The Little Divergence in early modern Europe is a well-documented phase of economic history, but the causes of Northwestern Europe’s surge ahead remain controversial (Broadberry Reference Broadberry2021a, 2021b). An influential tradition, dating back to the work of North and Weingast (Reference North and Weingast1989), attributes English success to institutional change: the development of a semi-representative democracy and constrained executive. Acemoglu, Johnson, and Robinson (Reference Acemoglu, Johnson and Robinson2005) also foreground the state, arguing that pre-1500 political institutions interacted with the rise of Atlantic trade to produce a pro-growth political environment in England and the Netherlands, but not Spain and Portugal. Limited Anglo-Dutch polities are contrasted with the “absolutist states” of Iberia, whose despotic rulers purportedly intervened arbitrarily in commercial activities. Recent research by Henriques and Palma (Reference Henriques and Palma2023), however, casts doubt on the start date of the institutional divergence: until the mid-seventeenth century, Spain and Portugal both compare favorably with England in the frequency of their parliamentary meetings, currency debasements, extraordinary taxes, and in their sovereign risk. Reconstruction of Portugal’s national accounts, moreover, reveals a dynamic early modern economy. The period of Portuguese reversal, in both absolute and relative terms, only began in the second half of the 1700s (Palma and Reis Reference Palma and Reis2019).
In this paper, we show that eighteenth-century Portugal was beset by a resource curse because of gold inflows from Brazil, whose outer provinces were remitting over 70 percent of global output to Lisbon by 1730 in a continued process that would last decades. Treasure imports, mainly private, were a sudden addition to real incomes, which were then spent on domestic products, driving up prices and wages in the non-traded sector, as had previously happened in Spain (Charotti, Palma, and Santos 2022).Footnote 1 This led to an appreciation of the real exchange rate and attracted productive factors from the increasingly uncompetitive traded sector, lowering output and raising costs. Portuguese consumers found imports cheaper and began to substitute these for home goods. The Dutch disease thus manifested itself in relative price movements: traded prices were internationally determined and remained stagnant by comparison with inflation in the non-traded sector. By constructing relative price indices for four Portuguese regions using archival data, we show that such movements did, in fact, occur during and in the wake of the gold shock. We further explore the temporal relationship between persistent real exchange Footnote 2 rate fluctuations and various indicators of economic development—real GDP and wages, the non-agricultural population share, and trade balance statistics—to test for long-term structural consequences. Our price data for the period between 1650 and 1800 shows that a real exchange rate appreciation of about 30 percent occurred during the eighteenth century, leading to a loss of competitiveness in national industry. Following this evidence regarding the economic aspects of the resource curse, we move to an overall assessment of the resource curse. The latter also includes a political component that resulted from the additional revenues due to the gold reducing the rulers’ incentives to negotiate and leading to a complementary process of state capture. We document these events and compute a counterfactual using a synthetic control method. This suggests that the price level in Portugal rose relative to other countries. By 1800, Portugal’s GDP per capita was over 40 percent lower than it would have been if Portugal had not been the first-stage receiver of Brazilian gold.Footnote 3
Our results are related to but contrast with much of the existing literature. As early as 1701, the colonial administrator D. João de Lencastre raised concerns that Brazilian gold would not benefit Portugal in the long run, but instead England or the Netherlands, just as had happened with Spain’s silver (Boxer Reference Boxer1962, p. 42). A decade later, an Italian Jesuit living in Brazil also expressed his opinion that the gold mines were detrimental to the development of Portugal, despite being sparse on details (Antonil Reference Antonil2011, p. 266). In the twentieth century, Vitorino M. Godinho argued that commercial crises related to poorer export-sector performance led to surges of successful industrial policy. In contrast, when colonial booms returned, the former was abandoned (Godinho Reference Godinho1955, pp. 258, 292). Relatedly, Macedo (Reference Macedo1982a, p. 200) argued that once the gold inflows diminished in the last quarter of the eighteenth century, the economy largely recovered. Footnote 4 However, this viewpoint is not supported by the most recent data on structural change and growth: Portugal’s economy never recovered from the industrial decline that occurred during the eighteenth century (Palma and Reis Reference Palma and Reis2019; Palma Reference Palma2020).
Contrary to views foregrounding the supposed long-run neutrality of money, the resource curse can have persistently negative effects on developing economies.Footnote 5 In the case of eighteenth-century Portugal, the gold inflows encouraged the substitution of foreign imports, and caused a severe contraction of the export sector, diminishing employment across a range of critical industries. These included both manufacturing and agriculture (Fisher Reference Fisher1963; Costa and Reis Reference Costa and Reis2017). When technological progress (or the potential for it) depends on the lagging traded sector—which, given the scalability of newly industrialized textiles, was the situation in early modern European economies—the contraction of these industries slows per capita income growth (Van Wijnbergen Reference Wijnbergen1984). Matsuyama (Reference Matsuyama1992) similarly suggests that the reorientation of activity away from a manufacturing sector characterized by externalities in production could have negative growth effects in a dual-sector open economy. Therefore, the effects of Dutch disease on economic dynamism can be profound and enduring.Footnote 6
In Portugal, a significant real exchange rate appreciation occurred during the first half of the eighteenth century. Footnote 7 The price of non-traded goods in terms of traded goods rose precipitously in all the regions of Portugal after around 1700, exceeding the pre-gold era level by over 30 percent following the 1720s. This coincided with the peak of gold output, which exceeded 140,000 kilograms in the decade 1741–1751 (TePaske Reference TePaske2010). Notably, the price movements coincided with a period of general price inflation and were driven primarily by accelerated increases in the non-traded sector. These developments are as predicted by the standard Dutch disease model, and diminished output of traded goods as imports increased and factors of production flowed into the non-traded sector.
As real exchange rates turned against Portuguese exports, the burgeoning expansion of the late seventeenth century stalled. Urbanization fell from 17.3 to 16.2 percent between 1750 and 1800, while the rural non-agricultural share—representing much of the nation’s handicrafts—remained static at around 29 percent; further declines would continue into the nineteenth century (Palma and Reis Reference Palma and Reis2019, p. 485). Real GDP and wages stagnated and then declined during the same period, only starting to recover more than a century later. Footnote 8 At a time when urbanization and per capita output were rising across a fast-commercializing Western Europe, Portugal’s decades of outright depression from the 1770s stand out in stark relief.
This paper contributes to literatures in both economics and economic history. Corden and Neary (1982) and Corden (Reference Corden1984) present the basic formulation of Dutch disease theory, but the model is limited by the fact that the distortion ends with the monetary stimulus; in Portugal, GDP per head continued to decline after gold imports had ended. Van Wijnbergen (Reference Wijnbergen1984), Krugman (Reference Krugman1987), Matsuyama (Reference Matsuyama1992), and Sachs and Warner (Reference Sachs and Warner1995) link resource abundance with reduced economic growth to varying degrees through a decline in the size of the traded sector, and Gylfason, Herbertsson, and Zoega (Reference Gylfason, Herbertsson and Zoega1999) explicitly tie real exchange rate appreciation to a contraction of skill-intensive manufacturing. Asea and Lahiri (Reference Asea and Lahiri1999) show that the increased marginal product of unskilled labor slows human capital accumulation by making education costlier. Among others, Forsyth and Nicholas (Reference Forsyth and Nicholas1983) applied the Dutch disease model to early modern Spain, while Stein and Stein (Reference Stein and Stein2000) considered a political channel. Palma (Reference Palma2020) and Charotti, Palma, and Santos (2022) emphasize both channels and review the literature related to early modern Spain. Our resource curse interpretation of the economic and political history of eighteenth-century Portugal suggests a path that largely repeats the earlier story of Spain.Footnote 9
HISTORICAL BACKGROUND
It is well known that countries with abundant natural resources often experience negative economic and political consequences. A resource curse has both an economic manifestation (Dutch disease) and a political one (known as the political or institutional resource curse), manifesting itself in reduced checks on executive power, poor governance, and weak rule of law. Footnote 10 In this section, we provide the historical context of Portugal’s experience, including a description of the gold arrivals, and a narrative concerning the evolution of the industrial sector over the eighteenth century. The most striking fact of early modern Portuguese economic history is that real GDP per capita fell by one-third between 1750 and 1800, and it stagnated during the first half of the nineteenth century, just as other parts of Western Europe were industrializing and entering modern economic growth (Table 1). Around 1800, Portugal had endured a century that left it with a barely existent industrial sector and inadequate political institutions. As Figure 1 shows, Portugal saw a decrease in the size of its non-agricultural sector just as each of its European rivals were either flattening (Italy) or experiencing structural transformation out of agriculture. The consequences of the status quo around 1800 would continue to be felt in the future, but they are the result of earlier economic and political developments, as we document in this paper.
Table 1 GDP PER CAPITA OF SELECTED WESTERN EUROPEAN COUNTRIES IN CONSTANT PRICES (“INTERNATIONAL” GEARY-KHAMIS DOLLARS OF 1990)
Sources: Data for England from Broadberry et al. (Reference Broadberry, Campbell, Klein, Overton and van Leeuwen2015). This corresponds to Britain from 1700. In the case of Holland, borders correspond to Holland until 1800 and the Netherlands for 1850; a benchmark for 1807 was used for the data prior to 1800 (van Zanden and van Leeuwen Reference van Zanden and van Leeuwen2012, p. 121), and the 1850 level is from Smits, Horlings, and van Zanden (2000). Data for France from Ridolfi and Nuvolari (Reference Ridolfi and Nuvolari2021, p. 419). Germany corresponds to the 1871 borders and the data comes from Pfister (Reference Pfister2022); Spain, from Prados de la Escosura, Álvarez-Nogal, and Santiago-Caballero (Reference Prados de la Escosura, Carlos and Carlos2022). Portuguese data is from Palma and Reis (Reference Palma and Reis2019). “Italy” corresponds to northern and central Italy and comes from Malanima (Reference Malanima2013). Swedish data is from Krantz (Reference Krantz2017) for 1500–1560 and Schön and Krantz (Reference Schön and Krantz2015) for 1560–1850.
Figure 1 SHARES OF POPULATION OUTSIDE AGRICULTURE FOR EUROPEAN COUNTRIES
Sources: Keibek (Reference Keibek2017) for England; Palma and Reis (Reference Palma and Reis2019) for Portugal; Ridolfi and Nuvolari (Reference Ridolfi and Nuvolari2021) for France; Chilosi and Ciccarelli (Reference Chilosi and Ciccarelli2022) for Italy. Figures for Spain derived by back projecting from urban population shares adjusted to net out urban dwellers living on agriculture, using the average of the labor share in 1787 and 1797 as a benchmark, according to the population census for these two years (66.1 percent) (Prados de la Escosura, Álvarez-Nogal, and Santiago-Caballero Reference Prados de la Escosura, Carlos and Carlos2022). We thank Leandro Prados de la Escosura for this suggestion.
The Discovery and Import of Gold from Brazil
Around 1694, alluvial gold was discovered in a province of Portuguese Brazil known today as Minas Gerais. The original explorers were adventurers searching for native slaves and silver, and they sought for a time to keep the finding a secret. By 1697, however, the word was out, and an unprecedented gold rush wave of internal and external migration eventually swept the region (Magalhães Reference Magalhães2005). Gold was found in other surrounding regions of Brazil as well. As many as 50,000 prospectors and their slaves may have inhabited the mining towns as early as 1705 (Boxer Reference Boxer1969, pp. 455–57). Gold hunters arrived en masse from Portugal.Footnote 11 Before long, gold was flowing back to Europe in vast quantities, some in the form of taxes—usually the “royal fifth” (quinto) levied on domestic production—but mostly through remittances by the private contractors who predominated in the mines, which were privately owned. Footnote 12 The quantities produced were extraordinary, peaking at about 14,000 kg per year around 1750. Footnote 13 While local governors strove to regulate exports, smuggling was sometimes successful, especially prior to 1720; powerful merchants in the annual Lisbon fleets hid treasures in barrels, chests, and sugar bags, while informants were discouraged (Boxer, Reference Boxer1969, pp. 459–61). TePaske (Reference TePaske2010), estimating aggregate Brazilian production from 1690 to 1810, suggests that output stood at 40,000 kilograms during each of the first two decades of the eighteenth century, before exploding to 100,000 kilograms in the 1720s and over 140,000 in the 1730s. Of this total, more than 80 percent returned to Portugal (Costa, Lains, and Miranda 2016, p. 204). Three annual fleets shipped an average of 3 billion reis apiece, the proceeds of which were distributed to around 2,300 private individuals (Costa, Rocha, and de Sousa 2013, p. 1148; Costa, Rocha, and Brito 2018).
Most treasure was private. Footnote 14 It generally arrived already minted and about 72 percent was eventually re-exported (de Sousa Reference de Sousa2006, p. 244). The arrivals directly raised incomes considerably in the short term. The gold inflows lasted for around a century, however, and there were net negative effects still during this time frame. After a boom of around four decades in the first half of the eighteenth century, Portugal’s economy stagnated for around two decades, and then began a process of absolute and relative economic decline from the 1770s (Palma and Reis Reference Palma and Reis2019)—at a time when important gold inflows were still taking place (Costa, Rocha, and de Sousa 2013).
Portugal ran current account deficits (paid for in gold) with all of its major trading partners during the eighteenth century, particularly with England (Costa, Lains, and Miranda 2016; Fisher Reference Fisher1971, p. 197). Footnote 15 The deficit increased steadily throughout the first half of the eighteenth century, stabilizing at around 4,000 million reis per annum by the 1750s. Portuguese annual imports from England rose from 355,000 pounds in 1699–1702 to 1,300,000 pounds in 1756–60, while exports fell from a peak of 387,000 pounds in 1720 to 257,000 pounds over 1756–60.
The bulk of English goods imported to Portugal consisted of woolen textiles—never falling below 70 percent in any year before 1760, and at one stage reaching 84 percent—most of which were then re-exported to Brazil, where demand for durable goods had exploded with the expansion of mineral wealth. Footnote 16 In 40 of the 60 years between 1700 and 1760, the yearly London rate of exchange was at or below the gold export point of 5/5.75d (Fisher Reference Fisher1963, p. 223). A staggering 90 percent of exported gold went to financing the trade deficit with England and France, regularly departing for Plymouth in packet boats, which carried diplomatic immunity and were thus able to evade prohibitions set on metallic outflows (Francis Reference Francis1966, pp. 216–17). As Figure 2 illustrates, for decades, gold corresponded to 70 percent of the value of Portuguese exports—thus playing a crucial role in balancing the distorted current account. Footnote 17 The peak of gold remittances and the trough of the trade deficit essentially coincide during the early 1750s. By then, all Portuguese factories had failed except for a few highly subsidized cases (Pedreira Reference Pedreira2005, p. 196).
Figure 2 GOLD EXPORTS AS A PERCENTAGE OF TOTAL EXPORTS (INCLUDING GOLD)
Sources: Figures for gold derived from Costa, Rocha, and de Sousa (2013), cited with export figures in Costa, Lains, and Miranda (2016, p. 205).
Portugal’s increasing deficits with England must be seen in the context of the Methuen Treaty of 1703, which exchanged preferential rates on English textiles (23 percent) for a reduction in wine tariffs to one-third below those allotted to France.Footnote 18 While it is tempting to see the agreement as a codification of core-periphery relations between the two countries, establishing the former as a manufacturer and the latter as a primary product exporter, to do so would be anachronistic. Tariffs on English textile imports were not reduced but instead an official restatement of a secret clause in a 1654 peace treaty that guaranteed these same rates, which had not previously been enforced by customs officials. The new article introduced in the Methuen Treaty of December 1703 corresponded to the opening of the English market to Portuguese wines, and the whole treaty always had a military component to it; most immediately, it concerned Portugal’s participation in the Grand Alliance during the War of the Spanish Succession (Costa, Lains, and Miranda 2016, p. 140). The Treaty hence had military as well as economic implications, and it kept political conditions between the two powers stable, such that the effects of gold imports on the balance of trade could unfold according to largely economic factors. The signing of the Methuen Treaty in late 1703 hence mattered, but it was endogenous to the gold inflows that by then were arriving and were expected to continue to arrive in increasing quantities to Portugal (Macedo Reference Macedo1982a, p. 45). Indeed, when gold inflows shrank and trade accounts partially rebalanced from the early 1770s, the agreement was still legally binding (Costa, Lains, and Miranda 2016, p. 200). The Methuen Treaty thus contributed to the de-industrialization of Portugal’s economy, but it was merely a mechanism, not the ultimate cause.
A substantial share of the extracted treasure was retained within Portugal. Costa, Rocha, and Brito (2018, p. 1153) show that the domestic gold stock rose from 1.149 billion reis in 1720 (or 1.4 percent of nominal GDP) to more than 58 billion reis by mid-century (56.4 percent) (Palma and Reis Reference Palma and Reis2019; Reis Reference Reis2002). The growth rate of the money supply, 2.5 percent per annum, substantially exceeded output growth over the course of the eighteenth century; 80 percent of this total consisted of coins, most of which were private and many of which were invested in land—a fundamental factor in the non-traded sector (Costa, Rocha, and Brito 2018, p. 1153). Growing demand for land led the Pombal administration, once the gold inflows began to shrink, to eventually restrict property transfers to the Church and suppress entailments to improve market liquidity, measures which, though unsuccessful, reflect the rising value of a static resource. In the southern Alentejo region, rising meat prices—a non-traded good—led to wholesale conversion to husbandry from cereal cultivation, which remained comparatively inefficient (yield ratios for wheat and rye were below 1:4). Overall agricultural production, which had risen steadily since 1660 (at .72 percent per annum according to one estimate), declined from 1750 until the Napoleonic Wars. Olive oil and wine exports both flourished, but these were products with no ready substitutes, and their producers had gained access to foreign consumers through diplomatic means, including the 1703 Methuen Treaty (Costa, Lains, and Miranda 2016, pp. 172–73). These sectors did little to arrest the steady decline of Portuguese shipments abroad.
As expected, real wages rose during the gold boom period. Footnote 19 Data is scarce on manufactures, but even with state industrial policy intervention during the second half of the eighteenth century, industrial development was sluggish at best. By 1769, only four textile factories were operating in the entirety of Portugal, and outside the royal factories, the scale of production remained desperately small (Pedreira Reference Pedreira1994, p. 59). Even the larger, state-run institutions were, in fact, glorified systems of organizing and subcontracting to existing networks of cottage production. Faced with daunting initial investment costs and insufficient capital, the only sizeable enterprises to be established were erected behind heavy tariffs and import bans. Pombal’s “industrial policy” did not change the previous structure of production; of 200 “program” facilities reported in 1777, the majority were small-scale workshops (Macedo Reference Macedo1982a, pp. 155–60). Indeed, the proliferation starting after 1777 (jumping from 55 to 235 factories) was associated with a liberalization of the import-licensing system to allow minor firms to set up. Of the 180 applications for licenses between 1757 and 1832, 114 (63 percent) were by foreigners, bespeaking the absence of domestic industrial entrepreneurship and capital (Pedreira Reference Pedreira1994, p. 59). The small scale of these undertakings, many of which were located in the interior of the country where transport costs shielded them from international competition, slowed the adoption of mechanized technologies, a further contributing factor delaying the onset of the Industrial Revolution.Footnote 20
Portuguese Industry Prior to 1700
Assessing the true costs of Portugal’s contraction requires a plausible counterfactual for the evolution of the eighteenth-century economy in the absence of the discovery of gold in Brazil. The starting point for such a scenario is the expansion of the Portuguese economy in the second half of the seventeenth century, following the end of the Iberian Union with the restoration of a Portuguese dynasty in 1640 and the end of hostilities in 1668. At this stage, Portugal was on the cusp of a strong advance in per capita income and structural transformation away from agriculture. From 1650 to 1700, the percentage of the population employed in agriculture shrank from 63.7 percent to 58.9 percent, while the rural non-agricultural share rose from 23.9 percent to 28.6 percent. Per capita incomes grew once hostilities ended, from around 900 “international” G-K dollars of 1990 in the 1640–1660s to more than more than 1,200 by the early 1690s (Palma and Reis Reference Palma and Reis2019, pp. 485, 497).
Portugal had a venerable tradition in rural manufacturing.Footnote 21 By the late seventeenth century the country had a growing industrial sector that included clusters of textile, soap, iron, glass, and silk manufacturing, among others (Macedo Reference Macedo1982a, pp. 25, 28). By 1680, Portugal’s textile industry not only supplied the needs of Portugal and its colonies, but also exported valuable textiles to Castile (Pedreira Reference Pedreira1994, p. 26, 2005). Industrialization progressed steadily during the 1680s (Dias Reference Dias1954). A successful industrialization effort, promoted by Luís de Menezes, the Count of Ericeira, was in place for more than a decade, having started shortly after the end of hostilities with Spain in the late 1660s and being abandoned only after 1700 (Macedo Reference Macedo1982a, pp. 25, 29). Footnote 22 The industrialization effort had an associated vertical integration at the retail level and was located in many of the same regions that industrialized on a larger scale more than two centuries later (Macedo Reference Macedo1982a, p. 32). In addition to an import-substitution policy, prior to 1700, there was also much investment in increasing the efficiency of the manufacturing process, including through the construction of new factories in Covilhã, Manteigas, Estremoz, Vila Melo, Fundão, and Lisbon—with other factories having been planned, but never been constructed, in Porto and Portalegre (Meneses Reference Meneses2001; Macedo Reference Macedo1982a, p. 36).Footnote 23
Portugal had the advantage of Brazil as both an export market and supplier of agricultural raw materials and other inputs, as well as tropical products for re-export (including sugar).Footnote 24 Until the 1690s, the political will existed to make Portugal a mercantilist exporting power. The industrial development of the last three decades of the seventeenth century was simultaneous with an industrial policy based on import substitution. Footnote 25 This was implemented through a sequence of laws between 1672 and 1698. They were known as leis pragmaticas and supported by the writings of individuals such as Duarte Ribeiro de Macedo, who from the 1670s strongly advocated for industrialization, influenced by Colbertism and the writings of English mercantilists. Ericeira became a key politician who from 1675 targeted woolens, silks, glass, and iron-working, granting large production units monopoly status as “royal factories” to reap the potential gains of concentration and protection. Footnote 26 In 1677, Covilhã was the site of a new public-private partnership for textile production. A plant was established on the bank of a stream, weavers were contracted, and English technicians were brought in. Footnote 27 Rural putting-out handled the spinning, while the dyeing and carding occurred in-house. Three years in, 17 looms were operated by 23 weavers apiece, and the contractors employed a total of 415 people. The success of the enterprise led to expansion into the neighboring town of Manteigas and plans for its replication in Estremoz. Footnote 28 From the 1770s, the Crown encouraged the development of the silk industry, with thousands of mulberry trees planted in different areas of the country (Lourenço, Reference Lourenço2007, p. 308). Silk factories opened in Lisbon and elsewhere. There were also important advances in metallurgy, particularly in nail production (Hanson Reference Hanson1981). Vertical integration with respect to iron production was encouraged, even after the death of Ericeira (Hanson Reference Hanson1981).
By the late seventeenth century, then, Portugal had the rudiments of a successful industrial take-off. This fact was noticed by international observers, such as the French Ambassador in Portugal who commented on the successful industrialization effort that was happening (Lourenço Reference Lourenço2007, p. 306). Several “rural industrial districts” built on sophisticated putting-out networks existed, there was a large, mercantile capital city stimulating demand for manufactures, and Brazil provided a potentially enormous export market and supplier of raw materials. By 1680, Portugal’s textile industry was, in fact, not only supplying the needs of the internal market including Brazil but it was also exporting to Castile. Portugal never became an industrial, exporting power at this time—but if the 1670s–1690s trends had continued, this could have happened over the eighteenth century. Instead, in the first half of the eighteenth century Portugal’s industry lost its competitiveness and experienced serious difficulties (Pedreira Reference Pedreira2005, p. 194). The sources of such difficulties have been debated, with some claiming them to be related to the indiscipline of its workers or the actions of the Inquisition, while others claiming it to be related to the import of the gold coins (Pedreira Reference Pedreira2005, p. 194). We provide evidence for the latter as the fundamental cause. In its absence, Portugal could have used import-substitution on the motherland to combine proto-industrial networks with urban manufacturing, importing inputs and exporting manufactures to and from its privileged market of Brazil. Footnote 29 As it was, however, agglomeration economies with long-term consequences did not occur, and Portugal’s backward industrial structure persisted into the nineteenth century and beyond (Pedreira Reference Pedreira2005, pp. 205–06). Additionally, the deteriorating quality of political institutions had negative consequences for human capital accumulation, which we will explain later. Portugal would only industrialize and enter a process of modern economic growth from the mid-twentieth century—extraordinarily late by Western European standards.
DATA AND RESULTS
In this section, we provide details about our data construction, methodology, and empirical results. Testing for Dutch disease requires detailed information on the prices of a range of traded and non-traded goods, so as to establish whether an appreciation in the real exchange rate occurred. We utilize data from the Price, Wages, and Rents in Portugal 1310–1900 (henceforth PWR) project, which collects long-run data of the aforementioned variables from primary sources available in archives in Portugal. Footnote 30 The database relies on a variety of primary record sources, including hospitals, charitable institutions (Misericórdias), convents and monasteries, royal palaces, municipal bodies, and the University of Coimbra. The price indices to be discussed later are constructed by us from numbers converted from their original values in geographically varying Portuguese measures to a metric standard. Footnote 31 Our selected time ranges from 1650 to 1825. We construct indexes as follows:
where q k corresponds to the quantity purchased of good k, p k,t is the nominal price of good k at time t in monetary units (reis) per metric units, N is the number of goods 1,2,…,N, and i is an indicator concerning whether the good is traded (t) or non-traded (nt).
To build the traded and non-traded indices, we first had to choose a relevant set of goods. We used the set of goods and weights employed by Allen (Reference Allen2001), in his respectability consumption basket, as modified for Portugal by Palma and Reis (Reference Palma and Reis2019). We then divided these goods into two categories, displayed in Table 2: traded and non-traded. No early modern good was fully traded, except in rare cases, but a range of products have been shown to have been either competitive enough in foreign markets (such as wine) or in domestic port cities (such as wheat). Costa, Lains, and Miranda (2016) show that nearly 80 percent of all imports were manufactures as of 1685, and that most of the remainder consisted of cereals, dairy products, and fish. The same source suggests that 914 million reis worth of wine and olive oil combined were exported, along with smaller but still significant quantities of salt. With this in mind, we selected wheat (substituting for maize, as dictated by historical consumption patterns and discussed later), olive oil, wine, linen, and candles as traded goods, and meat, hens, eggs, soap, and charcoal as non-traded. Live animals were difficult to move across national borders, and no evidence from the trade accounts suggested that animal products—beyond salted cod—were moved in significant quantities. Eggs and soap Footnote 32 were too fragile to be moved, while neither charcoal nor firewood appear in the lists of traded goods from Fisher (Reference Fisher1971) and Costa, Lains, and Miranda (2016), presumably because they were too heavy relative to their value.Footnote 33
Table 2 TRADED AND NON-TRADED CATEGORIES AND WEIGHTS
Sources: Quantities derived from Palma and Reis (Reference Palma and Reis2019), in turn based on the weighting scheme described in Allen (Reference Allen2001), with appropriate adaptations for Portugal.
Finally, we combine the traded and non-traded baskets into separate indices by using the Laspeyres method, following Allen (Reference Allen2001).Footnote 34 To smooth out volatile premodern commodity prices and examine long-term trends, we convert the indices into 11-year moving averages. Even so, spikes in the prices of heavily weighted goods remain visible in the overall picture. We follow Palma and Reis (Reference Palma and Reis2019) in increasing the fraction of maize in the grain content of flour as wheat becomes more expensive, reflecting production shares derived from tithes. Footnote 35 Finally, we construct a whole Portugal series combining the baskets from the four regions, with the traded and non-traded figures consisting of population-weighted averages of the values for each location. We depict the real exchange rate appreciation as an increase in the price of non-traded relative to traded goods.
Empirical Results
Figure 3 plots the ratio between the indexed prices of the traded and non-traded baskets from 1650 to 1825 for the entire country of Portugal. The vertical black line stands at 1694, coincident with the approximate start of gold inflows from Brazil. Several distinct phases are evident: first, a declining real exchange rate during the half-century prior to 1690; second, a steep appreciation (by over 30 percent) from around 1700 until 1720; third, a high plateau peaking after 1750; and fourth, a gradual but muted recovery after 1775. Significantly, the path of gold production in Portuguese Brazil (shown by the dashed gray line) is proportionate to the real exchange rate dynamics. The peak of Figure 3 follows soon after the peak year of output, in which more than 14,000 kilograms were extracted.
Figure 3 RATIO OF NON-TRADED TO TRADED PRICE INDICES FOR PORTUGAL, 1650–1820
Notes: The vertical line in 1694 marks the approximate beginning of gold imports to Portugal. The price index shown here corresponds to an 11-year moving average.
Sources: Price data from PWR. Gold production in Portuguese Brazil from TePaske (Reference TePaske2010).
Two trends stand out: a short, sharp appreciation followed by persistently expensive non-traded goods, which even by 1800—as gold imports dwindled—remained 20 percent above traded prices relative to 1694 levels. Close examination reveals that the recovery is largely due to the rising price of wheat during the last half of the eighteenth century. Given the high weight placed on cereal grains, despite the partial maize substitution that occurred, the movement of wheat prices led to an increase in the overall price level. However, this figure tends to overstate the extent to which the traded index rose; the price of wheat increased due to the simultaneous growth of population after 1730 and stagnation of agricultural output (concomitant with the transfer of farmland to husbandry discussed previously). Gradually substituting maize for wheat as the latter grew increasingly expensive—as is both historically accurate and theoretically plausible—mutes this distortion. Portugal’s major exports, wine and olive oil, saw essentially flat nominal prices throughout the eighteenth century. Linens—for which English textiles could be substituted—followed a similar pattern, only appreciating (as most goods did) at the time of the Napoleonic Wars. On the non-traded side, beef, pork, eggs, hens, and charcoal all rose sharply in price in the first decade of the eighteenth century, coincident with the arrival of gold imports.
The slow recovery of the real exchange rate during the second half of the eighteenth century points to a persistent effect of continuous gold imports and suggests that the Portuguese economy was slow to adjust to this shock. Elevated non-traded prices are a strong inducement for an influx of productive factors from import-competitive industries, especially as wage and price inflation squeezed profits in the traded sector. This is broadly consistent with the pattern of increasing trade deficits prior to 1750 and the subsequent collapse in product and real wages. Given the extent to which the rising price of wheat drives the recovery of the traded index, imputing the shift to any historical break is hazardous; yet the partial recovery of Portugal’s competitive position coincides with the diminution of the country’s trade deficit. It remains unclear to what extent this was simply related to lower gold inflows or Pombal’s aggressive import-substitution program. In any event, the fact that non-traded industries remained expensive into the nineteenth century aligns well with the observed trends of Portugal’s economy: by 1850, all early modern per capita income gains were lost, and income per capita was at a level similar to that of 1527 (Palma and Reis Reference Palma and Reis2019).
Figure 4 shows the real exchange rate for Lisbon over the same period. Footnote 36 Unsurprisingly, the trends correspond closely with the national version—weighting by population gives much predominance to the capital and its hinterland. Note that the population of the city of Lisbon in fact even after the 1755 earthquake remained nearly four times that of Porto, the next-largest city. Footnote 37 This is fortunate, as the price data for Lisbon is the most complete and least volatile. The Lisbon figure undergoes several distinct phase changes—a falling exchange rate before 1694, a dramatic appreciation after 1710 (by 30–40 percent), and a recovery after 1750. This should be expected of the nation’s commercial hub: Lisbon, as the principal site of gold offloading and foreign trade, would have most completely absorbed the monetary shock and had prices that more closely reflected “exogenous” world market values.
Figure 4 RATIO OF NON-TRADED TO TRADED PRICE INDICES FOR LISBON, 1650–1820
Note: The vertical line in 1694 marks the approximate beginning of gold imports to Portugal.
Sources: Price data from PWR. Gold production in Portuguese Brazil from TePaske (Reference TePaske2010).
An Augmented Synthetic Control Approach
To provide empirical evidence for this counterfactual conjecture and establish the viability of the Portuguese economy in the absence of gold, we use the augmented synthetic control method (Ben-Michael, Feller, and Rothstein Reference Ben-Michael, Feller and Rothstein2021). This allows for the construction of a counterfactual, built from a weighted average of control units that match Portugal’s pre-treatment outcomes. The estimated impact of the gold is then calculated as the difference in post-treatment outcomes between Portugal and its synthetic doppelganger. In so doing, we show that Portugal’s price level and GDP per capita diverged sharply from its neighbors during the period of the shock and that, more consequentially, this trajectory was inferior to that which would have occurred in its absence.
The synthetic control method employs a two-step approach to comparing a treated unit with a donor pool, from which components are grafted to create a plausible counterfactual. First, the treated unit is compared with the donor pool in a pretreatment period (before 1694 in our case) in order to observe which combination of units from the donor pool best replicates the behavior of the treated unit. In our case, Portugal is compared with other European countries that are used to construct a counterfactual for the post-treatment period. The procedure is designed to ensure that the outcome variables imitate both the levels and the trend of the outcome before treatment occurs. Second, after the treatment (the influx of Brazilian gold) has taken place, we observe how the behavior of the treated unit differs from that of the donor pool. The behavior of the donor pool in the post-treatment period is understood as the counterfactual of how the treated unit would have behaved in the absence of the event. Our counterfactual does not correspond to a situation in which no gold existed in Brazil, but instead to one in which Portugal was a second-stage receiver of the gold windfall. This is because the observational data for donor countries corresponds to the historical scenario in which second-stage receivers potentially benefited from the gold (Palma Reference Palma2018, Reference Palma2020; Chen, Palma, and Ward Reference Chen, Palma and Ward2022).
This procedure requires an appropriate donor pool to construct an optimal combination of weights chosen to minimize the pre-shock differences between Portugal and a synthetic counterfactual. In our analysis, this is built using long time series from a donor pool formed by several Western European countries for which such data exist: England/Britain (Broadberry et al. Reference Broadberry, Campbell, Klein, Overton and van Leeuwen2015), Italy (Malanima Reference Malanima2002, Reference Malanima2011), Germany (Pfister Reference Pfister2022), France (Ridolfi and Nuvolari Reference Ridolfi and Nuvolari2021), Spain (Prados de la Escosura, Álvarez-Nogal, and Santiago-Caballero Reference Prados de la Escosura, Carlos and Carlos2022), and Holland (van Zanden and van Leeuwen Reference van Zanden and van Leeuwen2012). Footnote 38 By design, causal inference using this method is valid in settings where an excellent fit on pre-treatment outcomes is possible (Abadie, Diamond, and Hainmueller Reference Abadie, Diamond and Hainmueller2015); when this is infeasible, the Synthetic Control Method (SCM) approach must be modified to adjust for pre-treatment fit. This leads us to the Augmented Synthetic Control Method (ASCM) on which we rely here.
A Counterfactual Portugal
We now report the results of our synthetic control analysis. We follow Ben-Michael Feller, and Rothstein (Reference Ben-Michael, Feller and Rothstein2021) and use the ASCM, which controls pre-treatment fit while minimizing extrapolation, to estimate the long-run impact of Brazilian gold on Portuguese economic development. Formally, we consider a panel data setting with i = 1,…,N units observed for t = 1,…,T time periods. Let W i = 1 be an indicator that the unit i received the treatment at time T 0 < T, and W i = 0 means that the unit i never received the treatment. To simplify the notation, we define the pre-treatment period as T 0 = 1,….,T 0, and the post-treatment period as T = T 0 + 1,…,T. Additionally, we follow the convention that i = 1 is the unit that receives the treatment. Therefore, the observed outcomes of interest are:
The treatment effect is Y
1T
(1) − Y
1T
(0). Following Ben-Michael, Feller, and Rothstein (Reference Ben-Michael, Feller and Rothstein2021), we define the ridge-augmented SCM estimator as 
, where the weights 
solve:
where γ is the vector of weights being optimized, 
is the vector of SCM weights, and λ
ridge
is the hyperparameter that determines the amount of extrapolation (with the level of imbalance). The second term in Equation (2) is introduced in ASCM equations to penalize deviations from the SCM weights.
Footnote 39
We follow the cross-validation approach proposed by Ben-Michael, Feller, and Rothstein (Reference Ben-Michael, Feller and Rothstein2021) and the “one-standard-error” rule (Hastie, Tibshirani, and Friedman Reference Hastie, Tibshirani and Friedman2009) to select λ
ridge
. We show the resulting weights in the Online Appendix (Table B.3).
In Figure 5, the price level is shown to increase significantly above the synthetic counterfactual immediately in the decades following the chosen treatment date of 1694, concurrent with the growing influx of gold imports. In turn, Figure 6 shows a dramatic divergence between the synthetic and actual GDP measures for Portugal during the second half of the eighteenth century. Footnote 40 GDP declined in absolute terms during much of the second half of the century and ended up over 200 percent lower than the counterfactual path in 1800. Finally, Figure 7 shows the trajectories of the synthetic counterfactual and actual GDP per capita series for Portugal for the treatment, pre-treatment, and post-treatment periods. As predicted, the two corresponded closely prior to the gold shock. During the treatment period, however, distinct divergences appear in the actual and counterfactual outcomes. Following a brief dip after 1700, the historical series booms during the early years of the gold shock, exceeding the actual series around 1750. From this peak, per capita income tumbled sharply, falling well below the simulated trend line by the end of the century.
Figure 5 SYNTHETIC CONTROL RESULTS FOR PORTUGUESE CPI, 1640–1800
Notes: The vertical line in 1694 marks the approximate beginning of gold imports to Portugal. As a robustness check, in Online Appendix Figure B.21 we show the same figure using an unweighted mean, rather than the synthetic counterfactual, and demonstrate that the result is largely unaffected.
Source: See Table 1.
Figure 6 SYNTHETIC CONTROL RESULTS FOR PORTUGUESE GDP, 1640–1800
Notes: The vertical line in 1694 marks the approximate beginning of gold imports to Portugal. As a robustness check, in Online Appendix Figure B.22 we show the same figure using an unweighted mean, rather than the synthetic counterfactual, and demonstrate that the result is largely unaffected.
Source: See Table 1.
Figure 7 SYNTHETIC CONTROL RESULTS FOR PORTUGUESE GDP PER CAPITA, 1640–1800
Notes: The vertical line in 1694 marks the approximate beginning of gold imports to Portugal. As a robustness check, in Online Appendix Figure B.23 we show the same figure using an unweighted mean, rather than the synthetic counterfactual, and demonstrate that the result is largely unaffected.
Source: See Table 1.
Such patterns precisely fit expectations for an economy beset by a resource curse. During the initial decades of the gold inflows, the economy grew faster than the counterfactual, bolstered by rising real wages due to the influx of imported gold. From the mid-century, however, the accumulated structural weaknesses began to manifest, and stagnation from the mid-1750s was soon followed by outright decline from the 1770s (Palma and Reis Reference Palma and Reis2019). By 1800, GDP per capita was over 40 percent below the counterfactual, as previously mentioned: with the difference between the two series being both economically large and statistically significant.Footnote 41
THE GOLDEN ORIGINS OF PORTUGAL’S DIVERGENCE
The movements exhibited in relative prices and the real exchange rate align with predictions from the Dutch disease theory. In turn, the worsening political environment during the eighteenth century is consistent with an institutional resource curse. It remains to connect these factors more closely with Portugal’s observed economic decline and divergence from Western Europe. The causes of the ailment—whose symptoms included trade deficits and stagnating output—fall into proximate and ultimate categories. The former comprises real exchange rate appreciation, industrial decline, and institutional deterioration; these were driven by colonial gold inflows through a resource curse mechanism. While real exchange rate appreciation, increasing trade deficits, and political challenges are characteristic results of a resource curse, further work will be needed to understand the effects of the gold boom on Portuguese economic performance.
Annual output data are unavailable for most individual industries except agriculture, so a variety of substitute measures must be employed, in particular, to differentiate between the direct consequences of gold via its effect on the economy and its indirect effect via a political resource curse that negatively affects institutions. Footnote 42 Notice however, that the latter effect is not independent of the former: the expansion of the non-tradables sector (e.g., landed and Church-related interests) led to the formation of powerful lobbies as had happened earlier in Spain (Charotti, Palma, and Santos 2022). In Portugal, during the first half of the eighteenth century, King João V (r. 1706–1750) spent a significant proportion of the tax revenues generated by the Brazil windfall on frivolous expenses, including the construction of the huge Palace-Convent of Mafra and an expensive and ostentatious embassy to Pope Clement XI. During this time, there was limited interest in promoting domestic manufacturing, with political priorities focused on maximizing revenues related to gold extraction in Brazil (Pedreira Reference Pedreira1994, pp. 41–42, 2005).
As discussed earlier, between 1720 and 1750, the Portuguese trade deficit rose from 2 billion reis (2.4 percent of nominal GDP) to more than 4 billion (3.9 percent), reaching nearly 6 billion (6.1 percent) in 1756 in the wake of the Lisbon earthquake. Footnote 43 The foreign accounts remained in the negative until the last decade of the eighteenth century, notwithstanding the efforts of Pombal and his successors after 1777 to boost domestic industry and prohibit imports. In any event, Costa, Lains, and Miranda (2016, p. 197) suggest that the restoration of the trade balance was the result of falling domestic consumption, given that exports did not substantially increase. This is also consistent with the evidence of Palma and Reis (Reference Palma and Reis2019) regarding falling real incomes in the last decades of the eighteenth century, well before the Napoleonic invasion in 1807. Footnote 44 Even the suppression of Brazilian manufacturing by a decree of 5 January 1785, and Pombal’s earlier efforts to exclude English “interlopers” from the colonial trade through the assignment of monopoly rights failed to lift exports considerably, and if their value rose at the end of the century, this was at least partly the result of a pan-European wave of price inflation. So long as the golden tide flowed into Lisbon, Portugal was doomed to involution.Footnote 45
Did individual traded and non-traded industries fit the general pattern? Not much reliable data exists on manufacturing, but what limited figures can be adduced are not suggestive of general expansion. In 1769, only 15 factories existed across the country at large; just four of these produced textiles, the growth sector of the British Industrial Revolution (Pedreira Reference Pedreira1994, p. 59). Metallurgy, silk throwing, and dyeing were all neglected, while cotton, woolens, and linen each received only one new venture. Given that between 1736 and 1740 Portugal was purchasing half of the English textile industry’s exports (Fisher Reference Fisher1971, p. 144), demand for manufactured goods clearly outstripped the supply response of domestic industry. There was no push, as in England, toward modernization and mechanization; rather, the traditional putting-out system that had persisted since the sixteenth century was nearly universal. Larger-scale operations—either state-run or state-sponsored—merely subcontracted tasks to peasant production networks. Only by the turn of the century, once gold imports had nearly ceased, does any evidence appear of output and exports marginally increasing. However, the picture was not optimistic, given the low base from which industry restarted, and the fact that Brazil—subjected to a colonial monopoly until 1808—was the primary recipient of Portugal’s manufactures. The share of population outside agriculture, which had risen quickly until 1750, then fell from 46.5 percent in 1750 to 33.1 percent a century later (Palma and Reis Reference Palma and Reis2019). Footnote 46 The technical stagnation noted previously, therefore, would be a predictable consequence of this contraction. While Lisbon and Porto did experience strong urban growth during the eighteenth century, they were “consumption cities” dependent on the provision of non-traded services. Gollin, Jedwab, and Vollrath (Reference Gollin, Jedwab and Vollrath2016) find that consumption cities, by contrast with manufacturing-based “production cities,” tend to proliferate in countries with high levels of natural resource exports. Lisbon, home to the Crown and Court, proved a particularly voracious consumer of goods funded by gold imports. While aggregate demand increased, Lisbon’s status as a hub for foreign trade ensured that much of this was siphoned off into purchases of foreign manufactures (Fisher Reference Fisher1963).
Viticulture is a potential confounding factor. Both wine and olive oil saw increased production, with the former doubling in cultivated areas during the 1700s. Over 70 percent of wine output was exported to England, while olive oil was also traded, especially with Brazil. The foreign demand for these products boosted wine prices throughout the latter half of the eighteenth century. Yet the nature of the industry should give the skeptic pause. First, port wine had no close international substitutes; beer was not treated as such, as evidenced by the behavior of British consumers, while the Methuen Treaty undercut French competition. Thus, even if costs increased and opportunities arose to switch capital and labor into the non-traded sector, wine producers were able to compensate with higher prices. Diplomacy protected the winemaker. Furthermore, the growth of a traded subsector is consistent with Dutch disease. According to the Corden-Neary decomposed “Paradox Model” (see Online Appendix A), the most capital-intensive lagging industries might raise production at the expense of competing trades. Wine was surprisingly capital-intensive, and small farmers were unable to muster the investment necessary to compete with the nobility and urban elites who dominated the sector. Thus, the decrease in available labor might well have tended to increase the output of viticulture, and the evidence suggests that both cereal and wasteland were extensively converted for this purpose.
Cereal production, meanwhile, suffered an unambiguous decrease in output during the second half of the eighteenth century (Reis Reference Reis2017). Total factor productivity fell over this period, and yields remained both low and stagnant by European standards (averaging 1:4, and never exceeding 1:7). Though the introduction of maize proved increasingly successful, Portugal continued to import increasing amounts of grain, peaking at 151.6 tons worth (as discussed earlier) by 1800. Lisbon’s foreign consumption rose from an already-high 55 percent in 1729 to a colossal 72 percent by 1778 (Costa, Lains, and Miranda 2016, p. 183). Though this figure was smaller for the country at large (5.5 to 7 percent), it was still considerably more than England’s 3 percent, and was impressive in an age of debilitating overland transport costs. Indeed, 5 to 12 percent of Britain’s rising exports to Portugal were consisted of wheat throughout the period (Fisher Reference Fisher1963, p. 222). Due to low productivity and profits, vast tracts of land were switched over to pastoral husbandry, with grain production restricted to poor soils (Costa, Lains, and Miranda 2016, p. 180). Since the raising of cattle and sheep is more capital-intensive, this outcome should be expected within a Dutch disease model. Furthermore, the contraction of the sector would have hindered productivity advances, and the changes in techniques that revolutionized agriculture in eighteenth-century England were absent in contemporary Portugal.
Resource Curse: Economic and Political Consequences
Gold directly augmented incomes—of some people, and in the short run. Increasing gold production in Brazil enriched Portuguese nationals in the colonies, who either remitted their funds home or exchanged them directly for the durable goods arriving on the thrice-annual treasure fleets. Footnote 47 Newly expanded incomes would necessarily increase demand in the non-traded sector, causing a real appreciation and a consequent withdrawal of resources from the lagging sector through the spending effect. Footnote 48 A smaller though still significant contraction might follow from the emigration of hundreds of thousands of Portuguese prospectors and other individuals working on their supply lines, which would constitute a resource movement effect. With the collapse of the traded export sector and the increased purchasing power of the currency, exports would decline and imports surge, increasing the trade deficit. Gold re-exportation would then be required to pay the outstanding balances, as the proceeds of domestic industry no longer earn foreign exchange.
In the long run, the decades-long influx of gold would have continued to compress the traded sector. Skill-based productivity advances, which depended on industry and cereal agriculture, were persistently stifled, removing the principal driver of expansion. After an initial boom, therefore, GDP growth rates slowed, and eventually reversed, with critical sectors of the economy being replaced by imports and with the practical consequences of worsening political institutions. Though wine prices and exports apparently flourished, this can be reconciled with the “decomposed” traded-sector model discussed in Online Appendix A, with land as an additional mobile factor. This aligns with the general pattern of farms being transferred to relatively land- and capital-intensive industries, especially viticulture and husbandry, throughout the course of the eighteenth century. Costa and Reis (Reference Costa and Reis2017) estimate that the 1750–1800 period was the only extended span in Portuguese history in which there was a trade deficit in the agricultural sector. The growing success of English textiles in both the home and Brazilian markets, meanwhile, indicates that domestic production was increasingly not competitive.
As noted in the second section, Portugal’s increased efforts to sustainably boost industrial production failed. Yet, the implementation of import-substitution policies during the second half of the eighteenth century hints at contemporary recognition by Pombal that gold imports (and re-exportation) led to economic involution. Some writers of the period were in fact at least vaguely aware of the negative consequences of gold imports (Antonil Reference Antonil2011; da Silva Reference da Silva2013). Footnote 49 The tariffs, bans, and state entrepreneurship of the Pombal era were coupled with the capture and protection of the Brazilian market on behalf of domestic merchants. In the second half of the century, manufacturing in the colony was banned outright, and, as mentioned, efforts were made to reduce the influence of “interloper” English shippers and producers by granting monopoly rights to chartered companies. The Company of Grão-Pará and Maranhão exploited state credit to establish primary product plantations in cacao, rice, coffee, and cotton, for which they exchanged manufactured exports, which rose to 43.1 percent of all goods sent to the colony by 1800 (Costa, Lains, and Miranda 2016, pp. 206–08). There is no basis for attributing Portugal’s economic ineptitude to a historically poor representation of merchant interests due to comparatively backward political institutions going back to 1500 or earlier (Henriques and Palma Reference Henriques and Palma2023). But it is nonetheless true that the merchant push of the Pombal era was state-led and associated with rent-seeking, crony capitalism, and state capture (Madureira Reference Madureira1997; Costa, Lains, and Miranda 2016, pp. 215–16). Accordingly, negative structural transformation occurred despite concerted state efforts to boost industry at the expense of Brazil. Overall, the resource curse meant that Portugal’s industry not only lost competitiveness but also missed out on the expansion opportunities that, in the absence of Brazilian gold, would have occurred in the eighteenth century and likely continued via agglomeration and other spillover effects (e.g., related to human capital) into the future.
In addition to the direct economic consequences of de-industrialization and missed agglomeration opportunities, Portugal also suffered a political resource curse due to Brazil’s gold. Footnote 50 As late as the second half of the seventeenth century, Portugal’s political institutions, namely courts and parliaments (Cortes), had limited the power of the ruler (Henriques and Palma Reference Henriques and Palma2023). The additional revenues of the Crown from the Brazilian gold windfall and the military protection guaranteed by the Methuen Treaty of 1703, which was itself a result of this, meant that rulers no longer needed to negotiate, and parliaments stopped meeting altogether, leading to reduced executive constraints and a lack of checks and balances.Footnote 51
Previously active in providing checks and balances against executive power, the Cortes never met during the entire eighteenth century. Power accordingly became absolutist, with consequences that peaked during the rule of Sebastião José de Carvalho e Melo (1699–1782), who became Count of Oeiras in 1759 and Marquis of Pombal in 1770. He was the de facto ruler from around 1755 to the death of King José I in 1777. Perhaps most disastrously, Pombal’s educational reforms in association with the expulsion of the Jesuits, who opposed his absolutism, drastically reduced the number of students, and the industrialization effort he promoted was ultimately unsuccessful (Madureira Reference Madureira1997; Macedo Reference Macedo1982a; Pedreira Reference Pedreira2005, pp. 205–06, Romeiras and Leitão Reference Romeiras and Leitão2022). Pombal was responsible for large numbers of political prisoners and for instrumentalizing the Inquisition to pursue his political goals, such as by placing his brother at its head. Footnote 52 He also centralized censorship in his hands (Real Mesa Censória). Footnote 53 His rule was characterized by organized state predation and capture, putting an extractive system in place that continued after his fall in 1777.Footnote 54
Pombal’s unchecked policies were disastrous for human capital formation. There were around 20,000 students in pre-university education until 1759, a figure that Pombal’s reforms from 1759 reduced sharply (Leitão Reference Leitão2007, p. 88). Footnote 55 The same absolute number of pre-university students was only reached again in the 1930s, when the overall population was considerably larger (Romeiras Reference Romeiras2019; Leitão and Romeiras Reference Leitão and Romeiras2015; Palma, Reis, and Zhang Reference Palma, Reis and Zhang2020). Portugal still had a numeracy rate close to the most advanced countries in Europe during the first half of the eighteenth century, but this comparative situation would change drastically in the second half of the eighteenth century (Stolz, Baten, and Reis 2013, pp. 562–64). By 1773, less than a third of the existing primary schools had a teacher (Fernandes Reference Fernandes2006, p. 608). By 1800, literacy rates in Portugal—measured by the percentage of adults who could sign their names—were well behind other parts of Western Europe (Reis Reference Reis2005, p. 202).Footnote 56 The fact that it was the nature of the existing political institutions and not the Catholic religion, which by itself prevented industrialization is suggested by the fact that the second European country to industrialize and enter modern economic growth was Belgium. Footnote 57 The lack of explanatory power of the Catholic religion is also reflected in the absence of observable differences with respect to other variables. For example, there is no evidence of different social norms at the family level, meaning that women were more discriminated against in Southwestern Europe than in Northwestern Europe—a factor that could have led to different fertility practices causing low human capital accumulation and other inferior development outcomes (Palma, Reis, and Rodrigues Reference Palma, Reis and Rodrigues2023). Finally, recent research has shown that Portugal had a high level of state capacity, even when compared with other Western European polities (Costa, Henriques, and Palma Reference Costa, Henriques and Palma2024); hence, its problem lies elsewhere.
CONCLUSION
The decline of Portugal was due to the structural effects of massive gold imports from Brazil, which peaked around the 1740s. In this paper, we have focused on this resource curse. It had an economic aspect (the de-industrialization of the economy) and a complementary political resource curse operating through channels such as additional royal revenues. On the economic front, increasing imports of gold coincided with a strong appreciation of the real exchange rate, which only began to recover with the decline of remittances during the 1770s. Price data obtained from archival sources shows a sharp appreciation of the real exchange rate consistent with the onset of Dutch disease and the loss of competitiveness of the national industry. The prices of traded goods remained for the most part stagnant, except for wheat, whose price alone among grains tended upward during long spells of the eighteenth century. Imports rose throughout the first half of the eighteenth century, with gold remittances and the deficit peaking simultaneously, while exports remained stagnant or declined outright, despite deliberate efforts at import substitution. In sum, there was a sharp rise in the prices of the non-traded sector—chiefly primary products—leading to withdrawn resources from the traded sector, and hence reduced output and employment in manufacturing and grain agriculture. While the economy boomed in the short run, the longer-term effects of the gold shock were negative—contractions in industry and cereal production slowed the accumulation of technical progress, causing stagnant growth in subsequent years. Income growth was permanently reduced. Over the second half of the eighteenth century, the country de-industrialized, and the percentage of the population working outside of agriculture declined from around 46 percent in 1750 to only 33 percent a year later.
In proffering the resource curse as a cause of Portuguese retrogression, we also hope to show that initial institutions were not to blame for this Iberian episode of the “Little Divergence.” Contrary to the arguments presented in Acemoglu, Johnson, and Robinson (Reference Acemoglu, Johnson and Robinson2005), Portugal was ruled by a political system not meaningfully distinguishable from England’s until the second half of the seventeenth century (Henriques and Palma Reference Henriques and Palma2023). Portugal’s culture and family-level organization were also not different in relevant dimensions from that of England or the Netherlands (Palma, Reis, and Rodrigues Reference Palma, Reis and Rodrigues2023). Instead, a resource curse provides a compelling alternative conception of the Portuguese problem. Dutch disease caused industry to become less competitive. With access to additional revenues from gold, the Crown no longer needed to negotiate, meaning that, unlike in previous centuries, the Parliament did not meet in the eighteenth century, with power becoming absolute (Palma Reference Palma2020; Henriques and Palma Reference Henriques and Palma2023). The resource curse interrupted the economic and institutional improvement that had been taking place in Portugal during the second half of the seventeenth century. Given the backwardness of institutions and human capital, Portugal’s continued divergence was by then inevitable. Once the gold was gone, Portugal was left with no viable industry, and with a backward institutional and educational system. It then systematically diverged relative to Western Europe, until a partial recovery began to take place only from around the mid-twentieth century (Palma and Reis Reference Palma and Reis2019, Reference Palma and Reis2021; Amaral Reference Amaral2019).
Open access
