A future in another city? How migration costs shape young people's college choices: Guest post by Merve Demirel

This is the first in this year’s series of posts by PhD students on the job market. A central puzzle in economics is how massive income disparities over space can persist. If the gains from migrating are so large, why do people stay in places with fewer opportunities? Economists often infer that there must be costs that prevent them from leaving. Some of these costs, such as the monetary cost of transport, can be observed and counted. But estimates of these costs are usually not high enough to explain why people don’t move, suggesting other costs may matter more. Other costs—feeling unsafe, lonely or unhappy away from home—can be harder to observe and quantify. As a result, we know less about their importance in shaping migration decisions, and whether people anticipate them accurately. This matters for policy design: if people stay put because of excessive fears of what they will experience if they migrate, then information could help alleviate these fears, and a contact who has already moved may be the best person to provide it. But if their beliefs about the costs of migration reflect true conditions at the destinations, policy that aims to encourage migration may instead need to make life away from home safer and easier. To understand the expected costs of migration, and whether expectations change through contact with a migrant peer, in my job market paper I ran two experiments with 1,682 high-school students in two districts of Maharashtra, India, together home to more than seven million people. I followed them through their final year of school as they made their first major migration decision, and one of the most consequential of their lives: where to attend college. My first experiment, a choice experiment, reveals that migration costs are large, rise sharply with distance, and are tied to students’ expectations about safety, happiness and loneliness away from home. My second experiment evaluates a mentoring program that connects high-school students with college students who had left home to study, showing that such a contact can raise applications to colleges elsewhere in students’ own district from 18 to 23 percent, but not to the major education hubs farther away, where mentors study. What does educational migration offer? Students in my study come from 52 schools, which are broadly similar to public schools in the two districts. They are preparing to apply to college, and interested in learning more about their options. Many can continue studying at home, but looking further away, beyond their district can expand the number of programs they could typically qualify for by tenfold. Many of the most selective programs are in the state’s major education hubs, such as Pune and Mumbai, which also offer better-paid jobs. Students understand the potential returns to migration. More than half expect studying away to lead to jobs paying higher salaries after graduation, even after accounting for living costs. Yet the majority still prefer to study near home, even among those who expect to earn more elsewhere. This points to migration costs high enough to outweigh those gains. Experiment one: how large are migration costs? The magnitude of the salary gains that students would be willing to give up to avoid migration reveals the value of the perceived costs of migration in monetary terms. The first experiment therefore presents students with pairs of hypothetical colleges and asks which they would choose if admitted to both. The colleges’ locations and their graduates’ expected earnings vary experimentally, and tuition and living expenses stay fixed within the student’s budget. Since the survey also served as the application to the mentoring program, and responses were used to match students with mentors, students had an incentive to answer truthfully. On average, students are indifferent between colleges at home and away only if migrating more than doubles their expected salary, with women requiring about three times as much as men. The required increase rises with distance: students see moving to a nearby city within their district and moving to major educational hubs several hours away as substantially different choices (Figure 1). To unpack these costs, I also asked students what they expect life to be like in each destination. Away from home, they expect to feel less safe and less happy. They also anticipate lower parental approval, weaker ties with family, fewer friends, and greater difficulty managing daily life. Women are particularly concerned about safety, parental approval, and what others would think. Experiment two: can connections to a migrant peer decrease expected migration costs? How well-informed these fears are is unclear. Students have few sources of information about colleges and student life elsewhere. Nearly 30 percent know no one to ask for college advice, and half have no relatives or friends in other destinations. Those who do know someone in a city expect to be happier and safer in that city, but they may differ in other ways from those who do not, so this pattern cannot tell us how providing a connection would change beliefs and choices. That is the question the second experiment asks. For this experiment, I developed a phone-based peer mentorship program to help students make more informed college choices. I randomly offered half of the students a place in the program, which ran throughout the academic year. Each student was matched with a same-gender college student who had left home to study in Pune or Mumbai, two of the most common destinations for those who move outside of the district, where possible in the field and the city that the student was most interested in. They spoke about college choice and applications, but also about other aspects of moving to that city: housing, transport, safety, budgeting, making friends, and what daily life in a new city could look like. Students offered mentorship became more likely to apply to study away from home, by four percentage points from a control mean of 34 percent, and to move away, by three points from 22 percent. A pre-specified index of educational migration that combines overall progress in higher education with migration rose by 0.16 standard deviations. However, the induced applications and moves stayed within the home district, where migration costs were smaller to begin with (Figure 2). There was no corresponding increase toward the more distant hubs, where the mentors study. The effects were driven by men, whose applications away from home rose by ten percentage points. Women’s did not change, which may not be surprising given the much larger migration costs they expect. Beliefs at the end of the experiment follow the same pattern: mentorship lowered what men expected it would cost to live in the nearest city. It did not improve students’ expectations about life in the more distant cities. Among women, those expectations became, if anything, less favorable. What this means for policy Migrating for education could help reduce spatial inequality and ensure that students access the best educational opportunities. But the two experiments show us that migration costs are a major obstacle to these moves. Light-touch interventions providing information on living costs, or scholarships, may be effective for the short moves, where migration costs are small. Moving farther to large cities is different. It is where the potential gains are largest, with access to more selective programs and better-paid jobs. But students expect much larger practical, social and emotional costs when moving far. Talking to a peer who had left home to study neither improved those expectations nor made a distant move more attractive. A natural response to large migration costs is to bring education closer to students, for example by expanding colleges in smaller towns. But to the extent that good jobs require migration after graduation, investing only in local education may leave students in places with fewer opportunities. If policy instead aims to connect students to educational opportunities farther from home, this might mean lowering migration costs at the destination, such as providing safe housing, practical help with moving, or support that continues after arrival. Merve Demirel is a PhD student at the Institute for International Economic Studies, Stockholm University.