In 2005 Florida introduced Voluntary Pre-Kindergarten (VPK), a free, universal preschool initiative. Florida’s implementation of the VPK program represents the most rapid expansion of state-funded preschool in the United States. In its first year, the VPK program served nearly half of the four-year-olds in the state. Nine years later the program serves more than 80 percent, and is the largest state-funded preschool program nationwide.

Proponents of VPK touted preschool’s promise to increase student achievement and eliminate persistent achievement gaps. With the program in its ninth year, there has been no rigorous evidence released testing whether VPK has delivered on this promise. Given the large investments that states are making into their early childhood programs and that those investments are typically argued for by claiming they would pay dividends later in life, it is essential to examine whether participation is having a lasting impact as children progress through school. What is currently known is derived from simple descriptive analyses comparing outcomes for participants to those of non-participants. These results suggest the positive “effects” observed at kindergarten entry (readiness rates have been 12-15 percentage points higher for participants) are still detectable at third grade albeit meaningfully reduced (scores on statewide assessments are higher among participants and among VPK completers than non-completers). Of course, evaluating the effectiveness of the VPK program by such simple comparisons is problematic given the bias introduced by non-random self-selection into preschool.

Our study provides such evidence for Florida’s VPK program by employing instrumental variable estimation to examine non-cognitive outcomes between kindergarten and third grade (attendance, exceptionality classification, and grade-level retention) and third grade performance on Florida’s statewide standardized reading and mathematics assessments.

We use two instruments to isolate the exogenous portion in the variation in participation. The first instrument is an indicator equal to one if the student reached the age of four after VPK was implemented. This pre-post instrumental variable uses the state-level shock in the availability of state-subsidized preschool between 2004-05 and 2005-06 to identify the exogenous variation in public preschool participation. The second instrument exploits the steady year-to-year expansion of the VPK programs within local communities as new providers entered the VPK marketplace.

Our analysis leverages longitudinal, student-level data obtained from Florida’s K-20 Education Data Warehouse. With these data we observe all students enrolled in any of the state’s public schools over a twelve-year period between 2001-02 and 2012-13. We identified eight cohorts of four-year-olds (four prior to VPK and four after) and have followed them as they move into kindergarten and progress through the third grade.

Our findings not only have direct implications for Florida but also have broader relevance. Rather than speaking to the impacts of a highly-regulated and expensive early childhood intervention, our study measures the effects of the full-range of preschool programs in Florida. The results are particularly informative to states trying to expand their existing early childhood infrastructure to provide universal access.