We reviewed the higher education bubble in two previous posts (see here and here). Today’s post is looking at the fuzzy thinking driving the bubble, as well as the larger issue of which the bubble is merely a symptom. But first things first – I need to congratulate my Duke Blue Devils for assuming their rightful place atop the college basketball rankings! And a preemptive apology to my thesis adviser at Northwestern, who may not care much for this post.
At its core, the bubble in higher education is being driven by the belief that a college education teaches skills that are vital to building a successful career, and that higher levels of education are tied to increased worker productivity. In other words, skills learned in school lead to higher productivity, which is rewarded by employers in the form of higher earnings for the well-schooled. Here’s the assumed relationship:
In Economics there’s the concept of asymetric information, where one party to a transaction has better information than the other party. Virtually all job markets have imperfect information. To cope with this asymetry, buyers look for shortcuts – ways to infer quality – using a process called signaling. In the case of the job market, employers look at both the level (how much) and quality (where) of education to signal the quality of the applicant.
In my professional career, I’ve seen consistently that higher levels of educational attainment lead to higher levels of income. This is especially true with workers under 40, who matured professionally as the higher education mania spread in earnest. We see this relationship in the chart below, based on 2010 Census data:
There’s a clear correlation between levels of education attainment and higher incomes. The more interesting questions involve causality – does higher education cause higher income, and if so, how, why, and should it?
An Important Distinction: Correlation vs. Causation
One of the more prevalent learning disabilities in organizations today is confusing correlation with causation, the notion that umbrellas cause rain or diet soda causes weight gain or even that a rock keeps away tigers. In a professional setting, I’ve seen this confusion lead companies to mistakenly justify expensive marketing campaigns, brand overhauls, even major changes to their strategic positioning. Let me share just one example.
Several years ago we were asked to optimize the Marketing performance of a large multi-channel business. The most successful marketing channel was catalog, with more than 100 million copies being sent to millions of customers every year. Like many direct-response companies, the team tracked if someone received a catalog and then placed an order online or in a store. Based on that key assumption, more and more catalogs seemed the way to go. But, we asked, did the catalog really drive the sale, or was it the umbrella that happened to be around when it was most likely to rain?
There happens to be a very easy way to figure this out (though surprisingly few companies do). We split the customer file into 2 groups. One group received the normal series of catalogs; the second group received no catalogs. Here’s what we saw in terms of sales and profit between the two groups (numbers disguised but directional).
Turns out nearly every customer who received the catalogs and ordered would have ordered without one. The company had mistakenly focused on one variable (received catalog) and seen a correlation with the desired action (bought something). But when we controlled for the impact of the catalog, we saw it was largely unnecessary. Could the same thing be happening with higher education and the job market?
The Assumptions Linking Education & Income
We saw that the relationship between education and income is tied through the assumed impacts of greater skill development and then higher productivity. Turns out, researchers have begun questioning these assumptions. One study showed that a majority of students gain few to no job-related skills while in college. A second study (one of many) showed that there was a negligible impact on worker productivity from higher levels of education. Finally, a Princeton study showed that when you take a student’s inherent skills into account, universities have minimal impact on future earnings. The first 2 studies challenge the assumption that higher educational attainment causes skills and productivity, respectively. The last study severs altogether the assumed causal link between educational attainment and higher income.
The chart above makes this point – there are causal factors related to the “level-of-education decision” which are also related to someone’s level of “skill.” For example, ambition is a causal factor that may lead both to greater consumption of education and to a job-related skill such as hard work. This is where the signalling function, unable to make such fine distinctions within causal factors, mistakenly attributes a skill to more education.
This isn’t to say that higher educational attainment doesn’t cause higher incomes; it does. But the assumptions underlying that relationship are wrong. We’ve assumed that something we can easily measure – educational attainment – is a handy proxy for the thing we really want to measure, productivity. But shortcuts can sometimes lead you astray. If we found ways to identify and measure actual causes of productivity, we might, among other things, stop wasting money on unnecessary education.
Why the Higher Education Bubble Isn’t THE Problem
At first glance, what’s driving the bubble is demand from students. However, the real driver is employers adding, “Degree required, advanced degree preferred,” willy nilly to every job description. If employers modified or severed the relationship between hiring and college education by better understanding the impact of a degree on what they care about – productivity – we could arrest the dangerous spirals of higher costs and poorer returns within the higher education market.
However, the bubble in higher education is caused by a bigger problem; a misunderstanding of what drives productivity and innovation. For example, an increasing amount of research (see here and here) shows the outsourcing of manufacturing has a direct impact on our capacity to innovate. This shift from “sweating” to “thinking” sounds compelling individually, but taken collectively is perhaps the biggest factor driving the oversupply of higher education. As I mentioned in my recent posts on strategy (here and here), high-level strategy (thinking) is founded on a lot of messy digging (sweating). If you don’t get your hands dirty (literally and figuratively), you won’t capture the insights that lead to long-term innovation and productivity.
This is an issue increasingly being studied and given attention. With the direct and opportunity costs we face, the sooner we confront and address these issues, the more confident we can be that our education dollars are being wisely invested, that our kids and grandkids aren’t being saddled with debt they can’t hope to repay, and that the asset that is our workforce – regardless of collar – is properly utilized.
Today’s post on the higher education bubble is the first of three updates of last year’s post. This post will incorporate additional research and insights to help explain how we got where we are today. Subsequent posts will examine ways to deflate the bubble and long-run implications for the higher education industry.
When we look at higher education today, there are 4 key players: Schools, Government, Parents, and Employers. Let’s quickly review how each contributes to the education bubble.
To understand how Schools impact the bubble, we have Bowen’s Rule:
- The dominant goals of institutions are educational excellence, prestige, and influence.
- There is no limit to the amount of money an institution would spend on these goals.
- Each institution raises all the money it can, and spends all it raises.
- The cumulative effect of the preceding is toward ever increasing expenditure.
This rule is not contested by any research, shows how non-profit schools simply substitute wasted spend for profits, and explains why we might not expect schools to contribute to a solution. Former Harvard president Derek Bok states, “Universities share one characteristic with compulsive gamblers: there is never enough money to satisfy their desires.” Per Robert Martin: “Higher education finance is a black hole that cannot be filled.” Or Ronald Ehrenberg: “Administrators are like cookie monsters…They seek out all the resources that they can get their hands on and then devour them.”
To understand the role of Government funding on the bubble, we have Bennett’s Hypothesis 2.0:
- Individually, each college is trying to improve.
- More revenue is very useful in the quest for improvement.
- An increase in financial aid gives colleges the option to raise revenue by raising tuition.
- Most colleges succumb to this temptation; therefore, higher financial aid = higher tuition.
Economics professor Richard Vedder looks at the rapid rise in tuition the past 30 years. “What happened? The federal government has started dropping money out of airplanes.” Economist Bryan Caplan agrees. “It’s a giant waste of resources that will continue as long as the subsidies continue.” Argues Jonathan Robe, “Cheap student loans (backed by the government) make students far less price conscious, and enable schools to raise tuition because they know if the student can’t pay, the buck gets passed on to the taxpayers.”
Parents send their kids to college so they will learn valuable skills while increasing their lifetime earnings potential. Regarding skill development, in Academically Adrift Richard Arum finds that 1/3 of students gain no measurable skills in college, and for the rest, the gains are minimal. Regarding future earnings, data does conclusively show that earnings increase as the level of education increases. Of course, correlation isn’t causation. To this point, Princeton economist Alan Kreuger finds that when controlling for student’s talent, top-tier colleges add no measurable value to lifetime earnings.
It is unlikely that parents would voluntarily reduce the amount of education they demand for their children, given incentives today. Peter Thiel argues that parents are terrified about what will happen to their children if they don’t go to college. Parents are “Paying for college because it’s an insurance policy against falling out of the middle class.”
Finally, Employers, who have also benefited from the bubble. Increasingly business looks at college degrees as a minimum job requirement, not because of any benefit from the degree (or requirements of the job), but simply because they need an easy way to screen applicants (Doesn’t have a degree? Must be lazy or dumb). Argues professor Richard Vedder, “Employers seeing a surplus of graduates are just tacking on that (degree) requirement.” The result is credential creep into many jobs which manifestly do not require a college degree. Employers have little motivation to change the way they operate – they’re hiring overqualified people at bargain rates.
So where do we stand today? Outstanding student debt recently passed $1 trillion, based on costs increasing 900% over the past 30 years (12x more than the price increases that led to the housing bubble). Meanwhile, wages of college-educated workers are unchanged and a recent study found that 53% of bachelor’s degree holders under 25 are under- or unemployed. This combination has led to student default rates of up to 15%. With that number increasing every year, it seems likely that some sort of correction is inevitable. The scale and scope of this correction depends on where things go from here. That’s a topic for the next post.
My exposure to the higher education market was positive – thanks to incredibly generous financial aid offices, I received my BA at Northwestern owing just $5k, and graduated debt free with my MBA from Duke. However, as I’ve spent more time in the labor market, and as the cost of higher education has defied gravity, I’ve become increasingly doubtful about the economic value of a college degree. My goal with today’s post is to shed light on what I see as a flawed assumption: That a college degree is invariably a wise investment.
There are strong similarities between the housing bubble and what we see today in the higher education market. For starters, both saw a rapid rise in prices. Between 1997 and 2006, home prices more than tripled. In higher education, since 1988, tuition and fees have increased more than 300%. Both run-ups were fueled by cheap loans made possible by an artificially low interest rate environment, lax or non-existent lending standards, Principal-Agent problems (originators of the loans bear little risk from non-payment and hence have low incentives to make good loans and high incentives for making more loans), and the mistaken belief that value (prices of homes, return on education) will always increase. And lastly, in both markets the government decreed that more was always better, and created and adjusted policies to support this aim.
In the housing market, the last few years were fueled in part by the sub-prime market – having run out of even marginally qualified loan applicants, originators went to the bottom of the barrel – and when these loans defaulted, the market collapsed (The economy collapsed because of massively leveraged bankers gamblers). Does the recent growth of for-profit colleges, largely unqualified to educate but very adept at siphoning taxpayer-financed loans, signify that the irrational exuberance in the education market approaches a reckoning?
There’s a vicious cycle that appears to be driving the runaway costs. Starting with a policy of increasing college attendance, the government makes subsidized loans available, increasing the amount that students can pay for school. Recognizing this, schools are able to increase tuition and fees, which over time prices many students out of the education market. The government and private lenders respond by increasing financial aid available and the escalation continues. To break this cycle, just one assumption needs to be proven false. The one I’m interested in is the assumption by students and parents that taking on ever higher levels of debt to achieve a degree is a profit-maximizing decision.
So let’s see why everyone is convinced we are better off with more college graduates. First let’s look at trends in costs over the past 30 years, compared to the ability to pay.
It’s clear from this chart that costs have dramatically outpaced the median family income and overall rate of inflation. But costs, of course, are only one part of the equation. How do salaries compare across the different levels of degrees?
Comparing after-tax earnings, we see that there’s a significant rise in earnings as a student goes progressively higher into the education field. This seems to make sense – a degree signifies competence, which is then rewarded in the labor market with a higher income.
Now the last thing to understand is the actual return to the degree. During the time a student is working towards a degree, not only are many racking up debt that may take years to repay, but they’re sacrificing money they could have made if they had immediately entered the workforce instead of attending school. The chart below incorporates both higher earnings as well as direct and opportunity costs associated with education. It is created from public sources of data as well as from my own calculations and assumptions.
Along the bottom is the highest degree attained, ranging from a high school diploma to a professional degree (MBA, JD, etc.) The columns signify how much more money the person earns compared to holders of the degree immediately to the left. So for instance, a high school diploma earns you $300k more than not getting a diploma, while attending some college earns you $100k more than someone just with a high school diploma. Finally, the number on top of each column is the age at which that degree holder has accumulated more earnings than the degree to the left (except for professional degrees; these are compared to a bachelor’s degree). For instance, after earning a bachelor’s degree, it won’t be until you’re 33 that your cumulative earnings exceed the cumulative earnings of someone earning a degree at a community college. This is due to direct costs (i.e., debt) and opportunity costs (out of the workforce for several more years).
Currently, about 35% of adults have some sort of college degree. President Obama has indicated that he’d like to see this number hit 60%. The Gates Foundation has an even loftier target: 80%. Based on the previous chart, it’s understandable that people think more is better. According to this logic, we could increase worker pay across the entire economy if those with only high school diplomas all went back to school to get their undergrad degrees. Assuming that these workers returned to the workforce and spent their higher incomes, GDP would more than double. Magic!! Of course, this fantastical result shows the faulty logic underlying the more-education-is-always-better argument.
I suggest that what we need is more jobs requiring a college education, not more graduates. There’s no hidden reservoir of millions of high-paying, college-level jobs just waiting for qualified applicants to fill them (nor were there prior to the recession). In countries like Korea, where 80% of high school graduates enter college, fewer than half can find full-time jobs, and fewer than one third believe their jobs require a college degree. How can anyone believe this is really the model we want to follow? The promise of a better life that leads many people to pursue higher education is betrayed when a significant share of degree holders work in jobs for which a degree was unnecessary. Just because the job description says ‘degree required’ doesn’t make it so. Public policy appears focused on creating more college graduates without a clear understanding of what those folks will do and where they will work.
One of the problems with taking the charts above at face value is that they confuse average with marginal. Generally speaking, efficient investment continues until the marginal benefit equals the marginal cost. In this case, think of marginal benefit as lifetime earnings and marginal cost the expense students/society incurs to realize that benefit. Unfortunately, like in many markets, as more of something happens, the marginal benefit decreases. So when we see the historical average returns of a college degree, we are wrong to assume that those returns will remain positive today and in the future.The reality is that as we add more and more ‘sub-prime’ students and provide them ‘sub-prime’ educations at a premium cost, we can expect to see less and less return. Another way to think of this: Suppose a company offers the first 10 shoppers $200, then charges the next 10 shoppers $100. Stating that the average shopper receives $50 free, while technically true, disguises the fact that half the population would have been better off not shopping at all.
Our failure to distinguish differences on the margin is resulting in a massive distortion in the market for credentials. At the same time, the cost associated with supplying the marginal student an education is higher than average (generally because they require more loans). As we expand the proportion of students attending college, the marginal benefit diminishes rapidly while the marginal cost increases. This is a poor recipe for positive return, and I’ve graphed this below.
But wait, there’s more! Because as we add more degree holders to the job market, unless we’ve added the requisite number of jobs needing a degree, we’ve watered down the value of everyone’s degree. This results in a downward shift of the marginal benefit curve. Concurrently, because of ever spiraling costs, the associated debt burden, and resultant increase in delinquencies and defaults, the more students we add to the higher education market, the more ‘education’ costs for everyone. This results in an upward shift of the marginal cost curve. Taken together, this shifting massively expands the loss associated with the over-consumption of education.
Most folks would agree that the higher education market is, at best, inefficient. Others would argue that it’s broken. So whatever view you take, there’s room for improvement. There are 4 main actors on the stage, and I’d like to focus on each of them: Schools, Government, Business, and Students.
The schools clearly play a major role in creating and sustaining the problem. However, for several reasons it seems unlikely that they will contribute directly to a solution. For one thing, outside of government bureaucracy, there may be no more opposed-to-change culture than academics. Second, and more important, asking schools to focus on efficiency and profit/loss ignores the basic business model they use. That many schools turn a profit doesn’t mean that that is what they look to optimize. Instead, their focus is on higher rankings, exclusivity, etc. – what they’d refer to as their brand, and a prestigious one at that.
I do think that schools have a point – the power of the brand is important. I’ve done my share of job interviews and by and large what gets me in the door is the power of my personal brands: Northwestern, Duke, Apple, McKinsey. During the interviews, it’s rare that I actually discuss what I learned – it’s like the brands themselves define my worth, independent of what the brands added to my worth. From this perspective, schools are being completely rationale when they focus on ways to increase their brand equity: Always higher prices, greater selectivity, renowned professors who rarely teach, and world-class facilities
Like the housing market mess, well-intentioned government policies sometimes create and often exacerbate problems in the marketplace. As we’ve already seen, the government does this in the market for higher education by advocating a college degree at all cost, by distorting demand via artificially low interest rates, and by increasing prices through loan guarantees and raising borrowing limits in lockstep with tuition increases.
Here’s what I think should change. First, work with businesses to create a post-college accreditation test. If all college graduates are evaluated objectively, business will possess a powerful gauge of how potential candidates compare. This would also be an effective way to assess how well academic institutions are preparing their students, and likely lead to academic changes which better focused on results. High-caliber graduates from less prestigious schools would be less likely to be penalized for having the wrong brand name on their resume, which would serve to also reduce the demand for the most selective (and expensive) colleges.
In response to the fraud being perpetrated by some for-profit colleges like the University of Phoenix (graduation rate of 4%!), where two of three borrowers is delinquent or has defaulted, the government should force those colleges to set aside a significant portion of the $28 billion in federal aid they receive every year, to be used to pay back the government for the inevitable wave of delinquencies and defaults. These colleges are essentially playing in the sub-prime education market, willing to fleece anyone who can sign an ‘X’ on the application for federal financial aid. The rare graduate will have $20k+ of debt and a degree that means next to nothing.
Lastly, accept that the education market is not a free market – a higher price will not lead to competition which ultimately reduces prices while increasing supply. You add fuel to the fire by increasing loan amounts in an inelastic marketplace. Indeed, recent research has found little evidence that Federal grants and loans do much to increase college enrollment whatsoever; they just raise prices.
The first thing business needs to do is stop the credential inflation. Over the past 30 years, jobs which used to require a high school diploma now typically require a bachelor’s or even a master’s degree, despite the jobs themselves having undergone little change.
Understanding the skills required for a role and matching that to candidates’ backgrounds and abilities is preferable to assuming a college degree is highly correlated with job performance. To the extent it is, that’s more likely due to the candidate’s inherent qualities than what they learned in school. We see students who tip a toe in the labor market, find a lack of demand for their current level of education, and head right back to school to gain an additional degree (and presumably, a competitive edge). Riding out the storm by adding another degree (and additional debt) makes sense if there’s a perceived value to the degree. Business can help students make that decision by properly assessing and communicating the true value of education attainment.
Reducing the unnecessary focus on degrees will decrease the incentive for students to pursue degrees that will have little impact on their careers. Either students will self-select into majors or degrees which correspond with jobs truly needing a degree, or they’ll simply demand less education.
I remain conflicted on the advice I’d provide students/parents: Do you make an investment which appears rationale, but only because of market irrationalities? The folks who flipped houses in the 2000 – 2006 timeframe would have a much different answer than folks who got into the housing market at its peak.
Princeton economist Alan Krueger has recently published a working paper (#563) that estimates the return of attending an elite college. He reports a striking finding: When student ability is controlled for in the regression, ‘the returns to college selectivity fall substantially and are generally indistinguishable from zero.” Put another way, very talented students go to very selective colleges, and the future returns that those students experience are due to their talent, not the college.
If these findings can be reproduced with further research, it’s difficult to overstate their implications. Much is made of the breadth of academics, the outstanding faculty, and the high-value contacts that students can find at elite institutions. What if for many students, none of that matters? As a profit-maximizing parent, if your student could have equivalent returns on your $200k investment at Princeton or your $40k investment at your state’s best public university, might that alter your investment decision? If you’re a business, might you be willing to hire a local state school graduate over a Princeton graduate, if you knew that the state school graduate had been admitted to Princeton?
Based on the data and most recent research, I think students and parents should consider the following when making decisions on higher education. First, don’t attend college if you’re not sure it’s right for you. As we’ve seen, dropping out of college means that the student’s cumulative returns don’t exceed a high school graduate’s until they’re almost 50. Wait a semester, or a year, however long it takes to decide that college is the right choice. Second, under no circumstances should a freshly-minted high school graduate attend a for-profit university. Would you trust an institution with a 4% graduation rate and booming growth in profits to provide an effective education for your student? Third, as a parent, research the majors and degrees where demand is currently or projected to exceed supply, and funnel your students into those areas when appropriately matched with the students interests and abilities. Understand that a mediocre engineer or computer scientist will have superior employment returns to a world-class history major. And lastly, understand that while the higher education market is in some ways irrational, until that becomes understood, you need to play by the current rules. If you can afford to send your student to a selective university, strongly consider doing just that. At least for now, the power of that brand is more important than your student’s abilities.