If you want to work for one of Wall Street’s leading investment banks, law offices or consulting firms you’d better graduate from a prestigious college like Harvard, Princeton or Yale. These elite firms overwhelmingly hire graduates from these highly ranked colleges.
Why does college reputation matter so much to employers and students? Economists argue that reputation matters because it conveys information about the expected quality of a firm’s product.
When product quality is hard to ascertain, consumers seek information to make purchasing decisions. This is particularly true for “experience” products such as a bottle of wine, healthcare or a college education. Physical examination of these kinds of products does not reveal their quality—beautiful labels and modern athletic facilities do not predict whether a wine will taste good or whether a college will provide a good education.
Experience goods are typically purchased based on reputation or recommendation. This is why consumers are willing to pay more for wines that are highly rated by the likes of Robert Parker, why referrals are crucial when choosing doctors and why prospective students and parents pay so much attention to US News and World Report college rankings. Unlike clothing or electronics, which typically come with money-back guarantees, wine, healthcare and college education are products you need to experience to decide whether they ultimately represent a good value proposition.
So reputation matters a lot for students choosing which college to attend. Exactly why reputation matters so much to employers is much less clear, however.
In a new paper we aim to explain why. Our premise is that college reputation matters insofar as it conveys information about the abilities that employers can reasonably expect from graduates of a given college. Colleges with a strong reputation may just be cherry-picking students that will succeed no matter which college they attend. Colleges may actually do something to students that employers value. Or colleges may do a mix of both.
We set our study in the country of Colombia. College markets in Colombia and the US have many common features such as a similar mix of public and private colleges, and a tendency for students to attend colleges located near their homes. One key difference between the two countries is that the Colombian government introduced a nationwide college exit exam in the mid 2000s. This exit exam provided new information about students’ abilities upon graduation. Our key conjecture is that if college reputation stems from cherry picking, when the exit exam becomes available employers should rely less on reputation when they hire graduates and more on other measures of student ability.
Our study takes advantage of the staggered roll out of the exit exams across 55 fields like accounting, dentistry, economics, and law. This rollout leads to different levels of exam taking across college majors, and allows us to analyze how exit exam affects the earnings return to reputation and the earnings return to incoming student ability as measured by college entry scores. If employers use college reputation to infer individual ability, then the introduction of the exit test should reduce the return to reputation and increase the return to admission scores.
We find that college reputation matters to employers because it is a cheap and easy way to access a large pool of pre-screened candidates—all students at the top colleges have high college entry scores. In other words, we find strong support for the cherry-picking hypothesis. But we also find that colleges with a strong reputation also do something good to students. While we cannot pinpoint exactly what it is, colleges with a strong reputation may provide better learning environments, have extensive alumni networks or do something else.
College reputation, however, is an imperfect measure of student ability. Once exit exams become available, we find that employers rely less on college reputation and more on individual student ability to set earnings. In other words, the exit exams make it more likely for employers to discover potential stars at lower ranked schools.
Given the similarities in the Colombian and US college markets, we conjecture that when American firms use college reputation to decide whom to hire, they rely on an imperfect measure of student ability. Introducing exit exams may, therefore, provide employers with a more accurate characterization of student ability, thus making it easier for firms to identify the best candidates regardless of college reputation.
Some firms are already aware of their excessive reliance on college identity to infer student ability, consulting giants Deloitte and Ernst & Young are switching to college-blind hiring processes. This change represents an effort to emphasize a wider array of student accomplishments and to encourage applications from talented candidates from lower-ranked colleges.
With top colleges in the US admitting about 6 percent of applicants, introducing a college exit exam would be great news for the majority of students nationwide. States with large college systems—think of California’s College System with its three tiers of state-financed colleges and over 70 private colleges— may be well placed to pioneer the introduction of an exit exam, such as the existing Collegiate Learning Assessment. Doing so will likely promote greater student effort and lead to a more diverse and competitive pool of job applicants from across the state.
This post was co-authored with Bentley MacLeod, Professor of Economics, and International and Public Affairs at the Department of Economics at Columbia University, Evan Riehl, PhD Candidate at Columbia University, and Miguel Urquiola, Professor of Economics and International Affairs at the SIPA and Economics Department at Columbia University. For more information about the research featured in this post please see:
MacLeod, Bentley, Evan Riehl, Juan E. Saavedra and Miguel Urquiola. 2015. “The Big Sort: College Reputation and Labor Market Outcomes.” National Bureau of Economic Research Working Paper w21230 and CESR/Schaeffer Working Paper 2015-013