Imperial China’s eighteenth-century quota: how an affirmative-action experiment reshaped the exam empire
Home > Imperial China’s eighteenth-century quota: how an affirmative-action experiment reshaped the exam empire
In this blog post, Melanie Meng Xue of the London School of Economics & Boxiao Zhang of the Renmin University Of China introduce their new working paper, “The Short- and Long-Run Effects of Affirmative Action: Evidence from Imperial China,” which is based on a paper previously presented at an Economic History Society Annual Conference.
—
In 1712 the Qing emperor Kangxi confronted a political headache that would sound familiar to any admissions officer at a modern university. Candidates from the prosperous lower Yangzi dominated the all‑important civil‑service examinations, while scholars from poorer or frontier provinces seldom secured the coveted jinshi degree that opened the door to officialdom. To redress the imbalance, the throne rewrote a single rule: at the final, metropolitan stage of the triennial exam, the pass rate was fixed for every province in proportion to the number of locals who reached Beijing. The prosperous lower Yangzi found itself capped while poorer or frontier provinces suddenly had space to succeed.
In our paper, we exploit an extraordinary database of more than 16,000 successful candidates between 1650 and 1840. We observe exam year, birthplace, and final exam rank for each successful candidate. We supplement the dataset with information on bureaucratic appointments and administrative ranks for a subset of successful candidates over their careers. To investigate the reform’s long‑run spillovers, we collect data from Republican China (1912–1949) documenting local elites—university graduates, educators, and public figures. To measure human‑capital outcomes in the late twentieth century, we use the Chinese census of 1982 to calculate, for each prefecture, the share of individuals completing primary, secondary, and tertiary schooling and the average social and economic status of individuals’ occupations.
More winners, same standards
Provinces that gained slots saw a clear rise in jinshi per capita. Because earlier county and provincial exams were untouched, the pipeline of effort upstream did not dry up. Crucially, the new jinshi performed just as well once they entered government. They began careers at slightly higher ranks, were as likely to secure central or revenue‑rich posts, and were no more associated with local unrest than their peers. In short, the quotas uncovered under‑represented talent without diluting quality.
A new geography of inequality
As the reform decreased inequality between provinces, it produced higher inequality within them. Districts with stronger pre‑existing human‑capital stocks captured most of the extra places, while those with weaker stocks remained marginal. Our inequality indices fall across the empire as a whole but rise sharply inside the “winning” provinces. Redistribution, in other words, led to a new geography of inequality.
After abolition: a fading but durable legacy
The keju system was scrapped in 1905 and its provincial balancing rule vanished with it. In the early Republic the once‑favoured provinces began to supply a smaller share of the country’s emerging élite—whether measured by entrants to the new national universities, students sent overseas, or the biographical “who’s‑who” directories of the day—a sign that the direct boost had disappeared.
Yet the earlier learning push still left a mark. In the census of 1982, prefectures from provinces that received a larger 1712 shock record significantly higher secondary‑ and tertiary‑completion rates and higher average ISEI/SIOPS occupational‑prestige scores, even after controlling for pre‑1650 human‑capital stocks, population density, geography and other historical fundamentals. A rule enforced for two centuries was still shaping education and work almost eighty years after its repeal.
Lessons beyond the Qing
The eighteenth‑century experiment underlines that where a selection process has multiple filters, moving the last one can broaden representation without blunting earlier incentives. It also reminds us that affirmative action rarely eliminates inequality; it redistributes it and may even sharpen local hierarchies. Finally, temporary interventions can leave long shadows—but as the Republican‑era reversal shows, their benefits are not self‑sustaining. For policymakers still debating quotas three hundred years later, the Qing case provides both encouragement and caution in equal measure.
To contact the authors:
