The Fat Middle

There are those ideas, like Darwinian evolution or the basics of information theory or game theory, that cannot be unthought. They are powerful enough to change one's Weltanschauung, and having done so mark the mind like a growth ring. I suppose that should be the aim of a liberal arts education. We could probably do a better job of selling it that way.

Lately I've been orbiting the gravitational well of an idea that, while not on the order of calculus or market theory, makes a lot of things click into place for me personally. I call it The Fat Middle.

Most things in life are ruled by combinatorics. We experience and try to make sense of many dimensions of sensory data, which if our brains didn't automatically data compress for us, would overwhelm us with complexity. The very molecules we're made of are combinations of more permanent bits--those atoms hypothesized by Democritus about 1500 years ago. Our civilization comprises a vast economy with many, many moving parts. It's easy to get lost in the permutations.

Every day we figure out how to get from where we are to where we want to be, literally or metaphorically. It's that part in the middle where the opportunity lies. Not for you and me usually, but for someone who spends all day thinking about the middle. The middlemen.

Having a way to negotiate the way between trying to sell a house and having sold it, for example, is valuable. Finding someone who knows about applicable laws and contracts, and is hooked into a network designed for moving real estate is a valuable thing.  The creation of such a Middle is like pushing a volatile molecule up the energy scale until it cracks open to spill out far more energy than you put into it.

This Middle provides a public service, whether it's selling houses, buying a good book, or finding a job. Cities sprang up along trading routes (often on rivers) because of the value of the Middle. Paved roads were investments in the Middle where the rivers couldn't go.

Phase One: Creating the Middle

Basically, if you can find a way to be of use between a supply and a demand, you create the Middle, and this is nice because it naturally wants to be standardized. Roads are narrow, standardizing routes. Realtors may compete with one another, but they all cooperate when it comes to making a sale.

Here are some Middles:

• Google: between you and everything on the Internet
• Facebook: between you and your friends
• eBay: between you and the world's flea market
• Music companies: between you and music (pre-Internet)
• Newspapers: between customers and businesses (pre-Internet)
• Higher Education: between high school grads and good jobs
• Medical Establishment: between you and medicines or medical services

The last two of these are not informational, but still play the role of the Middle by providing access for a price to exclusive services and benefits. There didn't used to be a Middle for either education or medicine. You'd simply go to a teacher and pay tuition or take a chicken to the witch doctor, respectively. But standardization and massive organization have created a Middle for accommodating a huge demand.

Phase Two: Getting Fat

Once the Middle is standardized, it becomes a gatekeeper as well as access path. Tolls get higher on the road, until the price is as high as the market will bear.

As an example, consider an entrepreneur who builds a bridge over a river, meeting a demand for transportation. He makes enough money on tolls to maintain the bridge and have enough left over for a nice income. But as the use of the bridge becomes the standardized Middle, and travelers depend on it more and more, he realizes that he controls a natural monopoly. So he increases the price until the total revenue levels out, fattening the Middle.

It's interesting to note what is allowed to get fat and what isn't. Power companies in the US are regulated by the government, presumably as a nod to the fact that providing electricity is a public good. Roads are mostly free to drive on, paid for indirectly through taxes. Access to water provided at very low cost to most of us. On the other hand, access to professional medical care is not a concern of the government in the same way. Education falls in between, given the range of opportunities between public and private schools.

Uncontrolled, the Middle will make itself fat and continue to secure its position. Media outlets are particularly well suited for the latter task, since they are between news and consumers (or were until lately).

We used to pay more than $50/month for a wired telephone line. Then along came the Internet and Skype, and hey, I get that basically for free now. I can only conclude that most of that $50 was Fat Middle for the phone company. Don't cry for the phone company, though; they can still get away with charging a dime for every text message you send--all Fat Middle for them.

Those companies in the Middle can use their profits to tilt legislation their way, like the music industry and copyright laws, or the recent assault on net neutrality. I think net neutrality cannot hang on much longer because of this pressure. Or the cozy relationship between banks and regulators that helped create the overindulgence in consumer debt.

Another technique to fatten the Middle is to over-promise the usefulness of it. Think of it as a "Gold Rush" effect. This is more suitable to some services than others. The creation and marketing of "maintenance" drugs is one example, or Monsanto's attempt to corner the seed market with one-use varieties, so you can't grow your own. Soon, Facebook will be plugged into so many facets of your personal life that you won't know how to live without it, right? A bachelor's degree is absolutely essential to getting good job. Nobody questions that. Why?

Phase Three: Revolution

Once the Fat Middle has squeezed out every bit of extra economic value, it no longer performs a public service. This is an extreme position that few Middles may ever achieve due to the particulars of the betweenness. In the bridge story, if the owner charges so much to pass over the bridge that it would be economically more viable to take the ferry, but there are new laws outlawing such "unsafe" transportation, the fattening has ripened.

Technology has a history of dining on the Fat Middle. Toll roads are mostly gone in the US. Newspapers can't make fat revenue by cornering the advertisement market, and the music producers can't isolate musicians from their audience and charge a rich premium for access. AT&T can't charge me 30 cents a minute to talk to my parents in Illinois anymore (when that bill came, I immediately switched to another phone company. I got a letter from AT&T a couple of months later saying "we're sorry! we want you back!" Fat chance.).

The technology solution is bound to rearrange the Middle, but it will again solidify. Costs and laws around the Internet will slowly evolve under the pressure of the rich rewards dangling there, until you pay per click to Facebook and Youtube, and don't even think about starting a web business without going to the bank first. I hope I'm wrong.

The Fat Middle of Higher Education

In the July 20th Chronicle article "Learning From Socrates and Adam Smith on Financing Universities"
Richard Vedder writes:

If one were to allocate faculty salaries for instruction to account for the non-instructional dimension of university service, faculty compensation for instructional services often is well under 20 percent of revenues raised by institutions, and almost never as much as 50 percent. 

 So what? The point being made here is that the ostensible principal raison d’etre of most universities—the education of our youth—is really a small part of university activities. Put differently, if the faculty salary for instruction to institutional revenue ratio were to rise to, say, 50 percent, by reducing the non-instructional dimension of university spending, the total cost of educating students would fall dramatically—to roughly the levels found in many other industrialized nations in the world. 

This is an argument that universities (public and private) are too fat.

On October 20th, Vedder wrote another article "Why Did 17 Million Students Go to College?", citing two apposite bits of information. Both indicate a "Gold Rush" effect for higher education: that the Middle between high school graduates and good jobs has been oversold by the supplier:

Over 317,000 waiters and waitresses have college degrees (over 8,000 of them have doctoral or professional degrees), along with over 80,000 bartenders, and over 18,000parking lot attendants. All told, some 17,000,000 Americans with college degrees are doing jobs that the BLS says require less than the skill levels associated with a bachelor’s degree.

He goes on to cite a paper about the marginal return on higher education investment:

This week an extraordinarily interesting new study was posted on the Web site of America’s most prestigious economic-research organization, the National Bureau of Economic Research. Three highly regarded economists (one of whom has won the Nobel Prize in Economic Science) have produced “Estimating Marginal Returns to Education,” Working Paper 16474 of the NBER. After very sophisticated and elaborate analysis, the authors conclude “In general, marginal and average returns to college are not the same.” (p. 28)

Unfortunately, the paper itself is behind a paywall (the irony...). A low marginal return means that more investment is hard to justify: we're at the point of "diminished returns."


In a nutshell, the argument is that higher education has become a Fat Middle, and is ripe for revolution. The form of that revolution isn't hard to fathom: low-cost, high quality, online programs. The way Vedder puts it is:

Higher education is on the brink of big change, like it or not.

Expect to see more analyses like the one recently from The Wall Street Journal, where "Putting a Price on Professors" by Stephanie Simon and Stephanie Banchero refers to:

A 265-page spreadsheet, released last month by the chancellor of the Texas A&M University system, amounted to a profit-and-loss statement for each faculty member, weighing annual salary against students taught, tuition generated, and research grants obtained.

This creates a certain kind of business-speak dialog:

"Every conversation we have with these institutions now revolves around productivity," says Jason Bearce, associate commissioner for higher education in Indiana. He tells administrators it's not enough to find efficiencies in their operations; they must seek "academic efficiency" as well, graduating more students more quickly and with more demonstrable skills. The National Governors Association echoes that mantra; it just formed a commission focused on improving productivity in higher education.

A part of this is just the current demonizing of education that seems to be in vogue, and the narrow viewpoint that colleges are like factories with a uniform input that should be able to "six-sigma" end product. But there are valid points to be made about the cost versus return of post-secondary degrees too, particularly focusing on those costs that have little to do with the instructional mission.

It's ironic that the business viewpoint would be used to make this particular criticism of higher education, which (if they are correct) is only doing the same thing that, say, the drug companies do. But for some reason, education is seen as a public good and pharmaceuticals or access to medical care are not.

Opportunity

It's not all broom and doom. You can find a Middle of your own and enlarge your fortune. Here's one idea. One of the most frustrating problems in dealing with corporations is getting problems solved through their customer service departments. Some are good (like our local Time-Warner office), and some aren't. Standardization and transparency would work wonders in this area. So all you have to do is create a "complaint engine" that becomes the standard interface between companies and their customers for resolving disputes, with public ratings showing response time, resolution rate, and individual comments.  At first, companies will hate it. Then the more progressive ones will see the advantages and start asking for plugins so they can feed directly into their PeopleSoft (or whatever) systems. Then you can sit back and watch the Middle fatten.

The Internet (while it lasts) creates enormous potential for Middle solutions. Even within a university, there are opportunities, like creating an institutional document repository.

Pricing Higher Ed

My last post included a link to "Admission, Tuition, and Financial Aid Policies in the Market
for Higher Education" by Epple, Romano, and Sieg from 2003. In the paper, they test economic models against actual data and reach some very interesting conclusions about how pricing works. One of the assumptions is "In our model, colleges seek to maximize the quality of the educational
experience provided to their students."

I thought about this for a while. It's not obviously true, is it? I'm trying to remember how many meetings I've sat in where someone talked about the quality of educational experience. Of course, in many small ways programs, individual instructors, chairs, and so on do bits and pieces that impact this quality. And the SACS Quality Enhancement Plan is supposed to turn this into a visible project.

But by and large, I think most of my meeting time has been spent on solving problems, grinding away at the routine bureaucracy, or (once in a while) trying to make the bureaucracy work better. Of course, outcomes assessment is supposed to lead to continual improvements in the quality of education, but it would be a wonderful thing if board meetings were opened with the sentiment: we're here to improve the quality of educational experience.

As it turns out, I'm in the middle of a project to improve the "experience" part of that by helping organize strategic planning action items along those lines, and I'm going to start using that language.

In the article, the authors give some dependencies for quality:

1. peer ability of the student body
2. a measure of peer-student income diversity
3. instructional expenditures per student
   

Quality is relative, and two of the dependencies listed above are intuitive: students don't want to attend classes populated with students who are all less able than themselves. They also perceive the institution's ability to spend money in the classroom. This one is reflected in college rankings too (see "Zza's Best Liberal Arts Schools"), which probably has some affect on decisions. The second dependency, however, is surprising to me.

They see a distinct stratification that bestows economic benefits to the top schools:

Colleges at low and medium quality level have close substitutes in equilibrium and thus a limited amount of market power. Admission policies are largely driven by the “effective marginal costs” of educating students of differing abilities and incomes.

Colleges with high quality have more market power. These colleges do not face competition from higher-quality colleges. Hence, they can set tuitions above effective marginal costs and generate additional revenues that are used to enhance quality.

This suggests a Darwinian struggle for schools at the low and mid-levels of means and quality. In a catch-22, they lack the pricing power to enhance their position much. But once breaking through a ceiling, it becomes easier. At least that's my interpretation.

On the subject of price, the authors illuminate the second dependency (financial diversity):

We also find that colleges at all levels link tuition to student (household) income. Some of this pricing derives from the market power of each college. This allows colleges to extract additional revenues from students that are inframarginal consumers of a college. However, as noted above, our empirical findings suggest that market power of lower and middle ranked colleges is limited. This suggests that pricing by income may be driven by other causes.

I found an explanation of what an "inframarginal consumer" in another source "The inframarginal consumer is willing to pay more for the good than is the marginal consumer." So, if your college has a good market position, you can charge a premium. But the authors argue that that this isn't the whole story:

In this paper, we then also explore the role that income diversity measures play in determining college quality. Our findings here indicate that colleges and students believe that the quality of a student’s educational experience is enhanced by interacting with peers from diverse socioeconomic backgrounds.

Obviously there are many reasons for wanting a diverse student body, but the authors propose to actually use that as a factor that contributes to the price model. This begins to make more sense in Section 6 of the paper, where they verify empirically that college quality increases with income diversity, stating that "To attract students from lower-income backgrounds, colleges give financial aid that is inversely related to income as detailed below." While this is no doubt true for some institutions, others have a more directly self-interested reason for giving need-based aid: to increase enrollment in those students who couldn't otherwise afford to attend. I talked about the revenue-generating effect of this "gap filling" in "The Power of Discriminant Pricing."

Also in Section 6, they make an observation about college size:

Absent scale economies, peer effects and endowments create a force for colleges to reduce size to increase student quality–in the limit maximizing quality by admitting a handful of brilliant students and lavishing the entire endowment on educating those students. The countervailing effect of scale economies is captured in our cost function primarily by the c3 term in the cost function.

This outlines a good strategy for an elite school: keep it small because it's easier to maintain a high level of average student quality, but not so small that the economies of scale drive up costs unreasonably.

A hundred points of SAT is worth between $4688 and $10363 in merit aid (in 2003), according to the model output. The difference depends on what tier of college the applicant applies to.

Conclusions: First, remember I'm not an economist. But the paper is clearly written, and you can skip the mathy bits easily enough. The model presented has errors, as the authors describe, but the approach seems to lead to some insights, like the relationship between size and quality, the effect of financial diversity on institutional quality, and price sensitivity by student ability and income. I have not delved into all of these in my notes above. I don't know how hard it would be to simulate their model numerically to actually use it to build your policies (e.g. by running scenarios), but it's probably worth showing it to your IR office. And if you have an economics department handy, maybe they can shed some light as well.