The Higher Education Cost Spiral: January 21, 2018 Snippets
As always, thanks for reading. Want Snippets delivered to your inbox, a whole day earlier? Subscribe here.
This week’s themes: the one-way ratchet of escalating costs in higher education, and how Cryptomove hopes to solve one of the hardest problems of the 21st century
Welcome to back to our Snippets series on Cost Disease, where we try to figure out the logic behind certain industries — medicine, urban infrastructure, and higher education, for example — getting more expensive every year. Last week, we looked at a phenomenon called Baumol’s Cost Disease, which proposes a mechanism where productivity gains in manufacturing and tech-heavy industries have a secondary consequence of cost increases in service-heavy industries. This week, we’re going to ask the obligatory follow-up question: is Baumol’s Cost Disease really what’s going on here? Or is there another culprit we need to consider?
Let’s start by looking at the data for one of the worst-offending industries, higher education. With already-high tuition prices for private higher education rising 5% each year, the obvious place to look for evidence of Baumol’s Cost Disease is in faculty salaries. If Baumol’s were indeed the culprit, we would expect faculty wages — particularly in areas like medicine or engineering, with plenty of high-paying employment opportunities outside of academia — to rise accordingly. But that’s not what’s happening. For the most part, universities have done the opposite in recent years: they’ve been increasingly filling teaching roles with adjunct and part-time professors, and salaries for full-time faculty members have stayed pretty flat. But tuition keeps going up, and colleges for the most part are non-profit organizations, so the money must be going somewhere.
The data tells us that higher education is costing more each year not because employees are getting atypical pay raises, but rather because more and more employees are getting hired each year. So, although maybe there’s a little bit of Baumol’s reflected in nicer perks and benefits for academic faculty, it clearly doesn’t tell most of the story. There must be some reason why higher education has been such a significant job creator over the last few decades: what could be driving this trend of more employees on college campuses, particularly in service and administrative roles?
The author and economist Robert Martin, who has written extensively about costs in higher education, concludes that the dominant force driving cost increases is not Baumol’s Cost Disease but rather something else called the Bowen Effect. Originally proposed by H.R. Bowen in 1980, it describes what Martin calls the “Revenue-to-Cost Spiral”: a phenomenon whereby the obligation to provide as much service as possible with the resources available (as will be the case for any non-profit organization), which in theory should incentivize efficient use of resources, in practice leads to runaway cost increases.
The revenue-to-cost spiral in higher education | Robert E Martin
Measuring Baumol and Bowen effects in public research universities | Robert E Martin & R Carter Hill
As Martin explains, spiralling cost increases in higher education don’t come from incompetent management, but rather from the fact that university leadership has an obligation to provide as much service as possible with the resources they have, thereby ensuring that surplus revenues are continually funnelled into new initiatives as opposed to covering future cost escalations in existing programs. He writes, “Administrators have an incentive to spend every dollar that is available. They cannot spend more without appearing incompetent, but spending less would forgo the opportunity to make some members of their constituencies happier. Whatever is available is spent. … Revenues are the lid on expenditures during each period. Since existing resources are frozen in place, the resources required for the new initiatives that arise each year can only come from new revenues. Some of these new initiatives will be driven by competition; most will be the pet projects of faculty, administrators, or board members. In business, new initiatives are chosen on the basis of their expected return. In higher education, priorities are determined by internal politics. Because all the revenues are spent each year, costs typically increase as a result of these initiatives. If money is given to faculty to start a new program, for instance, that additional money will be expected in the next academic year as well.”
The result is a one-way ratchet of cost increases: so long as money is available, it will always be spent; so long as it’s spent, total costs will go up. And the societal cost is immense. Student debt is turning into a full-blown crisis in the United States, with total outstanding debt having rocketed past $1 trillion and half of borrowers not able to begin paying off principal until they’re 35. Why does it feel as though there are no levers we can pull to stop this, or at least slow it down?
As the buyers of a college degree, we’re paying for a lot of things beyond education: for the social status signal; for four years of community building and networking; for societally-endorsed time and space to figure out who you are and what you want to become. This matters because when we shop around for colleges and ultimately decide for which school we’re willing to sign away decades of student loan repayments, we’re effectively pondering: “I don’t know what will happen in the next fifty years of my life, but this college education is supposed to make it better across the board.” And indeed, it does. The earnings gap between college graduates and those without a degree has never been higher, and the gap is accelerating. So when we pay for a college degree, the leverage we get from that degree is so high that once we’ve committed to buying one, it makes sense to try and get as much out of that degree as possible: consume as much as you can, while you’re there. Universities, with their mandate to provide as much service as they can for each revenue dollar they have (and to increase their reputation and ranking by doing so), are more than happy to oblige. And that’s how we end up with hundreds of extracurricular clubs and fully-staffed student service buildings and fancy athletic facilities that no one actually needs, but we end up with anyway. It’s also why higher education job growth continues at the pace that it does, even though wage growth hasn’t kept up: all these services are really labor intensive.
Overall we get a cycle that looks like this:
As we’ll see next week, higher education isn’t the only industry where this cycle happens. When a service provider has a mandate to provide as much service as possible, and the consumer has an attitude of “I’ve committed; this will pay off for a long time in the future; I should try to get as much out of this as possible”, the inevitable outcome is rising costs every year. Combine that cycle with a seemingly unlimited piggy bank of funds to draw from, and it’s no surprise why a college degree, and also a mile of subway track or a night in the emergency room, get more expensive every year.
The Awl shut down this week, and it prompted a lot of reminiscing:
One of the great treasure troves of posts on The Awl that’s really worth rereading was John Hermann’s The Content Wars, which has some essential reading on social media and Twitter in particular:
Access Denied | John Hermann, The Awl
Mutually Assured Content | John Hermann, The Awl
The Genuine Article | John Hermann, The Awl
Tomorrow’s Internet Turns 20 | John Hermann, The Awl
The especially creative:
Michael B Jordan takes on his first blockbuster role — as a villain | Jason Gay, WSJ
Upright Citizens Brigade and the case for paying improvisors | Seth Simons, Paste Magazine
Notable buildings:
What’s at stake with Amazon’s new HQ? Ask Newark | Issie Lapowski, Wired
How the Big Kids spend their money:
The Apple Cash FAQ | Horace Dediu, Asymco
Expanding our global infrastructure with new regions and subsea cables | Ben Treynor Sloss, Google
Amazon, Netflix should consider buying movie theatres | Matt Pressberg, The Information
Other reading from around the Internet:
Maybe the real superintelligent AI is extremely smart computers | Scott Alexander, Slate Star Codex
How to Design Social Systems (Without Causing Depression and War) | Joe Edelman
The Tether Conundrum: a quick backstory | Tony Arcieri
What other industries teach us about investing | Morgan Housel
US doctors plan to treat cancer patients using CRISPR | Emily Mullin, MIT Technology Review
As we previewed last week, we have another exciting Series A announcement to share with you: Cryptomove is joining the Social Capital family.
Let’s start with the fundamental problem: attack-and-defence asymmetry. There are some kinds of conflict where the defender has a huge advantage over an attacker: think an archer atop a castle wall in the middle ages, or a machine gunner in a trench in World War I. In these cases, the defender has an asymmetric advantage: it might take five, ten or fifty attackers to successfully overtake a single defender, and in fact may not even be worth trying at all. But today’s world does not work that way. Modern conflict in the 21st century looks much more like terrorism: instances of a single attacker, armed with some massively powerful weapon, able to cause enormous damage to an orders-of-magnitude bigger opponent. One need only remember the horrible, chilling statement from the IRA after their failed assassination of Margaret Thatcher 34 years ago: “Today, we were unlucky, but remember, we only have to be lucky once — you will have to be lucky always.” This is the fundamental problem of attack-and-defence asymmetry, and it should scare us.
Fast forward to the 21st century, and we can appreciate how this particular kind of attack-and-defence asymmetry has become one of the biggest challenges facing the next decade of software and the Internet. Our computing and web infrastructure has gotten so complex that malicious attackers who want to break into a protected environment can effectively always find a way in, if you give them enough time. The more complex the system, and the more time you give an attacker to study a system, the more inevitable it is that they’ll successfully break in and obtain their target. The security community has tried many defensive tactics: perimeter defences, proactive detection, deception tactics, and all kinds of other methods all help, but they don’t do anything about the fundamental asymmetry. This is a huge challenge: until we solve it for good, no data can ever really be considered safe.
Cryptomove, led by the father-and-son team of Mike and Boris Burshteyn, reverses this fundamental asymmetry with an approach that’s both ingeniously simple and also devilishly complicated. The idea is this: if you’re trying to protect the crown jewels from being stolen, don’t just leave them sitting there. Keep them constantly moving, constantly being scrambled and re-scrambled, moved around continuously, in a way where they are never stable or intact at rest. Why is this so important? It’s more than just an additional obstacle for attackers: it flips the attackers’ advantage into an advantage for the defender. The longer the time period and the more complex the attack surface, the harder it will be for attackers to be able to figure out how to approach the target. What used to be an “anti-network effect” for security is now flipped around: a moving defence system whose difficulty of attack compounds as you add more data is a true paradigm shift in security that’s desperately needed.
Cryptomove has a fascinating origin story: it started when Boris Burshteyn, the CEO (and father) of the founding team was working on a new kind of distributed programming model, both at the language and the compiler level. It was one of the rare and wonderful instances of a technology in search of a problem that found exactly the right problem: the incredibly tricky challenge of continuously moving and mutating distributed and encrypted data in a way that keeps the target in a state of very high entropy all the time. Cryptomove was born, and since then they’ve developed a very impressive roster of paying customers, ranging from the Fortune 20 to the Department of Homeland Security. As of 2017, they’re now set to work seamlessly with AWS, Docker, Azure, Box, and other essential enterprise platforms, and with their Series A behind them we can’t wait to see what’s next. We’re thrilled to be working with them, and if the future is anything like we believe it will be, Cryptomove has the potential to be one of the most important companies of the next many years.
Have a great week,
Alex & the team from Social Capital
