What Needs Explaining?

Latin America was not underdeveloped by any conventional measure (like GDP per capita) until sometime between roughly 1750 and 1850 (see Table 1). The regions conquered and colonised by Europeans attained unprecedented levels of productivity during the colonial era.Footnote ² In the first century after the European invasions, productivity gains occurred chiefly in the mountainous interiors of the viceroyalties of New Spain and Peru, where European technology and organisation succeeded in creating an entirely new industry, deep-shaft mining and ore processing.Footnote ³ Substantial gains in productivity were also achieved in Mesoamerica and the Andes in domestic-use agriculture. As European disease and abuse decimated indigenous populations, per capita output rose.Footnote ⁴ The demographic catastrophe facilitated the introduction of European crops and livestock. Additional gains in productivity occurred when those who survived found work, resettled, or were ‘congregated’ on the most productive lands. Over the long run, however, the productivity gains associated with the first century of Spanish rule in the Americas tended to diminish. Peruvian silver production collapsed at the end of the sixteenth century and never fully recovered. Mexican production stagnated for most of the seventeenth century, falling by half in per capita terms as population recovered.Footnote ⁵ At the end of the colonial era precious metals production, still based largely on seventeenth-century technology, represented less than 10 per cent of GDP in Mexico and the Andes.Footnote ⁶ The revival of the indigenous population had by then put an end to rising productivity in agriculture as well.

Table 1. GDP per capita, 1500–2001 (in Geary-Khamis International Dollars of 1990)

Sources: Maddison, The World Economy, op. cit. except for Mexico, in John Coatsworth, ‘Mexico’ in Joel Mokyr (ed.), The Oxford Encyclopedia of Economic History (New York 2003), pp. 501–7; on Colombia 1800, from Salomon Kalmanovitz, ‘El PIB de la Nueva Granada en 1800: El auge colonial, estancamiento republicano’, Revista de Economía Institucional, vol. 8, no. 15, (2006), pp. 161–83; and other 1800 from John H. Coatsworth, ‘Economic and Institutional Trajectories in Nineteenth-Century Latin America’ in John H. Coatworth and Alan Taylor (eds.), Latin America and the World Economy Since 1800 (Cambridge, MA, 1998), pp. 23–54.

Notes: (1) Mexico only in 1500 and 1600; 1700 is the unweighted average of Cuba and Mexico; 1800 includes Argentina, Brazil, Chile, Colombia, Cuba, Mexico, and Peru; thereafter Argentina, Brazil, Chile, Colombia, Mexico, Peru, Uruguay, and Venezuela. (2) Includes Albania, Bulgaria, Czechoslovakia, Hungary, Poland, Romania, Yugoslavia. (3) 57 countries. (4) 1929.

As the initial productivity gains from the Spanish conquest began to diminish in the early seventeenth century, European technology and organisation achieved new and even greater successes with the introduction of cane sugar and the importation of millions of enslaved Africans to grow and process it. By the late seventeenth century, the most productive economies in Latin America, and perhaps the world, were the slave plantation islands in the Caribbean where sugar exports accounted for a large proportion of GDP (as much as 30 to 40 per cent), and GDP per capita reached or exceeded European levels.Footnote ⁷ By the late eighteenth century, temperate zone colonies like Argentina and Chile were also doing well with predominantly European populations lured to the New World by the promise of high wages and relatively easy access to land.Footnote ⁸ In the frontier zones on the periphery of European settlement, missions organised by Jesuit priests induced thousands of indigenous foragers into productivity-enhancing transitions to sedentary agriculture, although (as is now known) this change precipitated severe declines in the biological welfare of the affected populations.Footnote ⁹ The European impact may have been more positive beyond the frontier, where indigenous nomads from the pampas to the Great Plains tamed wild horses, descended originally from animals that had escaped from European herds, and used them to increase the efficiency of their hunts. By the mid-nineteenth century, the Plains Indians in the United States, descendants of the reduced number of indigenous people who survived European diseases, were among the tallest and thus apparently the best nourished people in the world.Footnote ¹⁰

To summarise, we now know that three historic transformations – the opening of the Americas to trade with the rest of the world; the transfer of European technology, organisation, flora and fauna; and two major demographic catastrophes (the destruction of indigenous populations and the forced migration of millions of Africans) – raised per capita GDP in the Americas well above the achievements of the most advanced pre-Columbian societies.Footnote ¹¹ Once the new European enterprises were established, however, productivity tended to stagnate, with fluctuations in output largely determined by natural resource discoveries or volatile world export demand.Footnote ¹² Conquest and colonisation thus produced what economists sometimes refer to as Smithian growth, that is, gains from trade based on adapting existing technology and organisation to exploit accessible natural resources. Latin America's newly discovered comparative advantages produced an impressive growth in productivity. For 1700, we have rough estimates of GDP per capita for Argentina, Barbados, Cuba and Mexico.Footnote ¹³ All but Mexico had more productive economies than that of British North America, and Mexico was only slightly behind.

Much of the recent work on the political economy of the Americas concludes that Iberian colonialism failed to create dynamic societies that could independently generate technological or organisational innovation.Footnote ¹⁴ As long as economic performance in the rest of the world remained similarly constrained, Latin America's abundant and accessible natural resources kept it at, or near, the top of the heap. Between the mid-eighteenth century and the second half of the nineteenth century, however, most of Latin America fell behind the North Atlantic economies. At first, the Anglo-American advantage may have come largely from the cumulative effects of institutional changes that made property rights more secure and increased the efficiency of market transactions.Footnote ¹⁵ By the end of the eighteenth century, however, the first Industrial Revolution was pushing GDP per capita to heights beyond levels ever before attained, first in Britain and then in the United States and much of north-western Europe. All the mainland colonies except for Argentina had fallen well behind by 1800. Cuba, with its sugar economy booming, kept pace with the US economy for another quarter of a century.

Most of Latin America could not keep up and would not have been able to do so, even without the costly independence wars that wrought havoc from 1810 to the 1820s. As the Industrial Revolution accelerated, Latin America stagnated. By 1850, Latin America's PPP-adjusted GDP per capita had fallen to only a quarter that of the United States and Britain. In the short span of a century or so, Latin America had become ‘underdeveloped’.

Then, happily, sometime between the 1850s and the end of the nineteenth century, most of Latin America began growing and continued to do so for the next century at roughly the rate of the mature economies of the North Atlantic. Angus Maddison's estimates show that in 2000 the relative gap in GDP per capita between the United States and the eight largest economies of Latin America was about what it had been in 1900, though taking into account slower growth in the smaller economies throughout the twentieth century and slower growth everywhere after 1982 raises the gap from slightly over four to one to nearly five to one today.Footnote ¹⁶

These are the long term trends that the new work on the political economy of Latin American development seeks to explain. Why did Latin America fall behind between roughly 1700 and 1850? How did sustained economic growth get started in the late nineteenth century? And why has Latin America's economic growth been so slow -- no faster than that of the United States – over the past 100 years. Since the new answers we are getting to the second and third of these questions depend to a large extent on how the first is answered, this paper will concentrate on the first question, leaving the last century or so to brief comments at the end.

The Fall from Development

Let us start, then, with how and why Latin America fell from economic grace between roughly 1750 and 1850. Three different, but overlapping and partially complementary, theories seem to be on offer. Each seeks to explain contemporary backwardness by contrasting Latin America's troubled past with that of other, more successful former colonies, chiefly the United States.Footnote ¹⁷ The first theory, elaborated by Stanley Engerman and Kenneth Sokoloff in a seminal essay, relies chiefly on differences in natural resource endowments.Footnote ¹⁸ The second, offered in a series of influential papers by Daron Acemoglu, Simon Johnson and James Robinson, explains Latin America's difficulties as the result of the ‘extractive institutions’ imposed by European colonialism on subject populations.Footnote ¹⁹ Both the natural resource and colonialism models posit the early development in Latin America of unequal societies in which small elites of European settlers created institutions that failed to protect the property rights of majorities, in contrast to the robust protections secured by European settlers in British North America and other areas where European settlers constituted the majority of colonial populations. A third theory, recently reformulated by James Lange, James Mahoney and Matthias vom Hau, argues against ‘an exclusive focus on domestic conditions within colonies’ and focuses on ‘the identity of the colonizing nation’ expressed in the ‘differences in the economic models’ imposed by Britain and Spain on their colonies.Footnote ²⁰

The first theory, that of Engerman and Sokoloff, makes the claim that differences in factor endowments, that is, both natural resources and labour supply (the latter partly a consequence of imperial immigration and labour mobilisation policies), produced high concentrations of land ownership in the Spanish and Portuguese colonies. Their essay, however, tends to treat labour flows to the Americas (as well as capital flows) as responsive to opportunities for profitable exploitation of natural resources, so their argument relies mainly on geography rather than policy. They argue that exploiting Latin America's natural resources required the creation of large slave plantations in the tropics and large grain and cattle haciendas elsewhere. The resulting concentrations of wealth in the Iberian colonies led to the creation of unequal societies in which settler elites exploited the majority indigenous or imported slave populations. To facilitate these arrangements, colonial regimes created appropriately exploitative institutions which persisted long past the end of colonial rule. These colonial institutions protected the property rights of elites, but denied such protections to powerless subordinate classes. As Engerman and Sokoloff put it, ‘the relatively small fractions of their populations composed of whites, as well as their highly unequal distribution of wealth, may have contributed to the evolution of political, legal, and economic institutions that were less favorable toward full participation in the commercial economy by a broad spectrum of the population’.Footnote ²¹ In contrast, the natural resource endowments of the Mid-Atlantic and New England colonies in British North America favoured small-scale mixed agricultural enterprises, a more egalitarian distribution of wealth (especially land), and thus the creation of institutions that protected the property and human rights of broad sectors of the population.

The Engerman and Sokoloff reliance on natural resource endowments to explain alleged differences in the distribution of wealth between Latin America and British North America has not held up well to scrutiny. Resource endowments and wealth distribution varied as much within as between Latin America and North America. Sugar plantations tended to be large in Cuba and small in Brazil. Grain estates in central Mexico produced the same mix of crops and animals as the small farms of Pennsylvania. In the southern slave colonies of British North America, land ownership near the coast and along rivers tended to be high, in contrast to frontier zones everywhere and the New England colonies. In much of the interior of Latin America peasant agriculture persisted and even expanded well into the nineteenth century, while in other regions large cattle estates sometimes covered thousands or even millions of hectares. Moreover, what little quantitative evidence there is does not suggest that ownership of land, or other assets for that matter, was more concentrated in Latin America than in the United States (see Table 2). Lyman Johnson and Zephyr Frank, for example, reached the conclusion in their analysis of Rio de Janeiro and Buenos Aires in the early nineteenth century that ‘wealth inequality was not significantly higher in the cities of the South Atlantic prior to 1860 when compared with cities in the North Atlantic’.Footnote ²²

Table 2. Distribution of wealth, 18th–19th centuries

Using other measures of inequality does not work either. For example, class differences in net nutrition, as reflected in the height attained by adult males, were probably greater in Britain, the seat of the Industrial Revolution, than anywhere in the western hemisphere in the early nineteenth century. At that time the adult male members of the English titled aristocracy stood five inches (12.5 centimetres) taller than the population as a whole.Footnote ²³ None of the growing number of studies on the biological standard of living in Latin America has turned up class differences in net nutrition as large this either for the colonial era or the nineteenth century.Footnote ²⁴

The second theory, offered by Acemoglu, Johnson and Robinson, starts differently but relies on a similar institutional argument. Inequality was crucial, they assert, but it originated not in the geography of natural resource endowments and economic inequalities, but in historic inequalities of status and power. Wherever small settler elites dominated large populations of native or African descent, they created ‘extractive institutions’ that deliberately excluded majorities from power and failed to protect their property and human rights. As in Engerman and Sokoloff, the bad institutions created in such circumstances tended to persist through long periods of time – through half a millennium in fact. Unlike Engerman and Sokoloff, though, Acemoglu and his colleagues subjected their hypothesis to statistical testing in two fascinating papers. The first showed a significant inverse correlation between contemporary GDP per capita and mortality rates in colonial regions. Areas of high settler mortality, they argued, tended to attract a small number of Europeans willing to risk death for the opportunity to exploit local populations. These areas developed ‘extractive’ institutions inimical to long term economic growth. Low mortality regions, on the other hand, attracted larger numbers of European settlers, who soon outnumbered indigenous populations and insisted on institutions that did a better job of protecting the rights of most (that is, European) citizens. A second paper by the same authors took a different tack, showing that a ‘reversal of fortune’ occurred between 1500 and the twentieth century, as formerly ‘developed’ regions of the non-European world (measured mainly by urbanisation rates) became objects of plunder and conquest, and thus the imposition of bad institutions, while less developed regions in 1500, with their majority settler populations, became the most developed 500 years later.

To sum up, these two theories tend to the conclusion that Latin Americans were expelled from the Garden of Developed Economies because of the original institutional sins of their European rulers and exploiters, and because of their subsequent failure to repent. This conclusion has considerable aesthetic – if not theological – appeal. Logic and evidence, however, make it problematic. The rough estimates of GDP per capita for Argentina, Brazil, Chile, Cuba, Mexico and Peru in 1800 (see Table 1) suggest that productivity differences between the richest and poorest Latin American colonies in 1800 were nearly as great as for the richest and poorest regions of the entire world, that is, a ratio of roughly three to one. This would not be so inconvenient for the institutionalists were it not for the fact that variations in legal and political inequality seem to have been entirely unrelated to economic success or failure. Cuba, with its increasing commitment to slavery, was the wealthiest of the colonies in 1800, while Brazil, also a slave colony, was among the poorest. Argentina, with its small population composed mainly of European migrants who were attracted by high wages and access to land, fits the model well, but its GDP per capita in 1800 was below that of Cuba.

It appears, therefore, that if colonial institutions constrained economic development, they did not do so until after the colonial era. Acemoglu, Johnson and Robinson offer a way round this conundrum by suggesting that, bad as they were, colonial institutions did not impede economic performance until the end of the colonial era. In the early nineteenth century, they argue, Latin America faced an ‘opportunity to industrialize’, but failed to seize this opportunity because to do so would have required security of property rights for a broad cross section of the population. ‘Early industrialization’, they say, ‘appears to require both investments from a large number of people who were not previously part of the ruling elite and the emergence of new entrepreneurs’.Footnote ²⁵ In this interpretation, Latin Americans are permitted to stay in the Garden for a time after tasting the forbidden fruit, but must leave later on, when they are too busy consuming apples, as it were, to learn how to drive tractors and apply artificial fertilisers.

Before embracing this laudably egalitarian heresy, however, it would be useful to take a closer look at the failure of the Latin American economies to respond to the alleged opportunities that the Industrial Revolution made available to them. There are three problems with the Acemoglu, Johnson and Robinson explanation: the omitted variable (geography), the missing villain (settler elites), and the wrong time (by the time of the Industrial Revolution, it was already too late).

First, geography imposed limits on the potential for economic growth throughout most of Latin America. Acemoglu et al. claim to have refuted this proposition by showing that the reversal of fortune occurred without any corresponding change in landscape, climate, disease environment, or natural resources. But if geography did not change, technology did. The concentration of pre-Columbian populations in the highlands occurred because the disease environment was less deadly and the soils richer in minerals, both of which helped cities grow larger. But globalisation after 1492 devalued these advantages and made others far more important. The Industrial Revolution began in England, where the most remote places were located no further than thirty miles from the coast or a navigable waterway. It then spread to a former colony, the United States, whose river systems, later supplemented with canals, still carry immense quantities of freight among regions and out to rest of the world. Latin America had no comparable advantages in water transport, except in the vast Amazon basin where the thin soils and lack of mineral resources made commercial development unprofitable. Sugar plantations in the Americas had to be located no more than fifteen miles from a port, which is why sugar was most abundantly and profitably produced on islands. There were too few navigable waterways to make the exploitation of most of Latin America's natural resources feasible until new transport technologies, mainly railroads and steamships, appeared in the nineteenth century. In this sense, natural resource endowments were, indeed, crucial to the growth, or lack of growth of the colonies.

Second, the Spanish and Portuguese colonial societies were less unequal, social hierarchies far more contested, and settler elites therefore far less powerful than is suggested in the new work, and much of the old historiography. In the Amerindian colonies of Mesoamerica and the Andes, for example, the creole settler elite did not control the indigenous peasantry's access to land. Colonial officials and magistrates were explicitly and repeatedly instructed to issue land titles to indigenous villages and to defend village lands against usurpation and encroachment. In contrast to Britain and the European continent, most of the arable land in the predominantly indigenous colonies of Spanish America belonged to indigenous peasants and their communities until long after independence. In the Andes, moreover, most of the highland pastures with their immense llama herds also remained under indigenous ownership and control throughout the colonial era.Footnote ²⁶ Even on the slave plantations of Brazil and the Caribbean, rebellion and resistance were endemic.

Nor did settler elites dominate colonial administrations, control policy making, or shape institutions just as they pleased. Surrounded by indigenous populations with independent access to subsistence, or by frequently rebellious slaves on the tropical coasts and islands, economic elites in the Americas had little choice but to seek protection in exchange for loyalty. This historic compromise between colonial elites and the imperial state constituted one of the keys to the longevity of Spanish and Portuguese rule in the Americas. Colonial elites endured, mostly in silence, an imperial regime that occasionally expropriated wealth by defaulting on the public debt, debasing the currency, and imposing forced loans and predatory taxes, especially in time of war. They acquiesced in the colonial trade monopolies that forced all trade to pass through Spain or Portugal. They paid consumption and excise taxes, tolerated royal monopolies (estancos), conformed to endless regulations for fiscal and other purposes, and had no choice but to accept a legal system that failed to define property rights clearly or to provide an efficient court system to enforce them. These hardships imposed on colonial ‘elites’ were at least as consequential for economic performance as those imposed on the majority populations of the Americas, since most of the colonial population earned too little to invest in the economic future of their societies.

Linking Latin America's contemporary backwardness to colonial inequality also assumes some mechanism of continuity, or path dependence. The highly unstable fortunes of the continuously revolving settler elites of the colonial era and the nineteenth century do not appear to have constituted such a mechanism. Membership in the colonial elites, and the accumulated wealth that defined them, were both highly unstable. In periods of rising urban demand for food and reasonably secure conditions in the countryside, the ownership of rural estates tended to be stable; when the surrounding conditions turned unfavourable, estate ownership tended to revolve and the estates themselves often fragmented into smaller units. For example, in the Valley of Oaxaca, an area of relatively stable hacienda boundaries and ownership patterns, William Taylor's data (see Table 3) showed that haciendas were sold to new owners on average only once every 37.1 years between 1580 and 1639, but changed hands much more frequently during the succeeding century – once every 14 years from 1640 to 1739. From 1740 to 1819, when population and demand for farm products were again growing steadily, sales of haciendas dropped to once every 28 years. Political instability and economic decline, after independence in 1821, produced a new wave of ownership turnovers. Similar patterns can be observed throughout the Americas. Wherever feasible, profitable estates expanded at the expense of unprotected public or indigenous lands and landowners assumed the risks of managing production directly, expelling tenants and even reducing the size of the garden plots assigned to resident peones. In periods of downturn, or when rural conditions became insecure, however, bankruptcies and forced sales tended to increase, donations of land to the Church rose, and estates were rented out to tenants or broken up for sale in smaller lots. This fragmentation occurred throughout much of highland Latin America in the seventeenth century and again in the first decades after independence.Footnote ²⁷ Fortunes accumulated in mining and commerce were even less stable; miners and merchants often bought up estates because they believed them to be less risky, though less profitable, investments.Footnote ²⁸

Table 3. Ownership turnovers on seven haciendas in the Central District of Oaxaca

Source: William B. Taylor, Landlord and Peasant in Colonial Oaxaca (Stanford 1972), Appendix A.

Third, Latin America's main problem at the end of the colonial era was not that its institutional environment prevented it from importing the Industrial Revolution. The deeper problem was that the Portuguese and Spanish empires had failed to adapt to the revolution in property rights that had produced a Commercial Revolution and the sustained economic advance evident in Britain and the Netherlands at least a century earlier. The Iberian failure to modernise property rights and other institutions affected elites as well as commoners. As generations of Spanish and Portuguese policymakers understood, the survival of their empires depended crucially on the maintenance of a fragile equilibrium in which the authority of weak and distant monarchs depended as much on keeping settler elites insecure in their rights and properties as it did on keeping the lid on discontent from below. By clinging to absolutist principles and colonial hierarchies of race and caste, the Iberian world had already lost the opportunity to make an Industrial Revolution on its own.

This failure had two main dimensions. The first was the failure to reform the property rights regime to encourage enterprise and innovation. Modernising legal codes, judicial systems, fiscal burdens, commercial regulation and governing structures took decades of civil strife to accomplish after independence. Contrary to Engerman and Sokoloff, economic growth did not require institutions that encouraged the poor to invent and invest. It required institutions that made it possible for people of means to do so by lowering transactions costs and reducing risks. Institutional modernisation did coincide with popular demands when liberal movements turned against caste and slave systems, or managed to mobilise peasant or artisan support by means of reversible concessions. But it required an immense effort to consolidate the abolition of communal property, expel squatters from private and public lands, expropriate Church property held in mortmain, reform commercial and civil codes, and render judicial systems properly deferential to elite interests, especially in the face of popular indifference, and even resistance.

The second was the failure to increase state capacities. The weak Iberian states focused their resources and energies on defending territory, repressing internal rebellions, and extracting sufficient tax revenues to do both. They had little energy or resources left to invest in public services, physical infrastructure, or human capital. The instability that followed independence undermined even these limited capacities. The state capacities required to facilitate economic growth were not those demanded by popular majorities. Contrary to Acemoglu et al., economic growth did not require less extractive institutions designed to encourage the exploited to engage in more productive economic activities. Rather, it required an immense effort to exclude majorities from political power and influence, consolidate the dominance of desperately precarious elites, and provide reasonable security for elite (and foreign) investors.

These tasks were left to the independent states created in the 1820s, but by this time the Anglo-American Industrial Revolution had already begun its transformation of the international economy. By dramatically lowering the cost of importing manufactured goods and raising the benefits of specialising in the export of natural resources, the developed countries made industrialisation more difficult in the rest of world. Thus, Latin America was not offered an opportunity to import the Industrial Revolution at the end of the colonial era. Rather, it was offered what Jeffrey Williamson and others have called ‘de-industrialization’, that is, irresistible incentives to abandon whatever local craft and manufacturing industries had developed by the end of the colonial era to specialise more intensively on producing raw materials.Footnote ²⁹ To resist such pressures, Latin America would have needed strong and effective governments committed to promoting modern industry, that is, not less inequality and exploitation, but perhaps much more of both, including subsidies to business and efforts to keep wages down. Until late in the nineteenth century or later, any country or colony, that had not already made its own industrial revolution, faced insuperable difficulties trying to import one from elsewhere.Footnote ³⁰

The last of our three theories on the fall from development locates the source of Latin America's economic backwardness in the deficiencies of Spain's ‘mercantilist’ economic model. The Lange, Mahoney and vom Hau theory relies on the distinction between the ‘mercantilism’ of Spain and the liberalism of the British. The chief virtue of this theory is that it helps to focus attention on the external institutional and policy constraints on economic performance in the New World colonies. As Gabriel Tortella and I argued in 2002, of four major institutional or policy constraints that inhibited economic growth in both Spain and Mexico, including the unwieldy Spanish legal system, the political risk of confiscation or other losses, the slow rate of human capital formation, and anti-trade policies, ‘none appears to be related in a straightforward way to factor endowments or the distribution of wealth or income’.Footnote ³¹ All these problems affected all of the Iberian colonies, but they originated in the institutions and policies that Spain and Portugal imposed and adapted as necessary in their colonial possessions.

The main problem with the approach of Lange, Mahoney and vom Hau is that it risks confusing classification (mercantilist versus liberal models) with the evolving, often contradictory intentions of policy makers, or even with historical processes of cause and effect.Footnote ³² Latin American economies indeed suffered, as they suggest, from a long list of economically counter-productive institutions and policies designed mainly to raise revenues or to balance contending interests in colonial societies. But this was due less to mercantilist dogma than to the structure of the Iberian empires and the challenges that Spain and Portugal faced in running them. As early as the sixteenth century, when the rulers of Spain and Portugal became aware of the vast territories that their subjects were claiming for them, they faced three essential tasks: how to make their possessions secure from external threat; how to impose stable governance that would minimise the risk of defection; and how to raise revenues sufficient to achieve these two goals and perhaps other dynastic ambitions as well. External security required naval forces and loyal settlers in territories likely to be coveted by other powers or populated by potentially rebellious subordinates. Stable governance could only be achieved by policies that used settler elites, along with colonial officials and the Church, to keep indigenous and slave populations in order, while at the same time limiting their power and maintaining a credibly powerful imperial deterrent against internal rebellion of any kind. Imperial revenues, the linchpin of the two Iberian empires, could only be maximised by specialising in extracting them wherever opportunities arose.

All of the early modern empires, absolutist or parliamentary, had to meet these challenges while facing severe material, political, and even cultural constraints. The British government, like its Iberian counterparts, met these challenges pragmatically over time. The repertoire of British imperial management, like that of Spain and Portugal, included restricting colonial trade to the empire; taxing, regulating or prohibiting colonial production that might have competed with home industry; intervening directly to fix prices and wages for the common good; leaving vast territories and populations ungoverned; and excluding native and slave populations from access to the rights and privileges accorded to Europeans. Lange et al. claim that liberal or British colonialism aimed at promoting trade and markets and thus tended to impose such institutional constraints most intensely on less complex societies where native elites could be easily displaced, and less intensely where pre-existing indigenous institutions would have made it difficult and costly to do so. Mercantilist empires, on the other hand, aimed at ‘extracting resources through the use of a dependent labor force without having to alter radically pre-existing economic structures’, and ignored less complex societies where there were few if any precedents for coercive exploitation of labour. In the long run, they suggest, liberal empires work best for economic development.

Geography and population dynamics make this scheme problematic. The Iberians did not seek out complex societies to conquer and administer. They sought riches in whatever circumstances and by whatever means they could discover. In the densely populated highlands of Mesoamerica and the Andes, the first and preferred method was indirect rule, extracting resources through pre-existing institutions, just like the British in India. The Spaniards failed and the British succeeded, not because they sought different ends, but because the extraordinary mortality of the indigenous populations in the Americas destroyed or severely damaged the pre-existing indigenous institutions. Indirect rule became impossible. Neither the differences between Spain and Britain in ruling complex societies, nor the similarities elsewhere, had much to do with the initial intentions or models imposed by the imperial powers.

The Latin American economies fell behind their North Atlantic counterparts between 1750 and 1850 not because they concentrated wealth in too few hands, subjugated indigenous peoples, enslaved Africans, and developed institutions that failed to protect the property rights of the many. They stagnated because the weak governments of the Spanish and Portuguese empires understood that survival depended on avoiding potential challenges as much from settler elites – too secure in their property and civic rights – as from indigenous and slave rebels demanding rights of their own. When the imperial deterrent against rebellion that had maintained this shaky imperial equilibrium for three centuries collapsed after 1808, the Latin American economies faced the worst possible circumstances. They inherited the manifold weaknesses of the colonial state, but now lacked the imperial deterrent to rebellion. Thus, to the inherited burdens of colonialism were now added the costs of civil and international wars. Outside the Garden, all was quicksand. Small wonder it took decades to create the conditions required for economic growth to begin.

How Did Growth Begin (Part One)?

As the settlements at Vienna and the Pax Britannica launched a half-century of peace in the North Atlantic, Latin America sank into an epoch of political turmoil and violence. State capacities declined everywhere, the more so where armed conflicts were prolonged. Struggles to secure or resist independence required the mobilisation of resources and manpower, but independence did not end the fighting. The key insight of the Lange et al. model is the authors' insistence, similar to that of Acemoglu et al., on the connection between colonial institutions and the timing and intensity of institutional modernisation in the post-independence era. The duration and depth of the post-independence civil conflicts depended on the nature of colonial social conditions: conflicts tended to deepen and last longer in places where the power and status of settler elites was most challenged from below. International competition often exacerbated the persistent internal conflicts.Footnote ³³

Because the post-independence conflicts were so costly, they narrowed the political economic options available to the contending interests and factions. Major elements of the colonial institutional legacy became, in the language of twenty-first century politics, non-viable. Caste systems could not be enforced, state monopolies perished, slave systems weakened and collapsed, archaic property rights (entail, mortmain) disappeared, legal exemption from civil and criminal jurisdiction for military officers and priests (called fueros) fell, church and state separated, and internal direct taxes (like the indigenous tribute) had to be abandoned. As this happened, governments shifted to reliance on customs revenues, and taxes on high-value, location-fixed activities, such as urban consumption and mining. This destructive phase of Latin America's nineteenth-century institutional modernisation overlapped somewhat with a second, more constructive phase that coincided in most cases with the onset of economic growth. Second-phase reforms included modern constitutions and legal codes, especially civil and commercial codes, followed usually by banking, insurance and mining laws, along with fiscal, tariff and public debt reforms (see Table 4). Finally, economic growth made possible and facilitated a considerable increase in state capacities, reflected in unprecedented investments in public goods, including infrastructure, education and, eventually, public health.

Table 4. Dates of First Reformed Civil, Commercial, and Mining Codes

Notes: (1) Short-lived codes. (2) Colombian provincial codes adopted during indicated periods. (3) The Brazilian code was restrictive rather than liberalizing. (4) Constitution of 1891 ceded subsoil rights to surface owners.

Source: Adapted from Alan Dye, ‘The Institutional Framework’, in Bulmer-Thomas et al. (eds.), The Cambridge Economic History of Latin America, vol. 2, Table 5.2, p. 189.

Most of the first phase changes occurred with independence or shortly thereafter in the southern cone countries (Argentina, Chile and Uruguay), and Costa Rica, where the population was overwhelmingly European. In these cases, post-independence conflicts centred on the spatial distribution of political power between provincial and central governments. These fights between centralists and federalists stopped when central governments managed to mobilise sufficient resources to impose centralised control (Chile, Costa Rica) or when nominally federalist governments were able to claim the country's custom revenues (Argentina, Uruguay). In all of these cases, the colonial institutional legacy disintegrated soon after independence and mostly without provoking major conflicts. These economies managed to begin growing even before the constitutional settlements that consolidated the new order.Footnote ³⁴

However, where settler elites dominated large slave or indigenous populations, colonial institutions persisted and Latin America's initial experience with independence was neither peaceful nor progressive. Creole elites in Cuba and Puerto Rico preferred colonial status and opted for the comforting presence of Spanish garrisons. The Brazilian colonial regime declared independence in order to avoid changes in the old order. In the Andes (Bolivia, Peru, Ecuador, Colombia) and Mesoamerica (Mexico, Guatemala, El Salvador, Nicaragua), settler elites formed the core of conservative resistance to change, aided by the Church hierarchy and by the new national armies recruited from the officer corps and ranks of the Spanish colonial army. They struggled to restore state capacities (including indispensable sources of state revenues from monopolies, internal customs, and the old systems of state regulation, taxes and fees) and to protect themselves, the Church and the villages from liberal assaults on ancient property rights and proprieties. Eventually, this agenda was defeated everywhere in Latin America, less by liberal doctrines than by the liberal drift inherent in the collapse of the already weak state capacities inherited from the colonial regime during the periods of protracted conflict that followed independence.

It is now possible to account for the Acemoglu et al. results that show a correlation between colonial mortality rates and modern underdevelopment and a ‘reversal of fortune’ that struck down economies that were prosperous in 1500, and elevated those that were not. The answer lies in an intervening variable that amends, but does not reject these results, and that has the additional advantage of opening the black box connecting settlement patterns and pre-colonial wealth to modern levels of GDP per capita. The intervening variable is the speed with which Latin America states undertook the process of institutional modernisation after independence. The pace was quickest in the temperate zone colonies, populated mainly by European settlers and their descendants, notably the southern cone countries (Argentina, Chile and Uruguay) and Costa Rica, the least developed of the Central American republics in the early colonial era. The slowest to modernise institutions included Brazil, the only former slave colony where no political change occurred with independence, as well as Bolivia, Peru and Mexico, where elites struggled to restore colonial stability, not only to protect old ethnic (caste) and social privileges, but also to avert a recurrence of destructive rebellion from below. Thus the pattern of resistance to institutional change in the nineteenth century correlates well with the areas of high settler mortality (the slave-importing tropics) and with the areas most developed prior to the Conquest (the highland regions of Mesoamerica and the Andes).

The pace of nineteenth-century institutional modernisation, with its socio-economic correlates, performs better as a predictor of long term economic performance than the colonial extraction and exploitation proposed by Acemoglu et al. The two indicators they propose (the size of the European population and the extent of pre-colonial urbanisation) yield contradictory results when applied to Latin America. High mortality rates characterised the tropical slave colonies, which would imply later backwardness (a moderately good prediction for Iberian and French slave ex-colonies in the twentieth century, but not the eighteenth). At the same time, the relative backwardness of the tropics at the time of European colonisation would predict a reversal to more advanced economies later (a good prediction only for the twentieth-century British ex-colonies, most of which have attained a higher GDP per capita than most of Latin America). European settler mortality was not unusually high in the temperate plateaus and highlands of Mexico and the Andes (which would suggest better economic outcomes), but these were precisely the areas where the reversal of fortune is most evident today. Colonial institutions made possible the commercial exploitation of people and resources throughout the Iberian New World (and brought some of the slave islands to unprecedented heights of productivity), but their very success made it difficult to modify or abandon them when the empires that gave them life and defended them disintegrated after 1808. Thus, we have an alternative explanation for the success of the two Acemoglu et al. models. They did not identify the source of bad institutions in the colonial era, but they did identify the pattern of resistance to institutional modernisation in the nineteenth century.

How Did Growth Begin (Part Two)?

Towards the end of the nineteenth century, institutional modernisation, political stability and sustained economic growth were evident throughout most of the Latin America).Footnote ³⁵ Four different but overlapping explanations for this happy turnaround have been offered by the new work in political economy. First, historians and economists have long pointed to rising international demand for Latin American exports and the concomitant increase in international capital flows to produce them. As Jeffrey Williamson has pointed out, however, international economic conditions were even more favourable to export growth throughout the first half of the nineteenth century, as the Industrial Revolution precipitated favourable shifts in the terms of trade and the rapid fall of shipping rates made it more and more profitable to produce for export.Footnote ³⁶ Moreover, railroad technology, tested successfully in Great Britain in 1830s, quickly became available for export. Thus, a favourable external economic environment was not enough to get growth started.

The second theory points to institutional modernisation, in addition to the rising demand for exports, as the key to success. Beginning in the 1860s, the destruction of colonial institutions, as a result of civil strife, along with a stunning series of liberal military and political victories, effectively settled the major institutional issues in contention. The major achievements of the liberal victories throughout Latin America included the abolition of caste and slave systems, the liquidation of archaic ‘corporate’ property rights in land, and the enactment of new civil and commercial codes. These were all achievements that redefined property rights and created new legal and social spaces for private enterprise.

This alteration, in what Douglass North once referred to as ‘primary’ institutions, may help to explain why a third theory also makes sense. Several authors, including Adam Przeworski and Carolina Curvale (2005) and Douglass North, William Summerhill and Barry Weingast (2000), have rightly identified the onset of economic growth with the development of mechanisms for ‘securing political order’.Footnote ³⁷ This third theory argues that Latin America achieved economic growth by creating political stability, that is, mechanisms for resolving conflicts among elites and between elite interests and those of peasants and workers. Both sets of authors suggest that democratic mechanisms work better for economic growth, echoing the egalitarianism of the theorists discussed earlier. None appears to distinguish between the more foundational aspects of institution building, which spread public goods widely, and the creation of political institutions for resolving what we might call second-order disputes among interest groups.

Finally, a fourth group of theorists now argue that neither democracy nor stability is required for economic growth. In an important recent book, Stephen Haber, Noel Maurer and Armando Razo have modelled what they called the institutionalisation of a kind of ‘crony capitalism’ or ‘Vertical Political Integration’ (VPI) in Mexico, in which the government of Porfirio Díaz (1876–1910) discovers how to make commitments that it will not expropriate property sufficiently credible to members of the economic elite that they are induced to invest in the financial institutions and productive activities that the country badly needs.Footnote ³⁸ Their model could be applied profitably not just to Mexico, but to other countries as well. However, as Thomas Passananti has pointed out in his exhaustively researched and theoretically sophisticated re-analysis of the early banking history of the Porfiriato, the historical process of developing business confidence occurred in a persistently uncertain environment in which the government manipulated wealthy investors to serve developmental goals, while the members of the economic elite bribed, lobbied, married and deployed family members to stay ahead in what was a much more complex set of interrelated and repetitive games.Footnote ³⁹

As foreign investors (who are curiously missing from the analysis of Haber et al.) and elites made profitable deals, increasingly stable Latin American governments succeeded in adding and enforcing restrictions on the franchise that excluded all but tiny minorities. And as export-led economic growth took off throughout Latin America in the late nineteenth century, economic inequality increased. Peasant and public lands, made newly accessible and valuable – by railway links to regional and external markets – passed into the hands of landlords, politicians, land companies and plantations.Footnote ⁴⁰ Meanwhile, immigration and migration tended to keep wages down as economies grew.Footnote ⁴¹ The returns to scarce capital and skills rose precipitously. The nineteenth century ended, in other words, by committing the sins that much of the new political economy erroneously attributed to the colonial era: relatively high economic inequality, dominance of government by narrow economic elites, exclusion of competing interests and groups from political influence, and ‘bad’ institutions that failed to protect the property and human rights of majorities. Unfortunately for our theorists, and for the region, the nature and timing of Latin America's sinning clearly indicates that it was good, not bad, for economic growth. The conditions that Engerman-Sokoloff and Acemoglu, Johnson and Robinson saw as blocking economic growth were in fact the conditions that made it possible.

Why Has Latin America Grown So Slowly?

We come, finally, to our last question. Latin America's economic growth over the past century or so has not been fast enough to achieve convergence with the productivity and living standards of the developed world. But it has been sustained despite high levels of political and social instability. Stephen Haber and his colleagues have shown that VPI helps to explain how the Mexican economy continued to grow during and after the Mexican Revolution, despite continuing political and social turmoil. Formal and informal deal-making between elite investors and Mexican officials continued, they argue, under the new regime. The key element in each case was the existence or creation of some third-party enforcement mechanism that satisfied elites and governments that the deals would be honoured. Such arrangements, as Noel Maurer has suggested, are often suboptimal, in other words they make investment feasible, but, as in the case of the Banco Nacional de México in 1885, they do so by institutionalising rent-seeking at a significant cost to society.Footnote ⁴² What got growth started, tended to restrain it.

In most countries, the eruption of populist movements and parties especially in the 1930s, tended to extend the benefits of VPI to formerly excluded groups, including unionised workers. But corporatist incorporation did not change the model. In many cases, it did not even promote stability, but instead demonstrated to the still excluded how much they could gain from forcing their way into the game. Import substituting industrialisation worked well in this model of political economy, but collapsed under its own weight, with help from the US Federal Reserve, in the early 1980s.

It is possible that the political economic model, which facilitated Latin American growth at the end of the last century, has run its course. Had the Cold War not succeeded in excluding the left from power for nearly half a century, the region might have avoided prolonged experimentation with militarist VPI, or what the literature once called ‘bureaucratic authoritarianism’.Footnote ⁴³ The still open question in most of Latin America is whether a transition can be made to a new kind of political economy in which credible institutions effectively guarantee property and civic rights, as well as access to public goods, to every citizen. And whether such regimes will be able to improve on the mediocre performance of the past 100 years.