Skip to main content

College Choice

 Last update: September 11, 2018

My Ph.D. dissertation concerns the determinants of schooling choices. In one paper, I study how the education experience of a sibling impacts college choices. In another, I measure the returns to Swedish college fields and evaluate if applicants sort by comparative advantage.

I use a unique data set of all applications to Swedish universities, and identify causal relationships by exploiting lotteries used to break ties in the application process.

All projects were pre registered at the OSF before I was given access to the data, to avoid any risks of accidentally exploiting the many degrees of freedom in research design that are available when using a large data set such as this.

Sibling Spill-overs

In “Sibling Influence on College Choice” I study how preferences for college are influenced by the education experience of siblings. It answers the question; If your sibling studies business at Stockholm School of Economics, are you more likely to apply there? The answer is yes, up to 50% as likely!

It is well known that children with educated parents are more likely to go to university. Sociologists often stress the importance of familial influence, citing the lack of role models as one important reason for the persistent inequality in schooling. But how exactly does the familial transmission of education preferences work? Is it just that people are from either college-going families or not, or do the experiences of family members matter for what one decides to study?

Finding the right college education is not easy, especially without enough information. College applicants have been shown to not know enough about either the costs or the benefits of different degrees to make good decisions. Could an older sibling help fill this information gap?

To study this question I collected a large data set of admissions to Swedish universities. By looking at applicants who have scores exactly at the cutoff, where a lottery determines if they receive an offer to a program or not, I can isolate the effect of a sibling’s admission. Around 20% of sibling pairs decide to study at the same university. But most chose to do so because they have correlated preferences. They would go there even if the other sibling was not admitted. I isolate the spillover effect and see that the likelihood that an individual applies to a specific school increases by about five percentage points (or 25%) if their sibling has already been admitted.

By studying different groups of siblings separately, I identify a number of interesting patterns. Siblings are five times more likely to follow each other into the same institution compared to the same field of study. The effect is stronger when both siblings are male but does not vary with the education level of the parents, or with the popularity of the program. I conclude that the patterns that I observe indicate that siblings follow each other mostly because it is convenient.

Returns to College Fields

It has been said that all economist try to measure the returns to education at some point in their career. “Relative Returns to Swedish College Fields” is my attempt. The paper is a methodological replication Kirkebøen et al. (QJE, 2016), where I estimate the economic returns to different fields of study in Sweden.

Sweden has the lowest college premium in the developed world. At the age of 30, the difference in yearly earnings between individuals with a college education and those without is only $6,000. However, among university degree holders, the variability between fields of study can be almost three times as large.

It is complicated to properly separate the effect of education from the fact that the most productive students attend the best schools. In Sweden individuals with humanities degrees earn $22,000 on average, while those with degrees in the best paid fields have wages that are twice as high. Is this because the returns to humanities are low, or are humanities students just less productive? By studying applicants at the margin, I am able to control for this selection bias. I find large differences in returns between fields.

Medicine, Engineering and Business often yield premia of over $10,000 per year compared to other fields. Humanities, on the other hand, frequently has a large enough negative payoff that the applicant ends up earning less than those who do not go to university at all.

When compared to other countries, Swedish applicants are more likely to apply to fields where they can expect a negative payoff. With the lowest education premia in the developed world, even the poor returns to humanities are not that discouraging. It is possible that Swedish students simply care less about future earnings when choosing what to study. Instead, they might prioritize the consumption value of college and study topics that are fun or interesting, rather than tailored to the demands of the labor market.

Work in Progress

  • Relative Returns to Swedish College Fields

    This paper studies the economic returns to fields of college education, using a unique Swedish data set of applicant preferences and university admissions. Field payoffs eight years after initial application vary substantially by field, but not nearly as much as in Norway (Kirkebøen et. al., 2016). Medicine, engineering, and business degree holders enjoy positive returns of around $10 000 per year, when compared to most other fields. Humanities, which is the worst paying field, incurs losses of almost as much. These results are causally identified using a regression discontinuity approach and confirmed by admission lotteries. In contrast to many other countries, the findings indicate that Swedish students do not necessarily prefer fields where their potential earnings indicate that they would have a comparative advantage.

  • Sibling Influence on College Choice

    How are college choices shaped by the experiences of peers? Using registry data of applications to Swedish universities, I study how an individual's education experience influences the college applications of their siblings. Tie breaking lotteries at admission margins provide causal identification. I find that successful admission to a specific institution-program combination increases the likelihood that a sibling ranks that specific combination as their most preferred option from 2 to 3 percent. Siblings are five times as likely to follow one another to the same institution compared to the same field. The effect is stronger when both siblings are male but does not vary with the education level of the parents or with the popularity of the program. The observed spillovers seem to mainly be driven by a demand for convenience rather than the transmission of information between siblings.