Which country really has the cleverest students?
Higher education has a strong sense of hierarchy.
And high-profile international league tables are a very public form of this pecking order.
While these might measure a whole range of factors - from reputation and staff ratios to research output - what they do not compare is the ability of students who have been taught in these universities.
But the OECD, in its annual Education at a Glance, has published test results comparing the ability of graduates in different countries.
And it shows a very different map of higher education than the ranking tables, which are dominated by US and UK universities, such as Harvard, MIT, Stanford, Oxford, Cambridge and UCL.
The OECD tested literacy skills among graduates - and the high-flyers were not in the US or UK, but in Japan and Finland.
These figures, based on test results rather than reputation, show a very different set of nationalities from the usual suspects.
The OECD's top 10 highest performing graduates
- New Zealand
- United States
None of the countries in the top places make much of an appearance in conventional university rankings.
But while the names of US Ivy League universities are familiar around the world, Norwegian and Australian universities seem to be turning out more capable graduates.
In the QS World University Rankings, there were 32 US universities in the top 100, but only one from New Zealand.
But graduates from New Zealand are higher achieving than their US counterparts.
There is also the question of cost - and the return on investment in higher education for both students and taxpayers.
The Dutch university system, with low fees, outperforms the United States and England, which charge much higher tuition fees.
Scotland and Wales are not included in this OECD measure, but Northern Ireland is in 14th place.
It casts a light too on how an efficient school system might not translate into success in higher education.
South Korea and Singapore, both high achievers at school level, are below average in the graduate rankings.
And what does it mean for the value of university degrees in countries such as Italy, Spain and Greece, who are languishing at the bottom of these graduate test results?
Andreas Schleicher, the OECD's education director, says the results show ability levels can "vary hugely among people with similar qualifications".
They might all have degrees, but "there are major differences in the quality of higher education".
"When it comes to advanced literacy skills, you might be better off getting a high school degree in Japan, Finland or the Netherlands than getting a tertiary degree in Italy, Spain or Greece," says Mr Schleicher.
These OECD test results may be completely different from conventional university rankings, but the two sets of findings are not incompatible, says Ben Sowter, director of the QS World University Rankings.
While the OECD has compared standards across national higher education systems, the university rankings are focused on an elite group of individual universities.
Mr Sowter says if every university in the US was measured in rankings, it would show "they have a share of the worst as well as the best".
The US has a highly polarised education system, but that is not apparent from a ranking system that focuses only on the top.
QS World University Rankings 2016-17
- Massachusetts Institute of Technology (MIT)
- Stanford University
- Harvard University
- University of Cambridge
- California Institute of Technology (Caltech)
- University of Oxford
- University College London
- ETH Zurich
- Imperial College London
- University of Chicago
The success of a country such as Finland in the quality of its graduates could owe as much to its school system as its universities, Mr Sowter says.
And it is likely to be "harder to run a bad university in Finland than in the US".
But Mr Sowter says the OECD findings highlight a longstanding question about priorities for higher education.
Should countries invest in making sure there is a good overall standard - or should they focus on cultivating a few world-leading institutions?
There would be a good economic case for arguing for a consistently high standard across all universities rather than a landscape of peaks and valleys.
University rankings can highlight differences between individual institutions, but Mr Sowter says they cannot be used to evaluate how well a higher education system is performing.
The OECD runs the Pisa tests, which compare standards in secondary schools in more than 70 countries.
And there was an attempt by the think tank to set up a higher education version, so that comparisons could be made between individual universities.
But universities, particularly in the US, were not at all keen, and there seems little imminent sign of university league tables based on the quality of the students they produce.
Even though there might be scepticism about how international league tables are calculated, there is no avoiding their importance, when universities have to compete as much as brands as academic institutions.
When Oxford was named for the first time this year at the top of the Times Higher Education world rankings, it was headline news.
But according to the OECD tables, maybe the celebrations should be among graduates in Japan and Finland.