Outcomes of early childhood education: Literature review
This literature review was commissioned by the Ministry of Education to provide policy makers with a synthesis of research that analyses the impact of early childhood education (ECE) for children and families.
Author(s): Linda Mitchell, Cathy Wylie and Margaret Carr, New Zealand Council for Educational Research.
Date Published: May 2008
This report is available as a download (please refer to the 'Downloads' inset box). To view the individual chapters please refer to the 'Sections' inset box.
Chapter 7: Economic returns from ECE
This review has focused on ECE outcomes for participating children and their parents. A third party interested in ECE outcomes is government. In this section, we describe the main findings of studies of economic returns from ECE for children, parents and government. We used Penn et al.'s (2006) definition of economic returns as referring to
...the outcomes for children (and for mothers of participants) to which a cost has been assigned, including long-term social integration or mental or physical health, rates of incarceration, remedial education, teenage pregnancy rates, employment and earnings (p. 1).
Cost benefit analysis was the main tool used to determine economic returns from ECE in the studies that we reviewed. Masse and Barnett (2003) defined cost-benefit analysis in this way:
As informed by economic theory, our perspective is that education is both a social good that confers immediate benefits and an investment good that confers personal and social benefits well into the future (Becker, 1964; Haveman & Wolfe, 1984). Benefit-cost analysis involves estimating the monetary values of streams of cost and benefits in order to measure the program's net value as a social investment (p. 6).
This requires a detailed estimation of programme costs and identification and estimation of the value of benefits or effects. Meadows (2001) set out the seven steps involved in cost benefit analyses:
- Measure the costs;
- Measure the effects;
- Estimate the monetary value of the effects;
- Adjust for the effects of inflation by placing all costs and benefits in terms of constant values;
- Discount future costs and benefits to take account of the opportunity costs of the use of the original resources;
- Identify the distribution of costs and benefits across different groups (children, their families, the local community and the wider public); and
- Undertake sensitivity analyses (p. 15).
Cost benefit analyses are limited by the range of benefits to which a cost can be assigned, and the length of follow up. For example, "quality of life" indicators such as wellbeing are valuable outcomes of ECE (Penn et al., 2006), but are not amenable to quantification in numerical or monetary terms. Pascal and Bertram (2001) argued wellbeing and happiness need to form an important part of cost benefit analysis. Long-term evaluations enable quantification of benefits into adulthood, but these studies are few. Cunha et al.'s (2005) model of the economics of investing in human capital, discussed in the Introduction, shows that early investment has to be followed up by later investment in order for the early investment to be productive.
Karoly et al. (2005) identified other methodological issues that influence levels of comparability between studies: methods of discounting dollars; whether outcomes are measured or projected; whether benefits should be weighted according to valued outcomes (they were not weighted in any of the studies reviewed); what costs are projected; and use of adjusted or non-adjusted effect sizes.
Economists (e.g., Cleveland & Krashinsky, 1998; Meadows, 2001) have also pointed out that cost-benefit analyses should include a measure of the benefits that would have been received if the resources had been devoted to the next best policy alternative. Is investment in ECE programmes a good investment compared with alternative investments? Only one study that we reviewed offered such an analysis.
This section draws predominantly on a review by Karoly, Kilburn and Cannon (2005) of cost benefit analyses of intervention programmes in terms of public expenditure in relation to children's participation in ECE, the EPPI-Centre systematic review (Penn et al., 2006) that covers three of the six studies used by Karoly et al., a cost benefit analysis of the Abecedarian Project by Masse and Barnett (2003), and a recent review with recommendations and cautions for policy makers (Barnett & Ackerman, 2006). We also consider two studies (Muller Kucera & Bauer, 2002; Pricewaterhouse Coopers, 2004) discussed in the recent OECD review of ECE in 20 countries (OECD, 2006) which used survey data to estimate revenue through taxation and/or GDP from parental use of childcare to support employment. Four further studies projected the monetary value from providing high-quality universal ECE (Cleveland & Krashinsky, 1998; Dickens, Sawhill, & Tebbs, 2006; Karoly & Biglow, 2005; Oppenheim & MacGregor, 2002).
We examine evidence on the benefits of ECE attendance and then give a synthesis of cost-benefit findings.
Employment-related benefits have been studied in U.S. intervention studies, studies using survey data and studies estimating costs and benefits of providing universal quality ECE. Outcomes can be measured for the mothers of children participating in the ECE programme and for participating children when they reach adulthood. Employment for mothers and programme participants in turn contributes to economic returns for government through gains in taxation.
|Study Type: Intervention Studies|
|Maternal Employment & Earnings Outcomes:|
Abecedarian study: higher paying jobs at child age 12, 15 and 21. Teen mothers self-supporting (E=70%, C=58%), teen mothers post-secondary training (E=46%, C=13%). Estimated higher earnings for Abecedarian mothers $3,750 per year compared with control group mothers. No other study costed maternal earnings outcomes directly.Infant Health and Development study: higher level of employment – average one month during the 3 years of project.No employment difference for Perry Preschool Project mothers.
|Cost Benefit to Govt:||-|
|Child Participant Employment & Earnings Outcomes in Adulthood:||Abecedarian: in skilled jobs at age 21 (E=70%, C=40%).Chicago Child-Parent Centres projections for lifetime earnings and tax revenues based on differences between participants and control group in high school completion age 21.Perry Preschool Project Employment rate age 19 (E=50%, C=32%), Employment rate at age 27 (E=71%, C=59%), Monthly earnings at age 27 (1993$) E=$1219, C=$766).Received public welfare at age 27. E<C; E=59%; C=80%|
|Cost Benefit to Govt:||Tax benefits: Abecedarian - $16,460 (projected),Chicago Child-Parent Centres-$3,300 (projected), and Perry Preschool Project ($6,566) (actual).Tax contribution differential for Perry Preschool children for each year up to age 40 and projected to age 65: $30,146 (male), $37,191 (female).|
|Study Type: Studies Using National Survey Data|
|Maternal Employment & Earnings Outcomes:||Outcomes in studies in Zurich (Muller Kucera & Bauer, 2002) and Quebec (Baker, Gruber, & Milligan, 2005) from provision and/or reduction in cost of ECE. In Zurich, estimated income effect CHF 44.1 million. In Quebec, labour force participation of married women rose by 7.7 percentage points.|
|Cost Benefit to Govt:||Zurich study: Cost CHF 18 million, additional tax revenues CHF 29 million.Quebec study: taxes from new labour supply after expansion near-free ECE covered 40% of new subsidies.|
|Child Participant Employment & Earnings Outcomes in Adulthood:||-|
|Cost Benefit to Govt:||-|
|Study Type: Study estimating costs and benefits of expanding universal quality ECE provision and paid parental leave and providing a home care allowance|
|Maternal Employment & Earnings Outcomes:||(Pricewaterhouse Coopers, 2004) study assumed increase in female employment rates towards rates in Sweden/Denmark, which offer universal early years provision and paid parental leave. Estimated average effect on maternal earnings 3 percent for those working fulltime, 1 percent for those working part-time.|
|Cost Benefit to Govt:||-|
|Child Participant Employment & Earnings Outcomes in Adulthood:||Estimated average effect on life-time earnings 3 percent for participants who benefit from additional early years education. Takes into account fact that there may be lower benefits for better-off children than for families in disadvantaged areas, but that nevertheless studies have shown benefits of good quality ECE for all children.|
|Cost Benefit to Govt:||Overall, outcomes of maternal employment contribute to rise of 1.1 percent GDP and from participants' lifetime earnings 0.4 percent of GDP. Total 1.5 percent. Note: subject to uncertainties – plausible range of economic benefits 1–2 percent GDP. Excludes important social benefits.Estimated cost to government around 2.2 percent GDP.Benefits would largely offset costs in monetary terms.|
Chapter six of this review discussed evidence that decreasing ECE costs for parents and offering accessible ECE services can encourage maternal employment. This section examines evidence of outcomes of maternal employment.
Three intervention studies, the Abecedarian Project, the Infant Health and Development Project, and the Perry Preschool Project compared the employment status of mothers of participating children with the employment status of control groups. These were summarised by Masse and Barnett (2003) and Karoly et al. (1998). Two of these found employment advantages for participating mothers. Abecedarian mothers had higher paying jobs than the control group mothers at child participant ages of 12, 15 and 21. IHDP mothers had a higher level of employment (one month on average) compared with the control group during the three years of the programme (Karoly et al., 1998).
Studies using national survey data:
Two studies, in Quebec and Zurich respectively (Baker et al., 2005; Muller Kucera & Bauer, 2002), showed increased revenue from taxation generated through availability of child care and maternal employment. These were not direct studies of ECE experience and outcomes.
The Zurich study (Muller Kucera & Bauer, 2002) was based on individual household data comparing the weekly working hours and composition of households that used child care in 1999 with comparable households not using child care. It evaluated and quantified costs and benefits of use of childcare. The study found that the city's public investment of CHF 18 million annually in child care is offset by at least CHF 29 million of additional tax revenues and reduced public spending on social aid. Use of child care was estimated to result in the rate of maternal hours worked almost doubling, especially for single-headed households with one or more children.
Baker et al. (2005) found employment expanded after the introduction of near free child care for 4-year-olds and then younger age groups from 1997 in Quebec, but at a slower rate than provision of ECE places. This was partly accounted for by parents who had already been employed moving from informal to formal ECE (i.e. they were already in the labour market). The taxes generated by the new labour supply covered about 40 percent of the cost of the child care subsidies, less than the Zurich study. This difference could be accounted for by the different measures used — taxes from the entire labour supply using childcare in Zurich, and taxes from the new labour supply resulting from expansion of childcare in Quebec.
Study estimating costs and benefits of expanding universal quality ECE:
In Britain, PricewaterhouseCoopers (2004) estimated that expansion of free ECE services for 2, 3, and 4 year-olds in the U.K., and improvement in the quality of ECE services through upskilling the workforce, would contribute to a rise in GDP from between 1 percent and 2 percent through higher rates of maternal employment and increased lifetime employment. This rise in GDP would broadly offset incremental costs to government. PricewaterhouseCoopers did not include estimates of social benefits, such as any impact on income distribution, child poverty, remedial education, improved health or lower crime rates: it is likely that such estimates would substantially increase the value of benefits at least for disadvantaged children, as was found in U.S. intervention projects discussed above. These are estimates: the authors expressed some concerns about the reliability of available cost data and uncertainties around precise cost estimates.
No differences in maternal employment were found for Perry Preschool Project mothers compared with their control group. One difference between the Perry Preschool Project and other intervention studies was that the Perry Preschool Project was part-time; the others offered full-time year round childcare (Masse & Barnett, 2003). Barnett and Ackerman (2006) noted that services operating only part-day can cause potential constraints to employment. Likewise, the operating times of the Perry Preschool Project may have been an employment constraint.
Value to government of maternal employment
Only the Abecedarian Project directly measured the value of employment-related outcomes arising from the employment of mothers of participants in the project. The higher earnings for these mothers were estimated to be $3,750 per year compared with control group mothers (Masse & Barnett, 2003).
The PricewaterhouseCoopers study assumed the average effect on maternal earnings of a "vision" of providing good quality free ECE for 2, 3 and 4 year olds in the U.K., wrap around care for school age children, a home care allowance, and paid parental leave would be three percent for those working fulltime while their children are aged 1–7 years, and one percent for those working part-time. This was assumed on the basis that a movement would be seen towards the average Swedish and Danish employment rates (where ECE, paid parental leave and wrap around care are similar to the "vision" for the U.K. that was analysed) and of differences between the U.K. and these countries in work culture. It also took into account the effect of avoiding long career breaks for mothers. Effects on maternal employment of extended parental leave and enhanced childcare for school age children were assumed to cancel each other out on the basis that reductions in those working because of paid parental leave and the home care allowance would be offset by greater employment of mothers of primary school children. Short-term, the share of extra economic benefits was calculated at 0.9 percent of GDP, based on Office of National Statistics figures on the proportion of gross income paid in direct and indirect tax. Long-term, extra benefits from parental lifetime earnings were projected to be a further 0.2 percent of GDP.
Differences for population groups
The only studies assessing participants' gains in maternal employment directly were the three intervention studies, which are targeted to low income groups. Therefore no comparisons across different groups were possible. However, the PricewaterhouseCoopers study did estimate the employment rate for women from all income groups, not just low income groups, would rise.
Participant employment benefits (lifetime earnings)
Higher educational attainment for the children participating in ECE may be an economic benefit when they reach adulthood because of the extra productivity of the educated person. The monetary value of taxation offers a measure of this effect. However, the costs of attaining higher levels of attainment should also be taken into account.
Only the Perry Preschool Project and Abecedarian Project measured participants' employment related outcomes directly.
Nores, Belfield, Barnett and Schweinhart (2005) analysed data on outcomes for Perry Preschool Project participants at the age of 40, compared with the control group. Earnings profiles were calculated from self-reported data about the current job and work history. There were clear earnings gains for both male and female participants compared with control group participants who did not experience the intervention. The differential in tax contributions for male participants was + $30,146 and for female participants was +$37,191 over the course of a working life. The earnings were calculated for each year up to age 40 and projected to age 65. Fringe benefits worth 29.2 percent of salary were included.
When they were aged 21, 70 percent of the Abecedarian participants were engaged in skilled jobs (above entry level positions) compared with 40 percent of the control group (Ramey & Ramey, 2004).
Study estimating costs and benefits of expanding universal quality ECE:
In Britain, PricewaterhouseCoopers (2004) estimated the earnings benefit for participants from expansion of free ECE services for 2, 3, and 4 year-olds in the U.K. and improvement in the quality of services through improving ECE service staff qualifications. They estimated that lifetime earnings for those children who benefit from additional early years education and care would be on average three percent higher than they would have been without this provision. The estimated economic benefits were 0.4 percent of GDP from the participant's lifetime earnings.
Differences for population groups
The only studies assessing participants' employment gains directly were for children in targeted intervention studies. The differential for earnings was higher between females in the Perry Preschool Project and control group females than was the differential for males.
Employment-related benefits not costed
Other potential employment-related benefits from investing in universal ECE that have not been costed were described by Karoly and Bigelow (2005): labour force recruitment; participation rates among the working age population; workforce performance; and long-term economic growth, competitiveness, and equality.
Three intervention studies, the Abecedarian Project, the Infant Health and Development Project, and the Perry Preschool Project measured educational benefits of their interventions in terms of the proportion of children in the intervention and control groups who were retained in grade and/or required special education assistance. These studies also measured differences in rates of high school completion, high school graduation, and/or post-secondary education. These constitute a cost to the taxpayer, but the cost is counterbalanced by tax returns from participants having greater lifetime earnings.
In all three studies, a smaller proportion of the experimental group were retained in grade or required special education assistance compared with the control group. A higher proportion completed high school, graduated from high school, or gained a post-secondary qualification. The table below is from Penn et al., 2006, p. 18.
|By age 9||E<C; E=8%, C=9%|
|(yrs) through age 14||E<C; E=0.6, C=0.9|
|(yrs) through age 18||E<C; E=0.7; C=1.5|
|(yrs) retained in grade by age 15||E<C; E=31%, E=55%|
|By grade 9 (age 14)||E<C; E=25%, C=48%|
|Time spent in by age 19||E<C; E=16%, C=28%|
|Years spent in to age 27||E<C; E=1.1, C=2.8|
|Completion||E<C; E=49%, C=39%Δ||E<C; E=50%; C=39%||E<C; E=67%, C=51%|
|Graduation to age 27||E<C; E=66%, C=45%|
|Post-secondary||E<C; E=33%, C=28%†||E<C; E=36%, C=13%§|
The size of savings was summarised by Penn et al. (2006). Different data was collected for each study which may explain some differences in cost savings.
Reduction in education services|
(special education, grade retention)
Reductions in criminal activities
Lower public spending on justice is another outcome that has been attributed to participation in good quality ECE. The table below (from Penn et al., 2006, p. 20) compares programme outcomes for criminal offending for three U.S. intervention studies: the Perry High Scope, Chicago Child-Parent Centres, and Abecedarian studies comparing the experimental group with the control group. The incidence of crime for both experimental and control groups is high in the Perry High Scope study even though there are differences between the groups.
|Ever arrested by age 27||E<C; E=57%, C=69%|
|Lifetime arrests through age 27||E<C; E=2.3, C=4.6|
|Delinquency rate at age 13-14||E<C|
|Crime rate age 16-21||E=C|
|Juvenile court petitions through age 17||E<C; E=16%, C=26%|
|Violent offences by age 17||E<C; E=9, C=15.3|
|Child abuse and neglect||E<C; E=5%, C=10.3%|
The value of savings in criminal justice costs and reduction in tangible losses to crime victims are estimated as highest for the Perry High Scope Project and nil for the Abecedarian Project. Different measures of outcomes and different measures to estimate savings were used for the three studies.
Perry High Scope1
|Reduction in criminal justice cost||$10,195||$6,085|
|Reduction in tangible losses to crime victims||$10,690||$4,859|
Total benefits in relation to costs
Table 11 presents the results of cost benefit analyses for five intervention programmes that included ECE, synthesised by Karoly et al. (2005, pp. 109–111). As Karoly et al. pointed out, programmes with longer follow-up periods (the Abecedarian, Chicago Child-Parent Centre, and Perry Preschool) tended to have higher estimates of total benefits. This is because the longer term studies could measure outcomes at adulthood that were more readily translated into dollar benefits, such as higher earnings (benefiting the participant and the taxpayer), reduced reliance on welfare programmes (a benefit to taxpayers), and reduced contact with the criminal justice system (a benefit to taxpayers and other members of society). This is illustrated by the Perry Preschool Project benefit-cost ratios which increased with participants' age (a return of just over $17 for every dollar invested at age 40).
The short-term intervention study, Infant Health and Development Project (IHDP), showed no reduction in government costs at age 8. The IHDP used test scores as its child outcome measure to which a cost was assigned. The Cognitive outcomes section of this review has noted that the IHDP intervention finished when children were aged 3, and the lighter birth weight children who did not benefit may have needed a longer period of intervention or have been neurologically impaired. It is also the most expensive ECE programme analysed ($49,021 per child). We have not included the Comprehensive Child Development Programme findings also reported by Karoly et al. (2005) because of flaws in the experimental design for the evaluation (see for example, Goodson, Layzer, St Pierre, Bernstein, & Lopez, 2000).
(last follow-up age)
|Follow-up during elementary school years|
|IHDP (age 8)||49,021||0||0||0||0||-49,021||-|
|Follow-up to early adulthood|
|Abecedarian (age 21)||42,871||na||na||na||138,635||95,764||3.23|
|Chicago CPC (age 21)||6,913||22,715||19,985||6,637||49,337||42,424||7.14|
Perry Preschool (Age 27)|
(excl. intangible crime costs)
Perry Preschool (Age27)|
(incl. intangible crime costs)
Good Quality ECE for Low|
Income 3 & 4 year-olds5
|Follow-up to middle adulthood|
|Perry Preschool (age 40)||14,830||61,866||191,288||191,288||253,154||238,324||17.07|
Studies estimating costs and benefits of expanding universal quality ECE:
Four studies analysed the costs and benefits of providing universal, high quality early childhood education for all children or high quality early childhood education targeted to low income families. In all four studies, projected benefits exceeded the costs. Cost: benefit ratios were 1:2 or greater in three studies, and the effect on GDP in the fourth study was 1.34 to 4.02 percent.
A Canadian study, carried out by economists from the University of Toronto (Cleveland & Krashinsky, 1998), used research undertaken during the late 1980s and 1990s on children's academic success and mothers' employment in the labour force to estimate benefits and costs of providing high-quality ECE for all preschoolers in Canada aged between 2 and 5. Using conservative estimates, they calculated that providing high-quality ECE for all these preschoolers in Canada would cost approximately $5.3 billion per year and the value of additional benefits to children and parents would be about $10.6 billion per year. Benefits to children through participation and to mothers from employment were about equal. For society, the benefits arise from increased productivity, higher generation of tax revenues, decreased social assistance and health costs, and improved citizenship.
The economics of paid parental leave versus infant ECE programmes are relevant to discussion of economic impacts from parental workforce participation. High quality ECE for babies is equally, or possibly more, expensive than paid parental leave, as Cleveland and Krashinsky (1998) note, and vital bonding occurs between parents and children in the early months. Access to paid parental leave significantly improves infant health. Ruhm (1998), for example, has provided evidence of a strong negative relationship between duration of parental leave and post-neonatal mortality between the first and fifth birthday. Cleveland and Krashinsky (1998) argued that key elements of child care policy should include universal provision of good quality ECE for all children aged 2 to 5 and a well-designed family benefit and parental leave policy20 for parents with babies. On the other hand, there may be a negative cost for the labour market outcomes of mothers using extended paid parental leave. Parents are likely to return to work earlier and place children in childcare where there are generous childcare subsidies and meagre paid parental leave (Waldfogel, 2001). Use of parental leave is associated with increases in women's employment long term, but is also associated with reductions in their relative earnings (Ruhm, 1998).
In the U.S., Oppenheim and MacGregor (2002) used actual average per-child costs of the Head Start programme to calculate costs of providing high-quality preschool education to all 3 and 4 year olds from low-income families in the United States in general, and especially four states (Arkansas, Louisiana, Mississippi, and Texas). They used information from intervention studies on returns associated with school measures (including higher levels of education, reduced grade retention, and need for special education), crime, earnings, taxation, welfare, and employment to estimate a cost: benefit ratio of 1:2.4 overall.
Karoly and Bigelow (2005) used results from the Chicago Child–Parent Centre programme to estimate the cost benefits of providing good quality ECE21 to a single cohort of California's 4-year-olds. In making their estimates they projected greater benefits for disadvantaged children, lesser benefits for advantaged children, and greater benefits for children attending no ECE programme and moving to an ECE programme than for those already in ECE. They also factored in the kind of preschool a child would have attended without a universal programme. Their estimates are conservative. They include fewer grade retentions, less need for special education, reduction in child abuse or neglect, reduced costs to the criminal justice system, and increased earnings by parents (in the shorter term) and children (in the longer term). They do not include improved labour market outcomes for parents, improved health and wellbeing of children, reduced reliance on public welfare, lower crime, neglect, child abuse, and "the intergenerational transmission of favourable benefits" (Karoly & Bigelow, 2005, p. xxvi). They found that California would expect to gain between $2 and $4 for every $1 invested.
Dickens et al. (2006) used results from the Perry Preschool programme to determine the magnitude of the effect of high-quality universal ECE for 3- and 4-year-olds on Gross Domestic Product. They adjusted the Perry Preschool programme results for probable attenuation from delivering the programme to less disadvantaged children, and estimated effects based on a range of assumptions. They estimated that by 2080, investment in high-quality ECE could raise the U.S. GDP by between 1.34 and 4.02 percent. This analysis concentrates on the fact that the Perry Preschool Program led to participants having an additional 0.9 years of education compared with non-participants. The analysis did not include other benefits found in the Perry Preschool Program, such as reduced rates of teenage pregnancy and lower rates of criminal activity. The models used to project GDP took account of expected less than universal take-up levels of enrolment and the differing quality of ECE currently received. They also included roll-on benefit or dynamic feedback in the increase in physical and human capital over time.
Assessment of findings
Generalisability from longitudinal studies
Evidence for long term actual cost benefits of early childhood education emerged from three longitudinal intervention studies, the Abecedarian, Perry Preschool and Chicago Child-Parent Centre projects. Penn et al. (2006) raise issues about the generalisability of these three long-term intervention studies given their samples of disadvantaged children from low-income African American families in the U.S., and the context of the times when they were initiated (1960s and 1970s). They question the relevance of studies outside the U.S.
New Zealand also has children from low income families, with parents having low levels of formal qualification and relatively high levels of unemployment (and welfare assistance), and who are less likely than other children to participate in centre-based ECE. However, New Zealand's cultural context is quite different from the U.S. and the context of the times is also different. In addition, the model of schooling, which forms the basis for calculations of costs and benefits, is different from schooling in New Zealand. For example, New Zealand does not hold children back a level (grade retention). In two of the intervention studies the benefits from reduction in costs of crime are calculated to be very high. Incarceration rates are higher in the U.S. than any industrialised country, at 701 per 100,000 people. New Zealand's rate of 155 places it seventh highest in the OECD (Statistics New Zealand, 2005). We would therefore expect some lesser gains than the U.S. from reduction in costs of crimes.
Economic data and analytic methods
Differences in the measures of costs and benefits are some of the reasons for variations in cost-benefit ratios in the studies reviewed. Programmes with longer term follow-ups tend to have higher estimates of benefits and cost-benefit ratio because in adulthood some benefits are easier to measure. The studies with longer term follow ups were also intensive targeted interventions. As we have also discussed, there may be many intangible benefits of early childhood education which are difficult to cost. These are reasons why it is important to regard cost-benefit analyses as only estimates. Penn et al. (2006) have argued that headline figures, such as a saving of $7 for every dollar spent, should be treated as tentative. "The evidence is suggestive rather than incontrovertible" (p. 28).
At the same time, Karoly, Kilburn and Cannon (2005) have argued that cost-benefit analyses tend to provide conservative estimates because data limitations preclude measuring the economic benefits associated with all potential benefits. They noted linkages between the level and growth of human capital and the rate of economic growth:
Investments in early childhood interventions, to the extent that they raise eventual educational attainment and other measures that are valued in the workplace, can help raise the overall skill level in the economy and contribute to the economic success of the country (p. 118).
Such macroeconomic benefits were not costed in most of the studies reviewed.
Barnett and Ackerman (2006) summarised the extent to which variations in the populations served, programme characteristics, and the educational and community context of the ECE service alter the magnitude of educational benefits. Their analysis indicated that gains are highest for children who are poor, but children from middle-income homes also benefit, creating a potential for much larger net benefits from universal programmes. Contrasts in programme quality are likely to explain gaps between potential returns and actual performance of programmes. Barnett and Ackerman also argued that universal programmes within a community will be expected to produce larger gains because of previously unrealised peer effects, i.e. "if everyone in a classroom has experienced preschool, classroom climate will change, median ability will rise, and dispersion in ability will narrow, with those at the bottom gaining most" (Barnett & Ackerman, 2006, p. 94). In addition, these authors highlighted benefits of access to acceptable ECE for maternal employment. Constraints to realising this benefit are the programme schedule (times and hours) and cost. They cited an analysis by Blau (2001) estimating full government funding of child care would result in a 9.5–10 percent increase in overall employment.
Barnett and Ackerman (2006) proposed the following reasons why investing in universal ECE is likely to be more efficient than targeting funding to disadvantaged children:
- Targeting is inaccurate, especially in relation to a status such as maternal employment and poverty that can change fairly often. Costs associated with administering and monitoring targeting can be high.
- Substantial benefits to children's learning extend to children from middle and higher socioeconomic backgrounds, not just low socioeconomic backgrounds. Peer effects from enrolling children from a variety of socioeconomic backgrounds are positive.
- Free public ECE offered universally may increase the child care options of those not in the labour force to participate in employment thereby providing further benefits. Timetabling and costs are key issues in providing support for employment.
These authors also discussed how to enhance the benefits of current services:
- Large additional gains at modest additional costs can be achieved by improving quality in services that are of low-quality.
- The ECE goals should include a range of child outcomes, or results will be inefficient. These should include cognitive and non-cognitive outcomes, not just academic outcomes.
- Broader policies, such as social, economic, and school policies affect economic returns from ECE. These need to be considered alongside ECE policies.
Cleveland (2006) also points to the very positive impacts of high-quality ECE on cognitive and language development from other U.S. studies, Andersson's (1989, 1992) Swedish studies, and the English EPPE study for children from across the socio-economic range. All the studies he cited have been included in earlier chapters in this review, as well as additional studies providing further substantial evidence of positive impacts of ECE. Cleveland argued that evidence from what we know about good quality statistical studies should be included in cost benefit analyses. He argued that it is reasonable to infer, given the evidence of positive outcomes from ECE participation for children from all socio-economic levels, that there are likely to be economic benefits from universal programmes in other countries. The studies reviewed provide evidence that supports this inference.
Applicability of benefits to New Zealand
The government savings and benefits most commonly measured in the general population studies and the targeted intervention studies would also apply in New Zealand:
- increased maternal and participant employment resulting in higher income tax payments (from the participating parents in the short term and children from low income families in the long term);
- reduced welfare assistance (for the parent in the short term and children in the long term), including unemployment benefits;
- reduced costs for special education interventions; and
- reduced burden in the criminal justice system.
The same is true for those often listed, but less often measured:
- reduced property loss, pain, injury and death from a reduced crime rate;
- reduced rate of child abuse and neglect;
- reduced need for late-childhood interventions, which are more costly and less effective;
- life-long improvement in level of cognitive, attitudinal, and achievement skills;
- peer group benefits if more children start school with improved behaviour in the classroom and better skills (cognitive and attitudinal);
- improved nutrition and health, reducing healthcare costs (including reduced substance abuse and teen pregnancy);
- "multiplier" effects on those around the children, and the next generation; and
- increase in GDP.
The rates of return derived vary considerably, depending on the assumptions used about the costs, benefits, and varying rate of benefit across the whole population. However, even the most conservative estimates show cost-savings benefits for providing high-quality ECE for all children, whether they include all the benefits listed above, or merely look at the effect of improved qualifications on the GDP as a whole.
International evidence demonstrates that investing in good quality ECE can bring cost savings and benefits to governments and economies as well as to children and families. Although cost benefit analyses measure only some quantifiable outcomes, these find that through provision of good quality ECE services, employment and tax revenues are increased, and savings are generated in educational and social expenditure.
Economists have cautioned that social, economic, and school policies also affect ECE cost benefit returns.
Good quality ECE has greater benefits for children from low socioeconomic families, but children from middle and high socioeconomic families also gain, as our earlier chapters have demonstrated. Economists, using conservative estimates of benefits based on recent relevant research evidence of ECE effects for children from across the socio-economic spectrum, have projected cost benefits of offering universal high-quality ECE for 2-, 3- and/or 4-year-olds. Most of the economic evaluations of ECE programmes have shown that benefits of public spending exceed the costs. Gains are not realised or are not as great if the ECE is of poor quality. Without considering the opportunity costs of that spending, comparing the investment to other types of early years intervention or alternative policy options, these evaluation findings tend to suggest that public spending for ECE programmes will result in good returns in terms of maternal employment, higher levels of the participant's lifetime earnings, reductions in usage of special education services, lesser criminal activity and reduced use of social services that are expected to have a flow-on effect to the economy.
- Sweden's policy, which Cleveland and Krashinsky favour, is for 16 months of publicly funded paid parental leave divided between mother and father.
- Programme is voluntary for all age eligible 4-year-olds, maximum class size 20, staff:child ratio 1:10, ECE qualified teaching staff.