Tab 5.0 Calculating the costs and potential benefits of early childhood development projects Tab 5.0 Calculating the costs and potential benefits of early childhood development projects



The financial requirements for early childhood development programs that would address the needs of the majority of preschool-age children in Latin America are vast, and the financial resources available quite limited. So it is not surprising that politicians inevitably ask, How much do these programs cost?

Answering this question is not so simple. The costs of ECCD programs vary as widely as do the programs' characteristics (box 5.1):

  • Goals (desired coverage, inputs, processes, desired outcomes).

  • Activities (contents or number of services, such as education, health, nutrition, and family services).

  • Models (center-based, home visiting, mass media campaigns).

  • Technologies (paraprofessionals or professionals, high or low ratios of children to staff and of staff to supervisors).

  • Duration (hours per day, days per year, number of years).

  • Population served (a large or small group of beneficiaries, the age and health status of children).

  • The program context (concentrated or dispersed settlement, high or low per capita incomes).

The annual cost per child, or unit cost, of 27 programs in Latin America illustrates the range: the ratio between the highest and lowest cost is over 120 (figure 5.1). [need electronic version of this figure] The median is $376 per child. Or close to a dollar a day. (See annexes A and B for the characteristics of these programs and their unit costs.)

While we know that there are serious measurement errors reflected in the magnitudes underlying Figure 5.1 (see below), these are unlikely to negate important implications of the data for those 27 programs, which can be taken as roughly indicative of the range and shape of the distribution of the universe of such programs in the region.

First, differences across corresponding programs must be comparably diverse, so that when we are talking about ECCD programs in an analytical context, we need to ask what kind of ECCD program we are talking about.

Second, although the overall range is huge, the bulk of the sample of programs falls in a far narrower interval: two-thirds of the observations are between US$142 and US$500 per child per year-a range of only 3.5 times. Thus, programs with unit costs falling outside that range are atypical and deserve special scrutiny.

Third, if ample coverage is a goal, the highest cost programs, say, those above US$1,500, are unaffordable for all our clients, and even those above the US$500 benchmark, are unaffordable for most countries in the region, at least under prevailing social priorities.

We turn to the matter of measurement errors in cost comparisons. There are important differences in the way costs are commonly calculated. Some costs are based on budgets rather than on actual expenditures. Some include prorated administrative costs of the larger system in which a program is situated; others do not. Some programs calculate the cost per participant based on the number of people enrolled; others use the number of people who actually attend. Some include estimates for contributions in kind; others do not. Moreover, the cost figures that can be extracted from international comparisons do not provide an accurate guide to budgeting for an early childhood program in a new setting, even if the same model is used. Salary levels may be very different in that setting, for example, or the beneficiary population more dispersed. In view of the wide variation in unit costs, in programs, and in the assumptions on which cost data are based, the most promising approach for estimating the costs of an ECCD program is to derive them from its components (Moran, Myers, and Zymelman, 1997).

Box 5.1 Characteristics and costs of three early childhood development programs

The annual cost per child of the PRONOEI program in Peru in 1984 was estimated to range between $39 and $84 in the four states in which the program had been implemented; the average cost per child was $55 (Myers and others 1985). The basic program model, technology, age of the children enrolled, and program duration were the same across the states, and all programs were located in rural areas. The main reason that the cost per child differed was economies of scale: the size of enrollment varied across the states. In addition, some states included a community development component, which increased the cost per child, while others did not.

The annual cost of the Colombian home day care program (Hogares de Cuidado Diario) in 1992 was estimated at about $313 per child (Castillo, Ortíz, and González 1993). This program has both day care and child development components (while PRONOEI is centered on a child development goal closely linked to preparation for school). In 1992 the caregiver-child ratio was 1 to 15 in the Colombian program (compared with about 1 to 30 in the PRONOEI model), with parents providing volunteer help on a rotating basis. Children remain in the day care home eight hours a day (compared with three hours a day in the PRONOEI program).

The yearly cost per child of the home day care program in Venezuela during the oil boom was very high, estimated at about $2,500 (de Ruesta 1978). Although the basic model was the same as that of the Colombian program-home day care-the caregiver-child ratio was 1 to 5, children were cared for 12 hours a day, and a variety of supervisory and social services were included that were not part of the Colombian program.

Note: All dollar amounts in the box are in constant 1994 dollars.