Primary School Construction in Chile
Introduction
In November 1964, Chile announced the Special Plan for Primary Instruction to expand access to public schooling in grades 1-8. A national enrollment census of children aged 7-15 identified communities where schools or classrooms were needed, and 3,539 new classrooms were built in 1965 alone. New classrooms were staffed by teachers from an accelerated 32-week training program and supplied with textbooks and instructional materials.
Lucas and McEwan (2026) estimate the plan’s long-run effects with a difference-in-differences design. Children born before 1949 were too old to benefit, while each cohort born from 1949 to 1959 gained an additional year of potential exposure. Treatment intensity varies across 79 departments (roughly equivalent to US counties) with the share of 7-15 year-olds not enrolled in school in the 1960 census. The analysis compares individuals born in more-treated departments (above the population-weighted median treatment intensity) with those born in less-treated departments.
Using census samples pooled from 1982 to 2017, the paper finds that individuals in the fully exposed 1959 cohort born in the most-treated department completed 1.1 more years of schooling than those born in the least-treated department, and were 10 percentage points more likely to be employed and 6 percentage points more likely to work in a higher-skill occupation. Schooling gains were larger for women, closing the gender gap in attainment. Mothers’ exposure also improved their children’s grade progression and completed schooling; fathers’ exposure had no detectable effect.
The paper uses these estimates to estimate the MVPF of the extra school construction in more-treated departments, including spillovers to the next generation.
MVPF = 10.1
Net Cost
Program expenditures combine capital and current costs, with cost figures drawn from Schiefelbein (1976).
For capital costs, the paper annualizes a 1969 construction cost of 700 escudos per square meter over a 30-year building life and applies it to the square meters built in excess of a linear interpolation between 1964 and 1969. Current costs, covering teacher salaries, instructional materials, and other recurrent spending, are the increase in classroom operating costs between 1964 and 1970, applied to each program year from 1965 to 1975. The paper treats 25.6 percent of national expenditures, the difference between the two groups’ shares, as the additional spending in more-treated departments, spreads each year’s spending evenly across the 11 birth cohorts eligible that year, and divides by the cohort size in more-treated departments. This gives additional expenditures of USD 320 (CLP 271,733) for the 1959 cohort.
These costs are partially offset by value-added tax on consumption out of the earnings gains described below. With a marginal propensity to consume of 0.5 (Barrero et al. 2020) and a value-added tax rate of at least 18 percent in most years, the additional revenue is USD 152.
The children of the average 1959-cohort individual generate USD 35 of additional secondary-school expenditures and USD 25 of additional value-added tax revenue.
The net cost is therefore (USD 320 + USD 35) – (USD 152 + USD 25) = USD 178.
Willingness to Pay
Willingness to pay is the present value of increased after-tax lifetime earnings for the average individual born in 1959 in a more-treated department and for that individual’s children. Because the census data contain no earnings, the paper converts schooling gains into earnings.
For the first generation, the paper estimates a gain of 0.051 years of schooling per year of exposure in more-treated relative to less-treated departments, or 0.56 years for the 1959 cohort. It multiplies this by a 15.4 percent return to a year of schooling, estimated in the 2006 CASEN household survey including adults with zero earnings, so that it also captures effects on labor force participation. The gain is applied to age-earnings profiles for ages 18–65 from the same survey, adjusted for lower earnings in more-treated departments and 0.5 percent annual real wage growth. The result is an additional USD 1,698 (CLP 1,443,359) in earnings, of which USD 1,546 remains after value-added tax.
Children of mothers born in 1959 are assumed to be born in 1986. The paper estimates a gain of 0.031 years of schooling per year of maternal exposure, applies the same return to schooling to age-earnings profiles constructed for the children’s cohort, and scales the gain by the share of women in the first generation (0.52) and their mean number of surviving children. The children’s after-tax earnings gain is USD 241.
Willingness to pay is therefore USD 1,546 + USD 241 = USD 1,787.
MVPF Summary
The paper reports an MVPF of 10.1 for the 1959 cohort, and a population-weighted average of 12.7.
The paper calculates an MVPF for each cohort born from 1949 to 1959. Counting the first generation only, the MVPF is 9.2 for the 1959 cohort and ranges from 8.7 to 13.7 across cohorts, with a population-weighted average of 11.7.
The paper also reports an MVPF of 4.1 with an 8 percent return to schooling, 6.6 with a 7 percent discount rate, 12.5 with a 20-year building life, and 2.6 under a joint worst-case scenario. The program breaks even (an MVPF of 1) only at a return to schooling of 2.5 percent.
References
Barrero, Alfonso, Markus Kirchner, Camilo N. Pérez, and Andrés Sansone (2020). “Estimación del impacto del Covid-19 en los ingresos de hogares, medidas de apoyo y efectos en el consumo.” Minutas Citadas en Recuadros IPoM, División Política Monetaria, Banco Central de Chile. https://www.bcentral.cl/documents/33528/2688278/Minuta_Citadas_en_Recuadros_del_IPoM_Dic_2020.pdf
Lucas, Adrienne M., and Patrick J. McEwan (2026). “Building opportunity: the intergenerational effects of Chilean school construction.” Journal of Public Economics, 260: 105702. DOI: https://doi.org/10.1016/j.jpubeco.2026.105702
Schiefelbein, Ernesto (1976). “Diagnóstico del sistema educacional chileno en 1970.” Universidad de Chile, Facultad de Ciencias Económicas y Administrativas, Santiago.
Sistema Nacional de Inversiones (SNI), Ministerio de Desarrollo Social y Familia (2024). “Precios Sociales (Reporte Annual).” Downloaded July 25, 2024 from https://sni.gob.cl/storage/docs/Informe_precios_sociales_2024_SNI-Chile.pdf
Policy Features
- Category
- Education
- Sub-Category
- Child Education
- Beneficiary Type(s)
- Children and Youth, Primary school enrollees, Secondary school enrollees, Students
- Country of Implementation
- Chile
- Year of Implementation
- 1965
- Empirical Method
- Difference in Differences
- Research Type
- Primary
- Peer Reviewed
- Yes
- MVPF Publication Link
- doi.org/10.1016/j.jpubeco.2026.105702