My oldest son was born in 1994. I was an assistant professor at Wellesley College at the time. Two years later, I received an offer to take a one-year leave of absence from Wellesley and work as a senior economist in the White House Council of Economic Advisers. I accepted the position, and my family moved to Washington, D.C. My portfolio at the CEA included labor market, education, and welfare policies, and a big issue at that time was how to design tax policy to help reduce the cost of higher education. One result was the introduction of tax-deferred college savings accounts, so-called 529 plans, in 1996.
Families want to know what college is going to cost them, not some average that may or may not reflect their situation.
My son was 2 years old then, and we had another son early the next year. I understood the benefits of 529 plans very well, so I opened accounts for each of them and started funding them immediately.
When my older son was 14 years old, I wanted to know whether I needed to continue making those contributions. Had I already saved enough for college? As a professor, I make a good living, but not so much that paying for college would be easy. Knowing how much I needed to save required knowing how much college would cost. I wondered whether our family would be eligible for any financial aid and, if so, how much.
That is when the problem started. As hard as I looked, it became obvious that figuring out if my son would be eligible for financial aid was impossible. All college websites posted the “cost of attendance,” a formal term that included all costs for one year, including tuition, room and board, books, and other miscellaneous expenses. Federal law required that COA be reported. That number was often big — around $60,000 even back then at many private colleges.
But each college’s admissions and financial-aid web pages also made bold claims — “our school is affordable!” Their websites would include a page, www.ourschool.edu/affordable. It would tell me that the college offered generous financial aid and that a large fraction of their students (50 percent? 80 percent?) of their students received that aid. Student testimonials noted that it would have been impossible for them to have attended the college without such generous aid. The websites included statements like “The average student receiving financial aid pays $20,000.”
If my family were eligible for financial aid, would we pay the average amount? How much would we have to pay?
It was impossible to get a personalized answer, and it occurred to me that if I could not figure this out, as an economist who had worked on higher-ed policy, then surely many other parents could not figure it out either.
For students from lower-income backgrounds, this murkiness would pose a far more significant impediment in their college search process. Why risk searching for colleges and falling in love with one or more of them without some understanding of whether it was even remotely financially feasible to attend?
New federal legislation went into effect in 2011 — on account of the 2008 Higher Education Opportunity Act — mandating the use of “net price calculators” at every institution that receives federal funding for financial aid. These tools are, generally speaking, not user friendly. They typically require complicated tax information. Most people do not like doing their taxes (and often use software like TurboTax or hire others to do it for them). Even those who do their own taxes might not fully understand foundational tax concepts. Net price calculators may ask for adjusted gross income, untaxed income, or “adjustments to income.” That terminology turns people off.
Real affordability is sacrificed in the name of perceived affordability.
Making matters worse, the net price calculators aren’t always helpful. I conducted an exercise using net price calculators at 200 randomly selected four-year residential institutions. Many institutions based their awards on data that were out of date (including two that provided no indication of the date, and another that provided five-year-old data). Some required the user to leave the website to get additional information. Two asked about the race of the applicant. Three were nonfunctional over the several-week period during which I conducted this exercise, including one very large public university with tens of thousands of students. Another flashed a warning that my connection to the website was not fully secure, a serious problem for a website that requires users to input their financial data.
The introduction of net price calculators made it possible for a family to get an estimate of what a college would cost them, but it is still a difficult and anxiety-producing task.
The Obama administration also made other efforts toward improve pricing information. The College Scorecard enables users to enter the name of a college and easily obtain extensive data regarding its characteristics. Among other information, it provides data on the average net price paid by students at that institution. This would seem to be a very good thing, but the average net price captures both how expensive a college is and the composition of its students. Two institutions that would charge an individual applicant exactly the same amount may have substantially different average net prices if the financial status of their students differed. The college with a higher-income/higher-asset student population would appear to be more expensive, and so comparing by average net price would be misleading.
The College Scorecard also provides average net price data for students with family incomes in five different income bands. These data help address the problem that differences in student composition create, but issues still remain in their interpretation. A fundamental problem is the use of averages rather than medians for a statistic that is affected by outliers. In this case, a family’s asset holdings contribute to net price, and that distribution is heavily skewed. This bias similarly affects statistics on the average net price among all students.
The ultimate problem with the College Scorecard data, though, is that families want to know what college is going to cost them, not some average that may or may not reflect their situation. Many institutions have reported average statistics on their own web pages for quite some time to no avail — students do not believe average statistics. What they want to know is how much is this college going to cost me?
How much does college really cost? That question is difficult to answer because all students pay a different price (sort of like airline seats). Most do not pay the sticker price. How much does a student whose family makes $50,000 per year pay? What about $100,000 or $200,000? Assets matter as well, so what we really want to know is how much does a student with, say, a family income of $50,000 and assets valued at $75,000 pay? How much do other families with different financial circumstances pay? I call this more specific approach the individual net price. It is what we need to know.
Of course, it is impossible to estimate and report individual net prices for all students at all institutions. That would be an insurmountable task. But we can take representative examples of students with different levels of income and assets. I embarked on such an exercise, focusing on families with income and assets at the 10th, 25th, 50th, 75th, and 90th percentiles of income and assets. I calculated these values using data from the 2016 Survey of Consumer Finances for families with children approaching college age (13 to 17) at the time the survey was conducted.
I then entered the resulting financial values into the net price calculators, adopting a standard set of assumptions about the students and their families (living on campus, married parents, no other siblings in college, etc.). I conducted this exercise at 200 randomly selected four-year residential institutions, that can be broken down as follows: public flagship/R1; other public; high-endowment private; and other private. The limitations of the net price calculators may have introduced errors in estimating net prices at each college, but taking averages across institutions within a specific category is likely to yield relatively accurate results. An important advantage of using these data is that they represent what institutions themselves communicate to prospective students regarding their individualized net price.
The results of this analysis are shown in the chart below. They reveal that the financial-aid system reduces the price of attending college below the sticker price for almost all students, and that the price reduction is substantial for those from lower-income families. For instance, at public flagships that charge a sticker price around $30,000, on average students in the bottom quartile of the income distribution would pay a net price (including loans and work-study) of less than half that amount.
Beyond that, one other pattern that emerges is that the income/assets-to-cost slope is much steeper at high-endowment, private colleges. At these institutions, a student at the 10th percentile of the income/asset distribution pays a net price of around $9,300 (again, including loans and work-study). This jumps to $58,000 at the 90th percentile. At non-flagship, non-R1 publics, the same 10th percentile family pays more ($12,000), while the 90th percentile family pays less ($20,000).
What are we to make of this data? Is the amount that a student is expected to pay an amount they can afford? The total amount they can afford includes a cash payment that the financial-aid system estimates, however imperfectly, through a statistic called the Expected Family Contribution (EFC). I simulated that amount using an algorithm I designed, based on the work that I do with MyinTuition, an online tool that dozens of institutions use to provide ballpark financial-aid estimates based on a small number of financial inputs. I also assume that students will take out a $5,500 federal student loan (the maximum allowed for a first-year student) and contribute $2,500 from a work-study job. I label the sum of my estimated EFC and assumed loan and work-study values the affordable net price.
I deem a college affordable to a specific student if the affordable net price is greater than the individual net price that the student is charged. I calculated the affordable net price using the same percentiles of income and assets as reported earlier.
What I found is that there are, indeed, trouble spots in the financial-aid landscape. For students at the 50th percentile of income and assets or above, most institutions are “affordable,” based on my definition. That does not mean that paying for college is easy — it still may be a struggle for these families — but in the end they are likely to be able to make it work.
For students at lower levels of income and assets, however, this is not true. Students at the 25th percentile of the income/asset distribution or below are charged $4,000 to $6,000 more than they can afford at public institutions. Private institutions without large endowments are even less affordable for these students. Those institutions provide financial aid that falls about $12,000 to $13,000 short of what these students would need to afford them. In either case (public or non-highly-endowed private), households with limited economic resources (incomes below $37,000 and limited assets), face an insurmountable hurdle.
Private institutions with large endowments mainly do not suffer from this problem (they charge lower-income students an amount close to what they can afford). This may not be surprising since many colleges in this category meet full need, meaning students pay what they can afford regardless of income or assets. Such large-endowment private colleges, though, represent a small minority of colleges and are typically only an option for the highest achieving students.
Higher ed’s affordability problem results from unintended economic constraints. At public institutions, the state sets both the in-state and out-of-state sticker prices. A well-intentioned desire to maintain affordability results in lower sticker prices for state residents, but these lower prices reduce the resources available to the institutions. Combined with inadequate levels of direct state funding, these institutions have insufficient funds to provide enough financial aid to lower-income students. Higher sticker prices for out-of-state and international students help, but not nearly enough. Real affordability is sacrificed in the name of perceived affordability at these institutions.
This problem also applies to private institutions without large endowments, since they face fierce competition from publics. Price competition prevents them from charging a lot more than what public institutions charge higher-income students, whose prices are capped by the state at relatively low levels. Their sticker prices may be considerably higher than those at public institutions, but merit aid is used to lower it for many of their higher-income students. This pricing model effectively restricts revenue at these less-wealthy private institutions while also making their pricing opaque. Because these institutions also have limited endowments from which to draw additional financial support, they also struggle to provide the financial aid lower-income students need.
Private institutions with large endowments have the market strength and financial resources to offer much more financial aid and more affordable pricing. They face the least market competition, as there are far fewer of these kinds of colleges. The high level of demand to enroll in these institutions, combined with the lower level of competition between them, enables them to charge higher-income students a lot more than lower-income students. Spending from the endowment further enables these institutions to subsidize the prices charged to lower-income students.
We need a way to ease the anxiety and lift the weight off the shoulders of parents and their children who are considering college. A college education is an investment that generates substantial returns. It does not need to be free. It just needs to be affordable — and not just for the rich, but for everyone.
This essay is adapted from the author’s book A Problem of Fit: How the Complexity of College Pricing Hurts Students — and Universities (University of Chicago Press).
