Research Highlights Podcast
January 14, 2026
Diversifying college applications
Nageeb Ali and Ran Shorrer discuss the optimal strategy for applying to college.
Source: TeroVesalainen
Guidance counselors generally advise college applicants to diversify their applications across schools they believe to be safeties, matches, and reaches. Yet, prevailing economic theories of school choice suggest that such hedging strategies are suboptimal and that applicants should focus on applying to the best schools they have a chance of getting into.
In a paper in the American Economic Review, authors S. Nageeb Ali and Ran I. Shorrer show how incorporating correlations among admissions decisions rationalizes the motive to hedge. Their findings highlight the tradeoffs applicants face under realistic assumptions and may offer insights into the optimal design of admission processes.
Ali and Shorrer recently spoke with Tyler Smith about how the admissions process can be correlated and the implications for students.
The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.
Tyler Smith: Your paper is motivated by a very simple question: what is the best strategy for applying to college. Can you tell us a little bit about how you hoped to shed new light on this question?
Nageeb Ali: Let me start off by highlighting some of the prime considerations that a college student is going through. The first thing that a college applicant has to decide is which colleges she likes and which she dislikes. The second dimension that an applicant cares a lot about are the odds of getting into various colleges. They're uncertain about that. Of course, they might have some information about their class rank and about their standardized test score, but they know it's a gamble. And so they will turn to guidance counselors. And guidance counselors generally give advice that you apply to some reach schools, some match schools, and some safety schools. You diversify your portfolio. And that is in some ways the question this paper is asking: what tractable model for this portfolio problem—choosing which colleges to apply to—would give a solution in which you diversify your portfolio in that way. That's what the paper sought to answer and to try to better understand.
Smith: Why did economists think that diversifying wasn't the best strategy?
Ali: This question gets at where the literature was before our paper came onto the scene. Let me back up a little bit and talk about why this portfolio problem is difficult. When an applicant is applying to a bunch of colleges, they know how much they like each of them, but they’re not sure which ones they’re going to get into. They also know that when choosing which schools to apply to, they can't go to more than one college simultaneously. In contrast, when you put money into the stock market, one of the things that makes that portfolio problem somewhat nice is that you're getting your average returns across all your different assets. For applicants, a college is valuable to apply to if they get into that school and they’ve been turned down by all other schools that they prefer. So that's a really weird kind of portfolio problem where a college is valuable only under special circumstances. To make headway on that problem, there was a seminal paper by Hector Chade and Lones Smith called “Simultaneous Search,” published in Econometrica, which had an ingenious solution to the problem. However, for that solution to work, they needed to assume that from the applicant’s perspective, getting into these schools was stochastically independent. What does that mean? It means from the applicant’s perspective, whether they’re rejected by a college or not doesn't affect their odds of getting into a different college. That was their assumption, that things are independent, and out of that assumption came the conclusion that applicants should not apply to safety schools. Instead, they should view this as a gamble where they take multiple shots and apply to multiple aggressive, reach schools, hoping that at least one of them works out. So, the standard theoretical predictions could not rationalize the applications to safety schools that we see in practice.
Smith: Can you explain some of the ways that the admissions process can be correlated across schools?
Ran Shorrer: In fact, the genesis of this paper comes from one such application. I was sitting in a seminar by one of my favorite development economists, Kehinde Ajayi, and she was discussing high school admissions in Ghana. In Ghana, students apply to schools before taking a national exam. If you do very well in this exam, you're going to be accepted to any place you apply to. And if you do poorly, then you'll be rejected everywhere. That's a very mechanical way, which is quite common across the world, to correlate admission decisions across schools. You can think about many other correlating factors that exist outside of national exams. For example, your letters of recommendation—typically you don't get to see them, but if your letters are good, that's going to affect your admission chances across many schools. You could also think that some kids don't understand where they stand very well, and colleges are better equipped at assessing their file than they themselves are. And all of these things mean that when one college accepts or rejects a student, that carries information for them about their odds at other places.
Smith: How do you approach solving for the optimal allocation between schools under correlation?
Shorrer: The idea was basically this: Let's say I'm applying and I can apply to only two schools. How should I think about my second school? Well, I know that the second school is only relevant if the one I like most rejected me. Basically, I should ask myself, given that this school is only relevant if my first choice rejects me, what would I learn if that school rejects me? So I should choose my second school as if I already have been rejected by my first choice schools. And, like a mirror image, when I choose my top school, I should ask myself what my insurance is. If I'm rejected by the top school, I still have this other shot. So, actually, I should aim more aggressively because I have an outside option. So, we have two forces. One is the outside option effect that says you should be more aggressive at the top—you should go for reaches because you have the safety as insurance. And, at the same time, you should be very defensive with your safeties because you already got rejected by all your reaches and your matches, and that is very bad information.
One lesson that comes from this is that when one is being told to diversify one's portfolio, it's really with this idea that you don't want to apply to too many schools that are comparable within the same tier, but you want to diversify across tiers to some degree.
Nageeb Ali
Smith: What broad lessons do you think applicants can take away from this work?
Ali: Our work gives a rationale for the advice that applicants are getting from guidance counselors that, in particular, applying to colleges is a huge problem of uncertainty with really high stakes or seemingly high stakes. Maybe it doesn't matter too much where one goes—one can get a great education at so many schools in the United States—but at least for the applicants, it's a stressful part of their lives. One lesson that comes from this is that when one is being told to diversify one's portfolio, it's really with this idea that you don't want to apply to too many schools that are comparable within the same tier, but you want to diversify across tiers to some degree.
Smith: Do you think there are any lessons for how education policymakers or administrators design the application process? Should they implement processes where there's less correlation?
Ali: This is actually touching upon the next paper that Ron and I are writing alongside a fantastic PhD student, Salvador Candelas. In that paper, we're actually highlighting the benefits that correlation introduces for applicants. The central idea is that even though correlation makes the portfolio problem a little bit trickier, whether you have correlation or not affects the thresholds or standards that colleges use for admissions. In a world in which you have more correlation, colleges are going to have to compete more to be able to attract applicants. We show that by competing more, a higher fraction of applicants are going to be able to get into better choices than when we have too much independence. So even though, as I said, it might seem one punch line from this paper is that correlation makes things harder, actually, in equilibrium, things wash out in a way that it makes applicants better off.
Shorrer: I want to tie this back to something that Najeeb said before about financial portfolios and how they're different from school application portfolios. When I apply to a bunch of colleges, if I get accepted to seven of them, eventually I'm only going to get one spot. So, actually, even though there is this fallacy in our mind of thinking that if some people win more, maybe it's at the expense of others, that's actually not the right intuition. From the policymaker perspective, you want to maximize choice. You want to have low yield rates. You want to have a lot of kids accepted to many colleges and let them sort according to their geographic preferences rather than being compelled to go to the one school that accepted them. From the policy perspective, I think our new paper suggests that we really want to push down the yield rates rather than having them so high, which is something that schools are incentivized nowadays to do.
♦
“Hedging When Applying: Simultaneous Search with Correlation” appears in the February 2025 issue of the American Economic Review. Music in the audio is by Podington Bear.