XIVth International Economic History Congress

Helsinki, Finland, 21-25 August 2006

Session 97

Settler Economies in World History

___________________________________________________________________

Five Hundred Years of European Colonization:Inequality and Paths of Development

Stanley L. EngermanUniversity of Rochester and NBER

and

Kenneth L. SokoloffUniversity of California, Los Angeles and NBER

This paper is preliminary, and not for quotation without permission. The paper is intended forpresentation at the International Economic History Association meetings in Helsinki, August 2006. Arevised and more polished version will be available at the meetings.

The study, if not the practice, of colonialism is again in fashion. Over the last few years, the

institution has enjoyed a revival in interest among both scholars and the general public. The European

record has perhaps been the subject of most intense study, but an appreciation of diversity is evident in

the attention devoted to the colonization practices of the Mongols, the Chinese, the Russians, the

Ottomans, the Incas, and the Aztecs. One reason for this reexamination may be sentimentality for a

simpler ordered world, as a number of these new accounts cast colonial empires in a more favorable

light than has generally been customary. Deepak Lal, for example, argues that those nations that

established empires merit praise, as their creations normally brought about lower levels of conflict and

costs of carrying out long-distance trade, as well as promoted greater prosperity in the affected

societies.1 Niall Ferguson highlights progressive sides to Britain’s oversight of her colonies, such as the

introduction of efficient civil services and rule of law, as well as the abolition of slavery. 2 The image of

kinder and gentler imperial powers also has some foundation in the work of Lance Davis and Robert

Huttenback, who in their meticulous and detailed estimates found that Britain was not nearly so

aggressive or successful in extracting returns from its colonies as she could have been, and indeed that

her Empire generated little in the way of returns for the home country overall.3

Quite a different motivation, however, has been behind the recent proliferation of studies by

economists of the European effort to colonize most of the rest of the world.4 Inspired by the goal of

improving understanding of the processes and institutions of economic growth, these scholars have been

attracted by the quasi-natural experiment generated by a small number of European countries

establishing many colonies across a wide range of environments. The logic is that the historical record

of these different societies can be analyzed to determine whether there were systematic patterns in how

their institutions or economies evolved with respect to initial conditions. For example, have colonies

with a British heritage, or those in a particular sort of physical environment, realized more economic

progress over time than their counterparts have? In other words, the history of European colonization

provides scholars with a rich supply of evidence, or a laboratory, that can be used to study economic

performance and the evolution of institutions over the long run. Because some of the characteristics of

the colonies were in place at or near the time of settlement, and thus can reasonably be treated as

exogenous, many economists have been hopeful that the data generated by their later development can

be used to get at causal relationships or mechanisms.

1 Lal 2004. Parts of this paper draw on our working paper, “Colonialism, Inequality, and Long-Run Paths of Development.”2 Ferguson 2003.3 Davis and Huttenback 1986.4 For examples of what has become a substantial literature, see Engerman and Sokoloff, 1997 and 2002; Acemoglu, Johnson,and Robinson 2001 and 2002; and Easterly and Levine 2003.

2

Our investigation began with a question. Why was it that for at least two hundred and fifty years

after the Europeans arrived to colonize the so-called New World, most observers regarded the English,

French, Dutch, and Spanish settlements on the northern part of the North American continent as relative

backwaters with limited economic prospects, and that the flows of resources to the Americas mirrored

that view? The simple answer is that per capita incomes, especially for those of European descent, were

higher in at least parts of the Caribbean and South America than they were in the colonies that were to

become the United States and Canada well into the late-18th and early 19th centuries. Looking back from

the vantage point of the early twenty-first century, however, it is clear that the real puzzle is why the

colonies that were the choices of the first Europeans to settle in the Americas, were those that fell behind

-- and conversely, why the societies populated by those who came later and had to settle for areas

considered less favorable have proved more successful economically over the long run. Another issue

suggested by this experience is why it was not until 250 years after settlement began in the Americas

and Asia, a period after many of the nations of the Americas had gained their independence, that there

was a second burst of colonial expansion into much of Africa, the Pacific islands, and other parts of

Asia.

A traditional and popular explanation for these intriguing patterns credits the success of the

North American economies to the superiority of English institutional heritage, or to the better fit of

Protestant beliefs with market institutions.5 However, proponents of this interpretation generally neglect

the implications of the fact that various British colonies in the New World evolved quite distinct

societies and sets of economic institutions, despite beginning with roughly the same legal and cultural

background and drawing immigrants from similar places and economic classes. Impressed with how the

evidence seemed inconsistent with the notions that British heritage or Protestantism was the key factor,

we instead offered an alternative explanation of the divergent paths of development among the societies

of the Americas. We highlighted how the great majority of European colonies in the New World came

to be characterized early in their histories, primarily because of their factor endowments, by extreme

inequality in the distributions of wealth, human capital, and political influence. We argued, moreover,

that these initial differences in inequality were of major import, because societies that began with great

inequality tended, as compared to the small number -- including those that came to make up the U.S. and

Canada -- that began with relative equality and homogeneity of the population, to evolve institutions that

contributed to the persistence of substantial inequality and generally poor records of development over

5. For example, see North 1988 and Coatsworth 1993 for discussions of why the English institutional heritage helped Canadaand the United States in realizing economic growth. For general discussions of the role of institutions in worldwide economic

3

the long run.6

What led to such substantial differences in inequality across colonies? Briefly put, extreme

inequality arose in the colonies of the Caribbean and in Brazil, because their soils and climates gave

them a comparative advantage in growing sugar and other lucrative crops that were produced at lowest

cost during the 17th, 18th, and 19th centuries on very large slave plantations. These colonies soon

specialized heavily in their comparative advantage, and with the consequent importation of enormous

numbers of slaves, their populations came to be composed of a small elite of European descent with the

dominant share of the population (generally 85 percent or more) consisting of black slaves, or (later)

non-white freedmen and their descendants. Extreme inequality in wealth and human capital came to

characterize much of Spanish America as well. The inequality arose here from the endowment of large

populations of Native Americans surviving the initial impact with the diseases the Europeans brought

with them, and from the Spanish practices (which were influenced by pre-existing Native-American

organizations in Mexico and Peru) of awarding claims on land, native labor, and rich mineral resources

to members of the elite (whose number were limited by restrictive immigration policies). A few areas

that had relatively small native populations, such as Argentina, Uruguay, and Costa Rica, were less

affected however. In contrast, the societies of the northern part of North America developed with

relative equality and population homogeneity, as there were relatively few Native Americans on the east

coast where the colonies were established, and the climates and soils favored a regime of mixed farming

centered on grains and livestock which exhibited quite limited economies of scale in production.

Although the great diversity of settlement patterns and economic structures across the

Americas provide a particularly well suited context for the study of the impact of inequality on

institutional and economic development, the patterns in that part of the world may well have important

implications for the experience of societies established as European colonies elsewhere. With the

exceptions of Australia and New Zealand, European settlements in other parts of the world were not

based upon large numbers of European settlers who became the key productive laborers, but upon small

numbers who remained on the perimeter of the country and exercised control through military power or

political arrangements with the local rulers. For example, the Portuguese, Dutch, British, and French

sailed around the Cape of Good Hope at roughly the same time as they went to the Americas, to acquire

territories and control of large native populations in Asia. The numbers of European settlers were few

and they were generally involved in either political administration or in operating very large agricultural

growth, see North 1981.6. Engerman and Sokoloff 1997 and 2002. It is notable that the Latin American and Caribbean region continues to have the

4

units. And, as with the Americas, the Portuguese had a one-century lead over the British, the French,

and the Dutch in colonizing Asia. These settler populations were rarely directly employed in producing

commodities for sale in European markets, and their primary concern was more with control than with

the production of economic surpluses. As for Africa, the early European settlements on the coast,

mainly trading forts, were not able to exercise control over the native population because of the disease

factors as well as African military force. Even when Europeans were able to move inland during the

19th century, after the introduction of quinine, European domination was achieved with relatively few

settlers through arrangements with local powers or via military prowess.7 The last to be settled of the

European colonies were the Pacific Ocean islands, including Fiji and Hawaii. There too, and particularly

where sugar could be grown, Europeans accounted for only small proportions of the population. In

virtually all of these colonies, suffrage was restricted and expenditures on education and other public

services tended to be miniscule, reflecting (and contributing to) the magnitude of the inequality that

existed between those of European descent and others.

Almost everywhere Europeans settled during their grand epoch of expansion across the globe,

they did so with far higher levels of wealth, human capital (including literacy and familiarity with

technology and markets), and political influence or power than most of the residents native to the area

enjoyed. Thus, where the Europeans encountered large native populations who survived contact with

western diseases (as in Mexico, Peru, Indonesia, or India), their advantages in human capital and other

assets generally meant that Europeans did extremely well relative to the bulk of the natives, and that

there was great inequality. Where they moved into fairly empty or depopulated territories, however (as

in Australia, New Zealand, Canada, or the United States), relative equality tended to prevail. 8 The more

heavily populated colonies, or those in tropical areas that could quickly increase population by drawing

on imported slave labor, often had quite different comparative advantages than Europe (due to different

climates, valuable natural resources, and large native populations). As free populations were primarily

motivated by the prospects of economic returns, these areas generally attracted the greater number of

Europeans until the 18th century, when the greater opportunities associated with commercial grain

highest level of inequality in the world. See Deininger and Squire 1996.7 The main exception to this generalization is South Africa, but even here those of European descent accounted for about 20percent of the population.8 In those cases where the endowments were well suited to large-scale labor-intensive production of staples, slaves or contractlabor were often brought in to provide a labor force. The importation of slaves to the Caribbean basin to grow sugar is theoutstanding example of this, but the extensive use of contract labor from South Asia to augment the labor force, especiallywhere land was relatively abundant, after the emancipation of slaves provides another. See Engerman 1982, 1983, and 1986.

5

agriculture and industrialization shifted attention to mainland North America.9 Overall, the

phenomenon of European colonization generating many societies with extremely high degrees of

inequality, and rather few with low inequality, seems unlikely to have been confined to the Americas.

II

In chapter one of his Imperialism, John Hobson reproduced an interesting table first presented by

Henry C. Morris in his The History of Colonization, constructed from data presented in the Statesman’s

Yearbook of 1900. That Yearbook presents 1899 data on area, population, and related matters for each

independent country and their colonies, protectorates, and dependencies. Among the figures Morris

computed are the actual number of separate colonies that existed at that time, and the numbers belonging

to each colonizing nation. The total number of colonies listed was 136, with Britain having 50 and

France 33 (61 percent for the two together of the overall figure). In 1890, the domains of these two

main colonial powers accounted for nearly 65 percent of the more than 18 million square miles and

nearly 80 percent of the more than 400 million people living under colonial dominion (see Tables 1, 1a,

2, and 2a). Britain alone had 37 percent of the colonies, controlling a bit more than half of the land area

and roughly 70 percent of the population. Hobson pointed out that 39 of the British colonies were

annexed after 1870. Thus one-third of the land area of the British Empire, with one-quarter of its

population, was acquired between 1870 and 1900. By 1900, the land area of the Empire had nearly

quintupled since 1860, and the population more than doubled. France, whose extensive colonial

holdings had been markedly depleted though military defeats in the late 18th and early 19th centuries,

expanded its domain even more rapidly during the this period as the Europeans nations made major

moves into Africa and Southeast Asia; the land area and population encompassed by its colonies rose

between 1860 and 1900 by 15.5 and 16.7 times respectively.

Britain and France were far from the only European countries engaged in colonial enterprise.

Indeed, the scale of the colonization efforts carried out by the Europeans countries is striking not only

for the numbers of individual colonies and the vast land areas and populations involved, but also for the

wide range of colonizers. The European colonizers included Austria-Hungary, Belgium, Denmark,

France, Germany, Italy, the Netherlands, Portugal, Russia, Spain, Turkey, as well as Britain (see Table 3

for an incomplete accounting),and of these none other than Belgium held fewer than two colonies. Only

four had mother countries larger in area than their colonies10, and four had populations less than the

9 Engerman and Sokoloff 2002.10 Austria Hungary, Russia, Turkey, and the United States.

6

population of the colonies.11 Most of them had colonial domains that extended across two or more

continents. The United States would join this group in 1898, with the possessions it gained during the

Spanish-American War.12 Especially when one considers that many of the colonies the Europeans had

established, and long maintained, in the Americas were independent by the late 19th century, one cannot

fail to be impressed with how much of the globe -- both in land area and population -- had direct

experience of being a colony.13

The large numbers of both colonies and colonizing countries, and the wide range of geographic

locations and other characteristics, provide a wonderful basis for a detailed examination of various

contributions to the success (or lack thereof) of colonial enterprise. What are also extremely valuable

are the weak association between the identity of the colonizing country and the environmental variables,

as well as the major differences across colonies in the date of acquisition. That European colonization

of different parts of the world tended to occur in waves (and that many colonies experienced changes in

the identity of the colonizer) provides the contemporary investigator with rich information that allows

for study of whether characteristics that are conducive to success in one era may have different effects in

another. Of course the role of changes over time is a crucial and complex question. Did the era in

which a colony was established affect its institutions or overall performance? Was there, for example,

significant evolution in the world’s legal or moral standards -- such as the acceptability of slavery or of

indentured servitude -- that influenced the foundation of institutions? For Hobson, while the “year 1870

has been taken as indicative of the beginning of a conscious policy of imperialism, it will be evident that

the movement did not attain its full impetus until the middle of the eighties.” We might expect,

however, that earlier settlements -- including those in the Americas that had already gained their

independence after centuries under colonial rule -- were in areas judged to have more favorable natural

resources, factor endowments, or conditions more generally attractive or promising than those settled

later. Another issue in determining whether time of settlement made a difference, say to the

performance of African and several Asian colonies, is that these territories had previously been

politically independent and their longer periods to develop on their own may have had a long-term

impact.

There are of course other questions about how to evaluate the performance of colonies. Should

the focus be on the success of a colony over the period when a society was ruled by the colonizing

11 Britain, France, the Netherlands, and Portugal.12 China also merits inclusion in a list of colonizers of that era, and Morris includes her in his compilation.13 This assessment is reinforced by the observation that much of Europe was at some point colonized by societies from Asiaor the southern Mediterranean.

7

country or on the record of the society over the longer term -- including both the period after

independence as well as the conditions prior to colonization? Should, moreover, success be gauged by

the level of GNP per capita (or some other measure of material output or welfare), the income of the

elites, the income of the native population, or perhaps by the improvement in levels since the time of

colonization (as opposed to the absolute level)?

Colonies, as institutions, are a blend of the roles of the colonial power and of the resource and

endowment base of the colonized areas. The process of colonization entails a change in political control

of the land areas and the resources therein, with the settlers and/or colonizing nations acquiring the

political authority to design and legislate institutions. The specific forms of the economic and political

institutions that evolve will be influenced by the attributes of the controlling elite and of others

(including the native population) already resident, as well as by the impacts of resources, demographic

endowments, and other (including market) environmental conditions. The aims of settling nations are

generally economic in character, for the individual settlers or for the entire nations, although at times

they may be subsidiary to other goals, such as the attainment of prestige, the strategic benefits of

location, or the attempt to triumph in international rivalry. That the circumstances prevailing in the

world at the time of settlement will likely have a profound influence on the nature of the institutions in a

colony seems obvious. As already alluded to, the legality and technical feasibility of different types of

institutional arrangements change over time. For example, European nations abolished systems of

slavery and serfdom over the course of the 19th century, and indentured labor was outlawed at the start

of the next century. Thus, settlements established at different times faced rather different institutional

options for mobilizing labor from outside.

One of the critical factors in influencing institutions and the process of economic growth is the

nature of the population endowment at the time of settlement, and the overall ratio of land (and

resources more generally) to labor. The population density at the time of settlement, and the willingness

of the resident population to supply labor (i.e. work the land of others), will determine whether the

desired level of labor input (as reflected in the productivity of labor relative to some subsistence

standard) can be obtained from the existing population or whether labor should be attracted from

elsewhere. The institutions in colonies with high population density should accordingly differ from

those of colonies with lower population densities. Many colonial settlements were in areas with higher

population density relative to labor requirements, and these entailed limited European migration and

populations. For example, much of Latin America, particularly the areas settled by the Spanish, had

large Native-American populations, even after the heavy mortality after European arrival, and there was

8

limited need for in-migration by Europeans or African slaves.14 Other areas where immigration from

outside was quite limited include the overwhelming majority of colonies in Africa (except for South

Africa) and Asia (particularly India and Indonesia), where the populations contained very few

Europeans relative to the native populations.15

Some colonies with low ratios of labor to land (such as Australia) arose from small native

populations that had long characterized the territories, but others were based upon the nature of their

soils, climate, and comparative advantage in the world economy. Most prominently, colonies in the

Caribbean and the U.S. South, producers of export crops and settled beginning in the 17th and 18th

centuries, mobilized slave labor from Africa to meet the bulk of their labor needs, since there were small

Native-American populations and substantial economies to producing the key crops cultivated there on

large plantations with slave labor. The populations of the so-called ‘sugar islands’ in the Caribbean,

especially, were generally about 90 percent black slave after the early years of settlement. A major

change in institutions came with the ending of the slave trade and slavery by European powers, and such

areas -- not only in the Caribbean, but also in other sugar-growing areas of labor shortage such as Fiji,

Queensland, and Natal -- generally came to rely on indentured labor coming from poorer parts of the

world, particularly India, for their labor supply. Presumably, if slavery had continued to be permitted, it

would have been the predominant means of obtaining labor in these colonies.

A second type of area where labor was in short supply relative to land, but which had more

temperate climates and less need for plantation labor producing exports, were the key areas of British

settlement, including the thirteen colonies that became the United States, which developed out of

military conquest (vs. the Dutch) and purchase (from the French and Spanish): Canada (based on

conquest in the Seven Years War), Australia, New Zealand, and (with a smaller ratio of whites than the

others) South Africa (conquered from the Dutch). These areas sought primarily white labor, as they

were unable to pay the high prices for slaves prevailed in world markets because of the demand from

more tropical colonies. The labor these colonies attracted, generally from the metropole could take

several forms, free or convict.

There are several institutional variants of free labor institutions, with quite different means of

14 Indeed, early in the colonial period, the Spanish authorities adopted rather stringent restrictions on immigration. For morediscussion, see Engerman and Sokoloff 2002. Coercive measures were sometimes employed, however, to get the workers towork where settlers wanted them to, although these measures often resembled those that had been used by Native- Americansocieties before European arrival.15 Although some of these areas did not produce much for export, others appear to have been quite dependent on productionfor export markets. Indeed, in several of the latter cases -- such as Barbados and India -- there was overpopulation, and lowreturns to labor, leading to encouragement of out-migration.

9

controls and cost allocations. The most straightforward, of course, is where the free migrants made their

travel arrangements and were able to pay directly for the cost of their transportation. Many ‘free labor’

migrants, and especially those from Asia or from Europe prior to the 19 th century, were not so fortunate.

People in this latter group sometimes depended on subsidies that were offered by potential employers

(or intermediaries in labor markets) to attract labor to work on a specific crop or in a particular location.

Perhaps more often, they chose to become indentured servants, whereby they gave up control for a

specified period of years of where and when to labor in exchange for having transport costs paid for.

In the case of indentured servants who traveled from Europe to the Americas prior to the 19th century,

almost all workers remained once free (having worked off their commitment). Convict labor from

England was used to settle the thirteen colonies and, later, Australia. Although convict workers would

ultimately be freed, this institution of labor entailed the individual worker a giving up of his or her rights

for a period of years to serve punishment. Convict labor was also employed, to some degree, in settling

Portuguese and French colonies.

The relative scarcity of labor, as well as the numbers of those of European descent (white) as

compared to the numbers of natives and workers brought in as slaves or under contract arrangements

from Asia, varied enormously across colonies (see Table 4). Whereas whites were the great majority in

the colonies established in Europe, North America (as they were in former colonies such as the United

States) and in Australasia, and accounted for substantial shares of the population in a few others such as

the Cape of Good Hope, they generally constituted only a miniscule proportion in those colonies that

either had large aboriginal populations before the Europeans arrived or climates, soils, and other

resources conducive to the production of sugar and certain other commodities that could be produced at

low cost with slave or contract labor. Hence, regardless of the identity of the colonizing country, whites

numbered less than one percent of the population in the colonies in Asia and (with the exception of

South Africa) in Sub-Saharan Africa, and typically much less than twenty percent in the colonies of

North Africa, the West Indies, and Pacific/Oceania. Favored by generally higher (and more appropriate

to colonial societies) levels of human capital, wealth, and legal privileges/rights, whites typically

enjoyed far higher incomes and political influence than did non-whites. There is reason to believe,

therefore, that the circumstances that systematically contributed to the extreme variation in the ethnic

composition of the colonial populations, were powerfully related to the extent of economic and political

inequality that emerged early in the histories of the respective societies, and that such conditions

10

(including differences in the extent of inequality) often persisted over time.16

The stark contrasts in the degree of initial inequality that seem apparent among the European

colonies around the world allow scholars to study whether and how inequality affects the processes and

path of development. Whereas previous treatments of the impact of inequality on growth have often

been concerned with how savings or investment rates might be affected, we and other scholars who have

sought to use the natural experiment provided by colonization focus on the hypothesis that extreme

differences across colonies in the extent of inequality gave rise to systematic differences in the ways

institutions evolved, and in turn on paths of development.17 The argument is that greater equality or

homogeneity among the population led, over time, to more democratic political institutions, more

investment in public goods and infrastructure, and to institutions that offered relatively broad access to

property rights and economic opportunities.18 In contrast, where there was extreme inequality, political

institutions were less democratic, investments in public goods and infrastructure were far more limited,

and the institutions that evolved tended to provide highly unbalanced (favoring elites) access to property

rights and economic opportunities. The resulting differences in access to opportunities may be

important in accounting for the disparate records at long-term growth, because where processes of early

industrialization have been sustained (such as in Britain and the United States during the 19th century,

and even East Asia in the 20th), they have generally involved broad participation in the commercial

economy. Economies that only provided narrow access to opportunities might have been, and be, less

capable of realizing sustained economic growth.

There are a variety of mechanisms through which the extent of inequality in a society might

affect the character of institutions that develop. The avenue that typically receives the most attention

works through political inequality. When political power or influence is concentrated among a small

segment of the population, that group is able to shape policies or institutions to its advantage. We

expect members of such elites to act in their interest, for example, by inducing the government to make

investments and provide services they favor while being assessed for a less than proportionate share of

the costs, or to define and enforce property and other sorts of rights in ways that treat them in a

16 For a more extensive elaboration on this idea, focusing on the colonies established in the Americas, see Engerman andSokoloff 2002.17 For examples of the approach that highlight variation in savings rates with relative income or with rates of taxation, seeAlesina and Rodrik 1994 and Persson and Tabellini 19994. For those investigating the impact of inequality on institutionsmore broadly, see Engerman and Sokoloff, 1997 and 2002; Acemoglu, Johnson, and Robinson 2001 and 2002; and Easterlyand Levine 2003.

18 There are of course some classic expositions of these and similar ideas. See, for example Tocqueville 1835 and Turner1947.

11

preferential manner. Some activity of this sort is present in all societies, as the distribution of political

influence is never entirely equal, and those with more resources generally fare better in the competition

over influencing the government. But the extent and ultimate impact of such activity can vary even

across nominal democracies, especially when the right to vote depends on literacy or wealth (or other

attributes), or where ballots are not secret. The absence of democracy, or a situation when one class of

the population has the capability to impose its will by force if need be, is an extreme case of how

political inequality can lead to institutions that favor a narrow range of the population.

The importance of political inequality (or military might) often figures prominently in

discussions of how institutions are established in colonies. The presumption that those with a monopoly

of force or a dominant share of the votes get their way does not seem an unreasonable presumption.

Nevertheless, it is worth reflecting on the relevance of the well accepted modern adage: you can’t

always get what you want.19 No matter how much inequality there is in political influence or in any

other dimension, there are frequently constraints that inconveniently narrow the range of feasible

possibilities for the fortunate individual or class. The initial objects of the colonies established in the

Americas, and indeed elsewhere in the world, were generally the same -- to generate economic returns

for the respective European country. Although the goals may have been similar, the diverse

environments in which the colonies were located led to a variety of economic structures and institutions

as the colonizers sought to take best advantage of the different opportunities and challenges they faced

in the respective places. Miscalculations of the effects of various institutional designs, with resulting

unintended consequences, were, of course, not uncommon. The colonists came with similar

backgrounds and institutional heritages, but heterogeneity developed as they applied and adapted the

technologies and institutional heritages they brought with them to conditions quite unlike those in the

Old World. Moreover, the extent to which the metropolis, or any political authority, could effectively

specify the institutions prevailing in any colony varied with the local circumstances.

It is well known that in many of the Spanish colonies in Latin America, especially where

aboriginal populations were concentrated, a relatively small number of individuals were favored with

large grants of land and claims on labor and tribute from natives that long endured. Less fully

appreciated, however, is that there were also efforts to implant a European-style organization of

agriculture based on concentrated ownership of land combined with labor provided by tenant farmers or

indentured servants in many of the colonies of North America, as when Pennsylvania, New York,

Maryland, and Canada (the same could be said for Australia) were established. But these attempts

12

invariably failed; large landholdings unraveled because even men of ordinary means were able to set up

and flourish as independent farmers where land was abundant and cheap, labor was scarce, and scale

economies were absent. Despite William Penn having received the royal charter for Pennsylvania, and

accordingly having initial control of the territory, such conditions frustrated the attempts of this

fabulously wealthy member of the elite to replicate an English-style organization of agriculture in the

New World. As much as wealthy men such as Penn might have liked in an ideal world to institute

hierarchical institutions that greatly advantaged those of their class, their ability to attain that goal was

tempered by the need to attract more labor and more productive labor to their respective colonies; that is,

even landowners were desirous of taking steps that would attract more migrants from Europe and

elsewhere. Similarly, the Puritans who settled in the Massachusetts Bay Colony might have liked to do

as their brethren who chose to site their early 17th century colony on Providence Island (an island off the

coast of Nicaragua), and rely on Native American, indentured servants, or slaves to perform their

manual labor, but the cold harsh climate in New England would not support such a commercial

strategy.20 Indeed, a century later New Englanders despaired at being able to afford the high prices

slaves commanded in the international market.21

These cases suggest that political inequality alone was not sufficient for elites to obtain

institutions that greatly advantaged them with respect to government policies or access to property rights

and other kinds of economic opportunities. In some environments, even when political or military

power was highly concentrated in their hands, elites might have voluntarily, and without threat of

violent upheaval, found it in their interest to provide better conditions and treatment to the humble.

Although there are a variety of factors that might lead to such an evidently anomalous outcome, and

ways of characterizing them, the relative scarcity of labor seems in the context of the European colonies

to be a crucial one. Where labor was relatively scarce, as compared to land and other resources, political

inequality was not accompanied by economic inequality. In such circumstances, the lack of economic

inequality (or relative equality) circumscribed how far political elites could go in designing institutions

to advantage their members. In a situation where there was relative political equality, however,

economic inequality -- as reflected in the relative scarcity of factor in somewhat elastic supply – might

lead to institutions that greatly advantaged that scarce factor. Hence, economic inequality can

sometimes, in the sorts of conditions that are not uncommon in colonies or less-developed countries

19 Rolling Stones 1969.20 Kupperman 1993.21 As McManus 1973 (p. 23) makes clear, those in the northern U.S. were priced out of the market for slaves by the 1760s:“By 1764 Thomas Rich, one of Philadelphia’s leading traders noted that ‘the time is over for the sale of Negroes here.’”

13

(with an abundance of unskilled labor but a scarcity of capital and skilled labor), exert more of an

influence on the way institutions evolve than political inequality per se.

III

Comparative study of the record of the long-term development of the societies of the Americas

supports our hypothesis that there were empirical regularities in the ways strategic institutions evolved,

such that those that began as colonies with relatively extreme inequality were more restrictive in

providing access to economic opportunities and less oriented toward investing in public goods and

infrastructure than were those that began with relative equality or homogeneity among the population.

This pattern contributed to the long-term persistence of extreme inequality among the former group, and

may also help to explain why their long-term records of economic growth have been mediocre at best,

relative to those of the latter and especially relative to expectations during the era of European

colonization. The specific mechanisms that worked to produce the divergence in institutional and other

development are complex and difficult to discern, but it seems clear that they often involved factors

other than differences in the political power of the elite.

It has long been recognized that the conduct of elections, including who holds the right to vote, is

one of the most crucial of institutions. Varying the rules or organization of how votes are cast and of

who casts them can have a fundamental impact on the policy choices that the elected representatives –

who in some sense constitute the collective government of the electors – make. As governments

generally have a monopoly of power over certain important activities, there are often major implications

for how a society’s resources or wealth is distributed across the population, as well as for the pace of

economic growth. Given what is at stake, it should not be surprising that throughout history many have

fought and died over both the design of the rules and the outcomes of elections. Most of the societies of

the Americas had achieved independence from their colonial masters and were at least nominal

democracies by the early 19th century, and thus our estimates (see Table 5) of how broadly the franchise

was extended over time and of what fractions of respective populations actually voted in elections have

a direct bearing on the extent to which elites based largely on wealth, human capital, race, and gender

held disproportionate political power in their respective countries, and on whether and how initial

differences in such power or influence persisted.

The estimates reveal that until the 20th century it was common in all countries to reserve the right

to vote to adult males (white adult males until after the Civil War), but the United States and Canada

were the clear leaders in doing away with restrictions based on wealth and literacy, and much higher

14

fractions of the populations voted in these countries than anywhere else in the Americas. Not only did

the United States and Canada attain the secret ballot and extend the franchise to even the poor and

illiterate much earlier (restrictions that were reintroduced in the United States at the expense of blacks

and immigrants late in the 19th century), but the evolution of the proportion of the population that voted

was at least a half-century ahead of even the most democratic countries of South America (namely,

Uruguay, Argentina, and Costa Rica, which have generally been regarded as among the most egalitarian

of Latin American societies and whose initial factor endowments most closely resembled those of the

United States and Canada).

It is remarkable that as late as 1900, none of the countries in Latin America had the secret ballot

or more than a miniscule fraction of the population casting votes.22 The great majority of European

nations, as well as the United States and Canada, achieved secrecy in balloting and universal adult male

suffrage long before other countries in the western hemisphere, and the proportions of the populations

voting in the former were always higher, often four to five times higher, than those in the latter.

Although many factors may have contributed to the low levels of participation in South America and the

Caribbean, wealth and literacy requirements were serious binding constraints. Some societies, such as

Barbados, maintained wealth-based suffrage restrictions until the mid-20th century, while most joined

the United States and Canada in moving away from economic requirements in the 19th century.

However, whereas the states in the United States frequently adopted explicit racial limitations when they

abandoned economic requirements, Latin American countries typically chose to screen by literacy.

The contrast between the United States and Canada, on the one hand, and the Latin American

countries, on the other, was not so evident at the outset. Despite the sentiments popularly attributed to

the Founding Fathers, voting in the United States was largely a privilege reserved for white men with

significant amounts of property until early in the 19th century. By 1815, only four states had adopted

universal white male suffrage, but as the movement to do away with political inequality gained strength,

the rest of the country followed suit: virtually all new entrants to the Union extended voting rights to all

white men (with explicit racial restrictions and very favorable definitions for white immigrants of

residence generally introduced in the same state constitutions that did away with economic

requirements), and older states revised their laws in the wake of protracted political debates. The rapid

22 There is some controversy about whether Argentina had wealth and literacy requirements for suffrage. Whatever the case,the proportions of the population voting were very low in that country (1.8 percent in 1896) until the electoral reform law of1912. Those who point to the absence of such electoral restrictions at the level of the national government suggest that thelow voter participation was due to a failure of immigrants to change their citizenship and vote, as well as to the lack of asecret ballot. Others believe that restrictions on the franchise had, in fact, been enacted and were enforced at the provinciallevel until 1912. See the discussion in Engerman and Sokoloff 2005.

15

extension of access to the franchise in the frontier states, which were distinguished both by more equal

distributions of wealth and labor scarcity, not coincidentally paralleled liberal policies toward public

schools and access to land, as well as other policies that were expected to be attractive to potential

migrants.23 It is hard to avoid the conclusion that political equality was the result of economic equality,

rooted in labor scarcity, rather than the reverse. It is striking that pioneers in extending suffrage, such as

new states to the United States, Argentina, and Uruguay, did so during periods in which they hoped to

attract migrants, such that the rights to suffrage formed part of a package of policies thought to be

potentially attractive to those contemplating relocation.24 When elites—such as land or other asset

holders—desire common men to locate in the polity, they thus may choose to extend access to privileges

and opportunities even in the absence of threats of civil disorder; indeed, a polity (or one set of elites)

may find itself competing with another to attract the labor or other resources.25 Alternative explanations,

such as the importance of national heritage, are not very useful in identifying why Argentina, Uruguay,

and Costa Rica pulled so far ahead of their Latin American neighbors, or why other British colonies in

the New World lagged behind Canada.

Schooling institutions provide yet another appropriate and important test of whether societies

that began with extreme inequality exhibited different patterns of investment in public goods and of

access to economic opportunities. Increases in a society’s levels of schooling and literacy have been

related both theoretically and empirically to many socioeconomic changes conducive to growth,

including higher labor productivity, more rapid technological change, and higher rates of commercial

and political participation.26 Although many New World societies arising out of European colonization

were so prosperous that they clearly had the material resources to attain high rates of literacy by

establishing a widespread network of primary schools, only a few made such investments on a scale

sufficient to serve the general population before the 20th century. The exceptional societies, in terms of

leadership in investing in institutions of primary education, were the United States and Canada (see

Table 6). Virtually from the time of settlement, the populations of these countries seem generally to have

been convinced of the value of providing their children (females as well as males) with a basic

education, including the ability to read and write. It was common for schools to be organized and funded

at the village or town level, especially in New England. The United States probably had the most literate

23 See the extended treatment of these and related issue in Engeman and Sokoloff 2005.24 For the concern with attracting migrants in the U.S. and Argentina for example, see Engerman and Sokoloff 2005; Castro1971; Solberg 1970; and Adelman 1994.25. See Acemoglu and Robinson 2000 for the argument that in many western European countries, the franchise was extendedunder threat.26. See the discussion in Easterlin 1981.

16

population in the world by the beginning of the 19th century, but the common school movement, which

got under way in the 1820s (following closely after the movement to extend the franchise), put the

country on an accelerated path of investment in educational institutions. Between 1825 and 1850, nearly

every northern state that had not already done so enacted a law strongly encouraging or requiring

localities to establish free schools open to all children and supported by general taxes.27 Schools were

also widespread in early 19th century Canada, and although this northern-most English colony lagged

behind the U.S. by several decades in establishing tax-supported schools with universal access, its

literacy rates were nearly as high.28

The rest of the hemisphere trailed far behind the United States and Canada in primary schooling

and in literacy. Despite enormous wealth, the other societies of British colonial origin were very slow to

organize schooling institutions that would serve broad segments of the population.29 Similarly, even the

most progressive Latin American countries, such as Argentina and Uruguay, were more than seventy-

five years behind the United States and Canada in this regard. These societies began to boost their

investments in public schooling at roughly the same time that they intensified their efforts to attract

migrants from Europe, well before they implemented a general liberalization of the franchise. While this

association might be interpreted as providing for the socialization of foreign immigrants, it also suggests

that the elites may have been inclined to extend access to opportunities as part of an effort to attract the

scarce labor for which they were directly or indirectly competing. The latter perspective is supported by

the observation that major investments in primary schooling did not generally occur in any Latin

American country until the national governments provided the funds; in contrast to the pattern in North

America, local and state governments in Latin America were not willing or able to take on this

responsibility on their own. Most of these societies did not achieve high levels of literacy until well into

the 20th century. Fairly generous support was made available, however, for universities and other

institutions of higher learning that were more geared toward children of the elite.

Two mechanisms help explain why extreme levels of inequality depressed investments in

schooling. First, in settings where private schooling predominated or where parents paid user fees for

their children, greater wealth or income inequality would generally reduce the fraction of the school-age

population enrolled, holding per capita income constant. Second, greater inequality likely exacerbated

27. Cubberley 1920.28. See, for example, Phillips 1957; and Wilson, Stamp, and Audet 1970.29. Indeed, significant steps were not taken in this direction until the British Colonial Office began promoting schooling inthe 1870s.The increased concern for promoting education in the colonies may have been related to developments in GreatBritain itself. Several important expansions of the public provision of elementary education occurred during the 1870s,including the 1870 Education Act and the 1876 passage of a law calling for compulsory schooling through the age of ten.

17

the collective-action problems associated with the establishment and funding of universal public

schools, either because the distribution of benefits across the population was quite different from the

incidence of taxes and other costs or simply because population heterogeneity made it more difficult for

communities to reach consensus on public projects. Where the wealthy enjoyed disproportionate

political power, they were able to procure schooling services for their own children and to resist being

taxed to underwrite or subsidize services to others.

Indeed, this sort of interpretation is supported by examination of the structures of public finance

employed across the Americas over time. The general pattern throughout the hemisphere well into the

20th century was reliance by national governments on tax structures that targeted commodities or trade

rather than income or wealth.30 Stark differences existed across the societies of the Americas, however,

in the size and revenue sources of state/provincial and local governments. Local governments were

much smaller in Latin American nations than in the United States and Canada (see Table 7). They

accounted for only about 10 percent of total government revenue in Brazil, Colombia, and Mexico

throughout the 19th century (and in Chile, between 10 and 20 percent during the second decade of the

20th century, despite the absence of state/provincial governments). The contrast with the neighboring

societies in North America is dramatic. In both the U.S. and Canada, the local governments collected

more than half of overall government revenue from the middle of the 19th century onward. Even as late

as the 1930s, the share of local government revenue was nearly 40 percent in both the US and Canada.

This is quite significant in that the aggregate pattern of expenditures by local governments was

quite progressive in that the main priorities of local governments were (well into the 19th century)

schools, roads, and other infrastructure that generate broadly distributed social returns.31 Moreover,

their heavy reliance on the property tax, together with their large share of the government sector, made

for a rather progressive tax structure at both the local and national (all levels of government together)

levels.32 This pattern, characterized by the predominance of property and inheritance taxes accounting

for the bulk of the revenue collected by governments at all levels, endured into the early decades of the

30 The income tax introduced during the Civil War was ultimately ruled unconstitutional, and thus it was not until aconstitutional amendment that such a tax could be reinstituted.31 At 1900, local governments seem to have obtained well over 90 percent of revenue from property taxes. For furtherdiscussion of how the importance of the property tax as a source of state revenue varied over the 19th century, see Wallis2001.32 Any conclusions about just how progressive or regressive any particular tax structures are must, of course, take intoaccount the ultimate incidence of the taxes assessed. An analysis of incidence is beyond the scope of this paper.Nevertheless, we feel rather confident, especially for the 19th century, of following the convention of presuming that propertytaxes are more progressive in incidence than levies consisting of tariffs on imported goods and the revenues obtained fromstate monopolies on consumer commodities such as liquors and tobacco.

18

20th century (a similar pattern holds in Canada).33 In contrast, although the local governments in Latin

America raised relatively more of their revenue from taxes on property and income than did the

respective national or state/provincial governments, they relied much less on these sources than did their

North American counterparts. This, together with the markedly smaller size of local governments in

Latin American nations resulted in radically different, and much less progressive, aggregate tax

structures overall than in the U.S. or Canada.

Although there may be other explanations for these patterns in the evolution of tax institutions,

the evidence is consistent with the hypothesis that initial differences in the extent of inequality across

these societies contributed to the different decisions they made regarding how much revenue to raise, the

relative use of different tax instruments, the nature and size of state and local governments, and the types

and size of government expenditure programs. In general, the countries that began with more inequality

developed structures of public finance that relied relatively more on indirect taxes and placed less of a

tax burden on those with higher levels of wealth. This alone should have encouraged the persistence of

extreme inequality, but the stunted local governments, which are the same authorities most concerned

with public schooling, transportation, water/sewer projects, and other types of investment projects that

generate benefits for a broad spectrum of the population, also worked in the same direction. An

explanation for this pattern is readily available. With a radically unequal distribution of resources, elites

would bear most of any tax burden, especially one levied on wealth or income, and realize a smaller than

proportionate benefit, especially since they could procure for themselves and their families many of the

same services privately.

Land policy comprises a final example of the ways in which institutions may have contributed to

the persistence of inequality over the long run. Virtually all the economies in the Americas had ample

supplies of public lands well into the 19th century and beyond. Since the respective governments of each

colony, province, or nation were regarded as the owners of this resource, they were able to influence the

distribution of wealth, as well as the pace of settlement for effective production, by implementing

policies to control the availability of land, set prices, establish minimum or maximum acreages, provide

credit for such purposes, and design tax systems. Because agriculture was the dominant sector

throughout the Americas, questions of how best to employ this public resource for the national interest,

and how to make the land available for private use, were widely recognized as highly important and

often became the subject of protracted political debates and struggles. Land policy was also used as a

33 At 1902, property, death, and gift taxes accounted for more than 60 percent of total tax revenue to all levels of governmentin the U.S. combined. See Table 7.1 in Sokoloff and Zolt 2004.

19

policy instrument to influence the labor force, either by encouraging immigration through making land

readily available or by influencing the regional distribution of labor (or supply of wage labor) through

limiting access and raising land prices.

The United States never experienced major obstacles in this regard, and the terms of land

acquisition became easier over the course of the 19th century.34 The well-known Homestead Act of

1862, which essentially made land free in plots suitable for family farms to all those who settled and

worked the land for a specified period, was perhaps the culmination of this policy of promoting broad

access to land. Canada pursued similar policies: the Dominion Lands Act of 1872 closely resembled the

Homestead Act in both spirit and substance. Argentina and Brazil instituted similar changes in the

second half of the 19th century as a means to encourage immigration, but these efforts were much less

directed and thus less successful at getting land to smallholders than the programs in the United States

and Canada.35 In Argentina, for example, a number of factors explain the contrast in outcomes. First, the

elites of Buenos Aires, whose interests favored keeping scarce labor in the province if not the capital

city, were much more effective at weakening or blocking programs than were their urban counterparts in

North America. Second, even those policies nominally intended to broaden access tended to involve

large grants to land developers (with the logic that allocative efficiency could best be achieved through

exchanges between private agents) or transfers to occupants who were already using the land (including

those who were grazing livestock). They thus generally conveyed public lands to private owners in

much larger and concentrated holdings than did the policies in the United States and Canada. Third, the

processes by which large landholdings might have broken up in the absence of scale economies may

have operated very slowly in Argentina: once the land was in private hands, the potential value of land

in raising or harvesting livestock may have set too high a floor on land prices for immigrants and other

ordinary would-be farmers to manage. Such constraints were exacerbated by the underdevelopment of

mortgage and financial institutions more generally.36

Argentina, Canada, and the United States all had an extraordinary abundance of virtually

34. See Gates 1968 for a comprehensive overview of U.S. land policy. Discussions of Canadian land policy include Solberg1987; Pomfret 1981, pp. 111–19; and Adelman 1994, chap. 2.35. See Viotti da Costa 1985, chap. 4; Solberg 1987; Solberg’s essay in Platt and di Tella 1985; and the excellent discussionsin Adelman 1994.36. It is generally thought that the introduction of livestock to Argentina, when the Spanish first arrived in the 16th century,was the basis for widespread herds of feral cattle that were present during the 19th century and could virtually be harvested.Such production of animal products (hides and beef) was associated with scale economies and did not require much in theway of labor. These conditions may have increased the economic viability of large estates where labor was scarce and landabundant. In contrast, because the major crops produced in the expansion of the northern United States and Canada weregrains, whose production was relatively labor intensive and characterized by quite limited scale economies, the policy ofencouraging smallholding was effective. See Adelman 1994 and Engerman and Sokoloff 2002 for more discussion.

20

uninhabited public lands to transfer to private hands in the interest of bringing this public resource into

production and serving other general interests. In societies such as Mexico, however, the issues at stake

in land policy were very different. Good land was relatively scarce, and labor was relatively abundant.

Here the lands in question had long been controlled by Native Americans, but without individual private

property rights. Mexico was not unique in pursuing policies, especially near the end of the 19th and the

first decade of the 20th centuries, that had the effect of conferring ownership of much of this land to

large non-Native American landholders.37 Under the regime of Porfirio Díaz, Mexico effected a massive

transfer of such lands (over 10.7 percent of the national territory) between 1878 and 1908 to large

holders such as survey and land development companies, either in the form of outright grants for

services rendered by the companies or for prices set by decree.

In Table 8, we present estimates for these four countries of the fractions of household heads, or a

near equivalent, that owned land in agricultural areas in the late 19th and early 20th centuries. The

figures indicate enormous differences across the countries in the prevalence of land ownership among

the adult male population in rural areas. On the eve of the Mexican Revolution, the figures from the

1910 census suggest that only 2.4 percent of household heads in rural Mexico owned land. The number

is astoundingly low. The dramatic land policy measures in Mexico at the end of the 19th century may

have succeeded in privatizing most of the public lands, but they left the vast majority of the rural

population without any land at all. The evidence obviously conforms well with the idea that in societies

that began with extreme inequality, such as Mexico, institutions evolved so as to greatly advantage the

elite in access to economic opportunities, and they thus contributed to the persistence of that extreme

inequality.

In contrast, the proportion of adult males that owned land in rural areas was quite high in the

United States, at just below 75 percent in 1900. Although the prevalence of land ownership was

markedly lower in the South, where blacks were disproportionately concentrated, the overall picture is

one of a series of liberal land policies, leading up to the Homestead Act of 1862, providing broad access

to this fundamental type of economic opportunity. Canada had an even better record, with nearly 90

percent of household heads owning the agricultural lands they occupied in 1901. The estimates of

landholding in these two countries support the notion that land policies made a difference, especially

when compared to Argentina. The rural regions of Argentina constitute a set of frontier provinces, where

one would expect higher rates of ownership than in Buenos Aires. The numbers, however, suggest a

37. For further discussion of Mexico, see McBride 1923; Tannebaum 1929; and Holden 1994.

21

much lower prevalence of land ownership than in the two North American economies.38 Nevertheless,

all of these countries were far more effective than Mexico in making land ownership available to the

general population. The contrast between the United States and Canada, with their practices of offering

easy access to small units of land, and the rest of the Americas (as well as the contrast between

Argentina and Mexico) is consistent with the hypothesis that the initial extent of inequality influenced

the way in which institutions evolved and in so doing helped foster persistence in the degree of

inequality over time.

IV

There has long been debate over the impact of the European establishment of colonies around the

world that took place over centuries, beginning in the 1400s. Much of the controversy has been

concerned with issues such as how the long-term performance of the colonized areas and the colonizing

economies were affected by the exchange of resources and terms of trade between them and the

imbalance of military power. As we have argued here, however, one of the most fundamental

consequences of European colonization may have been in altering the composition of the populations in

the societies colonized. Because the efforts of the Europeans generally meant implanting communities

who were greatly advantaged over natives in terms of human capital and legal status, and because the

trajectories of institutional development were sensitive to the incidence of extreme inequality that often

followed, their activity had long lingering effects. Although more study is needed to identify all of the

mechanisms at work, it seems clear that colonies in the Americas with extreme inequality, as compared

to those with relative equality, were systematically more likely to evolve institutions that restricted

access to economic opportunities and to generate lower rates of public investment in schools and other

infrastructure considered conducive to growth. These patterns of institutional development, which tend

to yield persistence over time, may help to explain why a great many former European colonies that

began with extreme inequality have suffered poor economic outcomes.

38. Our work with the data from the 1914 census yields the same qualitative results. It is worth noting that the proportions offamilies that owned land are exaggerated by the 1895 census figures. A close examination of the manuscripts indicates thatdouble counting, in which both the husband and wife were listed as landowners, was prevalent in many parts of Argentina.

22

BIBLIOGRAPHY

Acemoglu, Daron and James A. Robinson. 2000. “Why Did Western Europe Extend the Franchise?Quarterly Journal of Economics 115 (Nov.): 1167-1199.

Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. “The Colonial Origins ofComparative Development: An Empirical Investigation.” American Economic Review 91 (Dec.): 1369-1401.

________________. 2002. “Reversal of Fortune: Geography and Institutions in the Making of theModern World Income Distribution, Quarterly Journal of Economics 117 (Nov.): 1231-1294.

Adelman, Jeremy. 1994. Frontier Development: Land. Labor, and Capital on Wheatlands of Argentinaand Canada. 1890-1914. Oxford.

Alesina, Alberto F. and Dani Rodrik. 1994. “Distributive Politics and Economic Growth.” QuarterlyJournal of Economics 109 (May): 465-490.

Castro, Donald. 1971. The Development of Argentine Immigration Policy, 1852-1914. Ann Arbor.

Coatsworth, John H. 1993. “Notes on the Comparative Economic History of Latin America and theUnited States.” In Walter L. Bernecker and Hans Werner Tobler, eds., Development andUnderdevelopment in America: Contrasts of Economic Growth in North and Latin America in HistoricalPerspective. New York.

Cubberley Ellwood P. 1920. The History of Education. Boston.

Davis, Lance E. and Robert A. Huttenback. 1986. Mammon and the Pursuit of Empire: The PoliticalEconomy of British Imperialism, 1860-1912. New York.

Deininger, Klaus and Lyn Squire. 1996. “A New Data Set and Measure of Income Inequality.” WorldBank Economic Review. 10 (September): 565-591.

Easterlin, Richard A. 1981. “Why Isn’t the Whole World Developed?” Journal of Economic History 41(March):1-19.

Easterly, William and Ross Levine. 2003. “Tropics, Germs, and Crops: The Role of Endowments inEconomic Development.” Journal of Monetary Economics 50 (Jan): 3-39.

Engerman, Stanley L. 1982. “Economic Adjustments to Emancipation in the United States and theBritish West Indies.” Journal of Interdisciplinary History 12 (Autumn): 191-220.

_______., 1983. “Contract Labor, Sugar and Technology in the Nineteenth Century.” Journal ofEconomic History 43 (September): 635-659.

_______., 1986. “Servants to Slaves to Servants: Contract Labour and European Expansion.” In P.C.Emmer, ed., Colonialism and Migration: Indentured Labor Before and After Slavery. Dordrecht.

Engerman, Stanley L., Stephen Haber, and Kenneth L. Sokoloff. 2000. “Inequality, Institutions, andDifferential Paths of Growth Among New World Economies.” In Claude Menard, ed., Institutions,

23

Contracts, and Organizations. Cheltenham.

Engerman, Stanley L., Elisa V. Mariscal, and Kenneth L. Sokoloff. 2002. “The Evolution of SchoolingInstitutions in the Americas, 1800-1925.” Unpublished working paper.

Engerman, Stanley L. and Kenneth L. Sokoloff. 1997. “Factor Endowments, Institutions, andDifferential Paths of Growth Among New World Economies: A View from Economic Historians of theUnited States.” In Stephen Haber, ed., How Latin America Fell Behind. Stanford.

___________. 2002. “Factor Endowments, Inequality and Paths of Development Among New WorldEconomies.” Economia 3 (Fall): 41-109.

_______. 2005. “The Evolution of Suffrage Institutions in the Americas,” Journal of EconomicHistory 65 (September): forthcoming.

Ferguson, Niall. 2003. Empire: The Rise and the Demise of the British World Order and the Lessons forGlobal Power. New York.

Gates, Paul W. 1968. History of Public Land Law Development. Washington, D.C.

Gibbins, H. de B. 1891. The History of Commerce in Europe. London.

Hobson, J.A. 1905. Imperialism. London.

Holden, Robert. 1994. Mexico and the Survey of Public Lands. The Management of Modernization,1876-1911. Dekalb.

Kupperman, Karen Ordahl. 1993. Providence Island. 1630-1641: The Other Puritan Colony. Cambridge.

Lal, Deepak. 2004. In Praise of Empires: Globalization and Order. New York.

Mc Bride, George McCutchen. 1923. The Land Systems of Mexico. New York.

McManus, Edgar J. 1973. Black Bondage in the North. Syracuse.

Morris, Henry C. 1900. The History of Colonization: From the Earliest Times to the Present Day.2 vols. London.

North, Douglass C. 1981. Structure and Change in Economic History. New York.

______________. 1988. “Institutions, Economic Growth and Freedom: An Historical Introduction.” InMichael Walker, ed., Freedom, Democracy, and Economic Welfare. Vancouver.

Perrson, Torsten and Guido Tabellini. 1994. “Is Inequality Harmful for Growth? Theory and Evidence.”American Economic Review 84 (June): 600-621.

Phillips, Charles E. 1957. The Development of Education in Canada. Toronto.

24

Platt, D.C.M. and Guido di Tella, eds. 1985. Argentina, Australia, and Canada: Studies in ComparativeDevelopment. 1870-1965. London.

Pomfret, Richard. 1981. The Economic Development of Canada. Toronto.

Rolling Stones. 1969. “You Can’t Always Get What You Want,” in Let It Bleed. London.

Sokoloff, Kenneth L. and Eric M. Zolt. 2004. “Inequality and Taxation: Evidence from the Americas”.Unpublished working paper.

Solberg, Carl E. 1970. Immigration and Nationalism: Argentina and Chile, 1890-1914. Austin.

______. 1987. The Prairies and the Pampas: Agrarian Policy in Canada and Argentina 1880-1913.Stanford.

Statesman’s Yearbook. 1900. Statesman’s Yearbook. New York.

Tannebaum, Frank. 1929. The Mexican Agrarian Revolution. New York.

Tocqueville, Alexis de. 1835. Democracy in America. Translated by George Lawrence and Edited byJ.P. Mayer. Reprinted in 1969. Garden City.

Turner, Frederick J. 1947. The Frontier in American History. New York.

Viotti da Costa, Emilia. 1985. The Brazilian Empire: Myths and Histories. Chicago.

Wallis, John Joseph. 2001. “A History of the Property Tax in America.” In Wallace E. Oates, ed.,Property Taxation and Local Government Finance: Essays in Honor of C. Lowell Harriss. Cambridge.

Wilson, J. Donald, Stamp, Robert M., and Audet, Louis-Philippe. 1970. Canadian Education: A History.Scarborough.

25

TABLE 1

REGIONAL DISTRIBUTION OF BRITAIN’S COLONIAL DOMAIN OVER THE 19TH CENTURY

1815 1860 1880 1890-91 1899

(000) (%) (000) (%) (000) (%) (000) (%) (000) (%)

EuropePop 340 0.3 387 0.3 175 0.1 191 0.1 204 0.1Land -- -- 1 0.0 0 0.0 0 0.0 0 0.0

AsiaPop 124,200 98.3 137,279 94.6 256,149 95.6 288,436 94.7 291,587 94.4Area -- -- 876 34.5 963 12.5 1,827 20.1 1,827 19.6

AfricaPop 244 0.2 836 0.6 2,718 1.0 4,963 1.6 4,932 1.6Area -- -- 130 5.1 278 3.6 342 3.8 368 3.9

AmericasPop 1,600 1.3 4,227 2.9 6,016 2.2 6,708 2.2 7,260 2.3Area -- -- 954 37.5 3,359 43.7 3,769 41.4 3,953 42.4

AustralasiaPop 25 0.0 2,401 1.6 2,877 1.1 4,417 1.4 5,009 1.6Area -- -- 580 22.8 3,084 40.1 3,175 34.8 3,176 34.1

____________________________________________________________________TOTAL BRITISH DOMAIN

Pop 126,408 145,129 267,935 304,716 308,992 Area -- 2,541 7,685 9,113 9,324

BRITISH DOMAIN AS PERCENTAGE OF ALL EUROPEAN COLONIES, c. 1890

Pop 70.5 %

Land 50.6

_____________________________________________________________________________

Notes and Sources: The land area is measured in square miles. Morris (1900), p. 88. These estimates of thecolonial domain of Britain do not include some areas that many might consider to have been colonies. Aprominent example is Ireland, which is not included because the Union Act of 1801 formed a kingdom of GreatBritain and Ireland into the United Kingdom. In 1891, Ireland has a population of 4,704,750, and a land area of32,531 square miles.

26

TABLE 1a

COLONIAL DOMAINS OF BRITAIN, c. 1899________________________________________________________________________

Area in Sq. Miles Population Date of(000) (000) Acquisition

________________________________________________________________________INDIA

British India 1,068.3 221,173 1601-1856Feudatory States 731.9 66,050 ---

_____________________________Total INDIA 1,800.2 287,223

EUROPEGibraltar -- 24 1704Malta and Gozo 0.1 180 1800

OTHER ASIAAden and Perim 0.1 42 1839Ceylon 25.3 3,449 1795Hong Kong 0.4 354 1842Labuan -- 6 1846Straits Settlement 1.5 512 1819

AFRICAAscension -- -- 1815Basutoland 10.3 250 1868-1883Cape Colony 276.8 1,788 1806Mauritius 0.7 338 1810Natal and Zululand 35.0 902 1824St. Helena -- 5 1651

WEST AFRICAGambia -- 14 1631Gold Coast 40.0 1,474 1661Lagos 0.9 86 1787Sierra Leone 4.0 75 1789

AMERICASBermudas -- 16 1609Canada 3,653.9 5,186 1763Falklands/So. Georgia Isl. 7.5 2 1833British Guiana 109.0 286 1803British Honduras 7.6 35 1670Newfoundland/Labrador 162.2 202 1497

WEST INDIESBahamas 4.5 53 1629Jamaica/Turks Islands 4.4 733 1655Barbados 0.1 190 1605Leeward Islands 0.7 128 1600sWindward Islands 0.8 155 1600sTrinidad and Tobago 1.9 274 1763-1797

27

AUSTRALASIAFiji 7.7 122 1874New Guinea 90.5 350 1884New South Wales 310.7 1,357 1788New Zealand 104.5 796 1840Queensland 668.5 499 1859South Australia 903.7 363 1836Tasmania 29.4 171 1803Victoria 87.9 1,177 1850Western Australia 975.9 168 1829

_______________________________________________

Total British Colonies 7,523.8 21,769Grand TotalBRITISH COLONIES 9,324.0 308,992AND INDIA

Notes and Sources: Morris (1900), pp. 88-89.

28

TABLE 2

REGIONAL DISTRIBUTION OF FRANCE’S COLONIAL DOMAIN OVER THE 19TH CENTURY

1815 1860 1880 1890-91 1899 (000) (%) (000) (%) (000) (%) (000) (%) (000) (%)

Asia Pop 179 35.9 222 6.6 3,334 44.3 18,000 51.1 22,679 40.2 Area -- 1.1 -- 0.1 69 9.2 201 8.4 363 9.7

Africa Pop 95 19.0 2,800 83.0 3,702 49.2 16,800 47.7 33,257 59.0 Area 1 6.0 186 76.9 625 83.3 2,129 89.2 3,320 88.7

Americas Pop 225 45.1 300 8.9 391 5.2 373 1.1 384 0.7 Area 16 92.9 48 19.8 48 6.4 48 2.0 48 1.3

Oceania Pop -- -- 50 1.5 94 1.2 72 0.2 82 0.1 Area -- -- 8 3.3 9 1.2 9 0.4 9 0.2______________________________________________________________________________

TOTAL FRENCH DOMAIN

Pop 499 3,372 7,521 35,245 56,402 Area 17 242 750 2,387 3,741

FRENCH DOMAIN AS PERCENTAGE OF ALL EUROPEAN COLONIES, c. 1890

Pop 8.2 % Land 13.3 %_____________________________________________________________________________

Notes and Sources: The land area is measured in square miles. Morris (1900), p. 88.

29

TABLE 2a

COLONIAL DOMAINS OF FRANCE, c. 1899___________________________________________________________________________

Area in Sq. Miles Population Date of (000) (000) Acquisition

___________________________________________________________________________ASIA

India 0.2 279 1679Annam 88.8 5,000 1884Cambodia 40.5 1,500 1862Cochin-China 23.1 2,400 1861Tonquin (with Laos) 210.4 13,500 1704

AFRICAAlgeria 184.5 4,430 1830Algerian Sahara 123.5 50 --Tunis 50.8 1,500 1881Sahara Region 1,684.0 2,500 --Senegal 120.0 2,000 1637Sudan 300.0 2,500 1880Ivory Coast 100.0 2,500 1843Dahomey 50.0 1,000 1893Congo and Gabon 425.0 12,000 1884French Guinea 48.0 1,000 1843Obok and Somali Coast 5.0 22 1864Réunion 1.0 173 1649Comoro Isles 0.6 53 1886Mayotte 0.1 11 1843Nossi-Bé 0.1 10 1841Madagascar 227.8 3,500 1896

AMERICASGuiana 46.9 22 1626Guadeloupe and Dependencies 0.7 167 1634Martinique 0.4 188 1635St. Pierre and Miquelon 0.1 6 1635

OCEANIANew Caledonia/ Dependencies 7.7 53 1854Other French Establishments 1.5 29 1841-1881

___________________________________________

Total All French Colonies 3,740.7 68,403

Notes and Sources: Morris (1900), pp. 88-89.

30

TABLE 3

COLONIAL DOMAINS OF SELECTED OTHER EUROPEAN POWERS, c. 1890_____________________________________________________________________________

Area in Sq. Miles Population(000) (000)

_____________________________________________________________________________DENMARK Europe

Far e Islands 0.5 11.2Greenland (coasts) 46.7 9.8Iceland 39.8 72.4

West IndiesSt. John and St. Thomas -- 15.4Santa Cruz (St. Croix) 0.1 18.4

Total DENMARK 87.1 127.2 % of All European Colonies, 0.5% 0.0% c. 1890 _______________________________________________

GERMANY Africa

Cameroons 130.0 500.0Damaraland/ Luderitzland 342.0 250.0German East Africa 344.8 1,755.0Tagoland 16.0 500.0

PacificBismarck Archipelago 18.2 188.0Kaiser Wilhelm’s Land 70.3 109.0 (New Guinea)Marshall Archipelago -- 10.0Solomon Islands 8.5 80.0

Total GERMANY 929.8 3,392.0 % of All European Colonies, 5.2% 0.8% c. 1890 ______________________________________________ITALY Africa

Eritrea 56.1 659.0Abyssinia 189.0 4,500.0Somali Coast 70.0 210.0

Total ITALY 315.1 5,369.0 % of All European Colonies, 1.7% 1.2% c. 1890 ________________________________________________

31

_____________________________________________________________________________

Area in Sq. Miles Population(000) (000)

_____________________________________________________________________________NETHERLANDS East Indies

Borneo 203.7 1,073.5Celebes 72.0 2,000.0Java (with Madura) 50.8 21,974.2Moluccas 42.4 353.0Sumatra 170.7 200.0Other Islands 29.0 235.0

West IndiesAll Islands∗ 0.4 45.1

Total NETHERLANDS 569.0 25,880.8 % of All European Colonies, 3.2% 6.0% c. 1890 _______________________________________________PORTUGAL Africa

Portuguese Guinea 11.6 150.0Angola 603.0 3,600.0Madeira 0.3 132.0Cape Verde Islands 1.5 111.0Mozambique 293.0 1,500.0Principe/St. Thomas’ Island 0.4 21.0

AsiaGoa, Daman, etc. 1.6 481.5Macao -- 66.0Timor, etc. 6.3 300.0

Total PORTUGAL 917.7 6,361.5 % of All European Colonies, 5.1% 1.5% c. 1890 ___________________________________________________SPAIN Africa

Canaries 2.8 288.0Gulf of Guinea 0.9 50.0N.W Saharan Seaboard and Span. Morocco 243.0 106.0

AsiaPhilippine Islands 114.3 5,561.2

Other Islands♣ 2.0 119.7

∗ Includes Buen Ayre, Curaçao, Oruba, Saba, St.Eustatius, and St. Martin.♣ Includes Caroline, Pelew, Marianne, and Sulu Islands

32

West IndiesCuba (and Pinos) 43.2 1,521.7Puerto Rico 3.6 784.7

Total SPAIN 409.8 8,431.3 % of All European Colonies, 2.3% 2.0% c. 1890 _________________________________________________

Notes and Sources: Gibbins (1891), Appendix B. Not all European countries with colonies are included here. Anoutstanding example is Belgium, and its Belgian Congo colony in particular. After 1883, the Congo was owned,essentially by King Leopold II, and it was not until 1908 that there was a legal connection with Belgium (theCongo, now Zaire, gained its independence in 1960). In 1900, its population was 8 million, of which 1630residents were of European descent. Its land area covered 900,000 square miles. The table does not specify thecolonies of Russia (including Finland), of Turkey (including Egypt, the Egyptian Sudan, Bulgaria, Crete, andSamos), and of Austria-Hungary, nor is it inclusive of quite all of the colonies of the European countries explicitlyincluded (such as Ireland for Britain, or Libya for Italy). In 1899, the Congo and these latter colonies together hadat least a land area of 3,279,494 square miles and a population of 59,221,660. If we include the Spanish coloniesthat were taken over by the United States in 1898, we estimate (somewhat conservatively in our view) that18,008,094 square miles of land were encompassed in European colonies c. 1890, and that roughly 432,306,000people resided in these areas. We use these latter figures as the basis for the calculations of the proportions of thetotal land area and population in European colonies reported here.

33

TABLE 4

COMPOSITION OF POPULATIONS IN EUROPEAN COLONIAL DOMAINS_____________________________________________________________________________

Non-Whites Whites Ratio of Whites to Others

_____________________________________________________________________________BRITAIN

1850

Europe 15 347,680 23,179.000Asia 97,356,000 60,162 0.001Australasia 155,000 131,800 0.850Africa 221,000 67,068 0.302North America 120,000 1,410,000 11.750South America 199,571 3,958 0.020West Indies 639,708 104,040 0.163 TOTAL 98,692,394 2,125,116 0.215

FRANCE1926

Africa (all) 32,883,000 1,331,400 0.040Americas (all) 492,500 48,500 0.098Asia 20,415,000 23,500 0.001Oceania 71,600 16,400 0.229 TOTAL 53,862,100 1,419,800 0.026

GERMANY1913

Africa 12,084,436 22,405 0.002Pacific/Oceania 961,000 6,454 0.007

ITALY1931

Africa 2,380,560 69,441 0.029

PORTUGAL1950

Africa 11,879,140 137,542 0.012

______________________________________________________________________________

34

TABLE 5LAWS GOVERNING THE FRANCHISE AND THE EXTENT OF VOTING IN SELECTED

COUNTRIES, 1840-1940

Lack of Secrecy WealthRequirement

LiteracyRequirement

Proportion ofthe

In Balloting PopulationVoting

1840-80--

Chile 1869 N Y Y 1.6%1878 N N N39 --

Costa Rica 1890 Y Y Y --Ecuador 1848 Y Y Y 0.0

1856 Y Y Y 0.1Mexico 1840 Y Y Y --Peru 1875 Y Y Y --Uruguay 1840 Y Y Y --

1880 Y Y Y --Venezuela 1840 Y Y Y --

1880 Y Y Y --Canada 1867 Y Y N 7.7

1878 N Y N 12.9

United States 185040 N N N 12.91880 N N N 18.3

1881-1920Argentina 1896 Y Y Y 1.8%41

1916 N N N 9.0Brazil 1894 Y Y Y 2.2

1914 Y Y Y 2.4Chile 1881 N N N 3.1

1920 N N Y 4.4Colombia 191842 N N N 6.9Costa Rica 1912 Y Y Y --

1919 Y N N 10.6Ecuador 1888 N Y Y 2.8

1894 N N Y 3.3

39 After eliminating wealth and education requirements in 1878, Chile instituted a literacy requirement in 1885, which seems to have been responsible for a

sharp decline in the proportion of the population who were registered to vote.

40 Three states, Connecticut, Louisiana, and New Jersey, still maintained wealth requirements at 1840, but eliminated them soon afterwards. All states

except for Illinois and Virginia had implemented the secret ballot by the end of the 1840s.

41 This figure is for the city of Buenos Aires, and likely overstates the proportion who voted at the national level.

42 The information on restrictions refers to national laws. The 1863 Constitution empowered provincial state governments to regulate electoral affairs.

Afterwards, elections became restricted (in terms of the franchise for adult males) and indirect in some states. It was not until 1948 that a national law

established universal adult male suffrage throughout the country. This pattern was followed in other Latin American countries, as it was in the U.S. and

Canada to a lesser extent.

35

Mexico 1920 N N N 8.6Peru 1920 Y Y Y --Uruguay 1900 Y Y Y --

1920 N N N 13.8Venezuela 1920 Y Y Y --Canada 1911 N N N 18.1

1917 N N N 20.5United States 1900 N N Y43 18.4

1920 N N Y 25.1

1921-40Argentina 1928 N N N 12.8%

1937 N N N 15.0Bolivia 1951 ? Y Y 4.1Brazil 1930 Y Y Y 5.7Colombia 1930 N N N 11.1

1936 N N N 5.9Chile 1920 N N Y 4.4

1931 N N Y 6.51938 N N Y 9.4

Costa Rica 1940 N N N 17.6Ecuador 1940 N N Y 3.3Mexico 1940 N N N 11.8Peru 1940 N N Y --Uruguay 1940 N N N 19.7Venezuela 1940 N Y Y --Canada 1940 N N N 41.1United States 1940 N N Y 37.8

Notes and Sources: Engerman, Haber, and Sokoloff 2000.

43 Eighteen states, 7 southern and 11 non-southern, introduced literacy requirements between 1890 and 1926. These restrictions were directed primarily at

Blacks and immigrants.

36

TABLE 6LITERACY RATES IN THE AMERICAS, 1850-1950

Year Ages Rate

Argentina 1869 +6 23.8%1895 +6 45.61900 +10 52.01925 +10 73.0

Bolivia 1900 +10 17.0

Brazil 1872 +7 15.81890 +7 14.81900 +7 25.61920 +10 30.0

British Honduras 1911 +10 59.6(Belize) 1931 +10 71.8

Chile 1865 +7 18.01875 +7 25.71885 +7 30.31900 +10 43.01925 +10 66.0

Colombia 1918 +15 32.0

Costa Rica 1892 +7 23.61900 +10 33.01925 +10 64.0

Cuba 1861 +7 23.8(38.5,5.3)*

1899 +10 40.51925 +10 67.0

Guatemala 1893 +7 11.31925 +10 15.0

Honduras 1887 +7 15.21925 +10 29.0

Jamaica 1871 +5 16.31891 +5 32.01911 +5 47.2

Mexico 1900 +10 22.21925 +10 36.0

Paraguay 1886 +7 19.31900 +10 30.0

Peru 1925 +10 38.0

37

Uruguay 1900 +10 54.01925 +10 70.0

Venezuela 1925 +10 34.0

Canada 1861 All 82.5Eng-majority counties 1861 All 93.0Fr- majority counties 1861 All 81.2

United StatesNorth Whites 1850 +10 96.9South Whites 1850 +10 91.5All 1870 +10 80.0

(88.5,21.1)*1890 +10 86.7

(92.3,43.2)*1910 +10 92.3

(95.0,69.5)*

*The figures for Whites and Non-Whites are reported respectively within parentheses.

Sources: Engerman, Haber, and Sokoloff 2000.

38

TABLE 7

DISTRIBUTION OF TAX REVENUES ACROSS LEVELS OF GOVERNMENT DURINGTHE 19th AND 20th CENTURIES: BRAZIL, CHILE COLOMBIA, MEXICO, CANADA,

AND THE UNITED STATES

NationalGovernment(%)

ProvincialGovernments(%)

Municipalitiesor other Local(%)

Brazil1826 30.8 69.2 01856 79.5 17.1 3.31860 78.2 18.2 3.51885/86 76.3 18.5 5.2Chile1913 85.8 -- 14.21915 82.7 -- 17.31920 85.3 -- 14.7Colombia1839 88.4 2.9 8.71842 91.8 1.6 6.71850 85.4 8.7 5.81870 46.6 30.8 22.61894 60 32 81898 66.7 28.6 4.8Mexico1882 69.1 19.5 11.51890 74.7 16.3 91900 67.3 19.8 12.91908 70.6 17.1 12.3Canada1933 42.5 17.9 39.61950 68.7 18.7 12.61960 62.8 20.7 16.4United States1855 25.5 17.4 57.11875 39.6 16.4 44.01895 36 14 501913 29.1 13.2 57.61927 35.5 18 46.51950 68.3 17.3 14.4

Sources and Notes: Sokoloff and Zolt 2004.

39

TABLE 8

LANDHOLDING IN RURAL REGIONS OF MEXICO, THE UNITED STATES, CANADA, ANDARGENTINA DURING THE EARLY 1900s

Country, year, and region Proportion of household heads who ownlanda

Mexico, 1910North Pacific 5.6North 3.4Central 2.0Gulf 2.1South Pacific 1.5Total rural Mexico 2.4

United States, 1900North Atlantic 79.2South Atlantic 55.8North Central 72.1South Central 51.4Western 83.4Alaska/Hawaii 42.1Total United States 74.5

Canada, 1901British Columbia 87.1Alberta 95.8Saskatchewan 96.2Manitoba 88.9Ontario 80.2Quebec 90.1Maritimeb 95.0Total Canada 87.1

Argentina, 1895Chaco 27.8Formosa 18.5Missiones 26.7La Pampa 9.7Neuquén 12.3Río Negro 15.4Chubut 35.2Santa Cruz 20.2Tierra del Fuego 6.6

Notes and Sources: Engerman and Sokoloff 2002.a. Landownership is defined as follows: in Mexico, household heads who own land; in the US, farms that are owner operated; in Canada, total

occupiers of farm lands who are owners; and in Argentina, the ratio of landowners to the number of males between the ages of 18 and 50.b. The Maritime region includes Nova Scotia, New Brunswick, and Prince Edward Island.