Teacher Retention: Estimating The Effects Of Financial Incentives In Denver

Our guest author today is Eleanor Fulbeck, who earned her Ph.D. in education policy from the University of Colorado at Boulder in 2011, and is currently a post-doctoral fellow at the University of Pennsylvania.

There is currently much interest in improving access to high-quality teachers (Clotfelter, Ladd, & Vigdor, 2010; Hanushek, 2007) through improved recruitment and retention. Prior research has shown that it is difficult to retain teachers, particularly in high-poverty schools (Boyd et al., 2011; Ingersoll, 2004). Although there is no one reason for this difficulty, there is some evidence to suggest teachers may leave certain schools or the profession in part because of dissatisfaction with low salaries (Ingersoll, 2001).

Thus, it is possible that by offering teachers financial incentives, whether in the form of alternative compensation systems or standalone bonuses, they would become more satisfied with their jobs and retention would increase. As of yet, however, support for this approach has not been grounded in empirical research.

Denver’s Professional Compensation System for Teachers ("ProComp") is one of the most prominent alternative teacher compensation reforms in the nation.* Via a combination of ten financial incentives, ProComp seeks to increase student achievement by motivating teachers to improve their instructional practices and by attracting and retaining high-quality teachers to work in the district.

My research examines ProComp in terms of: 1) whether it has increased retention rates; 2) the relationship between retention and school quality (defined in terms of student test score growth); and 3) the reasons underlying these effects. I pay special attention to the effects of ProComp on schools that serve high concentrations of poor students – “Hard to Serve” (HTS) schools where teachers are eligible to receive a financial incentive to stay. The quantitative findings are discussed briefly below (I will discuss my other results in a future post).

ProComp, financially backed by a federal TIF grant and a voter referendum, has gained national attention as one of the only teacher compensation reforms jointly conceived of and implemented by the district and local teachers’ union (Gonring, Teske, & Jupp, 2007). Starting January 1, 2006, incumbent teachers could opt into ProComp or continue to be paid based on the DPS single-salary schedule. Teachers who were hired on or after this date were required to participate in ProComp. ProComp incentives range from a low of $376 for meeting annual student growth objectives, to a high of $3,379 for earning an advanced degree or a specialty license. Teachers can also receive a Hard to Serve (HTS) incentive of about $2,400 for working at a high-poverty school (about half of DPS schools). On average, teachers who participate in ProComp earned roughly $4,700 in financial incentives, in addition to their base salaries.**

ProComp is an established, fully-funded, and comprehensive alternative teacher compensation reform. Effects of the program on retention rates should not be contaminated with implementation inconsistencies, inadequate resources, or incomplete information. If financial incentives help to retain teachers, one would expect to see evidence of this under ProComp.

My analysis, which uses longitudinal panel data (linking teachers to schools) from 2001-02 to 2010-11, suggests that teachers do respond to financial incentives, albeit at a seemingly low level.

The figure below summarizes results from a series of regression models. ProComp has had a small, positive effect on average annual retention rates (an increase of roughly 2% for all schools). That is, for a school with an average rate of retention of 87% before ProComp, one would expect to observe an average rate of retention of roughly 89% after ProComp.

Furthermore, analyses point to a greater impact on retention rates for schools with high ProComp participation and for the HTS schools where teachers are eligible for the retention incentive (increases of roughly 4% for high-participation schools and HTS schools). At most, these estimates indicate ProComp would yield the retention of an additional 156 teachers (of a population of 4,000 teachers) on average in a given year.

Analyses that explore the relationship between retention and school quality suggest that DPS schools that have had the biggest increases in retention since ProComp was implemented have not gotten significantly worse in terms of school quality, and in some cases, they may have gotten better.

There are, however, caveats to my analysis. The most substantial is the reliance on school-level observational data. Because analyses have been conducted with schools as the unit of analysis, estimates include teachers who do not participate in ProComp and thus are not part of the ProComp/HTS schools treatment. I address this limitation by providing a measure of “exposure” to the treatment (i.e., ProComp participation) for different schools.

In addition, there are factors besides ProComp, such as school climate, location, the principal, the economy, and the number of new teachers (to name a few) that may affect retention rates. Although my models account for a number of these factors, there are many factors (such as the principal) that were not available in the data, and have the potential to bias retention effect estimates.

Although findings suggest teachers do respond to financial incentives, their responses may be tempered by the nominal importance of financial compensation relative to other, non-pecuniary factors. As suggested by prior research (Boyd et al., 2011; Milanowski et al., 2009), the incentives teachers are more likely to respond to involve school leadership and working conditions. This study does not allow a comparative judgment between these various types of incentives, but it does suggest the value of an incentive plan – and evaluation thereof – that offers these non-pecuniary incentives as part of the menu.

This research comes at a time when there is much enthusiasm for and action aimed at improving retention via financial incentives, yet there exists only scant evidence about both the intended and unintended effects of such a policy approach.

This study suggests that incentives do indeed influence teachers’ retention decisions, although, again, effects are not especially large, and they vary by school characteristics.

Given high turnover rates and the shortage of high-quality teachers, especially in high-poverty schools, findings from this study could inform future efforts to improve retention and increase access to a high-quality education for our nation’s neediest students.

- Eleanor Fulbeck


* This research stems from a longitudinal evaluation of Denver’s ProComp that examined the effects of ProComp on teacher quality (measured by teacher value-added estimates on student test-score achievement), teacher retention, teacher salary distributions, and teachers’ attitudes and beliefs about alternative teacher compensation. This study focuses on retention; the effects of ProComp on other outcomes of interest are addressed in companion papers and an evaluation report submitted to the district.

** These figures reflect incentive amounts available to teachers who participated in ProComp during the 2009-10 school year. While most ProComp incentives are annual bonuses, knowledge-based incentives such as earning an advanced degree or specialty license are added to teachers’ “base” salaries.

Note: The research paper discussed in this post is adapted from Ellie’s dissertation, and it is currently under review for publication. You can download the full dissertation here.


It's kind of amazing that the bonus for getting a masters' degree (which we know has zero impact on student learning) is so much higher than the bonus for working in a high-poverty school. Was this a union priority?


Sorry, Stuart, but your negative comment doesn't fly. Do you have statistics to back up your "zero" claim? The point of this article is that financial incentives do not seem to have to expected affect on achievement.


I'm now a retired teacher after 27 years. I taught at the midschool and high school level. The majority of those years at two high poverty high schools. I made a living wage. What burned me out was an eroding lack of support by successive administrations and focus on test preparedness taking away from instruction time. Why don't they do studies on the differences of teacher satisfaction and retention when administrators back up the teachers on discipline and not bothering them on silly issues like requiring nonsense charts and grafts on you walls, anal retentive goals, etc.


Paying more for advanced degrees, training, or working at poor schools seems qualitatively different than paying for performance/test scores. Does this research suggest that one or the other led to the reported retention improvements?


Stuart: All the terms of ProComp were jointly agreed to by the district and union. As Gonring and his co-authors describe in their book, this is what has set ProComp apart and made it potentially more successful than more top-down alternative pay policies.

Demian: You are right that the incentives available under ProComp fall into different categories: 1) performance-based, 2) knowledge/skill-based, and 3) market-based. I did not attempt to tease apart which types of incentives teachers are most likely to respond to in my study summarized above but survey and interview data suggest (perhaps not surprisingly) that teachers are most supportive of knowledge/skills-based and market-based incentives.


Thanks for the additional info Ellie. Based on past research on performance based incentives (from Deci and others), I would expect that performance-based incentives would undermine intrinsic motivation and therefore hurt retention so I was initially surprised when I saw a summary of your research finding. However, given the relative weight of the knowledge/skills-based and market-based incentives and the survey/interview data, the result makes more sense. It would be nice to eventually tease apart the links, but this is still an interesting result. Thanks.