Does investment in school construction promote higher educational attainment—and do the effects improve students’ later lives and those of the next generation? This column examines the impact of over 61,000 primary schools built by the Indonesian government between 1973 and 1979, almost doubling the number in the country. The evidence shows that the men and women who accessed education provided by the construction program benefited from significant improvements in their educational and later life outcomes. So too did their children.
Governments in developing countries spend approximately one trillion dollars annually on education, and households spend hundreds of billions more on educating their children. What are the long-term returns on all this investment in terms of educational attainment, people’s later life outcomes, and overall economic development?
Public spending on education can take many forms, including teacher interventions, school input programs, cash transfer programs, and of course a common and direct form of education spending: building schools. While these programs are often motivated by the belief that increases in education will translate into higher economic development and growth, there is considerable debate about the effectiveness of schooling in boosting economic growth.
Indeed, it is often difficult to measure the effects because the choice of how much education to obtain is correlated with many different characteristics of individuals, households, and communities.
In recent years, progress has been made using ‘randomized experiments’. But a review of over 100 primary school interventions in developing countries finds that the vast majority of evaluations take place shortly after the intervention has ended. This is unfortunate because the ultimate objectives are improvements in later life outcomes and overall economic development. It remains unclear whether a program’s effects persist or fade out over time.
School construction in Indonesia
In recent work, we study the impact of one of the largest primary school construction programs ever completed on a wide range of long-term and intergenerational outcomes, many of which have not been studied before.
Between 1973 and 1979, the Indonesian government constructed over 61,000 primary schools, almost doubling the number in the country. This massive school construction program was partially funded by the oil boom that resulted in increased government revenue.
The program was designed with the specific goal of reducing regional disparities. As such, it targeted areas where primary school enrollment was low, especially in the outer islands of Indonesia.
Figure 1 presents a map of Indonesia showing the geographical distribution of the number of schools constructed in each district. On average, the program added over 200 schools per district or two schools for every 1,000 children of primary school age.
Figure 1: Spatial distribution of schools constructed between 1973 and 1979 in Indonesia
The government recruited and trained new teachers and paid their salaries for these newly constructed schools. But according to the World Bank, the percentage of teachers who met the minimum qualification of having an upper secondary school degree did not change over this period.
We use nationally representative 2016 data from Indonesia, which contain information on a wide range of outcomes related to education, employment, migration, living standards, taxes, and marriage outcomes. We use ‘difference-in-differences’ analysis based on variation across geographical districts in the number of schools built and across birth cohorts in whether they benefited from access to the new schools.
Long-term effects of education
Figure 2 summarizes the main findings and shows large improvements across a broad range of outcomes (each combined in an index). Unsurprisingly, school construction leads to improved educational outcomes. Previous work has shown this for men—and we confirm that it also improves women’s education.
As adults, men who benefit from school construction are more likely to be employed, to work in the formal sector, to work in the non-agricultural sector, and to migrate. Women who benefit from school construction are more likely to migrate and to have fewer children.
Households in which either parent benefits from school construction have higher living standards, better housing, more assets, and pay more taxes. While nutrition and health investments increase, we do not find any improvements in health outcomes. School construction leads to improved marriage market outcomes with spouses being more educated, more likely to be literate, and more likely to have migrated.
Figure 2: Effects of school construction on indexes of long-term and intergenerational outcomes
Figure 3 highlights the gender dynamics of the effects of school construction on education. It shows the change in the likelihood of completing at least a certain number of years of education. For women, the effects are concentrated in primary school; while for men, the effect extends throughout lower and upper secondary school.
Figure 3: Effects of school construction on the probability of a first-generation individual attending at least n-years of schooling
Intergenerational effects of education
Parents who benefit from school construction transmit those benefits to the next generation.
Figure 4 shows the likelihood of a second-generation child completing at least a certain number of years of school. We explore the effects depending on whether the father or mother benefits from school construction and whether their child is a son or daughter.
We find no effects during primary school because that is almost universal for second-generation individuals. But the effects extend throughout secondary and tertiary education. Relative to baseline levels, the largest impacts are in tertiary education with effect sizes that indicate a 20–25% increase in the likelihood of the second-generation child completing university.
Figure 4 also reveals important gender differences. Parents benefiting from school construction see larger impacts for their daughters than for their sons, and a mother benefiting from school construction has a larger intergenerational impact than does a father.
By extending the focus of earlier work on men to study the impact of school construction on women, we find gender differences for both first and second-generation outcomes. We explore the mechanisms that drive the intergenerational transmission of schooling. Marriage market outcomes appear to play a crucial role, particularly whether the spouse has completed primary school, is literate, works in the formal sector, or works outside agriculture.
Figure 4: Effects of school construction on the probability of a second-generation individual attending at least n-years of schooling
Rate of return and fiscal impacts of school construction
To quantify the policy implications, we conducted a cost-benefit analysis in which we created an accounting model to calculate the costs of school construction and subsequent discounted benefits for the government in terms of increased tax revenues and overall improved living standards for the Indonesian population.
Our cost-benefit analysis highlights that under all reasonable assumptions, school construction pays for itself in terms of additional expected tax revenues, not to mention the additional benefits of improved living standards.
We find that school construction leads to increased tax revenues that directly offset construction costs within 40 years. In addition, accounting for improved living standards of the Indonesian population reveals high internal rates of return, in the range of 13–21%, and benefits surpassing costs within 17–30 years after the schools were built.
Furthermore, given the observed intergenerational transmission of education, the likely long-term benefits are even larger. These results provide strong support for the cost-effectiveness of school construction interventions.