Earnings benefits of Tulsa's pre-K program for different income groups
Highlights
► Test-score percentile effects of Tulsa pre-K are similar across income groups. ► Predicted effects of Tulsa pre-K on adult earnings are 3–4 times program cost. ► Predicted % effects of Tulsa pre-K on earnings are greater for low-income groups.
Introduction
As state pre-K programs, both targeted and universal, have blossomed throughout the United States, public officials have sought hard evidence on these programs’ short-term and long-term effects. We now have substantial evidence on the short-term effects of targeted (Reynolds et al., 2011a, Schweinhart et al., 2005) and universal (Gormley et al., 2008, Henry et al., 2003) programs and the long-term effects of targeted programs (Heckman et al., 2010b, Reynolds et al., 2011a, Reynolds et al., 2011b), but limited evidence on the long-term effects of universal programs (Karoly & Bigelow, 2005). Because the first universal pre-K program in the United States was not established until 1997, we will need to wait many years to estimate the consequences of universal pre-K for adults. Even when long-term data are available on universal pre-K, analysts will face the challenge that universal pre-K programs lack experimental data.
It is possible, however, to make some informed projections, using recent data from Oklahoma's universal pre-K program. Using information from recent work by Chetty et al. (2011) on the link between early test scores and adult earnings, we estimate future earnings effects of pre-K for children who were and were not eligible for a free school lunch. These projections, though not without their limitations, enable us to estimate some of the long-term benefits of a high-quality pre-K experience to children from different socioeconomic strata.
Using data collected on children who were beginning pre-K and kindergarten in Tulsa Public Schools in the fall of 2006, we use a regression-discontinuity design to estimate treatment effects on average test-score percentiles. We combine these results with Chetty et al.’s findings to generate projected earnings effects and a partial cost–benefit analysis of the Tulsa pre-K program. To make our Tulsa estimates consistent with Chetty's results, we use a different test score metric than in previous research on Tulsa pre-K (Gormley, 2010, Gormley and Gayer, 2005, Gormley et al., 2005, Gormley et al., 2008). To allow for benefit–cost analysis by income group and half-day versus full-day pre-K, we go beyond previous Tulsa results in segmenting estimates by both income group and half-day versus full-day program. Our analysis suggests that the Tulsa pre-K program has substantial earnings benefits for each of the income (school lunch) and program-type (full-day versus half-day) subgroups examined. In each case, the benefit–cost ratio well exceeds 1, even though we only consider adult earnings benefits in this analysis. We also find that the percentage effects on expected future adult earnings are largest for lower-income children.
In the next section, we discuss the limitations of test scores for assessing program benefits and making comparisons across income categories. We then describe recent findings by Chetty et al., which suggest the shape of the relationship between test-score percentiles and adult earnings. We describe the Tulsa pre-K program. We also describe our regression-discontinuity estimation technique. Next, we present program effects using average test score percentile as our outcome. We then use these estimated program effects to generate projected adult earnings benefits for different income and program-type subgroups. These projected earnings benefits serve as the basis for a partial cost–benefit analysis of the Tulsa pre-K program. Finally, we discuss the limitations of our analysis and offer some concluding remarks.
Section snippets
Existing studies of preschool for different income groups
Should preschool be targeted at the disadvantaged, or universal? The answer depends on how benefits of preschool vary with a child's family income. Many benefits, such as higher adult earnings for former preschool participants, are long-term.
Long-term studies of preschool have focused on programs serving disadvantaged children, such as the Perry Preschool Program (e.g., Schweinhart et al., 2005), the Abecedarian Project (Campbell & Ramey, 2010), or the Chicago Child–Parent Center Program (
Overview
This section of the paper describe the Tulsa pre-K program and our data, outlines and defends our methodology for estimating test-score effects, and presents those test score effects.
Tulsa pre-K program
The Tulsa Public Schools pre-K program is a school-based, state-funded pre-K program for four-year-old children. Since 1998, Oklahoma's school districts have had the option of providing pre-K to all four-year-olds. Most school districts have participated, including Tulsa, the state's largest school district.
Overview
This section uses the estimated effects of Tulsa pre-K on the average test-score percentile for different income groups, combined with Chetty et al.’s estimates, to predict future earnings effects, which are then compared with program costs.
Predicted future earnings benefits
Even without much number-crunching, it can easily be seen that our Tulsa results and Chetty et al.’s results imply large forecasted effects on future earnings. The Chetty et al. results imply an annual earnings effect at ages 25–27 of around $73 per a
Conclusion
It is important to stress our findings’ limitations. First, we cannot be sure that kindergarten test scores in Oklahoma will translate into adult earnings in precisely the same way that kindergarten test scores in Tennessee have translated into adult earnings. Different job markets could have different returns to different skills. Long-term effects of better kindergarten class quality in Tennessee could be more or less durable than the long-term effects of a pre-K program in Tulsa. If so, we
References (46)
- et al.
Comparative benefit–cost analysis of the Abecedarian program and its policy implications
Economics of Education Review
(2007) - et al.
Methods for addressing missing data in psychiatric and developmental research
Journal of the American Academy of Child and Adolescent Psychiatry
(2005) - et al.
The rate of return to the high/scope Perry preschool program
Journal of Public Economics
(2010) - et al.
Later effects of preschool on low-risk children
Early Childhood Research Quarterly
(1989) - et al.
Regression discontinuity inference with specification error
Journal of Econometrics
(2008) - et al.
Does prekindergarten improve school preparation and performance?
Economics of Education Review
(2007) - et al.
Inside the pre-kindergarten door: Classroom climate and instructional time allocation in Tulsa's pre-k programs
Early Childhood Research Quarterly
(2009) One year of preschool intervention or two: Does it matter?
Early Childhood Research Quarterly
(1995)- et al.
Benefits and costs of investments in preschool education: Evidence from the Child–Parent Centers and related programs
Economics of Education Review
(2007) - et al.
The state of preschool 2011
(2012)