Comparative Politics - kellogg.nd.edu

working paper

419April2017

THEDEVELOPMENTALLEGACIESOFTHEBIFURCATEDCOLONIALSTATE:STATISTICALEVIDENCEFROMSIXTY-SEVENBRITISH,FRENCH,ANDPORTUGUESECOLONIES

OLUKUNLEP.OWOLABI

The Kellogg Institute for International Studies

University of Notre Dame 130 Hesburgh Center for International Studies

Notre Dame, IN 46556-5677 Phone: 574/631-6580 Web: kellogg.nd.edu

The Kellogg Institute for International Studies at the University of Notre Dame has built an international reputation by bringing the best of interdisciplinary scholarly inquiry to bear on democratization, human development, and other research themes relevant to contemporary societies around the world. Together, more than 100 faculty and visiting fellows as well as both graduate and undergraduate students make up the Kellogg community of scholars. Founded in 1982, the Institute promotes research, provides students with exceptional educational opportunities, and builds linkages across campus and around the world. The Kellogg Working Paper Series:

n Shares work-in-progress in a timely way before final publication in scholarly books and journals n Includes peer-reviewed papers by visiting and faculty fellows of the Institute n Includes a Web database of texts and abstracts in English and Spanish or Portuguese n Is indexed chronologically, by region and by research theme, and by author n Most full manuscripts downloadable from kellogg.nd.edu

Contacts: Elizabeth Rankin, Editorial Manager [email protected]

THE DEVELOPMENTAL LEGACIES OF THE BIFURCATED COLONIAL STATE: STATISTICAL EVIDENCE FROM SIXTY-SEVEN BRITISH, FRENCH,

AND PORTUGUESE COLONIES

Olukunle P. Owolabi

Kellogg Institute for International Studies Working Paper #419 – April 2017

Olukunle P. Owolabi is assistant professor of political science at Villanova University. His research explores the long-term developmental and political consequences of colonialism, with a particular focus on Sub-Saharan Africa and the Caribbean region. His work has been published in Comparative Politics and the APSA’s Comparative Democratization newsletter. He is currently completing the book manuscript “The Colonial Origins of (Under)development, Authoritarianism and Democracy: Africa and the West Indies in Comparative-Historical Perspective.” After earning his PhD from the University of Notre Dame in 2012, Owolabi returned to the Kellogg Institute for International Studies as a visiting fellow during the spring 2016 semester.

ABSTRACT This paper examines the impact of bifurcated colonial institutions—i.e. the use of distinctive legal codes for “native” vs. “settler” populations—for long-term development in 67 former British, French, and Portuguese colonies. Building on theoretical arguments by Mahmood Mamdani (1996) and Matthew Lange’s empirical research on the distinctive developmental legacies of direct vs. indirect British rule, I develop a new measure of legal-administrative bifurcation in French and Portuguese colonies. Consistent with Mamdani’s theoretical arguments, the statistical models in this paper demonstrate that bifurcated colonial institutions contributed to poor development outcomes and ineffective postcolonial governance among British, French, and Portuguese colonies alike. Regardless of the colonizing power, directly ruled colonies with a uniform and inclusive legal-administrative framework have better development outcomes than bifurcated colonial states that maintained distinctive “native” legal codes for indigenous populations. These results are robust to a variety of statistical controls, as well as to instrumental variable analysis, highlighting the enduring legacy of colonial institutions for human well-being and governmental effectiveness today.

RESUMO Este trabalho examina o impacto de instituições coloniais bifurcadas—isto é: o uso de códigos legais distintos para populações “nativas” e “colonas”—para o desenvolvimento no longo prazo de 67 antigas colônias britânicas, francesas e portuguesas. A partir dos argumentos teóricos de Mahmood Mamdani (1996) e a pesquisa empírica de Matthew Lange sobre as diferentes heranças desenvolvimentistas entre o domínio direto e indireto britânico, desenvolvo uma nova medida de bifurcação das instituições coloniais francesas e portuguesas. Concordando com os argumentos teóricos de Mamdani, os modelos estatísticos neste trabalho demonstram que as instituições coloniais bifurcadas contribuíram para os pobres resultados no desenvolvimento e a governança ineficaz pós-colonial das colônias britânicas, francesas e portuguesas. Sem importar a potência colonizadora, colônias governadas diretamente com estruturas legais-administrativas uniformes e inclusivas têm melhores resultados no seu desenvolvimento do que estados coloniais bifurcados que mantiveram códigos legais “nativos” para a população indígena. O resultados são robustos para múltiplos controles estatísticos, assim como uma análise de variáveis instrumentais, mostrando assim os duradouros legados das instituições coloniais para o bem-estar humano e a efetividade governamental atual.

RÉSUMÉ

Cet article examine l’effet des institutions coloniales doubles—l’usage de codes légales différents pour des “autochtones” vs. des “colonnes”—sur le développement à long terme dans 67 anciennes colonies Britanniques, Françaises et Portugaises. En me basant sur les arguments théoriques de Mahmood Mamdani (1996) et la recherche empirique de Matthew Lange sur les héritages en matière de développement laissés par l’administration du Britannique direct vs. indirect, je propose une nouvelle méthode pour calibrer les institutions juridique-administrative dans des colonies Françaises et Portugaises. Conformément aux arguments théoriques de Mamdani, les modèles statistiques de cet article montrent que le double système d’institutions coloniales a contribué à faible développement et l’inefficacité de la gouvernance postcoloniale. N’importe le pouvoir colonisateur, les colonies sous l’administration direct avec un cadre juridique-administratif uniforme et incluant ont des meilleurs résultats en matière de développement que les états coloniaux qui ont maintiennent des codes légaux différenciés pour les populations natives. Ces résultats sont cohérents avec divers contrôles statistiques, même avec l’analyse des variables instrumentales, soulignant l’héritage durable des institutions coloniales pour le bien-être humain et l’efficacité jusqu’au nos jours.

Owolabi 1

INTRODUCTION

The past two decades have seen a resurgence of empirical scholarship examining the

developmental and political legacies of European colonialism in different parts of the world (see

Acemoglu, Johnson, and Robinson 2001, 2002; Easterly and Levine 2013; Lange 2004, 2009;

Lange et al. 2006; Mahoney 2010; La Porta et al. 1998, 1999). There is widespread agreement

among scholars that European settler colonialism generated favorable development outcomes

relative to “extractive” colonialism (see Krieckhaus 2006; Acemoglu, Johnson, and Robinson

2001, 2002; Sokoloff and Engerman 2000) and that British colonial and/or legal institutions

generated favorable developmental outcomes relative to continental European colonial and/or

legal institutions (see La Porta, Lopez-de-Silanes, and Shleifer 2008; La Porta et al. 1998, 1999;

Brown 2000). More recently, a number of influential studies have highlighted the important

contribution of Christian, and especially Protestant missionaries, in promoting mass education

and postcolonial democratization in many parts of the Global South (see Woodberry 2012;

Lankina and Getachew 2012, 2013; Gallego and Woodberry 2010). Notwithstanding these

general trends, there is a surprising amount of unexplained developmental variation across

former colonies with limited European settlement, and this is as much the case among French

and Portuguese colonies as it is among British colonies.

The data in Table 1 highlight the significant variation in human well-being across sixty-

seven former British, French, and Portuguese colonies that experienced at least thirty years of

colonial rule ending after 1945.1 This paper includes former colonies that never gained

independence (such as Hong Kong, Macau, and France’s overseas departments and territories) if

their population exceeded 100,000 at the end of the colonial era. Apart from Hong Kong, these

dependencies are generally excluded from existing literature on the colonial origins of

development. Nevertheless, their omission significantly understates the developmental variation

                                                                                                                         1 This data sample excludes countries that experienced less than thirty years of British, French, or Portuguese rule, such as Lebanon, Israel/Palestine, and Syria. It also excludes British protected states, such as Qatar, Bahrain, and the United Arab Emirates, whose political elites retained control over domestic affairs during the colonial era (see Owolabi 2015). The data sample also excludes micro-states with fewer than 100,000 people at the end of the colonial era.

Owolabi 2

across former French and Portuguese colonies relative to former British colonies.2 Human well-

being is measured using the United Nations’ Human Development Index (HDI), which accounts

for per capita income, educational attainment, and life expectancy in 2015.

The HDI scores in Table 1 reveal two interesting trends. First, there is considerable

variation in human well-being among former British, French, and Portuguese colonies with

predominantly nonwhite populations. Within each imperial system, the 2015 HDI scores range

from very high to low, as exemplified by the developmental extremes of Singapore (0.912) and

Sierra Leone (0.413) as former British colonies, Martinique (0.859) and Niger (0.348) as former

French colonies, and Macau (0.892) and Mozambique (0.416) as former Portuguese colonies.

Second, although the mean HDI score among former British colonies is somewhat higher than

among former French or Portuguese colonies, the developmental variation across the three

colonial powers is surprisingly similar. In fact, the standard deviation in 2015 HDI scores among

former French and Portuguese colonies (at 0.155 and 0.162 respectively) is slightly larger than

among former British colonies (0.144). To explain the significant developmental variation

among former British colonies, previous studies have examined the distinctive developmental

and political legacies of indirect vs. direct British rule (see Lange 2004, 2009). Indirect British

rule institutionalized a bifurcated legal-administrative framework in which rural indigenous

populations were subject to customary laws enforced by traditional “chiefs,” whereas the

bureaucratic institutions of the modern state were primarily limited to the colonial capital and

areas of European settlement. Although other colonial powers such as France and Portugal also

established bifurcated colonial states (see Lange 2009, 199–204; Mamdani 1996), the

developmental consequences of legal-administrative bifurcation have never been empirically

tested outside of the British colonial empire.

                                                                                                                         2 Most existing studies significantly underestimate the extent of developmental variation in the former French and Portuguese colonial empires, because they exclude many successful colonies that never gained independence. Prior to 1946, France’s “old colonies” (i.e., French Guiana, Guadeloupe, Martinique, and Réunion) had similar patterns of socio–economic development and more extensive political rights than many of the British Caribbean colonies that became independent after 1962 (see Murch 1968, 548–550, and 1971). According to the data in Table 1, these former French colonies continue to outperform their British equivalents (such as Barbados, Jamaica, or Mauritius) in terms of human well-being, yet they are generally excluded from most empirical studies because of their current political status as overseas French departments (since 1946). At the same time, many empirical studies include Hong Kong as a former British colony (e.g., Fails and Krieckhaus 2010; Krieckhaus 2006; Lange 2004, 2009; Acemoglu and Robinson 2001, 2002), while excluding Macau, which experienced similar development under Portuguese rule.

Owolabi 3

TABLE 1

HUMAN DEVELOPMENT IN SIXTY-SEVEN FORMER BRITISH, FRENCH, AND PORTUGUESE COLONIES

HDI Scores (2015) British Colonies French Colonies Portuguese Colonies Very High (>0.800) [N=8]

Brunei, Hong Kong, Singapore

French Guiana, Guadeloupe, Martinique, Réunion

Macau

High (0.700–0.799) [N=12]

Bahamas, Barbados, Belize, Fiji, Jamaica, Malaysia, Mauritius, Sri Lanka, Trinidad and Tobago

Algeria, Tunisia, French Polynesia

Medium (0.550–0.699) [N=15]

Bangladesh, Botswana, Ghana, Guyana, India, Zambia

Cambodia, Congo-Brazzaville, Gabon, Laos, Morocco, Vietnam

Cape Verde, São Tomé and Principe, Timor-Leste

Low (<0.550) [N=32]

Gambia, Kenya, Lesotho, Malawi, Myanmar, Nigeria, Pakistan, Sierra Leone, Solomon Islands, Sudan, Swaziland, Tanzania, Uganda, Zimbabwe

Benin, Burkina Faso, Cameroon, Central African Republic, Chad, Comoros, Cote D’Ivoire, Djibouti, Guinea, Madagascar, Mali, Mauritania, Niger, Senegal, Togo

Angola, Guinea-Bissau, Mozambique

Mean HDI Score [Standard Deviation]

0.629 [0.144]

0.572 [0.155]

0.579 [0.162]

Min. 0.413 (Sierra Leone) 0.348 (Niger) 0.416 (Mozambique) Max. 0.912 (Singapore) 0.859 (Martinique) 0.892 (Macau) N=67 32 28 7

Dependent territories in italics.

Source: HDI scores from Wikipedia (2016). “List of Countries by Human Development Index,” https://en.wikipedia.org/wiki/List_of_countries_by_Human_Development_Index .

This paper seeks to fill this gap by developing a new measure of legal-administrative

bifurcation in French and Portuguese colonies. Although British indirect rule devolved more

local political autonomy to traditional chiefs than was commonplace in French or Portuguese

colonies, all three powers developed bifurcated legal-administrative institutions that restricted

the legal rights of rural indigenous subjects relative to colonial settlers. I argue that bifurcated

legal-administrative institutions undermined long-term development and postcolonial

Owolabi 4

governance by creating colonial states with limited bureaucratic capacity and limited

inclusiveness and by impeding the development of a uniform and inclusive legal system that

protects individual civil liberties. Because bifurcated colonial states lacked the bureaucratic

capacity to provide public goods such as education, sanitation, and a uniform legal system that

protected individual civil liberties (Lange 2009), they generally struggled to democratize after

independence (Owolabi 2015; Lange 2009; Mamdani 1996). Consequently, I expect bifurcated

colonial states to have worse developmental outcomes and less effective governance today

relative to states that established a bureaucratic and inclusive legal-administrative framework

during the colonial era.

After exploring the distinctive ways in which the British, French, and Portuguese created

bifurcated colonial states, this paper develops a new indicator that measures the extent to which

indigenous populations in French and Portuguese colonies were governed under distinctive

“native” legal codes prior to the Second World War. In the statistical models that follow,

bifurcated colonial legal institutions—whether British, French, or Portuguese—predict lower

rates of primary school enrollment in 1960, worse governmental performance (1996–2012), and

lower HDI scores in 2015. These results are robust to statistical controls for colonial (i.e.,

European and Asian) settlement, ethnic diversity, geographic region, religion and missionary

presence. They are also robust to instrumental variable analysis. Furthermore, these empirical

results are consistent in separate samples of British and non-British colonies, as well as in the

combined sample of sixty-seven former British, French, and Portuguese colonies. Consequently,

these results demonstrate that Mamdani’s (1996) theoretical arguments regarding the negative

developmental consequences of the bifurcated colonial state can be generalized beyond the

British colonial empire.

THE BIFURCATED COLONIAL STATE, INEFFECTIVE GOVERNANCE,

AND UNDERDEVELOPMENT

In his detailed ethnographic study of the African colonial state, Mahmood Mamdani (1996) notes

that the bureaucratic apparatus of the colonial state was primarily limited to the administrative

capital and areas of European settlement, while rural, indigenous populations were governed

under distinctive “customary” laws. In general, urban and settler populations were subject to

European laws, which protected their property rights and individual civil liberties, while

Owolabi 5

provincial district officials oversaw a network of local intermediaries (known as “chiefs” in

British colonies, chefs de canton in French colonies, and régulos in Portuguese colonies) who

enforced “customary” laws that applied only to indigenous subject populations. In rural,

indigenous districts, local state authority operated along highly patrimonial lines, as local chiefs

combined legislative, executive, judicial, and administrative authority and were responsible only

to the local district administrator (Mamdani 1996, 56–61). Mamdani emphasizes the bifurcated

nature of this authority structure, noting that “customary justice was dispensed to native

[subjects] by chiefs and commissioners,” whereas European laws generally protected the civil

liberties of non-natives (Mamdani 1996, 109).

Because indigenous intermediaries were responsible only to the colonial authorities who

appointed them, chiefs often abused their powers for self-enrichment, through forced labor (on

their private properties) or forced contributions. Furthermore, the definition of “custom” often

changed over time to suit the interests of chiefs and the needs of the colonial state. Often, new

laws regarding taxation, forced labor, agricultural cultivation, and restrictions on the free

movement of peoples, had little to do with traditional customs. In fact, chiefs often manipulated

the so-called “customary laws” for extortion and self-aggrandizement (Mamdani 1996, 122–

125). Consequently, Mamdani refers to this patrimonial authority structure as “decentralized

despotism,” given the bureaucratic weakness and limited capacity of the modern state (Mamdani

1996, 56–61).

Mamdani (1996) argues that legal-administrative bifurcation generated poorly integrated

states with limited bureaucratic capacity and inclusiveness, which undermined their long-term

development. He also argues that the bifurcated colonial state institutionalized a highly

patrimonial authority structure that has persisted in many African countries long after

independence. Despite the important theoretical contributions of Mamdani’s (1996) work, the

empirical test of his argument is primarily limited to a detailed longitudinal analysis of urban vs.

rural development outcomes in South Africa and Uganda.

Legal-Administrative Bifurcation in the British Colonial Empire To date, the most rigorous test of Mamdani’s theoretical framework has been Matthew Lange’s

empirical research on the distinctive developmental and political consequences of direct vs.

indirect British rule (Lange 2004, 2009). Lange measures the extent of indirect rule in each

Owolabi 6

British colony on the basis of the percentage of legal cases heard in “customary courts”

controlled by traditional “chiefs” c. 1955. As such, the extent of indirect rule ranged from over

80% in Malawi, Nigeria, Sierra Leone, and Uganda, to approximately 50% on the Indian

subcontinent, to 0% in directly ruled British colonies such as Barbados, Hong Kong, Jamaica,

Mauritius, and Singapore (Lange 2009, 48). In directly ruled colonies, by contrast, Europeans

constructed “centralized, territory-wide, and bureaucratic legal-administrative institutions” that

resembled the type of state domination that emerged in early modern Europe (Lange 2009, 4).

Directly ruled British colonies also developed strong legal protections for individual civil

liberties and property rights based on British common law.

Lange argues that indirect British rule institutionalized poorly integrated states with

limited bureaucratic capacity, limited infrastructural power, and limited inclusiveness (see Lange

2009, 47–49). A high degree of indirect rule also suggests that British common law, with its

strong protections of individual civil liberties and property rights, was only weakly

institutionalized during the colonial era, which likely had adverse consequences for long-term

development and postcolonial governance. Perhaps not surprisingly then, Matthew Lange’s

statistical models demonstrate that the extent of indirect rule consistently predicts worse

development outcomes (in terms of per capita GDP, educational attainment, and infant mortality

rates), less effective postcolonial governance (1996–2005), and lower mean democracy scores

(1972–2005) than in directly ruled British colonies, which developed more effective bureaucratic

and inclusive state structures (Lange 2009).

Legal-Administrative Bifurcation in French and Portuguese Colonies

French and Portuguese colonial administration was considerably more rigid and centralized than

Britain’s and generally less tolerant of traditional authority structures. Consequently, many

scholars regard French and Portuguese colonial administration as synonymous with “direct” or

assimilationist rule (see Bernhard, Reenock, and Nordstrom 2004, 231; Njoh 2000; Young 1994,

Ch. 4; R. Collier 1982, 81–87). Yet, despite their republican bias toward administrative

centralization and cultural assimilation, France and Portugal also developed bifurcated legal-

administrative structures in colonies with significant indigenous populations. The data in Tables

A2 and A3 clearly demonstrate that only a handful of French and Portuguese colonies developed

an inclusive and integrated legal-administrative framework that extended citizenship rights to the

Owolabi 7

majority of the population (see Appendix). Consequently, I avoid the standard practice of

equating British administration with “indirect rule” and French and Portuguese administration

with “direct” or assimilationist rule. In each colonial empire, there were directly ruled colonies

with a uniform and inclusive legal-administrative framework, as well as bifurcated colonies that

maintained a distinctive legal-administrative framework for rural, indigenous populations.

Direct rule was only fully implemented in the “old” French colonies of Guiana,

Guadeloupe, Martinique, and Réunion, where a uniform legal-administrative structure and

inclusive citizenship rights were established following the abolition of slavery in 1848 (France

1944; Murch 1971). This had profound consequences for the extension of civil liberties and

political rights, as well as the expansion of public schooling after 1880. In fact, the “old” French

colonies have had elected representatives to their local assemblies and the French national

parliament since the introduction of universal male suffrage in the 1870s, and they also benefited

from the Jules Ferry laws, which made secular primary education free and compulsory for all

French citizens (Murch 1971, 32–36). Beginning with Algeria, however, all remaining French

colonies developed bifurcated legal-administrative institutions that denied citizenship rights to

their indigenous subject populations. The French indigénat legal code was first implemented in

Algeria during the 1870s but later extended to West and Equatorial Africa, Indochina, and the

Pacific islands during the final decades of the nineteenth century (Merle 2002).3 In Morocco and

Tunisia, Arab Muslims were under the authority of their traditional indigenous rulers and lacked

civil and political equality with metropolitan French citizens (Lawrence 2009, 28; France 1944).

Portugal also established distinctive “native” labor and criminal codes for indigenous subject

populations in Africa and Timor. These were officially codified as the regime do indigenato in

1929 (Duffy 1961).4 By contrast, Portuguese citizenship rights were extended to the entire

populations of Cape Verde, Macau, and Portuguese India (i.e., the Indian state of Goa) during the

nineteenth century, and to São Tomé and Principe after 1954 (Duffy 1961; Portugal 1960). Like

Britain, France and Portugal generally established bifurcated legal institutions in colonies with

significant indigenous populations and where the territorial control of the colonial state remained

limited until after 1880.

                                                                                                                         3 Mann (2009) also offers a critical evaluation of the French indigénat legal code in West Africa. 4 For a comprehensive overview of the Portuguese indigenato legal code, see Ferreira (1974, Ch. 2) and O’Laughlin (2000).

Owolabi 8

The French indigénat and Portuguese indigenato legal codes empowered local officials to

impose summary justice on indigenous populations, who were not protected by French or

Portuguese civil law. French and Portuguese officials appointed “administrative chiefs” to collect

taxes and forcibly recruit labor for public works and even private enterprises. Whereas British

indirect rule often supported traditional authority structures, French and Portuguese officials

frequently appointed indigenous soldiers, policemen, or interpreters to serve as “administrative

chiefs.” These administrative appointees often enforced the most repressive aspects of the

colonial state (i.e., tax collection, labor recruitment, and “summary justice,” which relied heavily

on fines and corporal punishment) in rural communities (see Mamdani 1996, 125–128;

O’Laughlin 2000; Mann 2009). Despite the differences in British vs. continental European

approaches toward indigenous administration (see Ajayi and Crowder 1971; Crowder 1964),

French and Portuguese “native” codes also created bifurcated colonial states with limited

bureaucratic capacity and where indigenous subjects were not protected by European civil laws

and citizenship rights. This had profound consequences for the protection of property rights,

individual civil liberties, and access to public services such as education. Consequently, this

paper explores whether the development of bifurcated colonial states and the extensive use of

“native” legal codes in French and Portuguese colonies had similar negative consequences for

long-term development and postcolonial governance to those of indirect rule in former British

colonies.

THE LONG-TERM DEVELOPMENTAL LEGACIES OF

BIFURCATED COLONIAL STATES

Building on previous arguments outlined by Lange (2009), Mamdani (1996), Acemoglu and

Robinson (2012), and others, I expect bifurcated colonial institutions to generate worse

developmental outcomes than those in directly ruled colonies with a uniform legal-administrative

framework based on British common law or French/Portuguese civil law. I outline three distinct

causal mechanisms linking the bifurcated colonial state to poor developmental outcomes: limited

bureaucratic capacity, limited inclusiveness, and weak political legitimacy/rule of law. First,

existing studies suggest that governments are rarely able to efficiently provide public goods and

services such as roads, schools, healthcare facilities, and sanitation without a capable

bureaucratic state apparatus that can implement policies throughout the national territory (see

Owolabi 9

Lange 2009; Krieckhaus 2006; Kohli 2004). The bureaucratic-administrative capacity of most

bifurcated colonial states remained extremely limited outside of the principal urban centers, and

many postcolonial governments maintained a strong pro-urban bias in their developmental

policies while extracting tax revenues and profits from small rural producers. This practice was

particularly widespread in Sub-Saharan Africa (see Bates 1981), where state policies have tended

to reinforce the urban-rural divide established during the colonial era.

Second, Acemoglu and Robinson (2012) argue that inclusive economic and political

institutions promote better development outcomes than extractive institutions that limit

individual liberties and property rights. The legal-administrative institutions of the bifurcated

colonial states were anything but inclusive, as they generally denied citizenship rights to rural

indigenous populations, who were legally regarded as native “subjects” (see Mamdani 1996).

Consequently, Matthew Lange argues that indirect [British] rule “obstructed inclusiveness to a

greater extent than direct [British] rule” by empowering local elites to serve as “gatekeepers”

between the state bureaucracy and the general public (Lange 2009, 39). The legal recognition of

“customary laws” also undermined the institutionalization of British common law, which is

known for its strong protections of property rights and individual civil liberties (La Porta et al.

1998, 1999). Similarly, bifurcated French and Portuguese colonies had highly restrictive

citizenship laws that excluded the vast majority of the population, thereby denying their

protection under French or Portuguese civil law (see Appendix, Tables A2 and A3). In the

French and Portuguese colonial empires, entry to public schools was mainly limited to

metropolitan citizens prior to 1946 and 1961 respectively, resulting in vast differences in

educational attainment. School enrollment was significantly higher in colonies with inclusive

citizenship rights. On the eve of the Second World War, for example, 48% of children in

Guadeloupe, Martinique, and Réunion were enrolled in primary schools, compared with just

2.0% in French West Africa and 2.4% in French Equatorial Africa (Benavot and Riddle 1988).5

And at the end of the Second World War, there were more children enrolled in primary schools

in Martinique, Guadeloupe, and Réunion than in all of French West Africa, where the population

was twenty times greater (France 1949). And because human capital is a central component of

socioeconomic development, the stark differences in educational attainment during the colonial

era likely affected subsequent patterns of socioeconomic development and political governance.                                                                                                                          5 These estimates are based on population-weighted data from France (1944).

Owolabi 10

Third, the distinctive, and often highly repressive “native” legal codes in bifurcated

colonial states likely had lasting negative consequences for postcolonial governance due to the

fact that the rule of law was only weakly institutionalized during the colonial era. Bifurcated

colonial states did not develop a uniform legal code during the colonial era, and most of them

failed to democratize after independence (Owolabi 2015; Mamdani 1996). The impartial and

uniform enforcement of the rule of law is an important component of effective governance

(O’Donnell 1993), which is essential for successful long-term development (Lange 2009, 36;

Kohli 2004; Englebert 2000). In the absence of an effective and consolidated state, political

leaders often pursue policies that enhance their political power rather than promote effective

broad-based development (see Englebert 2000; Bates 1981). Due to these corrosive effects on

political legitimacy and governance, I expect bifurcated colonial states to continue to suffer from

significant developmental deficits over the longue durée.

EMPIRICAL STRATEGY AND DATA ANALYSIS

The following section examines the developmental consequences of bifurcated colonial legal-

administrative institutions for educational attainment at the end of the colonial era (c. 1960),

postcolonial governance (1996–2012), and overall human well-being in 2015. The data sample

consists of sixty-seven British, French, and Portuguese colonies that experienced at least thirty

years of European rule ending after 1945, where European settlers made up less than one-quarter

of the population, and whose population exceeded 100,000 inhabitants at the end of the colonial

era. In contrast to most existing studies, I include former colonies that never gained political

independence, as long as they meet the conditions outlined above. The small number of cases

limits the number of statistical controls in each model, although some models control for

competing explanations of long-term development, including colonial settlers, ethnic

fractionalization, religious composition and/or missionaries, and geographic region. Other

influential statistical analyses of the long-term developmental consequences of colonialism—

including Acemoglu, Johnson, and Robinson 2001; Lange 2004, 2009; Kenny 2013; and

Frankema 2012—are also based on small samples of thirty-two to sixty-four countries.6

                                                                                                                         6 Matthew Lange’s statistical analysis of the developmental and political consequences of direct vs. indirect British rule is based on a small sample of thirty-three to thirty-nine former British colonies (Lange 2009). Paul Kenny (2013) tests his hypothesis about the impact of colonial state-building and

Owolabi 11

Measuring Legal-Administrative Bifurcation in British, French, and Portuguese Colonies Given the distinct ways in which different European colonizers interacted with pre-existing

authority structures in bifurcated colonies, I use different indicators to measure the extent of

legal-administration bifurcation in British vs. French and Portuguese colonies. Because British

administrative policy tended to empower traditional chiefs, I use Matthew Lange’s indirect rule

indicator to measure the extent of legal-administrative bifurcation in former British colonies.

Lange (2009) operationalizes indirect rule as the percentage of total legal cases that were heard

in “customary courts” maintained by traditional “chiefs” during the mid-1950s. The extent of

indirect rule ranged from 0% in directly ruled colonies such as Barbados, Hong Kong, Mauritius,

and Trinidad, to 50% on the Indian subcontinent, to more than 80% in Malawi, Nigeria, Sierra

Leone, and Uganda (see Appendix, Table A1).

Comparable data on “customary courts” are not available for French and Portuguese

colonies, where colonial authorities developed their own “native” legal codes that denied

citizenship rights to indigenous subject populations. Because the bureaucratic capacity of the

bifurcated colonial state was primarily limited to urban areas, I assume that the French and

Portuguese “native” codes were primarily enforced in rural areas. Consequently, I calculate the

extent of legal-administrative bifurcation in French and Portuguese colonies by multiplying the

percentage of the population with indigenous legal status c. 1945 by the percentage of the

population living in rural areas in 1955.7 This creates an index that ranges from 0% in directly

ruled colonies with inclusive citizenship rights (such as Cape Verde, Macau, and the “old French

colonies”), to 60–70% in North African colonies, to more than 80% in French Indochina and

many parts of Sub-Saharan Africa (see Appendix, Tables A2 and A3). This suggests that French

                                                                                                                                                                                                                                                                                                                                                                                                       decolonization on current perceptions of corruption using a sample of thirty-two former British colonies. And Ewout Frankema’s (2012) statistical analysis of the development of colonial education is based on a small sample of forty-two countries in Sub-Saharan Africa. Although larger samples enable a more rigorous test of empirical patterns with more statistical controls, each of the studies above nonetheless demonstrates clear and consistent patterns, despite their limited sample size. Even Acemoglu, Johnson, and Robinson’s (2001) path-breaking study on colonial settler mortality, institutions, and economic development only uses a sample of sixty-four countries, and the settler mortality rates are actually imputed for more than half of the countries in the sample (see Albouy 2012). 7 Due to the availability of colonial census data, I use French census data for 1945 (see France 1949) and Portuguese data for 1950 (see Portugal 1960). The percentage of the population living in rural areas in 1955 is obtained from the United Nations’ World Population Prospects, http://data.un.org.

Owolabi 12

and Portuguese colonies with a bureaucratic legal-administrative structure and inclusive

citizenship rights (such as Guadeloupe, Martinique, or Macau) are the functional equivalent of

directly ruled British colonies such as Barbados, Hong Kong, or Singapore, which developed

strongly bureaucratic states with an inclusive legal-administrative framework based on British

common law. At the same time, British colonies that relied heavily on indirect rule (such as

Nigeria or Sierra Leone) are functionally equivalent to French or Portuguese colonies that

enforced distinctive “native” legal codes on rural, indigenous populations (such as Benin or

Guinea-Bissau).

Dependent Variables This paper examines the impact of both measures of legal-administrative bifurcation on three

distinct developmental outcomes in thirty-two former British colonies, followed by thirty-five

French and Portuguese colonies. First, I examine primary school enrollment rates in 1960, as a

measure of bureaucratic capacity and human capital at the end of the colonial era. Since the end

of the Second World War, virtually all political leaders have regarded public education as one of

the most important goals of the modern state, as public schools are important for building a

national political community and strengthening citizens’ loyalty to the state (see Ulsaner and

Rothstein 2016, 233–235; Ramirez and Boli 1987; Benavot and Riddle, 1988). Yet, most

colonial states did not promote mass education unless they had already developed inclusive

citizenship rights and/or legal systems prior to the Second World War. Consequently, the

colonies with the highest literacy rates in 1960 were generally those that developed a

bureaucratic state with inclusive legal rights, such as directly ruled British colonies and French

or Portuguese colonies with inclusive citizenship rights (see Owolabi 2015). Consequently, I

expect indirect British rule, or the use of “native” legal codes in French or Portuguese colonies to

predict lower rates of primary school enrollment in 1960.

Postcolonial governance (1996–2012) is measured by averaging the World Bank’s (2016)

data on governmental effectiveness, the rule of law, voice and accountability, political stability,

and the control of corruption. These governance indicators have a global mean of 0, with most

countries ranging from -2.5 (extremely ineffective governance) and +2.5 (extremely effective

Owolabi 13

governance).8 Because bifurcated colonial states enforced extractive and repressive legal systems

for indigenous subject populations, I expect both measures of legal-administrative bifurcation—

i.e., British indirect rule and French/Portuguese native codes—to predict worse governmental

performance between 1996 and 2012.

Finally, I use the United Nations’ Human Development Index (HDI) scores as a broad

measure of human well-being in 2015. For ease of interpretation, HDI scores are multiplied by

100, with higher scores demonstrating higher levels of human well-being. Once again, I expect

bifurcated legal-administrative institutions to predict worse developmental outcomes among

British and non-British colonies alike.

Control Variables

The regression models control for various contravening factors that might also affect long-term

development and/or postcolonial governance. These include the presence of colonial settlers,

ethnic fractionalization, religious composition or the historical presence of Protestant

missionaries, and geographic region. Existing studies demonstrate that European settlement

contributed significantly to long-term development, although scholars debate whether this is

primarily due to the political and economic institutions demanded by settler populations (see

Acemoglu, Johnson, and Robinson 2001; Acemoglu, Gallego, and Robinson 2014; Easterly and

Levine 2013) or the human capital that settlers brought with them (Engerman, Mariscal, and

Sokoloff 2009). Asian settlers may have also contributed to successful long-term development,

given that they were generally governed under bureaucratic Western institutions rather than

“customary law” (Mamdani 1996). Furthermore, Asian settlers and labor migrants were often

attracted by economic opportunities in various colonies. By contrast, many scholars have argued

that high levels of ethnic fractionalization have adverse effects for democratic stability (Horowitz

1985) and the provision of public goods (Habyarimana et al. 2007; P. J. Collier 2000; Easterly

and Levine 1997). Consequently, some of the statistical models control for ethnic

fractionalization in each colony.

Recent empirical scholarship has also examined the developmental consequences of

colonial-era missionary activity, as Christian missionaries often established denominational                                                                                                                          8 The World Bank governance indicators do not cover French Polynesia and Guadeloupe, which are excluded from the data results in Table 6.

Owolabi 14

schools, which facilitated the development of civil society and, in some cases, postcolonial

democracy. Previous studies suggest that Protestant missionaries brought more developmental

benefits than Catholic missionaries, given their greater emphasis on education, mass printing,

and promoting bible literacy in indigenous languages. Furthermore, Protestant missionaries often

exercised greater independence from colonial state authorities than their Catholic counterparts

(see Woodberry 2012; Gallego and Woodberry 2010). In general, Protestantism is associated

with positive development outcomes (see Grier 1997; Bollen 1979), whereas Islam may have

hindered long-term development by slowing the spread of Western education (Frankema 2012).

Furthermore, because Islamic law does not recognize the separation of church and state and

prohibits certain capitalist economic practices, this may also impede socioeconomic development

and postcolonial democracy (see Anckar 2011; Hadenius 1992; Huntington 1991). Consequently,

many statistical models control for the percentage of Muslims or Protestants in each colony, or

the historical presence of Protestant missionaries.

Some statistical models also include regional “dummies” for countries located in the

West Indies or Sub-Saharan Africa to measure the impact of unmeasured factors that might have

shaped the developmental outcomes in each region (e.g., small island states in the West Indies;

landlocked states, low population densities, or lack of access to navigable rivers in Sub-Saharan

Africa). This is important because many existing studies demonstrate unexplained negative

developmental outcomes in Sub-Saharan Africa,9 whereas Caribbean states and territories have

surprisingly high levels of educational attainment and postcolonial democracy (see Benavot and

Riddle 1988; Dominguez et al. 1993; Edie 1994). Other models control for Portuguese

colonization, which may have had worse developmental consequences than French or British

colonization, given Portugal’s unique status as a relatively impoverished, semi-peripheral

colonial power with a repressive authoritarian regime from 1926 until 1974. Finally, some

models control for “dependent territories”—i.e., former colonies that never gained independence

after the Second World War. After all, it is reasonable to expect that Europeans would be most

likely to maintain their most “successful” colonies as overseas dependencies (e.g., Hong Kong,                                                                                                                          9 Including a dummy variable for Sub-Saharan African countries is extremely common in large-N statistical models (e.g. Easterly and Levine 1997; Lange 2004; Owolabi 2015), as many empirical studies demonstrate poor development outcomes in this region, even after controlling for colonial legacies, ethnic fractionalization, and other factors. The Africa dummy is typically insignificant, however, after controlling for environmental factors (see Sachs and Warner 1997) or measures of state legitimacy (see Englebert 2000).

Owolabi 15

Macau, and the French Caribbean and Pacific territories), so I expect to find superior

developmental and political outcomes in these territories. Summary statistics for the dependent

variables, focal independent variables, and control variables are found in Table 2.

Diagnostic Tests

For each statistical model, I inspected the bivariate correlations and variance inflation factors of

the independent variables, to ensure that models do not suffer from high levels of collinearity.10

Almost all of the regression models have homoskedatistic error terms, and although a few

models have one or two outlying or influential observations, I re-estimated each model using

robust regression (Stata’s rreg command) and with robust standard errors. The results are broadly

consistent, so only the OLS (or 2SLS) results are presented here. I also ran various diagnostic

tests to check for over-identification, under-identification, and weak instruments in the 2SLS

models. All of these tests can be rejected at conventional levels of statistical significance.

EMPIRICAL RESULTS

Colonial Education

The regression models in Table 3 support the hypothesis that bifurcated colonial institutions are

associated with lower rates of primary school enrollment at the end of the colonial era. Among

thirty-two former British colonies (Models 1–5), the indirect rule variable has the expected

negative sign and is statistically significant at p<0.05 in every model.11

                                                                                                                         10 None of the variance inflation factors exceed 4.5, demonstrating high levels of tolerance for the independent variables. 11 Woodberry’s (2012) missionary data do not cover dependent territories, so these are omitted from Models 4 and 9. Similarly, Maddison’s (2006) GDP data do not cover the Solomon Islands (omitted from Model 5), French Polynesia or Timor-Leste (missing from Model 10).

Owolabi 16

TABLE 2

SUMMARY STATISTICS

Dependent Variables N Mean Standard Deviation

Min. Max.

Net Primary Enrollment Rate (1960) 67 41.66 26.34 4 100 Mean Governance (1996–2012) 65 -0.35 0.73 -1.56 1.40 HDI Score (2015) 67 60 15.23 34.80 91.20 Focal Independent Variables % British Indirect Rule 32 36.44 31.75 0 93 % French/Portuguese Native Code 35 68.06 34.19 0 96 % Bifurcated 67 52.96 36.45 0 96 Key Control Variables Ethnic Fractionalization 67 0.55 0.27 0 0.93 % European Settlers 67 2.17 3.78 0.1 25 % Asian Settlers 67 6.46 16.90 0 81.2 % Catholic (1900) 67 10.44 24.20 0 99.8 % Protestant (1900) 67 5.67 15.64 0 81.2 % Christian (1900) 67 20.76 34.69 0 100 % Muslim (1900) 67 20.86 31.89 0 99.9 % Protestant (2000) 67 10.21 12.31 0 54.5 % Muslim (2000) 67 28.42 35.32 0 99.1 Protestant Missionaries per 10,000 pop. (1923)

59 0.91 1.60 0 9.05

Years Exposure to Protestant Missions (1960)

59 95.52 56.77 0 254

Logged per capita GDP (1950) 64 6.86 0.74 5.67 9.29 Logged per capita GDP (1950) 64 7.58 1.04 5.96 10.07 West Indies 67 0.13 0.34 0 1 Sub-Saharan Africa 67 0.57 0.50 0 1 Excluded Instruments: Colonial Occupation 67 0.70 0.46 0 1 Years since Colonization 67 150.1 128.1 38 500

Logged Population (1900) 67 6.46 1.92 2.94 12.35

The results for former French and Portuguese colonies are strikingly similar. In Models

6–10, the use of “native” legal codes also depresses 1960 primary school enrollment rates in

French and Portuguese colonies, significant at p<0.05 or p<0.01. The similarities between the

two samples of colonies are striking, as this represents a stark contrast from previous empirical

studies that demonstrate superior levels of education and literacy in former British colonies

Owolabi 17

relative to non-British colonies (see Brown 2000; Grier 1997). Nevertheless, in both samples

(i.e., Models 1 and 6) the extent of legal-administrative bifurcation explains 45% of the variation

in 1960 primary enrollment rates, and the magnitude of this coefficient is virtually identical. In

the absence of legal-administrative bifurcation, these models would predict 70% primary

enrollment in a directly ruled British colony, compared with 68% in a directly ruled French or

Portuguese colony. Nevertheless, given that the median extent of legal-administrative bifurcation

was 60% in British African colonies and 88% for French and Portuguese African colonies, this

would drop the predicted primary enrollment rate to 39% for the median British Africa colony

and 24% for the median French or Portuguese African colony.

After accounting for the impact of bifurcated legal institutions, most of the competing

explanations for variation in 1960 primary school enrollment rates are statistically irrelevant or

inconsistent between the two samples. In fact, the only control variables that affect 1960 primary

enrollment rates at statistically significant levels are the percentage of Protestants in 1900 and the

number of Protestant missionaries per capita in 1923, supporting Robert Woodberry’s (2012)

theoretical arguments. Catholicism is only advantageous for primary school enrollment in British

colonies, where colonial authorities and Catholic missionaries faced competition from Protestant

missionaries (see Gallego and Woodberry 2010; Woodberry 2012). By contrast, the percentage

of Muslims in 1900 depresses 1960 primary enrollment in former British colonies, while having

no effect in French and Portuguese colonies. This may reflect Muslim opposition to

denominational Christian schools in former British colonies (see Frankema 2012) as well as the

fact that British administrators often restricted Christian missionary activity in predominantly

Muslim areas (see Tibenderana 1983).12 The remaining control variables are either statistically

irrelevant or insignificant.

Postcolonial Governance In Table 4, the statistical models support the hypothesis that bifurcated colonial institutions

contributed to ineffective postcolonial governance. Both British indirect rule (Models 1–3) and

French/Portuguese “native” codes (Models 6–8) predict worse governance scores today,

                                                                                                                         12 Tibenderana (1983) provides detailed historical evidence of this from northern Nigeria, which continues to lag behind southern and Christian parts of the country in terms of educational attainment and economic development. The lasting consequences of this policy are particularly severe for girls’ education.

Owolabi 18

significant at p<0.05. This suggests a high level of continuity in institutional quality from the

colonial era until today. The magnitude of this effect is similar in both samples, although the use

of bifurcated legal codes predicts an even higher percentage of the variation in the quality of

governance among former French and Portuguese colonies (Models 6–8) relative to former

British colonies (Models 1–3). This likely reflects the particularly arbitrary and repressive nature

of French and Portuguese “native” legal codes, which were not based on precolonial customs or

traditions. Consequently, their long-term consequences for the rule of law may have been even

more pernicious than in former British colonies, where democracy was more likely to survive

after independence (see Bernhard, Reenock, and Nordstrom 2004; R. Collier 1982). Another

possible explanation is that the greater presence of Protestant missionaries in the British colonial

empire mitigated some of the negative developmental consequences of indirect British rule

relative to bifurcated French or Portuguese legal institutions. After all, Woodberry (2012)

suggests that there were nearly eight times as many Protestant missionaries per capita in the

British colonies in this sample, relative to French or Portuguese colonies.13 The negative effects

of bifurcated colonial legal institutions on postcolonial governance remain statistically

significant after controlling for colonial settlers (Models 2 and 7), ethnic fractionalization, and

religious composition (Models 3 and 8). None of these control variables are statistically

significant at conventional levels.

                                                                                                                         13 Because Woodberry’s (2012) missionary data do not cover former colonies that never gained independence, these models control for the extent of Protestantism in 2000 rather than the historical presence of Protestant missionaries in 1923.

Owolabi 19

TABLE 3

COLONIAL EDUCATION (OLS RESULTS)

VARIABLES

Net Primary Enrollment Rate (1960) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

British Indirect Rule (%) -0.52** (0.10)

-0.30* (0.13)

-0.37** (0.08)

-0.50** (0.11)

-0.35* (0.16)

French or Portuguese “Native” Legal Code (%)

-0.50** (0.10)

-0.49** (0.11)

-0.43** (0.15)

-0.26* (0.12)

-0.39** (0.13)

Ethnic Fractionalization -2.45 (13.15)

4.13 (11.93)

% European Settlers 3.65* (1.38)

-0.88 (0.68)

% Asian Settlers 0.29 (0.15)

4.22** (1.14)

% Catholic (1900) 0.63* (0.27)

0.05 (0.17)

% Protestant (1900) 0.35* (0.15)

0.55** (0.23)

% Muslim (1900) -0.28** (0.10)

-0.15 (0.09)

Protestant Missionaries per 10,000 pop. (1923)

4.01* (1.61)

23.34** (8.00)

Years Protestant Missions x 100 (1960)

-0.05 (0.07)

0.09 (0.07)

Logged per capita GDP (1950)

7.84 (5.07)

5.80 (5.70)

West Indies 17.29 (9.74)

5.18 (14.69)

Sub-Saharan Africa 7.17 (9.01)

-8.04 (7.41)

Portuguese Colony -11.60 (6.70)

-12.62 (8.83)

Constant 69.69** (4.93)

53.38** (8.35)

62.64** (5.27)

69.81** (11.54)

2.14 (38.22)

67.64** (7.38)

61.91** (8.44)

64.16** (13.48)

37.65** (11.07)

27.56 (42.67)

Observations 32 32 32 31 31 35 35 35 28 33 R-squared 0.45 0.60 0.75 0.58 0.58 0.45 0.63 0.59 0.49 0.60

Coefficients in parentheses, **p<0.01, *p<0.05.

Owolabi 20

TABLE 4

POSTCOLONIAL GOVERNANCE

VARIABLES Mean Governance Score, 1996–2012

(1) OLS (2) OLS (3) OLS (4) OLS (5) 2SLS (6) OLS (7) OLS (8) OLS (9) OLS (10) 2SLS

British Indirect Rule (%) -0.016** (0.004)

-0.012* (0.005)

-0.014** (0.004)

French or Portuguese “Native” Legal Code (%)

-0.016** (0.002)

-0.012** (0.004)

-0.014** (0.003)

% European Settlers 0.053 (0.053)

-0.008 (0.015)

% Asian Settlers 0.006 (0.006)

0.004 (0.041)

Ethnic Fractionalization -0.400 (0.494)

-0.291 (0.311)

% Protestant (2000) 0.010 (0.010)

-0.002 (0.013)

% Muslim (2000) -0.002 (0.005)

-0.000 (0.002)

Logged per capita GDP (2001)

0.645** (0.096)

0.587** (0.122)

Logged per capita GDP (2001)—Instrumented

0.673** (0.138)

0.915** (0.188)

West Indies 0.140 (0.251)

0.125 (0.240)

0.780* (0.346)

0.528 (0.373)

Sub-Saharan Africa 0.397 (0.228)

0.432 (0.249)

0.347 (0.214)

0.660* (0.258)

Portuguese Colony 0.178 (0.195)

0.057 (0.207)

Dependency 0.492 (0.368)

Constant 0.370* (0.171)

0.062 (0.278)

0.418 (0.312)

-5.487** (0.818)

-5.725** (1.165)

0.600** (0.155)

0.31 (0.30)

0.685** (0.179)

-5.090** (0.975)

-7.663*** (1.486)

Observations 32 32 32 31 31 33 33 33 32 32 R-squared 0.39 0.43 0.46 0.69 0.69 0.66 0.69 0.68 0.64 0.54

Standard errors in parentheses; 2SLS models use British indirect rule (Model 5) or French/Portuguese native code (Model 10) as the excluded instrument; Centered R-squared values reported for 2SLS models, **p<0.01,*p<0.05.

Owolabi 21

Models 4 and 9 test the (likely) possibility that variation in governance scores for 1996–

2012 might reflect current socioeconomic conditions rather than bifurcated legal institutions

from the colonial era. Existing studies demonstrate a positive association between economic

development and effective governance (e.g., Lipset 1959; Hadenius 1992; Przeworski et al.

2000), even if scholars disagree about the causal direction of this relationship. In Models 4 and 9,

economic development is positively associated with governmental effectiveness (significant at

p<0.01),14 although existing studies demonstrate that per capita income is itself shaped by

institutional quality (see Acemoglu, Johnson, and Robinson 2001). Consequently, Models 5 and

10 use bifurcated colonial institutions to instrument for per capita income in 2001. These 2SLS

models are substantively similar to the corresponding OLS models, suggesting that economic

development is an important causal mechanism linking colonial institutions to postcolonial

governance. The highest postcolonial governance scores are found in four directly ruled British

colonies (i.e., Singapore +1.4, Barbados +1.26, Hong Kong +1.21, and the Bahamas +1.12),

which have some of the most impressive development records in the Global South. Directly

ruled French and Portuguese colonies (e.g., Réunion +0.92, Martinique +0.79, or Macau +0.62)

do not perform quite as well as their British equivalents, yet their governmental performance

greatly exceeds that of bifurcated French or Portuguese colonies (such as Guinea -1.22,

Cameroon -0.93, or Angola -1.23).

Human Well-Being in 2015

The statistical models in Table 5 demonstrate that bifurcated legal-administrative institutions

from the colonial era are associated with lower levels of human well-being today. Both British

indirect rule (Models 1–3) and French/Portuguese native codes (Models 6–8) have the expected

negative sign, significant at p<0.01. In the bivariate models, bifurcated colonial legal institutions

predict 66% of the variation in current levels of human well-being among former British colonies

(Model 1) and 61% among former French and Portuguese colonies (Model 6). This is an

enormous impact that has been overlooked by existing studies on the colonial origins of long-

term development (e.g., Fails and Krieckhaus 2010; Krieckhaus 2006; Acemoglu, Johnson, and

                                                                                                                         14 Maddison’s per capita GDP data do not cover the Solomon Islands and Timor-Leste, which are omitted from Models 4 and 5 and Models 9 and 10 respectively.

Owolabi 22

Robinson 2001, 2002). These statistical models demonstrate empirically that the negative

developmental consequences of bifurcated legal institutions in the British colonial empire (see

Lange 2009) are consistent among French and Portuguese colonies as well. The statistical

controls for European settlement are insignificant in these models, suggesting that European

settler minorities were more likely to favor extractive institutions that protected their political

and economic privileges rather than inclusive institutions that protected everybody’s property

rights and individual civil liberties (see Fails and Krieckhaus 2010). This casts some doubt on

earlier studies that suggest a positive and linear relationship between European settlement and

long-term development (see Easterly and Levine 2013). Similarly, the statistical controls for the

percentage of Christians and Muslims in 1900 are statistically insignificant.15 Overall, these

models suggest that current levels of human well-being were primarily shaped by colonial legal-

administrative institutions and that the magnitude of this effect is striking and robust to statistical

controls for colonial settlement patterns or religious composition.

Nevertheless, because current levels of human well-being might be better explained by

contemporary political institutions, Models 4 and 9 examine the relationship between

postcolonial governance and human development in 2015. Not surprisingly, effective political

governance (1996–2012) predicts higher 2015 HDI scores among British and non-British

colonies alike.16 2SLS Models that use bifurcated colonial institutions as an instrument for

                                                                                                                         15 These models control for the percentage of Christians in 1900 rather than the historical presence of Protestant missionaries, because Woodberry’s missionary data do not cover dependent territories. Presumably, however, the colonies with the highest percentage of Christians in 1900 also had high levels of missionary exposure. Controlling for the historical presence of Protestant missionaries reduces the sample size considerably but generates substantively similar results. British indirect rule and French/Portuguese “native” codes predict worse human development outcomes (significant at p=0.00 and p=0.05 respectively), whereas the missionary controls are generally insignificant. 16 These models also control for ethnic fractionalization and the impact of being located in the Caribbean or Sub-Saharan Africa. The statistical significance of the Africa dummy may reflect the impact of diseases such as malaria (in Equatorial Africa) or HIV/AIDS (in Southern Africa), which are more prevalent than in other parts of the developing world. According to UNICEF, roughly half of the world’s HIV-infected population lives in Southern Africa and East Africa. In nine Southern African countries, the HIV infection rate exceeds 10% of the total populations (http://www.unicef.org/esaro/ 5482_HIV_AIDS.html). Malaria is also a major killer in Sub-Saharan Africa. Although this mosquito-borne disease is also fairly common in parts of Southeast Asia and Latin America, Sub-Saharan Africa accounts for thirty-eight of the forty countries with the highest rates of mortality from malaria (http://www.worldlifeexpectancy.com/cause-of-death/malaria/by-country/). Nevertheless, effective governance can go a long way toward reducing the death rate from malaria—even eradicating the disease altogether. In the British, French, Dutch, and US Caribbean territories, which have a similar climate to

Owolabi 23

postcolonial governance generate substantively similar results (see Models 5 and 10

respectively). This suggests that bifurcated colonial institutions affect the quality of postcolonial

government, which in turn shapes human development outcomes.

Endogeneity Concerns So far, these statistical models have assumed the extent of legal-administrative bifurcation to be

an exogenous factor that was not shaped by pre-existing conditions in the various colonies. This

assumption seems unlikely, given that European administrators only enforced “native” codes in

territories with significant indigenous populations, which were arguably more difficult for them

to control. Furthermore, colonies acquired prior to 1815 were far less likely to have bifurcated

legal-administrative institutions during the early twentieth century than those colonized after

1850. As Mamdani (1996) suggests, the bifurcated colonial state developed with the significant

territorial expansion of European colonial empires during the nineteenth century. Because these

“omitted” factors might in fact explain the empirical results in Tables 3–5, I estimated six 2SLS

models that use colonial occupation (i.e., countries and territories in which indigenous peoples

made up more than three-quarters of the total population in 1975), the duration of colonial rule

prior to 1960, and the size of population in 1900 as instruments for the extent of legal-

administrative bifurcation during the colonial era. These instruments should not directly affect

postcolonial outcomes except through their impact on colonial legal-administrative institutions.

For the first time, British, French, and Portuguese colonies are combined into a single sample.

These results should be interpreted cautiously, given the different measures of legal-

administrative bifurcation in British vs. non-British colonies. At either extreme—i.e., directly

ruled colonies that lacked native codes and highly bifurcated colonial states with high levels of

indirect British rule, or the widespread use of French/Portuguese “native codes”—should not

significantly bias the regression results. In the middle ranges (i.e., 35% to 65%), however, my

estimates of legal-administrative bifurcation in French and Portuguese colonies might not map

onto Matthew Lange’s measure of indirect rule.

                                                                                                                                                                                                                                                                                                                                                                                                       many parts of West Africa, malaria was largely eradicated after the Second World War (see Rawlins, Hinds, and Rawlins 2008).

Owolabi 24

TABLE 5

POSTCOLONIAL DEVELOPMENT

VARIABLES HDI Scores, 0–100 (2015)

(1) OLS (2) OLS (3) OLS (4) OLS (5) 2SLS (6) OLS (7) OLS (8) OLS (9) OLS (10) 2SLS British Indirect Rule (%) -0.37**

(0.05) -0.30** (0.06)

-0.37** (0.06)

French or Portuguese “Native” Code (%)

-0.36** (0.05)

-0.23** (0.08)

-0.29** (0.08)

% European Settlers 0.48 (0.69)

0.30 (0.36)

% Asian Settlers 0.16* (0.07)

1.04 (0.76)

% Christian (1900) -0.00 (0.06)

0.08 (0.08)

% Muslim (1900) -0.04 (0.07)

0.00 (0.05)

Mean Governance Score (1996–2012)

12.35** (1.64)

10.85** (3.30)

Mean Governance Score—Instrumented (1996–2012)

22.07** (4.68)

22.33** (5.77)

Ethnic Fractionalization -0.67 (4.43)

14.65 (8.43)

-8.17 (7.16)

4.55 (9.26)

West Indies -0.76 (3.55)

-8.74 (5.29)

4.14 (8.02)

-11.06 (10.59)

Sub-Saharan Africa -13.03** (2.46)

-11.86** (3.85)

-13.27** (4.10)

-15.90** (4.58)

Portuguese Colony 0.59 (0.37)

-2.27 (4.22)

Dependency 10.53 (8.28)

Constant 76.34** (2.33)

71.14** (3.62)

76.97** (3.82)

71.52** (2.25)

67.00** (3.98)

81.67** (3.78)

69.15** (6.89)

75.65** (6.96)

74.76** (3.81)

76.59** (4.21)

Observations 32 32 32 32 32 35 35 35 35 35 R-squared 0.66 0.71 0.66 0.74 0.63 0.61 0.70 0.62 0.73 0.61

Standard errors in parentheses; 2SLS models use British Indirect rule (Model 5) or French/Portuguese Native Code (Model 10) as the excluded instrument; Centered R-squared values reported for 2SLS models, **p<0.01,*p<0.05.

Owolabi 25

TABLE 6

INSTRUMENTAL VARIABLE MODELS

VARIABLES Net Primary Enrollment Rate (1960)

Mean Governance Score (1966–2012) HDI Score (2015)

(1) OLS (2) 2SLS (3) 2SLS (4) OLS (5) 2SLS (6) 2SLS (7) OLS (8) 2SLS (9) 2SLS % “Native” Code

-0.53** (0.07)

-0.01** (0.00)

-0.30** (0.04)

% “Native” Code—instrumented

-0.64** (0.08)

-0.34** (0.10)

-0.02** (0.00)

-0.01** (0.00)

-0.36** (0.05)

-0.30** (0.07)

British Colony -3.89 (5.15)

-7.41 (5.23)

-6.94 (4.24)

-0.13 (0.17)

-0.28 (0.18)

-0.26 (0.20)

-2.95 (3.01)

-5.63 (3.21)

-5.96* (2.97)

Portuguese Colony -20.81** (7.57)

-22.23** (6.27)

-23.95** (5.89)

-0.03 (0.22)

-0.09 (0.22)

-0.11 (0.23)

-3.00 (3.81)

-3.98 (3.76)

-4.18 (3.37)

% European Settlers -0.08 (0.50)

0.01 (0.02)

0.34 (0.27)

% Asian Settlers 0.23* (0.11)

0.01 (0.00)

0.11 (0.07)

Ethnic Fractionalization -10.67 (7.83)

-0.49 (0.29)

-3.01 (4.25)

% Protestant (1900) 0.34** (0.12)

0.01 (0.01)

0.07 (0.06)

% Muslim (1900) -0.27** (0.06)

-0.00 (0.00)

-0.04 (0.03)

West Indies 8.07 (7.20)

0.00 (0.30)

-9.35* (4.15)

Sub-Saharan Africa -0.61 (4.30)

0.26 (0.16)

-8.30** (2.31)

Dependency 0.48 (0.23)

0.23 (0.31)

0.35 (0.33)

11.47* (4.59)

7.47 (4.88)

6.98 (4.38)

Constant 73.98** (5.83)

81.23** (6.27)

73.14** (6.99)

0.39 (0.23)

0.69** (0.25)

0.57 (0.38)

76.56** (3.92)

81.56** (4.54)

84.85** (5.59)

Observations 67 67 67 65 65 65 67 67 67 R-squared 0.57 0.55 0.76 0.54 0.52 0.59 0.68 0.67 0.80

Standard errors in parentheses; Centered R-squared values reported for 2SLS models, **p<0.01,*p<0.05. Excluded instruments in 2SLS models: colonial occupation; logged number of years since colonization (1960); logged population in 1900.

Owolabi 26

The OLS models in Table 6 demonstrate that bifurcated colonial legal institutions predict

worse developmental outcomes in the combined sample of sixty-five to sixty-seven British,

French, and Portuguese colonies, significant at p<0.01. These results are substantively similar to

the split-sample results in Tables 3–5, suggesting that combining British and non-British

colonies into a single sample does not bias the regression estimates. The instrumented measure

of bifurcated legal institutions is negatively associated with each outcome variable, confirming

that the duration of colonial rule, the size of the population in 1900, and a predominantly

indigenous population shaped the extent of legal-administrative bifurcation during the colonial

era, which in turn had negative developmental consequences for primary school enrollment in

1960, postcolonial governance from 1996 to 2012, and human well-being in 2015.

SUMMARY AND DISCUSSION

The statistical models in this paper provide overwhelming empirical support for Mamdani’s

(1996) hypothesis that bifurcated colonial institutions hindered long-term development and

undermined effective postcolonial governance. While previous empirical research by Matthew

Lange (2009) demonstrates the negative developmental consequences of bifurcated legal

institutions in former British colonies, this paper demonstrates similar outcomes for French and

Portuguese colonies with limited European settlement. British indirect rule devolved significant

local autonomy to traditional indigenous elites, generating poorly integrated states with highly

restrictive legal systems and limited bureaucratic capacity. This limited their ability to promote

broad-based development and establish effective governmental institutions after independence

(Lange 2009). The use of distinctive “native” legal codes in French and Portuguese colonies had

similarly pernicious effects on the bureaucratic capacity and inclusiveness of the colonial state,

with similarly negative developmental consequences. By contrast, directly ruled colonies,

whether British, French, or Portuguese, developed bureaucratic institutions within a uniform and

inclusive legal-administrative framework that promoted governmental effectiveness and

successful development outcomes over the longue durée.

Consider, for example, the significant developmental gap between the “old” French

colonies that established a uniform legal-administrative framework and inclusive citizenship

rights during the nineteenth century (i.e., French Guiana, Guadeloupe, Martinique, and Réunion)

and other French colonies that maintained bifurcated legal-administrative institutions until after

Owolabi 27

the Second World War. In 1935–40, the primary school enrollment rate in the “old” French

colonies was 48%, compared with only 13% in Algeria, 11% in Indochina, and only 2% in

French West and Equatorial Africa (Benavot and Riddle, 1988). This gap persisted for several

decades. By 1960, the median adult literacy rate in the “old” French colonies was 75%,

compared to 14% in all remaining French colonies (United Nations 1963, 1967). If we examine

life expectancy as another indicator of human well-being, the developmental gap between the

two groups of French colonies is equally large and enduring. In 1955–1960, the median life

expectancy in the “old” French colonies was 56.7 years, compared with 40.5 years in bifurcated

French colonies. Although life expectancy has increased significantly across the board during the

past five decades, the “old” French colonies enjoyed an even larger development advantage (with

a median life expectancy of 78.8 years, compared with 59.7 years in other French colonies) by

2005–2010.17 Although some of the developmental success of the old French colonies might be

due to their current political status as overseas departments of the French Republic (which

enables them to share some of the benefits of the French welfare state), the historical evidence in

this paper suggests that this developmental advantage originated as a result of inclusive legal and

political institutions established during the colonial era. After all, the developmental gap between

the two groups of French colonies had already emerged prior to 1946, and it remains stubbornly

persistent today. Not surprisingly then, the statistical models in this paper demonstrate worse

developmental outcomes in bifurcated colonial states compared to directly ruled colonies that

established a uniform, bureaucratic, and inclusive legal-administrative framework. And this

pattern remains consistent regardless of the colonizing power.

In many ways, the conceptual framework and empirical analysis of this paper challenge

the dominant narratives of leading empirical studies from the previous decade. Many of these

studies highlight the distinctive developmental consequences of “settler” vs. “extractive”

colonialism (see Acemoglu, Johnson, and Robinson 2001; Sokoloff and Engerman 2000;

Krieckhaus 2006; Easterly and Levine 2013), the developmental advantages of British colonial

and/or legal institutions (see La Porta, Lopez-de-Silanes, and Shleifer 2008; La Porta et al. 1998,

1999), or the developmental advantages that resulted from Christian (and especially Protestant)

evangelization (see Woodberry 2012; Lankina and Getachew 2012, 2013). Nevertheless, these

                                                                                                                         17 Author’s calculations based on data from the United Nations’ (2016) World Population Prospects.

Owolabi 28

studies do not account for the use of distinctive “native” legal codes, which had negative

developmental consequences in British, French, and Portuguese colonies alike.

Building on Matthew Lange’s (2004, 2009) research on the distinctive developmental

legacies of direct vs. indirect British rule, the statistical models in this paper confirm Mahmood

Mamdani’s (1996) hypothesis that bifurcated colonial institutions had negative developmental

consequences in British and continental European colonies alike. Regardless of the colonial

power, the use of distinct legal-administrative institutions for rural indigenous populations is

consistently associated with poor development outcomes and ineffective postcolonial

governance, as demonstrated by the experience of countries such as Sierra Leone (British),

Guinea (French), and Guinea-Bissau (Portuguese). By contrast, colonies that developed a

uniform, bureaucratic, and inclusive legal-administrative institutional framework, such as

Barbados (British), Martinique (French), and Macau (Portuguese) have much better

developmental outcomes and governing institutions today. Consistent with other historical-

institutionalist literature (e.g., Lange 2004, 2009; Acemoglu, Johnson, and Robinson 2001,

2002), the statistical models in this paper suggest that legal-administrative institutions

established during the colonial era continue to impact current developmental outcomes and the

institutional quality of government today.

Owolabi 29

APPENDIX

MEASURING LEGAL-ADMINISTRATIVE INSTITUTIONS IN BRITISH, FRENCH, AND PORTUGUESE COLONIES

TABLE A1

LEGAL-ADMINISTRATIVE INSTITUTIONS IN BRITISH COLONIES

Colony Start of Colonization

Type of Colonization

Bifurcated Colonial State

% Indirect Rule, 1955

Bahamas 1717 Forced Settlement N 0 Bangladesh 1756 Occupation Y 50 Barbados 1627 Forced Settlement N 0 Belize 1638 Other N 0 Botswana 1885 Occupation Y 43 Brunei 1888 Other Y 1 Fiji 1871 Other Y 55 Gambia 1888 Occupation Y 37 Ghana 1874 Occupation Y 65 Guyana 1623 Forced Settlement N 0 Hong Kong 1842 Other N 0 India 1757 Occupation Y 49 Jamaica 1655 Forced Settlement N 0 Kenya 1886 Occupation Y 59 Lesotho 1884 Occupation Y 50 Malawi 1891 Occupation Y 82 Malaysia 1873 Other Y 6 Mauritius 1715 Forced Settlement N 0 Myanmar 1826 Occupation Y 16 Nigeria 1885 Occupation Y 93 Pakistan 1857 Occupation Y 50 Sierra Leone 1896 Occupation Y 81 Singapore 1819 Other N 0 Solomon Islands 1893 Occupation Y 52 Sri Lanka 1580 Occupation N 0 Sudan 1898 Occupation Y 73 Swaziland 1894 Occupation Y 49 Tanzania 1890 Occupation Y 75 Trinidad and Tobago 1592 Forced Settlement N 0 Uganda 1893 Occupation Y 80 Zambia 1890 Occupation Y 60 Zimbabwe 1895 Occupation Y 40

Sources: The initial year of colonization is based on data from Lange and Dawson (2009), except the Bahamas, Belize, Brunei, Hong Kong, and the Solomon Islands, which I obtained from the Worldmark Encyclopedia of Nations, 2004; mode of colonization is based on the author’s coding; colonial state institutions are coded as “bifurcated” if the legal system is partially based on customary law; the extent of indirect rule is based on data from Lange (2009, 48).

Owolabi 30

TABLE A2

LEGAL-ADMINISTRATIVE INSTITUTIONS IN FRENCH COLONIES— “NATIVE” LEGAL CODE

Colony Onset of Colonization

Type of Colonization

Pop. in 1900 (thousands)

% European and/or

Assimilated, 1945

% Native Code, c.

1955

Algeria 1830 Occupation 4600 12.3 65 Benin 1900 Occupation 620 0.1 93 Burkina Faso 1898 Occupation 1400 0.1 96 Cambodia 1884 Occupation 2500 0.1 90 Cameroon 1889 Occupation 2620 0.1 68 Central African Republic

1900 Occupation 770 0.1 81

Chad 1910 Occupation 1700 0.1 95 Comoros 1886 Occupation 70 0.5 93 Congo, Rep. (Brazzaville)

1891 Occupation 540 0.2 68

Cote d’Ivoire 1893 Occupation 1000 0.1 84 Djibouti 1888 Occupation 20 3.2 55 Fr. Guiana 1604 Other 21 100.0 0 Fr. Polynesia 1844 Other 37 14.5 0 Gabon 1880 Occupation 280 4.5 86 Guadeloupe 1635 Forced Settlement 182 100.0 0 Guinea 1849 Occupation 990 0.2 91 Laos 1893 Occupation 1500 0.1 92 Madagascar 1896 Occupation 2580 1.4 90 Mali 1898 Occupation 1300 0.1 90 Martinique 1635 Forced Settlement 208 100.0 0 Mauritania 1900 Occupation 220 0.1 96 Morocco 1912 Occupation 5200 3.8 70 Niger 1922 Occupation 910 <0.1 95 Réunion 1650 Forced Settlement 173 100.0 0 Senegal 1895 Occupation 1000 1.2 68 Togo 1884 Occupation 470 0.1 92 Tunisia 1881 Occupation 1600 7.4 62 Vietnam 1859 Occupation 11,000 0.1 87

Sources: The initial year of colonization is based on data from Lange and Dawson (2009), except for Comoros, French Guiana, French Polynesia, Guadeloupe, Martinique, and Réunion, which I obtained from the Worldmark Encyclopedia of Nations, 2004; mode of colonization is based on the author’s coding; population in 1900 is obtained from the World Christian Encyclopedia, 2001; the percentage of European and/or “assimilated” persons in the population is obtained from France (1949); the % native code is estimated by multiplying the percentage of the population with indigenous legal status in 1945 (France 1949) by the percentage of the population living in rural areas in 1955 (United Nations, 2016).

Owolabi 31

TABLE A3

LEGAL-ADMINISTRATIVE INSTITUTIONS IN PORTUGUESE COLONIES

Colony Onset of Colonization

Type of Colonization

Pop. in thousands

(1900)

% European and/or

“Assimilated” (i.e., civilisado),

1950

% Native Legal Code,

1955

Angola 1576 Occupation 2970 3.3 88 Cape Verde 1460 Forced Settlement 70 100.0 0 Guinea-Bissau 1885 Occupation 120 1.6 87 Macau 1557 Other 64 100.0 0 Mozambique 1885 Occupation 2600 1.6 95 São Tomé and Principe

1485 Forced Settlement 40 72.1 40

Timor-Leste 1520 Occupation 370 1.7 89

Sources: The initial year of colonization is based on data from Lange and Dawson (2009), except for Cape Verde, Macau, São Tomé and Principe, and Timor-Leste, which I obtained from the Worldmark Encyclopedia of Nations, 2004; mode of colonization is based on the author’s coding; population in 1900 is obtained from the World Christian Encyclopedia, 2001; the percentage of European and/or “assimilated” persons in the population is obtained from Portugal (1960); the % native code is estimated by multiplying the percentage of the population with indigenous legal status in 1950 (Portugal 1960) by the percentage of the population living in rural areas in 1955 (United Nations, 2016).

These data clearly demonstrate that Britain introduced some degree of indirect rule in

most colonies with significant indigenous populations (see Table A1). Similarly, France and

Portugal enforced distinctive “native” legal codes in colonies with significant indigenous

populations (see Tables A2 and A3). Consequently, all three powers developed bifurcated legal-

administrative institutions in the majority of their colonies. In general, territories that lacked

significant indigenous populations, and those that were colonized prior to 1815, were less likely

to maintain bifurcated legal-administrative institutions into the twentieth century. The data in

Table A1 are largely consistent with Matthew Lange’s historical overview of British colonization

(see Lange 2009, Ch. 2). Direct rule was commonly used in British colonies acquired prior to

1815, including the forced settlement colonies of the Caribbean and Indian Ocean region (where

white planters imported large numbers of African slaves and/or Asian indentured laborers), and

Ceylon (present-day Sri Lanka). Britain also established direct rule in colonies with significant

commercial or geo-strategic importance, such as Hong Kong and Singapore. By contrast, indirect

Owolabi 32

rule was widely used in “occupation” colonies with significant indigenous populations.18 Indirect

rule was first formalized on the Indian subcontinent following the 1857 Sepoy Mutiny. It later

became the dominant form of colonial administration in British African territories acquired after

1880 (see Lange 2009, Ch. 2).

Tables A2 and A3 demonstrate similar geographic and temporal variation in the French

and Portuguese colonial empires. Forced settlement colonies (such as Martinique, Guadeloupe,

and Réunion) and commercial city-states with significant geo-strategic importance (such as

Macau) were more likely to develop bureaucratic and inclusive state structures that resembled

direct British rule. By contrast, the French and Portuguese also developed bifurcated state

structures in colonies with significant indigenous populations, and especially in colonies

acquired after 1830. The extension of “native” legal codes was particularly widespread in French

and Portuguese colonies with significant indigenous populations and limited urbanization.

French colonial census records distinguish between “European and assimilated” persons subject

to the French civil code and indigenous persons subject to traditional authorities (in Morocco and

Tunisia) or to the French indigénat legal code (see France 1944; France 1949). Similarly,

Portuguese colonial census records distinguish between European and assimilated persons (i.e.,

civilisados) who were recognized as Portuguese citizens vs. native populations subject to the

indigenato legal code (Portugal 1960), which was used in all Portuguese occupation colonies

until 1961.

                                                                                                                         18 I define forced settlement as the mode of colonization based on the mass import of African slaves and/or Asian indentured labor, whose descendants make up more than three-quarters of the postcolonial population. By contrast, colonial occupation refers to European domination over territories in which indigenous peoples comprise more than three-quarters of the postcolonial population. In previous work, I demonstrate that forced settlement generated superior outcomes for colonial literacy and postcolonial democracy as a result of nineteenth-century reforms that extended metropolitan legal rights to emancipated populations following the abolition of slavery. By contrast, most “occupation” colonies maintained bifurcated legal institutions into the twentieth century, with negative consequences for educational attainment and postcolonial democracy (see Owolabi 2015).

Owolabi 33

REFERENCES

Acemoglu, Daron, and James A. Robinson (2012). Why Nations Fail: The Origins of Power, Prosperity, and Poverty. New York: Crown Business.

Acemoglu, Daron, Francisco Gallego, and James A. Robinson (2014). “Institutions, Human

Capital, and Development,” Annual Review of Economics, 6: 875–912. Acemoglu, Daron, Simon Johnson, and James A. Robinson (2001). “The Colonial Origins of

Comparative Development: An Empirical Investigation,” American Economic Review, 91(5): 1369–1401.

______ (2002). “Reversal of Fortune: Geography and Institutions in the Making of Modern

World Income Distribution,” Quarterly Journal of Economics, 117(4): 1231–1294. Ajayi, J. F. A. and Michael Crowder (1971). “West Africa, 1919–1939: The Colonial Situation,”

in J. F. A. Ajayi and Michael Crowder (eds.), History of West Africa, Vol. 2, Ch. 14, pp. 514–541. New York: Columbia University Press.

Albouy, David Y. (2012). “The Colonial Origins of Comparative Development: An Empirical

Investigation: Comment,” American Economic Review, 102(6): 3059–3076. Anckar, Carsten (2011). Religion and Democracy: A Worldwide Comparison.

Abingdon, UK: Routledge. Bates, Robert (1981). Markets and States in Tropical Africa: The Political Basis of Agricultural

Policies. Berkeley: University of California Press. Benavot, Aaron, and Phyllis Riddle (1988). “The Expansion of Primary Education, 1870–1940:

Trends and Issues,” Sociology of Education, 61(3): 191–210. Bernhard, Michael, Christopher Reenock, and Timothy Nordstrom (2004). “The Legacy of

Western Overseas Colonialism on Democratic Survival,” International Studies Quarterly, 48(1): 225–250.

Bollen, Robert (1979). “Political Democracy and the Timing of Development,” American

Sociological Review, 44(4), 572–587. Brown, David S. (2000). “Democracy, Colonization, and Human Capital in Sub-Saharan

Africa,” Studies in Comparative International Development, 35(1): 20–40. Collier, Paul J. (2000). “Economics, Politics, and Economic Performance,” Economics and

Politics, 12(3): 225–245. Collier, Paul, and J. Willem Gunning (1999). “Why Has Africa Grown Slowly?” Journal of

Economic Perspectives, 13(3): 3–22.

Owolabi 34

Collier, Ruth (1982). Regimes in Tropical Africa: Changing Forms of Supremacy, 1945–1975.

Berkeley: University of California Press. Crowder, Michael (1964). “Indirect Rule—French and British Style,” Africa XXXIV(3): 197–

205. Domínguez, Jorge (1993). “The Caribbean Question: Why Has Liberal Democracy

(Surprisingly) Flourished?” in Domínguez, Jorge, Robert A. Pastor and R. Delisle Worrell (eds.), Democracy in the Caribbean: Political, Economic, and Social Perspectives. Baltimore: Johns Hopkins University Press.

Duffy, James (1961). “Portuguese Africa (Angola and Mozambique): Some Crucial Problems

and the Role of Education in Their Resolution,” Journal of Negro Education, 30(3): 294–301.

Easterly, William, and Ross Levine (1997). “Africa’s Growth Tragedy: Policies and Ethnic

Divisions,” Quarterly Journal of Economics, 112(4): 1203–1250. _____ (2013). “The European Origins of Economic Development,” National Bureau of

Economic Research, Working Paper 18162. Edie, Carlene (1994). Democracy in the Caribbean: Myths and Realities. Westport, CT: Praeger

Publishers. Engerman, Stanley, Elisa V. Mariscal, and Kenneth L. Sokoloff (2009). “The Evolution of

Schooling in the Americas, 1800–1925,” in David Eltis, Frank D. Lewis, and Kenneth L. Sokoloff (eds.), Human Capital and Institutions: A Long-Run View, pp. 93–142. Cambridge: Cambridge University Press.

Englebert, Pierre (2000). State Legitimacy and Development in Africa. Boulder: Lynne Rienner

Publishers. Fails, Matthew, and Jonathan Krieckhaus (2010). “Colonialism, Property Rights, and Modern

World Income Distribution,” British Journal of Political Science, 40(3): 487–508. Ferreira, Eduardo de Sousa (1974). Portuguese Colonialism in Africa: The End of an Era. Paris:

UNESCO. France, Ministère des Colonies (1944). “Annuaire Statistique des Possessions Françaises:

Années Antérieures à la Guerre.” Paris: Imprimerie Nationale Française. France, Ministère de la France Outre-Mer (1949). Annuaire Statistique de l’Union Française

Outre-Mer, 1939–46. Paris: Imprimerie Nationale Française.

Owolabi 35

Frankema, Ewout (2012). “The Origins of Formal Education in Sub-Saharan Africa: Was British Rule More Benign?” European Review of Economic History, 16(4): 335–355.

Gallego, Francisco, and Robert D. Woodberry (2010). “Christian Missionaries and Education in

Former African Colonies: How Competition Mattered,” Journal of African Economies, 19(3): 294–329.

Grier, Ruth (1997). “The Effect of Religion on Economic Development: A Cross National Study

of 63 Former Colonies,” Kyklos, 50(1): 47–62. Habyarimana, James, Marcartan Humphreys, Daniel Posner, and Jeremy M. Weinstein (2007).

“Why Does Ethnic Diversity Undermine Public Goods Provision?” American Political Science Review, 101(4): 709–725.

Hadenius, Axel (1992). Democracy and Development. Cambridge: Cambridge University Press. Horowitz, Donald (1985). Ethnic Groups in Conflict. Berkeley: University of California Press. Huntington, Samuel K. (1991). The Third Wave: Democratization in the Late Twentieth Century.

Norman: University of Oklahoma Press. Kenny, Paul D. (2013). “The Origins of Patronage Politics: State-Building, Centrifugalism, and

Decolonization,” British Journal of Political Science, 45(1): 141–171. Kohli, Atul (2004). State-Directed Development: Political Power and Industrialization in the

Global Periphery. Cambridge: Cambridge University Press. Krieckhaus, Jonathan (2006). Dictating Development: How Europe Shaped the Global

Periphery. Pittsburgh: University of Pittsburgh Press. Lange, Matthew (2004). “British Colonial Legacies and Political Development,” World

Development, 32(6): 905–922. _____ (2009). Lineages of Despotism and Development: British Colonialism and State Power.

Chicago: University of Chicago Press. Lange, Matthew and Andrew Dawson (2009). “Dividing and Ruling the World: A Statistical

Test of the Effects of Colonialism on Postcolonial Civil Violence,” Social Forces, 88(2), 785–817.

Lange, Matthew, James Mahoney & Matthias Vom Hau (2006). “Colonialism and Development:

A Comparative Analysis of Spanish and British Colonies,” American Journal of Sociology, 111(5): 1412–1162.

Owolabi 36

Lankina, Tomila, and Lullit Getachew (2012). “Mission or Empire, Word or Sword? The Human Capital Legacy in Postcolonial Democratic Development,” American Journal of Political Science, 56(2): 465–483.

_____ (2013). “Competitive Religious Entrepreneurs: Christian Missionaries and Female

Education in Colonial and Postcolonial India,” British Journal of Political Science, 43(1): 103–131.

La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer (2008). “The Economic

Consequences of Legal Origins,” Journal of Economic Literature, 46(2): 285–332. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny (1998).

“Law and Finance,” Journal of Political Economy, 106(6): 1113–1155. _____ (1999). “The Quality of Government,” Journal of Law, Economics and Organization,

15(1): 222–279. Lawrence, Adria K. (2013). Imperial Rule and the Politics of Nationalism: Anti-Colonial Protest

in the French Empire. Cambridge and New York: Cambridge University Press. Lipset, Seymour M. (1959). “Some Social Requisites of Democracy: Economic Development

and Political Legitimacy,” American Political Science Review, 53(1), 69–105. Maddison, Angus (2006). The World Economy. Paris: Development Center for the Organization

of Economic Cooperation and Development. Mahoney, James (2010). Colonialism and Postcolonial Development: Spanish America in

Comparative Perspective. Cambridge and New York: Cambridge University Press. Mamdani, Mahmood (1996). Citizen and Subject: Contemporary Africa and the Legacy of Late

Colonialism. Princeton: Princeton University Press. Mann, Gregory (2009). “What was the Indigénat? The ‘Empire of Law’ in French West Africa,”

Journal of African History, 50(3): 331–353. Merle, Isabelle (2002). “Retour sur le Régime de l’Indigénat: Genèse et Contradictions des

Principes Répressifs dans l’Empire Français,” French Politics, Culture, and Society, 20(2): 77–97.

Murch, Arvin (1968). “Political Integration as an Alternative to Independence in the French

Antilles,” American Sociological Review, 33(4), 544–562. _____ (1971). Black Frenchmen: The Political Integration of the French Antilles. Cambridge,

MA: Schenkman Publishing Company.

Owolabi 37

Njoh, Ambe (2000). “The Impact of Colonial Heritage on Development in Sub-Saharan Africa,” Social Indicators Research, 52, 161–178.

O’Donnell, Guillermo (1993). “On the State, Democratization, and Some Conceptual Problems:

A Latin American View with Glances at Some Post-Communist Countries,” World Development, 21(8): 1355–1369.

O’Laughlin, Bridgitte (2000). “Class and the Customary: The Ambiguous Legacy of the

Indigenato in Mozambique,” African Affairs, 99(394): 5–42. Owolabi, Olukunle P. (2015). “Literacy and Democracy despite Slavery: Forced Settlement and

Postcolonial Outcomes in the Developing World,” Comparative Politics, 48(1): 43–66. _____ (manuscript). “The Colonial Origins of (Under)development, Authoritarianism and

Democracy: Africa and the West Indies in Comparative Historical Perspective.” Portugal, Instituto Nacional de Estatística (1960). Anuário Estatístico do Ultramar, 1959.

Lisboa: Bertrand Irmãos, Lda. Przeworski, Adam, Michael Alvarez, José Antonio Cheibub, and Fernando Limongi (2000).

Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990. New York and Cambridge: Cambridge University Press.

Ramirez, Francisco, and John O. Boli (1987). “The Political Construction of Mass Schooling:

European Origins and Worldwide Institutionalization,” Sociology of Education, 60(1): 2–17.

Rawlins, S. C., A. Hinds, and J. M. Rawlins (2008). “Malaria and Its Vectors in the Caribbean:

The Continuing Challenge of the Disease Forty-Five Years after Its Eradication from the Islands,” West Indian Medical Journal, 57(5): 462–469.

Sachs, Jeffrey, and Mark Warner (1997). “Sources of Slow Growth in African Economies,”

Journal of African Economies, 6(3): 335–376. Sokoloff, Kenneth, and Stanley Engerman (2000). “History Lessons: Institutions, Factor

Endowments, and Paths of Development in the New World,” Journal of Economic Perspectives, 14(3): 217–232.

Tibenderana, Peter K. (1983). “The Emirs and the Spread of Western Education in Northern

Nigeria, 1910–1946,” Journal of African History, 24(4): 517–534. Ulsaner, Eric, and Bo Rothstein (2016). “The Historical Roots of Corruption: State-Building,

Economic Inequality and Mass Education,” Comparative Politics, 48(2): 227–248. United Nations (1963). Compendium of Social Statistics. New York: United Nations.

Owolabi 38

_____ (1967). Compendium of Social Statistics. New York: United Nations. _____ (2016). UN Data: A World of Information, http://data.un.org. Wikipedia (2016). “List of Countries by Human Development Index,” https://en. wikipedia.org/

wiki/List_of_countries_by_Human_Development_Index. Woodberry, Robert (2012). “The Missionary Roots of Liberal Democracy,” American Political

Science Review, 106(2): 244–274. World Bank (2016). Worldwide Governance Indicators, http://info.worldbank.org/

governance/wgi/index.aspx#home. Worldmark Encyclopedia of the Nations, 2004. Detroit: Thompson, Gale. Young, Crawford (1994). The African Colonial State in Comparative Perspective. New Haven:

Yale University Press.