A commenter on this blog argues that “universal pre-k” is too costly because of the large estimated costs of providing full-day preschool for all 3 year olds and 4 year olds. She argues that “According to UPK advocates, “Universal pre-k” means full day UPK for all 3 and 4 year old children in the nation…”

I can’t speak for all UPK advocates, but by “universal pre-k”, I do not necessarily mean a full-day program for all 3- and 4-year-olds. The program may be half-day or full-day, or may be for both 3- and 4- year olds, or only for 4-year-olds. In Investing in Kids, the specific preschool program whose effects I estimate is a half-day program for four-year-olds only.

I should also note that in general, state pre-k programs, including “universal” programs, are mostly for 4-year-olds. For example, in the 2009 State of Preschool Yearbook by NIEER, 25% of the nation’s 4-year-olds are in state-funded pre-k, versus 4% of all 3-year-olds. Furthermore, of the state programs (with some states having more than one program), 10 state programs are half-day programs, 10 state programs are full-day programs, and 31 state programs leave this to local option. But total reported spending per child on pre-k averages $4711.  This suggests that the majority of students in state-funded pre-k are in half-day programs for 4-year-olds only. So, in practice, state-funded pre-k seems to usually mean a half-day program for 4-year-olds.

Some of the best long-term evidence for pre-k is on half-day programs. Both Perry Preschool and the Chicago Child-Parent Center program are half-day programs. For Perry Preschool, about 80% of the children participated for both ages 3 and 4, whereas 20% participated only for age 4. The experiment did not find significant differences in effects between two-year and one-year participants, but the sample size for one-year participants is tiny. For the CPC program, about half participated for both ages 3 and 4, and half participated only for age 4.  The differences in estimated effects between one-year and two-year participants were in favor of two-year participants, but two-year participation did not come close to “doubling” the effects of preschool.

In chapter 5 of Investing in Kids, I consider the benefits and costs of expanding from my baseline preschool program, a half-day program for 4-year-olds only, to either a full-day program or a program that also served 3-year-olds.  For going from a half-day program to a full-day program, the only rigorous study is by Robin, Frede, and Barnett (2006). I used their results for test score effects to calculate the benefits for state economic development of an expansion to full-day preschool, versus the costs.  I calculate that the measured benefits for state economic development are less than costs.

However,  the economic development benefits of going from half-day to full-day preschool may exceed costs if full-day preschool has effects on adult earnings that go beyond what one would expect based on test score effects. Such effects are plausible, as much of preschool’s benefits are due to effects on “soft skills”.  In addition, full-day preschool is likely to have other benefits that are not “economic development benefits”, such as reduced crime. Full-day preschool may also improve access to preschool over half-day preschool.

Therefore, benefits may exceed costs in going from a half-day to a full-day preschool program. But it seems unlikely that benefits go up in percentage terms as much as costs when expanding from a half-day preschool to a full-day preschool.

I also consider the benefits and costs of expanding from a program for 4-year-olds only, to a program that includes both 3-year-olds and 4-year-olds. I base these estimates on a study by Arthur Reynolds of the Chicago Child-Parent Center program.  Using these estimates of test score effects of one year of preschool, versus two years of preschool, I estimate the economic development benefits to a state economy of expanding services to 3-year-olds, versus the added costs. I calculate that the estimated benefits exceed the costs.

However, although benefits exceed costs, the percentage increase in benefits is not as great as the approximate doubling of costs when adding in 3-year-olds. So, the ratio of economic development benefits to costs of the expansion is less than the benefit-cost ratio for the original program for 4-year-olds only. This has implication for state budget choices. Suppose that we could get some additional funds for state-funded pre-k. Suppose we had a choice between using those funds to add additional four-year-olds, with similar characteristics to the four-year-olds currently being served,  to the state-funded program, versus adding age 3 preschool for some children who would have received preschool services at age 4.  These results imply that the benefits from serving more four-year-olds exceed the benefits from doubling preschool years for some students.

Therefore, the empirical result suggests we get the highest benefit-cost ratio from a high-quality half-day program for 4-year-olds only. There are probably net benefits from expanding to 3-year-olds, and benefits may also exceed costs from expanding to a full-day for 4-year-olds. But the benefits per dollar probably are somewhat lower for these expansions.

One way of explaining this pattern of results is that there may be what an economist would call “diminishing returns” to time in preschool. Children learn more with more time in preschool, but what they learn per additional hour may decline with more time.  This may be particularly true when going from a half-day to a full-day preschool program.  There may be limits to how much pre-k students can learn during a day.

These conclusions are for comparisons of adding services to children from similar backgrounds. These results may not hold when the tradeoff is between adding more 4-year-olds from higher-income families to a state’s preschool program, versus offering more years of service to children from lower-income families. This choice is a more difficult one.  The net benefits versus costs would depend on the particulars of the situation.

The evidence seems clear that even a half-day preschool program for four-year-olds can have large benefits. It may seem amazing that a program that provides perhaps 500 hours of services per child can have large long-term effects. As discussed in a previous post, these long-run effects probably occur because preschool’s effects on “soft skills” tend to appreciate over time. A child who develops better social skills and self-confidence will find those skills reinforced by their experiences. Success will lead to future success, and skills will beget skills.