The Perfect and the Free

About the Author

Jacob Vigdor is Professor of Public Policy and Economics at the Sanford School of Public Policy at Duke University, and a Faculty Research Fellow at the National Bureau of Economic Research.

Curriculum Vitae

A dispatch from Budapest

Terminal 2 at Budapest's Ferihegy airport is an exemplar of the modern style.  Airy and spacious, with plentiful duty-free options and a serene business class lounge done up in light-toned wood paneling and leather furniture.

It's also empty.  It was clearly built as a showcase for Malev, the national airline, which went bust last year.  My flight to London, departing in 70 minutes, is the next item on the departure list.  I approached passport control to find no line whatsoever, just three officers shooting the breeze.

You don't hear much about Hungary in the contemporary stories of European economic woe -- it isn't one of the so-called "PIIGS."  But there is plenty of woe here.  Many of the locals I met in the course of two days in conference at Central European University worry about the nationalist government that recently took power and proceeded to change key elements of the national Constitution.  But even the best government in the world would face severe economic challenges.  Things aren't as bad here as in, say, Greece, but GDP growth has dipped into negative territory this year and even in the best of recent quarters Hungary's growth rate rarely reached half of Germany's.

Hungary has not yet joined the Euro -- but their one-time aspirations to do so are clear.  The 200-forint coin, pictured here, is clearly designed to emulate the 2-Euro coin.  The principal difference being that the 2-euro coin is worth approximately three times more.  This gives Hungary the option of devaluation as a means of boosting exports, an option that many of their counterparts lack.  Even with this option there is a clear sense of foreboding here.  On the way to the airport this morning my driver commented on how the Forint has once again lost value relative to the Euro.

Earlier this month I ran into one of my doctoral advisors, Martin Feldstein.  I asked him, "how are things?"  and his reponse was "Things are not good.  Things with me are fine, but things in general are not good."  The state of Europe was foremost among his concerns.  And with good reason.

 

Making the $10,000 Bachelor's degree work

I read this morning that a couple of public universities in Texas have responded to Rick Perry's call to offer a $10,000 4-year education.  That got me to wondering, what would it look like to offer a bachelor's degree for that much, without any form of endowment revenue or government subsidy?  I'm going to undergo this exercise with an "old school" mentality, without resorting to big money-savers like online courses.

The typical undergrad curriculum requires something like 32 courses over 4 years.  So our overarching goal is to offer education for about $300 per course per student.  In fact, if we did that exactly we'd use up $9,600 of the $10,000 leaving $400 per student for overhead stuff (administration) -- $100 per year per student.  I'm presuming that the $10,000 is not supposed to cover room and board.  So let's operate a commuter campus or expect room and board to be charged separately.

Starting with the education component, we basically just need two things to educate students: faculty and rooms.  I checked some room rental rates in the Austin, Texas area and found a 278-person theater available at the rate of $1,100 per 10-hour day.  Let's assume that represents something close to the cost of building and maintaining a lecture hall that size.  If we hold 6 75-minute classes per day in that room, the cost works out to $183.33 per class.  If each class meets 30 times -- including lectures and any in-class exams -- then we're looking at $5,500 over a semester.  Divided among 278 students, that comes to about $20 per student per class.  Since we're collecting $300 per student in revenue, that leaves $280 per student to spend on faculty.

The simplest strategy would be to offer faculty 278*280 = $77,900 to teach the course as an independent contractor.  No need to worry about benefits or anything like that.  That's a pretty nice sum to teach one course over one semester.  The faculty member might want some help with grading and that kind of stuff, which could be subcontracted.  Even if you paid 10 teaching assistants $5,000 each to help you, you are still looking at over $25,000 to teach one course, which is a much better deal than many adjuncts get.

We hire 144 of these faculty members and we've got our university.  Now, you might say that it would be a shame to have 100% of the classes at this university be so big, and a shame as well to hire people just for one class.  So let's do the following.  Let's hire full-time faculty and have them each teach the following load per year:

1 278-person lecture

3 15-person seminars

For the seminar courses, we'll need a small room.  So let's rent a 1,000 foot conference room (palatial for a 15-person class) at $20/sq ft per year.  If we use the room 250 days per year that costs $80/day (this means about one-third of our small classes will need to take place in the summer or other times outside the normal academic calendar).  Squeeze 6 classes in each day and your cost is $13.33 per class, $400 for the year (about $27 per student per class).  With 15 students, the revenue works out to $4500, and after paying for facilities we can up our faculty salary offer by $4100 for each seminar, or $12,300 total.

So no we are looking at paying $90,200 for each faculty member.  They will probably want offices, so we'll deduct $3,000 per year to lease them a 150-square-foot office.  They'll also want a benefits package which will probably soak up 20% of total compensation.  So the salary offer works out to a hair under $70,000.

That's probably more than enough to hire faculty in some fields, but not enough in others.  For faculty inclined to do research, we could offer to let them increase their own salary substantially so long as they find grants to pay it (and also cover the costs of their research space, equipment, etc., which we're not paying for).  A four-course load is not the best out there, but far from the worst.  Note that we've dropped the idea of paying for teaching assistants.  In the big lecture courses, we could work exclusively with computerized assessments (on what computers, you might ask... more on that in a moment), expecting the faculty to get involved in the messy work of assessment only in their small courses.  Or we could offer to hire TAs for faculty, deducting the expense from their compensation.

Our university offers 10,000 students 8 courses per year.  So we need 80,000 seats filled.  Each faculty member can teach 278 + 3*15=323 students.  Our university will therefore require a faculty numbering 248.  The student-faculty ratio will be 40:1.  That's big, but not unheard of.

Note that in any given year only 15% of the seats will be in small classes.  So the typical student will only have a 60% chance of getting into one small class over the course of their career -- and that's assuming we forbid any student from taking more than one small class.  We could potentially change that by taking some of our 278-person lectures and making them even bigger to compensate.

Now, how will we spend our $1,000,000 in overhead?  Let's hire a head honcho -- dean, president, chancellor, whatever you want to call him/her -- for $300,000 total compensation (including benefits).  I wouldn't do the job for that kind of money, but I'm sure there's at least one moderately competent person out there who would.  Add in 13 support staff members -- counselors for the students, accountants, etc -- at average compensation of $50,000 each.  Let's give the head honcho a 300-square-foot office and everybody else 150, all at $20/sq ft/year.  That leaves about $5,000 left over, which we can use to rent a small (450-sq ft) conference room for the administrators to have meetings in.

We haven't budgeted anything at all for things like custodial staff and security.  By my math, though, our space needs amount to just over 200,000 square feet, not counting the big lecture halls; given that bulk, we should have the capacity to negotiate with our landlord to include that in the cost of rent.

We have no on-campus restaurants, student meeting spaces, or anything like that.  I'm sure that once the University started up plenty of entrepreneurs would rush in to provide pizza, coffee, and lounge-type spaces of their own volition.  Universities surely have no comparative advantage in operating food and drink establishments.

We have no dedicated space for faculty to lounge around in or hold meetings in.  They would be welcome to use classrooms for meetings, at no charge, on the weekends or outside of class hours (before 8 AM or after 6 PM).  If they want a lunch meeting, one of the nearby food and drink establishments would probably be happy to offer them a private room for nothing more than the cost of lunch -- which they'd have to pay for themselves.

There is no graduate education on this campus.  Some grad programs are cash cows which could be used to help provide some of the amenities we'd otherwise leave out.  Doctoral training would be restricted to those fields where faculty can get outside funding to pay for it.

We have no library and no budget for computing equipment.  We'll expect our students to avail themselves of existing public libraries, to supply their own computers & software, etc.  Or perhaps these items could be available "a la carte."  

Somehow we'd need to do things like record grades, manage waitlists and enrollment in classes, etc, all with zero budget for computers or software.  We might have to knock out one of our administrators to do this, or bump up enrollments in the big classes.  Or just keep some big spreadsheets in our free "Google Docs" account.

We don't even have a budget for wifi on campus, but I went to a college that had no wifi (yeah, because it was 20 years ago, but so what) and I turned out just fine.  I'm guessing that most of our students & faculty will have 3G appliances anyway, and the restaurants across the street -- or maybe even on the ground floor of our buildings -- will surely provide wifi.  Maybe we can negotiate with our landlord to provide us wired internet for faculty in their offices.  Or they can write that expense into their research grants.  We have no budget for faculty phones.  Again, write that into your grant, take a deduction from your pay, or just use the cell phone that you almost certainly have.

We have assumed all instruction can take place in a large lecture hall or a 1,000 square foot conference room.  Labs and other specialized facilities probably cost more per square foot and need specialized equipment to boot.  If we're going to offer these things we'll to charge extra for them.  And that is in some sense fair -- why should an English major pay higher tuition to subsidize the lab equipment of the chemistry majors?

At my University, we have a pretty campus requiring lots of groundskeepers which presumably inspires the students & faculty.  At the $10k University, you'd just recruit some local elementary school kids to paint pretty murals, or maybe locate near a nice public park so long as it didn't make the rents too high.  Let me know if you think of another category of expense that isn't covered here.

This hypothetical University might correspond to your idea of hell.  Or it might seem viable.  To me, it seems to be the educational equivalent of Spirit Airlines.  It does make you wonder how some places charge up to 20 times the price and still can't cover their costs.

Two more charts for undergraduates who happen to be aspiring lawyers

In my previous post we noticed a pattern wherein college graduates with more practically oriented majors earned more money right out of the gate -- in their twenties.  One possible defense for the less-practical majors is that they provide you with thinking and reasoning skills that might not have much direct value in the marketplace, but really help you become a superstar when you couple them with some additional professional training.

So, let's look at lawyers.  I used the American Community Survey to collect information on just over 20,000 attorneys surveyed in either 2009 or 2010.  To be counted in my analysis you had to claim to be a lawyer and also to have received a postgraduate degree.  So, what exactly do lawyers major in when they are undergraduates?

A real hodgepodge of stuff.  The number one answer is political science, followed by business, history, English, and economics.  The "other" category is itself a hodgepodge of a great many things, from sciences to area studies and all points inbetween.  Note, among other things, that there are more lawyers with an engineering major than with a philosophy major.

Which of these majors best prepares you to be a star attorney with high earning power?  Here's a look at median earnings across these same undergraduate majors.

The engineers come out on top, with the typical engineer-turned-lawyer earning 50% more than the median philosophy or English major-turned-lawyer.  Economists do pretty well relative to most other disciplines, too.  To the credit of philosopher-lawyers, they do manage to close the gap with business majors, and come out well ahead of communications or education majors.  So it does appear to be true that the traditional liberal arts can put you ahead of certain more practically-oriented undergraduate programs if your goal is a professional career such as law.  But not all of them.

Two charts every college student should see

I've been noticing lots of commentary lately about the state of higher education in the United States: how there are many graduates having trouble finding an employer willing to put their "critical thinking" skills to work.  This prompted me to collect some data from the 2009 and 2010 editions of the American Community Survey, which asked college graduates to identify their major.  I collected information only for college graduates under 30 with no postgraduate degrees, with the goal of comparing labor market outcomes by major.  There are tons of different majors out there; I restricted my attention to the top 33 by enrollment (each one of these counts at least 1,000 20-something survey respondents in the ACS data).

First of all, here's a chart comparing the unemployment rates of different majors.  The unemployment rate is the fraction of people who want to be working who can't find work.  It doesn't count people who are not looking for work at all, e.g. because they are full time grad students.

For those who may be numerically challenged, the "unemployment rate" here is measured as a proportion rather than the more typical percent.  So for example, we see that former nursing majors have an unemployment rate around 3%, less than a third that of their comrades who studied commercial art or graphic design.

Now, let's look at some income information.  This chart shows median earned income (i.e., the income earned by the person in the middle when you rank people from highest to lowest) for college graduates with no postgraduate degrees currently in the labor market (either employed or unemployed looking for work).  Using the median prevents us from showing statistics skewed by one or two very high earners in a given category.

What do we learn from these two charts?  Those who learn to do something practical, such as nursing, elementary education, or electrical engineering, are much more likely to find work.  Engineering, computer science, and nursing top the earnings list.

What's wrong with philosophers?  They have a high unemployment rate, and the median philosophy major makes something equal to a full-time worker earning $10 per hour -- about the going rate for a line cook in my neck of the woods.  One thing you'd probably want to know is whether the philosophers are all in law school, or some similar program, setting themselves up for a significant career boost in the near future.  I'll try to assess that hypothesis in a later post.

For now, though, the message seems to be clear.  The job market values practical skills.  I'm sure it values critical thinking too, but you've got to do quite a bit of that to be a successful computer scientist, right?

Trayvon Martin, self-defense, and the law

Several years ago I served on a jury here in Durham County.  The case involved an altercation on a street corner, in broad daylight, in which one person fatally stabbed another with an ordinary paring knife.  There were numerous eyewitnesses -- school buses had only recently disgorged students who crowded around the combatants -- and the defendant openly admitted to the stabbing.  Yet we acquitted him.

The defendant -- whom we would later learn had a long criminal record, including a prior conviction in a similar stabbing incident -- claimed that he was acting in self-defense.  The defendant's prior convictions were deemed not admissable as evidence in court.  North Carolina's law regarding self-defense claims holds that in such cases the prosecution must prove beyond a reasonable doubt that the defendant was NOT acting in self-defense.  In other words, a defendant doesn't have to prove the self-defense claim; the prosecution must prove the opposite.  This, as it turns out, is very hard to do.

In the Trayvon Martin case, it is very hard to have sympathy for the shooter, George Zimmerman.  It may also be difficult for many to sympathize with the local law enforcement officials who investigated the case and decided not to arrest Zimmerman.  Considering the facts of the case, though, it would appear that there is little to no chance that Zimmerman can be convicted of anything.  Personally, I subscribe to the point of view that Zimmerman was a trigger-happy wacko inclined to shoot first and ask questions later.  I also suspect that he did engage in racial profiling.  But those viewpoints would probably disqualify me from serving on a jury in the case.  Even trigger-happy wackos are entitled to the essential protections of law.

Compare the Zimmerman-Martin altercation with the one I described earlier.  It occurred in the evening, rather than broad daylight.  If there were eyewitnesses, they were some distance away.  The shooter has certainly engaged in some questionable behavior before, but appears to have no prior felony convictions.  And it occurred in a state with a self-defense law broader than North Carolina's.  All these factors suggest that obtaining a conviction would be even more difficult in this case than it was in the case that I deliberated.

Local law enforcement finds itself in a nightmare scenario.  They are caught between an outraged public and the law.  This event may lead to a reconsideration of the law, and that might be a good idea.  But so long as the law makes any concessions for individuals acting in self-defense, and considers defendants innocent until proven guilty, there will be more cases like this.  Each one is tragic.  But there is no method of eliminating such tragedies without inviting others.  

Will publishing perish?

My line of work is research.  Research, for the most part, is not something you can sell on the private market.  Therefore, one can't evaluate the quality of a professor's research using sales, profits, or anything like that.  Everything is based on reputation -- what do important people think about the quality of your work?

Traditionally, the list of "important people" started with editors -- editors serving academic book publishers or journals.  The editors, in turn, rely on the judgment of other people they trust -- reviewers -- to help them decide whether a submission was good enough to publish.  And publication was the key to getting your work on the desks of other important people -- subscribers.  The journals and presses who provided this review service managed to break even, or in some cases turn tidy profits, by exploiting their monopoly power over content.

Today, the review process is no longer the gateway to dissemination.  And this, in turn, threatens the viability of the entire system.  The widespread availability of journal articles -- for example, on researchers' web pages -- dilutes monopoly power.  The ease of electronic distribution has encouraged new startups -- PLoS in the sciences, for example -- who offer content at no charge.  Some academic disciplines have de-emphasized journal publications.  This further endangers the financial standing of traditional publishers.

Not only is the review process unnecessary for dissemination, it is also not entirely necessary to establish the quality of work.  Instead of inferring the quality of work by whether some journal editor liked it, we can count up the number of times it has been downloaded, or the number of times it has been cited in subsequent work.  The wisdom of crowds replaces the sometimes idiosyncratic tastes of one person.  All in close to real time.  The economic necessity of journals & presses, in sum, has been mostly eliminated by advances in technology.

There's still one service that the journals can provide that the web can't.  They can help authors to improve the quality of their work.  They can point out broader implications that the author might not have recognized.  They can help improve exposition, so that readers understand points more completely. And they can suggest directions for further work.

These services are of clear value to society.  Who should pay for them?  The standard economic answer is "whoever receives the benefits."  Some of them accrue to society at large; but the clearest beneficiary is the author, who earns a better reputation through the process.  So this suggests that the future of academic publishing involves shifting costs from subscribers to authors.

The future is already here.  The PLoS model -- which has enabled the organization to break even -- involves charging nothing to subscribers, but imposing fees on authors of published work (to the tune of nearly $3,000 per article, albeit with waivers for those without the ability to pay).  This might not be perfectly optimal -- some of the benefits associated with the review process accrue to authors whose work is ultimately not published by the journal.  But it is probably closer to the optimum than the traditional model of charging little or nothing to authors, but outrageous subscription fees.

Are public teacher ratings good or bad?

First it happened in LA.  Then New York.  Thanks in large part to the efforts of media organizations, the nation's largest school districts have publicly released information on teachers' ability to improve student test scores.  The releases have been controversial, to say the least.

There are some good things to say about these information releases.  Generally speaking, parents are not very sophisticated about picking schools for their children.  They fixate on high test scores, not realizing that the #1 determinant of a school's test scores is not the quality of the school but rather how advanced the children are on the first day of kindergarten.  New York's teacher rating data, and to some extent the reports mandated by No Child Left Behind, help to expose the folly of this behavior.  There are some purportedly great schools that have decidedly mediocre effects on their students; there are also some apparently mediocre schools that are in fact doing amazing things with difficult-to-educate students.  School choice and competition have been theorized to have strong beneficial effects on school performance, but to realize these theoretical benefits we need parents to make sophisticated decisions that reward schools for what they do, not the students they recruit.  More information helps that.

A system based purely on test score improvement for math and reading teachers in grades 4 through 8, however, is not going to work in the long run.  It's too easy to game the system.  A principal who wants to look better will switch their mediocre teachers into grades K-3, or to middle school subjects without standardized tests.  Volumes of high-quality research have shown myriad other ways in which principals and teachers game the system when high stakes are attached to test scores.

The consequences of sticking the worst-performing teachers in the earliest grades could be grave.  Research indicates that the early grades may be a "critical period" in the development of positive attitudes towards learning.

A good information collection and release system could radically transform public education in the United States, particularly if the information is used to help teachers become better at what they do.  A bad information collection and release system could do the same -- but for the worse, not the better.  

Revisiting Race in College Admissions

So the Supreme Court has elected to hear arguments in the Fisher case (wherein a young Texan sued because UT-Austin declined to admit her).  This is the case that produced the amicus brief that landed my Duke colleagues Peter Arcidiacono and Ken Spenner in hot water.  Note that the same brief also cited some of my joint work with Peter... somehow that reference didn't generate any controversy here on campus.

The Supreme Court's last rulings on race in college admissions (in 2003) argued that there was a "compelling state interest" in ensuring diversity on college campuses.  In other words, the learning experience of all students is enhanced when diverse viewpoints are represented in classrooms and other settings.  That makes a lot of sense in theory.  My work with Arcidiacono (and Shakeeb Kahn) suggests that the argument breaks down in practice: that increasing representation of a group on campus by applying a different admission standard to that group does not increase the rate of meaningful interaction between members of different groups. Our study analyzed data on graduates of 30 selective colleges and found that graduates' self-reported frequency of interactions with members of other racial groups were not influenced by the percentage of those students on their campus (or within their major on their campus). In a separate study, Arcidiacono and I found that graduates' monetary earnings and life satisfaction bore little relation to the degree of racial diversity on their undergraduate campus or in their major.

In other words, the evidence doesn't really suggest that white kids gain anything substantial from being on a campus with a higher proportion of kids from other racial groups.  This does cast some doubt on the "compelling state interest" theory espoused in the 2003 Supreme Court decision.

Here's the thing you have to understand about college admissions.  The system is rigged against the disadvantaged -- whether they be racial minorities or poor whites.  In some ways the rigging is obvious (e.g. legacy admissions).  It's rigged in plenty of less obvious ways too.  For example, most colleges consider only your highest SAT or ACT score.  So you might as well take the test many times, since you're bound to get lucky sooner or later.  The policy therefore favors those who can afford to take the test early and often.  This is not exactly a secret -- I overheard a family discussing exactly this strategy at a fairly exclusive beach resort over the weekend.

In part because of this rigging (but also for other reasons), it is difficult to encounter students from lower-income families on selective college campuses.  To the uninitiated, they resemble country clubs with classrooms.  I don't mind if my students are wealthy, so long as they come to my classroom with a dose of humility and an eager passion to learn.  I've spoken to many faculty members at many places who lament the fact that we don't see many students like that.  Intsead we see students who look at courses as stumbling blocks on their way to something greater, not as opportunities to learn.

So, if it were up to me, I'd toss the entire system -- affirmative action and all.  I don't necessarily want the student who got the top test scores, or went to the best high schools.  I want students with fire in their belly.  I want students who react to negative feedback by getting better, not lashing out.  Those are hard things to assess in a prospective student, but those who start in a humble position in life and make the most out of it are giving us a valuable clue.

A grain of salt for Mooresville

I've been forwarded this article about the public schools of Mooresville, North Carolina, which have earned accolades for (a) providing MacBooks for every child in 4th grade and higher, at a cost of $1.1 million per year which they paid for by eliminating 65 jobs, and (b) witnessing improvements in student performance since the MacBooks were adpoted in 2009/10.  Seems like an easy cause/effect story, right?

So here's some freely available public data for you.  Between 2008/09 (the last pre-computer school year) and 2010/11 (the most recent year with test score data), Mooresville's eighth graders did improve in terms of reading proficiency, from 75% to 82%.  Their math proficiency actually declined from 93% to 91%.

Check out what happened to the district's third grade students -- who do not participate in the laptop program.  Their proficiency rate improved in both reading in math -- by 13 points in reading, from 71% to 84%, and 5+ points in math, from 90% to over 95%.

Oh -- and the third graders of 2009/2010, who got their laptops as fourth graders in 2010/11, saw their test performance decrease in both reading in math, with declines of 2 percentage points in both subjects year-over-year.  So yeah, Mooresville has had a increase in test scores, but much of it can be attributed to the fact that they're finding smarter 3rd graders, not to the fact that they're doing anything special for them as they age.

So while the Mooresville case makes for a good anecdote, I'm not budging from my bearish stance on the use of portable technology in schools.  I freely admit that this is a pretty cursory look at the data, but it is no less cursory than what you'll read in the papers.  This looks like a great potential project for an education policy researcher... we've got all the data you'd need to give it a proper look.

Fear and loathing in Chisago County

Interesting portrait of economic insecurity in yesterday's NYT.  Some would look at the people of Chisago County as hypocrites, denouncing government spending even as they accept it in so many forms, from school lunch subsidies to Medicare.  It is clear, however, that many in this position recognize the dissonance between their ideals and their reality, and are deeply troubled by it.  Even while they acknowledge their dependence on government benefits, they profess a willingness to sacrifice some of them for the good of the nation.  Perhaps this is the heartening part of the story: these middle Americans are not out to "occupy" anything.  They are not calling for somebody else to pay up or accept less so they can keep what they have.

This willingness to sacrifice is in itself somewhat puzzling -- it's the old "What's the Matter with Kansas" problem.  The most basic economic model of political behavior assumes self-interest: the huge proportion of Americans who are not billionaires should all support a tax on billionaires.  Under a self-interest model, a "billionaire tax" would pass with 99.99% of the vote.  What's more, we know of at least some billionaires who would support higher taxes on themselves.  Yet we are not likely to see a "Buffett tax" anytime soon.  Why not?

Cynical people would say that the billionaires have bought the process.  My own work has suggested a different dynamic.  People have difficulty projecting beyond their own portfolio of experiences.  In a place like Chisago County, where billionaires seldom tread, taxing the rich does not appear as the obvious solution to government fiscal problems.  As profiled yesterday, the residents are quite likely to witness instances where government spending makes only a small impact on someone's life.  Based on their personal experience, and their sincere feelings of guilt about taking more than they have put in, cutting spending becomes the right thing to do.

The hotbeds of support for redistribution happen to be the places where rich and poor are most likely to come in contact -- large cities on the coasts.  I don't think that is a coincidence.  The clearest sense of class (and racial) tension I've ever felt in my life occurred about 17 years ago, waiting for a Brooklyn-bound train at the Fulton Street subway station in New York's Financial District.  It was the early evening, and the station was crowded with commuters heading home.  The opposite platform served workers heading uptown -- an upscale and mostly white clientele.  I stood in a predominantly black crowd on the other side.  Several uptown trains passed before a Brooklyn-bound train entered the station.  The growing crowd on the downtown platform never became unruly, but I heard several grumblings to the effect that the transit authority clearly prioritized serving the wealthy uptown passengers over themselves.  I figured that the disparity in service was coincidental -- after all, those uptown trains need to come back down at some point.  But it's easy to understand, in this place where rich and poor come in closer contact than perhaps any other location on Earth, how people arrive at a different worldview than those of similar means living in homogeneous communities.