This appendix describes the methodology and logic used for the illustrative cost estimate that appears in Chapter 6, as well as details of the particular policy specifications applied in the illustrative example to estimate the total systemwide cost of attaining high-quality early care and education.
The illustrative cost estimate that appears in Chapter 6 was conducted in two parts. First, a center-based early care and education (ECE) cost calculator used by Brandon (2011) and Elicker, Brandon, and MacDermid (2016) was adapted and applied to estimate ECE costs per child-hour separately for each child-age group (infant, toddler, prekindergartner) and for multiple sets of policy specifications representing the four phases of implementation, which are characterized by increasing staffing quality standards. Home-based costs were estimated by applying the ratio of home-based to center-based prices by child-age groups derived from the National Survey of Early Care and Education (NSECE) (National Survey of Early Care and Education Project Team, 2016a), under the broad assumption that the ratio of prices to costs is a constant. These calculations correspond to the left-hand box in each of the two rows of Figure 6-1.
Second, to obtain the national (aggregate) cost estimates (right-hand boxes in Figure 6-1), the hourly costs derived in the first step were applied
to the estimated number of hours of ECE care used by U.S. children in each child-age group and family income category (the middle boxes in Figure 6-1). Two versions of this aggregate cost calculation were run: one applied current hours of ECE utilization (static estimate, corresponding to the top row in Figure 6-1); the other applied estimated changes in utilization patterns due to increased use of high-quality ECE services. Changes in usage patterns were defined by child-age group and family income category (dynamic estimate, corresponding to bottom row in Figure 6-1). For both static and dynamic estimates, affordable shares of income were specified for each family income category and applied to the estimated costs to estimate the potential family contributions (“family payments” in Figure 6-1, right-hand boxes) and the remaining subsidy costs (“net subsidy cost” in Figure 6-1). Details of these two parts—the hourly cost calculation and the aggregate cost calculation—are discussed next.
Components of the Cost Calculation
Hourly Cost Calculation
The overall logic of the hourly cost calculation is to apply a set of quality-related policy specifications to derive the number of full-time equivalent (FTE) staff for various positions and related qualification-compensation levels. A constant factor of 8 percent, as well as an adjustment factor, was also applied to generate a nonpersonnel increment for phase 1 (see discussion below of basis for these factors). The number of FTEs required was calculated using specified child-to-staff ratios, ECE hours per day, and days per year of ECE service operation. An increasing share of lead educators with bachelor’s-level degrees was specified at each of the four phases, and the specified mix of qualifications was applied to differentiate the number of FTEs required, by position. The mix of staff with bachelor’s-level degrees versus lower educational qualifications was varied by child-age group as shown in Table A-1. The estimated numbers of FTEs were also adjusted upward to account for nonchild-contact hours for staff to prepare, plan, and engage in other professional responsibilities, as well as release time for participation in professional development.1 Additional FTEs were included to account for noninstructional staffing categories: program direction and administration, coaches and mentors, reading and language specialists, and
1 Nonchild-contact time refers to activities essential to ECE professional activity performed without simultaneous responsibility for child supervision. It includes preparation and planning time, collegial sharing, educator team meetings, and time for completing child assessment reports and holding parent conferences.
TABLE A-1 Distribution of Child Contact Hours by Role and by Child-Age Group, Across Four Phases of ECE System Transformation
|Current Mix (NSECE data)||Phase 1||Phase 2||Phase 3||Phase 4|
|Infants (less than 12 months)a|
|Role and Qualifications||Share of ECE Staff Hours by Role|
|Assistant||Some college (%)||36||40||30||30||15|
|Toddlers (12–36 months)a|
|Role and Qualifications||Share of ECE Staff Hours by Role|
|Assistant||Some college (%)||36||30||20||20||10|
|Current Mix (NSECE data)||Phase 1||Phase 2||Phase 3||Phase 4|
|Prekindergartners (36–60 months)b|
|Role and Qualifications||Share of ECE Staff Hours by Role|
|Assistant||Some college (%)||24||20||25||15||5|
NOTES: BA+ = bachelor’s degree or higher; AA = associate’s degree; CDA = Child Development Associate certification; HS = high school diploma.
a Current mix uses data from the NSECE for children ages 0–3 years.
b Current mix uses data from the NSECE for children ages 3–5 years.
other nonprofessional supports.2 The matrix of all staff salaries by position (see discussion below) was then multiplied by the estimated FTEs to obtain total staff salary costs. A constant factor of 31.5 percent was added to salaries to cover benefits, including health care, retirement, payroll tax contributions, and paid leave (see discussion below).
Thus, the estimate for total onsite costs is the sum of salaries and benefits for FTEs and nonpersonnel costs, plus the 10 percent adjustment factor (see detailed descriptions below). These totals were then divided by the number of hours of operation and number of children served to derive costs per child-hour. Costs per child-hour were then multiplied by 2,080 hours per year (40 hours per week, 52 weeks per year) to derive an illustrative full-time, full-year cost per child.
Aggregate Cost Calculation
The aggregate cost calculation converts the unit costs per child-hour into (national) aggregate system costs. As noted above, the overall logic of the calculation is to multiply the estimated hourly costs of onsite direct service provision, for each phase separately, by the number of hours of center-based and home-based ECE used by each child-age group and family income category. This calculation yields an estimated provider cost of onsite service delivery. Two versions were run: one assumed the current pattern of service utilization;3 the second projected shifts in utilization across the four phases resulting from increased accessibility and quality, varied by child-age group and family income category.4 To estimate
2 Ratios applied for directors/administrators were derived from NSECE data on the distribution of center programs by size, with a specification of one director/administrator for every facility. Coach/mentor ratios were based on caseloads reflected in the literature (see discussion of onsite professional development below). Ratios for specialists were based on desirable caseloads, adjusted for the estimated percentage of children in the population with special needs. See Box 6-1 in Chapter 6 on estimating the share of children with special needs.
3 The aggregate current hours of ECE utilization by child age, family income, and ECE type are household-based and were derived from Latham (2017) using 2012 data from the NSECE Public Data Set. Thus, aggregate hours of utilized early care and education equal the number of households times the mean number of hours per week utilized (times 52), reduced by 5 percent to reflect summer decline in utilization; this result is calculated separately for each family income category, and within income category, for each child-age group and utilization rate for ECE type (center-based or home-based care).
4 For the dynamic estimate, total hours utilized were adjusted by three factors: an accelerating shift from home-based to center-based ECE utilization; increased share of families/children using paid ECE services; and increased use of ECE hours per day and per week. We based our estimates for these three factors on elasticities reported in past econometric studies, adjusted for more recent changes in utilization patterns and an expectation (assumption) that low- and middle-income families are more price-sensitive than higher-income families (see discussion in Chapter 6).
system-level costs, a constant factor of 8 percent was added to the aggregate service delivery costs at each phase, based on prior research by Brandon and colleagues (2004b) as discussed in Chapter 6. The aggregate service delivery cost was thus calculated as the aggregate hours of ECE services utilized (static or adjusted dynamic) multiplied by the cost per child-hour, as estimated by the hourly cost calculation. This yields an aggregate annual service delivery cost per child-age group and ECE type (center-based or home-based care) for each family income category.
In addition, the affordable family payment schedule described in Chapter 6 was applied to estimate the potential family contribution to the total ECE costs for each income category and the remaining share of total costs (total costs less family contribution) that would require financial assistance (i.e., the public/private subsidy). Family income category was the primary categorization, with child-age group and ECE type used to calculate the estimated family-level fee. This computation must be done by family income category because almost a third of children ages 0 to 5 years have a sibling in that age group. Therefore, family payments as a share of income cannot be counted separately for each child-age group because there would then be substantial double-counting. Within the service delivery cost, the percentage of income considered affordable by the analysis discussed in Chapter 6 was then multiplied by the aggregate income of families in that income category to estimate the total potential family contribution toward the cost of early care and education. The family contribution was limited to the cost of early care and education. Affordable-fee estimates were based on total utilization per families in each income category, aggregated across child-age groups and ECE type. This approach ensured that families with more than one child under age 5 years using early care and education were not assumed to be paying the calculated “affordable share” of income for each child but were instead paying one affordable share across all their children. The potential family contribution was then subtracted from the total cost to yield the estimated subsidy cost required to make high-quality early care and education accessible to families of all incomes. (See the section in Chapter 6 titled “Example Part II: Family Payments in a High-Quality System.”)
POLICY CHOICES AND ASSUMPTIONS
As discussed in Chapter 6, the Transforming report outlined a number of quality standards for ensuring the provision of high-quality early care and education for all children. These quality standards, or elements of quality, each affect the cost. For the illustrative example in this report, the committee has specified an array of quality-related elements, as well as other onsite costs—including operating hours and days, staff qualification mix
and compensation, child-to-adult ratio, complements of nonclassroom staff, staff supports for effective practice, and nonpersonnel costs—to estimate the total direct operating costs for providing high-quality early care and education. In addition, estimates of system-level professional development and quality assurance costs have been included in the total estimate of the cost of a high-quality ECE system (see Chapter 6). All costs were estimated in constant 2016 dollars (i.e., with no inflation adjustments), with estimates derived from pre-2016 data inflated to 2016 values using the National Average Wage Index.5 All specifications reflect an average across a wide range of programs, and specific values may vary due to particular program attributes or local circumstances.
As emphasized in Chapter 6, the specifications for this estimation exercise were chosen as part of the committee’s illustrative (and hypothetical) cost estimate. They do not represent recommendations, explicit or implied, of the committee.
Staff Qualifications and Compensation
The cost estimate reflects a steady increase from current levels, across the transition phases, of the share of staff with desired qualifications and a steady increase in wage rate linked to each level of education at each phase. Wages are assumed to reflect education level and are not varied by child-age group. The key salary levels are defined as paying ECE educators with a bachelor’s degree wages equivalent to child-family social workers with a bachelor’s degree by phase 2 and equivalent to kindergarten educators by phase 4, though not annualized for a full year (12 months) of ECE service.6 The intermediate steps (i.e., phases 1 and 3) were specified to be 90 percent of these phase 2 and 4 values, respectively.7 For staff with less than a bachelor’s degree, each level of education is specified to be a percentage of the next higher level, derived from center educator/caregiver salary data in the NSECE. Thus, an educator with an associate’s degree would be paid 75 percent as much as one with a bachelor’s degree; an educator with a Child Development Associate (CDA) certification or some college would earn 81 percent as much as one with an associate’s degree; an educator with a high school degree or less would earn 91 percent as much as one with a CDA certification or some college. The 2016 dollar values of educator salaries are shown in Table A-2.
7 Salary levels are derived from the 2016 U.S. Department of Labor’s Occupational Employment Statistics report. (See https://www.bls.gov/oes/ [December 2017]). Comparative data collected by the National Survey of Early Care and Education (NSECE) in 2012 have been adjusted upward to reflect 2016 levels, using the Social Security Administration’s wage index.
TABLE A-2 Example Average Lead Educator Salaries, Current and Estimated, Across Four Phases of ECE System Transformation (in 2016 dollars)
|Educator Position||Degree||Current Averagea||Phase 1 (year 3)||Phase 2 (year 6)||Phase 3 (year 9)||Phase 4 (year 12)|
|Aide||High school or less||$22,700||$23,638||$26,265||$27,594||$30,660|
NOTES: BA+ = bachelor’s degree or higher; AA = associate’s degree; CDA = Child Development Associate certification.
a Current average salaries are adjusted to 2016 dollars from the 2012 dollar amounts reported in the NSECE.
SOURCE: Current average salaries are from the NSECE. Current bachelor’s-degree salary levels for phase 2 are for child-family social workers and bachelor’s-degree salaries for phase 4 are equivalent to elementary school teachers (as reported in 2016 U.S. Department of Labor’s Occupational Employment Statistics report). Other salary levels were calculated by the committee following specifications based on NSECE data as explained in the text accompanying the table.
These salary levels are assumed to be applied consistently for all center-based care, eliminating the current disparity of wages by sponsoring organization (National Survey of Early Care and Education Project Team, 2013).
In the phase 4 specifications, lead educators’ salaries are equivalent to a kindergarten educators’ salaries for a 9-month contract as reported in the U.S. Department of Labor’s Occupational Employment Statistics report.8 If phase 4 bachelor’s degree–level salaries were set equivalent to 12 months at the monthly rate of the contract amount for kindergarten educators’ salaries, they would be about $74,000 rather than $55,460. In order for educators’ salaries to reach true parity with salaries for kindergarten-to-3rd-grade educators, as discussed in Chapter 6, ECE educators working a full year would need to have their salaries set equivalent to 12 months at the monthly rate of the contract amount, and the cost per child and total costs to the entire ECE system would be adjusted to account for this increase. If lead educator salaries were pegged to an annualized equivalent of the normal 9-month kindergarten educator salary, and if all other leadership and instructional salaries were adjusted in a similar fashion, it would add about 11 percent to direct service costs and about 10 percent to total system costs.9
9 This estimate assumes that the costs of benefits increase commensurate to salary amounts, and there is no increase in nonpersonnel costs as salaries increase.
Salaries for directors, coaches/mentors, specialists, and other staff would also be increased commensurately to maintain the salary relationships among levels of qualification and responsibility.
According to the 2017 Bureau of Labor Statistics National Compensation Survey, employee benefits as a share of salaries do not vary substantially by occupation, but they do vary substantially by the sponsoring organization, especially public school–based versus community-based centers (Bureau of Labor Statistics, 2017). Given this differentiation, the committee’s estimate assumes there will be a distribution across different types of organization. Therefore, a slight increase from the 29 percent for service employees (the category including child care workers) to the 31.5 percent for kindergarten through grade 12 (K–12) educators was applied to all staff across all four phases. Even with this small increase in benefits as a share of wages, the value of benefits would increase substantially, since substantial increases in wage levels are projected. The level of benefits as a share of salary is applied equally for all staff positions. This treatment assumes elimination of the current large disparity of benefit levels among ECE educators, depending on the organization sponsoring their ECE program (reported in Maroto and Brandon, 2012).
STAFF LEVELS AND STRUCTURE
ECE staffing levels and structures differ from those typically used in K–12 classrooms. Whereas a K–12 educator typically works alone (although some educators of younger children have the assistance of an aide or paraprofessional), teaching in early care and education is a collective effort because of the needs of very young children. As noted in Chapter 6, it is common for more than one educator to be in an ECE classroom or group, often one lead and one assistant educator, and some time will be allocated to non-child-contact time (for educators at each level), in order for the educators to complete other professional responsibilities and participate in professional learning. Therefore, the phase 4 estimate reflects the costs for a lead educator interacting with children for 75 percent of the day, with assistant educators or aides accounting for the remaining contact time. Table A-1 (above) shows the increasing mix, across the four phases, of contact time by staff of different qualifications for the roles we assumed for each age group.
The specified values in the cost estimation reflect phasing in, from current levels, the child-to-staff ratios for each child-age group recommended by the National Association for the Education of Young Children and taking into account current state requirements (see Chapter 4). Intermediate group-size levels were applied.
In addition to the mix of staff qualifications, the child-to-adult ratio is a critical factor in determining the number and cost of staff required
to serve a given number of children. Considering the standards discussed in Chapter 6, the committee’s cost estimates are based on the following average child-to-adult ratios: for infants, phasing down from 5:1 to 3:1; for toddlers, phasing from 5:1 to 4:1; for prekindergartners, phasing from 11:1 to 10:1. It should be noted that for prekindergartners, these ratios are higher than current average ratios as reported in the NSECE and are based on practice in some European countries, assuming that better-qualified and -supported staff will be able to practice effectively with more children (Kagan et al., 2002).10 If child-to-adult ratios are lower for prekindergartners, total direct service costs could be 10–20 percent higher than the estimates presented here.
OPERATING HOURS AND DAYS
For the hourly cost calculation, costs were computed on the basis of full-time (40 hours per week), full-year operation (52 weeks per year). For the aggregate cost calculation, two different measures of duration were applied. For the static analysis, the current average hours per week in each type of ECE setting by child-age group was applied. The weekly hours were multiplied by 52 weeks, then decreased by 5 percent to account for an anticipated decrease during the summer in children using ECE services.
The dynamic analysis assumes that as affordability improves, average hours would increase by 8 percent for low-income families, 6 percent for middle-income families, and 2 percent for upper-income families.11 Hours per week in paid ECE services average about 35 hours per week for infants and toddlers, so that the static cost estimates mostly reflect full-time ECE use. For prekindergartners, the average is about 27 hours per week, reflecting a mix of full-time and part-time prekindergarten programs. Whereas prekindergartners currently spend more time in unpaid ECE settings than younger children, the dynamic estimates assumed a higher participation in full-time programs for prekindergartners.
10 Comparability of child-to-adult ratios used in some European countries to ratios in the United States is unclear, especially given the higher percentage of children living in poverty and dealing with chronic stress in this country.
11 Low-income refers to families with a household income zero to 200 percent of the federal poverty level; middle-income refers to families with household income between 200 percent and 300 percent of of the federal poverty level; and affluent refers to families with a household income above 400 percent of the federal poverty level.
SUPPORTS FOR PROFESSIONAL RESPONSIBILITIES AND LEARNING
Staff supports are those ongoing costs that would be reflected in a center’s operating budget. They include paid nonchild-contact time, including time for professional responsibilities such as preparation and planning, professional learning and development, and coaching and mentoring.12 The costs for these supports in the committee’s estimate are allocated by adding to the number of FTE staff at different positions and salaries beyond the FTE staff required to meet child-to-adult ratios in classrooms or groups.
Paid time for additional professional responsibilities conducted without children present is accounted for in the estimate to reflect the fact that staffing costs encompass more than just direct teaching time. This factor includes paid time for preparation and planning, assessment, professional sharing and reflection, and engagement with families. The cost estimate applies additional time for these professional responsibilities to lead educators.13 As described in the Transforming report, lead educators are primarily responsible “for planning and implementing activities and instruction and overseeing the work of assistant teachers and paraprofessionals” (Institute of Medicine and National Research Council, 2015, p. 6).
Costs for ongoing professional learning include paid release time for professional development (onsite and offsite) and ongoing professional learning activities such as coaching and mentoring. Drawing on Isner and colleagues (2011), the committee estimated that the resources devoted to coaching and mentoring for lead educators, assistants, and aides would increase across phases, from 1 mentor per 35 FTE educators in phase 1 to 1 per 25 educators in phase 4. In addition, the share of staff participating in offsite professional development each year, requiring backfill with substitute educators, was assumed to be high during the transition but declines slightly over the phases (as educational qualifications are increased)—from 25 percent in phase 1 to 15 percent in phase 4. The hours per week of offsite professional development time for each participating staff member increases across the phases from 3 to 4 hours (see, e.g., National Council on Teacher Quality, 2012; U.S. Department of Defense, 2009).
12 Coaching and mentoring includes both the additional staff costs of hiring the coaches and mentors who perform the coaching or mentoring activity and the additional nonchild-contact hours for teachers to participate as recipients of the coaching or mentoring activity.
13 Costs would increase if assistant educators or aides were provided with paid time for additional professional responsibilities such as planning and preparation.
Nonclassroom staff includes administrative staff and instructional support staff, such as coaches, mentors, and trainers, reading and language specialists, special education consultants, and assessment specialists. The committee’s estimate includes increasing complements of such staff across each of the four phases because they will be instrumental in supporting the development of the ECE workforce, as well as in contributing to the educational attainment of children (see, e.g., Elicker, Brandon, and MacDermid, 2016). Support or consulting staff that can be paid from health or nutrition programs are not included here. Transportation, food service, and custodial staff are included. Staffing complements reflect an average across all centers of varying size and sponsorship. Thus, an average of one director/administrator per facility is specified, even though small centers are likely to have a part-time director and large centers are likely to have a full-time director plus an additional assistant director. No changes in the distribution of center size or structure are assumed because there is no available literature relating center size to various aspects of quality. Similarly, no overall changes in efficiency were assumed. Some changes, such as purchasing of goods and services by groups of centers, may increase efficiency if adopted on a large scale. Others, such as a shift to more school-based programs, may increase costs due to special features of those operations.
Nonpersonnel costs include facilities occupancy costs, such as rent and utilities; education equipment and supplies, including technology; office supplies; and food and kitchen supplies (Augenblick, Palaich and Associates, 2017). A constant amount of 8 percent (roughly $3,200 per year for infants and toddlers and $1,800 per year for prekindergartners) is added for nonpersonnel costs in phase 1 (Augenblick, Palaich and Associates, 2017; Brandon et al., 2004b; Elicker, Brandon, and MacDermid, 2016). Nonpersonnel costs are not expected to increase as a constant percentage as personnel costs increase and are held constant across the four phases; these costs therefore decline as a share of total costs over the four phases as personnel costs increase.
A constant 10 percent is added to staffing and nonpersonnel costs to reflect the need for providers to maintain a reserve to cover such inefficiencies as temporary drops in enrollment, delays in state reimbursement, or nonpayment by families.