Do uncertain labor markets affect Indonesian high school graduation rates?

Educational decision-making is complex. The share of U.S. high school students going to college has decreased from 70% in 2009 to 62% in 2022. Part of the reason is the higher cost of college. It reduces what economists call the skill premium: the return on investment in the form of higher salaries for college-educated workers.

That calculation focuses on average costs and benefits. But uncertainty—about the impact of AI on the U.S. labor market, for example—also plays a role. To estimate how uncertainty affects educational decisions in an emerging economy, recent AAE graduate Jacob Bills chose the example of Indonesia, where he lived as a teenager and young adult. Despite government investments in secondary education and growing numbers of Indonesians with a high-school diploma, 25% of students were still dropping out of grades 10 to 12 in 2014.

“Understanding why students decide not to complete upper secondary school is important as Southeast Asia’s largest economy continues to evolve after major political changes in the late 1990s,” says Bills, whose PhD advisor was Ian Coxhead. “I wanted to assess if uncertainty about the skill premium in regional labor markets played a role in these decisions.”

Indonesia was an ideal setting for the question of interest. First, it is an archipelagic nation of 288 million people, who inhabit more than 6,000 islands. This unique geography means that industries tend to cluster within islands, making regional labor markets more relevant than the national one. Second, the 1997 Asian financial crisis spurred Indonesia’s transition from three decades of President Soeharto’s dictatorship to a modern democracy. This included a devolution of economic and educational decision-making power from the national to regional and local governments.

Structurally, the Indonesian economy changed from 48% to 28% of workers in agriculture and 38% to 49% in the service sector between 1991 and 2019. Societal changes included a rapid rise of Indonesians with a high school diploma from one generation to the next: 75% of children born between 1980 and 2000 versus just 22% of their parents.

However, regional economies often struggled to create enough new jobs for higher skill levels, and the un(der)employment of young adults with high school or college degrees attracted substantial media attention. “This was a big enough to topic to be featured in a comedy film that became Indonesia’s 2010 candidate for the Academy Awards,” notes Bills.

Taken together, these changes created the kind of dynamic and uncertain regional labor markets whose impact on educational decisions Bills wanted to assess. Guided by Coxhead, who had studied a related question in Vietnam, Bills combined two datasets spanning 1999 to 2014, the period after the Asian financial crisis.

The first included three waves of the Indonesian Family Life Survey, which began in 1993 and follows a representative sample of Indonesians over the course of their lives. Bills analyzed data on the families, homes and educational histories of about 9,000 teenagers who reached the ages of 14 to 19 and completed at least grade 9 during the study period.

The second dataset combined annual government statistics about county-level unemployment rates and wages for skilled and unskilled workers. Knowing the county where the teenagers lived at each survey wave allowed Bills to merge individual with regional labor market data. As expected, the spread, or variance, of the wage distribution was higher for skilled than unskilled jobs and grew from 1999 to 2014.

“Acquiring specialized skills creates new and often higher-paid jobs but also narrows an individual’s career opportunities,” explains Bills. “Skilled jobs are more diverse than unskilled jobs, and some grow faster than others. This increases the variance of the skilled-wage distribution over time.”

He points to the example of Indonesia’s visa-on-arrival program to help attract tourists. It increased the demand for English-speaking workers to explain the visa process to visitors and drives up their salaries. This greater demand for a skill taught in grades 10 to 12 resulted in a higher mean and variance of the wage distribution.

The answer to the study’s main question? Neither a less certain labor market nor a higher skill premium affected the probability of students dropping out of upper secondary school. Instead, the most important determinants were the father’s education level, gender and family size.

Girls and boys whose father had graduated from high school or college were less likely to drop out; the mother’s education played a smaller role. Girls were more likely to drop out than boys, especially when they were the oldest sister with several younger siblings, while family size had no impact on boys. This was consistent with other studies. Bills also found that 25% of risk-averse students dropped out—perhaps perceiving an unskilled job as the safer option—compared to 17% of more risk-tolerant students.

A diversified and highly skilled workforce is necessary for an emerging economy like Indonesia to become and remain globally competitive. Thus, although dropout rates have fallen since 2014, continued efforts to reduce them may be warranted. These could include counseling, especially for girls who may be expected to care for younger siblings or accept an unskilled job to help provide for their families, and school surveys to detect early-warning signs, such as low motivation, bullying or poor mental health.

“It makes sense to me that parental values and resources have the biggest impact on a child’s education and future career choice,” Bills concludes. “I have also met many Indonesian parents without a high school education who really want their own children to do better. Thus, keeping parents and students informed about average starting salaries for a variety of skilled jobs may be one of the best ways for Indonesia to further increase its graduation rates.”