Excellence Before Office
The scholar-leader model and the long history of the university
In the 630s a Chinese monk named Xuanzang walked through the gate of Nalanda, in the Gangetic plain of what is now Bihar, at the end of a journey that had taken him across the Taklamakan and over the Hindu Kush. He had come for one teacher. Śīlabhadra, the abbot, was by then a very old man and the most authoritative living exponent of the Yogācāra school of Buddhist philosophy, and Xuanzang wanted to sit at his feet before he died. The monastery around them held thousands of students, a library that later pilgrims described with awe, and a daily programme of lectures and public debate in which a scholar's standing rose or fell on argument. Kings supplied the land and the revenue. The learned supplied the authority.
Xuanzang's own account, written after his return to the Tang court, records who ran the place and why. Rank at Nalanda followed mastery. Gatekeeper scholars examined new arrivals on doctrine before admitting them; the senior monks who governed the institution were those who had won their reputations in disputation; and students from China, Korea, Tibet and Central Asia travelled for years precisely because the teachers there were better than the teachers at home. Prestige attached to the scholars, and the institution borrowed it from them. That arrangement is the scholar-leader model in its oldest form, arrived at independently, in several civilisations, centuries before anyone in Europe drafted a university charter.
Before Bologna
Nalanda was itself a late flowering. Indian tradition places Taxila, on the road between the Indus and the Kabul valley, as a centre of advanced learning from at least the middle of the first millennium BCE, and associates it with the grammarian Pāṇini and the political theorist Kauṭilya. Taxila had no charter, no single campus and no corporate identity. It was a cluster of masters, each attracting pupils on the strength of his name, teaching the Vedas, grammar, logic, medicine, astronomy and statecraft. A student chose a teacher, and the teacher's eminence was the qualification. Later Buddhist foundations, Vikramaśīla and Odantapurī among them, followed Nalanda's pattern of scholar-monks governing scholar-monks until the Turkic raids of the late twelfth and early thirteenth centuries burned them out.
China built the same principle into the state. In 124 BCE the Han emperor Wu established the Taixue, an imperial academy for the Confucian classics that trained the men who would staff his bureaucracy. Its head carried the title of jijiu, the libationer, and the office went to recognised classical scholars. Over the following dynasties the Taixue gave way to the Guozijian, which combined a teaching academy with a directorate overseeing education across the empire, and emperors occasionally lectured there themselves, the point being that the person who governed learning should himself be learned. The system lasted, with interruptions and reforms, into the early twentieth century, longer than any European university has yet existed.
The Islamic world produced the institutions that most resemble the later European form. Al-Qarawiyyin in Fez, founded in 859 by Fatima al-Fihri, grew from a mosque into a centre teaching law, theology, grammar, logic, mathematics and medicine, and is widely counted as the oldest continuously operating institution of higher learning. Al-Azhar in Cairo followed a century later. Both worked through the halaqa, the circle of students around a master, and through the ijaza, a personal licence to teach that a master granted to a pupil who had proved his command of a text. The ijaza is the telling document. It located the right to teach in demonstrated mastery, certified scholar to scholar, and the men who came to head these institutions, the shaykhs of al-Azhar above all, were those who held the most ijazas from the most respected hands.
Three traditions, three continents, one rule. Whoever governed the house of learning had first to be learned. Rulers paid the bills and sometimes appointed the heads, but the legitimacy that made a student walk from Chang'an to Bihar came from the scholars, and it was the scholars who decided what counted as knowing.
The European inheritance
What Europe added, from the late eleventh century, was the corporation. At Bologna the students, and at Paris the masters, organised themselves into a universitas, a legally recognised body that could hold property, sue and be sued, set its own examinations and elect its own officers. The rector of Paris was chosen from among the masters of arts. Oxford and Cambridge, founded on the Paris pattern, vested their government in the masters and later in the fellows of the colleges. The corporate form gave the scholar-leader principle a legal skeleton it had never had in Nalanda or Fez, and that skeleton is what the modern world inherited: degrees, faculties, self-government, and an assumption that the person at the top had come up through the ranks of teaching.
The research university of the nineteenth century strengthened the assumption. Wilhelm von Humboldt's Berlin of 1810 fused teaching with original inquiry and made the professor, rather than the priest or the administrator, the central figure of the institution. When Daniel Coit Gilman founded Johns Hopkins in 1876 on the German model, he recruited leading scholars first and built the university around them. Harvard's James Bryant Conant, a chemist of real distinction, ran the university from 1933 to 1953 and remains the standard American example of a scientist who modernised an institution without ceasing to be a scientist. Through the twentieth century the administrative apparatus of universities grew enormously, with professional staff in finance, estates, student services and marketing. At the research-intensive institutions, though, the top jobs stayed with people who had earned tenure and a reputation, and the faculty senates and governance traditions that grew up alongside them made it costly to appoint anyone else.
What the evidence says
Amanda Goodall put the old rule to a modern test. She took the hundred universities at the top of the Shanghai ranking, looked up the lifetime citation record of each one's president or vice-chancellor, and found a monotonic relationship: the higher the university, the more cited its leader, with the leaders of the top twenty standing well clear of the rest. Correlation of that kind invites the obvious reply that good universities can afford good scholars. So she then followed British universities over time and found that those which appointed more highly cited leaders improved their subsequent performance in the Research Assessment Exercise, an effect that survived controls for the size of the institution and other factors. Her interviews with the leaders themselves suggested why. A scholar-president knows what first-rate work looks like, can tell it from busywork, is trusted by the people who do it, and can recruit and keep them.
The most recent rankings show the pattern holding at the very top. The 2026 Shanghai ranking places Harvard first, then Stanford, MIT, Cambridge, Berkeley, Princeton and Oxford; the QS ranking for the same year puts MIT first, with Imperial, Stanford, Oxford and Harvard behind it. The people running these institutions in September 2026 read like a faculty roll. Harvard's Alan Garber is a physician and health economist who spent twelve years as the university's provost before taking the presidency. Stanford's Jonathan Levin is an economist who won the John Bates Clark Medal and ran the Graduate School of Business before moving to the president's office. MIT's Sally Kornbluth is a cell biologist who was provost at Duke. Oxford's vice-chancellor, Irene Tracey, is a neuroscientist and Fellow of the Royal Society who directed the university's brain-imaging centre for a decade before becoming Warden of Merton College. Cambridge's Deborah Prentice is a psychologist who was provost at Princeton. Every one of them built a scholarly reputation first and was handed a large institution second.
Rankings reward research, and universities led by researchers will naturally do well on measures of research. That objection is fair, and it is also the point. The rankings are proxies for the thing the university exists to do, and the institutions that do it best are led by people who have done it themselves.
The model as practice: Purdue
The scholar-leader model is a pipeline as much as a principle, and Purdue University has built one of the more deliberate pipelines in the United States. The Big Ten Academic Alliance's Academic Leadership Program selects faculty with strong records and evident administrative promise for a year of campus seminars and multi-university workshops on running a research university. The Purdue Insights Forum prepares people for the specific jobs of associate head, department head, associate dean and centre director. Provost Fellows take on real projects in the provost's office, and the Susan Bulkeley Butler Center for Leadership Excellence supplies coaching and mentoring. None of these programmes teaches anyone to do research. They assume the research is done, and they teach budgeting, personnel, strategy and shared governance to people who already know what a laboratory or an archive is for.
The results show in Purdue's senior offices. Patrick Wolfe, a statistician, moved from the deanship of the College of Science to the provost's office. Mung Chiang, an electrical engineer who had won the Waterman Award and a Guggenheim Fellowship, came to Purdue from Princeton, ran the College of Engineering, and served as president from 2023 until Northwestern recruited him to its own presidency in July 2026. Purdue has not been shy about the outcome. In October 2025 a Canadian consultancy, Organizational Excellence Specialists, named it the best-managed university in the world, ahead of Cambridge, the National University of Singapore, ETH Zurich and Stanford, citing a decade of frozen tuition, financial stability, and high satisfaction among students and faculty.
Purdue also supplies the strongest objection to the model, and it deserves to be stated in full. The tuition freeze and the management reforms that the consultancy praised were the work of Mitch Daniels, a former governor of Indiana and federal budget director with no academic career at all, who was president from 2013 to 2022 and returned as interim president in July 2026 when Chiang left. If an outsider can produce the best-managed university in the world, the scholar-leader model looks less like a law and more like a preference. The answer is that Daniels ran Purdue with a scholar-provost and scholar-deans below him, that his own trustees chose an engineer to succeed him, and that the leadership pipeline was built precisely so that the next generation of heads would be academics with management training rather than managers with none. The hybrid case shows the model absorbing a strong outside leader while keeping the academic core in academic hands.
Why it works
The reasons are the ones Xuanzang would have recognised. A leader who has done first-rate research knows the difference between a finding and a press release, between a promising young scholar and a fluent one, between a department that is quiet because it is thinking and one that is quiet because it has stopped. That knowledge cannot be delegated. It shapes who gets hired, who gets promoted, which programmes get money and which get closed, and every one of those decisions is a judgement about excellence that a person without the relevant experience must take on trust from somebody else. Faculty know this, which is why they extend legitimacy to scholar-leaders and withhold it from others, and legitimacy is the currency in which shared governance is paid.
The model has costs, and honest institutions name them. Good scholars give up their research when they take these jobs, and some of them are poor managers, and some of them mistake their own discipline's habits for universal truths. Finance, estates, litigation and politics are real skills, and a chemist does not acquire them by being a chemist. The answer the best institutions have reached is selection plus training: choose the academics with evident judgement, then invest in them as seriously as the institution invests in a new laboratory. Purdue's programmes, the American Council on Education's fellowships, the Big Ten's academy and their equivalents in Britain, Europe and Asia exist for exactly this purpose. They take excellence as given and build capacity on top of it.
Xuanzang crossed a continent for a teacher. The universities that still draw students across continents are led, with few exceptions, by people who could have been that teacher. Two thousand years of practice, from Taxila to Cambridge, Massachusetts, point to the same conclusion: the institution whose business is knowing should be led by those who know, and who have proved it.