Evaluating the Impacts of Federal Subsidies for Early Childhood Education and Care

In recent years, there has been growing interest in policies that reduce out-of-pocket costs for early childhood education and care (ECEC) through subsidies or direct public provision. In addition to state- and local-level initiatives, federal proposals to subsidize care have been introduced as legislation or enacted temporarily in the last half-decade.

Assessing the impacts of these policies on both families and the federal budget is a challenge. Prices for paid formal care (like many forms of center-based care) are determined in part by supply and demand. Policies that increase demand for these forms of care by lowering families’ costs may lead to price increases if the ECEC sector is limited in its ability to expand to meet the new demand. These increases would then translate to higher out-of-pocket costs for unsubsidized families and greater spending on subsidies for eligible families by the government. At the same time, shifts in prices may lead to families switching from one form of care to another—decisions that also affect whether and how much parents work, which in turn influences tax revenues. A realistic model of the ECEC sector needs to account for all of these dynamics.

To that end, The Budget Lab (TBL) has developed a new open-source equilibrium microsimulation model for analyzing ECEC policy. The model is designed to capture these dynamics in the nationwide market for ECEC for children from birth through age 4, allowing us to estimate how new federal policy initiatives would affect:

• patterns of childcare usage and parental employment, which are closely related decisions for families;
• cost burdens for different types of families, accounting for changes in subsidy eligibility and market prices;
• the prices that childcare providers charge;
• the wages earned by care workers;
• the federal budget, including feedback effects on tax revenues from changes in parental employment; and
• children’s future economic outcomes, based on how a policy reform induces children to move between ECEC options with different estimated impacts on their future earnings.

TBL’s model is a structural equilibrium microsimulation model of parents’ decisions over employment and childcare (on the demand side) and of ECEC providers’ ability to meet this demand (on the supply side). Family-level demand for different care alternatives aggregates to market demand for different forms of ECEC (for example, center-based care vs. home-based care), and provider prices and wages adjust to clear the market. Policy reforms can shift the equilibrium, and the model calculates how those changes flow through to fiscal and distributional effects. Importantly, the model accounts for the fact that parents have many reasons for their choice of care for their children—such as their assessment of the type of care, geographic proximity, cost, convenience, and cultural fit—and does not assume parents will necessarily prefer one dimension of care over another.

Our model allows for parents to make different decisions at different child ages and different decisions for multiple children in the same household, including both employment hours (full-time vs. part-time) and forms of childcare. In other words, we put no restrictions whatsoever on the combination of work and care decisions that different parents can make at different times.  Full documentation for The Budget Lab’s ECEC model can be found on TBL’s website, with complete code available on GitHub.1

In this report, we use the model to analyze five illustrative policies for subsidizing ECEC, designed to highlight the tradeoffs policymakers face when crafting programs to increase access to affordable care:

1. Universal Subsidy: offer free childcare for low-income families and cap out-of-pocket childcare expenses for middle- and high-income families.
2. Income-Limited Subsidy: offer free childcare for low-income families and cap out-of-pocket childcare expenses for middle- and upper middle-income families, with high-income families ineligible for subsidies.
3. Expanded Care Credit: expand the Child and Dependent Care Tax Credit (CDCTC) to cover a greater share of childcare expenses and to allow low-income families to benefit fully from the credit.
4. Provider Tax Credit: partially subsidize labor costs for ECEC providers, making care provision at lower market prices financially viable.
5. Child Allowance: provide $1,000 per child for all families, regardless of childcare enrollment, parent employment status, or income.

Our choice of these five policies is meant to highlight different dynamics present in TBL’s model for a broad set of ECEC subsidy and tax proposals, and is not meant to be exhaustive of potential new policies. Other policies that are similar in structure to those we analyze but with greater or lesser generosity, or policies at similar levels of generosity but different structure, would generate different results that may not necessarily scale dollar-for-dollar with our findings. Because we focus specifically on subsidies in this report, we also do not explicitly model policies intended solely to raise the relatively low wages of workers in the ECEC sector, which has been an important focus of policy discussions. Such a policy would have implications for both costs and the quality of childcare (and thus impacts on affected children), and is one area for future analysis.

It is also important to note that TBL’s model cannot necessarily capture the complete range of effects that changes in ECEC policies can have on children, their families, and the federal budget. For example, and as discussed below, higher earnings for childcare providers can have positive effects on children’s outcomes through lower turnover (Grunewald, Nunn, and Palmer, 2022), a channel not directly captured by our model. Additionally, the estimates in this report do not include savings to the federal government from potential decreases in utilization of public assistance programs (such as Medicaid and SNAP) or from decreases in criminal justice involvement, nor do they include effects on state and local spending and revenues. In future versions of TBL’s ECEC model, we hope to expand its scope to include effects such as these.

Table 1 previews our findings.

This report is the first in a series of planned reports from The Budget Lab analyzing policies intended to help families of young children, including paid family and medical leave (PFML) and pre-K expansion, and comparing the relative impacts of each set of policies on outcomes for children, families, and the federal budget on a dollar-for-dollar basis.

In this report, we model five federal policy reforms designed to illustrate a range of approaches available to federal policymakers. These options vary in delivery mechanisms, progressivity, and treatment of different kinds of ECEC arrangements.

We describe each policy briefly below, with full details on exact policy parameters in the Appendix.

• Among these options, the Universal Subsidy and Income-Limited Subsidy policies would be the most generous for lower-income families.
  • For families of this size earning below roughly $70,000, each policy would cover 100% of ECEC costs, so the subsidy value is equal to the cost of care.
  • As income rises, discrete jumps in the subsidy rate would create benefit “cliffs”—ranges of income where small increases in income lead to large reductions in subsidies. (Within these subsidy rate ranges, there is an automatic income-based phase-out of the subsidy for a given price of care.)
  • The Income-Limited Subsidy option would create a larger cliff for families of this size when income nears about $230,000, at which point the family would lose subsidy eligibility entirely.
• The Expanded Care Credit line shows the incremental benefit of the expansion above and beyond the relief provided for under current law. Subsidies, which are largest at low incomes, phase down to lower levels at higher incomes.
• The Child Allowance, which is invariant to ECEC arrangement and does not scale with care costs, would provide $1,000 at all income levels for this single-child family.

Figure 1 above considers how subsidy generosity varies with family income for a given price of care. But some of these policies also differ in how much subsidy values change with the price of care, all else equal. For example, the Universal Subsidy covers all care costs above a fixed percentage of family income regardless of the size of those costs. In contrast, the Expanded Care Credit caps subsidy values at $4,000—regardless of whether care costs $5,000, $10,000 or $30,000. This means that, unlike the former policy, the latter policy does not automatically generate larger subsidies when care costs rise.2

Figure 2 highlights this dynamic by plotting the change in subsidy values by annual care cost. It shows these relationships at three different levels of income.

• The Universal Subsidy and Income-Limited Subsidy grow dollar-for-dollar with care costs above each family’s copay. Because copays are set as a fixed percentage of income, higher-earning families face a larger threshold of care costs before subsidies begin. For example, a family earning $200,000 must spend over $14,000 on care before receiving any subsidy, while a family earning $50,000 faces no copay at all.
• The Expanded Care Credit grows with care costs only up to its $8,000 expense cap, beyond which the subsidy is flat.
• The Child Allowance, a fixed payment defined in dollars, does not respond to higher care costs.

Table 2 compares the policy options across design-choice dimensions: the degree to which benefits are tailored based on need (“income targeting”); the extent to which fiscal costs automatically grow with higher care costs (“fiscal cost control mechanisms”); the rules governing which kinds of care are eligible for subsidy (“ECEC type eligibility”); and how the policy is implemented (“delivery mechanism”).

Each policy creates distinct incentives for families to place their children in different forms of ECEC, which in turn has implications for care worker wages, the prices families pay for childcare, and for long-run effects on children’s earnings in adulthood. Below, we report estimated impacts on enrollment patterns, classifying children as being primarily enrolled in one of four types of ECEC arrangement:

• Center-based care: care provided at a center-based facility, including both paid programs (like private daycares) and free programs (like Head Start or public pre-K).
• Home-based care: care provided in home-based (often called “family-based’) daycare programs.
• Other care: all remaining non-parental care arrangements, including regular care provided by nannies or relatives.
• Parent-only care: no regular attendance in any non-parental care arrangement.

Figure 3 shows ECEC enrollment patterns under baseline and under each reform scenario. We show results for 2030, since we assume—given the potentially large adjustments to ECEC enrollment, provider hiring, and other supply constraints—that the policy-induced changes to ECEC that we estimate will take multiple years to phase in.

• Our baseline estimates suggest that just over 40 percent of children ages 0-4 are not enrolled in any regular non-parental care arrangement. Just under one-quarter of children are enrolled in center-based care and just over one-quarter are in the “other” category. The remaining fraction of families utilizes home-based care.
• Direct subsidy policies (Universal Subsidy and Income-Limited Subsidy) would produce large shifts into formal care. We estimate that the fraction of children in center-based care would grow from less than one quarter to more than one third—an increase of almost 50 percent—driven by shifts out of both parent-only and other care arrangements.
  • This large shift is driven by policy design. Because each policy caps out-of-pocket expenses, families are largely insulated from higher sticker prices, making more-expensive arrangements like center-based care relatively more attractive.
• The Expanded Care Credit would also lead to families choosing less parent-only care and more center-based care. But unlike under the two direct subsidy policies, children would move into the “other” category, not out of it.
  • This difference is due to two factors. First, while the Universal Subsidy and Income-Limited Subsidy options restrict eligibility to formal care provided in centers or homes, care-type eligibility under the CDCTC is broader and includes options like nannies. Second, the expense cap under the CDCTC encourages families to choose cheaper options.
• The Provider Tax Credit policy works by subsidizing the labor costs of formal ECEC providers. Some of these cost savings would be passed on to consumers in the form of lower sticker prices for center- and home-based care (see the section below titled “Effects on ECEC Prices and Care Worker Wages” for more details). This translates to a modest enrollment shift into these forms of care, with outflows from parent-only and other forms of care.
• The main channel by which policy changes care arrangements is by altering relative prices across forms of care. Because the Child Allowance delivers identical payments regardless of which arrangement families choose, it leaves these relative prices unchanged. It is possible that extra disposable income would allow families to afford and thus “trade up” for higher-price care options, like moving from informal care to center-based care. But given economic evidence on the effects of income on parental decision-making, we find that these effects are negligible for a subsidy of the size we study: under the Child Allowance policy, ECEC enrollment patterns would look nearly identical to those at baseline.

Parental decisions about childcare are tightly bound up with decisions about whether and how much to work. This is especially true for mothers of young children, who typically bear a disproportionate share of caregiving duties within families.

It is thus unsurprising that the estimated changes in ECEC enrollment under each policy, described above, are also accompanied by changes in maternal employment. When subsidies reduce out-of-pocket childcare costs, the net return to work rises. Policies that reduce the cost of formal care should, all else equal, cause some stay-at-home mothers to instead enter the labor force and send their child to daycare or preschool.

Figure 4 shows estimated employment rates for mothers of children ages 0-4 in 2030 under each policy.

• The two copay cap subsidy policies (Universal Subsidy and Income-Limited Subsidy) would increase employment rates among mothers of young children by around 6 percentage points. Among those who start working, a little over half enter full-time work, with the remainder entering part-time work. These shifts are consistent with our large estimated impacts on patterns of care usage, with more than 10 percent of children moving out of parent-only and informal care arrangements and into formal ECEC.
• The Expanded Care Credit and the Provider Tax Credit would generate comparatively smaller boosts to maternal employment—a difference due largely to smaller average subsidies.
• Unlike the other policies, which change the relative attractiveness of work and nonwork (what economists call the “substitution effect”), the Child Allowance policy merely increases the income of all families—regardless of which decisions they make about care and work (the “income effect”). Empirical evidence shows that substitution effects are much larger than income effects.3 The Child Allowance policy, accordingly, has a negligibly small negative impact on maternal employment. 

When demand-side subsidies make formal care cheaper, more families try to enroll. Meeting this demand requires hiring more care workers, which in turn requires raising wages to attract workers to the ECEC sector.

Figures 5 and 6 present estimates of changes in care worker wages and employment, respectively.

• The Universal Subsidy and Income-Limited Subsidy policies, which are the most generous policies considered in this report, would generate the largest demand responses.
  • To meet this demand, the ECEC workforce would have to expand dramatically: providers would need to hire more than half a million full-time-equivalent workers, an almost 80-percent increase from baseline.
  • Attracting this many workers would require raising average hourly wages from about $17 to $20 for care workers without a college degree and from about $25 to $33 for college-educated workers.
• The Expanded Care Credit would generate more modest wage effects.
  • Because the maximum subsidy amount is smaller than in either the Universal or Income-Limited Subsidy policies, and because the design allows for recipients to use the credit for informal care options, the demand increase for formal care is smaller, and fewer additional workers are needed.
• As discussed below, the Provider Tax Credit reduces ECEC sticker prices by subsidizing labor demand by providers. This price decline generates some additional enrollment (and thus hiring and wage increases), but less than under the copay cap policies.
• The Child Allowance, for reasons discussed in the “Enrollment Effects” and “Employment Effects” sections, produces negligible changes in demand for formal care. Wages and ECEC sector employment thus stay nearly unchanged from baseline levels.

Increases in wages for ECEC care workers may have important effects on children’s outcomes beyond those our model captures (see below). Research has found that lower care worker wages are associated with higher turnover, which in turn can have meaningfully negative impacts on children’s future outcomes, especially for children in lower-income households (Grunewald, Nunn, and Palmer, 2022). To the extent that the subsidy policies we analyze here decrease care worker turnover through their effects on higher wages, they could, in turn, improve outcomes for these children in areas such as social development and language and literacy skills (Hamre et al., 2014).

As discussed in the previous section, increased demand for care resulting from subsidies requires ECEC providers to raise wages to attract more workers. Labor is the largest cost faced by ECEC providers; the Department of Labor reports that care worker compensation represents 60 to 80% of provider operating costs. Higher wages, then, lead providers to raise prices to cover their increased costs. These higher prices, in turn, offset some of the initial increase in demand.

In this section, we present estimates of how these dynamics play out in equilibrium. Understanding these results helps contextualize other results in this report. For example, subsidy-driven price hikes increase fiscal costs and erode benefits for unsubsidized families.

Figure 7 shows estimated impacts on market (i.e. pre-subsidy) prices in the ECEC sector.

• As discussed in the previous section, the Universal Subsidy and Income-Limited Subsidy policies would generate large increases in care worker employment and wages.
  • These wage hikes, about 15 percent and 30 percent respectively, would translate to increases in sticker prices of between 9 and 17 percent, depending on the type of care.
• The Expanded Care Credit would generate more modest wage effects, leading to smaller effects on market prices.
  • Moreover, the CDCTC’s expense cap of $8,000 per child means that subsidies do not grow with price above that cap—so unlike under the copay cap policies, families are not fully insulated from price increases. This limits the feedback loop between subsidies and prices: as prices rise, families bear more of the increase, which dampens demand and further limits upward pressure on wages and prices.
• As noted above, the Provider Tax Credit reduces providers’ input costs by covering part of the wage bill, allowing firms to offer the same amount of care slots at lower sticker prices. Even with increased demand offsetting some of these decreases, market prices are lower in equilibrium.
• Finally, the Child Allowance produces negligible changes in ECEC market prices.

Beyond their immediate impacts on family budgets, maternal employment decisions, and the federal budget, ECEC policies may also affect children’s long-run economic outcomes. Empirical research finds evidence that enrollment in certain types of early childhood care and education improves performance on cognitive assessments. These gains are associated with higher earnings in adulthood. Therefore, policies that allow families to switch from ECEC arrangements that yield relatively low earnings in adulthood to those that yield relatively high earnings in adulthood may have effects that go beyond the immediate impacts on family budgets.

Because our model tracks how policy affects ECEC enrollment patterns, we can translate those changes into estimates of long-run earnings impacts for children. We do so by mapping policy-induced changes in care arrangements to changes in children’s test scores (using different estimates for younger and older children), then converting those test score gains into projected adult earnings.4 These estimates should not be taken as normative judgments about what constitutes “better” childcare. Instead, they are meant to incorporate the existing empirical literature on average effects on one aspect of children’s long-run outcomes: later-in-life earnings.

(A full description of how we measure the impact of ECEC arrangements on cognitive assessments and how we translate these impacts into long-run effects on earnings can be found in the documentation for the ECEC model.)

Figure 8 shows the average estimated change in earnings at age 27, per child, by quintile of the parent income distribution. These are averages across all children in each quintile—including those whose care arrangements do not change under the policy—so they reflect both the size of the earnings gain for children who do switch and the share of children in each quintile who are induced to switch.

• The Universal Subsidy and Income-Limited Subsidy policies, which produce the largest enrollment shifts toward center-based care, would generate the largest estimated earnings gains.
  • Benefits would be concentrated among children from lower- and middle-income families who, due to the progressive structure of the policies, are the most likely to move from parent-only or informal care into formal center-based programs.
  • Under the Income-Limited Subsidy policy, future earnings would actually fall slightly for children of top-quintile parents. This is because some higher-income parents, who do not qualify for the subsidies which insulate families from rising prices, would shift away from care options that yield relatively higher earnings in adulthood.
• The Expanded Care Credit would produce very small changes in children’s future earnings.
  • As discussed in the “Enrollment Effects” section, the CDCTC’s broad care-type eligibility and expense cap encourage families to choose cheaper informal arrangements, so fewer children transition into care types yielding higher future earnings.
• The Provider Tax Credit would also generate positive earnings effects by reducing sticker prices for formal care.
  • Because these price reductions are not targeted to lower-income families (as is the case for some of the other policies), enrollment shifts are more evenly spread across quintiles than under the copayment cap policies. This translates into average earnings impacts that rise with parent income, since children of higher-income parents earn more on average in adulthood.5
• The Child Allowance, which does not alter the relative prices of different forms of care, would produce negligible effects on children’s future earnings through the channels we consider in this report.6

Table 3 shows estimated direct federal budget effects of each policy over the next three decades. These estimates account for the direct cost of subsidy payments (or tax credits), net of changes in tax revenue driven by shifts in parental employment. They also reflect the effects of changes in enrollment and changes in sticker prices.

Importantly, these fiscal effect estimates do not include other potential indirect savings to the federal government, such as decreases in utilization of public assistance programs (such as Medicaid and SNAP) or decreases in criminal justice involvement, nor do they include potential labor productivity gains beyond the children directly affected by the policies. The estimates also do not include effects on state and local spending and revenues.

• The Universal Subsidy and Income-Limited Subsidy policies would have similar budget effects despite the latter imposing an income restriction. The reason for the small difference is that only a small minority of parents of young children earn above the income eligibility threshold (250% of median income), and among those who do earn that much, the Universal Subsidy policy only covers the costs above 7 percent of income—and most care options generate costs of less than 7 percent of income for high-income families. (This similarity in effects is seen across all aggregate outcome metrics for these two policies.)
• The Expanded Care Credit and Provider Tax Credit, policies that are less generous in aggregate compared to the copay cap subsidies, generate correspondingly smaller fiscal impacts.
• As noted above, the Child Allowance policy does not meaningfully shift enrollment or employment behavior. Therefore, the policy’s budget cost largely reflects a simple mechanical accounting exercise: $1,000 (adjusted for inflation) times the number of children aged 0-4.

Aside from these multi-decade budget estimates, another way to analyze fiscal impacts is to perform a net present value (NPV) calculation, where the direct costs of the policy are weighed against indirect fiscal benefits from the policy’s longer-term economic impacts. In the case of ECEC policy, the direct fiscal costs are the subsidy outlays; the indirect fiscal benefits are higher tax payments from children who earn more in adulthood after being enrolled in forms of ECEC that are estimated to have a larger impact on children’s future earnings. Because these tax revenues accrue decades from now, they need to be discounted when compared to the near-term costs.

We apply the child earnings estimates described in the previous section to The Budget Lab’s projected population of tax filers in future years, accounting for intergenerational income mobility patterns so that children who benefit from ECEC policy are accurately matched to future earnings profiles and tax rates. Combining this information, we get projected annual changes in tax revenue due to policy-driven changes in human capital development starting two decades from now (when today’s children enter the workforce).

Table 4 shows how these fiscal costs and benefits, appropriately discounted to 2026, stack up under each policy.7 Calculations are done for a single cohort of children exposed to the policy reform environment from birth through age four. It reports the share of upfront net cost of the policy (including tax changes from parental employment changes) that are eventually recouped in the form of higher federal tax revenues decades from now—a kind of fiscal return-on-investment (ROI) metric we call the federal tax revenue recovery ratio. Policies that shift enrollment towards forms of care with a larger estimated impact on children’s future earnings and that target benefits towards children of higher-income parents—who themselves tend to earn more and face higher tax rates in adulthood—will generally see higher federal tax revenue recovery ratios. (We stress that this is not a social ROI metric but rather a narrower financial calculation concerning the federal budget.)

• None of the policies “pay for themselves” in fiscal terms—that is, all recovery ratios are less than one.
• The policy with the highest recovery ratio is the Provider Tax Credit policy. This finding is the result of two features of this policy:
  • First, because it focuses on formal, licensed types of care, the policy encourages shifting into forms of ECEC with a greater estimated impact on children’s future earnings.
  • Second, relative to other policies which disproportionately impact low-income families, these enrollment shifts are concentrated among children of higher-income parents. These children, on average, have higher incomes and face higher tax rates, leading to relatively more tax revenue in the future.
• The Universal Subsidy and Income-Limited Subsidy policies, like the Provider Tax Credit policy, shift enrollment towards ECEC arrangements with higher impacts on children’s future earnings. The typical beneficiary, however, has lower income as an adult than that of the Provider Tax Credit policy, leading to lower fiscal benefits.
• The Expanded Care Credit policy has a very low federal tax revenue recovery ratio because it does not necessarily target forms of care with larger impacts on children’s future earnings in its subsidy design.
• Because it generates negligible behavioral responses, the Child Allowance policy generates essentially zero long-term fiscal benefits.

How does each policy affect families across the income distribution? In this section, we introduce The Budget Lab’s preferred way of measuring policy-induced changes in a given family’s well-being, which we call the economic value of the policy for that family. We then show how this measure varies across the income distribution for each of the five policies we consider.

Economists often quantify how a proposed policy would change an individual’s (or family’s) well-being by asking, “How much additional income would the individual (or family) need in order to be exactly as well-off today as they would be if the policy were enacted?”8 For many policies, it is straightforward to answer this question: for example, if we were to impose a $1 tax on all individuals, every individual would be made worse off by exactly $1.

In other cases, like for the policies we study in this report, the answer is less straightforward. This is primarily because families can (and do) shift their utilization of different forms of childcare in response. This means that simply computing the change in out-of-pocket childcare spending due to the policy is not a good measure of the change in families’ well-being. Suppose, for example, that a family shifts from unpaid childcare provided by a grandparent to center-based childcare once the latter is subsidized. In this case, their out-of-pocket childcare spending would go up—but they could very well be better off because a more-preferred but previously unaffordable childcare arrangement is now available to them.

For this reason, we prefer to use a metric we call economic value—which takes into account both changes in out-of-pocket prices faced by families and changes in use of different forms of childcare—when quantifying policies’ impacts on families across the distribution.9 In a companion blog post, we explain in detail how we calculate this metric and how it relates to other common distributional metrics, such as changes in after-tax-and-transfer income and changes in out-of-pocket spending.

Figure 9 shows the average economic value for each policy for families in each quintile of the income distribution, as well as for the top 10, 5, and 1 percent.10

• The Universal Subsidy, Income-Limited Subsidy, and Expanded Care Credit policies all show a similar pattern of progressivity: economic value is higher (in both dollar and percent-of-income terms) for families in the middle and at the bottom of the distribution, while those towards the top are made slightly worse off (largely due to increases in market prices for paid forms of childcare).
• Economic value is highest in dollar terms for the Provider Tax Credit policy for families at the top of the income distribution, who tend to use higher-priced forms of care and who thus benefit most from the policy’s reductions in market prices. When expressed as a share of baseline income, however, the policy’s benefits are higher towards the bottom of the distribution than at the top.
• The economic value of the Child Allowance policy is exactly equal to $1,000 per child for all families. Because this flat value does not vary with income, the policy has greater benefits as a share of baseline income for those at the bottom of the income distribution.

The authors are grateful to the following external reviewers for their thoughtful comments and feedback on an earlier draft of the report: Aviva Aron-Dine, Lena Bilik, Lauren Bauer, Caleb Brobst, Barbara Chow, Este Griffith, Rob Grunewald, Suzanne Kahn, Sadie Kliner, Joshua McCabe, Rebecca Gomez, Zach Moller, Rebecca Nathanson, T.V. Ninan, Will Raderman, Natalie Renew, Abigail Sanchez, and Elizabeth Wilkins.

The authors especially wish to thank Jessica Brown, Chloe Gibbs, Aaron Sojourner, and Matthew Wiswall, whose advice and guidance were invaluable in the development of TBL’s ECEC model and who also commented on an earlier draft.

The views expressed in this report are not necessarily those of any external reviewer(s).