Hoxby and Avery begin to help us understand why by evaluating the merits of several urban legends, two of which are repeated by seemingly every elite college admissions officer in the country:
(1) "We don't have more Pell recipients on campus because there are simply too few students from low-income families academically qualified to get in here." In other words, bright poor kids just don't exist in large numbers.
(2) "We don't have more Pell recipients on campus because every qualified student from a low-income family is already taken by another great college." This is just a twist on the first claim.
As I've long suspected, and have argued at my own institution for many years, the data prove both of these statements wrong. First, there are plenty of very bright students from low-income families graduating from high school. In fact, very high-achieving, high-income students only outnumber high-achieving, low-income students by 2.5:1. According to the authors, "there are at least 25,000 and probably something like 35,000 low-income high achievers in the U.S." Second, many of these amazing students aren't attending any great college. In fact, they aren't even applying.
So there's the main punchline of their new paper: high-ability low-income students do exist, and they aren't in college because they haven't applied.
Moving on from there, Hoxby and Avery stroll onto shakier ground. The title of the New York Times article their hypothesis about why these students aren't applying to top colleges: "Better Colleges Failing to Lure Talented Poor." The authors suggest that problem resides with admissions officers who aren't using appropriate methods to recruit these students. They go through many analyses to show why this is likely a problem, and their arguments are compelling. Every enrollment manager needs to consider them.
But this is hardly a full accounting of every possible explanation. I'd like to offer this headline instead: "Better Colleges Charge a Lot of Money and Scare Off the Talented Poor."
Everyone knows that a top college comes with a high sticker price. Hoxby and Avery dismiss this by noting that the net price is often lower at top colleges because they also provide more financial aid. But assuming that net price matters more than sticker price is a major assumption; it is guided by economic theory but rarely ever tested. My own work suggests that one reason it may not bear out in practice is that responding to sticker price, rather than net price, is not merely a reflection of information asymmetries (e.g. lower-income families are less likely to know about the availability of financial aid), but also a reflection of differences in the extent to which families trust the major social institutions to actually help them.
Hoxby and Avery do not acknowledge this. They frame the decision not to attend a top college as irrational, writing that "added to the puzzle is the fact that very selective institutions not only offer students much richer instructional, extracurricular, and other resources, they also offer high-achieving, low-income students so much financial aid that the students would often pay less to attend a selective institution than the far less selective or nonselective post-secondary institutions that most of them do attend....[the students'] choices are odd because although the private match colleges might offer fewer scholarships that are explicitly merit-based, they offer much more generous need-based aid so that the student would pay less to attend and would enjoy substantially more resources. Furthermore, it is almost never sensible for a low-income student to apply to a single private, selective college: he may be able to use competing aid offers to improve the aid package he gets from his most preferred college." Clearly, if only students from low-income families used the net price calculator, they'd opt to apply-- right? How do we know this?
Hoxby and Avery also dismiss the idea that sticker shock explains the lack of low-income students at expensive colleges because they observe few income differences in the behaviors of college applicants. As in other studies, among students who apply to college, low and high income students behave similarly.
So what? This simply means that the effects of sticker shock may affect the application decision but not the enrollment decision, and this makes utter sense. Trust in schools plays a major role in families' educational decisions. The fact that over the last 30 years the sticker price at most top colleges and universities has skyrocketed, rising more rapidly than inflation, and making headlines, leaves little reason for people with little financial strength to trust in them. Moreover, those sticker prices rise even after their kids enroll, their financial aid packages change from year to year, the FAFSA must be refiled again and again, and grants are often replaced with loans once a school has lured the student in the door. Why would anyone trust these places? Why would smart students who've watched their families suffer in an exceedingly wealthy society like this one trust the system to give them the money required to make attendance at an elite college possible? Especially when they won't be "shown the money" until after they have applied, been accepted, and are on the brink of enrollment? Furthermore, we have the gall to expect them not only to trust these schools but to believe that it's possible to negotiate with them? The act of applying is an act of trust and faith in what's widely understood to be a highly rigged system, so looking at the behaviors of applicants post-application makes little sense.
The Hoxby and Avery view of the world is pervasive among researchers, and leads many economists to suggest that raising tuition is no big deal if you can simply "hold students harmless" with financial aid. We suffered through a regime of that leadership here at UW-Madison several years ago, and while many departments are enjoying the extra revenue from the tuition, the representation of first generation students on campus is falling. None of our institutional reports examines the impacts of our tuition hikes on applicant behavior. By solely comparing our figures to college applicants, or worse yet enrolled students, you can convince yourself that tuition hikes do no harm. But when you see such clear evidence that talented poor students stay out of the applicant pool, you really have to wonder. Do they simply not know that your college exists, or how to apply? Or might they actually believe that they cannot afford it?
They aren't wrong to feel this way. Let's take the case of UW-Madison, where most professors and students on campus suggest that our price is still affordable because we distribute financial aid. Consider the net price of college attendance, taking the sticker price and subtracting all grant aid. (My team's research in Wisconsin reveals that students do not think of loans as financial aid, do not feel they are 'help' and endure quite a bit of stress from them.):
My graduate student Robert Kelchen produced these figures for the 3,487 first-time, full-time, degree-seeking freshmen who are in-state students. The first thing to note that we can only calculate net price for the 1,983 students receiving Title IV aid. This means that just under half of all in-state Madison freshmen do not file the FAFSA. Is that because they aren't qualified for aid, or because they do not know they can or should? Second, look at the net price faced by students according to their family income:
$0-$30k: $6,363 (n=212)
$30L-$48k: $10,098 (n=232)
$48K-$75k: $15,286 (n=406)
$75-$110k: $19,482 (n=542)
$110+k: $20,442 (n=591)
Ok, so now you look at these numbers and think, hmm. Is $6,363 a good deal for a year of college?
Well, let's consider that relative to their family income. Here is the cost burden relative to the midpoint income in each bracket:
110+k: 18% or less
There you have it. At UW-Madison we expect the poorest families to contribute far more of their family income to attend college than we expect from the richest families. Yes, the poorest folks get more aid, but it's far from proportional to their need. Why would we expect them to buy into such a system?
Researchers, and especially economists, need to get out in the world and talk to real families and students who do not enjoy the benefits of tenure and stable incomes. Only 16% of parents nationwide will ignore the cost of tuition when helping their kids choose a college, according to a new survey, and I'm betting this is even less common among poorer families. The survey also shows that one in five parents refuses to take on any debt for their kids to go to college, and among families earning less than $3,000 a month, 25% aren't sure if they'd take on any debt, and only 28% are willing to borrow $20,000 or more. Barely one in five of these poor families think it's reasonable for their child to accumulate even $20,000 in debt over four years.
Yes, there is a fair bit of financial aid out there, but it increasingly comes in the form of loans or scholarships with strings attached which can be easily lost, creating financial instability. An alternative interpretation of Avery and Hoxby's results is that rather than taking a chance on expensive institutions they can't trust, many high ability low-income students are revealing a preference for lower-priced institutions. They might occasionally apply to one pricey dream school, but it's more likely reflecting a quick momentary aspiration, rather than an actual plan. They live in the 21st century where affordability is an indicator of quality, not the 1980s where the pricier the college or restaurant, the better it must be. They know they are living proof that the nation isn't a meritocracy; they see their families work hard every day and get nowhere. Ignoring their savvy and discounting it as irrational, providing them with net price calculators rather than investing in high-quality free public options-- well, it's actually sort of insulting. It's time to consider the possibility that getting smart striving young people in search of upward mobility to apply to and attend good colleges will require transforming our system to focus public investments on ensuring real access and opportunity for all Americans, rather than subsidizing the choices of the middle class. The current system of voucher-provided financial aid is now decades old and according to this study, it may be failing. It's time to consider a restart.
Note: I thank Zakiya Smith for a helpful clarifying Twitter conversation that led to some revision.
Postscript. As it turns out, Hoxby and Sarah Turner have been conducting a very important randomized trial on the effectiveness of several intervention strategies aimed at helping more talented poor students enroll in college, and I've finally gotten a look at the forthcoming results. They show that a combination of better information about college graduation rates, information on the net price of various colleges, and fee waivers for applications is effective at boosting application and enrollment rates of the talented poor substantially. Moreover, their strategy is cheap! This is wonderful news. But what does it tell us about the importance of sticker shock-- the topic of this blog? Unfortunately, not much. In this case, the intervention most clearly aimed at addressing sticker shock is a net price calculator. That sort of intervention assumes that the "problem" of sticker shock is mainly informational. As I said above, I think that the issue of sticker shock involves more than information, it also involves trust. Being told that a college is likely to give you aid is not the same thing as getting the aid. In their trial, the authors find that the impacts of the net price calculator are significant but not as important as the other interventions. Again, it would be wrong to interpret this to mean that sticker shock is unimportant, since this is not a test of sticker shock. It is a test of a net price calculator. In contrast, the fee waiver information they provide is real money, on the table, with a high probability of receipt. They find that this has larger effects than the net price calculator, and I think that supports the idea that a real concern about money actually being available is inhibiting the actions of these talented poor students. For sure, it's not the only factor affecting their behavior, but it's worth attending to since even with this entire package of Hoxby/Turner interventions, a substantial fraction of the income gap in college behaviors remains among talented students. Moreover, a large income gap persists among less-talented, but nonetheless very important, students who aim to attend and graduate from college.
Postscript 2. The numbers I quoted above are clearly out of data based on UW-Madison's website. The situation seems to have worsened.