Life is a Journey of Learning: Your Universities Didn't Get Worse. The Scoreboard Ran Out of Seats

Every August, the same ritual plays out. The Shanghai Ranking, formally the Academic Ranking of World Universities (ARWU), publishes its list of the world's top 500 universities, and within hours the headlines write themselves. Another German university has dropped off the list. Another American state flagship has vanished. Japan is in free fall. Editorials demand reform, ministers promise reviews, and the story hardens into conventional wisdom: Western universities are in decline.

There is just one problem with this story. When you look at what actually happened to the universities that fell off the list, most of them were not getting worse. Many were getting better.

A new working paper, prepared for submission to the Journal of Higher Education Policy and Management, took the trouble to check. Instead of reading the ranking counts and jumping to conclusions, the authors asked a more careful question: through what mechanism, exactly, did universities in Europe and the US lose their top-500 positions between 2012 and 2025? The answer changes how you should read every ranking headline you will see for the next decade.

A game of musical chairs, played with 500 seats

Start with the one fact from which everything else follows. In 2003, the first year of the ARWU, mainland China had 9 universities in the top 500 (Cao & Yang, 2023). In 2025, it had 101 (ShanghaiRanking Consultancy, 2025), a jump also flagged in international monitoring of Chinese higher education (British Council, 2025). That is not a typo. In two decades, China went from a marginal presence to more than a fifth of the entire table, with the steepest acceleration coming after 2015, the year China announced its Double First-Class excellence initiative (Cao & Yang, 2023).

Now consider what kind of instrument the ARWU top 500 actually is. It is a fixed-size table. There are exactly 500 seats, no more, no fewer. Every time a new university enters, another one must leave. Not because it failed. Because the chair it was sitting on was needed.

The ranking's own architects have described its construction and its limits: institutions are scored on six indicators of research output and prestige, including Nobel and Fields awards among alumni and staff, highly cited researchers, papers in Nature and Science, indexed publications, and per-capita performance (Liu & Cheng, 2005). Because the published table has a fixed size, it is what economists call a zero-sum instrument, and it has a strange and underappreciated property: a country's count of ranked universities can fall even when every single one of its universities improves, as long as other universities improve faster or enter the table for the first time (Liu & Cheng, 2005). Roughly 90 new Chinese entries since 2003 mathematically required roughly 90 exits by universities somewhere else. Those exits landed, overwhelmingly, in Europe, the US, and Japan.

The question is whether those exiting universities were pushed, or whether they fell. And that question, it turns out, can be answered with data.

Two stories, one test

The researchers set the problem up as a clean contest between two hypotheses.

Under the decline hypothesis, the one implied by nearly every headline, universities that dropped out of the top 500 should show deteriorating scores in the years before their exit: fewer highly cited researchers, fewer papers in Nature and Science, shrinking publication output.

Under the displacement hypothesis, exiting universities should show stable or even improving scores, and their exit should coincide with new entrants arriving at the same rank bands. They did not fall down the table; the table's entry price went up.

To run this test, the team assembled a harmonized dataset of every ARWU edition from 2003 to 2025, using a two-layer collection strategy (intercepting the raw data feeds behind the ranking's web tables, then cross-checking against the rendered pages) and tracking each institution's scores on all six ARWU indicators (Liu & Cheng, 2005). Because ARWU publishes exact ranks only for the top 100 and broad bands below that, the analysis uses rank-band transitions: for every pair of consecutive years, each university is classified as stable, moving up, moving down, entering, or exiting, and every exit is attributed to either score deterioration or threshold displacement.

The verdict is not close. The universities in Europe, the US, and Japan that fell out of the top 500 did not, as a group, show declining absolute scores before exit. Their publication and citation indicators were flat or gently rising. What changed was the score required to hold rank 500, which climbed steadily, year after year, as new entrants accumulated. The pattern matches displacement, not decline.

Why Germany lost and Britain didn't

Here is where the story gets genuinely interesting, because the displacement wave did not hit everyone equally, and the pattern of who got hit reveals the mechanism with unusual clarity.

In 2012, the UK and Germany had almost identical counts in the ARWU top 500: 38 and 37 universities respectively. By 2025, the UK had 37 (ShanghaiRanking Consultancy, 2025). Germany had approximately 23, a preliminary figure with intermediate years still being verified against archived editions. Nearly identical starting points, radically different outcomes. Did British universities outperform German ones by that margin? No. They were simply sitting in different parts of the room.

Chinese entrants arrive at the bottom of the table, in the 301-500 bands, and expel whoever sits at the margin. A country with a dozen universities parked around rank 350-480 offers a wide target. A country whose universities sit at ranks 5, 20, and 60 offers none. The UK cohort sits high in the table; the German cohort sat low. That positional difference, not any difference in research performance, determined their fates. Japan, with a similarly bottom-heavy profile, lost roughly 43 percent of its ranked universities. The US lost the most in absolute terms, around 40 slots, drawn heavily from public flagships outside the elite.

But why was Germany's cohort sitting low in the first place? The paper identifies three structural reasons, and none of them is a performance failure.

First, egalitarian system design. For most of the postwar period, German policy deliberately spread funding broadly across comprehensive universities rather than concentrating it in a few flagships. The result is a system with many good universities and few spectacular ones. The ARWU, which rewards the concentration of elite indicators in single institutions (Liu & Cheng, 2005), is close to a worst-case measuring device for such a system.

Second, and this is the detail most rankings coverage misses entirely, much of Germany's best research does not happen in universities at all. It happens in the Max Planck, Helmholtz, Fraunhofer, and Leibniz institutes, which never appear in the ARWU because they are not universities. Nobel Prizes won there, highly cited researchers employed there, Nature and Science papers published there: none of it counts toward any German university's score. In fairness, the same blindness applies to the Chinese Academy of Sciences on the other side of the ledger.

Third, indicator profile. German mid-tier universities earned their ARWU points disproportionately on publication volume, which is precisely the indicator where Chinese universities gained first and fastest (Zhang, Patton, & Kenney, 2013). Sitting near the rank-500 threshold with a publication-heavy profile made them the most exposed institutions in the entire table.

The wave was funded, deliberately, for decades

None of this happened by accident. The displacement wave was built, and the construction schedule is a matter of public record.

China ran three successive national excellence programs (Cao & Yang, 2023). Project 211, launched in 1995, financed research infrastructure at over a hundred institutions. Project 985, launched in 1999, concentrated resources on just 39 flagship universities, with Peking and Tsinghua each receiving 1.8 billion RMB in the first phase alone (Zhang, Patton, & Kenney, 2013) and the elite C9 League absorbing about half of all funding in every phase (Cao & Yang, 2023). Both were replaced in 2016 by the Double First-Class Initiative, under which cumulative central investment reached roughly 167 billion RMB, about US$24 billion, by 2024 (Times Higher Education, 2024).

Europe was not idle. The EU Framework Programmes grew from roughly US$15 billion under FP4 in the mid-1990s to over US$100 billion under Horizon Europe (European Commission, 2021, 2024), with universities and research organisations taking about two thirds of Horizon 2020 funds (European Commission, 2016), and a proposed near-doubling to roughly US$190 billion under FP10 for 2028-2034, currently under negotiation (European Commission, 2025).

An honest caveat: these figures are not directly comparable, and neither program type constitutes the majority of university funding, which still comes from stable core budgets. But that is exactly why targeted programs punch above their weight. They are the marginal, competitive, strategically directed money that decides which laboratories get built and which fields expand. And the timing tells the mechanical story: what maps onto China's entries into the top 500 is not the recent spending peak but Project 985's concentration on 39 flagships from 1999 onward, whose output crossed the ranking threshold with a lag of roughly a decade (Zhang, Patton, & Kenney, 2013).

What this means for the rest of us

Rankings shape policy. Governments respond to ranking movements with funding decisions and reform programs, a dynamic documented at length by Hazelkorn (2015). Which is why misreading them is expensive.

Three lessons follow from this study. First, ranking counts are not performance indicators for systems whose universities cluster near the threshold. Reading them that way invites misdiagnosis, and misdiagnosis invites the wrong reforms (Hazelkorn, 2015). A German-style system can maintain or improve output while shedding ranked universities every single year.

Second, exposure to displacement is a design feature, not a performance grade. Egalitarian systems, and systems that house elite research outside universities, will always look worse in institution-level zero-sum rankings than concentrated systems of equal or even lower total output. The ranking measures the shape of the system as much as its quality (Liu & Cheng, 2005; Hazelkorn, 2015).

Third, if a country actually wants ranking visibility, the Chinese experience shows how it is produced: sustained, concentrated, marginal investment in a defined set of institutions, with results arriving about a decade later (Zhang, Patton, & Kenney, 2013; Cao & Yang, 2023). That is a legitimate policy choice. But it is a choice, and it should be made deliberately, not stumbled into by panicking over headlines.

So the next time you read that your country's universities are "falling behind" because a few of them slipped off a list, remember the musical chairs. The music sped up, the room got more crowded, and someone spent twenty years and tens of billions of dollars buying new chairs for other players (Times Higher Education, 2024). That is a real and consequential story about the shifting geography of global science. But it is not a story about your universities getting worse. The data says most of them didn't.

References

British Council. (2025, August 22). More Chinese institutions rank high globally. https://opportunities-insight.britishcouncil.org/short-articles/news/more-chinese-institutions-rank-high-globally

Cao, Y., & Yang, R. (2023). China's academic excellence initiatives. International Higher Education, 115, 12-14.

European Commission. (2016). Horizon 2020: Research and innovation funding, making a real difference. Publications Office of the European Union.

European Commission. (2021). How Horizon Europe was developed. https://research-and-innovation.ec.europa.eu/funding/funding-opportunities/funding-programmes-and-open-calls/horizon-europe/how-horizon-europe-was-developed_en

European Commission. (2024). Ex-post evaluation of Horizon 2020, the EU Framework Programme for Research and Innovation (COM(2024) 49 final). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52024DC0049

European Commission. (2025). Proposal for a regulation of the European Parliament and of the Council establishing Horizon Europe, the framework programme for research and innovation, for the period 2028-2034 (COM(2025) 543 final). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52025PC0543

Hazelkorn, E. (2015). Rankings and the reshaping of higher education: The battle for world-class excellence (2nd ed.). Palgrave Macmillan.

Liu, N. C., & Cheng, Y. (2005). The Academic Ranking of World Universities. Higher Education in Europe, 30(2), 127-136. https://doi.org/10.1080/03797720500260116

ShanghaiRanking Consultancy. (2025). 2025 Academic Ranking of World Universities. https://www.shanghairanking.com/rankings/arwu/2025

Times Higher Education. (2024, November 8). China sets new priorities for Double First-Class universities. https://www.timeshighereducation.com/news/china-sets-new-priorities-double-first-class-universities

Zhang, H., Patton, D., & Kenney, M. (2013). Building global-class universities: Assessing the impact of the 985 Project. Research Policy, 42(3), 765-775. https://doi.org/10.1016/j.respol.2012.10.003

Learn more:

Life is a Journey of Learning

24/08/2026

Your Universities Didn't Get Worse. The Scoreboard Ran Out of Seats