Higher Education Strategy Associates

Category Archives: access

September 01

The Tennessee Promise

So, yesterday I talked about a big increase in access in the UK, which seems to have little to do with tuition fees.  Today, let’s talk about a developing story in the United States, where a lowering of net prices seems to have had a big impact on access.

You may recall that in the US over the last couple of years, there has been a growing movement for free community college, something that President Obama picked up on earlier this year.  But before Obama picked up this baton, free community college had already been introduced in Republican Tennessee, where governor Bill Haslam had turned something called “the Tennessee Promise into law in 2014.

Technically, the Tennessee Promise is not “free tuition”.  It’s only available to students entering straight from high school (which is a bit weird in terms of design, but whatever).  Students have to be full-time, maintain a 2.0 average, meet regularly with a mentor, and perform eight hours of community service per term.  And technically, what it does is reduce your tuition to zero after all other forms of aid and scholarship are taken care of (this is what is known in the business as a “last dollar” scholarship).  If you apply for the award and meet the terms, government will cover your tuition to the point where your net price is zero.  For a good number of people, this means free tuition with minimal strings attached, so let’s just call it free tuition.

Now, you might expect that with this kind of incentive, enrolment might rise a bit.  And you’d be right.  According to very early results, the number of freshmen is up 29.6% over last year.  Obviously this is a pretty impressive result, but before we get too excited, we should probably find out a little more about where these new students are coming from.  Are they “new” students, or are they mostly students who would have gone to a 4-year college, but have chosen 2-year instead?  And what about students’ financial background?  If you’re poor enough to be anywhere near maximum Pell grant ($5,775), the Tennessee Promise provides no additional aid, because tuition at Tennessee Community Colleges is about $4,000.  So it may well be that what the Tennessee Promise is doing is providing aid to people higher up the income ladder.  This is a little inefficient, but since (as I noted back here) community college students tend to come from poorer backgrounds anyway, this is not as regressive as it would be if it were implemented at 4-year colleges.

We should be able to answer these questions in a few weeks (yes, Canadians, in some places data is available in weeks, rather than years).  Even though Tennessee does not track applicants by income the way the UK does, the state’s excellent annual Higher Education Fact Book does contain two pieces of data that will help us track this.  The first is college-going rates by county, which will help us understand whether the jump in participation is concentrated in higher- or lower-income counties, and the second is the percentage of students who are Pell-eligible.  I’ll keep you up-to-date on this when the data is out.

The most intriguing possibility here is that rates of attendance for Pell-eligible students might be rising, even though the Tennessee Promise provides no actual added benefit for many of them.  It may well be that simply re-packing the way we frame higher education costs (“it’s free!”) matters more than the way we actually fund it (“your tuition is $4,000, and you also have a grant for $4,500”).

This would have significant policy ramifications for us in Canada.  As we noted last year in our publication, The Many Prices of Knowledge, many students at Canadian community colleges face an all-inclusive net price that is negative, or very close to it.  Similarly, poor first-year university students in both Ontario and Quebec have negative net prices.   No one knows it, because we package aid in such a ludicrously opaque fashion, but it’s true.  And if the Tennessee data provides evidence that the packaging of aid matters as much as the content, then it will be time for Canadian governments to re-evaluate that packaging, tout de suite.

August 31

An Interesting Story about Access in the U.K.

Remember how, in 2012, tuition in England rose by about $10,000-$12,000 (depending on the currency exchange rate you care to use) for everyone, all at once?  Remember how the increase was only offset by an increase in loans, with no increase in means-tested grants?  Remember how everyone said how awful this was going to be for access?

Well, let me show you some interesting data.  The following comes from UCAS, which, at this time of year, does daily (yes, daily!) reports on “accepted applicants” (that is, applicants who have been offered a place at universities for the term commencing in a couple of weeks).  Figure 1 shows what’s happened to student numbers from families in the lowest income quintile since 2011, which was the year before the tuition increase.

Figure 1: Number of Accepted Applicants from the Lowest Income Quintile, England, 2011-15














Big increase, right?  Over three years, it amounts to 19.8%.

“Oh well”, say the zero-tuition true believers, “this doesn’t prove anything.  What really matters is what happened to students from higher income backgrounds.  Surely, being less bound by financial constraints, their numbers grew even more”.

In a word: nope.  The rate of accepted applicants increased by more than three times faster for students from the bottom quintile (quintile 1) than it did for those from the top (quintile 5).  Of course that’s partly because they have a lot more room to grow: there are still about three times as many accepted applicants from the top quintile as the bottom quintile.  But the point is: contrary to expectations, the gap is closing.

Figure 2: Change in Number of Accepted Applicants by Income Quintile, England, 2011-2015, Indexed to 2011














“Ok”, say the skeptics; “let’s look at counterfactuals: what’s going on in neighbouring countries, where policy didn’t involve a massive tuition fee increase?  What about Wales, where tuition stayed at a little over £3,000, or Scotland where tuition is free (for Scots: English kids still have to pay the £9,000)?”

Fair question.  Figure 3 shows what happened to students from the lowest income quintile in all three countries: in Scotland, rates of accepted applicants are up by 28%, in Wales by 21%, and in England by 17%.

Figure 3: Change in Rate of Accepted Applicants, England, Scotland, and Wales, 2011-15, Indexed to 2011














“A-HA!”  Say the usual suspects.  “Clear evidence that free is better!”  Well, maybe.  But before declaring victory, why not look at rates of accepted applicants for low-income students across these three countries?   That is: what percentage of all youth from the bottom income quintile actually reach the stage of being “accepted applicants”?

Figure 4: Accepted Applicants from Bottom Quintile Families as a Percentage of all Bottom Quartile Youth, England Scotland, And Wales, 2011-2015














Quite a different story, isn’t it?  Turns out that in horrible, vicious, neo-liberal, £9,000 tuition England, 18% of lowest-income quintile youth apply, and are admitted to university.  In idyllic, equality-loving, £0 tuition Scotland, the figure is not much more than half that, at 10%.  So let’s just say that the evidence claiming fees explain participation rates, and changes thereof, is pretty limited.

But getting beyond the issue of fees, I think there’s a bigger story here.  Right across the UK, regardless of tuition fee regime, there is a massive uptick in participation from low-income students over the last couple of years.  Clearly, something is going right there with respect to low-income students.  Is it a change in aspirations?  Expectations?  Academic preparation?  As far as I know, no one has published on this – I have a feeling everyone was so keyed on explaining expected declines in participation that no one was set up to explain the opposite.  But whatever is going on, it’s a success, and other countries would do well to learn from it.

June 01

The Cost of Moving the Needle

One of the things about increasing post-secondary participation is that the cost of improving access increases all of the time – as you get closer to universality, the students you want to attract becoming increasingly marginal, academically, and require greater investments in order for them to succeed.

A really good example of this comes from the City University of New York (CUNY), which recently completed an evaluation of its Accelerated Study in Associate Programs (ASAP), which is meant to encourage completion among students taking Associate (i.e. 2-year) degrees at the school, on a full-time basis.  The program design is pretty much what you’d expect: it tops-up financial aid so that net cost is zero, plus throws in textbooks and a transit pass.  Each student gets a personal advisor/mentor/coach, as well as career counselling.  Participants get grouped together in small classes (25 students, or less), and the classes are block-booked so that students can take all their courses either in the morning, afternoon, or evening (of great assistance to students with work or family responsibilities).

This is not a cheap program.  At the time the program was being evaluated, it cost over $4,000 per student, per year, though the cost later fell towards $3,500 per student as the program ramped-up.  In the context of US 2-year colleges, such as the nine CUNY community colleges at which this program was implemented, and where per-student expenditure is about $8,000 per student, this is a heck of a lot of money.  But it works.  MDRC, one of the world’s top social science research organizations, evaluated the project recently using a random-assignment experiment, and found that ASAP’s effects on a range of outcome measures were “the largest it (had) found in any of its evaluations of community college reforms”.

The evaluation (executive summary available here) showed that 40.1% of program participants graduated within three years, compared to just 21.8% of students from the control group, and 25.1% had enrolled in a four-year college by semester 6, compared to just 17.3% in the control group (though many American community colleges offer more technical programs, the colleges at which the program was implemented mostly offered arts programs designed as pathways to 4-year colleges, so this metric is actually quite important, because completion without continuation to a 4-year college is of substantially lower value to the student).

Now, that’s a pretty impressive-sounding statistic: for $4,000/year, ASAP can almost double the graduation rate.   But let’s not get ahead of ourselves: in fact, it takes $14,000, spread over 3 years, to achieve this effect.  And even with a doubling, the program is really only affecting one-out-of-five students; one-fifth of students would have graduated anyway, and another three-fifths still don’t graduate.  So to produce one extra graduate, you actually have to spend something in the neighbourhood of $50,000 or so (it’s not actually 5 x $14,000, because you stop spending money once a student drops out).  That’s a lot of money to get one extra graduate, especially for a general Associate degree, where both public and private returns are quite low.

This is by no means a criticism of ASAP: it’s a good program delivering excellent results.  But it does go to show how much money it takes to move the needle on degree completion.  That’s not all going to come from new government sources; it’s going to require changes in institutional business models to reduce costs in order to put more money into things like counselling, advising, and support.

November 17


If I could ban one word from higher education discussions, it’s “affordability”.  It’s a word without precision, and, particularly when used as a synonym for “accessibility”, it’s downright misleading and harmful.

The worst is when someone uses the raw price of a good – in this case tuition – to indicate “affordability”; as in: “tuition went up 5% last year, and that makes it less affordable”.  This is simply asinine.  When the price of milk or gas goes up, we don’t wring our hands about the “affordability” of milk or gas.  We don’t do this for two reasons.  The first is that “affordability”, as a concept, is a ratio and not a point. It’s a function not just of price, but of available resources.  If people were serious when talking about affordability, they would be talking about it in terms of fractions, not prices.

(This of course raises a question – what should we use as a denominator?  When I talk affordability, I tend to use mean or median family income, because nearly all students entering post-secondary education for the first time are drawing on family resources to do so.  The Canadian Centre for Policy Alternatives tends to use much smaller numbers as a denominator, like whatever the minimum wage happens to be.  I get where they’re coming from on this – many students, as they get older, pick up more of the burden of their education costs [though they also tend to earn significantly more than minimum wage].  My number will tend make the fraction fairly small.  Their number will make it look large.  Who’s right?  It depends; to some extent, we both are.)

Which brings us to the second issue: there are people for whom a night out at the movies is affordable, and others for whom it is not.  For some people a Mercedes S-500 is affordable, for others (most of us) it’s not.  Demand curves slope downward, and affordability matters at the margin, not the average.  Most people are simply not affected by an increase in price.  Even in the largest tuition increase in history – the English tuition hike of 2012, where tuition rose by nearly $9,000 – the net effect on applications was only about 5%.  To the extent that affordability affects accessibility, the issue is always about how it affects student at the margin, not how it affects the average student.

That’s why student aid is important.  Student aid helps the students at the margin (or at least it does so everywhere outside Ontario, where “needy” has been re-defined by a vote-grubbing government as anyone with income under $160,000).  Having grants offsetting higher costs is precisely the way affordability concerns should be dealt with – provided you think that affordability is an access issue.

The problem is, for most people the question of affordability is about almost anything other than accessibility.  For most, it’s about making sure that whoever is paying for tuition has more money in their pocket to have a better “quality of life”.   Parents – you deserve that second vacation each year rather than paying tuition!  Students – you should have smaller loan repayments on your way to being the upper-middle class of tomorrow!

Affordability – as a ratio – is thus an important concept in the way we design student aid to help students at the margin.  But the way most people try to explain the concept, and the purposes for which they deploy the concept, are either wrong or disingenuous.  We need to talk a lot more about access and a lot less about affordability.

September 02

Higher Education as a Positional Good

In policy circles, we talk a lot about whether education is a public or a private good (it’s both), and what the implications are for pricing.  But one thing we don’t talk enough about is the extent to which education is a positional good.  And that’s a problem because our decisions on this topic have serious implications for the way we fund higher education.

What’s a positional good?  It’s a good that derives part of its value from the fact it’s valuable, but not everybody has it.  It’s kind of like when a particular article of clothing becomes “cool” – if too many people wear it, it ceases to be cool.  Status goods are a zero-sum game – every time someone else gets something I have, its value to me decreases.

Now think about education.  If too many people get a Bachelor’s degree, its value as a signal of skill goes down, even if everyone’s still gaining the same set of skills.  Think about what that means: the consequence of education being a positional good is that as access to higher education increases, the value of a “plain” Bachelor’s degree decreases, and degree-holders have to find new ways to distinguish themselves.

One way to do this is to focus less on the credential and more on where it was obtained.  Thus, one common pattern in higher education is that as access increases, so too does stratification.  Harvard degrees, for instance, have increased significantly in prestige as access to education has improved. We haven’t seen much of this phenomenon, because – as Joseph Heath recently pointed out – in Canada our most prestigious institutions accommodate a pretty large proportion of our student body.  Combine McGill, Toronto, and UBC and you’ve got about 14% of the entire population (the equivalent for, say, Yale/Harvard/Princeton is about 0.1%).  Since our top institutions don’t confer (much) exclusivity, Canadians look for higher education distinction in the collection of additional degrees.  Hence the explosion in professional Master’s degrees and (to a lesser extent) PhDs.

What makes higher education so weird as a field of policy is that it’s pretty much the only type of status good that governments subsidize.  And that’s really quite weird when you think about it.  What idiot would try to promote universal access to something that, by definition, not everyone can have?

We justify subsidizing degrees because to some extent they do raise skills and productivity, which is good for everyone, not just the people who get the degrees.  But the fact is, when it comes to private returns, in the early career phase at least, what matters is the positional value of the degree.  And not everyone can get into Harvard, or into a PhD program.  If they could, those goods would lose all meaning.  Scarcity is what makes them valuable.

The cry of people who say that there are too many spots in higher education/ law school/ teacher training are really making an argument that there aren’t enough status goods to go around.  That’s in part the consequence of subsidizing education to improve access – you’re bound to get excess demand.

As they say in the computer business, that’s a feature, not a bug.

March 25

The Cost of Expanding Access in Poor Countries

I’ve been dealing a lot with issues of access in Africa (specifically, Senegal and Uganda) over the past couple of months.  And I think I’m coming to the conclusion that there are some situations where it flat-out doesn’t make any sense to expand access.

If you’re a producer of good and services, the main advantage of poor countries is that labour is cheap.  This is why manufacturing has, over the years, drifted to lower-wage countries – first Mexico, then China, and so on.  But universities don’t work that way.  Academics are significantly more mobile than other workers; If university pay falls behind in Ghana they’ll move to Nigeria or South Africa; if it falls behind in South Africa, they’ll move to the UK or Australia.  So to keep them, salaries have to be well above local norms.  Scientific equipment is sold at a global price, as are journals and periodicals (price reduction schemes do exist for Africa, but universities in places like the Balkans or the ‘Stans are pretty much out of luck on that), which is a huge burden for poorer countries.

As a result, the price differential between rich countries and poor countries for producing university graduates is substantially less than it is for producing widgets.  You can see this most easily if you express countries’ expenditures per student on higher education as a fraction of GDP/capita.  In advanced OECD countries, that number is usually in the region of 30%; in Africa, it is frequently over 100% (and even with that disparity, it’s not even close to buying a similar end-product).  It’s quite simply enormously expensive for governments in this situation to expand higher education.

The natural instinct of higher education policy wonks in this situation is always the same: pile on more resources.  If government can’t afford it, let fee-paying students (either in public or private universities) make up the difference.  And that works, up to a point.  But you still run up against the same problem: the cost structures of those institutions aren’t that different from those of public universities, and the troubles the government has in raising money for public services is mirrored by the troubles individuals have in finding well-paying jobs to pay for that education.

Student loans are sometimes mooted as a solution to the problem, but the repayment problems are enormous.  In Africa, for instance, it’s fairly typical that the cost of a year of study is equal to about 40-50% of an entry-level salary.  That means that even if a graduate does find a job right away, their outstanding debt will be on the order of 150%-200% of their income.  Not sustainable.

This isn’t a question of public vs. private.  It is simply a question of return on investment.  At certain levels of development, there are points beyond which you either have to radically reduce the cost of higher education (perhaps via intensive use of MOOCs, as the Kepler project in Rwanda is doing), or you have to say “enough is enough”, because the return isn’t there.  It’s politically difficult to do, but as with any good, one needs to acknowledge when marginal costs start exceeding marginal benefits.  This may be one of those cases.

November 26

The Changing Median Student

Unless you’ve been in some sort of cave for the last decade, you’ve probably heard conversations about students, which begin with the phrase, “Today’s students are… less engaged/less able to write/weaker at math/not as curious/not as academically inclined…”  The obvious question of “compared to when” is usually left unanswered, and depends to some extent on the age of the person doing the kvetching; solipsistically, I always assume they’re talking about 25 years ago (when I started university). Usually, the implication is that students are coming out of high school less university-ready than they used to be, and hence, that high schools aren’t doing as good a job as they used to.

So, is it true?  Well, to start, we have to come to grips with what we mean by “students”.  The student body of the 1980s looks a lot different than the student body of 2013 – and I don’t just mean in terms of the changing ethnic or age mix.

Take a look at the following graph, which shows the change in participation rate of 18-21 year olds in Canada, over time.

Participation Rate (in percent) of 18-21 Year Olds, Canada, 1980-81 to 2010-11













Cool, huh? We tripled access rates in thirty years. Remember this the next time you hear that accessible education is dead.

Now, assume for the moment that the fraction of Canadians who go to university are all at the top-end of the academic ability scale – that is, it was the 10% most academically able students in university in 1980-81, and it’s the top 30% now.  I know that’s not literally true, but it’s true-ish, so bear with me.  Today’s median student is therefore in the 85th percentile of academic achievement. Prior to 1988, when access rates were below 15%, we would not have even let that student into university.

Given this, it would be very surprising if today’s students weren’t, on average, academically weaker than those from 25 years ago.  But it doesn’t follow that students coming out of high school are getting worse.  To compare like-to-like you need to ask the question: is the average student of the top-half of this year’s university class better-or-worse prepared than the average student of 1988?  Pose the question that way, and I think you get a different answer.

Expanding access to universities brought in a new type of student whose academic orientation was different – but the curriculum and the student support system didn’t always change to match their abilities and expectations.  These are completely understandable reasons for the persistent disconnects and the complaints about “unprepared students” – but they don’t support the thesis that high schools are doing their job any less well than in the past.

October 28

Debt, Tuition, and Inequality

A few weeks ago, I noted on Twitter that back when I was in student politics (24 years and counting) we opposed tuition hikes because we feared their negative effect on access.  Back then, fees hadn’t increased in real terms in almost two decades, so there wasn’t much evidence either way on the issue.  More than two decades on, the evidence has accumulated, and, on the whole, it turns out that what we believed back then was mostly mistaken: despite much higher tuition, more students than ever are attending PSE, and more low-income students than ever are attending PSE.

As a result, it’s increasingly ridiculous to use access arguments against tuition increases: basically you’re reduced to slogans like, “tuition reduces access!  And it’s a disgrace more people are paying it every year!”  Given this, I asked Twitter peeps what the best grounds were for opposing higher tuition and debt.  The most common answer was some variation of “tuition/debt causes inequality”, on the grounds that graduates with debt accumulate assets more slowly than students without debt.

Superficially plausible, maybe, but still wrong.  Inequality doesn’t exist “because of” borrowing.  Everybody pays equal tuition fees; debt is incurred by those who don’t have the cash up front to pay for it.  Blaming student aid for inequality is just blaming student aid for the fact that some students come from poorer families, and others from richer ones.  You don’t have to condone inequality to realize that it’s a silly proposition.

So, students who go in rich don’t accumulate debt, and therefore end up richer at the other end; students who go in poor do accumulate debt, and therefore remain poorer at the other end.  But that would be true regardless of how fees are set.  If you reduce fees for all, everybody is made better by the same amount.  Poorer families end up with less debt, richer families get to keep more cash-in-hand.  For the abolition of fees to reduce inequality, it would have to create some kind of behavioural change among richer parents that would make them less likely to pass that extra money on to their kids.  And how likely is that?

These are all pretty basic observations, yet they rarely make it into the debate about tuition and debt.  Partly, it’s because inter-generational transfers complicate the analysis, but I think it’s mostly because debt makes people squirrelly.  As Margaret Atwood and David Graeber have pointed out in recent books, the concept of debt is tied-up with an enormous amount of cultural baggage, which makes it difficult to talk about in purely economic terms.  For the next couple of days, I’ll be unpacking some of these cultural issues, and how they affect our discourse on debt.

More tomorrow.

June 17

A Dreadful Argument About Tuition Fees

I see that the Canadian Centre for Policy Alternatives has just released a new paper by Hugh MacKenzie called The Impact of Taxation on the Higher Education Debates. It’s worth a read because it sets out the argument against higher fees in the most respectable terms possible – certainly more respectable that anything student groups themselves have come up with.

It is still, however, a pretty crap argument.

The spiel runs like this: the lazy talking point about how higher education subsidies mean that “the poor pay for the education of the rich” isn’t actually true. Our system of income taxes is quite progressive; the top income quartile pays about 69% of all taxes and the bottom quartile pays for just one percent. As a result, on a net basis, the bottom income quartile gains from the current system because they make up more than 1% of all university students.

MacKenzie is dead on about this. It’s where he goes from there which is problematic. He continues by arguing that its OK that subsidies for education, on aggregate, end up disproportionately with the children of wealthy families, because on average, the wealthy pay most of the taxes. Effectively, he argues that as long as the top income quartile doesn’t claim more than 69% of the expenditures of any program, and as long as the lowest income quartile gets more than 1% of the expenditures, the program is ipso facto progressive. And since higher education meets that test, it’s OK to spend more money on it.

This, to put it mildly, is an interesting definition of the term “progressive”. Education Savings Grants, which mostly go to the rich, would be considered progressive according to this definition. So, too, would the Bush Tax cuts (h/t Stephen Gordon).

Surely the relevant issue isn’t whether public subsidies can pass an unbelievably weak test such as this. The issue is how to ensure that subsidies are directed to those who need them most. This of course is what most people who argue for subsidy reform argue; higher fees for the rich, lower net fees for the poor through increased grants. For McKenzie this option is never considered; whether through ignorance or deliberate omission, he never mentions the billion or so in student grants which permit this kind of pro-poor price-discrimination.

MacKenzie self-righteously claims that anyone who disagrees with his views on progressiveness is just “appropriat(ing) the language of fairness” while “arguing against public policies whose goals include equalization of opportunity”. So be it. If arguing against providing rich families with bigger subsidies because they pay more taxes makes me anti-progressive, I’ll wear that label with pride.

The question is – why wouldn’t CCPA do the same?

May 16

Think Big?

With all the chat recently about reducing unit costs through ever-larger instructional units (e.g. MOOCs), it occurred to me that the world already has a lot of models for this.  They just aren’t in the developed world.

University World News recently carried a very interesting article regarding a new higher education master plan in Nigeria.  One of the plan’s key elements is to construct a half-dozen “mega-universities” – each with 100-150,000 students – to soak up the rising demand for higher education.  On the one hand, this plan is self-evidently mad: large Nigerian universities are already a violent and lawless mess, plagued with cults such as the Black Axe and the Supreme Vikings (I wrote about them a couple of years ago: here); surely these new, even larger campuses will face even bigger gang problems.  On the other hand, you can sort of see where Nigeria’s coming from on this.  Thanks to some truly staggering levels of corruption, the ability of Nigeria to use public funds to meet demand for higher education is quite small – currently just $1.4 billion to cover expenses at 33 federal universities.  So the solution is simple – go big, and keep unit costs low.  Just like MOOCs.

Actually, the way access has been increased in much of the developing world is through strategies like this.  The world’s largest universities are Open Universities – Indira Gandhi in India (3.5 million), and Anadolu in Turkey (2 million).   The largest residential schools are ones with multiple constituent campuses.  The reigning world champion here is Islamic Azad University in Iran – a private school with 350 locations, 1.5 million students, and a very significant endowment of contested legality (I don’t buy the $200 billion number, but it’s substantial nonetheless).

What about single-campus institutions?  On the Indian subcontinent, there are a handful (e.g. Delhi, Pune) which boast enrolments of 400K plus, but most of those students are not residential – rather, they study at a college somewhere, and simply take the Delhi or Pune exams.  For really big schools, you need to go to places like the University of Buenos Aires (300K plus) or UNAM in Mexico City (250K plus).  The University of Cairo, at about 150K, is the biggest in Africa; it’s also generally considered the continent’s best school outside of South Africa, which may explain Nigeria’s attraction to the model.

William Gibson once said that the future is already here; it’s just unevenly distributed.  So it is.   These mega-institutions can provide some lessons about the perils and promises of uber-massification through mega-universities.  We probably shouldn’t ignore them just because they’re happening offline and in poor countries.

Page 1 of 3123