Why did sustained per-capita growth begin in northwestern Europe? Why not in the Yangtze delta, the Mughal heartland, or the Ottoman core, when all three were comparably wealthy and commercially sophisticated as late as 1750? This is the central question of modern economic history. Four major accounts compete: Kenneth Pomeranz's coal-and-colonies thesis, Robert Allen's high-wage economy, Joel Mokyr's culture of useful knowledge, and the institutionalist account of Acemoglu, Robinson, and North. This chapter walks each at full strength, then takes a position.
Named literature: Pomeranz (2000); Allen (2009, 2017); Mokyr (2002, 2009); Acemoglu, Johnson & Robinson (2001); North & Weingast (1989); Wong (1997); Broadberry et al. (2018); Bolt & van Zanden (2020); Pamuk (2018); Rosenthal & Wong (2011); Wrigley (1988).
The previous chapter ended with the early modern world an integrated commercial system but not yet a divergent one. By 1750 the world's major economic cores remained, on the measurable margins, broadly comparable. By 1850 they did not. This chapter walks the question of why.
Until the late 1990s, the standard textbook narrative assumed that Europe's economic lead was deep-rooted: that something about European institutions, culture, or geography had made northwestern Europe richer than Asia long before the Industrial Revolution. The California school (Pomeranz, Wong, Frank, and others) forced a correction. On the measurable margins that matter for explaining where sustained growth emerged, the leading economic cores of Eurasia were comparable well into the eighteenth century. The divergence is real. The assumption that it was prefigured is not.
The comparison needs the right unit. "Europe versus Asia" is too coarse—it averages medieval Scandinavia with Venetian commerce, and Tibetan plateaus with Yangtze delta trade networks. Pomeranz's methodological innovation was to compare the most advanced core region within each major civilization. For Europe: England and the Low Countries. For China: the Yangtze delta—Suzhou, Hangzhou, Nanjing, and the dense commercial network of Jiangnan. For India: the Mughal heartland around Delhi and the upper Indo-Gangetic plain. For the Ottoman Empire: the core around Istanbul, Izmir, and western Anatolia. These are comparable cores: the places where each Eurasian civilization concentrated its economic energy.
What does comparability look like in the data? Robert Allen's welfare-ratio series measures a worker's annual income as a multiple of a bare-subsistence consumption basket—food, fuel, shelter, and clothing at the cheapest locally available prices. A welfare ratio of 1.0 means the worker earns exactly enough to sustain a family at bare subsistence; 4.0 means four times that level. Allen's data places London workers at approximately 3.5 to 4.0 subsistence baskets across most of 1500–1700. Beijing sits at roughly 1.5, Delhi at about 1.1, and Istanbul between 1.5 and 2.0.
That looks like a gap. But the welfare-ratio method is designed to absorb cross-cultural variation: baskets are locally priced, caloric thresholds reflect local diets, and the comparison measures distance from the same floor. London's higher ratio partly reflects the higher cost of Northern European subsistence: more fuel, more animal protein, more clothing against cold. The substantive question is whether these differences indicate fundamentally different economic systems or local variants within a shared Eurasian range. The California school's answer is the latter. The strongest version of their claim is not "they were identical" but "neither core had broken away from the Malthusian productivity frontier that constrained all pre-modern agrarian economies."
Urbanization tells a similar story. The Yangtze delta was among the most urbanized regions on earth in 1500: roughly 10–15 percent of its population lived in cities of ten thousand or more, comparable to England and the Low Countries. The Chinese gazetteer evidence, reconstructed by Rozman and Skinner, reveals a sophisticated urban hierarchy—market towns feeding regional centers feeding the great Yangtze metropolises. Istanbul was one of the world's largest cities through the sixteenth and seventeenth centuries, and the Ottoman core sustained a network of commercial cities from Izmir to Aleppo. The Delhi-Agra corridor under the Mughals supported dense craft production and long-distance trade.
| Metric | Yangtze delta (rice paddy) | England (enclosed arable) |
|---|---|---|
| Yield per acre | Higher — intensive multi-crop irrigated system; 2–3 harvests per year in favorable areas | Lower — single-crop rotation; Norfolk four-course from mid-18th c. |
| Output per worker | Lower — labor-intensive cultivation requiring large seasonal workforce | Higher — less labor per unit of output; enclosure reduced labor needs |
| Market integration | Deep — Jiangnan regional markets moved grain prices across hundreds of km within weeks | Deep — integrated grain markets by 1700; canal network from mid-18th c. |
| Overall productivity position | At the Malthusian frontier: extracting maximum calories from available land | At the Malthusian frontier: extracting maximum value per worker from available land |
Agricultural productivity, the foundation on which all pre-industrial economic life rested, was at the Malthusian frontier in both the Yangtze and English systems c. 1750. Rice-paddy agriculture in the Yangtze delta produced higher yields per acre than English enclosed-arable farming: the intensive, irrigated system extracted more calories from the same land. English agriculture produced more output per worker, reflecting the higher cost of labor that would become central to the divergence story. Neither system held a clear overall advantage. Each operated at the limit of what its technology and labor configuration could sustain.
Commercial sophistication reinforces the pattern. The Yangtze delta had merchant networks spanning the Jiangnan region, credit instruments that preceded European bills of exchange by centuries, and market integration deep enough to arbitrage grain prices over hundreds of kilometers. The Mughal heartland operated the hundi system—an efficient network of bills of exchange—alongside complex land-revenue arrangements that functioned as fiscal infrastructure for a continent-spanning empire. The Ottoman core combined the mukataa tax-farming system with extensive guild networks and Mediterranean trade integration.
These regions were not "the same." They had profoundly different political structures, demographic patterns, ecological constraints, and cultural orientations. The comparability claim is narrower: on the economic margins that matter for explaining where sustained per-capita growth first emerged, the leading cores of Eurasia stood within a comparable range through most of the early modern period. What diverged, when, and by how much is the next question.
Allen's welfare-ratio series is this chapter's central data exhibit. Four cities, four centuries, one chart, and the shape tells the story.
Between 1500 and 1700, the four lines (London, Beijing, Delhi, Istanbul) move within the same broad band. London's welfare ratio fluctuates between 2.5 and 4.0, dipping after the sixteenth-century price revolution and recovering toward 1700. Beijing holds steady around 1.5, reflecting the stability of the late-Ming and early-Qing commercial economy. Delhi sits lower, around 1.0–1.5, with wider uncertainty reflecting a sparser documentary record. Istanbul fluctuates between 1.5 and 2.0, disturbed by the 1620 currency debasement that compressed real wages for a generation.
Around 1750, the lines separate. London begins pulling upward, from roughly 3.6 in 1725 to 4.5 by 1820 and then steeply higher through the nineteenth century. Beijing holds steady or edges downward. Delhi stagnates. Istanbul drifts. By 1820, what had been a loose band of comparable welfare ratios has become a fan. By 1870, the fan has become a chasm.
The same pattern appears in urbanization curves. England's urbanization rate begins climbing around 1750, from roughly 20 percent to over 50 percent by 1850, while the Yangtze delta's urbanization stagnates or declines slightly from its early modern peak. GDP-per-capita reconstructions from the Maddison Project (Bolt and van Zanden) confirm the shape: Britain's GDP per capita roughly doubled between 1500 and 1820, while Chinese and Indian GDP per capita moved sideways. Then the nineteenth century produced an explosive divergence in which northwestern Europe pulled away by a factor of four or more within a single lifetime.
Two things about the timing deserve attention. First, the divergence in real wages is visible by roughly 1750–1800, before the most dramatic phase of industrialization. London's wages were pulling away from Beijing's before the textile factories of Manchester were running at scale. The wage divergence partly precedes the Industrial Revolution; it is a cause as much as a consequence. Second, the GDP-per-capita divergence looks different from the wage divergence in phasing. GDP series start their dramatic separation around 1820, a generation later than wages. The GDP map begins at 1820 not because the data is irrelevant before that date, but because the Maddison-Project reconstructions are sparse before it. This is a limitation of the evidence, not of the phenomenon.
The sources behind these numbers matter. Allen's welfare-ratio series for London rests on building-trades wage records: robust, continuous, well-documented. His Beijing series draws on rice prices and skilled-labor wages from Qing dynasty archives, supplemented by Broadberry's independent Asian GDP estimates. Şevket Pamuk's Istanbul data uses Ottoman silver-wage records, with explicit caveats about the distortion introduced by currency debasements. The pre-1700 data for Delhi has the widest uncertainty band, as documentary records are sparser and the welfare ratio depends on basket assumptions drawn from a later period.
None of this invalidates the comparison. The divergence shape is robust across metrics and across different scholars' reconstructions. Real wages, urbanization, GDP per capita, energy consumption per capita: all show the same pattern of comparability followed by separation, with the inflection between 1750 and 1820. The question is no longer whether a divergence happened. It is why it happened where it did.
Pomeranz's argument bites because it specifies what made northwestern Europe distinctive, not merely that it was distinctive. If the comparable-cores frame establishes that the leading regions of Eurasia were on a similar productivity frontier c. 1750, then the divergence has to be explained by something that changed the configuration in one place and not the others. Pomeranz identifies two things: coal and colonies.
Britain sat on accessible coal seams, concentrated near population centers and navigable waterways rather than buried in remote interiors. The coalfields of Northumberland, South Wales, Yorkshire, and the Midlands were within economic reach of the cities that would industrialize. Coal is energy. Before coal, every economy on earth ran on the annual flow of solar energy captured through agriculture: food for human labor, fodder for animal power, wood and charcoal for heat. This flow imposes a ceiling. Population growth pushes against the land needed to produce food and fuel; urbanization concentrates demand that rural hinterlands cannot expand to meet. The Malthusian ceiling that constrained all pre-modern economies was fundamentally an energy ceiling.
Coal broke it.
British coal output grew from roughly 3 million tons in 1700 to 15 million by 1800, a fivefold increase that preceded and enabled the steam engine's adoption. By 1850, output reached 50 million tons; by 1900, 225 million. Each ton released the energy equivalent of acres of timber that didn't have to be grown. Wrigley calculated that Britain's coal use by 1800 was equivalent to the energy output of a hypothetical forest covering several million additional acres, land that Britain did not have and could not grow.
The Yangtze delta had no equivalent. China's coal deposits are concentrated in Shanxi and the northern interior, hundreds of kilometers from the commercially advanced coastal and riverine cities. The geographic mismatch between where China's coal sat and where China's commercial sophistication lived meant that the Yangtze delta could not replicate Britain's energy breakthrough. This is Pomeranz's first leg: coal was a geographically contingent windfall, not a reward for institutional superiority.
The second leg is colonies. The New World, and the Atlantic plantation economy built on it, provided what Pomeranz calls ghost acreage, a structural extension of Britain's productive base across the ocean. West Indian sugar plantations, Caribbean cotton, North American timber, and the Atlantic slave trade that transported over twelve million enslaved Africans between 1500 and 1870 gave Britain access to land, resources, and labor that could not be reproduced domestically.
Ghost acreage matters for a specific structural reason. Without the New World, Britain's late-eighteenth-century economy would have faced the same land constraint the Yangtze delta faced: rising population pressing against finite agricultural hinterland, with no prospect of breaking the Malthusian ceiling by domestic means alone. The plantation economy relieved that pressure through sheer resource volume. Cotton fed the Lancashire mills. Sugar fed London's working class. Timber built the ships. The resources flowed because an Atlantic trade system, built over three centuries of colonization, had organized them to flow.
The thesis is not that coal and colonies "caused" the Industrial Revolution in some simple, monocausal sense. Pomeranz's claim is structural: without the coal endowment and without the colonial resource base, the economic configuration that made sustained per-capita growth possible in Britain would not have obtained. The Yangtze delta, which matched Britain on nearly every other economic margin, lacked both. That lack, geographic and structural rather than institutional or cultural, is what makes the divergence explicable.
Allen's argument does not replace Pomeranz's. It specifies the mechanism.
If Pomeranz explains why Britain had the right configuration (coal plus colonial resource base), Allen explains how that configuration produced the innovations that drove the takeoff. His key insight is deceptively simple: innovation responds to prices. When labor is expensive relative to capital and energy, firms have economic incentive to replace workers with machines. When labor is cheap, they do not. Britain c. 1750 had the highest wages and cheapest energy in the world. That factor-price configuration, unique among the comparable cores, made labor-saving, capital-using technology profitable in Britain in a way it was not anywhere else.
| Core | Real wage level | Energy cost | Wage ÷ energy index (London = 1.0) |
|---|---|---|---|
| London | High (4.0 subsistence baskets) | Low (abundant, accessible coal) | 1.00 |
| Amsterdam | High (3.5–4.0 baskets) | Moderate (peat, imported coal) | 0.75 |
| Beijing | Low (1.5 baskets) | High (distant coal; charcoal) | 0.15 |
| Delhi | Low (1.0–1.5 baskets) | High (wood, dung, charcoal) | 0.10 |
| Istanbul | Moderate (1.5–2.0 baskets) | Moderate-high (wood, imported fuel) | 0.20 |
Allen's data shows the configuration in stark terms. London's real wages c. 1750 were two to three times higher than Beijing's in any consistent basket. British coal was the cheapest energy source of any major economy. The ratio of wages to energy costs was higher in Britain than in any of the other cores by a factor of three to five. When a British manufacturer weighed whether a machine that replaced ten workers with one worker plus coal-fired power was worth the investment, the arithmetic favored the machine. When a Chinese manufacturer in Suzhou faced the same question with low-wage workers and no cheap coal, the arithmetic pointed the other direction.
This is factor-price-induced innovation: the direction of technical change follows from relative factor prices. The formal version (developed in Economics Chapter 13) shows that when the wage-to-capital-cost ratio exceeds the marginal rate of technical substitution at the production frontier, labor-saving innovation becomes the cost-minimizing path. The historical version is more concrete.
| Year | Indian hand-spinning (pence/lb) | British steam spinning (pence/lb) | Competitive position |
|---|---|---|---|
| 1780 | 6 | 12 | Indian hand-spinning cheaper by 2× |
| 1790 | 6 | 8 | Gap narrowing; steam still more expensive |
| 1800 | 6 | 4 | Inflection: steam wins on cost |
| 1810 | 5 | 2 | Steam cheaper by 2.5× |
| 1815 | 5 | 1.5 | Steam cheaper by 3×; Indian spinning uncompetitive |
Indian hand-spinners c. 1780 produced yarn at roughly 6 pence per pound, competitive with British hand-spinning and well below the cost of the early, crude steam-powered alternatives, which ran at about 12 pence per pound. The machines were expensive relative to what they replaced. But British wages were high enough that the capital investment paid for itself through labor savings in a way that Indian wages, at a fraction of British levels, could not justify. Each generation of steam technology became incrementally cheaper because each generation was incrementally profitable: the ratchet worked because someone was willing to fund the next iteration.
The inflection came around 1800–1810. The cost of steam-spun yarn fell from 12 pence to 8 to 4 to under 2 pence per pound across three decades of iterative improvement. By 1815, steam-spun British yarn was cheaper than Indian hand-spun yarn for the first time. The machines won, not because they were "better" in the abstract, but because the economics of British factor prices had made each generation of improvement worth funding.
Allen's thesis thus sits within Pomeranz's. Coal gave Britain cheap energy. The Atlantic economy gave Britain high wages through the demand for labor in a trade-enriched economy. Together they produced the factor-price configuration that made the specific pattern of British innovation (labor-saving, energy-using, capital-intensive) economically rational. Without the configuration, the inventions were unprofitable. Without profitability, they were not adopted, not iteratively improved, and not scaled.
Two more accounts of the divergence. Each describes something real about northwestern Europe c. 1750. Each faces the same problem when held against the comparable-cores frame.
Joel Mokyr's case begins with the observation that Europe's advantage was cognitive, not primarily economic or institutional. What he calls the Industrial Enlightenment, a cultural formation that emerged in northwestern Europe between roughly 1660 and 1800, turned tinkering into systematic improvement. The Royal Society (founded 1660), the Encyclopédie (published 1751–1772), the Birmingham Lunar Society (from 1765), the Manchester Literary and Philosophical Society (from 1781): these were nodes in a network that connected theoretical knowledge to practical improvement in a way with no exact counterpart elsewhere.
Mokyr distinguishes propositional knowledge (understanding how nature works) from prescriptive knowledge (knowing how to make things). His thesis is that the Industrial Enlightenment created a feedback loop between the two. Propositional advances suggested prescriptive applications; prescriptive failures prompted propositional inquiry; and a Republic of Letters, the pan-European network of correspondence, publication, and intellectual exchange, allowed ideas to travel from one practitioner to another faster than in any other civilization. The steam engine was improved not just by craftsmen tinkering with cylinder tolerances but by natural philosophers investigating the physics of heat. The institutional culture that valued and disseminated such investigations was, Mokyr argues, distinctively European.
The case is serious. Europe's scientific institutions, publication networks, and craft-to-science feedback channels were genuinely distinctive by 1750. Mokyr's strongest evidence comes from the specific texture of improvement: the incremental, documented, communicated character of British technical change in the eighteenth century, in which each improvement was built on a published or circulated predecessor. The question is whether this cognitive infrastructure was distinctive enough, and exclusively enough, to carry the full explanatory weight.
Daron Acemoglu, James Robinson, and Douglass North make a different structural argument: institutions determine economic outcomes. Secure property rights, parliamentary checks on executive power, independent courts, and a political system that distributed economic opportunity broadly rather than concentrating it in an extractive elite: these are the preconditions for sustained investment and innovation. Where institutions were inclusive, economies grew. Where they were extractive, elites captured surplus and blocked the changes that would have enabled growth.
The sharpest empirical exhibit for the institutional account is the Acemoglu, Johnson, and Robinson (AJR) settler-mortality regression, documented in the economics textbook (Chapter 18, §18.3). AJR use the mortality rate of European settlers in colonial territories as an instrumental variable for institutional quality. Where settlers survived, they built inclusive institutions; where they died quickly, they built extractive ones. The first-stage relationship is strong (F-statistic above 20), and the two-stage-least-squares estimate implies that a one-unit increase in institutional quality causes a roughly 0.9 log-point increase in GDP per capita.
Additional natural experiments reinforce the account. North versus South Korea, East versus West Germany, pre- versus post-reform China: all demonstrate that institutional divergence produces income divergence among populations that share geography, ethnicity, and initial conditions. North and Weingast's account of the Glorious Revolution, the 1688 settlement that constrained the English crown's fiscal power and made government debt credible, offers a specific historical mechanism linking institutional change to financial development in Britain.
| Core | Civil / legal administration | Fiscal capacity | Commercial law |
|---|---|---|---|
| NW Europe | English common law; Dutch commercial courts | Parliamentary fiscal control; tax farming → state collection | Bills of exchange; partnership law; limited joint-stock |
| Yangtze delta | Qing Code; magistrate courts; merchant guild arbitration | Land tax + salt monopoly; granary system | Merchant networks; huipi bills; guild self-regulation |
| Mughal heartland | Qazi courts; mansabdari administration | Zamindari revenue; jizya; trade tolls | Hundi exchange; caravan credit; sarrafs (bankers) |
| Ottoman core | Kanunname + Sharia; kadi courts | Timar land grants; mukataa tax farms | Ahdname trade treaties; guild regulations |
Both accounts run into the same evidence. Song-dynasty China (960–1279) operated civil-service examinations, commercial codes, sophisticated credit instruments, and agricultural extension services that matched or exceeded European institutional sophistication of the same period. The Ming and Qing legal codes regulated property rights, commercial disputes, and inheritance in ways that created functional security for investment, even if the forms differed from English common law. The Mughal heartland administered zamindari land-revenue and mansabdari military-administrative systems of enormous fiscal capacity. The Ottoman Empire governed commercial life through the kanunname alongside Sharia, with kadi courts providing dispute resolution and ahdname treaties governing trade relations with European states.
These are not the marks of institutionally unsophisticated economies. If institutions explain the divergence, the explanation must account for why East Asian and West Asian institutional sophistication (real, documented, and functional) did not produce the same breakthrough. The AJR regression addresses variation among colonized territories, but that is a different question from the one this chapter asks: not "among former colonies, which ones grew faster?" but "among the leading cores of Eurasia, why did one break out?" Mokyr's cultural account faces an analogous challenge. Chinese print culture predated Gutenberg by centuries. Song-era examinations selected for propositional knowledge. Islamic scientific traditions in optics, astronomy, and algebra represented genuine theoretical achievement.
These accounts describe real conditions. They do not establish geographic exclusivity at the scale the comparable-cores comparison demands.
Pomeranz's coal-and-colonies thesis is the strongest single account of why northwestern Europe diverged from the other Eurasian cores. Allen's factor-price model specifies how it worked. Mokyr and the institutionalists describe real second-order conditions. None of the three, alone, is the answer.
The coal-and-colonies thesis wins on geographic specificity. Coal deposits near industrial centers explain why Britain and not the Netherlands, which had comparable institutions and commercial sophistication but lacked coal. They explain why Europe and not China, which had coal but not where its commercial economy was concentrated. They explain the timing: the divergence coincides with the scaling of coal-based energy and the maturation of the post-1750 colonial economy, not with any institutional revolution or cultural awakening that can be dated to the same decades. And they explain the absence of equivalent breakthroughs where the configuration was missing. The Yangtze delta operated on a comparable economic frontier. It had no coal within reach and no ghost acreage. It did not industrialize.
Allen supplies the microfoundation. A structural account of why Britain had the right configuration is incomplete without a mechanism connecting that configuration to the specific innovations that drove growth. Factor-price-induced innovation provides it: high wages plus cheap energy equals economic incentive for labor-saving machinery. The mechanism is testable, quantifiable, and works at the level of individual industries. It also explains why the innovations were iteratively improved: each generation was profitable, so the next generation was funded.
Mokyr and the institutionalists sit at a different level. The Industrial Enlightenment was real: northwestern Europe's science-craft feedback networks were distinctive by 1750, and they contributed to the speed and direction of technical change. Institutional quality mattered: secure property rights and credible government reduced investment risk. But neither establishes the kind of geographic exclusivity the comparable-cores evidence demands. Song China had institutional sophistication; the Islamic world had science-craft interaction; Mughal and Ottoman commercial law provided functional property security. The question is not whether European culture and institutions were good (they were), but whether they were sufficiently different from Asian alternatives to carry the divergence's full weight. The evidence says they were not. They operated as enabling conditions necessary in some form for any pre-modern commercial economy to reach the comparable-cores frontier, rather than as the differentiating cause.
A reader carrying this forward should hold three things. No monocausal answer is complete, including this chapter's preferred one. Pomeranz plus Allen explains the British case more convincingly than any rival, but the coal-and-colonies thesis does not explain every feature of the subsequent spread of industrialization: Belgium's early adoption, America's distinctive path, Japan's Meiji-era breakthrough. The contingency is genuine: if coal had been located in the Yangtze delta rather than in Northumberland, or if the Ottoman Empire rather than Spain had colonized the New World, the divergence might have run differently. To call the outcome contingent is not to call it random. It is to say that the explanation is structural and geographic rather than civilizational.
The next chapter narrates what happened in Britain c. 1760–1840, the Industrial Revolution as a discrete event, building on the conditions this chapter has argued were unique. The relationship between the chapters: this one explains why Britain; that one explains what happened in Britain. The comparative case comes first because without it, the Industrial Revolution looks inevitable rather than contingent, and the contingency is the point. For the modern policy dimension of this evidence (why some countries remain poor today and what the divergence implies for development economics) see Economics Ch. 20 and Big Question #2: Why are some countries rich and others poor?