Quantitative Methods and Economic Statistical Sources for African History
Summary and Keywords
The mode of enquiry in African economic history has changed quite radically in recent years. In 1987, Patrick Manning surveyed practices and databases in African economic history and compared empirical strategies of scholars who studied the African past. Current practice, which A. G. Hopkins called “new African economic history,” incorporates econometric methods. The specific methods chosen and the types of source material used have implications for what kind of questions are asked and how they can be answered. The dominant mode of research in current African economic history, responding to some of the new challenges posed by econometric work by economists, is to create new data sets and databases that allow more consistent analysis of economic change over time.
An African Economic History Renaissance
More than three decades ago, in 1987, Patrick Manning published his renowned survey of the state of African economic history research and analysis. Manning argued that there was a great need for empirically rigorous, long-run historical perspectives, both for explaining the present and analyzing the past.1 In policy universes, often only a cursory glance towards the economic past is deemed pertinent to decision making. Taking postindependence 1960 as the default empirical starting point, the deep economic past has been sorely neglected in contemporary economic planning in Africa and yet is profoundly relevant to understanding long-term development on the continent.
Furthermore, Manning argued, the enduring 1960 “divide” had precipitated three separate academic literatures that had hitherto not entered into a sufficiently substantive dialogue. The first, colonial economic literature, typically covers the period 1900–1960, after which the postcolonial economic literature comes into effect. Running parallel to this economic corpus is the relatively nascent but increasingly consolidated discipline of economic history, spanning precolonial and colonial periods and indeed challenging such divides. There was much, Manning argued, that economic history could bring to bear on the economic presentism that dominated the time in which he wrote.
These different literatures were predictably characterized by different traditions, units of analysis, sources of data, and research methodologies. The post-1960 economic literature, for instance, saw a movement toward the nation-state as the primary unit of analysis, reinforcing the pre-1960 divide with its own focus on empire and trade. Whereas economic history held more in common methodologically with colonial economic literature, it differed greatly in perspective, primarily vis-à-vis the merits of colonial rule. It was against this backdrop that Manning asked how economic history might aspire to convincing insights to past economic structures and how these insights might be communicated effectively to policymakers. Economic history’s potential to speak meaningfully to other disciplines was great. This was especially true of African economic history, which was traditionally more interdisciplinary than other fields of study and varied in its influences and intended audiences.
More recent developments, however, have been shaped by the converse: the increasing contribution of other disciplines to studying economic history. Marking the formation of the African Economic History Network, Jerven et al. duly highlight the notable resurgence of economic history since Manning’s earlier observations.2 This resurgence has been buoyed by “new” postindependence history with its newly opened archives, but also by the entry of other disciplines and their attendant methods, notably that of economics, into the fray. The authors specifically note how A. G. Hopkins’s “new” economic history, whereby history is subjected to the key principles of economics, spurred healthy debate around the discipline as a whole, “resuscitating” it from neglect.3
Jerven et al. press for the need to move beyond the obvious methodological divisions towards a genuinely interdisciplinary approach to economic history.4 This is both possible and necessary given that economists and historians ask different questions that lead to different treatments of data and indeed different modes of theory making. Such differences are not irreconcilable and could be a potential source of strength. Indeed, interdisciplinary perspectives temper the urge to identify singular, overdetermined root causes to African economic development. Rather than focusing on defining differences, all disciplines with an interest in economic history share similar challenges in aggregating quality, consistent, and long-run data with which to work.
It is this endeavor that Jerven et al. place as central to the rationale of the nascent network. Indeed, they argue, attaining “reliable and valid data” is (and should be) as much the central concern of African economic scholarship as it is elsewhere in the social sciences.5 They go on to audit the current state of numerous data sources, metrics, and their supporting literatures, from familiar macrolevel indices to those that focus on the individual or the firm. They seek to pool assets and remedy gaps through a common industry of data generation. This will enable economic history to remain empirically grounded while ideally engaging in more theory generation than had characterized the discipline to date. This would be duly spurred on by the influence of other disciplines as well as the growing capacity for, and confidence in, such an undertaking outside of Western hubs of influence.
Jerven et al. summarize, as Hopkins did before them, the importance of economic history for understanding contemporary economic development, the latter acting as an engine for new and improved work on the former.6 For Hopkins, recent contributions from economists served to revitalize a discipline seemingly in decline, suffering from a lack of “big arguments” in its defining debates. Historians had arguably focused on more peripheral issues, whereas economists placed big questions around poverty and economic development as central to research agendas. Thus, big questions duly precipitated “big arguments” that were lacking in economic history, which had lost its “cutting edge.”7
Hopkins focuses on two of the most prominent debates, wherein econometrics were brought to bear on explaining long-run economic development in Africa: the influential “reversal of fortune” and “ethnic fractionalization” theses. The reversal of fortune thesis, drawing on the seminal work by Acemoglu et al., holds that Africa was relatively prosperous in 1500, before colonization took place.8 Easterly and Levine’s ethnic fragmentation work holds that relatively high ethnic diversity directly impeded economic development, leading to divergent interests, competition, and ultimately increased violence.9 Both theses place social institutions, including “non-settler” colonial rule, the slave trade, and ethnicity, as central to understanding Africa’s “growth tragedy” in the modern era and above other commonly cited factors such as the continent’s natural endowments. Both theses have in turn spawned great debate and additional literatures, exploring the ideas in different contexts.
While welcoming fresh insights and ideas to economic history, Hopkins issues words of caution in bringing econometrics to bear on such areas. Hypotheses duly tested are only as good as the quality of empirical data, which is generally found wanting in the African context: “Regression analysis is only as robust as the numerical evidence it draws on.”10 Proxies, such as those that Acemoglu et al. use for prosperity, specifically population density and urbanization, can be deeply problematic in their assumptions.11 Techniques such as backward projection come with their own “heroic assumptions,” which are only stretched further when pushing deeper into the past. Concepts such as ethnicity are “slippery,” difficult to define and, thus, to isolate and measure.12 The rush to generalize predictably masks contextual nuance and the importance of the particular.
Austin holds similar reservations about the ventures of economists into theorizing on the deeper past.13 He poses questions about the reliability of quantitative data for all time periods and the deep past in particular. More fundamentally, and echoed by Hopkins, Austin asks questions concerning the seeming lack of conceptual space offered to the agency of Africans, specifically in the reversal of fortune thesis.14 This thesis encompasses the perennial assumption that Africans were somehow undone by a mercantile world not of their own making, and that global history is determined by the rise of the West over the rest.15 These two challenges, while profound, are not insurmountable, with Austin indicating how they might be mitigated with appropriate supporting historical literature. Most far-reaching in his critique, however, is the urge to “compress” history, whereby “so much history” is condensed into binary variables.16 This tendency occurs both vertically, across time periods, and horizontally, between different contexts. The starkest compression of history, for instance, infers an equivalence between degrees of colonial rule that should not and indeed need not be introduced for the thesis to stand.17
Despite their detailed misgivings, Austin and Hopkins agree that this new literature must be taken seriously. For Austin, there is merit in examining growth theories against long-term history.18 Similarly, according to Hopkins, such necessary scrutiny demands that historians move away from a fixation on smaller matters toward engagement in debates on notably larger ones.19 Theoretical parsimony can only aid the overarching goal shared by all those cited to this point, that history in turn be taken seriously by policymakers in contemporary economic planning. Thus, these authors agree on the intrinsic value of such theory making, despite its obvious shortcomings. Historians can and must engage.
For Fenske, however, the content of such big theories is less critical than their supporting methodologies.20 What sets recent developments in economic history apart, therefore, is not the attention-grabbing causal theories of Acemoglu et al. and Easterly and Levine but rather the quantitative, econometric methodologies themselves, with their carefully crafted assertions of causality.21 Hopkins’s focus on but two theories and their specificities, to the exclusion of other notable studies, serves to misrepresent what is truly “new” about economic history.22 Fenske duly audits the innovative methods of new economic history: regression analysis, fixed effects, instrumental variables, and regression discontinuity, as well as the various literatures that utilize them. Such literatures, he argues at length, are more specific, more nuanced, and more methodologically robust than Hopkins allows.
Fenske thus labors to demonstrate the breadth and richness of key econometric-derived literature regarding the reversal of fortune and ethnic fractionalization theses, countering Hopkins’s critiques regarding the robustness of models and the validity of their data. Fenske then supplements these two corpora with other burgeoning literatures that again rely on detailed and, for Fenske, crucially robust econometrics. He concludes that historians must generate both context-specific historiography and testable hypotheses that infer causal relationships if they are to enter these growing debates substantively. Only in this regard can they offer research of “external validity” as well as broader “policy relevance,” a conclusion destined to foster discord.23
Jerven intervenes in this growing “clash of disciplines” while being mindful of the enduring potential for conflict between economists and historians. He restates, foreshadowing later reconciliatory arguments, that there are of course important methodological differences between these two camps; in questions asked, the treatment of data, and the role of theory making. With this in mind, Jerven traces the “coming of age” of development economics. This demanded the expansion of the discipline’s historical horizons as both the availability of data and the demand to isolate the main drivers of Africa’s apparent economic stagnation grew from within contemporary policymaking circles: the specifically “African” effect. Jerven, as others before, unpacks the various, at times precarious, assumptions underpinning this endeavor, not least regarding the nature of “underdevelopment” and its measurement.24
Moreover, Jerven tackles Fenske’s somewhat self-affirming position head on, whereby the robustness of a historical argument is “subject only to econometric criteria,” stating that there are different tests of whether historical evidence bolsters validity.25 The question must return once again to the quality of data and the painstaking steps historians have undertaken to generate it in the African context. Arguments concerning the robustness of data models must not circumvent the fundamental issue of data quality. Such differences in assessing the validity of data, however, do not preclude the search for common ground between economists and historians.
Jerven concludes that genuinely cross-disciplinary work must hold commensurable assumptions, which demands a broadening of robustness checks beyond econometrics alone. This is not only possible but necessary in forging cross-disciplinary conversations. This conclusion echoes Austin and Hopkins and cautiously welcomes the entry of new disciplines, methods, and perspectives, specifically econometrics, into understanding African economic history. The focus on causal pathways may indeed provide fertile ground for cross-disciplinary conversations. The acceptance of these potential new insights, however, is qualified. Econometrics’s somewhat insular conception of robustness remains its chief analytical weakness, serving to circumvent full and frank assessments of data reliability. The core question, therefore, is how to forge core principles in this regard that span disciplinary divides.
How History Matters for Economists Studying the African Past
While respective disciplines ask different questions and demand different treatment of the data, it is important to note that the availability and nature of relevant data also shapes the way in which historical analysis is undertaken. Most notably, the study of poverty and growth in sub-Saharan Africa has been predictably linked to the availability of relevant data. Time-series data on economic growth, which is central to writing economic histories, default to 1960 as year 1, thus focusing on the postindependence period. Economic histories of poverty that use the dollar-a-day metric default to 1990 as year 1. Both these base years are artificially construed, contingent, and, indeed, misrepresentative. Both these examples spotlight how the availability of data restricts analysis to a particular time frame, privileging certain research questions over others.
Jerven has argued that the literature took this form due to three contributing factors. The first was conjunctural. The lack of growth in the 1980s and continued stagnation in the early 1990s shaped the prompting research questions. From the vantage point of 1995 the question of what caused a permanent growth failure in Africa seemed pertinent. It is questionable whether it would have made sense to approach the history of African growth in the same manner if one was writing from 2015 or from 1975. Yet the question has misleadingly remained “Why is Africa poor?”26 This has been the central research question for a great deal of empirical work in economics and economic history until very recently.27 African states were misdiagnosed and dismissed as incapable of development based on observations made during the 1980s and early 1990s.
The second factor was the importance of the statistical databases. The availability and the character of poverty numbers and economic statistics have shaped the historical analysis of poverty and growth in sub-Saharan Africa. Quantitative economic histories of economic growth are written using time-series data on economic growth. In this case year 1 has been 1960 and hence economists have focused on the postcolonial period. Histories of poverty using the dollar per day metric use 1990 as year 1. Both are, of course, artificial starting points, but data availability restricts us to this time frame, and what kinds of questions can be investigated.
The narratives of economic growth in African economies change dramatically if the starting point of 1960 is rejected. Notions such as “chronic growth failure” and the “bottom billion” would not hold if we discarded this short time frame. Similarly, narratives of trends in living standards in Africa are shaped by the current configuration of what constitutes poverty knowledge at the World Bank. The short history of poverty by numbers in Africa is misleading because it fails to contextualize the history of poverty in the 1990s and 2000s in a longer time perspective. Thus, extreme and chronic poverty became the “imagined fact” that prompted the narrative on Africa and poverty.
The dearth of immediately available data on trajectories of poverty has led researchers to take shortcuts when studying long-term historical roots of poverty in Africa. Rather than carefully outlining the trajectory in economic growth or poverty over past centuries, the economic literature is content treating GDP per capita as standing in for lack of growth in the past. Thus, according to this method, if you can correlate the income distribution today with some explanatory variable in the past, you have found the historical roots of Africa’s relative poverty.
Thus, the data availability has been particularly influential and the influence has been dramatically strengthened by the method of investigation. The starting point for what I have called “the quest for the African dummy” is arguably a 1991 article by Robert J. Barro.28 His essay exploring the causes of economic growth in a global sample of countries provided the template for the next decade of research.29 The investigation was made possible by the availability of the global data sets on economic growth, such as the Penn World Tables, the arrival of which was considered to be “an important statistical event” that expanded the boundary for empirical research.30 The combination of a new methodology and new data sources spurred a great amount of research. The articles and essays this research generated used the methodology of cross-country growth regressions in which the dependent variable was the average growth rate of per capita GDP. Researchers added different independent variables, or interactions of independent variables, to the initial baseline estimation in their search for new insights about what correlates with economic growth).31
In cross-country regressions, variation in average GDP growth at the country level is the dependent variable. The independent variables are the things that explain economic growth. Together they can explain why some countries have high average GDP growth and others have low average GDP growth. The terms “dependent” and “independent” are important because they depict the causal direction. GDP growth depends on type of policy, regime type, and so forth, and not the other way around. If it is the other way around, then the explanatory framework is compromised.32
Following the seminal publication of Barro’s model, which used cross-country regressions with global data sets on growth in GDP per capita and looked for correlations with other variables such as educational attainment, number of assassinations, black market currency rates, and more, a literature developed that was aiming to find the global “determinants of growth.”33 After Barro had controlled for many different variables, one central finding remained. There was a large and significant African continent “dummy variable.” A dummy variable takes the value 1 or 0; in this case it took the value 1 if the country was situated on the African continent and 0 if it was not. In the regressions, the dummy variable remained significant. Barro’s interpretation of the dummy was that the analysis had not yet fully captured the characteristics of a “typical” African country.34
Thus, a scholarly search for the right variables began and, while over the course of a decade of research many correlates with slow growth were found, the African dummy variable proved stubborn to remove. (I have labeled the literature that shares this motivation the “quest for the African dummy” and, as I have argued elsewhere, the quest was ultimately both unsuccessful and highly influential.)35
In a review of the empirical growth literature, Durlauf et al. referred to this scholarly production over the following decade as a “growth regression industry.”36 As early as 1998, Pritchett observed that the growth experience of most developing countries had been characterized by instability rather than stable growth and warned that the “exploding economic growth literature” was “unlikely to be useful.”37 The general growth regression literature has been described as disappointing. One assessment concluded that the “current state of the understanding about causes of economic growth is fairly poor” and that “we are in a weak position to explain why some countries have experienced economic growth and others not.”38
Whereas the ultimate outcome of the search for causes of slow growth in Africa was inconclusive, an unintended consequence was the idea that African economies were captured in a chronic failure of growth became an accepted stylized fact within economics. Ultimately, this led to growth econometricians, equipped with new models and data sets, searching for the cause of low income measured as GDP per capita today instead of the lack of growth over time.
The main contenders in the empirical literature searching for a root cause of underdevelopment can be organized in three chronological strands according to whether they emphasize the negative effect of initial conditions particular to Africa’s geographical characteristics, the decisive impact of the slave trade, or the effects of European colonization.39 Many of the contributions are variations of the same argument: that a particular historical event or factor endowment led to a particular institutional constellation that has had lasting economic effect. The big question in the literature remains which of these historical events had the decisive impact and through which transmission channels it continues to have an effect. A divisive issue in the economics literature is the relative importance of institutions and geography or, as was succinctly phrased in debate, “institutions rule” versus “institutions don’t rule.”40 The list of factors and variables are growing and are sometimes in conflict with each other. However, they are also often listed as a growing body of evidence that indicates that “history matters.” Indeed, the approaches to the past are quite different between economists and historians.
The list of suggested historical events, or aspects of them, that had decisive impacts on long-term patterns of economic growth is long and still growing. Debraj Ray suggests that inadequate national infrastructures can explain a permanent state of global inequality.41 Nunn concludes that while “the literature has made considerable progress in showing that history matters, what remains less well understood are the exact channels of causality through which history matters.”42
Ray, pointing to the work of both Acemoglu, Johnson, and Robinson and Sokoloff and Engerman, argues that the divergence in income levels in Africa today can be explained by “situations of stagnation in which the losers (or potential losers) control political institutions and shape the rules of the society. Losers defend an old system—likely one born under a colonial umbrella—and so impede progress.” He argues that even when winners are granted control, “they may block all redistributions that spread the growth process to other sectors.” To analyze the causes of divergence “it will be of great importance to build a useful taxonomy of institutional performance (and reactions to such a performance) depending on who has control.”43
This is an admirable ambition and seems to be one where some common ground can be reached. Building a useful taxonomy would certainly require a combination of methodologies, drawing on the expertise of economists, historians, anthropologists, and others. When asked to explain what path dependency really is, economists and business scholars reach for the famous example of the “QWERTY” keyboard.44 This keyboard was developed for typewriters to slow down the speed of typing so the keys in the typewriter would not become entangled. Today we use laptops and touchpad keyboards and could use more efficient designs, but path dependence means that we are stuck with the old inefficient way of typing. That economic and institutional pressure has succeeded in keeping this version of the keyboard makes intuitive sense. But it is a rule that has many important exceptions.
The importance of institutions in explaining policy choice in postcolonial African economies was perhaps best illustrated by Robert Bates.45 His influential model of political economy attempted to explain why some African countries tended to have a policy regime that favored agricultural exports while other national policies discriminated against them. Although marketing boards were a colonial innovation, how they were used after independence depended on the postcolonial political economy. Thus, the theory showed that path dependency via a nation’s colonial legacy did not apply uniformly, nor did it last permanently, as structural adjustment led to institutional reform. The irony of path dependency arguments is that they invariably turn out to have limited temporal validity.46 The lesson is that both the economic evidence and how institutions are classified need to be fully historicized.
Endowments, or initial conditions in a narrow sense, are not a good predictor of economic performance. As Anthony Hopkins put it: “Comparing the natural resources and climates of different parts of the world in order to draw conclusions about whether they stimulated or retarded the economic progress of particular societies is a tempting but unprofitable exercise—rather like trying to decide if life is more difficult for penguins in the Antarctic or camels in the Sahara.”47
Issues such as which technology a nation adopts and what investments it makes in physical and human capital need to be evaluated in light of its specific endowments and local conditions. This context should be understood before any assertions are made that irrational policies or poorly managed institutions have hampered economic progress. When internal factors—such as the design of political systems—are used in analysis across countries the comparison must be reciprocal. Pomeranz explains the main principle of reciprocal comparison as “viewing both sides of the comparison as ‘deviations’ when seen through the expectations of the other, rather than leaving one as always the norm.”48 The faults arise from the failure to do so, which demonstrates the fault of using a subtraction approach when explaining slow growth in postcolonial Africa.49 Education, technology, infrastructure, and institutions can be interpreted as growth-retarding or growth-enhancing only in their own physical context and only with respect to the relative development level of the economy being examined.
In the methodologies used in economic history and economic anthropology, the question of whether institutions retard growth needs be considered carefully in the social, historical, and economic context of those institutions. The optimal design of institutions or policies is not a universal standard, but changes in response to development level and physical constraints and efficient institutions are in part a result of, and not an initial condition for, economic development.
The most famous purveyors of the theory that institutions matter are Acemoglu and Robinson.50 Their theory is simple: successful nations have productive institutions. The nations that failed did so because their institutions were extractive. Productive institutions are pluralistic (or inclusive) and centralized. Acemoglu and Robinson contend that nations fail to develop when their institutions fail to exhibit either of these two characteristics. Yet even when development fails, institutions can persist when they benefit a few elites. Centralized, authoritarian regimes may produce limited growth, they argue, but resources in such regimes will not be redistributed in a way that makes for real development because redistribution may destabilize the extractive institutional settlement. This makes sense in theory, but the empirical evidence is built on shaky foundations.
To support their questionable correlations between colonizers, institutions, and income today, Acemoglu and Robinson point to the poorest place in Africa, the place that is also notoriously famous for having “poor institutions” today, namely Kongo. Or Congo, then Zaire, and then Democratic Republic of Congo. The authors do take care to note that present-day Congo does not cohere with the historical Kingdom of Kongo (which was located within present-day Angola), yet it is implicitly argued that the institutional heritage lingers in the Democratic Republic of Congo in this very discontinuous political history.
Their rhetorical question is “Why didn’t farmers in the pre-colonial kingdom of Kongo in sub-Saharan Africa adopt the plow?” Their answer is because of the institutions; that is, because “they lacked any incentives to do so.”51 Specifically, Acemoglu and Robinson argue that the fear of farmers that their crops would be appropriated due to an absolutist king’s control of output and manpower took away incentives for them to invest in tools that would increase their productivity. The slave trade, colonial rule, and the postcolonial regime of Mobuto Sese Seko all contrived to keep this region poor. Thus, Acemoglu and Robinson conclude, “The interaction of economic and political institutions five hundred years ago is still relevant for understanding why the modern state of Congo is still miserably poor today.”52
This rather general statement does have some truth in it. The Democratic Republic of Congo is poor today, and the rule of Mobuto was not exemplary. Despite the historical discontinuities in the theory Acemoglu and Robinson present, one might be willing to buy the correlation. But it is generally conceded that political institutions are not the primary factor that explain the slow acceptance of the plow in sub-Saharan Africa.53 In the tropical forest zone, including the Congo Basin, the prevalence of trypanosomiasis, a parasitic disease that causes emaciation, anemia, and death in farm animals, made it impossible to keep cattle. Thus, without draft animals the plow was not efficient. Furthermore, in most places in precolonial Africa, including Kongo, land was relatively abundant, and therefore investment in land was discouraged, not by excessive state intervention but by this very abundance of land. Finally, as many colonial administrators would later find out, the plow is not universally desirable. Tropical soils are fertile only at a shallow depth and using plows increases the risk of soil erosion. Thus, regardless of the kind of postcolonial regime that came to power, the plow did not make sense. And neither does the argument that an oppressive political regime was linked to the reluctance of farmers to change to adopt the plow in precolonial times.
One important motive for using historical variables is technical: to facilitate the use of instrumental variables to take care of the endogeneity problem that arises when factors that are supposed to affect a specific outcome depend themselves on that outcome. This was the difficulty encountered by scholars who were attempting to quantify the effect of aid, infrastructure, and corruption on development. Acemoglu, Johnson, and Robinson’s seminal contribution was to use European settler mortality rates as an instrumental variable for risk of capital expropriation.54
To understand what Acemoglu, Johnson, and Robinson did, we need to understand the concept of an instrumental variable. If you wanted to measure the effect of police on crime, you would instantly run into the problem of reversal causality. While police may reduce crime, crime increases police. So that if there are more police in one area, there is probably more crime there too. A clever way of solving such puzzles is to add a variable that is unrelated to crime but is related to number of police out on the streets. Weather could be such a variable. Hypothetically, if it is sunny there might be more police on the streets. In a regression framework one can therefore use the effect of weather on police, insert that variation in the second equation, and then measure this instrumental variable effect on crime.
Acemoglu, Johnson, and Robinson revived debates on colonization with their controversial reversal of fortune and colonial origins of comparative development theses.55 The former argues that the poorest non-European areas of the world five hundred years ago are now among the wealthiest and that, conversely, the formerly wealthiest areas now are among the poorest. Thus, the last five hundred years of economic development constitute a reversal of fortunes in non-European areas. This “reversal of fortune” is explained by European colonization. In the poorer areas, Europeans settled in greater numbers and invested in the creation of costly but “good” institutions in the colonies. The second thesis builds on the first and explains the current comparative development levels in the non-European world using an instrumental variable approach. Acemoglu, Johnson, and Robinson argued that the mortality of European settlers determined the numbers of settlers that the colony attracted. In turn, this determined the quality of the institutions that were set up in the colony.56 Specifically, the argument distinguishes between colonies where “extractive” institutions were introduced and colonies where “productive” institutions were established—the latter of which are the rich ex-colonies today.
The problem with measuring the economic effects of colonization is the possibility of reverse causality. For example, did colonizers choose resource-rich economies to settle in or did they create wealth where they settled? Acemoglu, Johnson, and Robinson attempted to get rid of the endogeneity of those two variables—colonization and income—by finding a variable that does not have a direct causal relationship with income today but determined the number of settlers during the colonial period.57 They suggest a variable of settler mortality, using historical data on the mortality rates of soldiers, bishops, and sailors who were stationed in the colonies from the 17th to the 19th century. Their argument is that although the mortality rates of European settlers more than one hundred years ago did not have any effect on GDP per capita today, mortality rates did have an impact on the development of institutions, or more specifically, legal and political institutions that would protect private property. Thus, Acemoglu, Johnson, and Robinson feel that they have shown that it is the quality of institutions that matter for development today. And because development today obviously does not have a causal effect on settler mortality a long time ago, there is no chance for causality reversal in their methodology. It is this innovative way of using historical evidence to avoid reverse causality that has given their article such a high standing among economics departments.
Yet their method has its critics.58 The data on settler mortality has been subject to criticism and has been shown not to be robust with regard to other justifiable data points in the settler mortality data.59 Deaton warns that those who use instrumental variables often imply that the variable is external to the question at hand when in fact it is simply exogenous.60 That is, the variable is external to the model yet not truly external to the question the model grapples with. This makes intuitive sense in the example of settler mortality because it was probably related to climate and diseases. These factors may also affect development today.61
The problem is that the story Acemoglu, Johnson, and Robinson present is built on the principle that the disease environment, particularly malaria, was deadly for settlers but does not have a detrimental impact on economic development today. Thus, if Sachs is right, then Acemoglu, Johnson, and Robinson must be wrong. Let’s push these two stories a bit further and think about policy implications. According to Sachs, as soon as malaria is eradicated or people are properly protected, incomes will increase and other institutional developments will follow.62 Thus, development funds that distribute malaria nets are being put to effective use. However, if Acemoglu, Johnson, and Robinson are correct, the individuals who are cured of malaria would still be held back by deficient institutions such as the inadequate protection of private property. In an economy with weak institutions it is likely that the agencies that distribute malaria nets would not be functioning efficiently. The robustness of these published findings on the relative importance of institutions versus diseases can be contested by assessing the coherence of the different explanations. The econometrics in both the Sachs model and the Acemoglu, Johnson, and Robinson model are internally robust, but judgment about the credibility of the two arguments, and thus the causality question, ultimately depends on which of the stories most convincingly explains the observed pattern.
Vantage Points: How Data Availability Has Shaped Historical Investigation
The availability and the character of poverty numbers and economic statistics have shaped the historical analysis of poverty and growth in sub-Saharan Africa. Quantitative economic histories of economic growth are written using time-series data on economic growth. In this case year 1 has been 1960. The availability of this evidence in part explains why economists have focused on the postcolonial period. If one writes histories of poverty, using the dollar per day metric, year 1 has been 1990. Both are, of course, artificial starting points, but data availability restricts us to this time frame and to the kinds of questions that can be investigated.
Similarly, narratives of trends in living standards in Africa are shaped by the current configuration of what constitutes poverty knowledge, particularly at the World Bank, and what does not. That may be a mistake because the brief history of poverty by numbers in Africa is characterized by gaps and inaccuracies in the underlying data, and because it fails to contextualize the history of poverty in the 1990s and 2000s in a longer time perspective. This has resulted in persistent chronic poverty becoming the “imagined fact” that prompted the narrative on Africa and poverty.
One of the central challenges is to connect the colonial period to the postcolonial period. When this is done, there is a conspicuous absence in the central narrative of the economic history of 20th-century Africa. In previous analytical models, the postcolonial period has been treated as an “outcome” (and a dismal one), whereas a historical event in the colonial or the precolonial period has been treated as a root cause of this outcome. The aggregate pattern, particularly in terms of poverty and growth—though it is also evident in taxation measures—is a prolonged period of growth from the 1890s into the 1970s. This gave way to widespread failure and decline in the 1980s, followed by two decades of expansion since the late 1990s. Thus, it seems increasingly evident that growth failure in the postcolonial period has taken up undue space in explanations of African economic development. The old pattern in the literature was to regard the postcolonial period as a dismal outcome, as represented by a low GDP per capita today, which should be explained by a colonial or precolonial event or variable. With new data sets, the boundary of investigation is pushed backwards to allow the evaluation of trajectories of economic development across the colonial and postcolonial period.
A striking example of the power of data sets and the importance of vantage point comes from the debate on global warming. It took so long to notice global warming because of weaknesses in the global mean temperature data set.63 In the 1970s, the data series went back only a few decades, and from that perspective it looked like temperatures had been falling since 1940s. But the current data set, which spans 1860–2010, shows that the temperatures are indeed trending upwards.
When you do not have a complete picture, your impression may be biased. Every cross-country growth regression starts with the goal of using a global sample, but because of incomplete data availability, some countries are excluded from the equations. The assumption that a country’s lack of ability or opportunity to conduct a household survey is not correlated with any other determinants of poverty and slow growth is a dubious one. Continentwide statements on recent economic growth and trends in poverty are based on observations from a small subset of the world’s nations. Sometimes the missing observations are “made up” at the international organizations who fill in the gaps when data sets are assembled and at other times economists extrapolate data from neighboring countries to fill in information that would not exist otherwise.64
These are key weaknesses in the data, and they shape how growth in Africa is explained. Data are missing for some countries and for most economists the story of growth starts in 1960. The GDP data records some growth decline in the formal sector, but it fails to capture economic change in the informal sector. Perhaps most important for many of the concepts that we think are theoretically related to economic growth, concepts such as “policies” and “institutions,” there is no direct quantitative data. Instead different survey rankings, subjective metrics, and various kinds of proxies are used to capture effects from “policies” and “institutions.”
Economic historians welcomed explorations of the historical roots of economic conditions, albeit with some reservations. Austin called contributions such as the those suggested by Acemoglu, Johnson, and Robinson a “compression of history.”65 In essence, the problem is that no GDP data sets with growth time series go further back than 1950. Studies of economic growth have thus been confined to the postcolonial period, simply because there are no data sets available for earlier periods. A telling example of missing data is provided by Artadi and Sala-i-Martin, who wrote an article about what they call the “economic tragedy of the 20th century.”66 One might reasonably expect to see some evidence from the earlier part of the century in their study, but the data marshalled covers only 1960–2002. Only about 40 percent of the century is covered, and the remaining sixty years are assumed to have had no importance.
Of course, information and evidence about long-term economic history do exist and there is even GDP data that goes further back in time. The most widely used resource and the one that currently provides the longest time series is the Maddison data set.67
Table 1. African and World GDP per Capita, 1–1950 ce
Source: Angus Maddison, “Historical Statistics of the World Economy: 1–2006 AD.” Groningen Growth and Development Centre, University of Groningen, The Netherlands, 2009. All values in constant 1990 International Geary-Khamis dollars. Note that the only African countries for which Maddison has individual income estimates for this period are Algeria, Egypt, Libya, Tunisia, and Morocco.
This data set includes annual international GDP per capita data for all African countries for the period 1950–2006, although it includes observations for the continent (as a whole) back to year 1. There are significant problems with this data. First, no actual GDP per capita estimates were prepared for the post-1950 period by country (or for the continent, for that matter). Second, the data were created using assumptions and projections. The GDP per capita estimates in Table 1 suggest that, on average, Africa was richer than other parts of the world in year 1. This is because the estimated income for Egypt was higher than elsewhere in the world. Other data implies that Africa gradually fell behind the rest of the world due to stagnation in income per capita from 1000 to 1820. From 1820 to 1913, according to this data set, income per capita increased in Africa but fell behind the world on average as incomes increased more rapidly in other places. From 1913 to 1950, Africa stopped falling behind the world average, according to these estimates. In this period the average GDP per capita growth in Africa was 1 percent. In this data set, population growth in Africa is also estimated at 1 percent, meaning that the total GDP of the economies in Africa were growing at an average rate of 2 percent over this period.68
The aggregate picture that is drawn in Table 1 does not reflect what is known about periods of export growth, state formation, and wealth accumulation in parts of Africa. There were large flows of commodities and factors of production, both internally and externally, during the Atlantic slave trade and the crash-crop revolution. Kingdoms rose and fell, colonial empires were established, and railways and mines were developed, and yet the GDP per capita measure in the Maddison data set barely blinks. The African GDP series from year 1 to 1950 is inaccurate and incredible. Moreover, it inadvertently distorts our view of history. The problem with only having one number for one continent with one observation every century or so is that it gives the impression that there is very little history to explain. It can also be taken to support the notion that Africa has always been poor and has been permanently stuck in growth failure.
Indeed, Bloom et al. use an earlier version of the Maddison data for that purpose. They conclude that for the past two centuries, “Africa’s poor economic growth has been chronic rather than episodic.”69 this paints a misleading picture of African economic history. Well-documented growth in African markets for currency, labor, and goods led to Smithian growth through specialization in some areas.70 Moreover, the Atlantic trade brought new technology, for instance the introduction of new cultigens that must have led to increased total factor productivity.71 In addition, the economic growth that occurred in Africa could not have happened without significant investments in perennial crops, land improvement, and transportation infrastructure. Finally, these growth episodes induced historically documented changes such as new markets in land and labor, strengthened states, and increased living standards. Lasting and recurring economic growth episodes spurred political and institutional change. It would be a mistake to interpret the economic history of Africa as if modern economic growth had not touched the continent.
But instead of focusing on such trajectories of growth and seeking to explain them, in the second generation of the growth literature economists took an analytical shortcut. In their defense it was necessitated by the paucity of evidence, but it did lead economists astray. The new regressions stopped using “growth” measured as the percentage rate of change in GDP as evidence and switched to explaining country-level variations in “income” as measured by GDP per capita today. When is today? For most of these regressions, “today” was the year 2000. As already noted, in principle the link made in the economic growth literature between growth and income is relatively straightforward: low income today must be the result of a lack of income growth in the past. The next step is more complicated. Because the evidence does not go very far back in time, the growth regression literature has not provided any evidence that actually explains any differences in long-term growth rates. We are meant to take a leap of faith that the distribution of GDP per capita in the year 2000 contains some cumulative difference that regressions can exploit to tease out some wisdom about why some countries have been successful and other have not.
In addition to data availability, the type of data, categories, and mode of observation matter as well. A central criticism of much of African history is that “the visions of Africa often derive from Europe and come still predominantly from the Western World. Our perception of the African past has always been a European perception.”72 Thus, when interpreting social and economic change in African societies, it is particularly important to assess the bias and subjectivity of the authors who produced these sources. Administrators and explorers preceded scholars in making observations, and their observations were made using pre-1900 categories. The early scholarly observations are also dated. Their racial and political views shaped the type of information gathered and how it was categorized. The basic question is whether the knowledge gained through these sources is at all useful. The discipline of African history has long recognized these problems, and economists who are seeking to contribute to the interpretation of the African past would do well to listen to the caveats of their historian counterparts.
Berman illustrates the problem nicely in a discussion on colonial control and the knowledge problems that the colonial administrators had. In colonial Kenya there was a need for political decisions regarding property rights.73 Colonial administrators had unsuitable concepts of these rights and inapplicable notions of “ownership” derived from their own experience, which did not fit Kenyan conditions. The arbitrary decisions that were made in allocating land to some groups rather than others still lingers in contemporary Kenya. But instead of learning from the inadequacy of these knowledge categories, they reappear as historical facts in regressions attempting to correlate institutions with income today.
Of particular curiosity is the use of regressions of Murdock’s “Human Relations Area Files” and “Ethnographic Atlas” in lieu of evidence on institutions without comment on its historical validity or its current standing as evidence in the discipline where it originated: anthropology.74 As Austin notes, its use in historical arguments is doubly ahistorical.75 The observations on institutions in the data set need to be historicized, not only by dating when the observation was made but also in terms of what is considered acceptable evidence today in the discipline of anthropology.76 Cogneau and Dupraz have provided an incisive critique of the use of the Murdock data base for causal inferences.77
One of the biggest challenges for African history research has been a dearth of reliable evidence. The ingenious solutions scholars and particularly historians have devised to overcome this challenge have become one of the strengths of the field. When the econometric study of Africa adopts a longer time perspective, the lack of quantitative evidence limits the types of questions that can be asked and answered, and we may end up with Austin’s “compression of history.”78 African economic data is limited both in availability and quality.79 Furthermore, the data sets are biased in two respects: we know less about economic change for the period before 1960 than we do for the period after 1960 and we know much more about exports than we do about production.80 Thus, using quantitative approaches exclusively may mean that we understate the importance of economic change in the early period and ignore internal economic dynamics in both the early and later periods.
In response to this dearth of the more traditional measures of development, especially prior to 1960, new and innovative methods are beginning to emerge.81 Anthropometric research such the work of Baten and Moradi connects data on the height and weight of a population with the population’s general well-being.82 This builds on the reality that poverty is correlated to malnutrition, which, in turn, is negatively correlated with physical stature. Thus, ceteris paribus, a taller population, will experience less malnourishment and poverty. Data from colonial army registers or slave manifests is being drawn upon in the growing field of the anthropometric history of Africa. This approach is advantageous for several reasons other than increased statistical availability. Because height is an outcome of well-being, it can be understood to be an outcome of development rather than an input. The same cannot be said of income, for example, which is used to purchase things that would drive individual well-being. Further, because the absence of malnourishment and its consequences are favorable objectives that are not context-specific, using the increasing absence of malnourishment as a proxy of “development” helps to avoid the problems that stem from using approaches that could be criticized for being driven by Western values.
An additional wealth of information is being drawn from colonial records in the forms of trade and wage statistics. In the case of the former, efforts are being made to construct reliable price series from the records of European traders and trading companies. This builds on existing work conducted by researchers including Curtin and Austin.83 On the other hand, wage statistics are being gathered from colonial blue books by others, including Frankema and van Waijenburg.84 This stream of research endeavors to generate data on real wages across time and distance. When measured against the price series that are being generated, one can calculate a subsistence wage (what wages were required for a minimum subsistence level of goods) and compare these across different contexts. De Zwart, Du Plessis and Du Plessis, and Fourie have all taken similar approaches to demonstrate that living standards in the 18th-century Dutch Cape Colony rivaled those in England and Holland at the time.85
At the macro level, the fulfillment of state functions is correlated with a more developed state. For example, the establishment of minimal (Acemoglu, Johnson, and Robinson used the term “extractive”) states in many areas stifled the development of institutions outside of those designed to control borders and impose taxes.86 This inhibited the further development of state institutions, an imposition that is suggested to have endured to this day.87
Also enduring are colonial institutions that delivered education. To this point, contemporary enrollment rates are linked to the delivery of education in colonial times—typically through missionary activity.88 Namely, the type of missionary activity (Protestant versus Catholic, for example) led to differences in outcomes such as literacy rates, which have also endured to the present day.
Tangential to this is work being done in the domain of political economy. Ray calls for a taxonomy of institutional performance (and reactions to it) viz-à-viz political elites.89 While economists have traditionally focused on the origins and persistence of these institutions, much could be learned through a more thorough understanding of the impact of these forces in the formation of contemporary sub-Saharan states.90 Building on the experience of the tigers of East Asia, the examination of the patrimonial state of sub-Saharan Africa is a growing field of interest.91
There has been a recent surge in quantitative research on long-term African economic development. Most of this literature focuses on establishing relationships between historical events and current income levels. This focus is a result of one basic data constraint. Estimates of national income and economic growth for African economies are only available back to 1950. As noted, this results in the economic growth literature evaluating long-run economic performance, often using evidence from a very recent past. Typically, the literature has focused on explaining variation in income per capita today, thus asserting causal relationships across time periods, but without being able to account for different trajectories of economic development, which results in what has been called a “compression of history.” Certainly, it would be an improvement to focus on explaining trajectories of African economic growth instead of explaining the lack of economic growth. The literature on African economic development has already moved away from a “compression of history” towards focusing on analyzing and evaluating trajectories of economic development during the colonial and postcolonial period in Africa.
Recently, there have been efforts to provide historical national accounts data for a range of African economies. Particularly for the Gold Coast and Ghana there is mounting direct and indirect evidence of periods of economic expansion and increases in living standards during the colonial period. Moradi used anthropometric measures and reported that heights in cohorts born in 1905–1920 (taken from army recruits for World Wars I and II) outgrew the cohorts born in 1880–1893 “by an astonishing 2 cm on average implying a considerable improvement in the physical quality of life.”92 Frankema and van Waijenburg collect nominal wages from colonial records and use price data from the same sources to get a measure of real wage developments. They find that real wages started growing in the 1910s and grew rapidly in the 1920s, and that growth was sustained into the 1960s.93 Bringing further supporting evidence of widespread economic development in the late 19th and early 20th century in the Gold Coast rather than a narrow export expansion, Austin argues that the expansion in production of cash crops did not cause a net deficit in food production in the area and that the increase in output was accompanied by considerable capital investment, in the form of planting and clearing land for cocoa.94 Moreover, Jedwab and Moradi have shown that cocoa production was stimulated by infrastructural investment, particularly in railroads, thus indicating that central government and large-scale capital investment had a conducive role to play in this period of expansion.95 In sum, there is mounting evidence that tends to question that colonial Ghana was stuck in a low-level equilibrium. While a country study like this cannot form an explicit test of the causality suggested in the cross-country studies literature, it suggests further steps toward research that establishes whether the causal claims of institutions on economic growth is consistent with trajectories of economic and institutional change.
A new description of growth and development since the 19th century underlines the need for new analytics. In his study of poverty in Africa, Iliffe spoke of conjunctural and structural causes of poverty.96 Conjunctural causes lead to a sudden change in poverty because of specific events such as failed harvests and disease. Alternatively, structural changes in poverty were changes in the political economy, such as property rights regimes or institutions of taxation. The analysis of trends in economic growth, taxation, and living standards to identify the relative importance of these types of factors, and the relative importance of “policy” versus “shock,” will facilitate the understanding of African economic history. Moreover, the combined availability of longer time-series data on growth, taxes, and living standards will allow for better comparisons of the political economy of growth.
Both conceptually and in the source material, the new evidence is focusing on states and state records. This is particularly so for aggregate economic performance; chiefly economic growth, real wages, and trends in real per capita taxes. This means that other themes that are central in African history, such as agrarian change, religious institutions, and gender, are missing in this account. However, the information available in the state records across time and space varies, which could lead to weakness in a study based on those records.97 Alternatively, by using the state record as a kind of fingerprint of the state and its activities, the record becomes a lens through which we can gauge state activity and capacity.
The statistical record that has been collected, systematized, and synthesized provides a quantitative basis to challenge and shape existing narratives on African economic development from the 19th century into the 20th century. However, the shape of the statistical record tells a story. The history of counting and registering people is as old as documented history. It is a central part of a state’s effort to govern people and goes hand in hand with taxation. The word “statistics” derives from “state” and when we look at statistical records we are primarily seeing the historical footprint of states. We see what the state knew about itself and what it cared to find out. We also see a chosen projected image, or what the state wanted others to know about it. It is particularly important to clarify what kinds of questions the state records can answer and when other sources will have to be used. Future African economic history research should aim to expand the “new economic history” beyond the colonial period and the use of the colonial records.
Acemoglu, Daron, and James A. Robinson. Why Nations Fail: The Origins of Power, Prosperity, and Poverty. New York: Crown, 2012.Find this resource:
Austen, Ralph. African Economic History: International Development and External Dependency. Oxford: James Currey Press, 1987.Find this resource:
Austin, Gareth. Labour, Land and Capital in Ghana: From Slavery to Free Labour in Asante, 1807–1956. Rochester, NY: University of Rochester Press, 2005.Find this resource:
Austin, Gareth. “Reciprocal Comparison and African History: Tackling Conceptual Eurocentrism in the Study of Africa’s Economic Past.” African Studies Review 50, no. 3 (2007): 1–28.Find this resource:
Austin, Gareth. “The ‘Reversal of Fortune’ Thesis and the Compression of History: Perspectives from African and Comparative Economic History.” Journal of International Development 20 (2008): 996–1027.Find this resource:
Austin, Gareth, and Stephen Broadberry. “Introduction: The Renaissance of African Economic History.” Economic History Review 67, no. 4 (2014): 893–906.Find this resource:
Hopkins, Antony G. An Economic History of West Africa. London: Longman, 1973.Find this resource:
Hopkins, Antony G. “The New Economic History of Africa.” Journal of African History 50 (2009): 155–177.Find this resource:
Iliffe, John. The African Poor: A History. Cambridge, UK: Cambridge University Press, 1987.Find this resource:
Jerven, Morten. Africa: Why Economists Got It Wrong. London: Zed Books, 2015.Find this resource:
Jerven, Morten. The Wealth and Poverty of African States: Economic Growth, Living Standards and Taxation in Africa Since the Nineteenth Century. Cambridge, UK: Cambridge University Press, forthcoming.Find this resource:
Manning, Patrick. “The Prospects for African Economic History: Is Today Included in the Long Run?” African Studies Review 30 (1987): 49–62.Find this resource:
Frankema, Ewout. “Colonial Taxation and Government Spending in British Africa, 1880–1940: Maximizing Revenue or Minimizing Effort?” Explorations in Economic History 48, no. 1 (2011): 136–149.Find this resource:
Frankema, Ewout, and Morten Jerven. “Writing History Backwards or Sideways: Towards a Consensus on African Population, 1850–present.” Economic History Review 67, no. 4 (2014): 907–931.Find this resource:
Frankema, Ewout, and Marlous van Waijenburg. “African Real Wages in Asian Perspective, 1880–1940.” Working Paper No. 24, Centre for Global Economic History, Utrecht University, The Netherlands, 2011.Find this resource:
Frankema, Ewout, and Marlous van Waijenburg. “Structural Impediments to African Growth? New Evidence from Real Wages in British Africa, 1880–1965.” Journal of Economic History 72, no. 4 (2012): 895–926.Find this resource:
Moradi, Alexander. “Confronting Colonial Legacies: Lessons from Human Development in Ghana and Kenya, 1880–2000.” Journal of International Development 20, no. 8 (2008): 1109–1110.Find this resource:
Moradi, Alexander. “Towards an Objective Account of Nutritional Health in Colonial Kenya: A Study of Stature in African Army Recruits and Civilians, 1880–1980.” Journal of Economic History 69, no. 3 (2009): 719–754.Find this resource:
(2.) Morten Jerven, Gareth Austin, Erik Green, Chibuike Uche, Ewout Frankema, Johan Fourie, Joseph E. Inikori, Alexander Moradi, and Ellen Hillbom, “Moving Forward in African Economic History: Bridging the Gap between Methods and Sources” (working paper no. 1, African Economic History Network, 2012).
(4.) Jerven et al., “Moving Forward,” 5.
(5.) Jerven et al., “Moving Forward,” 7.
(6.) Frankema, Ewout, and Marlous van Waijenburg, “Structural Impediments to African Growth? New Evidence from Real Wages in British Africa, 1880–1965,” Journal of Economic History 72, no. 4 (2012): 895–926; Hopkins, “New Economic History.”
(7.) Hopkins, “New Economic History,” 156.
(8.) Daron Acemoglu, Simon H. Johnson, and James A. Robinson, “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution,” Quarterly Journal of Economics 117, no. 4 (2002): 1231–1294.
(9.) William Easterly and Ross Levine, “Africa's Growth Tragedy: Policies and Ethnic Divisions,” Quarterly Journal of Economics 112, no. 4 (1997): 1203–1250.
(10.) Hopkins, “New Economic History,” 168.
(11.) Hopkins, “New Economic History,” 166; and Acemoglu et al., “Reversal of Fortune.”
(12.) Hopkins, “New Economic History,” 168–170.
(13.) Gareth Austin, “The ‘Reversal of Fortune’ Thesis and the Compression of History: Perspectives from African and Comparative Economic History,” Journal of International Development 20, no. 8 (2008): 996–1027.
(14.) Hopkins, “New Economic History”; and Austin, “Reversal of Fortune Thesis,” 998.
(15.) Hopkins, “New Economic History,” 175.
(16.) Austin, “Reversal of Fortune Thesis,” 998, emphasis added.
(17.) Austin, “Reversal of Fortune Thesis,” 1019.
(18.) Austin, “Reversal of Fortune Thesis,” 997.
(19.) Hopkins, “New Economic History,” 177.
(20.) James Fenske, “The Causal History of Africa: A Response to Hopkins,” Economic History of Developing Regions 25, no. 2 (2010): 177–212.
(21.) Acemoglu et al., “Reversal of Fortune”; and Easterly and Levine, “Africa’s Growth Tragedy.”
(22.) Fenske, “Causal History of Africa,” 179.
(23.) Fenske, “Causal History of Africa,” 187.
(24.) Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past,” Economic History of Developing Regions 26, no. 2 (2011): 111–124; Jerven et al., “Moving Forward.”
(25.) Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past”; and Fenske, “Causal History of Africa.”
(26.) This argument was first made in Morten Jerven, “Growth, Stagnation or Retrogression? On the Accuracy of Economic Observations, Tanzania, 1961–2001,” Journal of African Economies 20, no. 3 (2011): 377–394; and then fully developed in Jerven, Africa: Why Economists Got It Wrong (London: Zed Books, 2015).
(27.) Daron Acemoglu and James A. Robinson, “Why Is Africa Poor?” Economic History of Developing Regions 25, no. 1 (2010): 21–50.
(28.) Morten Jerven, “The Quest for the African Dummy: Explaining African Post‐Colonial Economic Performance Revisited,” Journal of International Development 23, no. 2 (2011): 288–307.
(29.) David Wheeler, “Sources of Stagnation in Sub-Saharan Africa,” World Development 12, no. 1 (1984): 1–23, could be considered an early forerunner in this debate.
(30.) Nicholas Stern, “The Economics of Development: A Survey,” Economic Journal 99, no. 397 (1989): 600. The first versions of the Penn World Tables were published in the 1970s, see Irving B. Kravis, Alan W. Heston, and Robert Summers, “Real GDP Per Capita for More Than One Hundred Countries,” Economic Journal 88, no. 350 (1978): 215–242, but the mainstream use of this data set is dated to the 5.0 version, which was published in 1991 (see Robert Summers and Alan Heston, “The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 1950–1988,” Quarterly Journal of Economics 106, no. 2 (1991): 327–368.
(31.) Steven N. Durlauf, Paul A. Johnson, and Jonathan R. W. Temple, “Growth Econometrics,” in Handbook of Economic Growth, ed., P. Aghion and S. Durlauf (Amsterdam: Elsevier, 2005), 599.
(32.) Arguably, the frameworks were compromised. See Jerven, Africa: Why.
(33.) Robert J. Barro, “Economic Growth in a Cross Section of Countries,” Quarterly Journal of Economics 106, no. 2 (1991): 407–443; and Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past”.
(34.) Barro, “Economic Growth,” 437.
(35.) Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past”; see also Pierre Englebert, “Solving the Mystery of the African Dummy,” World Development 28, no. 10 (2000): 1821–1835; and Jerven, Africa: Why.
(36.) Durlauf et al., “Growth Econometrics,” 599. See Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past” and Africa: Why, for a review of the contributions that focus on explaining slow growth in African countries.
(37.) Lant Pritchett, “Patterns of Economic Growth: Hills, Plateaus, Mountains, and Plains” (working paper 1947, World Bank Policy Research Series, Washington, DC, 1998).
(38.) Charles Kenny and David Williams, “What Do We Know About Economic Growth? Or, Why Don’t We Know Very Much?” World Development 29, no. 1 (2001): 15.
(39.) Sambit Battacharyya, “Root Causes of African Underdevelopment,” Journal of African Economies 18, no. 5 (2009): 745–780; Nathan Nunn, “The Importance of History for Economic Development,” Annual Review of Economics 1, no. 1 (2009): 65–92. The seminal contributions for each strand are, respectively, David E. Bloom, Jeffrey D. Sachs, Paul Collier, and Christopher Udry, “Geography, Demography, and Economic Growth in Africa,” Brookings Papers on Economic Activity no. 2 (1998): 207–296; Nathan Nunn, “The Long-Term Effects of Africa’s Slave Trades,” Quarterly Journal of Economics 123, no. 1 (2008): 139–176; Daron Acemoglu, Simon H. Johnson, and James A. Robinson, “The Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review 91, no. 5 (2001): 1369–1401; and Acemoglu et al., “Reversal of Fortune.”
(40.) Dani Rodrik, Arvind Subramanian, and Francesco Trebbi, “Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development,” Journal of Economic Growth 9, no. 2 (2004): 131–165; and Jeffrey Sachs, “Institutions Don’t Rule: Direct Effects of Geography on Per Capita Income” (working paper w9490, National Bureau of Economic Research, Cambridge, Massachusetts, 2003).
(41.) Debraj Ray, “Uneven Growth: A Framework for Research in Development Economics,” Journal of Economic Perspectives 24, no. 3 (2010): 50.
(42.) Nunn, “Importance of History,” 31.
(43.) Ray, “Uneven Growth,” 56–57. See also, Acemoglu et al., “Colonial Origins” and “Reversal of Fortune”; and Kenneth Sokoloff and Stanley Engerman, “History Lessons: Institutions, Factors Endowments, and Paths of Development in the New World,” Journal of Economic Perspectives 14, no. 3 (2000): 217–232.
(44.) As in Melvin W. Reder, “The Tension between Strong History and Strong Economics,” in History Matters: Essays on Economic Growth, Technology, and Demographic Change, ed., Timothy W. Guinnane, William A. Sundstrom, and Warren Whatley (Stanford, CA: Stanford University Press), 96–113.
(45.) Robert H. Bates, Markets and States in Tropical Africa: The Political Basis of Agricultural Policies (Berkeley: University of California Press, 1981).
(46.) Platteau argues that the fundamental problem in postcolonial Africa has been the “fluid political setup dominated by unregulated factional competition as well as by the instability of ruling coalitions.” Jean-Philippe Platteau, “Institutional Obstacles to African Economic Development: State, Ethnicity, and Custom,” Journal of Economic Behaviour and Organization 71, no. 3 (2009): 676.
(48.) K. Pomeranz, The Great Divergence: China, Europe, and the Making of the Modern World Economy (Princeton, NJ: Princeton University Press), 8.
(49.) The virtues of “reciprocal comparison” are well laid out in Pomeranz, “Great Divergence,” and Gareth Austin, “Reciprocal Comparison and African History: Tackling Conceptual Eurocentrism in the Study of Africa’s Economic Past,” African Studies Review 50, no. 3 (2007): 1–28. For a version of the argument of endowments and choice of technology relating to rice production in Asia, see Francesca Bray, The Rice Economies: Technology and Development in Asian Societies (Oxford: Basil Blackwell, 1986).
(51.) Acemoglu and Robinson, Why Nations Fail, 61.
(52.) Acemoglu and Robinson, 90.
(53.) Gareth Austin, “Resources, Techniques, and Strategies South of the Sahara: Revising the Factor Endowments Perspective on African Economic Development, 1500–2000,” Economic History Review 61, no. 3 (2008): 587–624.
(54.) Acemoglu et al., “Colonial Origins.”
(55.) Acemoglu et al., “Reversal of Fortune” and “Colonial Origins.”
(56.) The exclusion restriction in their instrumental variable regression is that the mortality rates of European settlers more than one hundred years ago have no effect on GDP per capita today, other than their effect through institutional development.
(57.) For a review and general critique of the use of instrumental variables, see Angus Deaton, “Instruments, Randomization, and Learning about Development,” Journal of Economic Literature 48, no. 2 (2010): 424–455.
(58.) Bloom et al., “Geography.”
(59.) Acemoglu, et al., “Colonial Origins”; Daron Acemoglu, Simon H. Johnson, and James A. Robinson, “A Response to Albouy’s ‘A Reexamination Based on Improved Settler Mortality Data’” (unpublished manuscript, Massachusetts Institute of Technology, Cambridge, MA, 2005); Daron Acemoglu, Simon H. Johnson, and James A. Robinson, “Hither Thou Shalt Come, But No Further: Reply to ‘The Colonial Origins of Comparative Development: An Empirical Investigation: Comment’” (working paper w16955, National Bureau of Economic Research, Cambridge, Massachusetts, 2011); David Albouy, “The Colonial Origins of Comparative Development: A Reexamination Based on Improved Settler Mortality Data” (working paper, Department of Economics, University of California–Berkeley, 2004); David Albouy, “The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data” (working paper w14130, National Bureau of Economic Research, Cambridge, Massachusetts, 2008); and David Albouy, “The Colonial Origins of Comparative Development: An Empirical Investigation: Comment,” American Economic Review 102, no. 6 (2012): 3059–3076.
(60.) Deaton, “Instruments.”
(61.) For example, Bloom et al., among others, have argued that malaria has direct causal effects on income today. See Bloom et al., “Geography.”
(62.) Sachs, “Institutions Don’t Rule.”
(63.) Mark Maslin, Global Warming: A Very Short Introduction (New York: Oxford University Press, 2004), 27.
(64.) Morten Jerven, Poor Numbers: How We Are Misled by African Development Statistics and What to Do about It (Ithaca, NY: Cornell University Press, 2013), 8–32. Indeed, sometimes country data are made up before the countries exist. See, for instance, the controversy on settler mortality data in the debate between Albouy, “Colonial Origins,” and Acemoglu, Johnson, and Robinson, “Colonial Origins.”
(65.) See Austin, “Reversal of Fortune,” on Acemoglu et al., “Colonial Origins,” and “Reversal of Fortune.”
(66.) Elsa Artadi and Xavier Sala-i-Martin, “The Economic Tragedy of the XXth Century: Growth in Africa” (working paper w9865, National Bureau of Economic Research, Cambridge, Massachusetts, 2003).
(67.) Angus Maddison, “Historical Statistics of the World Economy: 1–2006 AD.” Groningen Growth and Development Centre, University of Groningen, The Netherlands, 2009.
(68.) Maddison, “Historical Statistics.”
(69.) Bloom et al., “Geography,” 2.
(70.) Hopkins, Economic History.
(71.) Austin, “Resources,” 588; Morten Jerven, “A Clash of Disciplines? Economists and Historians Approaching the African Past,” Economic History of Developing Regions 26, no. 2 (2011): 111–124.
(72.) Jan Vansina, “Knowledge and Perceptions of the African Past,” in African Historiographies: What History for Which Africa? ed., Bogumil Jewsiewicki and David Newbury (Beverly Hills, CA: SAGE, 1986), 40.
(73.) Bruce Berman, Control and Crisis in Colonial Kenya: The Dialectic of Domination (East African Publishers, 1992).
(74.) Either the effect of institutions “X” or as an outcome “Y,” as in, respectively, Dirk Bezemer, Jutta Bolt, and Robert Lensink, “Indigenous Slavery in Africa’s History: Conditions and Consequences” (World Economic History Congress, Utrecht, 2009) or James Fenske, “Does Land Abundance Explain African Institutions?” (working paper no. 981, Economic Growth Center, Yale University, New Haven, Connecticut, 2009), “Ecology, Trade, and States in Pre-Colonial Africa,” Journal of the European Economic Association 12, no. 3 (2014): 612–649, and “African Polygamy: Past and Present,” Journal of Development Economics 117 (2015): 58–73. In e-mail correspondence regarding the “Human Relations Area Files,” London School of Economics anthropology professor Chris Fuller described them as “a blast from the past” and noted that apart from Jack Goody the files were not highly rated among anthropologists in the UK. Joseph Tobin summarizes the critiques of the database made in the 1960s and 1970s and notes that “critiques have grown rarer recently, not, I think, because the heirs to Boas, Benedict, Leach, and Geertz have grown less antagonistic to quantification and comparison, but because, if they think of HRAF at all, they tend to think of it as moribund.” Joseph Tobin, “The HRAF as a Radical Text?” Cultural Anthropology 5, no. 4 (1990): 478.
(75.) Gareth Austin, “Poverty and Development in Sub-Saharan Africa, c1450–c1900: Reflections on the Development of Economic Historiography” (paper presented at the annual meeting of the European Historical Economics Society, Geneva, September 4, 2009).
(76.) See Bezemer et al., “Indigenous Slavery.” Similar concerns apply to the use of the dated and crude proxies for “social capital,” as in the Adelman and Morris data set in Jonathan Temple and Paul A. Johnson, “Social Capability and Economic Growth,” Quarterly Journal of Economics 113, no. 3 (1998): 965–990, and “ethnic fractionalization,” as with the index computed from Atlas Naradov Mira in Easterly and Levine,”Africa’s Growth Tragedy.” Fenske notes that the latter data set has been coded by Posner to capture “politically relevant groups” (Fenske, “Causal History,” 183). That may raise more problems than it solves when considering it as historical evidence. See David N. Posner, “Measuring Ethnic Fractionalization in Africa,” American Journal of Political Science 48, no. 4 (2004): 849–863.
(77.) Denis Cogneau and Yannick Dupraz, “Questionable Inference on the Power of Pre-Colonial Institutions in Africa (PSE working paper nos. 2014–25, 2014).
(78.) Austin, “Reversal of Fortune Thesis.”
(79.) Morten Jerven, “Random Growth in Africa? A Report on the Quality of the Growth Evidence in East-Central Africa, 1965–1995,” Journal of Development Studies 46, no. 2 (2010): 274–294; and “The Relativity of Poverty and Income: How Reliable Are African Economic Statistics?” African Affairs 109, no. 434 (2010): 77–96. For further case studies of the problems arising from the quality of data in economic growth analysis, see, on Botswana, Jerven, “Accounting for the African Growth Miracle: The Official Evidence, Botswana 1965–1995,” Journal of Southern African Studies 36, no. 1 (2010): 73–94; on Kenya, “Revisiting the Consensus on Kenyan Economic Growth, 1964–1995,” Journal of Eastern African Studies 5, no. 1 (2011): 2–23; and on Tanzania, “Clash of Disciplines.” For an attempt to historicize the African national income data, see Jerven, “Users and Producers of African Income: Measuring African Progress,” African Affairs 110, no. 439 (2011): 169–190.
(80.) Morten Jerven, “Accounting for the African Growth Miracle”; “African Growth Recurring: An Economic History Perspective on African Growth Episodes, 1690–2010,” Economic History of Developing Regions 25, no. 2 (2010): 127–154.
(81.) These scholars are endeavoring to create more robust and nuanced time-series data detailing historical levels of national accounts in Africa. These are detailed more thoroughly in Jerven et al., “Moving Forward,” but are introduced here. Space precludes any more than a cursory introduction of these pursuits.
(82.) Jörg Baten, “Height and Real Wages: An International Comparison,” Jahrbuch für Wirtschaftsgeschichte, no. 1 (2000): 17–32; Alexander Moradi, “Nutritional Status and Economic Development in Sub-Saharan Africa, 1950–1980,” Economics and Human Biology 8, no. 1 (2010): 16–29. Eltis of course pioneered anthropometry in African history through his work on the precolonial period. His articles analyze the heights of people freed by the British Navy from slave ships after 1807. See Eltis.
(83.) Philip D. Curtin, Economic Change in Pre-Colonial Africa: Senegambia in the Era of the Slave Trade (Madison: University of Wisconsin Press, 1975); Gareth Austin, Labour, Land and Capital in Ghana: From Slavery to Free Labour in Asante, 1807–1956 (Rochester, NY: University of Rochester Press, 2005).
(84.) Ewout Frankema and Marlous van Waijenburg, “Structural Impediments to African Growth? New Evidence from Real Wages in British Africa, 1880–1940,” Journal of Economic History 73, no. 4 (2012): 895–926.
(85.) Pim de Zwart, “South African Living Standards in Global Perspective, 1835–1910,” Economic History of Developing Regions 26, no. 1 (2011): 49–74; Sophia du Plessis and Stan du Plessis, “Happy in the Service of the Company: The Purchasing Power of VOC Salaries at the Cape in the 18th century,” Economic History of Developing Regions 27, no. 1 (2012): 125–149; and Johan Fourie, “The Remarkable Wealth of the Dutch Cape Colony: Measurements from Eighteenth-Century Probate Inventories,” Economic History Review 66, no. 2 (2013): 419–448.
(86.) Acemoglu et al., Why Nations Fail.
(87.) See, for example, Ewout Frankema, “Raising Revenue in the British Empire, 1870–1940: How ‘Extractive’ Were Colonial Taxes?” Journal of Global History, 5, no. 3 (2010): 447–477; “Colonial Taxation and Government Spending in British Africa, 1880–1940: Maximizing Revenue or Minimizing Effort?” Explorations in Economic History, 48, no. 1 (2011): 136–149.
(88.) Francisco A. Gallego and Robert Woodberry, Christian Missionaries and Education in Former African Colonies: How Competition Mattered (Oxford: Oxford University Press, 2010); Nathan Nunn, “Religious Conversion in Colonial Africa,” American Economic Review 100, no. 2 (2010): 147–152; and Ewout Frankema, “The Origins of Formal Education in sub-Saharan Africa: Was British Rule More Benign?” European Review of Economic History 16, no. 4 (2012): 335–355.
(89.) Ray, “Uneven Growth.”
(90.) Jerven et al., “Moving Forward,” 17.
(91.) See, for example, Tim Kelsall, David Booth, Diana Cammack, and Frederick Golooba-Mutebi, “Developmental Patrimonialism? Questioning the Orthodoxy on Political Governance and Economic Progress in Africa” (working paper no. 9, Africa Power and Politics, 2010).
(92.) Alexander Moradi, “Confronting Colonial Legacies: Lessons from Human Development in Ghana and Kenya, 1880–2000,” Journal of International Development 20, no. 8 (2008): 1107–1121; “Towards an Objective Account of Nutrition and Health in Colonial Kenya: A Study of Stature in African Army Recruits and Civilians, 1880–1980,” Journal of Economic History 69, no. 3 (2009): 719–754.
(93.) Frankema, Ewout, and Marlous van Waijenburg, “Structural Impediments to African Growth? New Evidence from Real Wages in British Africa, 1880–1965,” Journal of Economic History 72, no. 4 (2012): 895–926.
(94.) Gareth Austin, “Vent for Surplus or Productivity Breakthrough? The Ghanaian Cocoa Take‐off, c. 1890–1936,” Economic History Review 67, no. 4 (2014): 1035–1064.
(95.) Rémi Jedwab and Alexander Moradi, “Transportation Infrastructure and Development in Ghana” (working paper 2011–24, Paris School of Economics).
(97.) See Jerven, Poor Numbers.