Essays in Education and Labour in Indonesia

This thesis consists of one introductory chapter and three essays in development economics, focusing on education investment and labour dynamics in Indonesia. The first essay analyses the impact of trade liberalisation in the form of plant activities on schooling options for children aged 16 years old who live in the same district as the new plant. This essay provides empirical evidence that the expansion of export manufacturing in Indonesia during the 1990-1996 boom and bust period changed children’s educational attainment. Using the variations across districts in the yearly change of export manufacturing employment per working-age person, this paper finds that school dropout rates increased as local export manufacturing industries expanded. The paper concludes that export expansion had the biggest negative effect decreasing years of schooling and the probability of graduating from senior high school. Additionally, the paper discovered a gender disparity in educational attainment in that the impact of export manufacturing employment had a more significant negative effect on females. The second essay employs machine learning techniques and data for individuals aged 18-20 from a nationwide household survey dataset in Indonesia to establish a predictive model of schooling duration. The machine learning approach identifies the key factors that influence schooling duration from high-dimensional factors covering socio-economic, health, parental, community, and behavioural aspects. The paper concludes that internet knowledge, possession of a computer and fridge, a savings account, the ability to read and write in the Latin alphabet, and unmarried status are among the top predictors of years of schooling in Indonesia. Our predictive model can help the Indonesian government develop an effective early intervention by detecting students who are more likely to drop out. The third essay investigates the impact of Program Indonesia Pintar (the Smart Indonesia Program), a conditional cash transfer for educational purposes, in Indonesia. During the COVID-19 outbreaks, governments shut schools to ensure continuity of education and break the virus's chain of transmission. During that period education was carried out entirely online. As a developing country, Indonesia faced increased social issues since many pupils needed access to online teaching and learning resources while Indonesian families were also experiencing decreasing income and increasing expenses during the pandemic, and this led to greater strains on disadvantaged households with increased risk of children dropping out from school. So, Indonesia expanded the PIP in attempt to reduce dropout rates among underprivileged students during the pandemic. However, the economic shock associated with the pandemic may also have forced students to use the PIP money for other short-term non-educational expenditures. This paper employs regression discontinuity design to evaluate the effects of the PIP distribution among households. I found that the PIP helped the groups below its cut-off (the 40th percentile of household expenditure) who received the PIP to have better internet access and need less government assistance for their internet data quota. However, the PIP recipients spent the PIP not just on education needs but also on other basic needs such as new clothes, accommodation, and transportation costs.