The Origins of Colonial Investmentsin Former British and French Africa*

Joan Ricart-Huguet�

April 10, 2018

Abstract

Recent literature has documented the economic and political consequences of colonialism,

but we have little systematic evidence about the origins of colonial investments despite their

long-term impact. Combining multiple data sources, I present evidence that they were very

unequally distributed across the districts of 16 British and French African colonies. How did

colonial states allocate their investments? Some focus on ethnic disparities in colonial treat-

ment, following a divide and rule logic, while others emphasize the role of natural resources.

I argue that observable geographic features led some places to become centers of pre-colonial

coastal trade, which increased later colonial investments not only in infrastructure but also

in health and education. Although the context was highly extractive, pre-colonial commerce

also helps explain the limited diffusion of investments within each colony. A de facto de-

centralized colonial state and European settlement are two of the mechanisms behind the

persistence of inequality over time.

Keywords: public investments, colonialism, inequality, Africa, trade, geography.

Word count: 11,949

*For useful discussions and comments, I thank Carles Boix, Evan Lieberman, Marc Ratkovic, LeonardWantchekon, Jennifer Widner, Nikhar Gaikwad, Phil Keefer, Isabela Mares, Omar Garcia Ponce, Bob Wood-berry, Tsering Wangyal Shawa, Bill Guthe, Torben Behmer, Brandon Miller de la Cuesta, Costantino Pischeddaand seminar participants at the Contemporary African Political Economy Research Seminar (CAPERS), at thePolitics and History Network, and at Princeton University. I am very grateful to Elise Huillery and Bob Woodberryfor sharing relevant data. Jeremy Darrington and Elizabeth Bennett helped me locate multiple data sources, whileSeth Merkin Morokoff, Luise Zhong and Bruno Schaffa provided excellent research assistance. Financial supportfrom the Mamdouha Bobst Center for Peace and Justice and the Princeton Institute for International and RegionalStudies is gratefully acknowledged.

�Department of Politics, Princeton University, Princeton, NJ 08540. Email: jricart@princeton.edu.

1

1 Introduction

The last two decades have seen a surge in the literature examining the economic and political

consequences of colonialism, notably in Africa, the Americas and South Asia.1 Existing claims

about the effects of colonial institutions, investments and practices are wide-ranging.2 Some of this

literature uses instances of arbitrary or “as-if” random variation in a particular aspect of colonial

rule to identify those consequences, such as spatial discontinuities in agricultural revenue collection

systems in India (Banerjee and Iyer, 2005), in extractive forced labor in Peru (Dell, 2010), or in

missionary school location in Benin (Wantchekon, Novta and Klasnja, 2015). However, colonial

investments are likely not random beyond some particular instances. Explaining their origins

is important because they fundamentally affected welfare and development during colonial rule

as well as contemporary economic development (Huillery, 2009). Yet, we know little about the

quantitative distribution of colonial investments and about the causes underlying their distribution.

Combining multiple original and existing data sources, I present systematic evidence that

colonial investments were very unequally distributed across and within 16 British and French

colonies in East and West Africa.3 This paper focuses on the large within-colony variation in

investments across colonial districts. This focus allows us to hold institutions constant, since in

East and West Africa they varied little within each colony. Investments in some districts were orders

of magnitude larger than in others in the same colony, also when taking population into account

(Figure 1). These district inequalities exist for the three main types of investments—education,

infrastructure and health—in both empires (Figures 4-6, p. 17) and they do not diminish during

the colonial period.

Why did Europeans invest much more in some districts than others? The most immediate ex-

planation concerns a divide and rule logic, whereby the European colonizer discriminated between

ethnic groups to prevent coordinated opposition to imperial rule. Existing literature provides

1See Nunn (2009) for a concise review on the long-term economic effects of colonialism.2For example, see Acemoglu, Johnson and Robinson (2001) and Mahoney (2010) on the effects of colonial

institutions; Huillery (2009) and Cage and Rueda (2016) on investments; and Guardado (2014) and Wilkinson(2015) on colonial practices.

3Using the current names, the sample comprises Benin, Burkina Faso, Cote d’Ivoire, Ghana, Guinea, Kenya,Malawi, Mali, Mauritania, Niger, Nigeria, Senegal, Tanzania, Sierra Leone, Uganda and Zambia.

2

Figure 1: Public health and education by district in two colonies (1910-1939)0

1020

3040

Guiglo

Tagou

anas

Inde

nie

Segue

la

Odienn

e

Assini

e

Bondo

ukou

Tabou

Agneb

y

Gouro

s

DaloaKon

g

Laho

u

Sassa

ndra

Man

Nzicom

oe

Bassa

m

Baoule

Lagu

nes

Public health staff

010

2030

4050

Kong

Man

Odienn

e

Segue

la

Gouro

s

Bondo

ukou

Tagou

anas

Baoule

Daloa

Guiglo

Laho

u

Nzicom

oe

Sassa

ndra

Agneb

y

Inde

nie

Tabou

Assini

e

Bassa

m

Lagu

nes

Public health staff per 100,000 people

Cote d'Ivoire

05,

000

10,0

0015

,000

Platea

uNige

r

Katsin

a

Benue

Adam

awaIlo

rin

Kabba

Ogoja

Bornu

Bauch

i

ZariaKan

o

Sokot

o

Camer

oonsW

arri

Abeok

utaIje

buOnd

o

Lago

s

Benin

Oyo

Onitsh

a

Owerri

Calaba

r

Students

050

01,

000

1,50

02,

000

Kano

Benue

Sokot

o

Bauch

i

Bornu

Ilorin

Platea

u

Ogoja

Katsin

a

Adam

awa

KabbaNige

rZar

iaOyo

Owerri

War

ri

Onitsh

a

Camer

oons

Abeok

utaOnd

oBen

in

Calaba

rIje

bu

Lago

s

Students per 100,000 people

Nigeria

Note: The graphs show the distribution of investments by district in two former colonies. The left graphs showthe raw number of public health staff (doctors and nurses) and students per colonial district while the right graphsadjust that number per 100,000 people. While the order of districts in the right graphs somewhat changes, inequalityremains remarkably similar.

qualitative and historical evidence that treatment was indeed unequal across groups and regions,

but the causes underlying this inequality remain unclear. Horowitz (1985) termed this inequality

the “ethnic distribution of colonial opportunity”: some groups and regions benefited more from,

or were less harmed by, colonialism than others. For instance, in 1960 “the Kakwa and Lugbara

[in Northwestern Uganda] had between them a single student enrolled in Makerere University”,

the first university in East Africa. “The Baganda [in Central Uganda], though only 16 percent

of the population, comprised nearly half” of the student body (Horowitz, 1985, p. 151, p. 239).

Investments were also unequal in other domains of the British and French colonial state, notably

3

economic activity (Cooper, 1996) and colonial army recruitment.4 It is less clear whether a par-

ticular ethnic or local characteristic can explain these inequalities systematically.

Recent research has shown that higher levels of pre-colonial political centralization across eth-

nic groups, a highly visible characteristic that can be used to divide and rule, improve current

socioeconomic outcomes.5 For instance, the British found it convenient to co-opt the existing

social structure of the the Buganda Kingdom and invested more in that region than elsewhere in

Uganda. However, there are opposite examples as well: because the Hausa States in Northern

Nigeria were more politically developed than the rest of the colony, Lugard (1922) decided they

should be ruled indirectly and hence received lower investments. I find that colonial district in-

vestments were not affected by pre-colonial kingdoms and I find little support for other possible

explanations, such as pre-colonial ethnic diversity and fragmentation (Alesina, Baqir and Easterly,

1999; Easterly and Levine, 1997) lowering investments across districts. Divide and rule was an

important ruling strategy, but it may have been context-specific rather than systematically relying

on some ethnic or regional characteristic.

I argue instead that focusing on the economic logic of colonialism allows us to better understand

variation in infrastructure, education and health investments. European colonialism was first and

foremost an exploitative enterprise with economic profit as one of its main motivations (Young,

1994; Suret-Canale, 1971; Darwin, 2012; Diamond, 2005), and natural resources were central to

the colonial enterprise. Curtin et al. (1995, p. 447) explains that “European capital was invested

where exploitable resources promised the most extractive returns” (Huillery, 2010, p. 271). For

instance, the British took control of Ghanaian gold and Sierra Leonean diamonds because they

enjoyed a first-mover advantage over the French.6 Wantchekon and Stanig (2015, p.5) show that

4The King’s African Rifles in British Kenya provide one example: “lacking the education and financial opportu-nities of their more fortunate Kikuyu neighbors, Kamba soldiers found kazi ya bunduki [military service] appealingbecause it was lucrative” (Parsons, 1999, p. 57). And while the French mission civilisatrice was in theory blind toracial or regional considerations, the conscription “system in French West Africa was biased against the rural, lessprivileged groups” (Sharkey, 2013, p. 156).

5Effects of pre-colonial political centralization on current outcomes include “light density at night, paved roads,immunization, literacy and infant mortality rates” (Bandyopadhyay and Green, 2016, p. 471; see also Michalopoulosand Papaioannou, 2013). These findings are consistent with Diamond’s (2012) claim that “the most important factorbehind [good institutions] is the historical duration of centralized government.”

6“With a great sense of the practical, the British had for a long time been snapping up the best coastal sectors.[...] Britain took possession of the territories with the richest resources and best future, although without any

4

“colonial infrastructure can be predicted [...] by the presence of extractive resources (mines and

quarries) but not by soil quality.” Using novel historical data, I also find that infrastructure is

predicted by natural resources. However, the effect is modest and does not spill over into education

or health investments, which served less extractive purposes.

The main reason why colonial administrators invested so much more in some districts than

others lies in the role played by pre-colonial coastal trade. Early commerce provides a common

origin to later investments in infrastructure, education and health in both empires, a finding

consistent with accounts of historical trade in the Indian and Atlantic oceans (Hourani, 1995,

p. 83; Curtin et al., 1995). Besides providing quantitative support for existing historical work, I

identify the causal effect of coastal trade by using natural harbors and capes as an instrumental

variable.7 Early explorers and traders in the Age of Sail possessed very limited information about

the territory (Foster, 1967), so locational fundamentals (Davis and Weinstein, 2002) such as natural

harbors and capes influenced where they landed and therefore the location of initial trading posts

between the 16th century and the 19th century—the period of slavery and the Triangle trade.8

This trade, of which slavery was an important component (Nunn, 2008), in turn explains later

colonial settlement and investments.9

The relevance of pre-colonial commerce is not restricted to these early coastal “development

hubs”: geodesic distance from early trading centers helps explain the limited diffusion of invest-

ments from these early enclaves to the rest of the colony. As Bates (1974, p. 464) put it, “originating

in ‘nodes’ or ‘central places’, modernity then spreads or ‘diffuses’ into the more remote regions

of the territory.” This diffusion is very limited because colonialism in most of East and West

Africa was short-lived and shallow compared to other regions such as South Asia, where existing

geographical ties: the Niger delta, basis of future Nigeria; the Gambia estuary, Sierra Leone and the Gold Coast”(Chi-Bonnardel, 1973, p. 50).

7Jha (2012) first used natural harbors as an instrument in South Asia. Capes (e.g. Dakar) are also a reasonablepredictor of trading posts in Africa, which has many fewer natural harbors (e.g. Mombasa).

8Europeans sold manufactured goods to Africa in exchange for slaves that were sold in the Americas. TheAmericas, in turn, provided cotton, sugar and other primary commodities to Europe.

9For most of East and West Africa, “pre-colonial” refers to the period starting in the 1500s up to the 1885 BerlinConference and “colonial” refers roughly to the period between 1885 and 1960.

5

research shows that trading companies were central and had a long-lasting influence on economic

development and interethnic tolerance (Jha, 2012; Gaikwad, 2014).10

The autonomy of local colonial administrators and European settlers are two important reasons

why pre-colonial trading areas received much higher colonial investments. First, the lack of a

central planner, and hence the high autonomy of administrators in an effectively decentralized

state, meant that redistributive or long-term strategic investments (e.g. the Uganda Railway)

to develop the periphery were the exception rather than the norm. Second, Europeans settled

in the few districts with basic infrastructure already in place and sometimes acted as a lobby

with interests in trade and agriculture (Gardner, 2012), likely contributing to regional economic

inequality. European settlers are hardly a root cause for variation in investments or institutions

(Acemoglu, Johnson and Robinson, 2001) but a mechanism to explain the persistent inequality

in investments. They only came to East and West Africa in any significant numbers in the 20th

century, 400 years after the Triangle trade had begun. Unlike in neo-Europes and many Spanish

American colonies, there were fewer than 10,000 settlers prior to 1940 in all 16 colonies under

study except for Kenya.11

Given a de facto decentralized colonial state and the spatial concentration of settlers, the early

advantages of pre-colonial trading centers persisted into the colonial period: they became centers

of economic activity that benefited from complementarities between investments in infrastructure,

education and health. The pattern is consistent with a logic of increasing returns (Krugman,

1991b, Pierson, 2000) and with urban agglomeration economies (Krugman, 1991a; Rosenthal and

Strange, 2004), albeit in a very different context: instead of firms in a private market, the agents

are colonial state administrators allocating public finances.

In sum, this paper makes three contributions. First, I show that pre-colonial commerce played

a central role for subsequent colonial development even in the very exploitative context of East

and West Africa, where slaves were a central export and commerce was much less developed than

10The British East India Company lasted almost three centuries while the British Royal Africa Company lastedonly one. French chartered companies in Africa such as the French West India Company or the Senegal Companywere not prominent or long-lasting either.

11According to census data collection in both empires, which had important limitations (Frankema and Jerven,2014).

6

in India (Jha, 2012; Gaikwad, 2014) or China (Jia, 2014). Investment diffusion within a colony

is also in part the result of early trade, suggesting that its importance extends beyond the first

trading hubs in each colony. Second, I build upon the literature in ethnic politics and economic

history. The former provides ample case study evidence of the differences in opportunities and

exploitation across regions and ethnic groups, in line with a logic of divide and conquer, while

the latter highlights the commercial motivation behind European imperialism. I argue that the

root cause for these differences is economic. Finally, I contribute a dataset with new sources on

colonial investments—which allow me to quantify spatial inequality—, historical natural resources

and geographic subnational data spanning colonies of two empires.

2 Locational fundamentals and pre-colonial trade

“It is not an exaggeration that between 1550 and 1800 Europeans learned virtually

nothing new about the lands beyond the African coastline. [...] By 1875, in fact,

European possessions in Africa still only comprised the coastal forts and trading

stations and a few tiny colonies.”

Foster (1967, p. 45, 51)

Locational fundamentals are observable geographic characteristics of a territory that “change

little over time—even if their economic meaning may have evolved. For example, there are ad-

vantages of being near a river [or] on the coast, on a plain instead of a mountain or desert, etc.”

(Davis and Weinstein, 2002, p. 1270). Locational fundamentals are important to understand early

spatial patterns of economic activity in pre-industrial contexts, where geography greatly affects

mobility and economic activity (Diamond, 2005).

Early European exploration and commerce in Africa was difficult partly because much of the

Western and Eastern coastline did not possess geographic features amenable to docking ships.

For instance, much of the Windward and Gold Coasts in present-day Cote d’Ivoire and Ghana are

comprised of shallow waters (Curtin et al., 1995). Given the absence of man-made docks, Europeans

landed where coastal geography was favorable. Europeans observed variation mostly in geographic

7

characteristics because knowledge about socioeconomic and other characteristics of the territory

was very limited until the 19th century (Foster, 1967; Chi-Bonnardel, 1973). Geography was

especially important during the Age of Sail (1500s-1800s), when navigation technology depended

much more on environmental factors and wind patterns than in the later Age of Steam (1800s-)

(Feyrer and Sacerdote, 2009). Hence, the few natural harbors and capes (locational fundamentals)

that existed were valuable to Europeans engaged in the trade, including the shipment of slaves

since the 1600s (Figure 2).12

Figure 2: Timeline of major colonial events in East and West Africa

Early exploration

1450 1600

Pre-colonial Triangle trade

1600 1850

1885Berlin Conference

SettlementColonial period

1890 1960

Note: Dates are approximate. Pre-colonial trade and colonization periods vary between colonies.

For instance, the Portuguese and later the French first landed in three places as they descended

through the Northwest African coast. One was Ras Nouadhibou (Cap Blanc), currently in the

border between Western Sahara and Mauritania. They also established trade in what would later

become the cities of Saint Louis and Dakar in Senegal because the former is a natural harbor

formed by the Senegal river mouth and the latter is in a cape (Cap-Vert).

Other factors influenced the location of trade during early exploration, such as the presence of

natural resources in the few areas that had long extracted them. The most famous cases in the

colonies under study are the diamond mines of Sierra Leone and the Gold Coast (current Ghana),

where the British disrupted a Trans-Saharan gold trade that had existed for centuries (Young, 1994,

p. 134). Even then we can see some geographic logic: two key locations where the British landed

were the natural harbor of Tagrin Bay in Freetown and Cape Coast in Ghana, a country without

natural harbors. East Africa has a longer history of navigation than West Africa because of Arab

explorations of the Indian Ocean and the Arab slave trade (Hourani, 1995, p. 83). However, an

12Young (1994, p. 103) provides a brief discussion of chartered companies in Africa and their relation to thecolonial state.

8

equivalent logic applies. Mombasa, a natural harbor, and Zanzibar, an island, became centers of

Arab-African trade centuries before the Europeans arrived in the 1500s. Horowitz (1985, p. 151)

suggests a similar logic to explain interactions between European colonizers and local populations

in the early colonial period in Africa and Asia: “groups located near the colonial capital, near a rail

line or port, or near some center of colonial commerce—the sitting of which was usually determined

by capricious factors, such as a harbor or a natural resource to be exploited—were well situated to

take up opportunities as they arose.” In sum, locational fundamentals led to differences between

certain areas in West and East Africa, which became engaged in Atlantic and Indian commerce,

and the rest of the territory.

2.1 Early trade and colonial investments

Proximity to trading centers became perhaps even more advantageous during the colonial period.

As Europeans increased effective control of the colonies in the early 20th century, trading posts

became central points for the international shipment of goods produced elsewhere in the colony,

strengthening linkages to the international economy, as Laitin (1982) explains for the case of

Western Nigeria. Lagos and other early commercial centers such as Abidjan and Freetown became

the colonial capitals, adding a political dimension to their economic preeminence within the colony.

Capitals were rarely chosen for their central geographic location within the colony, as it became the

case in independent Nigeria (Abuja) and Cote d’Ivoire (Yamoussoukro). For one thing, colonial

state borders were not even defined by the time the capital was chosen. While economic activity

spread, its persistent concentration in a few places in each colony had spillover effects on other

colonial investments, notably hospitals and schools such as the Ecole William Ponty in Goree

(Senegal) and Fourah Bay College in Freetown (Sierra Leone). In other words, investments in

education and health partly followed basic infrastructure investments. It was cheaper for colonial

officials to establish facilities such as schools and hospitals in places where basic infrastructure was

already in place and hence where initial costs were lower.

9

Similar patterns of concentration and persistence of economic activity have been theorized

elsewhere, originally in Krugman’s (1991b) application of increasing returns theory to economic

geography. 13 Similarly, each colony had a few districts with basic infrastructure—the industrial

core—while the majority was part of a periphery that was largely agricultural, pastoral or nomadic

and where colonial state reach was very limited (Herbst, 2000; Jackson and Rosberg, 1982). Be-

cause parts of the hinterland were not well-known even after colonial borders were defined, initial

costs of investment in a new location, including transportation costs, were compounded by un-

certainty about the future profitability of investments. An increasing returns logic, compounded

with a financially poor colonial administration that could not afford risky investments of uncer-

tain profitability, are two reasons why we should observe persistence in the spatial distribution of

economic activity during the colonial period. These two reasons also suggest that new investments

elsewhere in the colony might be a function of proximity to early commercial centers: all else equal,

investments in less remote areas are less costly and less uncertain.

3 Colonial state: structure and administration

The ‘command and control’ of this [British] empire was always ramshackle and

quite often chaotic. To suppose that an order uttered in London was obeyed round

the world by zealous proconsuls is an historical fantasy (although a popular one).

Darwin (2012, p. xii)

This section dispels the popular conception of the colonial state as a highly organized entity

with a central planner following clear investment rules. In reality, the colonial state in East

and West Africa was decentralized and district administrators often acted autonomously in both

empires. Decentralization and autonomy arguably favored the persistence of pre-colonial patterns

of economic activity into the colonial period. Since these claims are context-specific, I first explain

why decision-making was not centralized, as one might expect. I then reinforce that idea by

13He explains why countries develop an “industrialized core” while the rest remains an “agricultural periphery.”Krugman’s (1991b, p. 483) key insight is that, “in order to realize scale economies while minimizing transport costs,manufacturing firms tend to locate in the region with larger demand, but the location of demand itself depends onthe distribution of manufacturing.”

10

providing relevant background on the administrative structure and public finances of the British

and French colonial state.

3.1 The lack of clear investment rules

Colonial documents do not present a systematic investment strategy within or across colonies.

Investments were not proportional to district population (cf. Figure 1), since the needs of the local

population were not a major concern: colonial officials on the ground responded to their superiors

(upward accountability), not to the colonized (downward accountability). Investments were more

responsive to European settlers, but ultimately colonial officials were not accountable to the settler

minority either. The interests of the British Colonial Office or the French Ministry of the Colonies

did not necessarily coincide with those of settlers, itself a diverse group comprising administrators,

traders, farmers, missionaries and sundry Europeans (Darwin, 2012). Granting additional land to

European farmers, for instance, risked alienating the local population even further. Some settlers

in Kenya created the European Taxpayer’s Protection League to shift the tax burden (further)

away from Europeans to the local population (Gardner, 2012, p. 98), hoping to benefit from public

investments without paying for their cost. In sum, investments were not simply a function of the

settler population.

The lack of a systematic investment strategy is perhaps the most surprising aspect of colonial

policy if we use Weberian states as reference categories. It is less surprising if we consider that

communication and knowledge of the territory were also limited, which made policy coordination

difficult between the core and the periphery (Darwin, 2012, p. xii; Delavignette, 1968, p. 63).

British and French colonial officials routinely applied for grants in aid to the respective ministries

with the goal of conducting infrastructure projects and complement colonial tax and tariff revenues.

Yet, the metropole precisely strove to reduce the quantity of funds, especially of grants-in-aid,

available to colonial governments (Constantine, 1984, p. 14, p. 84; Gardner, 2012, p. 9), resulting

in financial uncertainty and short-term local decision-making.

11

While expenditures presumably responded to particular needs, “no explicit investment strategy

can be found in [French colonial] local budgets. Motivations reported at the beginning of each local

budget explain the general level of annual resources but do not motivate the spatial distribution of

public goods provision” (Huillery, 2009, p. 181). British local budgets present a remarkably similar

focus on detailed descriptions and administration rather than policy: “colonial tax and spending

patterns did not follow a similar logic throughout British Africa” (Frankema, 2011, p. 147) in part

because “[Britain] did not strive to apply a common financial policy to the various dependencies”

beyond “general instructions [...] from the Secretary of State for the Colonies” (Stammer, 1967,

p. 194).

3.2 Autonomous administrators in a decentralized colonial state

Examining British and French colonial institutional structures provides some further insights on

why the state was decentralized and decision-makers were rather autonomous. It also provides us

with a better understanding of the public finances of the colonial state and on their similarities

between these two empires. These similarities are not emphasized in the literature because much

historical work on colonial policy focuses on how French ideas of assimilation and direct rule

differed from British ideas of association and indirect rule (Crowder, 1964; Strang, 1994; Sharkey,

2013) and because studies of colonial finance and administration tend to focus on one empire

(Delavignette, 1968; Suret-Canale, 1971; Constantine, 1984; Gardner, 2012).14 At the top of the

colonial hierarchy, there were the Ministry of the Colonies in Paris and the Colonial Office in

London. These ministries sent a Governor to the colony that acted as the main link between the

metropole and the civil servants in the colony, which included administrators, teachers, judges,

engineers, doctors and nurses. Each district (or cercle) was led by a district head called District

Commissioner (Commandant de Cercle).

District heads stationed far from the colonial core had much latitude in local policy and im-

plementation because of physical distances and poor transportation and communication networks.

14There are historical studies spanning both empires on other issues, notably economic history (Hopkins, 1973)and labor policy (Cooper, 1996).

12

“The administrative organization was officially centralized but effectively decentralized,” making

district heads “the real chiefs of the French empire” (Huillery, 2009, p. 181; Delavignette, 1968).

They oversaw the administration, taxation, justice and other public services in both empires (Suret-

Canale, 1971, p. 72). District heads were also in charge of relations with village chiefs, and the

former partly relied on the latter for policy implementation and revenue collection. “Local chiefs

[in British colonies] had a guise of autonomy in their local jurisdictions, but they were actually

guided and supervised by the British colonial administrators” (Strang, 1994, p. 149). Similarly,

local chiefs in French West African villages were an autonomous and useful but ultimately subor-

dinate figure whose “influence [was limited] to small areas” (Huillery, 2009, p. 181). Lugard (1922)

devised his theory of indirect rule with the British Empire in mind, but its practice extended to

parts of the French empire.15 Since the colony was not an organized entity with a central plan-

ner and clear investment rules, I turn to quantitative data sources to understand what factors

underpinned the spatial distribution of colonial public investments.

4 Data

4.1 Sources

To the best of my knowledge, this paper presents the most extensive data on investments at

the colonial district level for French West Africa (collected by Huillery, 2009) and for the main

eight British colonies under the Colonial Office (original data collection): Benin (formerly Da-

homey), Burkina Faso (Upper Volta), Cote d’Ivoire, Ghana (Gold Coast), Guinea, Kenya, Malawi

(Nyasaland), Mali (French Soudan), Mauritania, Niger, Nigeria, Senegal, Sierra Leone, Tanzania

(Tanganyika), Uganda, Zambia (Northern Rhodesia). One important advantage of focusing on

these colonies is their rather homogeneous institutional structure within each empire (Section 3).

Unlike French Algeria or British Southern Africa, most of East and West Africa was not integrated

15Sharkey (2013) considers the difference in that regard between the two empires a matter of degree, consistentwith the move by Gerring et al. (2011, p. 378) “to understand systems of [direct and indirect] rule along a continuumthat reflects the degree of central control.”

13

into the French and British empires until late in the 19th century.16 Due in part to their similar

colonial experience, all 16 colonies became independent around 1960.

Another important advantage is that record-keeping procedures were very similar within each

empire, since all administrators reported to the Ministry of the Colonies in Paris and the Colonial

Office in London. Huillery (2009) collected the original French records for selected years in the

1910-1939 period.17 I collected British colonial records from 1915, 1920, 1927, 1928 and 1938

as a function of availability of disaggregated data. They often contain detailed information on

demographics, education, health, infrastructure investments and other activities (Figure A.9 shows

a page of a British Blue Book and a page of a French Compte Definitif). I georeference colonial

investments by taking advantage of detailed colonial maps with district boundaries, which enable

the use of districts as the units of analysis (Figure A.10 presents a map published by the Colonial

Office).18

Colonial records are not without problems, however. Data on infrastructure, health and ed-

ucation is presented at the town or district level in some years but is too aggregated in others

and table formats often change, making comparisons across years challenging. They also contain

data gaps, such that some data are available in some years but not others. Further, population

censuses were far from accurate. In the early colonial period, some data may have been educated

guesses rather than censuses in the way they are currently conducted (Cooper, 2005; Frankema

and Jerven, 2014).

To explain colonial investments, I collect a set of sources that provide information on the

physical, geographic, social and economic characteristics of the territory. To test whether ob-

servable ethnic characteristics affect colonial investments, I draw on Murdock’s (1959) dataset of

pre-colonial ethnic group characteristics because it provides useful proxies of pre-colonial economic

and political development, such as intensity of agriculture, settlement patterns, the size of local

16There are exceptions. Part of the hinterland in Senegal and Ghana had been colonized much earlier, whileZambia was incorporated as Northern Rhodesia into the Colonial Office only in 1924. Tanzania, then Tanganyika,was under German occupation prior to World War I and incorporated to the British Empire as a League of NationsMandate only in 1922.

17Technically the panel extends to 1956, but data are mostly missing post-1939.18Colonial district boundaries are relevant today. With some exceptions, many have changed little over time and

around 80% of them remain in place today. Districts today are often partitions of a larger colonial district.

14

communities and level of political centralization. I also code two pre-colonial ethnic characteristics.

I use two indicator variables to code whether the district was located in a pre-colonial kingdom or in

an acephalous society (Olson, 1996; Encyclopedia Britannica, 2016), thereby extending Huillery’s

coding for French colonies to British ones. The other is an ordinal variable for low, medium, or

high historical presence of Islam across districts, improving a rough map by Bartholomew (1913).

New data on trade and natural resources allows me to test economic explanations of colonial

investments. The sources for pre-colonial trading posts come from Slave Voyages (2013), Curtin

et al. (1995) and Huillery (2009). Data for natural harbors and capes comes from Ramsar (2016)

and Ports.com (2016), which contain information on multiple geographic characteristics.19 To

explain investment diffusion, I calculate the geodesic distance from each district capital to the

closest trading post. Natural resources data comes from Hubert’s (1922) map (Figure A.11), likely

the most comprehensive on the issue for pre-1940 West Africa. The other is a detailed worldwide

map by Kuhne (1927). I complement these two main sources with an early publication by the

United States Geological Services (USGS, 1921) that also has world coverage but is much less

detailed.

The data also contain several geological, physical and geographic controls. Huillery (2010) col-

lected several of these physical and geographic attributes for French West Africa, such as distance

between the district capital and the coast, altitude and the presence of navigable rivers—which

I complement with a map by C.S. Hammond (1921). I extend those variables to include British

colonies and collect additional ones. Altitude, for instance, is a rough proxy for disease envi-

ronment, notably for malaria (World Health Organization, 2016), but to better capture disease

environment I use a geocoded map of malaria prevalence around 1900 (Lysenko and Semashko,

1968) and FAO tsetse fly data (Alsan, 2015). Those data are important in tropical Africa, “often

referred to as the white man’s grave” (Darwin, 2012, p. 138).

19Tables A.4 and A.5 provide the list of natural harbors and capes as well as of trading posts and include testsof covariate balance for both indicator variables.

15

4.2 Descriptive statistics

Infrastructure investments levels varied widely by colony (Figure 3). For instance, if we consider

investments per capita, levels of investments in Ghana are similar to those in Guinea but much

higher than those of any other British colony.20 Since colonies received little aid from the metropole,

differences in budgets are largely due to differences in the revenue raising capacity of colonial

governments (Hopkins, 1973, p. 190).

Figure 3: Infrastructure expenditures in British and French colonies (1910-1939, in 1910 FRA)

02.

0e+

064.

0e+

06

Mala

wi

Kenya

Zambia

Sierra

Leo

ne

Ugand

a

Tanza

nia

Nigeria

Ghana

Total

01.

0e+

062.

0e+

06

Mala

wi

Kenya

Zambia

Sierra

Leo

ne

Nigeria

Tanza

nia

Ugand

a

Ghana

Per 1 million people

02.

0e+

064.

0e+

06

Niger

Burkin

a Fas

o

Mau

ritan

iaBen

inM

ali

Cote

d'Ivo

ire

Seneg

al

Guinea

Total

01.

0e+

062.

0e+

06

Niger

Burkin

a Fas

oM

ali

Mau

ritan

ia

Cote

d'Ivo

ireBen

in

Seneg

al

Guinea

Per 1 million people

Note: Expenditures in infrastructure per colony varied widely absolutely and per capita. Except for Ghana, percapita investments in British colonies were lower.

I quantify inequality in investments by computing Gini indices by colony (Figures 4-6).21 The

indices are calculated using district-level investments. While the literature computes regional Gini

indices (Milanovic, 2016), regions and districts are not individuals and hence the administrative

20The denominator includes African and European population, but because colonies had only a few thousandEuropeans the results are almost identical if we exclude Europeans.

21Figures A.3-A.8 present the data by district.

16

divisions of the state affect those indices.22 In the case of infrastructure expenditures, they are

around or even above 0.7 in the average British colony and around 0.6 in the average French

colony. They are usually around or above 0.4 for education (proxied by students per district)

and health (proxied by doctors and nurses per district).23 Inequality is especially high in British

colonies compared to French ones. This could be in part the result of the redistributive federal

budgets in French West Africa, while between-colony variation could be the result of many colony

and empire factors beyond the scope of this paper. To visualize the variation, Figure 7 maps

infrastructure expenditures along with the presence of the most valued natural resources by district.

The correlation pattern varies by colony, seemingly high in Ghana but nonexistent in Nigeria, and

overall it appears weaker than one might expect.

Figure 4: Infrastructure Gini indices by colony

0.2

.4.6

.81

British colonies French colonies

Nigeria

Kenya

Tanza

nia

Mala

wi

Ugand

a

Ghana

Sierra

Leo

ne

Zambia

Benin

Burkin

a Fas

o

Cote

d'Ivo

ire

Seneg

al

Mau

ritan

iaM

ali

Niger

Guinea

Infrastructure expenditures

0.2

.4.6

.81

British colonies French colonies

Ugand

a

Kenya

Ghana

Mala

wi

Tanza

nia

Sierra

Leo

ne

Zambia

Nigeria

Burkin

a Fas

oM

ali

Benin

Cote

d'Ivo

ire

Guinea

Seneg

al

Niger

Mau

ritan

ia

Infrastructure expenditures per capita

Note: To quantify inequality, each colony’s Gini index is calculated using district investments (rather than individualincome, which colonial data do not provide) as the units of analysis.

Infrastructure expenditures in a district-year amount to around 50,000FRA on average, in

1910 real French francs, in each empire (Table A.1). They are a bit more evenly spread in French

West Africa, as also shown by the figures plotting Gini indices. French per capita expenditures

(96,000FRA) are three times British expenditures (31,000FRA), but much of the difference is

22Interstate inequality could be altered by redrawing district boundaries. While it happened sometimes, I findthat over 80% of colonial district boundaries in the 1920s and 1930s persist today.

23Gini indices generally use individual income data to measure economic inequality, where 0 means perfect equalityand 1 perfect inequality. If investments were perfectly equal across districts, the Gini index would be 0. If onedistrict received all investments and the others received nothing, it would be 1.

17

Figure 5: Education Gini indices by colony

0.2

.4.6

.81

British colonies French colonies

Ugand

a

Nigeria

Ghana

Kenya

Tanza

nia

Sierra

Leo

ne

Zambia

Burkin

a Fas

o

Guinea

Cote

d'Ivo

ire Mali

BeninNige

r

Students

0.2

.4.6

.81

British colonies French colonies

Ugand

a

Ghana

Tanza

nia

Nigeria

Kenya

Sierra

Leo

ne

Zambia

Burkin

a Fas

oM

ali

Guinea

Cote

d'Ivo

ireNige

r

Benin

Students per capita

Figure 6: Health Gini indices by colony

0.2

.4.6

.81

British colonies French colonies

Ugand

a

Tanza

nia

Nigeria

Kenya

Ghana

Zambia

Mala

wi

Sierra

Leo

ne

Mau

ritan

iaNige

r

Cote

d'Ivo

ireBen

in

Seneg

al

Guinea

Burkin

a Fas

oM

ali

Health staff

0.2

.4.6

.81

British colonies French colonies

Ugand

a

Tanza

nia

Ghana

Kenya

Mala

wi

Zambia

Sierra

Leo

ne

Nigeria

Burkin

a Fas

o

Cote

d'Ivo

ireMali

Guinea

Mau

ritan

iaNige

r

Seneg

al

Benin

Health staff per capita

driven by low investments in British East African colonies—the French had no colonies in East

Africa. Within West Africa, expenditures per capita are “only” 40% higher in French colonies

(Table A.2). Investments are higher in West Africa for internal and external reasons. The region

was richer than East Africa already in the pre-colonial period and conquest started earlier, two

factors which may have resulted in a more developed tax and tariff system during colonialism.

Colonial records show that most expenditures were devoted to buildings and premises of various

sorts, such as residences of district officers in both core and remote areas, port authorities, and

various transportation expenditures such as railroads, roads, bridges and harbors (Table A.3).

18

Figure 7: Public infrastructure investments by district (1910-1939 average) and location of baseand noble metals (1922)

Note: This map illustrates the colonies under study, their levels of infrastructure investments, and whether a districtcontained some base metal (e.g. iron, zinc) and some precious metals or stones (gold, silver or diamonds). Thecorrelation appears to exist in some countries (Ghana) but not in others (Nigeria). There is much variation withincolonies, some perhaps unexpected. The northernmost district in Mali, formerly French Soudan (SOU), had highper capita investments because of Timbuktu.

The average British and French district received close to no investments in sewage and water

sanitation or in electricity and lighting, infrastructure systems that were too costly for meager

colonial budgets.

5 Results

This section begins by examining the role of pre-colonial trade on colonial investments. The focus

is on coastal colonies, where accidented geography provides us with some exogenous variation in

the establishment of pre-colonial trade. Next, I consider possible alternative explanations such as

natural resources and ethnic characteristics. I then examine the diffusion of colonial investments in

both coastal and inland colonies. Finally, I conduct some exercises to determine whether investment

19

inequalities in 1910 increase, remain stable, or decrease during the colonial period up to World

War II (1939).

I first model the relationship between coastal geography, one the one hand, and pre-colonial

trade and investment outcomes, on the other, in coastal colonies (Table 1). The reduced form

equation takes the following form:

log(Yij) = β0 + β1Gij + LTβ2k +NTβ3k + STβ4k + ηj + εij, (1)

where Y is the colonial investment of interest in district i in colony j. These are district infras-

tructure expenditures (adjusted to 1910 French francs), an indicator for the presence of a railroad

in the district, the number of students and the number of health staff. Infrastructure, education

and health outcomes are logged to increase normality. G stands for natural harbors and capes, L

stands for other locational fundamentals, N for natural resources and soil quality, S for colonial

district population, area and pre-colonial socioeconomic characteristics. η are colony fixed effects

that are included in all models. They account for colony-specific variation in issues ranging from

different baseline levels of investments to possibly different reliability of the colonial records across

colonies.

The two-stage least squares (2SLS) models follow an analogous logic to equation 1 but in-

strument pre-colonial trade. The exclusion restriction claim is that natural harbors and capes

affect colonial investments because they enabled pre-colonial trade in the first place. Many mecha-

nisms may account for the relationship, however, such as early investments in forts and European

settlement, which I explore later. The model is as follows:

Tij = γ0 + γ1Gij + LTγ2k +NTγ3k + STγ4k + ηj + υij (2)

log(Yij) = β0 + βIV Tij + LTβ2k +NTβ3k + STβ4k + ηj + εij, (3)

where T is the presence of a pre-colonial trading post and all other variables follow the notation

described above (equation 1).

20

Table 1: First-stage effect of geography on pre-colonial trade and reduced form effect of geographyon colonial investments

First-stage Reduced form

(1) (2) (3) (4) (5) (6) (7)Trading post Infrastructure Railroad Education Health

Natural harbor or cape 0.47** 0.47** 0.45** 1.64* -0.00 1.76** 1.10**(0.11) (0.12) (0.11) (0.66) (0.12) (0.48) (0.29)

Colony FE Yes Yes Yes Yes Yes Yes YesLocational

fundamentals Yes Yes Yes Yes Yes Yes YesNatural resources and

soil quality No Yes Yes Yes Yes Yes YesSocioeconomiccharacteristics No No Yes Yes Yes Yes Yes

Districts (N) 211 211 211 211 211 202 210R2 0.36 0.37 0.45 0.43 0.32 0.49 0.35A-P F statistic 16.66 16.03 15.10

Notes: † p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses. The first-stage models showthat districts with a natural harbor or a cape were more likely to have a center of pre-colonial trade. The reducedform models show that districts with a natural harbor or a cape received higher investments in all investmentsexcept for railroads, as expected. Given the highly unequal investment distributions, outcomes are logged to reducedependence on extreme observations. The number of observations varies by colonial investment because the datasetlacks student data for Mauritania and public health staff data for Conakry. The Angrist-Pischke first-stage F-statistic is between 12 and 16 in all models, above the Stock and Yogo convention of 10 for weak instruments. Thecontrols include the presence of a navigable river, indices for rugged terrain, tsetse fly and malaria, indicators forthe presence of noble and base metals, a soil quality index, African population, area and district socioeconomiccharacteristics (ethnic diversity, Islam prevalence, intensity of agriculture, settlement patterns, pre-colonial politicalcentralization and indigenous slavery).

Table 2: Second-stage results for the effect of pre-colonial trade on colonial investments

Infrastructure Railroad Education Health

(1) (2) (3) (4) (5) (6) (7) (8)OLS IV OLS IV OLS IV OLS IV

Pre-colonial trading post 1.97* 3.88** 0.01 0.03 1.75** 3.96** 1.19** 2.75**(0.78) (1.41) (0.12) (0.24) (0.42) (1.28) (0.30) (0.87)

Colony FE Yes Yes Yes Yes Yes Yes Yes YesLocational

fundamentals Yes Yes Yes Yes Yes Yes Yes YesNatural resources and

soil quality Yes Yes Yes Yes Yes Yes Yes YesSocioeconomiccharacteristics Yes Yes Yes Yes Yes Yes Yes Yes

Districts (N) 211 211 211 211 202 202 210 210R2 0.44 0.41 0.34 0.34 0.49 0.44 0.36 0.27

Notes: † p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses. The table presents OLS resultsalongside 2SLS (IV) results. The controls are identical to those listed under Table 1.

21

Results from Table 1 strongly suggest that accidented geography affected the location of posts

(first-stage models 1 to 3) and that it also affected infrastructure, education and health invest-

ments (reduced form models 4 to 7).24 Next, I estimate the effect of pre-colonial trading posts on

colonial investments using ordinary least squares (OLS) models alongside the second stage of the

two-stage least squares (2SLS or instrumental variable (IV) models). Since the outcome is logged

to reduce right-skew, the percent change in Y can be interpreted as 100βIV . The presence of a

pre-colonial trading post increases expenditures in infrastructure and the number of students by

about 200-300% and the number of health staff by around 100-300%, depending on the specifica-

tion. The magnitude varies but the effect is large across specifications. Pre-colonial trade does not

increase the likelihood of having a railroad. This null effect on railroads makes sense, since rail-

roads were inland investments, and suggests that a different logic may account for transportation

infrastructure, as discussed below, since railroads served extractive purposes further inland while

early trade usually took place in coastal areas.

5.1 Other explanations

Investments are likely explained by other factors beyond pre-colonial trade. I consider some plau-

sible complementary explanations, notably natural resources and ethnic characteristics in Table 3.

Explanatory variables in the table are also present in equations 1-3 as controls. They are included

simultaneously because pairwise correlations between them are always below 0.4, reducing multi-

collinearity concerns.25 Infrastructure expenditures are higher and railroads more likely in districts

with noble metals (gold, silver) or diamonds, where they are associated with a 68% increase in

infrastructure expenditures.26 However, the effect does not extend to non-extractive investments

(education and health) and base metals, a less valuable but abundant type of metal in West Africa

(Figure A.11), do not affect any investments.

24The instrument and the main predictor are balanced across a set of observable pre-colonial characteristics(Tables A.4 and A.5).

25The exception is a 0.59 correlation between distance from the coast and the tsetse fly index (Alsan, 2015).26Benin, Kenya, Malawi, Niger and Uganda did not have any gold, silver or diamonds as of 1920. The other 11

colonies had at least one of these three resources.

22

Table 3: The limited role of alternative explanations

Infrastructure Railroad Education Health

(1) (2) (3) (4) (5) (6) (7) (8)Coastal All Coastal All Coastal All Coastal All

Pre-colonial trading post 1.76** 1.64* 0.08 0.04 2.00** 2.07** 1.29** 1.31**(0.64) (0.64) (0.11) (0.11) (0.40) (0.39) (0.27) (0.26)

GeographyDistance to coast, 100km -0.18 -0.17� -0.02 -0.04* -0.18� -0.10 -0.09� -0.06�

(0.13) (0.10) (0.02) (0.02) (0.09) (0.07) (0.05) (0.03)Navigable river (1910) -0.85� -0.30 -0.24** -0.20** -0.15 -0.19 -0.10 0.06

(0.46) (0.40) (0.08) (0.07) (0.31) (0.24) (0.18) (0.14)Terrain ruggedness 0.50 0.77 0.05 0.05 0.64 0.32 -0.09 -0.05

(0.88) (0.69) (0.12) (0.10) (0.73) (0.54) (0.29) (0.22)Natural resources

Noble metals (1920) 0.68� 0.46 0.19* 0.20* -0.25 -0.22 -0.20 -0.13(0.36) (0.37) (0.09) (0.08) (0.41) (0.32) (0.21) (0.18)

Base metals (1920) 0.19 0.32 -0.04 -0.01 0.15 0.34 0.22 0.21(0.43) (0.41) (0.09) (0.07) (0.23) (0.22) (0.17) (0.14)

Soil quality index (2000) 0.03 0.19 0.04 0.02 0.25 0.08 -0.06 -0.04(0.22) (0.20) (0.03) (0.03) (0.19) (0.13) (0.09) (0.06)

Disease environmentMalaria prevalence (1900) -0.67� -0.39 -0.11* -0.09* 0.13 0.03 -0.19 -0.12

(0.36) (0.30) (0.05) (0.04) (0.30) (0.22) (0.14) (0.10)Tsetse fly index (1970) -0.22 -0.38 -0.04 -0.06 -0.19 -0.04 -0.08 -0.01

(0.44) (0.36) (0.08) (0.06) (0.29) (0.22) (0.19) (0.14)Ethnic and demographic

characteristicsEthnic fractionalization 0.57 -0.53 -0.41** -0.30* -0.82 -1.07� -0.42 -0.35

(1.08) (0.96) (0.15) (0.12) (0.70) (0.59) (0.39) (0.29)Political centralization 0.61 0.33 0.09* 0.05 0.14 -0.02 0.04 0.01

(0.39) (0.32) (0.05) (0.04) (0.24) (0.18) (0.13) (0.09)Prevalence of Islam (1910) -0.55 -0.56� -0.03 -0.04 -0.68** -0.74** -0.15 -0.17

(0.35) (0.33) (0.05) (0.05) (0.23) (0.22) (0.13) (0.12)African population, logged 0.71** 0.64** 0.10* 0.08* 1.11** 0.89** 0.45** 0.38**

(0.25) (0.23) (0.04) (0.03) (0.21) (0.18) (0.11) (0.09)

Colony FE Yes Yes Yes Yes Yes Yes Yes Yes

Districts (N) 211 312 211 312 202 288 210 311R2 0.42 0.44 0.31 0.31 0.47 0.57 0.33 0.47

Notes: † p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses. The number of observationsvaries by colonial investment because the dataset lacks student data for Mauritania and public health staff data forConakry. It also varies by whether the sample is restricted to coastal colonies or the full sample is included.

23

One of the main explanations for colonial inequality concerns the divide and rule strategy

(Horowitz, 1985). While the practice was common, it remains unclear whether there was a system-

atic heuristic behind it. For instance, did Europeans invest more in pre-colonial kingdoms because

they were more developed to begin with and hence initial costs were lower? This finding would be

consistent with existing literature, which shows that pre-colonial kingdoms provide better public

services and are more developed today (Gennaioli and Rainer, 2007; Michalopoulos and Papaioan-

nou, 2013; Bandyopadhyay and Green, 2016). Also, colonialists might have invested less in more

ethnically diverse districts if these were more difficult to control. I find very modest support for

these hypotheses using several proxies.

Using Murdock’s (1959) map of ethnic homelands, I construct an index of ethnic fractionaliza-

tion at the district level and find that it only reduces the likelihood of a railroad in the district.

The same applies if I use a raw measure counting the number of ethnic homelands per district.

Perhaps surprisingly, higher pre-colonial political centralization does not increase investments. An

alternative specification using a pre-colonial kingdom indicator is not significant across investments

either, and bivariate correlations are usually below 0.1. Pre-colonial kingdoms provide better pub-

lic services and are more developed today, but higher levels of colonial investments do not appear

to be the mechanism behind this finding. A simple ordinal measure for prevalence of Islam (mi-

nority, mixed or majority) suggests that more Muslim districts seem to have fewer students, but

causality remains unclear. In Nigeria, for instance, the Muslim North was ruled indirectly, and

hence investments were lower. However, (Lugard, 1922) suggests that the reason for indirect rule is

the North’s higher political centralization under the Hausa States. I also test other characteristics

such as settlement patterns of the ethnic group and indigenous slavery and find insignificant re-

sults. While divide and rule was a common strategy, especially in British colonies (Wucherpfennig,

Hunziker and Cederman, 2015), some of the main observable ethnic characteristics mentioned in

the literature do not predict investments systematically. These null results do not rule out alter-

native explanations. Rather, they suggest that “ethnic” divide and rule strategies may have been

context-specific instead of systematically relying on observable local characteristic.

24

Only two patterns emerge clearly from Table 3. First, African population is always posi-

tively correlated with investments. Investments may not have been blind to population and locals

migrated to more developed districts. Second, the table also presents a correlational but coher-

ent picture to account for the presence of railroads, the one investment that is not explained by

pre-colonial trade. They were built in districts without rivers, such that the two appear to be

substitutes; with natural resources, supporting an extractive logic; and that were less ethnically

diverse, less malaria-prone and more populated. Beyond these two patterns, alternative explana-

tions have a limited role in explaining colonial investments in infrastructure (other than railroads),

health and education. For instance, distance to the coast, a standard control, is negative across

investments as expected but not always significant. Similarly, coefficients for disease environment

proxies are usually negative but often insignificant.

5.1.1 Differences by empire

I consider alternative explanations further by splitting the sample by empire, since considering

them together may be masking relevant heterogeneity. While not the focus of this paper, I provide

some correlational results in light of the literature comparing both empires (Hopkins, 1973; Cooper,

1996; Lee and Shultz, 2012). Figure 8 presents coefficients from the same models as in Table 3 but

now split by empire. As expected, investments increase with population and pre-colonial trade in

both empires.

Colonialists held many racial prejudices in both empires. However, British investments seem to

discriminate more than the French based on district ethnic diversity or on an alternative measure

using raw number of groups in the district. They also invest less in districts with more Muslim

presence. We do not observe these correlational patterns in French colonies, consistent with the

notion that the British practiced divide and rule more than the French (Wucherpfennig, Hun-

ziker and Cederman, 2015). Finally, the estimates on pre-colonial political centralization are also

insignificant in both empires.

25

Figure 8: Effects by empire

Pre-colonial trading post

Ethnic Fractionalization Index

Pre-col. political centralization

Prevalence of Islam (1910)

African population, logged

Pre-colonial trading post

Ethnic Fractionalization Index

Pre-col. political centralization

Prevalence of Islam (1910)

African population, logged

-2 -1 0 1 2 -2 -1 0 1 2

Infrastructure Rail road

Students Health staff

French colonies British colonies

Regression coefficients

Note: These coefficients result from models identical to those in Table 3, but now splittingthe sample by empire. Confidence intervals shown at the 95% and 90% level.

5.2 Diffusion and persistence of colonial investments

How did investments diffuse within East and West African colonies from early enclaves to the rest

of the colony? To the limited extent investments spread across districts, I examine whether early

trade was relevant or instead its importance was circumscribed to these colonial development hubs.

Colonial state expansion was progressive and limited because of European financial and manpower

constraints. In this logic of gradual expansion, I test whether pre-colonial enclaves were important

departing points to expand inland since there is evidence that “agglomeration economies attenuate

with distance” (Rosenthal and Strange, 2004, p. 2120). I use linear models of the form:

log(Yik) = β0 + β1T + β2DT + β3DC + β4P + β5E + ηk + εik, (4)

where Y is the investment of interest, T is an indicator for pre-colonial trade post, DT is the

distance between the district capital and the nearest pre-colonial trading post, DC is the distance

between the district capital and the coast, P is logged population, η are country fixed effects. E

26

is the logged number of Europeans, a potentially relevant mechanism. I include a pre-colonial

trading post indicator and the standard coastal distance measure to examine variation between

districts without a trading post that is not already explained by coastal distance. I use geodesic

distances between district capitals advisedly because it eliminates the endogeneity of man-made

infrastructure such as roads and even local geography such as hills and rivers.

Table 4: Diffusion of investments (1910-1939) across districts within coastal colonies

(1) (2) (3) (4)Infrastructure Railroads Education Health

Pre-colonial trading post indicator 1.46* 0.04 1.92** 1.25**(0.66) (0.12) (0.41) (0.27)

Distance from post, in 100km -0.27** -0.05** -0.14* -0.06�(0.10) (0.01) (0.06) (0.03)

Country FE Yes Yes Yes YesDistance to coast Yes Yes Yes YesAfrican population Yes Yes Yes Yes

Districts (N) 211 211 202 210R2 0.38 0.20 0.42 0.30

Notes: † p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses.

Table 5: European settlers as a colonial investment diffusion mechanism

(1) (2) (3) (4)Infrastructure Railroads Education Health

Pre-colonial trading post indicator 0.38 -0.13 1.47** 0.79**(0.62) (0.12) (0.40) (0.20)

Distance from post, in 100km -0.21** -0.04** -0.10* -0.04(0.08) (0.01) (0.05) (0.04)

European population, logged 0.99** 0.13** 0.48** 0.43**(0.13) (0.02) (0.13) (0.05)

Country FE Yes Yes Yes YesDistance to coast Yes Yes Yes YesAfrican population Yes Yes Yes Yes

Districts (N) 200 200 191 199R2 0.53 0.31 0.48 0.49

Notes: † p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses.

Increasing distance from a pre-colonial trading post reduces all types of investments (Table

4). The results are not identified but suggest that the limited diffusion of colonial investments

we observe is partly a function of early trade, consistent with Gaikwad’s (2014) findings in India.

The negative effect of increasing distance from early trade likely goes through many channels, one

27

of which could be European settlers. They chose districts that were centers of economic activity,

which should compound even further the advantages of early trading locations (such as Dakar and

Lagos) and their surroundings (Thies and Abeokuta, respectively). Including European settlers

in the model supports the idea that they are a relevant mechanism across investments (Table 5).

The effect of distance from a pre-colonial post remains negative but is smaller across investments,

consistent with the idea that settlers influenced the diffusion of investments.

I also examine the persistence of public investment inequalities during the 1910-1939 colonial

period.27 Serial correlations show that inequalities are persistent in both empires for the three

main public investments (infrastructure, education and health), especially in education and infras-

tructure (Figures A.1 and A.2). To examine whether these disparities increase, I use autoregressive

models in the case of French West Africa (Tables A.6-A.8), where the panel structure is less unbal-

anced than in British colonies. Results are mixed in infrastructure, while in education and health

initial disparities either remain constant or increase. Overall, the exercises in this section do not

disentangle the mechanisms of persistence, as future research should. They show, however, that

the importance of early trade extends beyond initial enclaves; that disparities in public investments

tend to remain stable or increase; and that European settlers are a mechanism in this feedback

loop of weakly increasing disparities within colonies.

6 Conclusion

In this paper, I present evidence that investments in infrastructure, health and education within

British and French African colonies were strikingly unequal across districts. Using district-level

data, I quantify this inequality in investments by constructing a Gini index for each of the 16

colonies and each of the main three investment categories. Inequality was around 0.7 in British

colonies and around 0.5 in French colonies. Those regional inequalities are higher than in many

countries in recent decades (Milanovic, 2016).

27Pre-20th century differences in investments are difficult to measure mostly because systematic records for thepre-colonial period are scarce.

28

I provide evidence that pre-colonial trade, rather than natural resources or ethnic characteris-

tics, is the root cause for this high within-colony inequality in investments. In doing so, the focus

of the explanation shifts away from an identity logic to an economic logic. Divide and rule along

regions and ethnic groups was common, especially in British colonies (Wucherpfennig, Hunziker

and Cederman, 2015). However, I find little support for some of its observable implications. In-

stead of systematically relying on some ethnic or regional characteristic, divide and rule strategies

may have been context-specific. To identify the impact of trade, I instrument pre-colonial trading

posts with the presence of a natural harbor or cape—adapting Jha (2012) and Gaikwad’s (2014)

instrument for the case of South Asia—to increase confidence in the causality of the estimates.

Colony fixed effects increase comparability and account for institutional and other colony-level

factors.

The positive effects of early trade on public investments are not circumscribed to those trading

enclaves: infrastructure, education and health investments decrease the further a district is from a

pre-colonial trading post. Findings in developed countries showing that “agglomeration economies

attenuate with distance” (Rosenthal and Strange, 2004, p. 2120) may extend to colonies in their

early stages of state formation and development. Yet, I show that investment inequalities across

districts do not diminish during much of the colonial period (1910-1939). Those locations with an

early geographic advantage, typically colonial capitals and trading posts, became centers of colonial

economic activity and likely benefited from complementarities between investments, consistent with

a logic of increasing returns. I also provide some evidence that European settlers are a mechanism

behind this inequality. They came to East and West Africa in any significant numbers only in

the 20th century, but they reinforced this feedback loop between pre-colonial trade and colonial

investments.

More broadly, the findings speak to research that highlights the importance of pre-colonial

trade for current development in Africa and South Asia. The pre-colonial slave trade had nefarious

consequences at the “point of origin” or extraction (Nunn and Wantchekon, 2011). Perversely, it

increased colonial investments and development at the African “point of destination” where the

trade took place. This result is consistent with historical accounts in Africa (Curtin et al., 1995)

29

and with Gaikwad (2014), who shows that areas in India with pre-colonial trading ports are more

developed today. Compared to East and West Africa, Europe’s trade with India in textiles and

raw materials was much more advanced and industrialized. Yet, I find that pre-colonial trade

increased colonial development even in a highly extractive context where slaves a central export

until the 19th century.

There are at least two areas for further research. One concerns differences between empires.

Why were infrastructure, health and education investments in British colonies even more unequal

than in French ones? One reason could be institutional: French colonies had a redistributive federal

budget that aided poorer colonies while the British did not. Also, this paper finds weak evidence

that the British were more discriminatory in their investments insofar as, unlike the French, they

may have invested less in districts more ethnically diverse districts and in those with higher Muslim

presence. These patterns are consistent with existing literature, but they ought to be explored

further. A second area of future research concerns the consequences of colonial investments for

current welfare (Huillery, 2009). Existing research focuses on the long-term impact of colonialism,

but future research should aim to better understand and quantify the long-term impact of pre-

colonial trade and colonial investments on current outcomes in former British and French Africa.

References

Acemoglu, Daron, Simon Johnson and James A. Robinson. 2001. “The Colonial Origins of Compar-ative Development: An Empirical Investigation.” The American Economic Review 91(5):1369–1401.

Alesina, Alberto, Reza Baqir and William Easterly. 1999. “Public Goods and Ethnic Divisions.”The Quarterly Journal of Economics 114(4):1243–1284.

Alsan, Marcella. 2015. “The Effect of the TseTse Fly on African Development.” The AmericanEconomic Review 105(1):382–410.

Bandyopadhyay, Sanghamitra and Elliott Green. 2016. “Precolonial Political Centralizationand Contemporary Development in Uganda.” Economic Development and Cultural Change64(3):471–508.

Banerjee, Abhijit and Lakshmi Iyer. 2005. “History, Institutions, and Economic Performance: TheLegacy of Colonial Land Tenure Systems in India.” American Economic Review 95(4):1190–1213.

30

Bartholomew, John George. 1913. A Literary and Historical Atlas of Africa and Australasia.Edinburgh: John Bartholomew and Co.

Bates, Robert H. 1974. “Ethnic Competition and Modernization in Contemporary Africa.” Com-parative Political Studies 6(4):457–484.

Cage, Julia and Valeria Rueda. 2016. “The Long-Term Effects of the Printing Press in Sub-SaharanAfrica.” American Economic Journal: Applied Economics 8(3):69–99.

Chi-Bonnardel, Regine Van. 1973. The Atlas of Africa. New York: The Free Press.

Constantine, Stephen. 1984. The Making of British Colonial Development Policy 1914-1940. Lon-don: Frank Cass & Co. Ltd.

Cooper, Frederick. 1996. Decolonization and African Society: The Labor Question in French andBritish Africa. Cambridge and New York: Cambridge University Press.

Cooper, Frederick. 2005. Colonialism in Question: Theory, Knowledge, History. Berkeley: Uni-versity of California Press.

Crowder, Michael. 1964. “Indirect Rule: French and British Style.” Africa: Journal of the Inter-national African Institute 34(3):197–205.

C.S. Hammond. 1921. Hammond’s Business Atlas of Economic Geography. New York: C. S.Hammond & company, inc.

Curtin, Philip, Steven Feierman, Leonard Thompson and Jan Vansina. 1995. African History:From Earliest Times to Independence. 2nd ed. Pearson.

Darwin, John. 2012. Unfinished Empire. New York: Bloomsbury Press.

Davis, Donald R. and David E. Weinstein. 2002. “Bones, Bombs, and Breakpoints: The Geographyof Economic Activity.” American Economic Review 92(5):1269–1289.

Delavignette, Robert. 1968. Freedom and Authority in French West Africa. London: Frank Cass& Co. Ltd.

Dell, Melissa. 2010. “The Persistent Effects of Peru’s Mining Mita.” Econometrica 78(6):1863–1903.

Diamond, Jared. 2005. Guns, Germs, and Steel. New York: W. W. Norton & Company.

Diamond, Jared. 2012. “What Makes Countries Rich or Poor?” The New York Review of Bookspp. 1–7.

Easterly, William and Ross Levine. 1997. “Africa’s Growth Tragedy: Policies and Ethnic Divi-sions.” The Quarterly Journal of Economics 112(4):1203–1250.

Encyclopedia Britannica. 2016.URL: http://academic.eb.com

FAO/IIASA. 2012. “Harmonized World Soil Database (version 1.2).”.URL: http://webarchive.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML/

31

Feyrer, James and Bruce Sacerdote. 2009. “Colonialism and Modern Income: Islands as NaturalExperiments.” Review of Economics and Statistics 91(2):245–262.

Foster, Phillip. 1967. Africa South of the Sahara. London: MacMillan.

Frankema, Ewout. 2011. “Colonial taxation and government spending in British Africa, 1880-1940:Maximizing revenue or minimizing effort?” Explorations in Economic History 48(1):136–149.

Frankema, Ewout and Morten Jerven. 2014. “Writing history backwards or sideways: towards aconsensus on African population, 18502010.” The Economic History Review 67(4):907–931.

Gaikwad, Nikhar. 2014. “East India Companies and Long-Term Economic Change in India.”.

Gardner, Leigh A. 2012. Taxing Colonial Africa: The Political Economy of British Imperialism.Oxford and New York: Oxford University Press.

Gennaioli, Nicola and Ilia Rainer. 2007. “The modern impact of precolonial centralization inAfrica.” Journal of Economic Growth 12(3):185–234.

Gerring, John, Daniel Ziblatt, Johan Van Gorp and Julian Arevalo. 2011. “An Institutional Theoryof Direct and Indirect Rule.” World Politics 63(3):377–433.

Guardado, Jenny R. 2014. “Office-Selling, Corruption and Long-term Development in Peru.”.

Herbst, Jeffrey. 2000. States and Power in Africa. Princeton, New Jersey: Princeton UniversityPress.

Hopkins, Antony Gerald. 1973. An Economic History of West Africa. New York: Routledge.

Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley and Los Angeles: University ofCalifornia Press.

Hourani, George Fadlo. 1995. Arab Seafaring in the Indian Ocean in Ancient and Early MedievalTimes. Princeton: Princeton University Press.

Hubert, Henry. 1922. “Substances Minerales Utiles de l’Afrique Occidentale.” Photo-Metallo Millet(New York Public Library) .

Huillery, Elise. 2009. “History Matters: The Long-Term Impact of Colonial Public Investments inFrench West Africa.” American Economic Journal: Applied Economics 1(2):176–215.

Huillery, Elise. 2010. “The impact of European settlement within French West Africa: Did pre-colonial prosperous areas fall behind?” Journal of African Economies 20(2):263–311.

Jackson, Robert H. and Carl G. Rosberg. 1982. “Why Africa’s Weak States Persist: The Empiricaland the Juridical in Statehood.” World politics 35(1):1–24.

Jha, Saumitra. 2012. “Trade, Institutions and Ethnic Tolerance: Evidence from South Asia.”American Political Science Review 107(4):806–832.

Jha, Saumitra and Steven Wilkinson. 2012. “Does Combat Experience Foster Organizational Skill?Evidence from Ethnic Cleansing during the Partition of South Asia.” American Political ScienceReview 106(4):1–25.

32

Jia, Ruixue. 2014. “The Legacies of Forced Freedom: China’s Treaty Ports.” The Review ofEconomics and Statistics 96(4):596–608.

Krugman, Paul. 1991a. “History and Industry Location: The Case of the Manufacturing Belt.”The American Economic Review 81(2):80–83.

Krugman, Paul. 1991b. “Increasing Returns and Economic Geography.” Journal of Political Econ-omy 99(3):483–499.

Kuhne, F. 1927. “Map of the mineral deposits of the world. Berlin: Verlag von Dietrich Reimer(Ernst Vohsen) A.G.”.

Laitin, David D. 1982. “Capitalism and hegemony: Yorubaland and the international economy.”International Organization 36(04):687–713.

Lee, Alexander and Kenneth A. Shultz. 2012. “Comparing British and French Colonial Legacies:A Discontinuity Analysis of Cameroon.” Quarterly Journal of Political Science 7(4):365–410.

Lugard, Frederick. 1922. The Dual Mandate in British Tropical Africa. Edinburgh and London:William Blackwood and Sons.

Lysenko, A.Y. and I.N. Semashko. 1968. Geography of malaria: A medico-geographical study ofan ancient disease. In Medical Geography, ed. D. V. Borisevich. Moscow: n.d. p. 85.

Mahoney, James. 2010. Colonialism and Postcolonial Development. Spanish America in Compar-ative Perspective. Cambridge University Press.

Michalopoulos, Stelios and Elias Papaioannou. 2013. “Pre-Colonial Ethnic Institutions and Con-temporary African Development.” Econometrica 81(1):113–152.

Milanovic, Branko. 2016. Global Inequality: A New Approach for the Age of Globalization. Cam-bridge, Massachusetts: Harvard University Press.

Murdock, George Peter. 1959. Africa: Its Peoples and Their Culture History. New York: McGraw-Hill Book Company.

Nunn, Nathan. 2008. “The Long-term Effects of Africa’s Slave Trades.” The Quarterly Journal ofEconomics 123(February):139–176.

Nunn, Nathan. 2009. “The Importance of History for Economic Development.” Annual Review ofEconomics 1(1):65–92.

Nunn, Nathan and Leonard Wantchekon. 2011. “The Slave Trade and the Origins of Mistrust inAfrica.” American Economic Review 101(December):3221–3252.

Olson, James Stuart. 1996. The Peoples of Africa: An Ethnohistorical Dictionary. Westport, CT:Greenwood Press.

Parsons, Timothy H. 1999. The African Rank-and-File. Oxford: James Currey.

Pierson, Paul. 2000. “Increasing Returns, Path Dependence, and the Study of Politics.” TheAmerican Political Science Review 94(2):251–267.

33

Ports.com. 2016. “Ports.com.”.URL: http://ports.com/

Ramsar. 2016. “Ramsar Sites Information Service.”.URL: https://rsis.ramsar.org/

Roessler, Philip, Yannick Pengl, Kyle Titlow, Rob Marty and Nicolas van de Walle. 2016. “TheEmpty Panorama: The Colonial Origins of Spatial Inequality in Africa.”.

Rosenthal, Stuart S. and William C. Strange. 2004. Evidence on the Nature and Sources ofAgglomeration Economies. In Handbook of Regional and Urban Economics. Vol. 4 Elsevier Inc.pp. 2119–2171.

Sharkey, Heather J. 2013. African Colonial States. In The Oxford Handbook of Modern AfricanHistory, ed. John Parker and Reid Richard. Oxford: Oxford University Press chapter 8, p. 536.

Slave Voyages. 2013. “The Trans-Atlantic Slave Trade Database.”.URL: http://slavevoyages.org/

Stammer, D. W. 1967. “British Colonial Public Finance.” Social and Economic Studies 16(2):191–205.

Strang, David. 1994. British and French political institutions and the patterning of decolonization.In The Comparative Political Economy of the Welfare State, ed. Thomas Janoski and Alexan-der M. Hicks. Cambridge and New York: Cambridge University Press p. 395.

Suret-Canale, Jean. 1971. French Colonialism in Tropical Africa, 1901-1945. London: HurstPublishers.

USGS. 1921. World atlas of commercial geology. Part I. Distribution of mineral production. Wash-ington, D. C.: U.S. Geological Survey.

Wantchekon, Leonard, Natalija Novta and Marko Klasnja. 2015. “Education and Human CapitalExternalities : Evidence from Colonial Benin.” Quarterly Journal of Economics 130(2):52.

Wantchekon, Leonard and Piero Stanig. 2015. “The Curse of Good Soil? Land Fertility, Roads,and Rural Poverty in Africa.”.

Wilkinson, Steven I. 2015. Army and Nation: The Military and Indian Democracy since Indepen-dence. Harvard University Press.

World Health Organization. 2016. “Malaria.”.URL: http://www.who.int/ith/diseases/malaria/en/

Wucherpfennig, Julian, Philipp M Hunziker and Lars-Erik Cederman. 2015. “Who Inherits theState? Colonial Rule and Post-Colonial Conflict.” American Journal of Political Science 00(0):1–17.

Young, Crawford. 1994. The African Colonial State in Comparative Perspective. New Haven andLondon: Yale University Press.

34

Appendix A: Tables

Table A.1: Infrastructure expenditures (1910-1939 average) in East and West African districts (in1910 FRA)

Districts (N) Mean Std Dev Min Median Max

British coloniesInfrastructure expenditures 200 44,086 153,461 0 2,834 1,551,032Infrastructure exp. per 100,000 people 200 31,633 110,059 0 3,108 1,134,658

French coloniesInfrastructure expenditures 112 51,240 130,562 0 12,112 1,150,341Infrastructure exp. per 100,000 people 112 96,397 276,465 0 14,451 1,940,058

Note: Average expenditure per district was not very different across empires. However, per capita expenditureswere higher in French colonies. British colonies spanned East and West Africa, while the French did not havecolonies in East Africa.

Table A.2: Infrastructure expenditures (1910-1939 average) in West African districts (in 1910FRA)

Districts (N) Mean Std Dev Min Median Max

British coloniesInfrastructure expenditures 66 107,974 249,909 0 20,117 1,551,032Infrastructure exp. per 100,000 people 66 70,282 180,251 0 9,706 1,134,658

French coloniesInfrastructure expenditures 112 51,240 130,562 0 12,112 1,150,341Infrastructure exp. per 100,000 people 112 96,397 276,465 0 14,451 1,940,058

Note: Subsetting the data to West Africa shows that British expenditures were higher in the aggregate but stilllower on a per capita basis.

Table A.3: Infrastructure expenditures by category (1910-1939 average) in East and West Africandistricts (in 1910 FRA)

Districts (N) Mean Std Dev Min Median Max

British coloniesBuildings 200 18,364 79,394 0 1,175 974,900Transportation 200 10,159 25,484 0 219 168,122Sewage/water 200 8,784 45,625 0 0 487,848Electricity/lighting 200 6,779 41,014 0 0 451,596

French coloniesBuildings 112 22,688 87,264 0 5,854 907,381Transportation 112 20,535 49,458 0 3,423 276,964Sewage/water 112 7,238 30,153 0 409 220,133Electricity/lighting 112 780 3,735 0 0 32,367

Note: Unpacking infrastructure expenditures shows that sewage/water and electricity/lighting was nonexistent inthe median district. Investments were centered on buildings, presumably to affirm colonial presence, and trans-portation, presumably for revenue collection and extraction purposes.

35

Table A.4: Covariate balance between coastal districts with a natural harbor or cape (Yes) andthose without (No)

No Yes p-value

Number of ethnic groups in the district 4.257 3.095 0.167Ethnic Fractionalization Index 0.426 0.335 0.200Gathering 0.104 0.218 0.180Hunting 0.869 0.690 0.234Fishing 1.106 1.585 0.022Intensity of agriculture (none to irrigation) 2.936 3.279 0.143Settlement patterns (nomadic to complex) 6.248 6.626 0.400Political centralization (acephalous to kingdoms) 2.343 2.372 0.909Slavery (absence to prevalent) 2.268 2.262 0.972Prevalence of Islam (1910) 0.543 0.667 0.592Malaria prevalence index (1900) 3.104 2.927 0.419Tsetse fly prevalence index (1970) 1.822 1.603 0.420

Note: N=56. The table shows balance along a set of pre-colonial covariates except in fishing, measured as percentageof the population of the ethnic group engaged in fishing as defined in Murdock (1959). List of districts with anatural harbor: Boke, Calabar, Casamance, Conakry, Dakar, Dar es Salaam, Freetown, Lagos, Lamu, Mikindani,Mombasa, Owerri, Saint Louis, Sherbro, Sine Saloum, Tanga. List of districts with a cape: Ahanta, Baie du Levrier(Nouadhibou), Cape Coast, Dakar, Freetown, Keta, Warri.

Table A.5: Covariate balance between coastal districts with pre-colonial trading posts (Yes) andthose without (No)

No Yes p-value

Natural harbor or cape indicator (instrument) 0.265 0.545 0.034Number of ethnic groups in the district 4.412 2.909 0.069Ethnic Fractionalization Index 0.429 0.334 0.180Gathering 0.202 0.061 0.090Hunting 0.706 0.949 0.102Fishing 1.298 1.267 0.884Intensity of agriculture (none to irrigation) 3.168 2.906 0.261Settlement patterns (nomadic to complex) 6.275 6.565 0.515Political centralization (acephalous to kingdoms) 2.282 2.465 0.476Slavery (absence to prevalent) 2.251 2.287 0.836Prevalence of Islam (1910) 0.618 0.545 0.753Malaria prevalence index (1900) 2.911 3.233 0.134Tsetse fly prevalence index (1970) 1.663 1.858 0.472

Note: N=56. The table shows balance along a set of pre-colonial covariates except in the instrument, as expected,and marginally (p < 0.1) in the number of ethnic groups and in the importance of gathering as an economic activityas defined in Murdock (1959). List of districts with a pre-colonial trading post: Accra, Ada, Ahanta, Assinie,Bagamoyo, Bassam, Calabar, Cape Coast, Casamance, Conakry, Cotonou, Dakar, Freetown, Keta, Kilifi, Kilwa,Lagos, Mombasa, Ouidah, Porto Novo, Saint Louis and Tanga.

36

Appendix B: Persistence during colonial times

Education inequalities seem persistent: serial correlations for the number of students and teachers

in French West African districts are around 0.8 (Figure A.1) across years with sufficient data

between 1910 and 1939. In health staff, the correlation is above 0.5, and for infrastructure it is

around 0.95 for any of the three pairwise correlations. In the case of British colonies, the number of

schools, students and health staff also correlate at 0.8 or above, while for infrastructure investments

it is 0.57 (Figure A.2). The pattern is overall similar in both empires: those ahead in the 1910s

are still ahead at the eve of World War II, even if shocks such as the Great War and the Great

Depression reduced overall levels of revenue collection and therefore investment (Gardner, 2012,

p. 6).

Figure A.1: Persistence of public investments in French West Africa

1915

1928

1939

0

2

4

0

2

4

Teachers, logged

1913

1933

1936

4

6

8

4

6

8

Students, logged

1915

1916

1928

0

4

8

12

0

4

8

12

Infrastructure investments, logged

1915

1928

1939

0

2

4

6

0

2

4

6

Health personnel, logged

Note: The correlation matrices show continuity in logged levels of public investments overtime. Both X and Y axis use the same logged scale.

I test more formally whether investment patterns over time converged or diverged for the

case of French West Africa using autoregressive models with one lag (AR1) (equation 5). One

advantage of AR1 models, as opposed to simple serial correlations, is that the constant controls

for deterministic trends.28 I also examine whether initial levels of colonial investments (Iit0) predict

28In other words, the constant would capture a constant increase across districts due to inflation (already ac-counted for by using real 1910FRA), a larger budget, or other factors.

37

Figure A.2: Persistence of public investments in British East and West Africa

pre-1930

1938

0

5

Schools, logged

pre-1930

1938

0

5

10

Students, logged

pre-1930

1938

0

5

10

15

Infrastructure investments, logged

pre-1930

1938

0

2

4

6

Health staff, logged

Note: The correlation matrices show continuity in logged levels of public investments overtime. Both X and Y axis use the same logged scale.

subsequent differences (equation 6):

Iit = α + γIit−1 + βIit0 + εit (5)

4Iit = Iit − Iit−k = α + βIit0 + εit (6)

Investments (I) are indexed by district i and year t. Models 1 and 2 in Tables A.6-A.7 use levels

of investments, while models 3 and 4 are logged to reduce dependence on outliers. Equation 5 is

an AR1 process that includes the initial value (t0) of I in the time series to adjust for baseline

levels of I (models 1 and 3). Equation 6 considers changes in investments over time (models 2 and

4). β > 0 indicates increasing divergence and β < 0 indicates convergence when compared to the

distribution at t0. The approach of equation 5 is more rigorous than that of equation 6 but the

cost is a reduced sample size. Because the panel is unbalanced and incomplete, the previous value

It−k (equation 6) is not necessarily the previous year as in equation 5 but the nearest previous

year for which there is data available. I report both for completeness because there can be biases

in either modeling strategy given the gaps in the data.

Inequality was already high in the early 1900s, yet all models except one indicate either increas-

ing (β > 0) or constant (β = 0) disparities in educational, health and infrastructure investments,

38

Table A.6: Disparities in infrastructure investments per district (1915-1939, in 1910 FRA)

(1) (2) (3) (4)levels (eq. 5) FD (eq. 6) levels (eq. 5) FD (eq. 6)

Infrastructure investments, lagged 0.32**(0.04)

Infrastructure investments in 1915 0.27** -0.45**(0.09) (0.14)

Infrastructure investments, logged first lag 0.67**(0.09)

Infrastructure investments in 1915, logged 0.11 -0.12*(0.08) (0.06)

Districts (N) 181 195 181 195R2 0.68 0.22 0.87 0.29

Notes: �p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses for panel data models (1 and 3).

Table A.7: Disparities in educational investments per district (1915-1939)

(1) (2) (3) (4)levels (eq. 5) FD (eq. 6) levels (eq. 5) FD (eq. 6)

Teachers, lagged 0.86**(0.06)

Teachers in 1915 0.17** 0.20**(0.07) (0.06)

Teachers, logged first lag 0.64**(0.05)

Teachers in 1915, logged 0.31** -0.00(0.05) (0.03)

Districts (N) 374 632 374 632R2 0.97 0.28 0.97 0.01

Notes: �p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses for panel data models (1 and 3).

consistent with a logic of increase returns. Teachers per district in 1915 predict a later increase in

teachers, and the same seems to apply to health staff. The results on infrastructure investments

vary depending on the specification. Public facilities such as schools and hospitals likely benefited

from complementarities and economies of agglomeration more than basic infrastructure, part of

which was intended to create transportation networks with more remote areas in the colony.

39

Table A.8: Disparities in health investments per district (1915-1939)

(1) (2) (3) (4)levels (eq. 5) FD (eq. 6) levels (eq. 5) FD (eq. 6)

Health staff, lagged 0.40**(0.07)

Health staff in 1915 0.82** 0.08(0.16) (0.09)

Health staff, logged first lag 0.61**(0.06)

Health staff in 1915, logged 0.35** 0.06(0.05) (0.04)

Districts (N) 178 267 178 267R2 0.94 0.23 0.97 0.11

Notes: �p < 0.10, * p < 0.05, ** p < 0.01. Robust standard errors in parentheses for panel data models (1 and 3).

40

Appendix C: Figures

Figure A.3: Public expenditures by district in British colonies (1920-1938, in 1910 FRA)

050

0000

1.0e

+06

1.5e

+06

Krach

i

LawraW

a

Navro

ngo

Gonja

Voltar

iverHo

Ada

Wen

chi

Mam

pong

Mam

prus

i

Obuas

i

Sunya

ni

Bekwai

Saltpo

ndSef

wiKet

aBirim

Ashan

ti

Wes

tern

akim

Winn

eba

Akwap

im

Dagom

ba

Was

aw

Kumas

i

Capec

oast

Ahant

aAcc

ra

Ghana

05,

000

10,0

0015

,000

Centra

lkavir

ondoDigo

Elgeyo

EmbuLa

mu

Nandi

Ravine

South

nyer

i

Tanar

iver

Teita

Naivas

ha

Baring

o

Forth

all

Laiki

pia

North

kavir

ondoUas

inM

eru

South

kavir

ondo

Wes

tsuk

Turka

na

North

nyer

iKilif

i

Trans

nzoia

Kisum

u

Mac

hako

s

Mas

aikenKitu

i

North

ernf

ront

ier

Mom

basa

Kiambu

Nairob

i

Nakur

u

Kenya

010

,000

20,0

0030

,000

Karon

ga

Likom

a

Chinte

che

Mzim

baDow

a

Lower

river

Ncheu

Fortjo

hnsto

n

Dedza

Cholo

Kota

Blantyr

e

Mlan

je

Lilon

gwe

Zomba

Malawi

020

0000

4000

0060

0000

8000

00

Katsin

a

Sokot

o

KabbaW

arri

Bornu

Bauch

i

Ondo

Adam

awa

BenueNige

r

BeninIje

bu

Camer

oons

Calaba

r

OgojaIlo

rin

Owerri

Kano

Platea

u

Onitsh

a

Abeok

utaOyo

Zaria

Lago

s

Nigeria

050

,000

1000

0015

0000

Kono

Karen

e

Tonko

lili

Kenem

a

Sherb

ro

Bonth

e

Pujehu

n

Koinad

ugu

Kailah

un Bo

Bomba

li

Portlo

ko

Moy

amba

Freet

own

Sierra Leone

050

,000

1000

001500

00 2000

002500

00

Bihara

muloKilw

a

Mas

asi

NewalaPar

e

Miki

ndan

i

Mas

aitza

Rungw

e

Kondo

a

Tundu

ru

Singida

Hande

ni

Mus

oma

NjombeUfip

a

Mas

wa

Kwimba

UlangaIri

nga

Kaham

aRuf

iji

Bagam

oyo

Arush

a

Songe

a

Kigom

a

Usam

bara

Shinya

nga

Mbu

lu

Mpw

apwa

Tabor

a

Bukob

a

Mwan

za

Mbe

yaLin

di

Mos

hi

Dodom

a

Tanga

Mor

ogor

o

Man

yoni

Kilosa

Dares

salaa

m

Tanzania

010

0000

2000

0030

0000

4000

00

Wes

tnile

Karam

oja

Mas

aka

Lang

o

Ankole

Mub

ende

Teso

Kigezi

Toro

Budam

a

Bunyo

ro

Acholi

Centra

l

Busog

a

Men

go

Uganda

020

,000

40,0

0060

,000

80,0

00

Balova

le

Chinsa

li

Kawam

bwa

Lund

azi

Luwing

u

Man

koya

Mku

shi

Mpik

a

Mpo

roko

so

Mum

bwa

Namwala

Petau

ke

Senan

ga

Seren

je

Seshe

ke

Kalabo

Fortro

sebe

ry

Kasem

pa

Solwez

i

Isoka

Mon

gu

Kasam

a

Broke

nhill

Fortja

mes

on

Maz

abuk

a

Aberc

orn

Kalom

o

Living

stone

Lusa

ka

Ndola

Zambia

41

Figure A.4: Public expenditures by district in French colonies (1910-1939, in 1910 FRA)

050

,000

1000

0015

0000

Allada

Djougo

u

Ataco

ra

Ouidah

Moy

ennig

er

Savalo

u

Borgo

uM

ono

Coton

ou

Abom

ey

Porto

novo

Benin

05,

000

10,0

0015

,000

20,0

00

Kaya

Koudo

ugou

Tenko

dogo Say

Ouahig

ouya

Gaoua

Fada

Dedou

gou

Bobod

ioulas

so Dori

Ouaga

doug

ou

Burkina Faso

050

,000

1000

0015

0000

Guiglo

Gouro

s

Laho

u

Daloa

Bondo

ukou

Tagou

anas

Odienn

e

Inde

nie Man

Segue

la

Tabou

Sassa

ndra

Assini

eKon

g

Agneb

y

Nzicom

oe

Baoule

Bassa

m

Lagu

nes

Cote d'Ivoire

050

0000

1.0e

+06

1.5e

+06

Nzere

kore

Gueck

edou

Koum

bia

Kissido

ugou Pita

Siguiri

Mac

enta

Kouro

ussa

Forec

ariah

Beyla

Labe

Dabola

Boffa

Kindia

Boke

Mam

ou

Kanka

n

Conak

ry

Guinea

050

,000

1000

0015

0000

2000

0025

0000

Satad

ougo

u

Gourm

aGao San

Niafun

ke

Gound

am

Bafou

labe

Nema

Koutia

la

Mac

inaNior

o

Bougo

uni

Bandia

gara

Kita

Sikass

oNar

aM

opti

Segou

Tombo

ucto

u

Kayes

Bamak

o

Mali

010

,000

20,0

0030

,000

Assab

a

Guidim

aka

Gorgo

l

Adrar

Brakn

a

Tagan

t

Baiedu

levrie

r

Trarz

a

Mauritania

02,

000

4,00

06,

000

8,00

0

Bilma

Dosso

Konny

Nguigm

i

Tahou

a

Tessa

oua

Agade

z

Niamey

Zinder

Goure

Niger

010

0000

2000

0030

0000

4000

00

Dagan

aBak

el

Podor

Tivaou

ane

Casam

ance

Mat

am

Tamba

coun

da

Loug

a

Haute

gam

bie Baol

Thies

Dakar

Sinesa

loum

Saintlo

uis

Senegal

42

Figure A.5: Students by district in British colonies (1920-1938))0

1,00

02,

000

3,00

04,

000

Bekwai

Gonja

Krach

i

Lawra

Navro

ngoSef

wi

Wen

chi

Mam

prus

i

Wes

tern

akim Ho

Dagom

baWaAda

Sunya

ni

Was

aw

Obuas

i

Ashan

ti

Mam

pong

Voltar

iver

Saltpo

nd

Winn

ebaBirim

Kumas

i

Ahant

a

Akwap

imKet

aAcc

ra

Capec

oast

Ghana

05,

000

10,0

0015

,000

Digo

Elgeyo

Laiki

pia

Mas

aiken

Nairob

i

North

ernf

ront

ier

Ravine

Tanar

iver

Turka

na

Naivas

ha

Wes

tsuk

Baring

oLa

mu

Trans

nzoia

EmbuUas

in

NandiTeit

a

Centra

lkavir

ondoKilif

i

Nakur

uKitu

i

North

nyer

i

Mer

u

Kiambu

South

kavir

ondo

North

kavir

ondo

Mac

hako

s

Forth

all

Mom

basa

South

nyer

i

Kisum

u

Kenya

-11

Blantyr

e

Chinte

che

Cholo

Dedza

Dowa

Fortjo

hnsto

n

Karon

gaKot

a

Likom

a

Lilon

gwe

Lower

river

Mlan

je

Mzim

ba

Ncheu

Zomba

Malawi

05,

000

10,0

0015

,000

Platea

uNige

r

Katsin

a

Benue

Adam

awaIlo

rin

Kabba

Ogoja

Bornu

Bauch

i

ZariaKan

o

Sokot

o

Camer

oonsW

arri

Abeok

utaIje

buOnd

o

Lago

s

Benin

Oyo

Onitsh

a

Owerri

Calaba

r

Nigeria

02,

000

4,00

06,

000

8,00

0

Koinad

ugu

Pujehu

n

Tonko

lili

Kenem

a

Karen

e

Kailah

unKon

o

Portlo

ko

Bomba

li Bo

Bonth

e

Moy

amba

Sherb

ro

Freet

own

Sierra Leone

02,

000

4,00

06,

000

Bihara

mulo

Hande

ni

Kondo

a

Kwimba

Man

yoni

Mas

aitza

Newala

Njombe

Tundu

ruUfip

a

Mas

waRuf

ijiPar

e

Miki

ndan

i

Ulanga

Shinya

nga

Mbu

lu

SingidaKilw

a

Kilosa

Mus

oma

Kaham

a

Rungw

e

Usam

bara

Mor

ogor

o

Bagam

oyo

Songe

a

Mas

asi

Mbe

ya

Kigom

aLin

di

Bukob

a

Arush

a

Dodom

a

Mpw

apwa

Mos

hi

Iring

a

Mwan

za

Tabor

a

Tanga

Dares

salaa

m

Tanzania

05,

000

10,0

0015

,000

Karam

oja

Budam

a

Kigezi

Mub

ende

Acholi

Lang

o

Wes

tnile

Bunyo

roTes

oTor

o

Ankole

Centra

l

Busog

a

Mas

aka

Men

go

Uganda

020

040

060

0

Aberc

orn

Balova

le

Chinsa

li

Fortro

sebe

ryIso

ka

Kalabo

Kasam

a

Kasem

pa

Kawam

bwa

Lund

azi

Luwing

u

Man

koya

Mku

shi

Mon

gu

Mpik

a

Mpo

roko

so

Mum

bwa

Namwala

Petau

ke

Senan

ga

Seren

je

Seshe

ke

Solwez

i

Fortja

mes

on

Maz

abuk

a

Kalom

o

Living

stone

Broke

nhill

Lusa

ka

Ndola

Zambia

Note: British colonial records (the Blue Books) do not provide disaggregated education data for Malawi.

43

Figure A.6: Students in French colonies (1910-1939)0

500

1,00

01,

500

Djougo

u

Ataco

ra

Moy

ennig

er

Borgo

u

Allada

Savalo

uM

ono

Coton

ou

Ouidah

Abom

ey

Porto

novo

Benin

050

100

150

200

250

SayKay

a

Tenko

dogo

Koudo

ugou

Fada

Ouahig

ouya

Gaoua Dor

i

Dedou

gou

Bobod

ioulas

so

Ouaga

doug

ou

Burkina Faso

050

01,

000

Tagou

anas

Tabou

Gouro

s

Segue

la

Daloa

Bondo

ukou

Guiglo

Odienn

e

Inde

nie

Laho

uKon

gM

an

Assini

e

Nzicom

oe

Bassa

m

Agneb

y

Baoule

Sassa

ndra

Lagu

nes

Cote d'Ivoire

050

01,

000

1,50

0

Gueck

edou

Beyla

Mac

enta

Koum

biaBof

fa

Kouro

ussa

Forec

ariah

Boke

Kissido

ugou

Dabola

Nzere

kore

Pita

Siguiri

Kindia

Kanka

n

Mam

ouLa

be

Conak

ry

Guinea

050

01,

000

Nema

Satad

ougo

u

Gourm

a

Gound

am

Bougo

uniGao

Bafou

labe

Koutia

laSan

Mac

inaNar

aNior

o

Niafun

ke Kita

Sikass

o

Bandia

gara

Tombo

ucto

u

Segou

Mop

ti

Kayes

Bamak

o

Mali

-11

Adrar

Assab

a

Baiedu

levrie

r

Brakn

a

Gorgo

l

Guidim

aka

Tagan

t

Trarz

a

Mauritania

050

100

150

Bilma

Agade

z

Konny

Tahou

a

Nguigm

i

Niamey

Tessa

oua

Dosso

Goure

Zinder

Niger

01,

000

2,00

03,

000

4,00

0

Haute

gam

bie

Tamba

coun

daBak

el

Mat

am

Dagan

aBao

l

Tivaou

ane

Loug

a

Podor

Thies

Sinesa

loum

Casam

ance

Saintlo

uis

Dakar

Senegal

Note: Student data for Mauritania is missing in Huillery (2009)

44

Figure A.7: Public health staff by district in British colonies (1920-1938))

050

100

150

Ashan

ti

Krach

i

Mam

pong

Obuas

i

Lawra

Navro

ngoAda

Bekwai

Gonja

Mam

prus

i

Saltpo

ndSef

wiW

a

Wen

chi

Voltar

iverHo

Keta

Sunya

ni

Was

aw

Wes

tern

akim

Akwap

imBirim

Winn

eba

Capec

oast

Dagom

ba

Ahant

a

Kumas

i

Accra

Ghana

050

100

150

Centra

lkavir

ondoDigo

Elgeyo

Embu

Laiki

piaLa

mu

Mas

aiken

Naivas

ha

Nandi

North

kavir

ondo

North

nyer

i

Ravine

South

kavir

ondo

Tanar

iver

Wes

tsuk

Turka

na

Baring

o

North

ernf

ront

ier

Trans

nzoiaTeit

aM

eru

KiambuKilif

iKitu

i

Nakur

u

South

nyer

i

Forth

all

Mac

hako

s

Uasin

Mom

basa

Kisum

u

Nairob

i

Kenya

02

46

810

Chinte

che

Cholo

Dedza

Dowa

Fortjo

hnsto

n

Karon

gaKot

a

Likom

a

Lilon

gwe

Lower

river

Mlan

je

Mzim

ba

Ncheu

Zomba

Blantyr

e

Malawi

020

4060

80

Adam

awa

Bauch

i

Benue

Bornu

Camer

oons

Ijebu

Katsin

a

OgojaOnd

o

Sokot

oIlo

rinNige

r

Abeok

utaOyo

Benin

Kabba

Onitsh

a

Platea

uW

arri

Calaba

r

Owerri

ZariaKan

o

Lago

s

Nigeria

010

2030

4050

Kailah

un

Karen

e

Kenem

a

Koinad

ugu

Kono

Moy

amba

Portlo

ko

Pujehu

n

Sherb

ro

Tonko

lili

Bomba

li Bo

Bonth

e

Freet

own

Sierra Leone

020

4060

80

Bihara

mulo

Hande

ni

Kwimba

Man

yoni

Mas

aitza

Mas

asi

Mas

wa

Mpw

apwa

Newala

NjombePar

e

Tundu

ru

Kigom

aRuf

iji

Bagam

oyo

Iring

a

Mbu

lu

Miki

ndan

i

Singida

Ulanga

Kondo

a

Songe

aUfip

a

Arush

aKilw

a

Shinya

nga

Usam

bara

Rungw

e

Mor

ogor

o

Kaham

a

Mus

oma

Kilosa

Mwan

za

Dodom

a

Bukob

a

Mbe

yaLin

di

Tanga

Mos

hi

Tabor

a

Dares

salaa

m

Tanzania

050

100

150

200

250

Karam

oja

Mub

ende

Budam

aTor

o

Lang

oTes

o

Wes

tnile

Ankole

Centra

l

Mas

aka

Acholi

Busog

a

Kigezi

Bunyo

ro

Men

go

Uganda

02

46

Aberc

orn

Balova

le

Chinsa

li

Fortja

mes

on

Fortro

sebe

ryIso

ka

Kalabo

Kalom

o

Kasem

pa

Kawam

bwa

Lund

azi

Luwing

u

Man

koya

Maz

abuk

a

Mku

shi

Mpik

a

Mpo

roko

so

Mum

bwa

Namwala

Petau

ke

Senan

ga

Seren

je

Seshe

ke

Solwez

i

Kasam

a

Mon

gu

Broke

nhill

Living

stone

Lusa

ka

Ndola

Zambia

45

Figure A.8: Public health staff in French colonies (1910-1939)

010

2030

40

Allada

Ataco

ra

Borgo

u

Moy

ennig

er

Djougo

u

Savalo

uM

ono

Ouidah

Abom

ey

Coton

ou

Porto

novo

Benin

010

2030

4050

Fada

Dori

Kaya

Tenko

dogo

Ouahig

ouya Say

Koudo

ugou

Dedou

gou

Gaoua

Bobod

ioulas

so

Ouaga

doug

ou

Burkina Faso

010

2030

40

Guiglo

Tagou

anas

Inde

nie

Segue

la

Odienn

e

Assini

e

Bondo

ukou

Tabou

Agneb

y

Gouro

s

DaloaKon

g

Laho

u

Sassa

ndra

Man

Nzicom

oe

Bassa

m

Baoule

Lagu

nes

Cote d'Ivoire

05

1015

Gueck

edou

Mac

enta

Nzere

kore

PitaBey

la

Kissido

ugou

Koum

biaBof

fa

Kanka

nLa

be

Siguiri

Dabola

Forec

ariah

Mam

ou

Kouro

ussa

Boke

Kindia

Conak

ry

Guinea

020

4060

80

Bafou

labe

Nema

Bougo

uni

Gound

am

Koutia

laKita

Bandia

gara

Gao

Niafun

keNior

o

Segou

Mac

ina San

Gourm

aNar

a

Tombo

ucto

u

Satad

ougo

u

Sikass

oM

opti

Kayes

Bamak

o

Mali

02

46

Guidim

aka

Assab

a

Baiedu

levrie

r

Gorgo

l

Adrar

Brakn

a

Tagan

t

Trarz

a

Mauritania

05

1015

Tahou

a

Goure

Agade

z

Tessa

oua

Bilma

Nguigm

i

Dosso

Konny

Zinder

Niamey

Niger

020

4060

Haute

gam

bieBak

el

Podor

Dagan

a

Tamba

coun

da

Mat

amLo

uga

Baol

Tivaou

ane

Casam

ance

Sinesa

loum

Dakar

Thies

Saintlo

uis

Senegal

46

Figure A.9: Pages of a Blue Book page for Uganda, 1945 (left) and of a Compte Definif for Benin,1928 (right)

COMPTE DÉFINITIF 1928 xxxm

3 W S CRÉDITSS J O

DEPENSESS u «J ' parPu H S NATURE DES DEPENSESOBSERVATIONS<i pi <; AUTORISATIONX <; pq FAITESo < '

des dépensesa.

Cercle de Cotonou.II 17 Entretien des routes et ponts 10.700 » 10.565 70 Entretien des routes.

8 Entretien des marchés et caravansérails ; 3.000 » 2.717 82 Réparation ie la toiture du mar-9 Entretien des immeubles, gîtes d'étapes et : ché de cotonou et uadigconnage.puits 1.250 » 350 50 Réparation do l'école de fiodomey,

, . crépissage des murs et badigeonnage.Totaux. 14.950 » 13 634 02

17 2 2 Autres dépenses imprévues. —Aménagement

d'un champ d'aviation entre Kadjèhoun etGodomey 3 084 » 2 084 » Aménagement d'un champ d'avia-

tion entre Kadjèhoun et (iodomey.Totaux 3.084 » 2.084 »

20 2 1 Construction de ponts et de routes dans lescercles 25.200 » 24.773 50 Grosses réparations a la route de

, Cotonou.Totaux 25.200 » 24.773 50

Cercle de DjougouIII 7 Entretien des routes et ponts 37.000 » 19.608 » Travaux courants d'entretien des

routes Djougou-Sèméré-Djougou-Onklou sur tout leur parcourt).

8 Entretien des mai chés et caravansérails. 1.000 » 020 » Entretien courant des bâtiments dumarché et <lu caravansérail. Réfec-tion de toiture.

10 Entretien des cimetières 125 » »H Réparations et transformation des Tribunaux

indigènes 225 » 200 » Travaux d'entretien divers, maçon-~nerie et toitures.

Totaux 38.350 » 20.428 »

2 3 Construction de bâtiments pour logement defonctionnaires 4.000 » 610 » Construction d'un pavillon de 3

pièces avec vôrandah, toiture chaume.

Totaux 4 000 » 010 »

20 2 l Construction de pnnts et routes nouvelles.' 32.000 » 16.148 » Construction de la route de Djou-, gou a N'Dali.

Totaux 32.000 » 16.148 »

6 3 Travaux divers aux postes médicaux... 5 000 » 4.180 » construction d'un dispensaire ar, Djougou.

Totaux 5.000 » 4.180 »

Cercle du Borgou.11 1 7 PntrptiPii HP« rontP< Pt nnntt; k\ 000 » 49 876 » Entretien et amélioration desroules/ entretien aes roiue» et ponts -il.UUU » i».o/u

,iu cercle. Remise en état de la routeNikki. Kallé, route Guessou, Senendé.Construction d'un pont sur rivièreKala, réfection des ponts du cercle.

8 Entretien des marchés et caravansérails 1.000». 1-009»decgKcu^^en8e^ffr "^tion des toits des caravansérails des

9 Entretien des immeubles, gîtes d'étapes et subdivisions.puits .. . 4 750 » 4.697 » lîntretiendes immeubles du cercle. * '

et creusement des puits du poste.Construction d'un gîte d'étapes â

Bori.10 Fntfpripn HPS pimptipr'PS 250 » 235 » Crépissage et blanchiement des1U entretien aes Cimctieies ~ou » ~o<;

tombes, réfection du mur d'entou-, , „ , . . . . t m •i. raSe â Bem béréké.11 Réparation et transformation des Tribunaux ~, ,. . m .. , ..f.. ,„ . ocn „. -JKO » Transformation du Tribunal deindigènes 3oO » dou

Parakou. Réfection toiture, aména-

fementsalle d'audiences. Crépissage

1V, „ es murs.

47

Figure A.10: Colonial map of Nigeria (1948)

Figure A.11: Natural resources in West Africa

48