HESA

Higher Education Strategy Associates

Tag Archives: Expenditures

September 14

Notes on the Finances of China’s Top Universities

One of my distractions over the past summer has been to learn more about Chinese universities.  And, fortunately, this is becoming a lot easier as Chinese universities are starting to put more of their data online.  Today, I just want to take you through a bit of a tour of China’s top universities (roughly the equivalent of the US Ivy League), which are known as the “C9”, most of which now put their financial data online.

So let’s start just by looking at raw annual expenditures (I prefer using expenditures to income as a guide to a university size because it tends to be more stable year-to-year) at these top universities.  Figure 1 shows this by institutions for the 2015 calendar year.  Tsinghua leads the pack by a wide margin, at a little over RMB 13 billion.  Peking, Zhejiang and Shanghai JiaoTong are next at between RMB 8-9 billion Yuan, Fudan followed by Fudan Xi’an Jiao Tong at between RMB 5-6 billion.  The bottom positions are held by the two C9 universities which do not report to the higher education ministry: the University of Science and Technology of China (Chinese Academy of Science) and the Harbin Institute of Technology (Ministry of Industry and Information Technology) at RMB 3.4 billion and RMB 2.2 billion, respectively.

Figure 1: Expenditures, in Billions of RMBTop Chinese Universities, 2015

One interesting piece of information about these institutions is how little of their annual budget actually comes from government.  Figure 2 shows government appropriations as a percentage of annual expenditures (Harbin Institute of Technology is excluded because its financials do not distinguish between public and private sources of revenue).  As it turns out, top Chinese universities actually look a lot like Ontario ones in that they tend to get less than half their money from government.  That said, at most institutions student fees only account for about 15% of total revenue.

Figure 2: Government income as a % of total expenditures, Top Chinese Universities, 2015

Now at this point you may be wondering: RMB 13billion….is that a lot?  A little?  What’s the frame of reference here?  Well, fair enough.  Let’s put all this into US dollars, just so we’re clear.  And for reference, let’s throw in data for Harvard, Berkeley, U of T and UBC for 2015-16 for comparison.  To do this, I’m converting to USD at the mid 2015 exchange rate of RMB 6.21 = CDN $1.29 = USD $1.  The results are shown in Figure 3: By this measure, only Tsinghua is really up in the North American big leagues.

Figure 3: Total Expenditures, in USD, Top Chinese Universities plus US/Canada Comparators, 2015

But hang on a second.  What if we use purchasing power-parity instead of exchange rates?  Well, actually, this changes things more than you’d think.  If you convert the data at the mid-2015 Big Mac Index rate of RMB 3.55 = CDN $1.22 = USD $1.

Figure 4: Total Expenditures, in billions of USD at PPP, Top Chinese Universities plus US/Canada Comparators, 2015

Once adjusted for PPP, Tsinghua moves closer to Harvard, and the next three are more obviously in the big leagues, having all passed UBC.  Now in fact, PPP probably overstates universities’ buying power somewhat, because for many of the goods what universities purchase (top professors, scientific equipment, etc), the price is global rather than local.  So if you want to think about relative purchase power, a fair comparison between the institutions is probably somewhere between figure 3 and figure 4.

(If we were to do this from the perspective of “how big is each institution relative to the size and development of the economy” – that is, adjusting for GDP per capita, all the Chinese institutions would rise by a factor of four relative to American ones, i.e. Tsinghua would be three times as large as Harvard.

Now, what about dollars per student?  For this, I take the student numbers the institutions report to Quacquarelli Simons (QS) for use on its “top universities” website.  You can take these with a grain of salt: I can’t get QS’ numbers to line up with the data I have directly from any of these institutions, but it’s the most consistent thing we’ve got, so we’ll just have to live with it.

Figure 5: Expenditures per Student, in USD at PPP, Top Chinese Universities plus US/Canada Comparators, 2015

Now Tsinghua is much more clearly in an Ivy-League-approaching kind of position, with expenditures of over $100,000 per student.  That’s not near Harvard, which spends about twice that, but it is a full 25% higher than Berkeley and 150% higher than UBC and Toronto.  Even the Chinese second-tier trio of Shanghai Jiao Tong, Peking and Zhejiang are spending 50% more per student than the top Canadian universities.

In short, the top Chinese universities aren’t, as it is sometimes said, “rising”.  Financially, they’re already comfortably part of the world elite.

September 13

Some Curious Data From OECD Education at a Glance 2017

The OECD put out its annual Education at a Glance  publication yesterday.  No huge surprises except for the fact that they appear to have killed one of the most-used tables in the whole book (A.1.2, which compared tertiary attainment rates for 25-34 year olds by type of tertiary program – i.e. college v. university) which is an enormous bummer.  The finance data says what it pretty much always says: Canada is the #2 spender overall on higher education at 2.6% of GDP (just behind the US at 2.7%).  If you read my analysis last year, the story is still pretty much the same this year.

But there are some interesting nuggets buried away in the report nevertheless – stuff that other media won’t pick up.  I thought I would highlight two of them in particular which pose some thorny questions about Canadian statistical data and what we think we know about higher education.

Let’s start with the data on expenditures per pupil at the tertiary level.  Figure 1 looks at costs in Short-cycle Tertiary Education (meaning career-oriented, which in Canada’s case means community colleges)

Figure 1: Total Expenditures per Student, Colleges (or equivalent), Select OECD countries

Among major countries, Canada spends the most (from both public and private sources) on college or college-equivalent student.  A couple of countries actually do outspend us (the Austrians and – totally out left field – the Czechs), but the important point here is that our expenditures are nearly 40% above the OECD average.  And if you’re wondering why the UK and the US aren’t there, it’s because the former has no college equivalent and the latter chooses to not to report on colleges on the batshit crazy spurious grounds that even if you’re studying for a (college-equivalent) associate’s degree, the fact that this can be laddered up into a full bachelor’s means everything is really degree-level.  Nonsense, I know, but there we are.

Now, let’s do the same with universities:

Figure 2: Total Expenditures per Student, Universities, Select OECD countries

There’s not much in figure 2 we didn’t already know: US and Canada in terms of total expenditure per university student at the top with us over 50% above the OECD average and Korea way down at the bottom because the Koreans do everything in higher ed on a shoestring.

Now, one new little detail that OECD has added to Education at a Glance this year is that it splits out the portion of total expenditures (that is combine short-cycle and degree-levels)  which are devoted to R&D.  And this data is a little odd.

Figure 3: Total R&D Expenditures per Tertiary Student, Selected OECD Countries

There’s nothing obviously egregiously wrong with figure 3 – except for the data on the USA, which is bananas.  Read literally, it suggests that Canadian universities on average spend twice as much on R&D as American ones do and that’s simply false.

(The explanation, I think, is that Canada and possibly some other countries claim that all professors’ time spent on research – notionally 40% of time or thereabout – counts as “R&D”.  Whereas Americans claim that their universities – which only pay staff for 9 months a year with the rest of the time notionally off for research – do not count time that way, preferring to claim that the government is buying profs’ time with research grants.  Basically, they view universities as mailboxes for cheques to pay for staff time and so all that time money gets claimed as government expenditure on R&D, not university expenditure on R&D.  GERD, not HERD, in the innovation policy lingo.  I think, anyway).

What’s actually a little crazy about figure 3 is that the denominator is all tertiary students, not just degree-level students.  And yet we know that R&D money is pretty heavily concentrated (98%+) in universities.  In a country like Germany where over 99% of tertiary students are in degree-level institutions, that’s not a big deal.  But in Canada, about a third of our students are in short-cycle programs.  Which means, if you do the math, that in fact the R&D expenditures per university student are a little ways north of $9750.  Now here’s figure 3 again, with just degree-level students in the denominator.

Figure 4: Total R&D Expenditures per University Student, Selected OECD Countries

And of course, subtracting these numbers means we can revisit figure 2 and work out total non-R&D expenditures per student in universities.  Canada still remains 40% or so ahead of the OECD average, but is now similarly that far behind the US in per-student expenditure.

Figure 5: Total non-R&D Expenditures per University Student, Selected OECD Countries

Now, to be clear: I’m not saying OECD is wrong, or Statscan is wrong or anything else like that.  What I’m saying is that there appear to be major inconsistencies in the way institutions report data for international comparative purposes on key concepts like R&D.  And that this particular inconsistency means that Canada at least (possibly others) look a lot better vis-à-vis the United States than it probably should.

Just something to keep in mind when making comparisons in future – particularly around research expenditures and performance.

September 06

Canadian University Finance Statistics (2015-16 Edition)

The 2015-16 version of Financial Information of Universities and Colleges Survey (which, confusingly, doesn’t include community colleges) was released over the summer.  As in previous years I’m going to do a little summary of what it tells us about how income and expenditure has change over one year and five years.  Just so we’re all clear, all figures here are in real (i.e. inflation-adjusted) dollars.  And – caveat – comparisons with 2010-11 are a little weird because Quebec universities changed their fiscal year-end that year and only reported 11 months of data, meaning that nationally, reported expenditures for that year are probably about 1.5% lower than they normally would be.  This means that 5-year averages are probably inflated slightly compared to reality.

For starters, let’s look at total income by source, which was $34,385 million.  That’s down nearly 5% from the previous year, though the a little over 75% of the drop is due to a fall in endowment income (apparently everyone got hammered in 2015).  Income from governments fell by a little under 2%, nearly all due to reductions at the provincial level.  Over the past five years, revenue from government is down by a stonking 12.6%.  However, rising fee income mostly compensates for this: it rose by nearly 5% over 2014-15 and 27% over 2010-11.  For the most part, these increases are not coming from domestic student fees, they are coming from increases in international student enrolment.

Figure 1: Change in Total Income by Source, Canadian Universities, 2015-16

What’s really interesting about the total income numbers is how small the government numbers are becoming.  Already as of 2014, the university sector as a whole took in more money from non-governmental sources (fees, donations, sale of goods and services, etc) than it did combined from the federal and provincial governments.  On current trends, income from student fees will surpass provincial government grants to universities in 2020-21, and will pass combined federal and provincial contributions in 2024-2025.  At which point it would be fair to say we will have moved from a public university system to a publicly-assisted university system.

Now, on to the changes in income by Fund, which I show below in Figure 2.  This tells a slightly different story.  Operating income actually kept pace with inflation in 2015-16 and over a five year period actually increased by 8.8%.  Endowment income fell from about $1.5 billion to about 27 million, or a fall of roughly 98%, but this is an erratic income source and like I said last year was a bad year.  Capital expenditures are down substantially, but recall that in the base year of 2010-11 the feds were sacrificing billions to the Construction Industry Gods to keep the recession at bay; in fact, current capital expenditure is close to the 30-year norm.  The interesting piece is that sponsored research income is down 6% over the past five years.

Figure 2: Change in Income by Fund, Canadian Universities, 2015-16

On to expenditures by type.  Total expenditures are roughly unchanged either over one year or five years. If you’re wondering how this is possible when income is down, recall that most of the income drop is in endowment, which has very little impact on year-to-year spending since it’s supposed to all be salted away anyway.   But while total expenditures are unchanged, some fairly big line items continue to rise, over the medium if not the short term.  Academic salaries by 7.5%, salaries to non-academic staff 8.3%, total compensation (including benefits) 8.3% and scholarships – three-quarters of which go to grad students – up by a whopping 16.4% since 2010-11.  Total scholarship expenditures are now just shy of $2 billion, which means institutions are giving back to students over 20 cents of every dollar they collect from students; from domestic students the figure is closer to 30 cents.

Figure 3: Change in Selected Types of Expenditure, Canadian Universities 2015-16

 

Now you may well ask yourself: wait a minute.  Total expenditures are flat, but salaries and scholarships are rising.  So how does this balance? Well, simple enough: non-salary, non-scholarship expenditures have fallen by 14% in the last five years in constant dollars.  Some of that is just buildings not getting built (no loss, in the eyes of some), but other things are getting squeezed, too; notably, renovations, travel and printing.

Finally, let’s look specifically at what’s going on inside the operating budget (that is, excluding ancillary, capital, research and the like) which accounts for about 60% of the total.  Figure 4 shows that overall, operating expenditures rose by 14.3% over five years.  How is this possible when operating income only rose 8.8% you ask?  Mainly, because universities have been trimming margins: universities were running surpluses five years ago and mostly aren’t any more.

 

Figure 4: Changes in Operating Budget Expenditures, Canadian Universities, 2015-16

The big expenditure increases are in ICT and student services.  In the case of student services, an awful lot of that increase is scholarships.  In ICT, interestingly, the cost of equipment purchased has actually gone down: the increases are in staff costs, consulting contracts, professional fees and equipment rentals.  Make of that what you will.  The biggest piece of the pie – Instruction and non-sponsored research (meaning basically what it costs to run core academic functions), which takes up about half the operating budget – is up 11.7 % over five years.

So there you go.  Don’t say financial reports aren’t fun.

 

October 27

Fun With Library Statistics (Part 2)

Is there any part of the university that has been more transformed over the past decade than libraries?  One of the fascinating things about looking through old Canadian Association of Research Libraries (CARL) statistical reports is how many things weren’t counted, say ten years ago.  Expenditures on databases?  Not counted.  Logins to databases?  Searches or article requests?  Nope, nope.  Not that those things didn’t exist back then – they just weren’t central enough to university missions to be thought worth counting.

One thing which was counted back then (and is still counted today) is loans.  And if you want to get a sense of how libraries have changed in the past decade or so, check out this graph of changes in initial loans between 2004-05 and 2014-15, by institution.

Figure 1: Change in Numbers of Initial Loans, Canadian Research University Libraries, 2004-5 to 2014-5

ottsyd20161026-1

Nationally, initial loans are down 58% across 27 CARL universities over the past decade (Brock and Ryerson are not shown because they did not provide statistics in 2004-05), from 11 million per year to just 5 million per year.  Simon Fraser has experienced the lowest drop – just 24%.  At Concordia, the fall was 81%.  And these figures do not account for growth in student numbers: add those in and figures would drop another 20% or so.

In other words, libraries are decreasingly about books but increasingly about electronic resources.  And this has had impacts on employment.  Across all 27 institutions, FTE employment is down 11%, with the biggest falls at some of the most research intensive universities: Alberta down 35%, McGill down 33%, Queen’s down 26%.  Toronto bucked that trend with a fall of just 5%, and four institutions (Calgary, Carleton, SFU and Ottawa) actually saw increases in employment.  But overall we are seeing a decline in numbers.

But, as with professorial staff (with whom they are often grouped in the same collective agreements), librarians have seen a considerable upward drift in pay.  So while FTE employment is down, nationally, aggregate staff compensation is still up 3% in inflation-adjusted dollars (for comparison, aggregate professorial salaries are up about 39% over the same period).  But here again the variation at the institutional level is absolutely enormous, from +50% at Regina, to -22% at McGill.

Figure 2: Change in Real Expenditures on Library Staffing, Canadian Research University Libraries, 2004-5 to 2014-5

ottsyd20161026-2

So much for staffing: what about acquisitions (or, more broadly, materials)?  Good news here, these grew by 9% between 2004-05 and 2014-15, though again, there’s really not that many institutions which are close to the national mean.  Of particular interest here are the identities of the two institutions who saw the biggest increase in spending: namely, Memorial and Ottawa.  Probably not coincidentally, these are the two institutions currently engaged in the biggest rows about cutting back on periodicals (see here and here).

Figure 3: Change in Real Expenditures on Library Materials, Canadian Research University Libraries, 2004-5 to 2014-5

ottsyd20161026-3

These last two graphs raise the question: do institutions have consistent strategies with respect to allocating their budgets between staffing and acquisitions?  Are both budgets being raised (or lowered) in tandem or are schools cutting in one so as to invest more in the other?

Figure 4: Change in Staffing Budgets v. Change in Materials Budget (in real dollars), Canadian University Research Libraries 2004-05 to 2014-15

ottsyd20161026-4

Here’s the way to understand figure 4.  In the upper left quadrant, you have institutions which are increasing their staffing budgets, but decreasing their materials budget (at the very top left is Regina, which is +50% and -15%).  In the top right quadrant you see institutions which have increased both staffing and materials (the most notable example here is Ottawa, at +56% and +24%).  Moving on clockwise to the bottom right, you see institutions which have cut staffing budgets and increased their materials budgets, and here you have both McGill and Alberta cutting the former by about 20% and increasing the latter by 12%.  And finally, in the lower left quadrant, you have three institutions (Queen’s, Montreal and Windsor) which have seen real decreases in both staffing and materials.

Nationally, there seems to be very little correlation – positive or negative – between changes in one kind of spending and change in the other.  To the extent that library expenditures are strategic, institutions seem to be pursuing a wide variety of strategies in this area.  It would be interesting to correlate this data with user satisfaction surveys to see if any of them are more likelier than others to produce satisfactory outcomes.

October 26

Fun With Library Statistics (Part 1)

The Canadian Association of Research Libraries (CARL) recently issued its annual statistical report.  I thought I’d take the opportunity over the next couple of days to take a look at a few interesting patterns in library practices and expenditures.  They shed some interesting light on the pressures Canadian academic libraries face right now.

Some methodology here: CARL has 29 university members, from the very large U of T (almost 74,000 FTE students) to UNB (under 8,000 FTE students).  As a result, expressing figures in terms of “average per institution) (as CARL does) is kind of weird.  So I have chosen to display all institutional values on a “per student” basis.  This has some problems of its own (libraries don’t just service students, not all students use libraries with same intensity, etc.) but they are less severe than doing it on a per-institution basis.

Let’s start with Library expenditures. A little over half (55%) of library expenditures go towards salaries and benefits.  On average, across all institutions, CARL universities spend $432/student per year on this, but the range is enormous.  At the top end Memorial and Calgary were spending over $700/student, but at Sherbrooke, Brock and Ryerson the amount was under $300 per student.

Figure 1: Per Student Expenditures on Salary & Benefits, Canadian Research University Libraries, 2014-15

ottsyd20161025-1

Most of the rest of the library budget goes on “materials”, which is predominantly the acquisition of titles and periodicals.  Across CARL institutions, the average expenditure on this item is $365/student per year, but again there is huge variation around the mean. McGill and Memorial both spend over $600/student; Saskatchewan spends over $700/student.  At the other end, Ryerson, UQAM and Brock are all under $200 per student.

Figure 2: Per Student Expenditures on Library Materials, Canadian Research University Libraries, 2014-15

ottsyd20161025-2

As you can see, some universities end up towards the top of both lists (mainly those in provinces with a lot of natural resource revenues), and some end up towards the bottom (mainly those universities which aren’t actually all that research-intensive).  But the interesting thing to me is the relationship between those two graphs: what’s the ratio of spending on staff to spending on materials?  Well, nationally, CARL members spend just under 85 cents on materials for every $ spent on salaries.  Again, there’s quite a bit of variation.

Figure 3: Ratio of Expenditures on Library Materials to Library Staffing Costs, Canadian Research University Libraries, 2014-15

ottsyd20161025-3

 

At one end, you have McGill, which spends $1.38 in materials for every dollar they spend on staffing, which suggests either that they are a lean machine or that their collection specializes in some ludicrously expensive journals (it’s probably not a coincidence that all the institutions at the left hand side of the graph have medical schools).  At the other end, curiously, you have the two francophone Quebec universities where the spending ratio is extremely low: 50 cents on materials for every dollar in staffing at Université de Montréal and 48 cents at UQAM.

Might these differences in some ways be related to utilization rates?  One crude way of looking at usage is to look at turnstile counts, which are tracked at 23 of the CARL institutions.  Measured on a per-student basis, one again sees massive differences, the average across the country is 70 turns per student per year, but it ranges at the top end with Saskatchewan recording 100 turns per student per year to Laval, which only gets 24.  One might expect that universities with higher turnstile counts might need a slightly higher staffing count to deal with more users.  But in fact the relationship between the two is essentially non-existent.

Figure 4: Turnstile Counts vs. Materials:Staff Ratio, Canadian Research University Libraries, 2014-15

ottsyd20161025-4

That leaves us with a bit of a puzzle as far as why we see such different cost patterns at different universities.  There are presumably some relevant structural and historical reasons why you get this kind of spread: different universities need of different library services, costs may be elevated because of specialized holdings; for instance, MUN manages a Centre for the Study of Newfoundland which is undoubtedly costly and would likely elevate staff costs considerably).  So, one shouldn’t leap to conclusions about the efficiency of any particular library based on this kind of comparison; but at the same time this kind of data does allow us to ask much better questions about why each university’s library cost structure looks the way it does.

Not that Libraries have that much to answer for in this respect. As we’ll see tomorrow when we look at trends over time, libraries have been able to contain their costs far better than the universities to which they belong.

July 23

Go West

The key to understanding what post-secondary education is going to look like a few years down the road – say, 2017 – is to look at what is likely to happen to government funding.   We can’t know exactly what governments will spend on PSE, but we can know  how much money they are going to have available to spend simply by working out how much money each will likely have once health expenditures (which make up just 40% of the budget in most provinces) are accounted for.  Today, I’ll be doing this for the country’s 4 largest provinces, which make up 88% of the country’s population

With the exception of Ontario, most provinces’ governments move more or less in step with GDP growth.  Quebec and British Columbia keep their expenditures steady at around 20% and 15 of the economy, respectively.  Alberta’s fluctuate somewhat, usually in line with changes in hydrocarbon prices.  Since the Liberals took office in 2003, Ontario has been increasing the size of its government quite steadily from 13 to 18% of GDP.

 Figure 1 – Provincial Budgets as a Percentage of GDP, 2000-2013

 

Given this, it seems unrealistic to expect any of these governments to increase overall expenditures much faster than GDP growth.  Alberta and BC could conceivably inch up a bit after 2014 since their spending is currently slightly below the long term average.  For simplicity’s sake, though, let’s assume that growth in those two provinces will be restricted to growth in nominal GDP, which in both provinces is expected to average 4.5-5% for the foreseeable future.

Quebec and Ontario, meanwhile, can’t grow expenditures anywhere near that much because of their abysmal finances.  Quebec’s budget currently projects spending growth to be around GDP growth minus 1% out to 2017; in Ontario, program spending is frozen in nominal dollars through to 2017.

Now, the amount of money available for PSE (and other types of government spending) is limited by what happens to the health budget.  With the overall size of government more or less steady as a percentage of the economy, every time the health budget increases more quickly than GDP, the pool of money available for every other piece of spending – including PSE – must decline. In Quebec and British Columbia, the health budget has been growing at 5-6% annually for the past decade.  In Ontario, the figure is 7% and in Alberta it is a (frankly) ludicrous 9%. 

So, let’s assume that everyone can keep health care increases to just fractionally above expected GDP growth levels (say, 5% per year). Here’s what will happen to the pool of non-health dollars available in each province: 

 Figure 2 – Nominal Non-Health Dollars Available by Province, indexed to 2013.

 

Your eyes do not deceive you: that is indeed a 41% funding gap opening up between Alberta and Ontario over the next four years.  Given the assumptions above, non-health spending in Alberta can grow by 20% by 2017 and BC looks set for an increase as well.  Quebec should hold just about steady; Ontario, thanks to its need to get rid of its deficit without raising revenues, is going to see a fall of a little over 15%.

To be clear: I am not saying that PSE budgets will increase or decrease by these amounts.  What I am saying is that this is a good approximation of how the amount of funds available to PSE will evolve in each province over the next four years and that if historical funding patterns hold up, these kinds of changes in nominal funding are about what we can expect.  But politics still matter, and universities and colleges could still see increases to their budgets relative to the amount of available funding if they are smart in their lobbying (or cuts if they are not).  

A couple of years ago I said that differential patterns of higher education investment meant that the country’s intellectual centre of gravity was moving west rather quickly.  Notwithstanding recent cuts in Alberta and British Columbia, it seems to me that this trend can only pick up steam in the next few years.