Hostname: page-component-7b9c58cd5d-g9frx Total loading time: 0 Render date: 2025-03-14T13:34:35.785Z Has data issue: false hasContentIssue false

THE NEW ECONOMIC HISTORY OF AFRICA*

Published online by Cambridge University Press:  07 September 2009

A. G. HOPKINS
Affiliation:
University of Texas at Austin
Rights & Permissions [Opens in a new window]

Abstract

The purpose of this article is to promote the revival of African economic history. Poverty, the most pressing issue confronting the continent, has received world-wide publicity in recent years. Yet historians have continued to neglect the history of economic development, which is central to the study of poverty, in favour of themes that have their origins in the Western world rather than in Africa. However, there is now an exceptional opportunity to correct the balance. Unknown to most historians, economists have produced a new economic history of Africa in the course of the past decade. This article introduces and evaluates two of the most important contributions to the new literature: the thesis that Africa has suffered a ‘reversal of fortune’ during the last 500 years, and the proposition that ethnic fragmentation, which has deep historical roots, is a distinctive cause of Africa's economic backwardness. These arguments are criticized on both methodological and empirical grounds. But they are also welcomed for their boldness, their freshness and their potential for re-engaging historians in the study of Africa's economic past – not least because it is relevant to Africa's economic future.

Type
Research Article
Copyright
Copyright © Cambridge University Press 2009

INTRODUCTION

What has happened to the study of Africa's economic history? The subject emerged in the 1960s, colonized the greater part of the continent during the 1970s, and for a decade or so was thought to be essential to understanding the long-run causes of underdevelopment. By the close of the 1980s, however, economic history was in failing health; in the 1990s its public appearances were limited; today, it seems to have died, unlamented, from that most mortifying of scholarly ailments – neglect. Before tolling the bell, however, it is worth checking to see if resuscitation is still possible, and, if so, what useful purpose it might serve. This article sounds a trumpet. It aims to alert historians to missed opportunities, to attract them to the continuing possibilities of studying Africa's economic past, and thus to signal the prospective revival of the subject. The moment, moreover, is apposite: during the past decade, economists seeking solutions to Africa's development problems have created a new type of economic history. Their work is virtually unknown to specialists on African history and has yet to feature in any of the three most relevant scholarly outlets: this Journal, the International Journal of African Historical Studies and African Economic History.Footnote 1 It is time that historians took notice.

These very general statements need refining. The claim made here is not that the study of Africa's economic history has ceased: that would be an unjust judgement on research of the highest quality undertaken on both external relations, notably the slave trade, and the domestic economy since the 1980s. It is equally apparent, however, that economic history has lost status and visibility during the last twenty years. It is no longer considered to be at the ‘cutting edge’ of historical studies. A subject declines in relative importance when it ceases to generate the overarching propositions that capture the attention of a sizeable segment of the profession. The excellent research completed during the last twenty years has generally taken the form of micro-studies or thematic enquiries that are either unrelated or have yet to be connected.Footnote 2 The broadest studies with the largest theses were all produced by the first generation of scholars who entered African history at the time of independence. The following examples illustrate the point: my own Economic History of West Africa, published in 1973, has not been replaced; the last and only substantial economic history of sub-Saharan Africa, by Ralph Austen, appeared in 1987; Charles Feinstein's economic history of South Africa, though published recently, was completed just before his death at the age of 72.Footnote 3 The only book on poverty in Africa written by a historian, John Iliffe, appeared in 1987.Footnote 4 The most recent survey of the state of the subject was published by Patrick Manning, also in 1987.Footnote 5 In retrospect, it can be seen that 1987 marks not only the high point but also the culmination of an endeavour that began in the 1960s. It is now more than twenty years since historians themselves produced big arguments attempting to understand Africa's long-run economic development and continuing poverty.

The explanation of the relative decline of the subject lies in the convergence of trends that had entirely separate origins. Diminishing interest in African economic history was partly a product of the decay of economic history in general. This development, in turn, stemmed on the one hand from the withering of the radical inspiration of the 1960s and 1970s, and on the other from the process of cleavage in universities that divided history from economics and took the economic segment into technical complexities that escaped the interest, as well as the grasp, of most historians. A second and wholly different influence, the rise to prominence of postmodernism, shifted attention in the 1990s from material to cultural considerations. This was the price paid for enabling elements of the radical tradition to survive: Antonio Gramsci validated the connection to Marx but did so by emphasizing the importance of forms of cultural hegemony that harked back to Hegelian idealism and associated anti-positivist strains of thought. Inspiration drawn from Gramsci was reinforced by non-Marxist and predominantly literary sources, notably Edward Said's Orientalism, which pointed in the same direction.Footnote 6 The result turned a generation of young researchers away from materialism, macro-historical projects and what was seen to be naïve realism, and towards the study of how European representations of indigenous societies had been used to perpetuate colonial control. This approach, whatever its merits, was not one that encouraged the study of economic history.Footnote 7 Accordingly, the subject found itself caught between positivist economic theory, which was remote and unhistorical, and cultural analysis, which was removed from economic considerations and, though historical in intent, was sometimes defective in its use of historical sources. It is scarcely surprising that economic history languished.

In effect, historians abolished poverty by the simple device of taking it off the agenda. Poverty, however, has not been eliminated. On the contrary, the problem has deepened. Much of Africa has failed to achieve any growth in per capita incomes since 1960; some countries have seen incomes decline.Footnote 8 At the close of the twentieth century, the average life expectancy of a child born in sub-Saharan Africa in 1980 was only 48 years; a typical African mother had only a 30 per cent chance of seeing all her children survive to the age of five; daily calorie intake was only 70 per cent of that of Latin America and East Asia. It is no wonder that The Economist referred to Africa in 2000 as ‘the hopeless continent’ that would never escape recurring crises unless it developed, among other qualities, a broad and sturdy civic consciousness to control tribal rivalries, corruption and civil war.Footnote 9 Without a solution to these problems, the journal concluded, ‘the whole world might just give up on the entire continent’.Footnote 10 We do not have to accept this gloomy analysis to see that impoverished societies scattered across the continent are manifestations of human misery, sources of instability, and a moral reproach on a scale even greater than that represented by the slave trade in the nineteenth century.

Africa's poor economic performance and frequent human disasters have aroused increasing international attention. The slogan ‘Make Poverty History’ was devised to accompany a mass demonstration organized to influence the meeting of the G8 states in Scotland in 2005.Footnote 11 The campaign, led by Bob Geldof, was memorable for helping to raise public awareness of the plight of Africa as well as for holding the biggest rock concert in history. Momentarily, at least, Africa became an issue that disturbed the public conscience and provided a platform for celebrities to shine their light on what the explorer Stanley had labelled ‘the Dark Continent’. Subsequently, Geldof was quoted approvingly on the front cover of the Commission for Africa's report, Our Common Interest, which appeared in 2005, while Bono contributed a foreword to Jeffrey Sachs's book The End of Poverty, in the same year. More recently still, the publication in 2007 of Paul Collier's widely noticed volume, The Bottom Billion, added to the publicity given to the continuing drama of Africa's poverty. Collier suggested that it is no longer accurate to refer to the underdeveloped countries as if they were homogeneous, as used to be the practice when the term ‘Third World’ was in vogue. The 5 billion people in the underdeveloped world had ceased to share the same predicament. Since the 1980s, global poverty has been reduced for the first time in history. Now 4 billion people inhabit countries that are developing, even though their standards of living remain low. The serious, intractable problem, so Collier claimed, lies with the bottom billion human beings, most of whom live in Africa in conditions that defy established approaches to development.

Yet, even this unprecedented degree of publicity has so far failed to shift historians from their current preoccupations with questions that lie outside issues of poverty and development. The continuing neglect of economic history has both weakened the study of African history as a whole and distanced historians from important debates over the formulation of development policy. Nevertheless, historians now have an almost unprecedented opportunity to re-engage with Africa's economic history. While historians have been preoccupied with topics such as race, gender and identity, other specialists on Africa have turned to history to try to understand the origins of present-day poverty. Economists have been creating a new type of economic history of Africa for at least a decade. Other social scientists have begun to follow their lead. This enterprise is bold and wide-ranging. Its approach to history involves both careful use of secondary sources and heroic leaps of faith; its findings mix originality with rediscoveries of what is already familiar. Its flaws, however, do not diminish its importance, and it demands what it has yet to receive: the attention of historians themselves.

THE NEW INSTITUTIONAL HISTORY

The new work deals with a variety of themes and periods and is written by different sets of authors. Nevertheless, it shares a common founding assumption: the belief that institutions lie at the heart of development problems. Consequently, it is convenient for present purposes to gather the research undertaken so far under the title ‘new institutional history’. This term is already associated with the approach of economists such as Ronald Coase, and historians such as Douglass North, and has increased its influence on the economics profession during the last half-century.Footnote 12 The label ‘new’ was adopted to distinguish the initiative from the institutional school associated with Thorstein Veblen, Wesley Mitchell and John Commons that flourished early in the twentieth century.Footnote 13 Whereas the old institutionalists sought to incorporate economics and other social sciences into history, their successors aimed at subjecting history to the principles of economics. The endeavour, however, was not simply an exercise in intellectual imperialism. The new institutionalists were also strongly critical of the way mainstream economics had become abstracted from the real world. They planned to change the content and the approach of the subject by introducing themes that had received little attention in orthodox teaching. Initially, the chief focus was on property rights and the firm. Subsequently, the approach widened to include the provision of public goods (such as infrastructure, defence, law, education and health), transactions costs (such as information flows) and social relations, political processes and belief systems. Institutions, broadly conceived, are seen as influences that establish the rules of the economic game by shaping individual and social behaviour. Organizations such as firms, political parties and voluntary associations are the players who operate the system.

The perspective is evidently one that, in principle, is congenial to historians because it extends beyond the narrow focus of any single sub-branch of the subject. In the 1980s and 1990s, a handful of historians applied the new institutional approach to the history of Africa by exploring topics such as property rights, bargaining theory, credit-protection, rent-seeking and transactions costs.Footnote 14 The latest manifestation of new institutional history, however, has been drawn to Africa by the limited success of orthodox development theory and the deepening problem of poverty in the continent. When Patrick Manning surveyed the state of African economic history in 1987, he observed, correctly, that the World Bank and associated economists believed that history, or at least the slice of history that was relevant to development policy, began in 1960, Africa's year of independence.Footnote 15 This date was the starting point of the highly influential Berg Report, published in 1981, which criticized the interventionist policies adopted by African governments and recommended market-driven measures of the type that were gaining influence in the developed world at that time.Footnote 16 These measures were applied to Africa through programmes of ‘structural adjustment’ that became summarized, colloquially, as ‘getting prices right’.Footnote 17 The orthodoxy of the day took little account of the innovative work of scholars such as Robert Bates, Sara Berry and Patrick Manning, who were already exploring, in different ways, the relationship between history, institutions and development.Footnote 18

Attitudes began to change, however, as experience revealed the limits to market solutions. As Africa's ‘growth tragedy’ deepened, development experts were prompted to reconsider the role of public agencies and the political process in general. By the 1990s, a new catch-phrase, ‘bringing the state back in’, had come into vogue.Footnote 19 The scope of development policy widened to incorporate what became known as ‘governance’, and policy recommendations allowed institutions and organizations to come to the fore. Some specialists, looking for fresh inspiration, even suggested that the experience of East Asia might be more relevant to Africa than that of the United States.Footnote 20

The prominence attached to the role of institutions enabled the new economic history of Africa to claim the attention of development specialists. ‘Good’ institutions, as the literature calls them, matter because they are crucial to providing incentives for investment, which is essential for development. The presence of inappropriate or weak institutions, on the other hand, creates uncertainty among potential investors because it increases the risk of expropriation by governments and elites, distorts public investment by directing it to political ends, encourages corruption and promotes the likelihood of civil strife. The new research programme begins with the hypothesis that institutional variables are more significant than other variables in explaining development. As applied to Africa, the hypothesis further assumes that institutional change is gradual. This assumption calls for historical investigation because current institutions are the product of a past that may extend beyond the short period of colonial rule into precolonial history and even into prehistory. The principal technique, cross-country regression analysis, aims to measure the relative weights of the causes of the problem that needs to be explained.Footnote 21

The emphasis on institutions has to contend with a number of alternatives, the most prominent of which stresses the primacy of initial conditions, notably geography, climate and disease.Footnote 22 There are also important new lines of enquiry in development economics that are institutional in orientation but, so far, lack a prominent historical dimension. The work of Abhijit Banerjee, Esther Duflo, Michael Kramer and Christopher Udry, for example, has greatly elevated the status of micro-economic studies that test the effectiveness of policies derived from macro-economic assumptions.Footnote 23 The work reviewed here, however, justifies the priority it receives not only because it has done so much to inspire the new debate on Africa's continuing poverty but also because it has led the way in making historical evidence central to contemporary issues. The propositions advanced by the new economic history of Africa are well illustrated by the two most influential contributions made so far. The authors are noted economists who are closely involved with development issues and are connected to organizations that influence policy-making.Footnote 24 Their theories have now generated an extensive literature among social scientists.Footnote 25

REVERSING FORTUNES

The boldest and most comprehensive historical argument advanced by the new institutionalists is that formulated by Daron Acemoglu, Simon Johnson and James Robinson in 2002.Footnote 26 The central claim arising from their research is that many of the countries colonized by European powers, including Africa, were relatively rich in 1500 but have now become relatively poor.Footnote 27 The authors recognize the problem of securing adequate statistical data to measure income levels in diverse parts of the world in 1500. The solution they propose adopts two conjoined proxies for relative prosperity: population density and levels of urbanization. The article then assembles the historical information on both proxies and converts them into measurable indices. The reversal of fortune, as they term it, cannot be explained by geographical and other environmental influences because these are comparatively persistent features of the natural endowment. The reversal is to be understood instead as a consequence of the varied character of European expansion overseas and the differential effect it had on institutions.

The argument derived from this evidence can be summarized as follows. Parts of the world such as the Americas and Australasia, which were relatively poor in 1500, tended also to be sparsely populated (even before European influences were felt) and to have low levels of urbanization. These conditions enabled Europeans to settle and install appropriate institutions that provided incentives for investment. Areas such as Africa and Asia, on the other hand, which were relatively rich in 1500, inhibited white settlement. Their larger populations could resist European encroachment and also occupied epidemiological zones that imposed a high rate of mortality on Europeans. In such regions the European presence was either marginal, in which case unproductive institutions remained in place, or took the form of colonial rule, which introduced extractive types of economic control that were poorly suited to long-term development. The Atlantic slave trade is a particularly telling case: it reduced the population from a level that was previously relatively high and encouraged the development of institutions that were inimical to long-term growth. The greater part of the reversal, nevertheless, occurred in the nineteenth century, when favoured parts of the world had established appropriate institutions that could take advantage of the technological and other innovations of the period. Elsewhere, unproductive institutions, once established, tended to persist. Nevertheless, institutions can be long-standing without being immutable. The task of policy is to adapt inappropriate institutions to present development needs.

The reversal of fortune thesis has stimulated further contributions by social scientists. One set of comments focuses on the methodology underlying the exercise. The key assumption made by Acemoglu and his colleagues is that settler mortality can be used as an instrumental variable because it is sufficiently independent of other variables to carry causal weight in explaining differing levels of prosperity in the ex-colonial world today.Footnote 28 The merit of instrumental variables is that they are uncontaminated by the problem that needs to be explained and so are capable of overcoming common estimation difficulties, such as cross-correlation and reverse causation. Settler mortality qualifies because it explains why Europeans populated some colonies rather than others in the era before the medical revolution of the late nineteenth century, but has long ceased to have a direct effect on current levels of wealth in ex-colonial states. The causal link explaining the reversal of fortunes is provided indirectly by the varied quality of the institutions that arose from the distribution of European settlement throughout the world. This variation determined the strength of property rights, the concomitant risk of capital expropriation, and ultimately the level of economic development.

The key assumption has attracted a good deal of criticism. It has been pointed out that settler mortality is a weak instrumental variable: it identifies an external source of variation in the problem that needs explaining but may not itself account for a large part of the cause.Footnote 29 This thought has been pursued by Patrick O'Brien, whose comprehensive analysis of the historical data shows that the potential for economic development in colonial and independent states in Asia and Africa between 1815 and 1948 depended far less on the type of government present than on a multiplicity of other influences, notably natural and human resources and access to international markets.Footnote 30 An alternative approach has reasserted the importance of the geographical environment.Footnote 31 The claim here is not that geography creates an implacable destiny, but that its influence, too, is felt through institutions as well as, in some formulations, directly. Different resource endowments encourage different institutions: plantations may be favoured in one region, small-holdings in another; each tends to be associated with distinctive sociopolitical formations.Footnote 32 Some scholars have proceeded to devise alternative measures that expand the list of desirable institutions; others have considered how to rank the contributions that have been made to development by different kinds of institutions. To the extent that there is agreement on the state of the debate, it is accepted that institutions are entwined with one another and with non-institutional influences in ways that cannot usefully be disentangled by referring to first principles or assumed best practice, but need to be placed in specific contexts that reflect the particularities, including the history, of each case.Footnote 33

A further methodological problem arises from the application of cross-country regression analysis.Footnote 34 This technique has established itself as both a valuable and a controversial econometric tool. It holds out the promise of making global comparisons of the effects of policy on development, but it carries risks in addition to the standard warnings that accompany the use of linear and multiple regressions. One danger is that comparisons drawn between countries as different as, for example, India and Chad enlarge the causal possibilities and raise the chances of discounting differences that are deemed not to fit the purpose of the investigation. This risk may be countered by undertaking detailed country-based studies, but this is an arduous corrective that is not possible in a global enquiry of the kind mounted by Acemoglu and his co-authors. A second assumption, which is of particular concern to historians, is that differences of time are of little consequence, or, more precisely, that what happens between the starting point and the end point of the investigation is insufficiently important to alter the design of the experiment. If the period in question is short, the assumption may still hold. If it extends over five centuries, as it does in Acemoglu's study, doubts arise as to whether significant temporal changes, as well as spatial differences, are being overlooked. Heroic assumptions can produce striking results, but they can also gain strength at the expense of credibility.

Following the methodological issues, economists have made two important historical contributions to the reversal of fortune thesis that bear directly on the case of Africa. Ola Olsson has pointed out that bundling all colonialism into one undifferentiated period from 1500 to 1950 glosses over the considerable disparity between mercantilist and modern forms of imperialism.Footnote 35 In the first phase, he argues, ‘good’ institutions scarcely existed, even in Europe, so could scarcely have been an option overseas. In reality, no choice between productive and extractive institutions was available in 1500. By the late nineteenth century, however, both options were present; by then, too, medical advances had removed some of the constraints on European settlement, and disease was no longer the important influence on colonial policy towards Africa that it had been earlier. Olsson concludes that the two eras were so different that their trajectories cannot be driven by one variable, settler mortality. Instead, they initiated two paths of development, and hence created two institutional legacies.

Olsson's analysis of these legacies leads him to hypothesize that the duration of colonial rule had a positive effect on institutional quality but a negative effect on the ability of independent states with a long history of colonial subordination to sustain democratic forms of government. Colonies founded during the second wave of imperialism, on the other hand, were likely to have poorer institutions but greater chances of establishing democratic governments. The suggested explanation is that it takes time to establish durable institutions, whereas modern imperialism transmitted ideas of self-determination and civil rights more rapidly and more effectively than its agents intended. According to Olsson, India combines duration with liberal ideals, whereas Africa suffers both from its long contact with Europe, which failed to eliminate unproductive institutions, and from the short duration of modern colonial rule, which allowed insufficient time to install modern institutions. As far as Africa is concerned, Olsson's argument remains in broad agreement with Acemoglu's to the extent that both concur that colonialism presided over non-productive institutions. However, whereas Acemoglu and his colleagues believe that colonial rule entrenched extractive regimes, Olsson suggests that colonial rulers did not have enough time to dismantle the unproductive institutions that were already in place.

The second contribution, by Nathan Nunn, examines Africa's external slave trades to illustrate the extent of the reversal of fortune suffered by the continent.Footnote 36 Nunn's starting point follows Acemoglu's in assuming that levels of economic and social development in Africa before 1500 were similar to those of other ‘relatively prosperous’ regions of the world. The external slave trades installed unproductive, extractive regimes that were perpetuated by colonial rule through land expropriation, compulsory labour and regressive taxes. The countries that are the poorest in Africa today, Nunn claims, are those from which most slaves were taken. Not only that, the slave trades weakened ethnic ties and hindered the growth of broader ethnic identities and the development of large states. In these ways, Africa's history helps to explain its poor performance after independence. Nunn's argument, like that of Acemoglu and his co-authors, appears to have much in common with the dependency thesis. However, none of the contributors to the debate about the reversal of fortunes pursues that line of enquiry. The best inference that can be drawn from their analyses is that Africa's development problem arose, not from the penetration of capitalism, but from the presence of the wrong type of capitalism.Footnote 37

Evidently, Acemoglu and his co-authors have advanced a vast thesis that invites far more comment than space allows. Three major issues, however, can be listed here for future discussion. The first issue concerns the data base. Establishing levels of prosperity for Africa in 1500 is a daunting task.Footnote 38 Acemoglu and his co-authors rely on population density as a proxy for prosperity. Elsewhere, they use data on urban populations too, but in the case of Africa they recognize that the evidence for the precolonial era is too sparse to be relied upon.Footnote 39 The population figures they assemble, however, are insufficiently robust to carry the explanatory weight placed on them.Footnote 40 They are derived from a dated Atlas, which itself compiled the guesses of the day.Footnote 41 Nearly all such estimates are obliged to employ the technique of backward projection, which rests on unreliable colonial censuses and heroic assumptions about growth rates in earlier periods. The fact is that there are few dependable studies of precolonial populations in Africa. Thornton's remarkable pioneering essay, which used baptism records to estimate the population of the Congo in the seventeenth century, remains unmatched, even though it was published over thirty years ago.Footnote 42 It will remain so until a new generation of researchers decides to tackle the problem anew.

The comparative use of population evidence is equally problematic. Acemoglu and his colleagues include Africa with Asia as being relatively well populated and prosperous in 1500, as opposed to the Americas and Australia, which were sparsely populated and relatively poor. It is hard to justify this grouping on the evidence available. As noted earlier, there are good grounds for supposing that, for the greater part of its history, Africa was in general a land-surplus region, not one that was densely populated, and that the ratio of people to land did not begin to change until the twentieth century. The early history of the Americas and Australia, on the other hand, has been the subject of revisionist research that has greatly increased previous estimates of population levels. In citing Colin McEvedy and Richard Jones's judgement that precolonial Australia was ‘an unchanging palaeolithic backwater’,Footnote 43 Acemoglu and his colleagues have missed the research of the last thirty years, which has presented an entirely different account of aboriginal history in what became Australia and has caused population estimates for the period shortly before white settlement to be raised from 150,000–300,000 to 1–1·5 million.Footnote 44 The history of the Americas before Columbus has been similarly rewritten to include estimates of the size of the population that reach as high as 100 million, which, if accurate, exceeded the population of Europe.Footnote 45 The point, however, is not to argue that these figures are correct but to suggest that all such figures are extrapolations from limited and shaky data. They ought not to be used to generate statistical results that have the appearance of scientific precision but in reality are little more than guesswork.Footnote 46

Regression analysis is only as robust as the numerical evidence it draws on. Statistical methods achieve valid results where the problem under investigation can be closely defined and where reliable quantitative data, such as time series, already exist or can be assembled for all the variables. In these circumstances econometric techniques can give precision to historical judgements that are often covered by vague qualifiers, such as ‘largely’ or ‘partly’. Problems arise where the data are incomplete or ambiguous, and where causal precision is extracted from very broad assumptions. The risk of achieving spurious accuracy increases where numerical weights are assigned to variables that are themselves diffuse, inconstant or imperfectly separated. In such cases, elegance may need to bow to complexity, and conventional historical methods, though inexact, may be the appropriate option.

The second issue concerns the classification of colonies adopted by Acemoglu and his co-authors and followed by most of the contributors to the debate. The decision to divide colonies into two categories, one associated with white settlement and termed ‘productive’, and the other associated with white rule over indigenous societies and termed ‘extractive’, is an unexplained retreat from the accepted historiography of the subject. It harks back to the distinction drawn by John Stuart Mill (though he is not cited in this connection) between advanced colonies inhabited by white settlers and uncivilized colonies inhabited by barbarians.Footnote 47 Later in the nineteenth century, weighty intellectuals such as Wilhelm Roscher and Paul Leroy-Beaulieu formulated more extensive classifications of the European empires that were expanding in the world around them.Footnote 48 In the 1940s, one of the most eminent historians of empire, Keith Hancock, applied Frederick Jackson Turner's notion of the moving frontier to produce four categories of colony.Footnote 49 By the 1960s, development economists were also devising typologies that took account of significant internal differences among countries previously described generically as ‘backward’ or ‘underdeveloped’. Hla Myint's influential analysis, for example, distinguished between over-populated and under-populated regions, and between mining and plantation enclaves on the one hand, and ‘peasant’ export sectors on the other.Footnote 50 Each of these categories combined factors of production in different ways, generated different linkages, and hence had different parameters of development.

By the time the modern study of African history began in the 1960s, Mill's over-simple and less than objective classification had long been replaced by typologies that offered a more accurate representation of the diversity of colonial rule. A common typology used by historians today distinguishes between colonies of settlement, colonies of concessions and colonies of trade, which in the case of Africa can be illustrated by South Africa, the Belgian Congo and Nigeria, respectively. Since Acemoglu's analysis lacks even this modest degree of refinement, he is obliged to crush all of colonial Africa into the extractive mode that supposedly reinforced unproductive institutions.

This procedure seriously distorts the history of colonial Africa. Acemoglu's typology takes no account of colonies of settlement in Africa, which fall by default into the category of extractive regimes, whereas in other parts of the world his study treats them as agents of productive institutions. Although the difference can be explained by evaluating the relationships between settler communities and indigenous societies, it is sufficiently important to require a revision of the simple binary classification adopted by Acemoglu and accepted by other parties to the debate on the reversal of fortunes. The classification also mangles the history of non-settler Africa. In colonies of trade, such as Nigeria, the greater part of export production remained in the hands of the indigenous people and was organized through indigenous institutions. The history of successful African enterprise in producing export crops in the course of the past century and a half is now well known. It is wholly inconsistent with Acemoglu's claim that non-settler colonies administered extractive regimes that fostered unproductive institutions. It is equally at variance with William Easterly and Ross Levine's assertion, discussed in the next section of this article, that ethnic fragmentation is a major cause of retarded development. The presence of diverse ethnic groups did not impede the massive export growth that took place before independence. Fortunes expanded in the era following the abolition of the slave trades, and specialization reinforced by ethnic ties contributed to it.

The third cause of concern is the excessive abbreviation of history that all these arguments rely on.Footnote 51 This objection is not simply that of a specialist with an eye for detail. The assumption that it is acceptable to start in 1500 and fast-forward to the present without taking full account of intervening changes is to use history but also to negate its purpose. Colonial rule in Africa was not the same, either in scale or in policy, in 1500 as it was in 1900; to suppose that it was is to promote the purposes of measurement above those of historical accuracy. Olsson has performed a service for economists in distinguishing between mercantilist empires and those that were created or enlarged at the close of the nineteenth century.Footnote 52 This distinction, however, is only a beginning. It is now a commonplace among historians that the precolonial era was anything but static, and that its multiple movements involved both institutional change and shifting ethnicities. In the nineteenth century, to cite the well-studied case of West Africa, the process of adaptation from slave-trading to other types of exports brought about a fundamental realignment of the factors of production and induced major institutional changes.Footnote 53 Ordinary households began to participate directly in export-production for the first time; land in areas that were able to participate in the new economy began to acquire scarcity value; wage labour gradually replaced slave status within Africa. Institutional change continued under colonial rule, but not solely in response to colonial policies, as Acemoglu and his co-authors, and Nunn, assume. In colonies of trade, colonial rule was negotiated as well as imposed, and institutional change was the product of African initiatives as well as of exogenous causes, such as population growth.

ETHNICITY AND DEVELOPMENT

William Easterly and Ross Levine's contribution to the debate deals with a social institution, ethnicity, which they claim has particular relevance to Africa's development problems.Footnote 54 They argue that cross-country variations in the incidence of ethnic diversity explain a substantial part of the differences in public policies, political institutions and other causes of long-run development. They present data showing that there is a much greater degree of ethnic diversity in Africa than in other developing countries, notably in Asia. Their analysis goes on to suggest that ethnic diversity impedes agreement on the provision of vital public goods because different groups fail to see eye-to-eye on policies that benefit all members of society. Indeed, ethnic diversity creates incentives to adopt policies that harm development. It diverts investment to particular interests, encourages inter-ethnic competition for rents, promotes corruption and opens the way for violence. The outcome is inadequate education, backward financial systems, distorted foreign exchange markets, government indebtedness, deficient infrastructure, and political instability. The authors conclude that ethnic diversity alone explains between 25 and 40 per cent of the growth differential between Africa and East Asia.Footnote 55

Although much of the discussion generated by this thesis has focused on methodology rather than on history, the questions raised are directly relevant to historical enquiry too.Footnote 56 Easterly and Levine assumed that their data on ethnicity (which were derived from a Soviet source dating from the 1960s)Footnote 57 were unproblematic, at least as far as their particular project was concerned. Subsequent commentators, however, have recognized that ethnicity is a slippery and malleable term that is not easily reduced to units that can be measured accurately. Moreover, even when ethnicity is pinned down, it is still hard to establish convincingly that ethnic fragmentation, though present, is also a prime cause of policies that hamper economic development.

The principal research effort has been devoted to improving Easterly and Levine's measure of ethnic fragmentation. One approach has been to construct new and broader indices of what the literature terms ‘fractionalization’ to capture degrees of cultural diversity that the original index failed to represent adequately.Footnote 58 An alternative index, designed by Arcand and his co-authors, confirms the high degree of ethnic fragmentation in Africa but concludes that there is little evidence linking it to policies that reduce growth.Footnote 59 A further index, devised for Africa by Daniel Posner, links ethnicity to economic outcomes by including only groups that were involved in the political process rather than all ethnic groups in a given country.Footnote 60 Posner's new measure confirms Easterly and Levine's original conclusion that ethnic fractionalization is causally related to poor growth. Posner's index, however, has its own difficulties. It is not clear that Posner succeeds either in isolating the ethnic element from other variables, or in showing that the presence of politically influential ethnic groups is causally related to adverse public sector policies. In general, the improved indices show that there is a higher degree of ethnic fragmentation in Africa than elsewhere. In doing so, however, they also make it harder to distinguish ethnic, linguistic and religious variables from other influences, such as the geographical endowment and the political process, with which they interact.

Easterly and Levine themselves did not explore the historical basis of ethnicity in any detail, though they referred briefly to the role of colonialism in splitting ethnic groups at the time of the partition of Africa. Other social scientists, however, have followed their tentative lead and looked further into the past. One of the most interesting contributions, which so far few economists have considered, is by a political scientist, Pierre Englebert, whose commentary on ethnic fragmentation draws on evidence from precolonial Africa.Footnote 61 Englebert argues that contemporary institutions are more likely to be effective in promoting growth if they are congruent with indigenous institutions. High degrees of congruence underline the legitimacy of governments and enhance their ability to foster institutions that induce development.Footnote 62 Low levels of congruence have the opposite effect and create conditions that favour patrimonial politics, which lead to regionalism, clientelism and corruption. Englebert concludes that ethno-linguistic fragmentation has no significant effect on the quality of public goods provided by African governments. Societies can be ethnically homogeneous but still be dysfunctional, and ethnically heterogeneous but still capable of upholding the rule of law and developing progressive institutions. To the extent that Africa's dismal economic record is linked to the presence of poor institutions, Englebert finds that the main cause is the low incidence of legitimacy among states, which he attributes largely to the actions of colonial powers in imposing arbitrary boundaries on Africa at the time of partition.

The most ambitious historical contribution is by two economists, Pelle Ahlerup and Ola Olsson, who have advanced a theory to explain the global determinants of ethnic diversity.Footnote 63 Why is it, they ask, that the probability of any two randomly chosen individuals belonging to different ethnic groups is only 0·2 per cent in South Korea, whereas it is 93 per cent in Uganda? To answer this question, they begin by defining an ethnic group as one whose members have a shared history or ancestry that is typically associated with a homeland or founding migration and is bound together by language, religion and common customs.Footnote 64 They then trace the evolution of diverse ethnicities by making enterprising use of data from the human genome project and relating it to the two main hypotheses adopted by historians: one holding that ethnicity is primordial; the other that it is the constructed product of recent state formation.

The central argument runs as follows. Since human beings originated in East Africa and migrated from there to other parts of the continent and the world, the process of dispersion, which was built into hunting and gathering societies, was inevitably accompanied by fragmentation. Why was this? The answer, according to Ahlerup and Olsson, is that ethnic groups came into being principally to supply public goods, such as protection and a moral code governing rules of conduct. These goods preserved the cohesion, and hence the identity, of the group. Public goods, however, were not easily provided over long distances in the pre-industrial era. The organization and system of controls that underpinned the solidarity of the group depended on the continuing proximity of its members. As migration took place and the distance from centres of origin grew, the public provision that initially held the group together had to be recreated by forming, over time, new ethnic identities. It follows from this argument that ethnic diversity is related to antiquity, which is why the greatest diversity in the world today is found in Africa.

The rise of sedentary agriculture (the so-called ‘neolithic revolution’) held out the prospect of creating centralized states. Such states preferred homogeneity to diversity because they needed to create commonalities that would underpin the supply of public goods. A concentrated population tended to decrease ethnic diversity by reducing isolation and encouraging state formation. This trend reached its highest point with the rise of nation states in the Western world during the nineteenth and twentieth centuries. In Africa, however, the development of states was limited by conditions that made centralization difficult. Moreover – and here Ahlerup and Olsson add an imperial twist to their argument – the new European nation states had no incentive to create homogeneity when they expanded overseas; instead, they pursued a policy of divide and rule because this was the most effective way of controlling their colonies. Ahlerup and Olsson do not connect their analysis to the provision of public goods in Africa today, but in drawing attention to the antiquity of ethnic fragmentation they, like Englebert, emphasize the effect of colonial rule in reinforcing it.

What might historians contribute to this debate? The possibilities are as wide as the scope of the theories, which are capacious. Two central themes will be picked out here because they refer to research that the new institutional history has yet to incorporate. The first theme, ethnicity, has been the subject of lively historical discussion since it was invigorated by the literature on the invention of tradition more than twenty years ago. The starting point is the idea that ethnicity, like ‘tribalism’, was not a fixed, primordial institution but a construct that colonial rulers were active in establishing because they needed to delineate units of administration in the new states they were building in Africa.Footnote 65 This proposition would appear to support the emphasis placed by scholars such as Englebert, Ahlerup and Olsson on the consolidation and creation of ethnicity in the colonial period. However, the original thesis has undergone several mutations since it was formulated. The idea that traditions were invented is now thought to be top-heavy and mechanical.Footnote 66 Colonial rulers may have tried to solidify ethnicities, but Africans contributed to the process too. Moreover, traditions were not simply conjured into being but drawn from precolonial institutions that continued to evolve under colonial rule.Footnote 67 There are cases where identifiable ethnicities can be traced to the precolonial era and were then consolidated by colonial governments; equally, there are cases where ethnicity was fluid and even ambiguous in the precolonial period and remained so under colonial rule.Footnote 68 The current historiography greatly complicates the story told by the new institutional history. It is now extraordinarily difficult to sustain long-run, continent-wide generalizations about ethnicity.Footnote 69 The historical record suggests that, if development policy is to be guided by assumptions about ethnicity, it needs to proceed cautiously on a case-by-case basis and abandon the attempt to formulate one prescription for a large and diverse continent.Footnote 70

The second theme concerns the relationship between population density, ethnicity and state-building. It is apparent that there were large states in precolonial Africa, but it is equally clear that they were the exception rather than the rule.Footnote 71 No single state ever controlled more than a small part of the continent. In explaining why this was the case, the new institutional history has omitted the contribution made by historians to this subject, though it could also be said that historians themselves might have done more to invigorate these arguments in recent years. The fundamental historical proposition derives from the assumption that the land–labour ratio prevailing in Africa encouraged the extensive use of land, either because the population was small in relation to resources, or because the land was too poor to sustain permanent settlement.Footnote 72 Accordingly, populations tended to migrate or at least to fan out from a central place to the point where, as Ahlerup and Olsson suggest, the chances of establishing central control were weakened or even eliminated. A dispersed population could neither be taxed readily (unlike in India, where land taxes were common) nor provide the basis for a mass market. No doubt this interpretation needs refining and may even need refurbishing.Footnote 73 As it stands, however, it offers the beginning of an explanation of the relative underdevelopment of Africa, and one moreover that identifies limits to growth that existed before the advent of the Europeans.

It is apparent that the new debate on Africa's ‘growth tragedy’ needs substantial adjustment to take account of the relevant historical evidence. If low levels of population in Africa were caused by infertile soils or disease, geographical and allied determinants need to be incorporated fully into the discussion. If ethnic fragmentation is the product of dispersal, then social institutions – or at least this particular social institution – are partly the product of environmental conditions and not exogenous, independent variables. If this is the case, ethnicity should be seen as a solution rather than as a problem. It functioned in a variety of ways to support the community rather than to act as a weight on it.Footnote 74 The implication of this claim is that the role of ethnicity is contingent on historical circumstances. From the point of view of development, what matters, or what ought to matter, is the fit between a social institution and the problems it has to confront. Research into the history of ethnicity will go astray if it tries to find a set of characteristics that are thought to have eternal qualities. A historical defence of ethnicity as a functioning, constructive institution may be irrelevant or anachronistic if the issues of the twenty-first century are no longer those of the previous century.Footnote 75 In evaluating these questions, it should be remembered that ethnicity is one measure of identity and not its totality. Group cohesion has other markers, including economic interests and religious affiliations, which create multiple loyalties. The argument that ethnic fragmentation is responsible for unproductive institutions needs to demonstrate, more convincingly than it has done so far, that other causes of solidarity and motives for action can be discounted.Footnote 76 The study of history is not a substitute for policy-making but it can assist by plotting the evolution of changes in the relationship between social institutions and the issues they are designed to deal with.Footnote 77

RETROSPECT AND PROSPECT

It would be easy to conclude from this critical survey that the new economic history of Africa contains too many fundamental flaws to merit serious attention from historians. Much of the history emerging from the new thinking is undoubtedly over-schematic, based on incomplete use of the available literature, and compressed to the point where time very nearly stands still. It is apparent, too, as at least one of the participants has observed, that some contributions seem to be driven by the need to demonstrate a mastery of econometric techniques rather than by a concern to incorporate differences in the historical context that are essential if the problem under review is to be explained successfully.Footnote 78 The result is an oddly assorted mixture in which the latest technique, dazzling though it is, rests on an outdated view of how history, and the history of Africa in particular, is studied.

The imminent danger is that the new history will reproduce a familiar story in a fresh disguise – that of the rise of the West and the fall of the rest. The world portrayed by much of the research reviewed here is one that was made and remade by Europeans. Other societies had little to contribute to economic development. Indeed, in the case of Africa the indigenous people seem to have been adept at devising growth-retarding institutions, albeit with outside assistance. The early development economists, whose work is now neglected, knew better when they formulated different categories of underdevelopment and drew attention to the varied routes to development open to different societies. Economic historians amplified this understanding by emphasising the dynamic qualities of indigenous societies in Africa and by identifying specific configurations of poverty that reflected regional particularities. Given these advances, it is anachronistic to try to reduce the history of modern economic development to a question of determining the location of Western settlement. A similar, equally unrecognized, bias can accompany the treatment of ethnicity if it is portrayed as an exotic phenomenon that needs rectifying if development is to occur. Ethnicity is as universal as it is malleable. Fluid ethnic identities abounded in Western Europe, when the foundations of modern development were being laid, and were not abandoned even in the nineteenth and twentieth centuries, despite the exclusive claims of nationalism.Footnote 79

These observations touch upon an old set of arguments concerning the universality of Western economic theory.Footnote 80 Policy-makers have traditionally favoured a ‘best practice’ model, which begins by assuming that recommendations for development can be deduced from first principles. This assumption is theoretically pleasing, globally applicable and allows progress towards common goals to be measured and compared. The risk is that, in practice, development policy will define ‘good’ institutions as those that prevail in the country of origin: protecting property rights, for example, translates as protecting private property. This thought process carries forward the errors of modernization theory, which held, for example, that the most appropriate social unit for development purposes was the nuclear family rather than the extended family. A closely associated mistake is to compare alien societies to a textbook ideal rather than to a functioning reality. If, as is claimed, the United States is the most developed country in the world, it is also one that has marked ethnic divisions, intense regional loyalties, a notable level of political corruption, and considerable difficulty agreeing on the provision of public goods. This reality is rarely to the fore when policies for reforming Africa are being formulated.Footnote 81 Universal principles and ideal types will take different forms in different contexts. Understanding particular contexts means accepting that there are different routes to development and different concepts of what development means.Footnote 82 Theoretical purity can be achieved either by working with perfect data or by trying to make imperfect data fit technical perfection. Both options abstract from the problem they seek to analyse. If the institutional approach is to be pursued, it will need to step closer to realism and adopt ‘second-best practice’ itself.Footnote 83

Even if all these points are conceded, however, there is still a powerful case for treating the new literature seriously. There is widespread agreement that poverty is a central issue, if not the central issue, in Africa today and has been throughout the continent's long history. The influences that have attracted historians to other subjects are receding: a material return is now in sight. With Bob Geldof to provide the publicity and leading economists to light the way, circumstances favour the revival of economic history. The parsimonious explanations economists aim at achieving have the advantage of framing bold, clear hypotheses; the techniques they deploy offer a degree of precision that historians sometimes shy away from, even when it is attainable. Historians know as well as economists that it is much easier to be right about small matters than to be illuminating about big ones. The research summarized here deserves great credit for reopening lines of enquiry that are important for understanding both precolonial and colonial history. The degree of approval of, or dissent from, particular interpretations matters less at this preliminary stage of the discussion than the prospect that history will be taken seriously by policy-makers.

Modernization theory disowned history. All the elements needed to transform ‘traditional’ societies into ‘modern’ ones could be deduced from the pattern of variables formulated by Talcott Parsons.Footnote 84 Today, economists have turned to history to improve their understanding of global poverty. It is they, not historians, who have made the case for the discipline:

Even giving any substantial weight to centuries-old history may not be so appealing from a policy-making point of view, but the world is as it is. The world seems to be a place where the long eons of history still matter very much today.Footnote 85

The conclusions emerging from the new economic history of Africa, if acted on, have implications for policies affecting a wide range of institutional reforms. On this occasion, historians have a chance to play a part in contributing to policies that are informed by history and in this way to help to eliminate poverty among the bottom billion people in Africa. It matters that the new generation of historians accepts the invitation they have been given.

References

1 To the best of my knowledge, Gareth Austin is the only historian of Africa to have engaged with this literature: Austin, ‘The “reversal of fortune” thesis’. C. A. Bayly has commented on India and Africa, though primarily on the former: Indigenous and Colonial Origins of Comparative Economic Development: The Case of Colonial India and Africa (World Bank Policy Research Working Paper 4474, 2008).

2 Even so, some of the most notable thematic and regional studies were published more than twenty years ago: Patrick Manning, Slavery, Colonialism and Economic Growth in Dahomey, 1640–1960 (Cambridge, 1982); Manning, Francophone Sub-Saharan Africa, 1880–1985 (Cambridge, 1988); Paul E. Lovejoy, Transformations in Slavery: A History of Slavery in Africa (Cambridge, 1983); Fred Cooper, From Slaves to Squatters: Plantation Labour and Agriculture in Zanzibar and Coastal Kenya, 1890–1925 (New Haven, 1980); Cooper, On the Waterfront: Urban Disorder and the Transformation of Work in Colonial Mombasa (New Haven, 1987). John K. Thornton's Africa and Africans in the Making of the Atlantic World, 1400–1800 (Cambridge, 1992) only just qualifies as an exception.

3 A. G. Hopkins, An Economic History of West Africa (London, 1973); Ralph A. Austen, African Economic History: Internal Development and External Dependency (London, 1987); Charles H. Feinstein, An Economic History of South Africa (Cambridge, 2005). Paul T. Zeleza's A Modern Economic History of Africa, vol. I: The Nineteenth Century (Dakar, 1993) summarizes much of the work undertaken during the previous two decades.

4 John Iliffe, The African Poor: A History (Cambridge, 1987).

5 Manning, Patrick, ‘The prospects for African economic history: is today included in the long run?’, African Studies Review, 30 (1987), 4962.CrossRefGoogle Scholar

6 Edward W. Said, Orientalism (New York, 1978).

7 Rita Abrahamsen's defence of ‘postcoloniality’ and advocacy of its application to Africa, for example, is notably slim in treating development issues and economic matters in particular: ‘African studies and the postcolonial challenge’, African Affairs, 102 (2003), 189–210.

8 The general trends are agreed and are widely summarized. For the illustrations that follow, see Elsa V. Artadi and Xavier Sala-i-Martin, The Economic Tragedy of the Twentieth Century: Growth in Africa (National Bureau of Economic Research [hereafter NBER] Working Paper 9865, 2003); Easterly, William and Levine, Ross, ‘Africa's growth tragedy: policies and ethnic traditions’, Quarterly Journal of Economics, 112 (1997), 1203–4.CrossRefGoogle Scholar

9 The Economist, 13 May 2000, 23–5. I owe this reference to Professor John Lonsdale.

10 Ibid. 17.

11 The G8 states are the Group of Eight richest industrialized countries: the United States, Canada, Japan, Russia, Germany, the United Kingdom, France and Italy.

12 A comprehensive statement is Claude Menard and Mary M. Shirley (eds.), Handbook of New Institutional Economics (2005). The Ronald Coase Institute provides accessible guides to the subject and news of relevant conferences: www.coase.org.

13 David Secker, Thorstein Veblen and the Institutionalists (London, 1975), provides a broader account than its title might suggest.

14 Examples include: Hopkins, A. G., ‘Property rights and empire-building: the annexation of Lagos, 1861’, Journal of Economic History, 40 (1980), 777–98CrossRefGoogle Scholar; Alence, Rod, ‘The 1937–1938 Gold Coast cocoa crisis: the political economy of commercial stalemate’, African Economic History, 19 (1990–1), 77104CrossRefGoogle Scholar; Evans, E. W. and Richardson, D., ‘Hunting for rents: the economics of slavery in pre-colonial Africa’, Economic History Review, 48 (1995), 665–86CrossRefGoogle Scholar; Wariboko, Nimi, ‘A theory of the Canoe House Corporation’, African Economic History, 26 (1998), 141–72CrossRefGoogle Scholar; Lovejoy, Paul E. and Richardson, David R., ‘Trust, pawnship and Atlantic history: the institutional foundations of the Old Calabar slave trade’, American Historical Review, 104 (1999), 332–55CrossRefGoogle ScholarPubMed; Lovejoy, and Richardson, , ‘“This horrid hole”: royal authority, commerce and credit at Bonny, 1690–1840’, Journal of African History, 45 (2004), 363–92.CrossRefGoogle Scholar The best guide and fullest contribution to this literature is Gareth Austin, Labour, Land and Capital in Ghana: From Slavery to Free Labour in Asante, 1807–1956 (Rochester NY, 2005).

15 Manning, ‘Prospects for African economic history’; Manning, ‘African economic growth and the public sector: lessons from historical statistics of Cameroon’, African Economic History, 19 (1990–1), 135–70, is a model demonstration of the flaws that arise from poorly based historical assumptions. A broader view from the same standpoint is Hopkins, A. G., ‘The World Bank in Africa: historical reflections on the African present’, World Development, 14 (1986), 1473–87.CrossRefGoogle Scholar

16 World Bank, Accelerated Development in Sub-Saharan Africa (New York, 1981).

17 The literature on structural adjustment is now vast. Accessible guides include: David E. Sahn, Paul A. Dorosch and Stephen D. Younger, Structural Adjustment Reconsidered: Economic Policy and Poverty in Africa (Cambridge, 1997); and Giles Mohan, Ed Brown, Bob Williams and Alfred B. Zach-Williams, Structural Adjustment: Theory, Practice and Impacts (London, 2000). An evaluation with a strong historical dimension is Celestin Monga, ‘Commodities, Mercedes Benz, and structural adjustment: an episode in West Africa's economic history’, in Emmanuel K. Akyeampong (ed.), Themes in West Africa's History (Oxford, 2006), 227–64. For a broader view of the evolution of development thinking in Africa, see Rimmer, Douglas, ‘Learning about economic development from Africa’, African Affairs, 102 (2003), 469–91.CrossRefGoogle Scholar

18 Representative examples include: Robert H. Bates, Markets and States in Tropical Africa: The Political Basis of Agricultural Policy (Berkeley, 1981); Bates, Essays on the Political Economy of Rural Africa (Cambridge, 1983); Bates, Prosperity and Violence: The Political Economy of Development (New York, 2001); Sara S. Berry, Fathers Work for Their Sons: Accumulation, Mobility and Class Formation in an Extended Yoruba Community (Berkeley, 1985); Berry, No Condition is Permanent: The Social Dynamics of Agrarian Change in Sub-Saharan Africa (Madison, 1993); Chiefs Know Their Boundaries: Essays on Property, Power, and the Past in Asante, 1896–1996 (Oxford, 2001); Manning, Slavery, Colonialism, and Economic Growth; Manning, ‘The slave trade: the formal demography of a global system’, Social Science History, 14 (1990), 255–79. Robert Szereszewski, Structural Changes in the Economy of Ghana, 1896–1911 (London, 1965), though not strongly institutional in orientation, should also be mentioned as a pioneering (and now unjustly neglected) account of structural economic change.

19 Peter B. Evans, Dietrich Rueschmeyer and Theda Skocpol (eds.), Bringing the State Back In (Cambridge, 1985); World Bank, Governance and Development (Washington DC, 1992).

20 See, for example, Paul Mosley's reflections on how Africa's continuing poverty was expressed in the poverty of development economics, and his suggestion that a comparison with Asia might be illuminating: ‘Development economics and the underdevelopment of sub-Saharan Africa’, Journal of International Development, 7 (1995), 685–706. An influential comparison at that time was Robert Wade, Governing the Market: Economic Theory and the Role of Government in East Asian Industrialisation (Princeton, 1992).

21 An accessible introduction to quantitative methods is Pat Hudson, History by Numbers: An Introduction to Quantitative Approaches (London, 2000). A more advanced text is Charles H. Feinstein and Mark Thomas, Making History Count: A Primer in Quantitative Methods for Historians (Cambridge, 2002).

22 See in particular, Jeffrey D. Sachs, Tropical Underdevelopment (NBER Working Paper 8119, 2001) at: www.nber.org/papers/w8119; Sachs, Institutions Don't Rule: Direct Effects of Geography on Per Capita Income (NBER Working Paper 9490, 2004) at www.nber.org/papers/w9490; Olsson, Ola and Hibbs, Douglas A., ‘Biogeography and long-run economic development’, European Economic Review, 49 (2005), 909–38.CrossRefGoogle Scholar

23 See, for example, Banerjee, Abhijit V., ‘The two poverties’, Nordic Journal of Political Economy, 26 (2001), 129–41Google Scholar; Esther Duflo, ‘Field experiments in development economics’, in Richard Blundell, Whitney Newey and Torsten Persson (eds.), Advances in Economics and Econometrics (Cambridge, 2006), 322–48; Rohinde Pande and Christopher Udry, ‘Institutions and development: a view from below’, in Blundell, Newey and Persson (eds.), Advances in Economics, 349–412; Esther Duflo, Rachel Glennester and Michael Kremer, Using Randomization in Development Economics: A Toolkit (NBER Technical Working Paper 333, 2006).

24 Such as the World Bank and the US National Bureau of Economic Research, which describes itself as ‘the nation's leading non-profit economic research organization’.

25 In addition to the specific citations that follow, see the ‘Symposium on institutions and economic performance’, Economics of Transition, 13 (2005), 421–572.

26 Acemoglu, Daron, Johnson, Simon and Robinson, James A., ‘Reversal of fortune: geography and institutions in the making of the modern world income distribution’, Quarterly Journal of Economics, 17 (2002), 1231–94CrossRefGoogle Scholar; Acemoglu, Johnson and Robinson, ‘The colonial origins of comparative development: an empirical investigation’, American Economic Review, 91 (2001), 1369–401. Earlier versions of both articles appeared as Working Papers that stimulated discussion before their formal publication.

27 The authors are careful to explain that ‘by economic prosperity or income per capita in 1500, we do not refer to the economic or social conditions or the welfare of the masses but to a measure of the total production in the economy relative to the number of inhabitants’. Similarly, the reversal of fortune refers to ‘changes in relative incomes across different areas and does not imply that the initial inhabitants of, for example, New Zealand or North America themselves became relatively rich’. Acemoglu, Johnson and Robinson, ‘Reversal of fortune’, 1232 n.1.

28 See, for example, Pranab Bardhan, ‘Institutions matter, but which ones?’ Economics of Transition, 13 (2005), 499–532; Jutta Bolt and Dirk Bezemer, Understanding Long-Run African Growth: Colonial Institutions or Colonial Education: Evidence from a New Data Set (MPRA Paper 7029, 2008) at http://mpra.ubuni-muenchen.de/7029/; Glaeser, Edward L., Porta, Rafael La, Lopez-de-Silanes, Florencio and Shleifer, Andrei, ‘Do institutions cause growth?’, Journal of Economic Growth, 9 (2004), 271303CrossRefGoogle Scholar; Dani Rodrik, Getting Institutions Right, CESifo DICE Report (Munich, 2004); Rodrik, Second-best Institutions (NBER Working Paper 14050, 2008).

29 Rodrik, Dani, Subramanian, Arvind and Trebbi, Francesco, ‘Institutions rule: the primacy of institutions over geography and integration in economic development’, Journal of Economic Growth, 9 (2004), 131–65.CrossRefGoogle Scholar

30 O'Brien, ‘Colonies in a globalizing economy, 1815–1948’, in Barry Giles and William R. Thompson (eds.), Globalization and Global History (Hoboken, 2006), 224–65.

31 See especially, John W. McArthur and Jeffrey D. Sachs, Institutions and Geography: Comment on Acemoglu, Johnson and Robinson (2000) (NBER Working Paper 8114, 2001); Sachs, Tropical Underdevelopment; Sachs, Institutions Don't Rule; William Easterly and Ross Levine, Tropics, Germs and Crops: How Endowments Influence Economic Development (NBER Working Paper 9106, 2002); Ola Olsson, ‘Geography and institutions: a review of plausible and implausible linkages’, Working Papers in Economics (Department of Economics, Goteborg University), 106 (2003).

32 Kenneth L. Sokoloff and Stanley L. Engerman have elaborated this point in a series of important studies: ‘Institutions, factor endowments, and paths of development in the New World’, Journal of Economic Perspectives, 14 (2000), 217–32; Institutional and Non-Institutional Explanations of Economic Differences (NBER Working Paper 9989, 2003); Colonialism, Inequality and Long-Run Paths of Development (NBER Working Paper 11057, 2005).

33 Valuable examples of putting these principles into practice are the studies of land tenure by Pande and Udry, ‘Institutions and development: a view from below’; and Boone, Catherine, ‘Property and constitutional order: land tenure reform and the future of the African state’, African Affairs, 106 (2007), 557–86.CrossRefGoogle Scholar Pande and Udry deal with Ghana; Boone with the Ivory Coast, Uganda, South Africa and Tanzania. See also Berry, Sara, ‘Debating the land question in Africa’, Comparative Studies in Society and History, 44 (2002), 638–68CrossRefGoogle Scholar, which ought to be mandatory reading for development planners.

34 Two balanced accounts are: Levine, Ross and Renelt, David, ‘A sensitivity analysis of cross-country growth regressions’, American Economic Review, 82 (1992), 942–63Google Scholar; Roodman, David M., ‘The anarchy of numbers: aid, development and cross-country empirics’, World Bank Economic Review, 21 (2007), 255–77.CrossRefGoogle Scholar

35 Ola Olsson, ‘Unbundling ex-colonies: a comment on Acemoglu, Johnson, and Robinson, 2001’, Working Papers in Economics (Goteborg University), 146 (2004); Olsson, ‘On the institutional legacy of mercantilist and imperialist colonialism’, Working Papers in Economics, 247 (2007).

36 Nunn, Nathan, ‘Historical legacies: a model linking Africa's past to its current underdevelopment’, Journal of Development Economics, 83 (2007), 157–75CrossRefGoogle Scholar; Nunn, ‘The long-term effects of Africa's slave trades’, Quarterly Journal of Economics, 23 (2008), 139–72. I follow Nunn in using the plural, ‘trades’, to refer to all Africa's external slave exports.

37 Daron Acemoglu, Simon Johnson and James A. Robinson, Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution (NBER Working Paper 2001), 4–5. The published version of this paper (see n. 26) omits this reference.

38 An exchange on this subject has already taken place. See David Albouy, The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data (Working Paper CO4-138, University of California at Berkeley, Center for International and Development Economics Research, 2006); Daron Acemoglu, Simon Johnson and James A. Robinson, ‘Reply’ (2006), at http://econ-www.mit.edu/files/212.

39 Acemoglu and his co-authors wisely decided that evidence of precolonial urban populations was insufficient for their purposes. Many towns in precolonial Africa varied in size according to seasons and market days; others were established for defensive purposes and contained a high proportion of cultivators who commuted to their farms.

40 Austin provides a fuller appraisal of the frailty of the data bases created by Acemoglu and his co-authors and by Nunn in ‘The “reversal of fortune” thesis’, 5–8.

41 Colin McEvedy and Richard Jones, Atlas of World Population (New York, 1978). Kent G. Deng concludes, with respect to China, that there ‘is not a shred of evidence to support their claims’: ‘Fact or fiction? Re-examination of Chinese premodern populaton statistics’, Working Papers in Economic History (LSE), 76 (2003), 6 n. 2.

42 Thornton, John, ‘Demography and history in the kingdom of the Kongo, 1550–1750’, Journal of African History, 18 (1977), 507–30.CrossRefGoogle Scholar

43 Acemoglu et al., ‘Reversal of fortune’, 1256.

44 Noel G. Butlin, Our Original Aggression: Aboriginal Populations of Southeastern Australia, 1788–1850 (Sydney, 1983); Butlin, Economics and the Dreamtimes: A Hypothetical History (Cambridge, 1993); Derek J. Mulvaney and John P. White (eds.), Australians to 1877 (Sydney, 1987).

45 Charles C. Mann provides a convenient summary of recent research in 1491: New Revelations of the Americas Before Columbus (New York, 2005).

46 David Henige has provided what should be the last word on this subject: Numbers from Nowhere: The American Indian Contact Population Debate (Norman OK, 1998).

47 Sullivan, Eileen P., ‘Liberalism and imperialism: J. S. Mill's defence of the British empire’, Journal of the History of Ideas, 44 (1983), 599617.CrossRefGoogle Scholar

48 Finlay, Moses I., ‘Colonies: an attempt at a typology’, Transactions of the Royal Historical Society, 26 (1976), 167–88.CrossRefGoogle Scholar

49 W. K. Hancock, Survey of British Commonwealth Affairs, 1918–1939, vol. II: Problems of Economic Policy, 1918–1939, Part 1 (London, 1942), 1–28.

50 Hla Myint, The Economics of the Developing Countries (London, 1964); and for a similar insistence on the need to relate development policies to diverse regional realities, Gunnar Myrdal, Asian Drama: An Inquiry into the Poverty of Nations (London, 1972).

51 Nunn, in particular, tends to assume that Europeans dealing with Africa in the precolonial period had much the same degree of authority as in the colonial period: ‘Long term effects’; ‘Historical legacies’. See Austin, ‘The “reversal of fortune” thesis’, 9.

52 Olsson, ‘Unbundling ex-colonies’.

53 Hopkins, Economic History, ch. 4, and, for an extension of the argument, Hopkins, ‘The “new international economic order” in the nineteenth century: Britain's first development plan for Africa’, in Robin Law (ed.), From Slave Trade to Legitimate Commerce: The Commercial Transition in Nineteenth-Century West Africa (Cambridge, 1995), 240–64.

54 Easterly and Levine, ‘Africa's growth tragedy’.

55 Ibid. p. 237.

56 Alberto Alesina and Eliana La Ferrara provide an excellent guide to the literature: ‘Ethnic diversity and economic performance’, Journal of Economic Literature, 43 (2005), 762–800.

57 The ‘ethno-linguistic fragmentation index’, which has been much used by social scientists, was produced by Charles L. Taylor and Michael C. Hudson, World Handbook of Political and Social Indicators (2nd ed., New Haven, 1972), from Atlas Narodov Mira (Moscow, 1964).

58 Alberto Alesina, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat and Romain Wacziarg, ‘Fractionalization’, Journal of Economic Growth, 8 (2003), 155–94; Fearon, James D., ‘Ethnic and cultural diversity by country’, Journal of Economic Growth, 8 (2003), 195222.CrossRefGoogle Scholar

59 Jean-Louis Arcand, Patrick Guillaumont and Sylviane Guillaumont Jeanneney, ‘How to make a tragedy: on the alleged effect of ethnicity on growth’, Journal of International Development, 12 (2000), 925–38.

60 Posner, Daniel N., ‘Measuring ethnic fractionalization in Africa’, American Journal of Political Science, 48 (2004), 849–63.CrossRefGoogle Scholar

61 Englebert, Pierre, ‘Pre-colonial institutions, post-colonial states, and economic development in tropical Africa’, Political Research Quarterly, 53 (2000), 736CrossRefGoogle Scholar; Englebert, State Legitimacy and Development in Africa (Boulder, 2000).

62 Nicola Gennaioli and Ilia Rainer reach broadly the same conclusion but do not refer to Englebert (see n. 67): ‘The modern impact of precolonial centralization in Africa’, Journal of Economic Growth, 12 (2007), 185–234.

63 Pelle Ahlerup and Ola Olsson, ‘The roots of ethnic diversity’, Working Papers (School of Business, Economics and Law, Goteborg University), 281 (2007).

64 They follow here Robert H. Bates, ‘Ethnicity’, in David A. Clark (ed.), The Elgar Companion to Development Studies (Cheltenham, 2006), 167–73. On this definition, ethnicity is ultimately a cultural construct, however primordial its roots may be, whereas race is a physiological attribute.

65 Fortunately, there is now a comprehensive guide to the literature: Spear, Thomas, ‘Neo-traditionalism and the limits of invention in British colonial Africa’, Journal of African History, 44 (2003), 327.CrossRefGoogle Scholar The modern debate stems from Eric J. Hobsbawm and Terence O. Ranger (eds.), The Invention of Tradition (Cambridge, 1983). See also the pioneering work by Leroy Vail (ed.), The Creation of Tribalism in Southern Africa (Berkeley and Los Angeles, 1989).

66 See, particularly, Terence O. Ranger's auto-criticism: ‘The invention of tradition revised’, in Terence O. Ranger and Olufemi Vaughan (eds.), Legitimacy and the State in Twentieth-Century Africa (London, 1993), 62–111.

67 John Lonsdale provides an excellent summary in ‘Globalization, ethnicity and democracy: a view from “the hopeless continent”, in A. G. Hopkins (ed.), Globalization World History (New York, 2002), ch. 9.

68 For one such example, see Willis, Justin, ‘“Hukm”: the creolization of authority in condominium Sudan’, Journal of African History, 46 (2005), 2950.CrossRefGoogle Scholar Dorothy L. Hodgson examines shifts in colonial policy in ‘Taking stock: state control, ethnic identity and pastoralist development in Tanganyika, 1948–1958’, Journal of African History, 41 (2000), 55–78; Pius S. Nyambara traces changing ethnicities in the postcolonial period in ‘“Madheruka and Shangwe”: ethnic identities and the culture of modernity in Gokwe, northwestern Zimbawe, 1963–79’, Journal of African History, 43 (2002), 287–306.

69 Spear, ‘Neo-traditionalism’, 16–25.

70 This is the conclusion of the most thorough account of ethnicity and development in Africa yet completed: Marcel Fafchamps, Market Institutions in Sub-Saharan Africa (Cambridge MA, 2004), section VI.

71 Thornton, Africa and Africans, 102–7.

72 I hope it is not an exaggeration to say that the argument that follows was first formulated in Hopkins, Economic History. Elements of the argument were then applied to Africa as a whole by John Iliffe, Africans: The History of a Continent (Cambridge, 1995).

73 The most important refinements and extensions of the argument are Austin, Labour, Land, and Capital in Ghana; Austin, ‘Resources, techniques and strategies south of the Sahara: revising the factor endowments perspective on African economic development, 1500–2000’, Economic History Review, 61 (2008), 587–624. The wider implications of Austin's Labour, Land and Capital in Ghana are considered in Hopkins, A. G., ‘Making poverty history’, International Journal of African Historical Studies, 38 (2005), 513–31.Google Scholar

74 Mwangi S. Kimenyi points out, for example, that ethnic heterogeneity may affect the provision of public goods adversely but also stimulates the supply of private patronage goods: ‘Ethnicity, governance and the provision of public goods’, Journal of African Economies, 15 (2006), 62–99. Alberto Alesina, Reza Baqir and William Easterly show that ethnic fragmentation in the USA is linked to increased private patronage: ‘Public goods and ethnic divisions’, Quarterly Journal of Economics, 114 (1999), 1243–84.

75 Fafchamps, Market Institutions, section VI, provides a full account of the advantages and disadvantages of ethnic networks today.

76 It is interesting to find that some of the most persuasive evidence comes not from Africa, but from the United States. Alberto Alesina and Eliana La Ferrara have shown that public participation is lower in ‘racially or ethnically fragmented localities’: ‘Participation in heterogeneous communities’, Quarterly Journal of Economics, 115 (2000), 847. It is worth pointing out, however, that their findings are robust partly because they refer to race rather than to ethnicity, the former being more readily identifiable than the latter.

77 As exemplified by Gareth Austin's work, which has linked the growth of land scarcity to the increasing importance of ethnic identities in the twentieth century: ‘Sub-Saharan Africa: land rights and ethno-national consciousness in historically land-abundant economies’, in Stanley L. Engerman and Jacob Metzer (eds.), Land Rights, Ethno-Nationality, and Sovereignty in History (New York, 2004), ch. 11.

78 Bardhan, ‘Institutions matter’, 511.

79 Patrick J. Geary, The Myth of Nations: The Medieval Origins of Europe (Princeton, 2002); David R. Roediger, Working Towards Whiteness: How America's Immigrants Became White (New York, 2005).

80 An interesting recent comment on this ancient subject is Esther Duflo, ‘Poor but rational?’, in Abhijit Banerjee, Dilip Mookherjee and Roland Benabou (eds.), Understanding Poverty (New York, 2006), 367–78.

81 Some unintended and adverse consequences of institutional reform are identified by Bates, Robert H., ‘Institutions and development’, Journal of African Economies, 15 (2006), 1061.CrossRefGoogle Scholar

82 As argued by Patrick Chabal and Jean-Pascal Daloz, Africa Works: Disorder as Political Instrument (Oxford, 1999), who show how policies of structural adjustment, imposed to promote ‘good’ institutions, have led to unintended neo-patrimonial adaptations that nevertheless enable African states to function.

83 Rodrik, Second-best Institutions, illustrates the advantages of adapting best practice to local circumstances. For a less benign view of the same process, see Jean François Bayart, Stephen Ellis and Beatrice Hibou, The Criminalization of the State in Africa (Oxford, 1999).

84 First advanced in Talcott Parsons, Towards a General Theory of Action (New York, 1937), and elaborated subsequently.

85 Diego Comin, William Easterly and Erick Gong, Was the Wealth of Nations Determined in 1000 B.C.? (NBER Working Paper 12657, 2006), 23. Italics in the original.