Von Glahn (2019) identifies four modalities of the fiscal state across Chinese imperial history. Each corresponds to a distinct relationship between the state, agrarian production, commercial activity, and revenue extraction.
| Modality | Period | Revenue base | Key features |
|---|---|---|---|
| Militarist-physiocratic | Qin/Han (221 BCE–220 CE) | Conscript labor, uniform agrarian tax | Universal civil registration, household surveys, labor service rosters, submitted accounts relayed to center |
| Mercantilist | Emperor Wu of Han; Wang Anshi (1069–85) | State monopolies, centralized commercial revenue | Radical intervention: salt/iron monopolies (119 BCE), state loans, price stabilization, cadastral surveys |
| Synergistic | Tang-Song transition (c. 780–1279) | Indirect commercial taxation, commodity monopolies, negotiable instruments | Finance Commission, 1,993 commercial tax stations, progressive household ranking, salt licenses as bills of exchange |
| Providential | Ming-Qing (1368–1911) | Frozen agrarian quotas, minimal commercial tax | Neo-Confucian commitment to light taxation, dismantled legibility, land tax ~70%+ of revenue |
These categories are best understood as a periodization tool for organizing the fiscal regime's evolution over two millennia. The transition from the synergistic to the providential state is the central event: a shift from an apparatus that harnessed commercial growth through indirect taxation and negotiable financial instruments to one that deliberately retreated from commercial extraction in favor of frozen agrarian quotas.
The roots of Chinese state capacity extend well before the Qin unification. The Western Zhou (1046–771 BCE) established a proto-bureaucracy with parallel administrative divisions (Grand Secretariat, Six Armies, Royal Household), initially hereditary but gradually incorporating meritocratic promotion. During the Spring and Autumn period (771–476 BCE), competing states began experimenting with legal codes and centralized administration. The Warring States era (476–221 BCE) produced a wave of institutional innovation under competitive pressure: Li Kui's legal and agricultural reforms in Wei (c. 445 BCE), Wu Qi's military reforms in Chu (381 BCE), Shen Buhai's bureaucratic techniques in Han, and most consequentially Shang Yang's comprehensive reforms in Qin (356–338 BCE). The Qin state itself originated as horse breeders for the Zhou royal house: Feizi was awarded a fief in the Qin valley for his skill with horses, and the Ying clan maintained close ties to steppe peoples on the western frontier, later adopting cavalry tactics (after Zhao pioneered them in 307 BCE). Shang Yang, originally a minor official from Wei, built on earlier innovations to create the first fully centralized fiscal-military state: universal household registration, a twenty-rank meritocratic system tied to military achievement, abolition of hereditary aristocratic privilege, county-level administration, and standardized taxation. This apparatus provided the template the Qin unification imposed empire-wide in 221 BCE, and which the Han inherited and refined. A critical enabling technology was the early development and subsequent Qin standardization of a unified script, which made possible empire-wide bureaucratic correspondence, population registries, and the relay of submitted accounts from county to center (von Glahn 2016, pp. 87ff; Goody 1986, pp. 103–04).
Michael Mann's distinction between infrastructural power (the capacity of the state to penetrate society and implement decisions) and despotic power (the range of actions elites can undertake without routine negotiation) provides a useful lens for reading the data. The Song synergistic state possessed high infrastructural power: universal civil registration, annual household surveys, detailed records of landholdings, a Finance Commission with specialized career tracks, and nearly 4,000 commercial tax and revenue collection points. The Qing retained substantial despotic power (the emperor could decree policy changes at will) but had lost the infrastructural capacity to identify its population, assess landholdings, or mobilize new revenue streams. By the late eighteenth century, the quinquennial household surveys had ceased, no cadastral survey had been conducted since 1580, and the state could not even enumerate its subjects. Von Glahn's summary is direct: the legibility of society vital to the operation of the fiscal state had vanished.
| Infrastructure | W. Zhou (1046–771 BCE) | Warring States (476–221 BCE) | Qin/Han (221 BCE–220 CE) | Sui/Tang (581–907) | Song (960–1279) | Yuan (1279–1368) | Ming (1368–1644) | Qing (1644–1911) |
|---|---|---|---|---|---|---|---|---|
| Household registration (编户齐民) | ||||||||
| Cadastral survey (land records) | ||||||||
| Population census | ||||||||
| Meritocratic appointment | ||||||||
| Keju examination system | ||||||||
| Codified legal system | ||||||||
| Specialized fiscal bureaucracy | ||||||||
| Commercial tax network | ||||||||
| Salt/commodity monopoly admin. | ||||||||
| Military registration / conscription | ||||||||
| Standardized weights / measures / currency | ||||||||
| Paper money / financial instruments |
The Northern Song (960–1127) built a fiscal apparatus that was, by premodern standards, exceptionally sophisticated. Von Glahn (2016) documents its core features: a Finance Commission (Sansi 三司) reporting directly to the emperor, comprising departments of revenue, salt, and census, which institutionalized functional specialization and encouraged the development of professional expertise and distinct career tracks in fiscal administration. By 1077, a network of 1,993 commercial tax stations and 1,861 liquor revenue depots were in operation (von Glahn 2016, p. 296, citing Guo Zhengzhong 1990). This density of fiscal infrastructure would not be matched by any European state for centuries.
The salt monopoly evolved into a proto-financial system: after 1048, salt licenses were decoupled from frontier delivery, could be redeemed for money, and functioned as negotiable bills of exchange, spurring a secondary market in Kaifeng (von Glahn 2016, p. 298). Total state revenue, measured in silver equivalents, rose from approximately 1,319,000 kg in 997 to 3,190,000 kg in 1086 (von Glahn 2016, Table 6.7, via Hartwell 1988 and Peng Xinwei 1965). The monetization of revenue collection increased from 48% in coin in 997 to over 80% by the 1070s.
Military expenditures consumed over 80% of central government revenue, supporting 1.4 million soldiers at peak (von Glahn 2016, p. 293). This fiscal-military pressure drove innovation: the Wang Anshi reforms (1069–85) attempted to expand state revenue through rural credit (Green Sprout loans), price stabilization, labor service commutation, cadastral surveys, and state trade. The reforms were partially successful before political opposition led to their reversal.
The ideological foundation for the Ming fiscal reversal was laid during the Southern Song, in direct response to the Mongol threat. In 1237, the Mongols reinstituted a civil service examination in their northern territories, claiming legitimacy as a Confucian state. The Southern Song court responded by making Zhu Xi's commentaries official state orthodoxy. The Yuan dynasty then cemented this further: it was the Yuan court that first officially adopted the Four Books as the basis of the civil service examination in 1313, a practice observed until 1905. The Mongols thus transmitted to the Ming precisely the ideology that would undermine the activist fiscal state the Song had built.
Meanwhile, the Mongols did little to maintain the Song's fiscal infrastructure. They were largely content to maintain the existing taxation system in the south and interfered little in the private economy, relying on tax-farming through great landowners (von Glahn 2016, pp. 334–35). The commercial tax network, the Finance Commission, the salt license financial system: all atrophied during ninety years of Mongol rule.
Zhu Yuanzhang (Emperor Hongwu, r. 1368–98) set out to repudiate everything he associated with the Mongols, but the lesson he drew was catastrophic. In von Glahn's taxonomy, Hongwu skipped back over a thousand years of fiscal evolution to revive the militarist-physiocratic model of the Qin and Han: in-kind taxation, conscript labor, self-sufficient military farms, and tight state control. Von Glahn (2016, p. 337): Hongwu repudiated the Mongol heritage along with the market economy that had developed during the Tang-Song transition, marking a sharp rupture in the evolution of economic life, especially in Jiangnan.
| Date | Reform | Significance |
|---|---|---|
| c. 445 BCE | Li Kui reforms (Wei 魏) | First systematic legal code (Fajing 法经); agricultural incentives to maximize land use; Wei becomes strongest Warring States power |
| 381 BCE | Wu Qi reforms (Chu) | Curtailed aristocratic privilege; centralized military command; Wu Qi murdered upon King Dao's death, reforms partly reversed |
| 356 BCE | Shang Yang reforms (Qin) | Universal household registration, twenty-rank merit system tied to military achievement, abolition of hereditary privilege, county administration, standardized taxation. Shang Yang was a Wei defector who built on Li Kui's model. Foundation of the Qin fiscal-military state |
| 307 BCE | Zhao cavalry reform (胡服骑射) | King Wuling of Zhao adopted nomadic cavalry tactics and dress; first Chinese state to field cavalry. Qin adopted cavalry shortly after |
| 221 BCE | Qin unification | Standardized weights, measures, currency, and script across unified empire |
| 119 BCE | Han salt & iron monopoly | Sang Hongyang's state enterprises: centralized commercial revenue under Emperor Wu (mercantilist modality) |
| 485 CE | Equal-field system (N. Wei → Tang) | Land-allocation tied to tax obligation; universal registration as basis for extraction |
| 780 | Two-Tax Reform (Yang Yan) | Shift from labor service to monetary taxation; biannual assessment by wealth and landholding; beginning of Tang-Song fiscal transition |
| 960–1077 | Song fiscal expansion | Tea/salt/wine monopolies, commercial tax network (1,993 stations), Finance Commission, progressive household ranking |
| 1069–85 | Wang Anshi New Policies | State loans (Green Sprout), hired service law, price stabilization, cadastral survey; mercantilist deviation within synergistic framework |
| 1393 | Ming tax quota freeze | Lijia quotas fixed permanently; beginning of providential state's fiscal regime |
| 1581 | Single Whip Reform (Zhang Juzheng) | Consolidation of multiple levies into single silver payment; partial re-monetization, but no expansion of tax base |
| 1712 | Kangxi tax freeze (永不加赋) | Pledge to never increase the head tax; formalized elite fiscal preferences as permanent policy |
| 1723–35 | 摊丁入亩 (Yongzheng) | Merge poll tax into land tax; simplified administration but further narrowed the revenue base to land alone |
| 1853 | Lijin transit tax | Decentralized extraction during Taiping emergency; first new revenue stream in centuries, but collected by provincial authorities rather than the center |
| 1901–11 | Late Qing New Policies | Attempted fiscal modernization (modern customs, new commercial taxes, educational levies), but too late and too weak to reverse the structural deficit |
What emerged from the Ming reversal was a coherent regime with its own logic: what von Glahn (2019) calls the providential state. Neo-Confucian ideological commitments to popular welfare guided fiscal policy. The state should tax lightly, avoid interference in markets, and provide basic public goods (flood control, famine relief, granaries). Private commerce was left largely alone; merchant organizations assumed self-governance through guilds and brokerages. This was a favorable environment for Smithian growth (market expansion, regional specialization, division of labor). It foreclosed Schumpeterian growth (state-driven investment, mercantilist strategy, demand creation through military expenditure).
Von Glahn (2016, p. 15) frames the broader divergence through this distinction: during the late imperial era, China's rulers embraced the Neo-Confucian commitment to light taxation and minimal state intrusion. Although this had positive effects in encouraging Smithian dynamics, the weak infrastructural capacity of the state limited the potential for Schumpeterian growth of the kind happening concurrently in early modern Europe. European centralization of political power curtailed seigneurial and urban monopolies that had hindered market integration, while war-making exerted crucial influence on state formation and long-term economic development.
Wang's (2022) data shows China's tax/GDP ratio peaking at 17% in 997 (Northern Song), then declining through the Ming and Qing to reach 0.78% by 1849. England's ratio was comparable to China's in the early 16th century (~2–4%), then surged after the Glorious Revolution to 19% by 1806. Perkins (1967) independently estimates Qing central government revenues at the end of the nineteenth century at less than 100 million taels, between 1 and 2 percent of gross national product. In his assessment, this made the Qing government a financially weak instrument for the task of modernization, and the contrast with Meiji Japan was striking.
Karaman and Pamuk (2013) add an important dimension to the European side of the comparison. Their central finding is that the determinants of European fiscal capacity worked in interaction: warfare, economic structure (as proxied by urbanization), and political regime type were jointly significant. Representative regimes in more urbanized-commercial economies and authoritarian regimes in more rural-agrarian economies tended to translate war pressure into fiscal capacity gains. The Song fiscal state, which combined commercial sophistication with centralized military-fiscal pressure, was structurally analogous to the early modern European pattern. The Ming-Qing providential state, which combined minimal commercial extraction with the absence of sustained interstate military competition, was structurally analogous to the configuration that K&P find least conducive to fiscal state-building.
A caveat on the European evidence is warranted. Ogilvie (2023) observes that the association between fiscal capacity and economic growth is more ambiguous than the state capacity literature implies. England did not develop high fiscal capacity until nearly 1800, well after the onset of rapid economic growth and industrialization. High-capacity Prussia remained economically stagnant and industrialized late. The chart above shows the same pattern: the fiscal divergence between China and Europe widened in the seventeenth and eighteenth centuries, but the economic divergence it is taken to explain was driven by a few northwestern European states, not Europe as a whole. The Ottoman Empire, with fiscal capacity comparable to Qing China's throughout the period, remained closer to the Chinese pattern than the English one.
The formal bureaucracy shrank in per capita terms across Chinese imperial history even as the keju system that staffed it grew more sophisticated. The Han empire maintained roughly 130,000 ranked officials and clerks for ~58 million people. By the Qing, the number of formally appointed officials had fallen to ~20,000 for a population that reached 400 million. The 1,500 county magistrates serving as the lowest tier of formal state authority remained essentially unchanged from the Ming through the end of the dynasty, even as county populations tripled.
| Period | Formal officials | Ratio | Informal staff (est.) | Ratio incl. informal | Pop. (M) |
|---|---|---|---|---|---|
| Han, c. 2 CE | 130,285 | 1 : 440 | — | — | ~58 |
| Tang, c. 750 | ~18,000 | 1 : 2,900 | ~350,000 clerks | ~1 : 140 | ~52 |
| N. Song, c. 1080 | ~24,000 | 1 : 4,600 | large sub-official corps | — | ~110 |
| Ming, c. 1400 | ~15,000 | 1 : 4,400 | ~50,000–100,000 | ~1 : 600–1,000 | ~65 |
| Qing, c. 1700 | ~20,000 | 1 : 7,500 | ~1,000,000+ | ~1 : 140 | ~150 |
| Qing, c. 1800 | ~20,000 | 1 : 15,000 | ~1,000,000+ | ~1 : 300 | ~300 |
| Qing, c. 1850 | ~20,000 | 1 : 21,500 | ~1,000,000+ | ~1 : 400 | ~430 |
| England, c. 1780 | ~16,000 excise alone | 1 : 560 | broader state apparatus | — | ~9 |
| France, c. 1780 | tens of thousands | ~1 : 400–800 | venal officeholders, intendants | — | ~28 |
The Han figure of 130,285 includes all ranked personnel down to the lowest clerks (Bielenstein 1980). The Qing 20,000 counts only formally appointed, keju-certified officials. If Qing sub-official staff (yamen clerks 胥吏, runners, constables, private secretaries 幕友) are included, the ratio narrows considerably. Reed (2000) and Ch'ü (1962) suggest this informal apparatus numbered over one million. At that count the Qing c. 1800 had roughly 1 state-associated person per 300 subjects, comparable to the Han's 1:440 formal ratio.
The critical difference was in what this apparatus could accomplish. The Han's 130,000 personnel maintained universal household registration, annual surveys, submitted accounts, and a cadastral system. The Qing's million-plus sub-officials, lacking formal training, rank, or central oversight, supported a system in which the last cadastral survey was from 1580, the last household enumeration from 1772, and the land tax was assessed on Ming-era quotas bearing no relationship to actual landholdings. Headcount was not the binding constraint. Administrative capability was.
The fundamental puzzle is why no subsequent dynasty rebuilt the Song's synergistic fiscal apparatus. The answer requires distinguishing between ideological, structural, and administrative dimensions of the problem, because by the mid-Qing all three had become self-reinforcing.
1. The providential consensus. Zhu Xi's Neo-Confucianism, entrenched as state orthodoxy since the Southern Song's anti-Mongol cultural mobilization and cemented by the Yuan's adoption of the Four Books examination in 1313, made fiscal activism ideologically illegitimate. Wang Anshi's reforms became a cautionary tale of state overreach rather than a model of fiscal innovation. Von Glahn (2019) stresses that the providential state expressed a positive commitment to light taxation, agrarian paternalism, and minimal commercial interference that its practitioners understood as good governance. The commitment was sincere, coherent, and self-reinforcing.
2. Elite capture. The Tang-Song transition replaced a national aristocracy with locally rooted gentry families whose sons competed through the Keju examination. These localized elites had strong incentives to block fiscal reform that would increase their tax burden. Wang (2022, Chs. 5–6) documents how gentry-connected officials systematically opposed the Wang Anshi reforms and delayed the Ming Single Whip reform for over a century. The Kangxi emperor formalized elite fiscal preferences in 1712 with his pledge to never increase taxes, which remained policy until the dynasty's end.
3. The destruction of legibility. The synergistic state depended on the state's capacity to identify, classify, and tax its population. The Qin and Han empires had achieved a level of legibility (to use James Scott's term) that von Glahn (2019) argues was unmatched in Europe until the early modern period: universal civil registration, annual household surveys, detailed records of landholdings, labor service rosters, and census data relayed to the center. The Song household ranking system (grading property-owning households into five rural and ten urban ranks as the basis for progressive taxation) was a sophisticated extension of this tradition. The Ming-Qing fiscal reforms systematically dismantled it. The 1729 merger of the ding labor service levy into the land tax shifted the object of taxation from households to land, rendering civil registration meaningless. The Qing never conducted a universal cadastral survey, relying instead on the Ming survey of 1580. The quinquennial household surveys ceased entirely in 1772. By the late eighteenth century, the Qing government could not identify its population, had no current knowledge of landholdings, and possessed no administrative mechanism to capture new revenue even if ideological constraints were relaxed.
4. Frozen quotas as equilibrium. The Ming's 1393 tax quotas, inherited and frozen permanently by the Qing in 1713, created a self-reinforcing equilibrium that locked all three dimensions together. As population and economic output grew while quotas stayed fixed, the per capita tax burden fell mechanically. Real per capita grain revenue in 1766 was only 30% of the 1435–49 level (von Glahn 2019, Table 9). Government spending probably dipped below 3% of GDP, less than a third of the Song synergistic state's level. Any attempt to raise quotas would have faced opposition from the gentry class, would have required cadastral and demographic knowledge the state no longer possessed, and would have violated the ideology that legitimated Qing rule. The system was stable precisely because all the forces that might have disrupted it had been neutralized.
Comparative note: Russia. The late Qing's situation has a structural parallel in Tsarist Russia, which Mann (1986) identifies as the paradigmatic case of despotic power without infrastructural power. Russia's conditions made a different outcome difficult to imagine: a vast, sparsely settled territory with only 3% of the Asian population living in towns, no tradition of parish-level administrative penetration, few urban centers relative to the landmass, and persistent military threats on multiple frontiers that absorbed whatever fiscal capacity existed. The geographic and demographic conditions that had enabled dense administrative networks in Song China or post-1688 England (compact territory, high urbanization, concentrated commercial wealth) simply did not obtain. By the early nineteenth century both empires occupied structurally similar positions: autocrats who could decree but not implement, ruling over territories too large and populations too dispersed for the administrative apparatus they possessed. The parallel reinforces the point that state capacity is not a policy choice but a product of cumulative institutional development within specific geographic, demographic, and technological constraints.