riding, jointproduction, and monitoring when designing incentive systems, and theysuggest that a one-size-fits-all approach may not be the most effective whenimplementing incentive pay schemes within a school district.Given that team-based incentives in other contexts resulted in student

achievement gains, other features of the New York City program mayhave also contributed to its ineffectiveness. Neal ð2011Þ suggests that results

14 However, impacts are only significant in schools at the 10th percentile in thedistribution of number of teachers ðresults available upon requestÞ.



from economic theory offer valuable insights into optimal incentive design.For instance, an intervention in India utilized a piece-rate payment scheme:


teachers or schools received bonus payments for incremental improve-ments in student achievement ðMuralidharan and Sundararaman 2011Þ. Thisavoids threshold effects of schemes like the New York City bonus program,which dilute incentives for teachers for whom the probability of bonus re-ceipt is close to zero or one.Even so, many challenges in designing effective teacher incentive schemes

remain. Incentive pay programs that come about as a compromise betweenschool districts and teachers unions’ might contain incentives that are sodiluted they are destined to fail. Finally, the most important margin throughwhich teacher pay can improve student achievement may be the extensivemargin, or the decision to enter the teaching profession. Small-scale teacherincentive pay experiments cannot provide information concerning the gen-eral equilibrium effects of overall increase in teacher pay or movement to-ward performance-based compensation.Currently, the US government provides significant funding through the

Race to the Top program. Eligibility for Race to the Top funding dependson districts’ ability and willingness to link student achievement to indi-vidual teachers and use this data in teacher evaluations, but the programgrants districts a great deal of discretion in designing performance paysystems. In 2010, 62 school districts and nonprofit groups received over$400 million in funding from the federal Teacher Incentive Fund. Our re-sults underscore the importance of the structure of performance pay in ed-ucation. Policy innovations in this area should be carefully considered, tak-ing into account personnel economics theory and research.

References

Cullen, Julie Berry, and Randall Reback. 2006. Tinkering toward acco-lades: School gaming under a performance accountability system. In Ad-vances in appliedmicroeconomics, vol. 14, Improving school accountability,ed. Timothy J. Gronberg and Dennis W. Jansen. Oxford: JAI.

Fryer, Roland G. 2011. Teacher incentives and student achievement: Evi-dence from New York City public schools. Working Paper no. 16850,National Bureau of Economic Research, Cambridge, MA.

Gibbons, Robert. 1998. Incentives in organizations. Journal of EconomicPerspectives 12, no. 4:115–32.

Goodman, Sarena, and Lesley Turner. 2010. Teacher incentive pay andeducational outcomes: Evidence from the NYC bonus program. Work-ing Paper no. 10–07, Program on Education Policy and Governance,Cambridge, MA.

Herrmann, Mariesa A., and Jonah E. Rockoff. 2012. Worker absence andproductivity: Evidence from teaching. Journal of Labor Economics 30,no. 4:749–82.



Holmstrom, Bengt. 1982. Moral hazard in teams. Bell Journal of Economics13, no. 2:324–40.

420 Goodman/Turner

Holmstrom, Bengt, and Paul Milgrom. 1991. Multitask principal-agentanalyses: Incentive contracts, asset ownership, and job design. Journal ofLaw, Economics, and Organization 7 ðJanuaryÞ: 24–52.

Itoh, Hideshi. 1991. Incentives to help in multi-agent situations. Economet-rica 59, no. 3:611–36.

Jackson, C. Kirabo, and Elias Bruegmann. 2009. Teaching students andteaching each other: The importance of peer learning for teachers. Amer-ican Economic Journal: Applied Economics 1, no. 4:1–27.

Jacob, Brian A. 2005. Accountability, incentives, and behavior: Evidencefrom school reform in Chicago. Journal of Public Economics 89:761–96.

Jacob, Brian A., and Steven D. Levitt. 2003. Rotten apples: An investi-gation of the prevalence and predictors of teacher cheating. QuarterlyJournal of Economics 118, no. 3:843–77.

Lavy, Victor. 2002. Evaluating the effect of teachers’ group performanceincentives on pupil achievement. Journal of Political Economy 110, no. 6:1286–1317.

———. 2009. Performance pay and teachers’ effort, productivity andgrading ethics. American Economic Review 99, no. 5:1979–2011.

Lazear, Edward P., and Paul Oyer. 2012. Personnel economics. In Hand-book of organizational economics, ed. Robert Gibbons and D. John Rob-erts, 479–519. Princeton, NJ: Princeton University Press.

Miller, Raegen T. Richard J. Murnane, and John B. Willett. 2008. Doworker absences affect productivity? The case of teachers. InternationalLabour Review 147, no. 1:71–89.

Muralidharan, Karthik, and Venkatesh Sundararaman. 2011. Teacherperformance pay: Experimental evidence from India. Journal of Polit-ical Economy 119, no. 1:39–77.

Neal, Derek. 2011. The design of performance pay in education. InHandbook of economics of education, vol. 4, ed. Eric Hanushek, SteveMachin, and Ludger Woessmann. Amsterdam: Elsevier.

Neal, Derek, and Diane W. Schanzenbach. 2010. Left behind by design:Proficiency counts and test-based accountability. Review of Economicsand Statistics 92, no. 2:263–83.

Rockoff, Jonah, and Lesley J. Turner. 2010. Short-run impacts of account-ability on school quality. American Economic Journal: Economic Policy 2,no. 4:119–47.

Springer, Matthew G., and Marcus A. Winters. 2009. The NYC teacherpay-for-performance program: Early evidence from a randomized trial.Civic Report no. 56, Manhattan Institute.



norc at the university of chicago the university of...

Documents