Overcoming the Tyranny of History:

Evidence from Post-Apartheid South Africa∗

Paulo Bastos† Nicolas Bottan‡

March 2014

Abstract

We identify the interplay between coercive institutions and the distribution

of natural resource rents as an important driver of local development follow-

ing the end of apartheid, the system of racial segregation enforced in South

Africa between 1948 and 1994. Using data from the 1996 census, we document

large income gaps between communities located just-inside and just-outside

the former self-governing territories set aside for black inhabitants. Exam-

ining relative changes between 1996 and 2011, we find that spatial income

convergence was considerably stronger among marginalized communities with

higher initial exposure to resource rents. These results accord with standard

bargaining theory in which the dismantling of coercive institutions improves

the negotiating position of unionized workers in the mining industry. We

provide additional evidence that our results are not driven by competing ex-

planations.

Keywords : Coercion, natural resource rents, wage bargaining, histor-

ical development.

∗We thank Aaditya Mattoo and seminar participants at the World Bank for very helpful com-

ments. The views expressed in this paper are those of the authors and not of the World Bank.†Development Research Group, The World Bank, 1818 18th Street NW, Washington DC, United

States. E-mail: pbastos@worldbank.org.‡Department of Economics, University of Illinois at Urbana-Champaign, Illinois 61801, United

States. Email: bottan2@illinois.edu.

1

1 Introduction

A growing body of evidence reveals that current economic underdevelopment can

be traced back to specific historical episodes, with long-lasting impacts typically

arising through either domestic institutions or cultural traits (Hall and Jones, 1999;

Acemoglu et al., 2001; Dell, 2010; Nunn, 2008, 2009, 2013). Less attention has

focused on identifying specific channels by which a bad legacy can be turned into a

relatively more prosperous future. In this paper, we identify the interaction between

the dismantling of coercive institutions and the distribution of natural resource

rents as an important driver of the development path of marginalized communities

following the end of apartheid, the system of racial segregation enforced in South

Africa between 1948 and 1994.1

Apartheid legislation classified inhabitants into racial groups and institution-

alized segregation of residential areas, enforced by means of forced removals and

strict control of population movements. In 1970, black people were deprived of

their citizenship, legally becoming citizens of one of ten tribally based self-governing

homelands, four of which became nominally independent states.2 At the same time,

they were prevented from working outside the corresponding homeland, unless a

pass for a particular area was issued. Black inhabitants were granted inferior prop-

erty rights, while labor legislation severely restricted their pay and access to skilled

and semi-skilled occupations. Supply of public goods was segregated, with markedly

inferior provision to black people. In the aftermath of the 1994 democratic elections,

all homelands were legally reintegrated into South Africa, and the newly empowered

government assumed responsibility for the effective dismantling of coercive institu-

tions previously imposed on marginalized racial groups.

Using community-level data from the 1996 and 2011 population censuses, we

provide quantitative evidence on the legacy of the apartheid regime. In October of

1996, just over year after Nelson Mandela took office as the first President elected in

1Racial segregation in South Africa began in colonial times under Dutch and British rule. How-

ever, apartheid as an official policy was introduced following the 1948 general election. Although

the official abolishment of apartheid occurred in 1990, with repeal of the last set of the remaining

apartheid laws, the effective end of the regime is widely regarded as arising from the democratic

general elections of April 1994.2Throughout this paper we will adopt the term homelands to designate the ten self-governing

territories set aside for black inhabitants, noting however that these areas are also commonly

referred to as bantustans.

2

a fully representative election, communities located just-inside and just-outside the

former homelands exhibited large differentials with regard to the racial composition

of the population, income per capita and several other socio-economic indicators.

Fifteen years later, these gaps remained sizable in general, but income convergence

occurred at a different pace across local communities of the former homelands.

Our main interest lies in the evolution of these spatial income disparities in the

post-apartheid period, when coercive institutions imposed on marginalized racial

groups were progressively dismantled and labor union activity was empowered. Ex-

amining relative changes in real per capita income between 1996 and 2011, we

find that spatial income convergence was considerably stronger among previously

marginalized communities with higher initial exposure to resource rents – as mea-

sured by a larger share of employment in the mining industry. This finding holds

irrespective of whether we consider: (i) the universe of local communities; (ii) a

restricted sample of local communities located just-inside and just-outside the for-

mer homelands; or (iii) a sample composed only of communities from the former

homelands.

The evidence we document is in line with bargaining theory emphasizing the in-

teraction between coercive institutions and the distribution of resource rents through

collective wage negotiations. In the model, a domestic unionized producer and a

foreign producer supply a homogeneous commodity to a global integrated product

market. Imperfect competition in the product market generates rents that workers

may capture in the form of higher wages. The wage premia – defined as wages in

excess of what workers would earn elsewhere in the labor market – is determined

through Nash bargaining, and then the firm sets the employment level unilaterally

as part of its product game with the foreign firm. The advent of democracy, accom-

panied by several legislative interventions aimed at leveling the historically uneven

bargaining field, leads to significant improvements in the negotiating position of the

union, which is therefore able to capture a larger share of resource rents.

In this framework, unionized wages would also be expected to rise in the face

of increased resource rents following international price shocks. However, in the

absence of changes in institutions governing collective bargaining, these marginal

rents would accrue mainly to workers and mine owners from non-marginalized racial

groups, who had stronger bargaining power to begin with and resided outside the

former homelands. The evidence we provide reveals that initial exposure to resource

rents – as measured by the initial employment share in the mining industry – is an

3

important driver of local income for black communities in the former homelands,

not of local income in general. This points to a key role of the interaction between

coercive institutions and the distribution of rents, rather than just variation in the

volume of resource windfalls.

We provide further evidence on the specific mechanisms emphasized by the model

by examining complementary individual-level data from the 1996 population census.

Consistent with the existence of large resource rents in the mining industry, we

document a large wage premium associated with employment in this sector. In line

with the expected impacts of coercive institutions, we document a large income gap

against black male workers, who also benefit from a smaller wage premium when

employed in the mining industry. These patterns are robust to the inclusion of

controls for individual attributes and occupational categories.

While suggestive of the importance of coercion and rent-sharing in the resource

industry, the differential income path of former homeland communities in democratic

South Africa could at least in part reflect alternative explanations that operate at

the community-level. First, they might be explained by government interventions

aimed at addressing the historically uneven access to infrastructure.3 In the South

African context, Dinkelman (2011) shows that mass roll-out of the electricity grid

to rural households generated sizable employment gains among females and led to

pay rises among males. If improvements in access to infrastructure are correlated

with initial employment in the mining industry, our estimates would be biased.

Second, since the homelands were spread across the country, the effects of their

progressive integration in the South African economy might be expected depend

on the degree of proximity to markets and connectivity (Harris, 1954; Krugman,

1991; Hanson, 1996; Fujita et al., 2001; Redding and Sturn, 2008). Third, the

dismantling of Apartheid was naturally followed by sizable migration flows from the

former homelands. Heterogeneity in the demographic and skill composition of these

flows across locations might challenge our interpretation of the results.

To further strengthen the credibility of our empirical findings, and to assess the

role of these alternative explanations, we use the community-level data to provide

several additional pieces of evidence. First, we show that our results remain unal-

tered when accounting for changes in access to several infrastructure items, notably

electricity, piped water, refuse and phone. Second, we show that the main find-

3At the time of the 1994 democratic elections, over two thirds of black households did not have

access to electricity (Dinkelman, 2011).

4

ings hold when accounting for the role of expanded market access and connectivity,

while also providing evidence that greater proximity to markets was associated with

stronger income growth. Third, we offer some evidence that relative population

growth was similar in local communities in former homelands regardless of initial

exposure to resource rents. We show, however, that our results remain robust when

accounting for changes in the size and attributes of the local population.

Our paper contributes to a growing literature examining the long-lasting im-

pacts of historical episodes. Many studies established how these episodes affect

domestic institutions or cultural traits (Hall and Jones, 1999; Acemoglu et al., 2001;

Dell, 2010; Nunn, 2008, 2009, 2013). Our study contributes by identifying specific

factors that contributed to breaking the bad legacy of history. Additionally, our

paper relates to a literature studying the empirical relation between market access

and economic growth (for a survey of the literature see Overman et al. 2003). In

particular, the fall of the Apartheid in South Africa serves as a potential Natural

Experiment to examine the relation between increased market access and regional

income convergence. Even though there are papers making use of Natural Experi-

ments for identification (for example Redding and Sturn, 2008), fewer papers do so

in the context of developing countries.

The remainder of the paper proceeds as follows. Section 2 provides background

on the rise and fall of the Apartheid regime, and describes key features of the mining

industry. Section 3 outlines a simple theoretical model of coercion and rent sharing

to guide the empirical work. Section 4 describes the data employed, before section

5 provides descriptive statistics on the evolution of development. Section 6 provides

econometric evidence on the role of the interplay between the dismantling of coercive

institutions and exposure to resource rents in shaping the relative income path of

communities from the former homelands. Section 7 discusses and examines the role

played by several alternative explanations. The last section concludes.

2 Historical background

2.1 The rise and fall of apartheid

The apartheid regime was enforced through several pieces of legislation introduced

by National Party governments that ruled South Africa from 1948 to 1994. Apartheid

considerably strengthened the racial segregation that begun under Dutch and British

5

colonial rule.

The Population Registration Act of 1950 classified inhabitants into four racial

groups–white, black, coloured and indian– and introduced an identity card for all

adult citizens specifying their race. In the same year, the Group Areas Act insti-

tutionalized racial segregation of residential areas. Each race was allotted its own

territory, which was later used as a basis for forced removals.

Public goods provision was highly segregated. The Reservation of Separate

Amenities Act of 1953 made it possible to reserve municipal grounds according

to race, leading to separation of buses, hospitals, beaches and park benches. Black

people were offered services markedly inferior to those of white people, and to a

lesser extent to those of Indian and coloured people. The Bantu Education Act of

1953 introduced a separate education system for black pupils, directed to preparing

them for lives as a labouring class. In 1959 separate universities were created for

black, coloured and indian people, while existing universities were prevented from

enrolling additional black students.

Through the homeland system, the National Party government sought to divide

South Africa into separate nation-states. The Bantu Authorities Act of 1951 cre-

ated separate government structures for white and black citizens. The Promotion

of Black Self-Government Act of 1959 proposed self-governing Bantu units, which

would have devolved administrative powers with the promise of later autonomy and

self-government. The Black Homeland Citizenship Act of 1970 deprived black peo-

ple of their citizenship, who legally became citizens of one of ten tribally based

self-governing homelands. Figure 1 depicts the geographic location of each of these

ten homelands.

The homelands accounted for about thirteen percent of the land, a small share

compared to population. Four homelands were declared independent states by the

South African government: Transkei in 1976, Bophuthatswana in 1977, Venda in

1979, and Ciskei in 1981. In parallel with the creation of the homelands, there was

a massive programme of forced relocation: between the 1960s and 1980s, millions of

inhabitants were forced from their homes, many being resettled in the homelands.

The government aimed for a total removal of the black population to the homelands.

Black people were prevented from running businesses or being employed in white

areas, unless a pass for a particular area was issued. A black person working in a

white-designated area without a pass was subject to arrest and trial for being an

illegal migrant, which would frequently lead to deportation to the corresponding

6

homeland and prosecution of the employer. Labor legislation severely restricted the

levels of pay and access to good jobs by black people.4 Unionization of black workers

was illegal until the early 1970s.

Black labor unions formed in the 1970s and 1980s, such as the National Union

of Mine workers, assumed an increasingly prominent role in economic and political

protests from the mid-1980s. Although the official abolishment of apartheid occurred

in 1990, with repeal of the last set of the remaining apartheid laws, the effective end

of the regime is widely regarded as arising from the democratic general elections

of April 1994. In the aftermath of the 1994 democratic elections, all homelands

were legally reintegrated into South Africa, and the newly empowered government

assumed responsibility for basic service provision for all citizens. The country was

constitutionally redivided into new provinces. These are displayed in Figure 2.

The advent of democracy was followed by numerous legislative initiatives aimed

at dismantling coercive institutions and leveling the historically uneven bargaining

field. The bill of rights of the 1996 constitution prohibited the state from discrimi-

nating on any grounds, and contained an explicit focus on labor rights, notably by

imposing fair labor practices, the right to unionize and the right to strike. The new

constitution served as the basis for new pieces of labor legislation that effectively dis-

mantled coercive institutions governing labor relations (Labor Relations Act 1995;

Basic Conditions of Employment Act 1997; Employment Equity Act 1998), and

were later followed by programs of affirmative action directed to empower previ-

ously marginalized racial groups (Black Economic Empowerment 2003, 2007).

2.2 The mining industry

South Africa has some of the world’s largest mineral reserves, and is a leading

producer of a range of mineral commodities such as gold, platinum and diamonds

(US Department of Interior, 1996, 2011). In segregated South Africa, the prosperity

of the white minority was intimately linked with a system that generated large

4Several coercive institutions governing labor relations were introduced before the apartheid.

These institutions restricted wages, access to skilled jobs and the right to quit a job by black workers

(Mines and Works Act 1911; Native Labor Regulation Act 1911; Apprenticeship Act 1922). Black

workers could not unionize, nor engage in strike activity (Industrial Conciliations Acts 1924, 1937;

War measure 145, 1942). Coercive institutions were further strengthened by apartheid legislation,

which reserved specific jobs and occupations to specific racial groups and turned the homelands

into pools of cheap migrant labor.

7

profits through the exploitation of black workers, and in tuned used those profits to

secure exploitation, frequently through the police force (Clark and Worger, 2011).

Apartheid legislation preserved and expanded coercive institutions that had long

been a feature of the mining economy, when a powerful network of laws barred black

workers from skilled occupations (Mines and Works Act 1911). Black underground

workers in the gold mining industry were among the most exploited: according

to historical accounts, in 1972 their wages were about 18 Rands per month, less

than the two shillings six pence they had been making in 1902; and while African

miners earned monthly salaries of about 18 Rands, white miners made 400 Rands

per month.5

In 1996, the industry accounted for about 8% of GDP and 43% of exports (US

Department of Interior, 1996). These shares remained fairly similar by 2011 (US

Department of Interior, 2011). During this period, about 80% of mineral output

was destined to foreign markets (US Department of Interior, 1996; 2011). The

bulk of production and exports have been controlled by a relatively small number of

privately-owned mining investment houses (US Department of Interior, 1996). Black

ownership continues to be very limited today, despite recent legislative initiatives

aimed at increasing it, notably under the Black Economic Empowerment program.

From the early 2000s, the mining industry benefited from a favorable evolution of

international prices; for instance, the real price of gold in US dollars more than

tripled between 1996 and 2011. The industry is characterized by the highest degree

of unionization among all sectors–with unionization rates of over 70% both in 1996

and 2011 (see Table 1).

3 A theory of coercion and rent sharing

To guide our empirical analysis, we outline a bargaining model of unionized interna-

tional duopoly, drawing on McDonald and Solow (1981) and Brander and Spencer

(1988).6 We use the model to examine the interplay between coercive institutions

and collective bargaining in determining labor market outcomes in the natural re-

source industry. In addition, we examine if and how international price shocks,

5http://overcomingapartheid.msu.edu/sidebar.php?id=65-258-5&page=36There is a vast theoretical literature on unionized oligopoly, including contributions by David-

son (1988), Dowrick (1989), Mezzetti and Dinopoulos (1991), Naylor (1999), Lommerud et al.

(2003), Lommerud et al. (2006a,b), Bastos and Kreickemeier (2009) and Bastos et al. (2009).

8

driven by exogenous shocks to global demand or foreign marginal costs, might affect

domestic unionized wages in the context of this framework.

3.1 Model setup

Consider a natural resource industry comprising one unionized domestic producer

and one foreign producer, both supplying a homogeneous commodity to a globally

integrated product market. The industry of interest is small relative to the economy

as a whole. There are unspecified barriers facing new firms, and hence product mar-

ket rents are not eroded by entry. Competition is Cournot, implying that domestic

and foreign outputs are strategic substitutes.

The inverse demand function in the industry is given by

p = a− b(y + y∗) (1)

where a, b > 0, y is the output of the domestic firm and y∗ is the output of the

foreign firm.

Labor is the only variable factor of production. The marginal product of labor is

constant, and is normalized to unity so that we can discuss output and employment

interchangeably. Producers must also incur a fixed cost associated with natural

resource extraction.7 The domestic producer is unionized, while the foreign producer

can recruit workers from a competitive labor market. We assume that the labor

union cares both about wages and employment levels, and adopt a Stone-Geary

utility function to represent union preferences:

U = (w − w)θy(1−θ) (2)

where w is the negotiated wage; w is the reservation wage of unionized workers,

defined as what they would earn elsewhere in the labor market; and y is the level

of employment in the unionized sector. The parameter θ ∈ (0, 1) represents the

relative importance of wages over employment for the labor union.8

7This assumption approximates the short-run conditions of an industry operating with excess

capacity of a fixed factor of production (Dowrick, 1989).8Pemberton (1988) shows that (2) can be derived as the maximand of a “managerial union”

with union leaders who are interested in size (employment) and union members (represented by the

median worker) who are interested in excess wages, where θ corresponds to the relative bargaining

power of workers and leadership. Note that θ = 12 corresponds to a rent-maximizing union.

9

Domestic wages and employment levels can be described as the outcome of a two

stage game. In the first stage, the domestic union and the firm bargain over wages

through a Nash bargaining process taking as given the competitive wage abroad.

In the second stage, each firm chooses its output (and hence employment) taking

as given the wage rate and the output of the competitor. We solve by backward

induction.

3.2 Production

Domestic profits are given by π = (p − w)y − F , where F is the fixed cost of

production. The foreign producer can recruit workers from a competitive labor

market at the (exogenous) wage w∗. Hence its profits are given by π∗ = (p− w∗)y∗

−F ∗.9 Profit maximization leads to the reaction functions of each firm:

y =a− w

2b− 1

2y∗ (3)

y∗ =a− w∗

2b− 1

2y (4)

By solving (3) and (4), we may obtain the equilibrium output of each firm for

given wage rates at home and abroad:

y =a− 2w + w∗

3b(5)

y∗ =a− 2w∗ + w

3b(6)

As would be expected, domestic employment levels and profit rates decrease in w

and increase in w∗,while the converse happens with foreign employment and profits.

3.3 Collective bargaining

Assuming zero disagreement payoffs, the generalized Nash product can be expressed

as

N = Uβπ1−β = [(w − w)θy]β[(a− b(y + y∗))y − wy]1−β (7)

9The model could be amended to allow for a fixed cost of production, entering the firms’ profit

function linearly. This assumption would approximate the short-run conditions of an industry

operating with excess capacity of a fixed factor of production (Dowrick, 1989). This would not

make a substantive difference to the results derived.

10

where β ∈ [0, 1] denotes the relative bargaining power of the union. Maximization

of (7) with respect to w yields the following equilibrium negotiated wage:

w = (1− βθ

2− β)w +

βθ

2(2− β)(a+ w∗) (8)

3.4 Dismantling of coercive institutions

The advent of democracy in South Africa was followed by numerous interventions to

level the historically uneven bargaining field. Labor market outcomes were affected

via two channels. First, changes in legislation governing union activity directly cause

an increase in the bargaining power of unionized labor, β. Second, the abolishment of

mobility restrictions for workers in the former homelands, along with the elimination

of the restrictions on the type of jobs they would gain access to, lead to an increase

in the reservation wage of unionized workers, w. Straightforward computations

show that, all else being equal, both these forces contribute to an unambiguous

increases in the negotiated wage and union utility, while decreasing domestic profits

and employment levels.

3.5 International price shocks

Consider now the effects of movements in international prices, which we assume are

driven by exogenous shocks to global demand, a, and/or foreign marginal costs, w∗.

From (8) we observe that:

∂w

∂a=

∂w

∂w∗=

βθ

2(2− β)> 0 (9)

An increase in global demand or a positive shock to foreign marginal costs leads

to a higher international price. This leads to an increase in domestic production and

quasi-rents, and hence to a higher negotiated wage. From (9), we further observe

that these positive effects on union wages are strictly increasing in β. The more

powerful the union, the stronger its ability to benefit from these resource windfalls.

4 Data

For the main analysis in this paper, we build a community-level panel data set

using the community profiles from the 1996 and 2011 population census run by

11

Statistics South Africa. In its original format, the community profiles (at the sub-

place level) provide aggregated category counts for each variable in the census. South

Africa is divided into 9 provinces (equivalent to states), 266 districts and over 21,000

communities (sub-places). Each community has a population of 2,000 individuals

on average, though there is significant heterogeneity. The census includes data

on demographics, labor market (including employment, industry and salary), and

access to infrastructure, among other. Using cartographic data on communities and

former homeland boundaries we identified which communities were located inside

former homelands. In the next section we document the extent to which the data

are consistent with the historical accounts of the apartheid regime.

The geographic community definitions changed from 1996 to 2011. We use official

cartographic data to match communities over time. For the analysis presented in the

paper, we match the 2011 data to the 1996 data by the nearest centroid.10 We allow

for multiple matches, therefore there is not a 1-to-1 correspondence between local

communities. In the Appendix we present results for different matching techniques

and find that the results remain unchanged across definitions. Our final sample is

comprised of almost 16,000 communities, observed both in 1996 and 2011.

Our main variable of interest – growth of income per capita at the community-

level – is constructed by computing the difference in logs of population-weighted

income per capital between 2011 and 1996. Since data on income are grouped in

categories (e.g. no income, 1 to 4800 rand a year, and so on), we take the mid-

point of each category.11 All income values are expressed in December 2012 prices

(obtained from Statistics South Africa). To provide complementary evidence on the

specific mechanisms we emphasize, we also use data from the 10% sample of the

1996 population census.

5 The legacy of the apartheid regime: size and

evolution of local development gaps

We document the evolution of development differentials across communities located

just-inside and just-outside the former homelands. Communities are grouped in

10The centroid of the community is defined as the central point (i.e. latitude and longitude) of

the community polygon.11For example, category 1 to 4800 rand would take the value of 2400.5, and so on. The maximum

category of income is truncated at its value.

12

one-kilometer bins with respect to the linear distance to the former homeland bor-

der. Our aim is to verify the extent to which the data are consistent with numerous

historical accounts pointing to a tragic legacy of the apartheid regime through the

various mechanisms discussed above. The use of these descriptive diagrams focusing

on communities located just-inside and just-outside the former homelands makes it

possible to minimize heterogeneity with respect to geographic and climatic condi-

tions.12

Figure 3 documents the extent to which blacks were geographically segregated

as a result of the homeland system. The large differences observed in 1996 per-

sist after a period of 15 years, though the share of blacks increased in communities

just-outside of the former homelands which points to some migration flows from

homelands to the rest of the country. Similar persistent gaps can be observed in

other dimensions. Figure 4 shows levels of income per capita, where former home-

land communities fare significantly worse both in 1996 and 2011. Figure 5 shows

that, while per capita income grew overall during this period, growth rates were con-

siderably stronger among former homeland communities. This figure also shows that

there has been significant heterogeneity in income growth across former homeland

communities. Our main interest is in evaluating the extent to which the interaction

of the dismantling of coercive institutions and union activity contributed to explain

this heterogeneity in spatial income convergence.

6 Resource rents, coercion, and the path of local

income

In this section we provide visual and econometric evidence that the interplay between

the dismantling of coercive institutions and the distribution of resource rents played

an important role in shaping the development path of marginalized communities

following the end of Apartheid.

12This analysis is purely descriptive. The gaps documented should not be interpreted as the

precise causal impacts of apartheid institutions on local socioeconomic outcomes. While there

is no doubt that apartheid institutions contributed to explain the large development differentials

observed between communities just-inside and just-outside the former homelands, these gaps might

partially reflect other sources of underlying heterogeneity across communities.

13

6.1 Empirical strategy

Let yi,t denote the logarithm of real per capita income for community i in period

t (where t = 1996, 2011). We are interested in the drivers of spatial income con-

vergence among communities marginalized during apartheid, that is, in explaining

∆yi = yi,11 − yi,96.13 To this end, we estimate an equation of the form:

∆yi = α0 + θhomelandi + γminingi,96 + βhomelandi ∗miningi,96

+ φp + λXi,96 + δi + ∆εi (10)

where: homelandi is a dummy variable indicating whether community i belongs

to a homeland; and miningi,96 is the employment share of the mining industry

of that community in 1996, aimed at capturing initial exposure to resource rents.

Our main coefficient of interest, β, measures the difference in the average effect

of a higher initial exposure to resource rents for communities located inside the

former homelands relative to those outside. The error term would be comprised

of δi, capturing unobservable community characteristics, and εi a random error

term. In order to avoid potential biases as a result of the unobservable term δi, we

account for several initial (and current) factors included in matrix Xi,96. We include

initial community characteristics in 1996 such as real per capita income, distance to

border of the homeland, percentage of population with no education, average age,

percentage male, share of white population, share of indian population, percentage

connected to electricity grid, piped water, phone line (land line or cellphone), and

percentage with no refuse. All controls are interacted with homeland indicator to

account for different trajectories inside and outside of the former homelands.

In Figure 6 we present the average share of employment in mining in 1996 for

communities just-inside and just-outside of former homelands. There does not ap-

pear to be systematic differences on shares of employment in mining on either side of

the former homeland border. The main results of the paper are previewed in Figure

7, where we present income growth by share of employment in mining.14 Commu-

nities with a relatively high share of employment in mining which are inside former

homelands experience higher income growth rates. In addition, there does not seem

to be a systematic difference by share in mining outside of former homelands.

13For simplicity, we scale this variable by 15 in order to interpret it as the average yearly growth

rate.14A community is classified as having a high share in mining when the share is above the 90th

percentile.

14

6.2 Main results

To assess the extent to which spatial income convergence was stronger among former

homeland communities with higher initial exposure to resource rents we estimate

(10) by OLS. The resulting estimates are presented in Table 2, where Panel A in-

cludes all communities in South Africa and Panel B restricts the sample to those

communities located within 30 kilometers of the border (on either side).15 We ob-

serve that communities inside the former homelands tended to experience faster

rates of growth than communities outside. The main coefficient of interest, on

the interaction between the homeland indicator and the initial employment share

in mining, consistently suggests that income convergence was faster for homeland

communities with higher exposure to resource rents. These results are robust to

accounting for initial community conditions (in column 2), and including province

or department fixed effects to account for unobserved geographic heterogeneity. The

point estimates for the reduced sample in Panel B are very similar to those found

for the whole country. The estimates from column 4 imply that moving from the

median to the 95th (99th) percentile in share of employment in mining relates to an

additional income growth of around 1.1 (2.4) percentage points yearly.16

Table 3 presents estimates for different samples of communities with regards to

their distance to the border of former homelands from columns 1 through 4. In

column 5 we restrict the sample only to former homeland communities. Regardless

of the sample used, point estimates are similar to those reported earlier.

Even though we are controlling for a number of community level characteristics,

including distance to the homeland border and province (or district) fixed effects,

there may be other underlying unobservable heterogeneity at the community level

driving our results. To better account for any underlying heterogeneity that other-

wise cannot be measured, in Table 4 we use different functions of the distance to the

border of the homeland and different functions of the geographic coordinates (Dell,

2010).17 The results reported in this table show that our findings are robust to the

inclusion of these alternative measures.

As additional robustness, we employ an instrumental variables approach since

there is the possibility that employment in mining may be correlated to other un-

observable confounding factors that can potentially bias our coefficients of interest.

15This is the sample used in the descriptive graphs from the previous sections.16This would be moving from a median of 0.5% to 41.7% (93.8%) for homeland communities.17Function of the latitude and longitude of the centroid of each community.

15

We use the district-level average employment in mining in 1996 (excluding the com-

munity of observation) as a proxy for the proximity of a community to a mine. We

find that the average share of employment in mining is highly relevant in explain-

ing the share of employment in mining in a community (the weighted correlation is

0.686). In Table 5 we present the results for instrumenting community-level average

share of employment in mining and it’s interaction with the homeland dummy by

the district-level share of employment in mining in 1996 and its interaction with the

homeland dummy. The point estimates are qualitatively similar to those from the

estimates in Table 2.

6.3 Complementary evidence from individual-level data

To shed light on the underlying mechanisms driving the interaction between coercive

institutions and rent-sharing we use individual-level data from the 10% sample of

the 1996 South African population census. We restrict the sample to men between

the ages of 15 and 65, working as an employee (i.e. not self-employed or working in

a family business). In Table 6 we present results for regressing the log of individual

wage on indicator variables for black, working in mining industry and their inter-

action.18 In alternative specifications we account for individual attributes such as

age, highest attained education indicators, occupation indicators, and province or

district fixed effects.

Consistent with the presence of resource rents in the mining industry as result

of the high degree of unionization relative to other industries, we find a large wage

premium of around 40 percent associated to employment in this industry. At the

same time, there is a large income gap against black workers of approximately 50

percent - in line with the lasting impacts of coercive institutions in place before the

end of Apartheid. We also observe that black mining workers initially benefited less

from resource rents than non-blacks. In a future version of the paper we will include

similar data from the 2011 census in order to examine how the wage premium for

black mining workers evolved over time.19 The individual-level evidence presented

so far is fully consistent with the mechanisms emphasized in the main analysis of

the paper.

18Note that income is available in categories. We imputed the mid-point of each category and

truncated the highest category at its initial value (that is, 30,001+ rand was assigned value of

30,001).19To the date, this data was not yet released by Statistics South Africa.

16

7 Alternative explanations

Though our results are suggestive of the importance of the dismantling of coercive

institutions and rent-sharing in the mining industry, the heterogeneity in spatial

income across former homeland communities could in part be reflecting alternative

explanations. In this section, we discuss and examine the extent to which these

alternative factors could be driving our results. Importantly, the fact that we use

community-level data in the main analysis makes it possible to account for competing

explanations that operate at this level of aggregation. This would be unfeasible

with individual-level data: since black people are highly concentrated in space, any

factor influencing community-level income that is correlated with employment in

mining might confound individual-level estimates; and since individual-level data

lack precise geographic identifiers, these factors could not be accounted for in the

estimation.

7.1 Access to infrastructure

The period of analysis witnessed numerous government interventions aimed at ad-

dressing the historically uneven access to infrastructure. In line with historical ac-

counts, Figures 8 to 11 document large initial gaps between communities just-inside

and just-outside the former homelands with regard to access to the electric grid,

refuse, land or mobile phone and piped water.20 However, these figures also docu-

ment considerable convergence in access and use of these infrastructure items over

the period of analysis. If improvements in access to infrastructure are correlated

with initial employment in the mining industry, our estimates would be biased.

In order to account for these factors we include the respective infrastructure

variables for the years 1996 and 2011, both interacted with the former homeland

indicator.21 Table 7 presents the results, where in each column we include each

infrastructure variable separately and present the corresponding estimates for the

coefficients of those variables as well as our initial coefficients of interest. Note that

when accounting for changes in access to infrastructure, the point estimates for our

coefficients of interest does not change substantially with regards to our original

20Convergence in mobile phone access is likely the result of market rather than government

intervention since competition and decreasing costs made mobile phones affordable across the

country.21Using the absolute change or the percentage change in access yields similar results.

17

estimates.

When considering the estimates for the correlation between our infrastructure

variables and income growth, we find that our results are consistent with the existing

evidence. These estimates have to be interpreted with caution since there may be

other unobserved heterogeneity related to expansion of infrastructure. Dinkelman

(2011) evaluates the effect of rural electrification in KwaZulu-Natal, a former home-

land, on employment. The study finds negative correlations between electrification

and employment rates for males when estimating by OLS. This is consistent with

our results, where greater access to electricity does not have a positive relation with

income trajectories for former homeland communities.

The estimates for phone access suggests that having higher initial access relates

to higher rates of growth later on. On the other hand, 2011 levels of phone access for

former homeland communities does not appear to have contributed to higher income

growth, as it did in communities outside the former homelands. This may reflect

several explanations, such as unobservable confounding factors biasing coefficient

downwards, or lower effective use of mobile phones due to cost of service (given

the significantly lower levels of income) or higher costs of accessing markets in the

former homelands.

7.2 Market access and connectivity

Since the homelands were spread across the country, their progressive integration

in the South African economy might be expected to depend on their proximity to

markets and connectivity (Harris, 1954; Krugman, 1991; Hanson, 1996; Fujita et al.,

2001; Redding and Sturn, 2008). Failing to account for this mechanisms might bias

our estimates, notably if initial employment in mining were correlated with proximity

to markets and roads of the corresponding communities. Even though we initially

accounted for the distance of the community to the border of the homeland, this

measure does not necessarily capture the effect of the integration of former homeland

communities in the South African economy. Since we are interested in accounting

for the effect of increased access to markets, we calculate the percentage growth

in access to markets using the log-difference between the market access indexes for

after and before the end of apartheid based on Harris (1954) as a proxy. For each

18

community i we compute:

MAAfteri =∑j∈J

Yje−distij (11)

where J is the set of communities such that i 6∈ J , Yj is the aggregate income of

community j in 1996, and distij is the linear distance in kilometers between com-

munity i and j. To calculate the market access index before the end of apartheid

for community i, we only include former apartheid communities in the set J if i

was also a former apartheid community. If i was not a former apartheid commu-

nity, then J would only include communities outside of apartheid areas. Given that

we do not have data on aggregate community level income for before the end of

apartheid, this measure serves as a proxy since homelands were considered indepen-

dent countries and did not have free access to South African markets. Therefore

log(MAAfteri ) − log(MABeforei ) would capture the intensity of the growth in access

to markets. This measure is included in estimation of (10) along with MABeforei to

account for the initial conditions of access to markets.22 Figure 12 describes this

measure for the sample of communities near the former homeland borders. Com-

munities just-inside former homeland areas appear to have experienced a slightly

larger increase in access to markets compared to communities just-outside. Notice-

ably, the variance is substantially larger for communities located just-outside former

homeland areas.

In a similar way, we account for the connectivity of each community, defined

as proximity to main road. Formally, this measure is defined as the inverse of the

linear distance in kilometers of a community to the nearest main road (as defined

by Statistics South Africa). Therefore, communities closer to a main road will have

larger values for this measure.23 Figure 13 describes this measure and shows that

communities just-inside the homelands on average are farther away from the main

roads than communities just-outside.

Table 8 presents regression results that account for the introduction of these

variables. In columns 1 and 2 (5 and 6, for the reduced sample), we introduce the

standardized market access index and the proximity to main road respectively. In-

troducing these variables does not modify the point estimates for our variables of

interest substantially. The estimates for the percentage change in market access

22Excluding this variable or using the posterior exposure to markets does not significantly alter

estimates in the regression.23This measure is standardized.

19

in column 1 suggest that greater expansion in access to markets is associated with

higher income convergence. This positive effect is twice the magnitude for commu-

nities located in former homeland areas. At the same time, proximity to roads by

itself does not seem to be statistically significantly associated to income growth.

The change in access to markets could have affected communities differently

depending on how well connected they are to the transport system. To address this

issue, in column 4 we interact the change in market access with proximity to main

roads. The estimated coefficient on the interaction term indicates that the most

important factor for former homeland communities is the increased access to markets

in itself. Holding increased access constant, being closer to the main road network

does not have a statistically relevant relation to income growth. These results hold

for the reduced sample. Finally, note that the estimates for our variables of interest

remain unchanged throughout.

7.3 Migration

The dismantling of Apartheid was naturally followed by sizable migration flows from

the former homelands. As suggested above, heterogeneity in the demographic and

skill composition of these flows across locations might challenge our interpretation

of the results. In particular, if the least skilled workers from communities with high

initial exposure to resource rents migrate away from former homeland communities,

we may be finding that the increase in average income is driven by changes in com-

position rather than through rent-sharing and dismantling of coercive institutions.

To examine whether this is the case, we estimate (10) using average population

growth as the dependent variable (instead of average income growth).

The results are presented in Table 9. Consistent with the historical accounts,

we find that former homeland communities experienced a significantly lower growth

rate of population, likely driven by migration flows in the years that followed the

advent of democracy. The coefficient on the interaction is consistently similar in

magnitude and of opposite sign to of employment in mining alone. This suggests

that population growth in former homeland communities did not differ significantly

between communities with higher exposure to mining rents. These results seem to

hold for the sub-sample of communities within 30 kilometers of the former homeland

border as presented in Panel B. Even though the change in population appears to

have been similar between communities inside former homeland areas, it could be

20

possible that the skill composition of those migrating is different among communi-

ties. In a future version of this paper we will be able to analyze this in detail with

the individual-level sample for the 2011 census.

Finally, as additional robustness, in Table 10 we present regression results for our

baseline model (with income growth as dependent variable) accounting for changes

in observable attributes of the population and, in columns 4 and 5, for differential

population growth. We observe that accounting for changes in population attributes

or size does not change the point estimates for our variables of interest, suggesting

that the distribution of natural resource rents and the dismantling of an coercive

institutions are driving our results, rather than changes driven by heterogeneity in

population growth.

8 Conclusion

A growing body of literature reveals that current economic underdevelopment is

deeply rooted in specific historical episodes. Less attention has focused on uncov-

ering specific channels by which these negative impacts can be countervailed. In

this paper we have made two main contributions. First, we have quantitatively

documented the tragic legacy of the Apartheid regime, as revealed by the large and

lasting development gaps between communities located just-inside and just-outside

the former Homelands. Second, we have identified the interaction between the dis-

mantling of coercive institutions and the distribution of natural resource rents as an

important channel by which this bad legacy was has been turned into a relatively

more prosperous future – hence the use of the phrasing ”overcoming the tyranny of

history” in this paper’s title.

21

References

[1] Acemoglu, Daron, Simon Johnson and James A. Robinson (2001), “The Colo-

nial Origins of Comparative Development: An Empirical Investigation,” Amer-

ican Economic Review, 91, 1369–1401.

[2] Banerjee, Abhijit, Sebastian Galiani, Jim Levinsohn, Zoe McLaren and Ingrid

Woolard (2008), ”Why has unemployment risen in the new South Africa.”,

Economics of Transition, 16(4)715-740.

[3] Bastos, Paulo and Udo Kreickemeier (2009), ”Unions, competition and inter-

national trade in general equilibrium.”, Journal of International Economics,

79(2): 238–247.

[4] Bastos, Paulo, Udo Kreickemeier and Peter Wright (2009), ”Oligopoly, open

shop unions and trade liberalisation.”, International Journal of Industrial Or-

ganization, 27(6): 679–686

[5] Brander, James and Barbara Spencer (1988), ”Unionized oligopoly and inter-

national trade policy.”, Journal of International Economics, 15: 313–321.

[6] Clark, Nancy L. and William H. Worger (2011), South Africa: The rise and fall

of apartheid, 2nd Edition, Harlow: Pearson Education Limited.

[7] Davidon, Carl (1988), ”Multi-unit bargaining in oligopolistic industries.”, Jour-

nal of Labor Economics, 397–422.

[8] Dell, Melissa (2010), ”The persistent effects of Peru’s mining Mita.”, Econo-

metrica, 78(6): 1863–1903.

[9] Dinkelman, Taryn (2011), ”The effects of rural electrification on employment:

New evidence from South Africa.”, American Economic Review, 101(7), 3078-

3108.

[10] Dowrick, Steve (1989), ”Union-oligopoly bargaining.” Economic Journal, 99:

1123–1142.

[11] Fujita, Masahisa, Paul Krugman and Anthony J. Venables (2001), ”The Spatial

Economy: Cities, Regions, and International Trade,” MIT Press Books, The

MIT Press, edition 1, volume 1, number 0262561476, January.

22

[12] Hall, Robert and Charles Jones (1999), ”Why do some countries produce so

much more output per worker than others?” Quarterly Journal of Economics,

114: 83–116.

[13] Hanson, Gordon H. (1996), ”Localization economies, vertical organization, and

trade.” American Economic Review, 86(5): 1266–78.

[14] Harris, C.D., (1954), ”The market as a factor in the localization of industry

in the United States.” Annals of the Association of American Geographers 44:

315-348.

[15] Krugman, Paul (1991), ”Increasing returns and economic geography.” Journal

of Political Economy, 99(3): 483–99.

[16] Lommerud, Kjell E., Frode Lommerud and Lars Sørgard (2003), ”Unionised

oligopoly, trade liberalisation and location choice.” Economic Journal, 113:

782–800

[17] Lommerud, Kjell E., Frode Meland and Odd Rune Straume (2006), ”Globali-

sation and union opposition to technological change.” Journal of International

Economics, 68: 1–23

[18] Lommerud, Kjell E., Odd Rune Straume and Lars Sørgard (2006), ”National

versus international mergers in unionised oligopoly.” RAND Journal of Eco-

nomics, 37: 212–233

[19] McDonald, Ian M. and Robert Solow (1981), ”Wage bargaining and employ-

ment.”, American Economic Review, 71: 896-908.

[20] Mezzetti, Claudio and Elias Dinopoulos (1991), ”Domestic unionization and

import competition.” Journal of International Economics, 31: 79-100.

[21] Naylor, Robin (1999), ”Union wage strategies and international trade.” Eco-

nomic Journal, 109: 102–125.

[22] Nunn, Nathan (2008), ”The long-term effects of Africa’s slave trades,” Quar-

terly Journal of Economics, 123: 139–176.

[23] Nunn, Nathan (2009), ”The Importance of History for Economic Development,”

Annual Review of Economics, 1: 65-92.

23

[24] Nunn, Nathan (2013), ”Historical Development,” in: Aghion P, Durlauf S. (eds)

Handbook of Economic Growth, volume 2, North-Holland, forthcoming.

[25] Overman, Henry G., Stephen Redding and Anthony J. Venables (2003), ”The

Economic Geography of Trade, Production, and Income: A Survey of Empir-

ics,” in: Choi, E.K. and Harrigan, J. (eds) Handbook of International Trade,

353-387 Blackwell Publishing Ltd

[26] Pemberton, James (1988), ”A ‘managerial’ model of the trade union.” Economic

Journal, 98: 755-771.

[27] Redding, Stephen J., and Daniel M. Sturm (2008). ”The Costs of Remote-

ness: Evidence from German Division and Reunification.” American Economic

Review, 98 (5): 1766–1797.

[28] US Department of Interior (1996), US Geological Survey, Minerals Yearbook:

South Africa.

[29] US Department of Interior (2011), US Geological Survey, Minerals Yearbook:

South Africa.

24

Figure 1: Former homelands of South Africa

Note: Map generated by authors using shape-files by Statistics South Africa and The Directorate:

Public State Land Support.

Figure 2: Current provinces of South Africa

Note: Map generated by authors using shape-files by Statistics South Africa.

25

Figure 3: % Black population

.4.5

.6.7

.8.9

1%

Bla

ck

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

.4.5

.6.7

.8.9

1%

Bla

ck

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

Figure 4: Per capita income

020

4060

8010

0An

nual

per

cap

ita in

com

e (in

100

0's

of ra

nd)

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

020

4060

8010

0An

nual

per

cap

ita in

com

e (in

100

0's

of ra

nd)

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively. Per capita income in December 2012 prices (1 usd

= 8.6 rand). Homeland averages: 4,525 and 11,694 respectively. Non-homeland averages: 20,651 and 46,420 respectively

26

Figure 5: Average growth in per capita income

.04

.06

.08

.1.1

2.1

4Av

erag

e gr

owth

in p

er c

apita

inco

me

-30 -20 -10 0 10 20 30Distance to border in km

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

Figure 6: Share of employment in mining

-.05

0.0

5.1

.15

.2%

em

ploy

ed in

min

ing

-30 -20 -10 0 10 20 30Distance to border in km

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

27

Figure 7: Growth of income by share of employment in mining

0.0

5.1

.15

.2Av

erag

e gr

owth

in p

er c

apita

inco

me

-30 -20 -10 0 10 20 30Distance to border in km

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

28

Figure 8: % Connection to electric grid

0.2

.4.6

.81

% C

onne

cted

to e

lect

ric g

rid

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

0.2

.4.6

.81

% C

onne

cted

to e

lect

ric g

rid

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

Figure 9: % No refuse

0.1

.2.3

.4%

No

refu

se

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

0.1

.2.3

.4%

No

refu

se

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

29

Figure 10: % Land line or mobile phone

0.2

.4.6

.81

% L

and/

Cel

l pho

ne

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

0.2

.4.6

.81

% L

and/

Cel

l pho

ne

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

Figure 11: % Access to piped water (inside or outside household)

0.2

.4.6

.81

% P

iped

wat

er (i

nsid

e or

out

side

hom

e)

-30 -20 -10 0 10 20 30Distance to border in km

A: 1996

0.2

.4.6

.81

% P

iped

wat

er (i

nsid

e or

out

side

hom

e)

-30 -20 -10 0 10 20 30Distance to border in km

B: 2011

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively.

30

Figure 12: Change in access to markets

510

1520

Expa

nsio

n in

mar

ket a

cces

s

-30 -20 -10 0 10 20 30Distance to border in km

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively. The percentage increase in access to markets based

on Harris (1954) in plot.

Figure 13: Proximity to main roads

-.50

.51

1.5

Prox

imity

to m

ain

road

-30 -20 -10 0 10 20 30Distance to border in km

Notes: Each dot represents population-weighted community averages for 1 km bins. Negative distances inside Homelands.

Linear fit and 95% CI represented by line and shaded area, respectively. Proximity to main roads measured as the inverse

distance (in km) to the nearest main road (as defined by Statistics South Africa).

31

Table 1: Unionization rates by industry

% labor force unionized

1996 2011

Mining and quarrying 71.0 74.7

Agriculture, hunting, forestry and fish 7.0 6.1

Manufacturing 42.8 35.9

Electricity, gas and water supply 43.6 46.6

Construction 19.7 10.5

Wholesale and retail trade 24.5 17.5

Transport, storage and communication 45.9 28.6

Financial intermediation, insurance, etc 22.6 21.9

Community, social and personal services 41.1 56.9

Private households 12.1 0.3

Source: Own calculations based on 1996 October Household Survey and

2011-q1 Labor Force Survey.

32

Table 2: OLS estimates for relation between share in employment and average per capita

income growth

(1) (2) (3) (4) (5)

Dep. Var.: Average yearly difference in log per capita income

Panel A: All South Africa

Homeland (=1) 0.041*** 0.216*** 0.030*** 0.219*** 0.163***

[0.001] [0.023] [0.002] [0.023] [0.019]

% Employed mining 1996 -0.015*** -0.021*** -0.015** -0.036*** -0.039***

[0.005] [0.008] [0.006] [0.009] [0.012]

Homeland*% Employed mining 0.103*** 0.049*** 0.103*** 0.062*** 0.061***

[0.015] [0.010] [0.014] [0.010] [0.013]

Observations 15,883 15,852 15,883 15,852 15,852

R-squared 0.2 0.561 0.257 0.576 0.625

Panel B: Within 30km of Homeland border

Homeland (=1) 0.032*** 0.213*** 0.030*** 0.208*** 0.192***

[0.002] [0.031] [0.002] [0.031] [0.030]

% Employed mining 1996 -0.016 -0.040*** 0.015 -0.045*** -0.068***

[0.013] [0.015] [0.019] [0.016] [0.018]

Homeland*% Employed mining 0.067*** 0.071*** 0.056** 0.076*** 0.087***

[0.021] [0.017] [0.024] [0.017] [0.018]

Observations 9,594 9,582 9,594 9,582 9,582

R-squared 0.083 0.561 0.149 0.572 0.622

1996 Controls - Yes - Yes Yes

Province FE (9 provinces) - - Yes Yes Yes

District FE (266 districts) - - - - Yes

Notes: Panel B restricts sample to communities within 30km of the Homeland border, both inside and outside

the Homelands. 1996 controls include: distance to homeland border (in km), per capita income, percentage

of white and percentage of indian population, average age, percentage with no education, percentage of

male population, percentage of households with electricity connection, land or cellphone, no refuse and

piped water connection (either at home or public). All controls are interacted with Homeland indicator.

Regressions are weighted by community total population in 1996. Robust standard errors in parentheses.

***p<0.01, **p<0.05, *p<0.1.

33

Table 3: OLS estimates, robustness to sample selection

(1) (2) (3) (4) (5)

Dep. Var.: Average per capita income growth

Homeland (=1) 0.208*** 0.188*** 0.190*** 0.219***

[0.031] [0.025] [0.023] [0.023]

% Employed mining 1996 -0.045*** -0.030*** -0.029*** -0.036***

[0.016] [0.011] [0.010] [0.009]

Homeland*% Employed mining 0.076*** 0.061*** 0.059*** 0.062*** 0.029***

[0.017] [0.013] [0.011] [0.010] [0.006]

1996 Controls Yes Yes Yes Yes Yes

Province FE (9 provinces) Yes Yes Yes Yes Yes

Sample <30 km <60 km <90 km All SA only HL

Observations 9,582 12,002 13,759 15,852 10,340

R-squared 0.572 0.574 0.588 0.576 0.561

Notes: Columns 1-3 restrict sample to communities within XXkm of the Homeland border, both inside

and outside the Homelands. Column 5 restricts sample to communities inside homelands. 1996 controls

include: distance to homeland border (in km), per capita income, percentage of white and percentage of

indian population, average age, percentage with no education, percentage of male population, percentage

of households with electricity connection, land or cellphone, no refuse and piped water connection (either

at home or public). All controls are interacted with Homeland indicator. Regressions are weighted by

community total population in 1996. Robust standard errors in parentheses. ***p<0.01, **p<0.05, *p<0.1.

34

Tab

le4:

OL

Ses

tim

ates

,ro

bust

ness

tosp

atia

lco

ntro

ls

(1)

(2)

(3)

(4)

(5)

(6)

Dep

.V

ar.

:A

ver

age

per

cap

ita

inco

me

gro

wth

Hom

elan

d(=

1)

0.2

08***

0.2

11***

0.2

09***

0.2

08***

0.2

09***

0.2

14***

[0.0

31]

[0.0

31]

[0.0

31]

[0.0

31]

[0.0

31]

[0.0

30]

%E

mp

loyed

min

ing

1996

-0.0

45***

-0.0

46***

-0.0

46***

-0.0

47***

-0.0

50***

-0.0

55***

[0.0

16]

[0.0

16]

[0.0

16]

[0.0

16]

[0.0

17]

[0.0

17]

Hom

elan

d*%

Em

plo

yed

min

ing

0.0

76***

0.0

77***

0.0

77***

0.0

79***

0.0

80***

0.0

87***

[0.0

17]

[0.0

17]

[0.0

17]

[0.0

18]

[0.0

18]

[0.0

19]

1996

Contr

ols

Yes

Yes

Yes

Yes

Yes

Yes

Pro

vin

ceF

E(9

pro

vin

ces)

Yes

Yes

Yes

Yes

Yes

Yes

f(D

ista

nce

toH

Lb

ord

erin

km

)L

inea

rIn

ter

Cu

bic

Cu

bic

Inte

rC

ub

icB

Sp

lin

eX

,Y,X

Ycu

bic

X,Y

,XY

Sam

ple

<30

km

<30

km

<30

km

<30

km

<30

km

<30

km

Ob

serv

ati

on

s9,5

82

9,5

82

9,5

82

9,5

82

9,5

82

9,5

82

R-s

qu

are

d0.5

72

0.5

72

0.5

72

0.5

73

0.5

75

0.5

80

Note

s:S

am

ple

isre

stri

cted

toco

mm

un

itie

sw

ith

in30km

of

the

hom

elan

db

ord

er,

both

insi

de

an

dou

tsid

e.1996

contr

ols

incl

ud

e:

dis

tan

ceto

hom

elan

db

ord

er(i

nkm

),p

erca

pit

ain

com

e,p

erce

nta

ge

of

wh

ite

an

dp

erce

nta

ge

of

ind

ian

pop

ula

tion

,aver

age

age,

per

centa

ge

wit

hn

oed

uca

tion

,p

erce

nta

ge

ofm

ale

pop

ula

tion

,p

erce

nta

ge

ofh

ou

seh

old

sw

ith

elec

tric

ity

con

nec

tion

,la

nd

or

cellp

hon

e,

no

refu

sean

dp

iped

wate

rco

nn

ecti

on

(eit

her

at

hom

eor

pu

blic)

.A

llco

ntr

ols

are

inte

ract

edw

ith

Hom

elan

din

dic

ato

r.L

inea

rIn

ter:

incl

ud

esin

tera

ctio

nb

etw

een

hom

elan

din

dic

ato

ran

dd

ista

nce

.C

ub

ic(I

nte

r):

incl

ud

esd

ista

nce

squ

are

dan

dcu

bed

(inte

ract

ed

wit

hh

om

elan

d).

X,Y

,XY

:in

clu

des

lati

tud

e,lo

ngit

ud

ean

dth

eir

inte

ract

ion

.C

ub

icX

,Y,X

Y:

incl

ud

esla

titu

de,

lon

git

ud

ean

dth

eir

inte

ract

ion

,in

clu

din

gu

pto

the

3rd

pow

er.

Reg

ress

ion

sare

wei

ghte

dby

com

mu

nit

yto

tal

pop

ula

tion

in1996.

Rob

ust

stan

dard

erro

rsin

pare

nth

eses

.***p<

0.0

1,

**p<

0.0

5,

*p<

0.1

.

35

Table 5: IV estimates for relation between share in employment and average per capita

income growth

(1) (2) (3) (4) (5)

Dep. Var.: Average per capita income growth

Panel A: All South Africa

Homeland (=1) 0.039*** 0.226*** 0.028*** 0.229*** 0.181***

[0.002] [0.021] [0.009] [0.021] [0.020]

% Employed mining 1996 -0.013* -0.021*** -0.011 -0.038*** -0.045**

[0.007] [0.008] [0.009] [0.010] [0.020]

Homeland*% Employed mining 0.143*** 0.076*** 0.150*** 0.089*** 0.120***

[0.028] [0.015] [0.025] [0.016] [0.027]

Observations 15,873 15,842 15,873 15,842 15,842

R-squared 0.197 0.560 0.252 0.575 0.621

Panel B: Within 30km of Homeland border

Homeland (=1) 0.031*** 0.222*** 0.029*** 0.218*** 0.202***

[0.002] [0.030] [0.002] [0.029] [0.030]

% Employed mining 1996 -0.005 -0.038** 0.049* -0.041** -0.105***

[0.017] [0.018] [0.029] [0.019] [0.029]

Homeland*% Employed mining 0.079*** 0.096*** 0.083** 0.108*** 0.138***

[0.030] [0.022] [0.036] [0.023] [0.028]

Observations 9,588 9,576 9,588 9,576 9,576

R-squared 0.082 0.559 0.140 0.569 0.621

1996 Controls - Yes - Yes Yes

Province FE (9 provinces) - - Yes Yes -

District FE (266 districts) - - - - Yes

Notes: Panel B restricts sample to communities within 30km of the Homeland border, both inside

and outside the Homelands. 1996 controls include: distance to homeland border (in km), per capita

income, percentage of white and percentage of indian population, average age, percentage with no

education, percentage of male population, percentage of households with electricity connection, land

or cellphone, no refuse and piped water connection (either at home or public). All controls are

interacted with Homeland indicator. % Employed in mining and it’s interaction with homeland

are instrumented by the district-level average employment in mining in 1996 (excluding itself) and

its interaction with homeland. Regressions are weighted by community total population in 1996.

Robust standard errors in parentheses. ***p<0.01, **p<0.05, *p<0.1.

36

Table 6: OLS estimates, 10% 1996 Census sample

(1) (2) (3) (4)

Dep. Var.: Log(income)

Black (=1) -1.002*** -0.509*** -0.553*** -0.590***

[0.004] [0.003] [0.004] [0.004]

Mining (=1) 0.489*** 0.374*** 0.426*** 0.417***

[0.012] [0.010] [0.010] [0.011]

Black*Mining -0.247*** -0.082*** -0.093*** -0.114***

[0.013] [0.011] [0.011] [0.011]

Individual controls - Yes Yes Yes

Province indicators (9

provinces)

- - Yes -

District indicators (266

districts)

- - - Yes

Observations 378,125 357,240 357,240 357,240

R-squared 0.188 0.490 0.504 0.529

Notes: Sample of males between 15 and 65 years old, employed in

firm/company (not self-employed or family employed). Individual con-

trols include age, age squared, indicators for highest achieved educa-

tion, and occupation indicators. Robust standard errors in parentheses.

***p<0.01, **p<0.05, *p<0.1.

37

Tab

le7:

OL

Ses

tim

ates

,ac

coun

ting

for

chan

ges

inin

fras

truc

ture

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Dep.

Var.

:A

vera

ge

per

capit

ain

com

egro

wth

Infr

ast

ructu

revari

able

:E

lectr

icgri

dN

oR

efu

seP

hone

Pip

ed

wate

rE

lectr

icgri

dN

oR

efu

seP

hone

Pip

ed

wate

r

Hom

ela

nd

(=1)

0.2

02***

0.1

59***

0.2

65***

0.1

59***

0.1

70***

0.1

23***

0.2

65***

0.1

20***

[0.0

12]

[0.0

12]

[0.0

23]

[0.0

13]

[0.0

16]

[0.0

16]

[0.0

31]

[0.0

17]

%E

mplo

yed

min

ing

1996

-0.0

17**

-0.0

19***

-0.0

22***

-0.0

25***

-0.0

26**

-0.0

18

-0.0

27***

-0.0

23**

[0.0

07]

[0.0

06]

[0.0

07]

[0.0

05]

[0.0

12]

[0.0

11]

[0.0

10]

[0.0

12]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.0

35***

0.0

38***

0.0

41***

0.0

44***

0.0

48***

0.0

41***

0.0

50***

0.0

47***

[0.0

08]

[0.0

08]

[0.0

08]

[0.0

07]

[0.0

13]

[0.0

13]

[0.0

12]

[0.0

13]

Infr

ast

ructu

re1996

-0.0

09**

-0.0

05

0.0

35***

-0.0

13

0.0

01

-0.0

02

0.0

20*

-0.0

20***

[0.0

04]

[0.0

09]

[0.0

06]

[0.0

10]

[0.0

05]

[0.0

11]

[0.0

11]

[0.0

07]

Hom

ela

nd*In

frast

ructu

re1996

0.0

02

0.0

18**

0.0

61***

-0.0

13

-0.0

09

0.0

12

0.0

73***

-0.0

04

[0.0

05]

[0.0

09]

[0.0

11]

[0.0

10]

[0.0

05]

[0.0

11]

[0.0

14]

[0.0

07]

Infr

ast

ructu

re2011

0.0

44***

-0.0

66***

0.1

03***

-0.0

06

0.0

52***

-0.0

46***

0.1

45***

-0.0

08

[0.0

08]

[0.0

10]

[0.0

23]

[0.0

09]

[0.0

07]

[0.0

17]

[0.0

28]

[0.0

06]

Hom

ela

nd*In

frast

ructu

re2011

-0.0

53***

0.0

75***

-0.1

01***

-0.0

01

-0.0

61***

0.0

58***

-0.1

41***

0

[0.0

08]

[0.0

11]

[0.0

23]

[0.0

09]

[0.0

07]

[0.0

18]

[0.0

29]

[0.0

06]

1996

Contr

ols

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Pro

vin

ce

FE

(9pro

vin

ces)

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Sam

ple

All

SA

All

SA

All

SA

All

SA

<30

km

<30

km

<30

km

<30

km

Obse

rvati

ons

15,8

19

15,8

19

15,8

18

15,8

19

9,5

61

9,5

61

9,5

61

9,5

61

R-s

quare

d0.5

84

0.5

78

0.5

81

0.5

75

0.5

76

0.5

70.5

72

0.5

7

Note

s:C

olu

mns

5-8

rest

rict

sam

ple

tocom

munit

ies

wit

hin

30km

of

the

Hom

ela

nd

bord

er,

both

insi

de

and

outs

ide

the

Hom

ela

nds.

1996

contr

ols

inclu

de:

dis

tance

tohom

ela

nd

bord

er

(in

km

),p

er

capit

ain

com

e,

perc

enta

ge

of

whit

eand

perc

enta

ge

of

india

np

opula

tion,

avera

ge

age,

perc

enta

ge

wit

hno

educati

on,

perc

enta

ge

of

male

popula

tion,

perc

enta

ge

of

houesh

old

sw

ith

ele

ctr

icit

yconnecti

on,

land

or

cell

phone,

no

refu

seand

pip

ed

wate

rconnecti

on

(eit

her

at

hom

eor

publi

c).

All

contr

ols

are

inte

racte

dw

ith

Hom

ela

nd

indic

ato

r.In

frast

ructu

revari

able

sin

dic

ate

the

perc

enta

ge

of

house

hold

sw

ith

that

serv

ice.

Phone

indic

ate

s

perc

enta

ge

of

house

hold

wit

hla

nd

line

or

cell

phone.

Regre

ssio

ns

are

weig

hte

dby

com

munit

yto

tal

popula

tion

in1996.

Robust

standard

err

ors

inpare

nth

ese

s.

***p

<0.0

1,

**p

<0.0

5,

*p

<0.1

.

38

Tab

le8:

OL

Ses

tim

ates

,ac

cess

tom

arke

tsan

dco

nnec

tivi

ty

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Dep.

Var.

:A

vera

ge

per

capit

ain

com

egro

wth

Hom

ela

nd

(=1)

0.1

93***

0.2

19***

0.1

93***

0.1

932***

0.1

77***

0.2

08***

0.1

77***

0.1

781***

[0.0

23]

[0.0

23]

[0.0

23]

[0.0

227]

[0.0

30]

[0.0

31]

[0.0

30]

[0.0

298]

%E

mplo

yed

min

ing

1996

-0.0

39***

-0.0

36***

-0.0

39***

-0.0

389***

-0.0

55***

-0.0

45***

-0.0

55***

-0.0

552***

[0.0

09]

[0.0

09]

[0.0

09]

[0.0

091]

[0.0

17]

[0.0

16]

[0.0

17]

[0.0

165]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.0

62***

0.0

62***

0.0

62***

0.0

611***

0.0

81***

0.0

76***

0.0

81***

0.0

814***

[0.0

10]

[0.0

10]

[0.0

10]

[0.0

104]

[0.0

18]

[0.0

17]

[0.0

18]

[0.0

175]

%C

hange

inaccess

tom

ark

ets

0.0

01***

0.0

01***

0.0

005***

0.0

01**

0.0

01**

0.0

008***

[0.0

00]

[0.0

00]

[0.0

002]

[0.0

00]

[0.0

00]

[0.0

003]

Hom

ela

nd*%

Change

inaccess

0.0

01**

0.0

01**

0.0

010**

0.0

01***

0.0

01***

0.0

011**

[0.0

00]

[0.0

00]

[0.0

004]

[0.0

00]

[0.0

00]

[0.0

004]

Pro

xim

ity

tom

ain

road

-0.0

02

-0.0

02

0.0

160***

0-0

.001

-0.0

257

[0.0

03]

[0.0

03]

[0.0

051]

[0.0

02]

[0.0

02]

[0.0

277]

Hom

ela

nd*P

roxim

ity

tom

ain

road

0.0

02

0.0

01

-0.0

123**

00.0

01

0.0

294

[0.0

03]

[0.0

03]

[0.0

053]

[0.0

02]

[0.0

02]

[0.0

278]

%C

hange

inaccess

*-0

.0020***

0.0

032

Pro

xim

ity

tom

ain

road

[0.0

007]

[0.0

035]

Hom

ela

nd*%

Change

inaccess

*0.0

016**

-0.0

036

Pro

xim

ity

tom

ain

road

[0.0

007]

[0.0

035]

1996

Contr

ols

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Pro

vin

ce

FE

(9pro

vin

ces)

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Sam

ple

All

SA

All

SA

All

SA

All

SA

<30

km

<30

km

<30

km

<30

km

Obse

rvati

ons

15,8

51

15,8

52

15,8

51

15,8

51

9,5

82

9,5

82

9,5

82

9,5

82

R-s

quare

d0.5

82

0.5

76

0.5

83

0.5

83

0.5

80.5

72

0.5

80.5

8

Note

s:C

olu

mns

5-8

rest

rict

sam

ple

tocom

munit

ies

wit

hin

30km

of

the

Hom

ela

nd

bord

er,

both

insi

de

and

outs

ide

the

Hom

ela

nds.

1996

contr

ols

inclu

de:

dis

tance

tohom

ela

nd

bord

er

(in

km

),p

er

capit

ain

com

e,

perc

enta

ge

of

whit

eand

perc

enta

ge

of

india

np

opula

tion,

avera

ge

age,

perc

enta

ge

wit

hno

educati

on,

perc

enta

ge

of

male

popula

tion,

perc

enta

ge

of

house

hold

sw

ith

ele

ctr

icit

yconnecti

on,

land

or

cell

phone,

no

refu

seand

pip

ed

wate

rconnecti

on

(eit

her

at

hom

eor

publi

c).

All

contr

ols

are

inte

racte

dw

ith

Hom

ela

nd

indic

ato

r.

%C

hange

inA

ccess

tom

ark

ets

=lo

g(M

ark

et

Access

-Aft

er)

-lo

g(M

ark

et

Access

-Befo

re);

where

Mark

et

Access

isbase

don

Harr

is(1

954).

Pro

xim

ity

tom

ain

road

isth

ein

vers

e

of

the

dis

tance

(in

km

)of

com

munit

yto

neare

stm

ain

road

(as

defi

ned

by

Sta

tist

ics

South

Afr

ica).

Regre

ssio

ns

are

weig

hte

dby

com

munit

yto

tal

popula

tion

in1996.

Robust

standard

err

ors

inpare

nth

ese

s.***p

<0.0

1,

**p

<0.0

5,

*p

<0.1

.

39

Tab

le9:

OL

Ses

tim

ates

,ex

plai

ning

popu

lati

ongr

owth

acro

ssSo

uth

Afr

ica

(1)

(2)

(3)

(4)

(5)

(6)

(7)

Dep.

Var.

:A

vera

ge

popula

tion

gro

wth

PanelA

:All

South

Afr

ica

Hom

ela

nd

(=1)

-0.0

10***

-0.1

59**

-0.1

55**

-0.1

49***

-0.0

94

-0.1

33*

-0.0

45

[0.0

03]

[0.0

69]

[0.0

67]

[0.0

55]

[0.0

65]

[0.0

68]

[0.0

64]

%E

mplo

yed

min

ing

1996

-0.1

32**

-0.0

75*

-0.0

49

-0.0

72

-0.0

64***

-0.0

71*

-0.0

71**

[0.0

63]

[0.0

40]

[0.0

36]

[0.0

45]

[0.0

24]

[0.0

40]

[0.0

31]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.1

42**

0.0

94**

0.0

68**

0.0

73*

0.0

84***

0.0

96**

0.0

78**

[0.0

64]

[0.0

40]

[0.0

34]

[0.0

42]

[0.0

25]

[0.0

40]

[0.0

31]

Obse

rvati

ons

15,9

19

15,8

86

15,8

86

15,8

86

15,8

25

15,8

85

15,8

24

R-s

quare

d0.0

36

0.0

91

0.1

05

0.2

37

0.2

50

0.0

95

0.3

40

PanelB:W

ithin

30km

ofH

om

ela

nd

bord

er

Hom

ela

nd

(=1)

-0.0

07**

-0.1

40***

-0.1

25**

-0.1

34***

-0.0

78

-0.1

16**

0.0

28

[0.0

03]

[0.0

48]

[0.0

48]

[0.0

52]

[0.0

81]

[0.0

49]

[0.0

92]

%E

mplo

yed

min

ing

1996

-0.0

42

0.0

20

0.0

17

-0.0

17

-0.0

08

0.0

24

-0.0

44

[0.0

30]

[0.0

27]

[0.0

27]

[0.0

33]

[0.0

24]

[0.0

28]

[0.0

33]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.0

50

-0.0

04

-0.0

09

0.0

10

0.0

24

0.0

00

0.0

47

[0.0

31]

[0.0

28]

[0.0

29]

[0.0

35]

[0.0

26]

[0.0

30]

[0.0

34]

Obse

rvati

ons

9,6

09

9,5

97

9,5

97

9,5

97

9,5

65

9,5

97

9,5

65

R-s

quare

d0.0

03

0.0

67

0.0

76

0.1

44

0.1

49

0.0

75

0.2

25

1996

Contr

ols

-Y

es

Yes

Yes

Yes

Yes

Yes

Pro

vin

ce

FE

(9pro

vic

es)

--

Yes

--

-Y

es

Dis

tric

tF

E(2

66

dis

tric

ts)

--

-Y

es

--

Yes

2011

Contr

ols

--

--

Yes

-Y

es

Mark

et

access

and

pro

xim

ity

tom

ain

roads

--

--

-Y

es

Yes

Note

s:P

anel

Bre

stri

cts

sam

ple

tocom

munit

ies

wit

hin

30km

of

the

Hom

ela

nd

bord

er,

both

insi

de

and

outs

ide

the

Hom

ela

nds.

1996

(and

2011)

contr

ols

inclu

de:

dis

tance

tohom

ela

nd

bord

er

(in

km

),p

er

capit

ain

com

e,

perc

enta

ge

of

whit

eand

perc

enta

ge

of

india

n

popula

tion,

avera

ge

age,

perc

enta

ge

wit

hno

educati

on,

perc

enta

ge

of

male

popula

tion,

perc

enta

ge

of

house

hold

sw

ith

ele

ctr

icit

y

connecti

on,

land

or

cell

phone,

no

refu

seand

pip

ed

wate

rconnecti

on

(eit

her

at

hom

eor

publi

c).

All

contr

ols

are

inte

racte

dw

ith

Hom

ela

nd

indic

ato

r.For

defi

nit

ions

of

mark

et

access

and

pro

xim

ity

toro

ads

refe

rto

Table

7.

Regre

ssio

ns

are

weig

hte

dby

com

munit

y

tota

lp

opula

tion

in1996.

Robust

standard

err

ors

inpare

nth

ese

s.***p

<0.0

1,

**p

<0.0

5,

*p

<0.1

.

40

Table 10: OLS estimates, accounting for alternative explanations

(1) (2) (3) (4) (5)

Dep. Var.: Average per capita income growth

Panel A: All South Africa

Homeland (=1) 0.219*** 0.276*** 0.260*** 0.265*** 0.245***

[0.023] [0.028] [0.028] [0.028] [0.027]

% Employed mining 1996 -0.036*** -0.022*** -0.024*** -0.022*** -0.024***

[0.009] [0.007] [0.007] [0.007] [0.007]

Homeland*% Employed mining 0.062*** 0.052*** 0.050*** 0.050*** 0.048***

[0.010] [0.009] [0.009] [0.009] [0.009]

Observations 15,852 15,800 15,799 15,800 15,799

R-squared 0.576 0.654 0.658 0.66 0.664

Panel B: Within 30km of Homeland border

Homeland (=1) 0.208*** 0.250*** 0.224*** 0.244*** 0.215***

[0.031] [0.036] [0.036] [0.036] [0.035]

% Employed mining 1996 -0.045*** -0.027** -0.034*** -0.027** -0.033***

[0.016] [0.012] [0.013] [0.012] [0.013]

Homeland*% Employed mining 0.076*** 0.059*** 0.061*** 0.058*** 0.058***

[0.017] [0.014] [0.014] [0.014] [0.014]

Observations 9,582 9,553 9,553 9,553 9,553

R-squared 0.572 0.627 0.632 0.634 0.639

1996 Controls Yes Yes Yes Yes Yes

Province FE (9 provinces) Yes Yes Yes Yes Yes

2011 Controls - Yes Yes Yes Yes

Market accecss and Connectivity - - Yes - Yes

Population growth - - - Yes Yes

Notes: Panel B restricts sample to communities within 30km of the Homeland border, both inside

and outside the Homelands. 1996 (and 2011) controls include: distance to homeland border (in km),

per capita income, percentage of white and percentage of indian population, average age, percentage

with no education, percentage of male population, percentage of households with electricity connection,

land or cellphone, no refuse and piped water connection (either at home or public). All controls are

interacted with Homeland indicator. Market access index is standardized, index is based on Harris

(1954). Proximity to main road is the inverse of the distance (in km) of community to nearest main

road (as defined by Statistics South Africa). Regressions are weighted by community total population

in 1996. Robust standard errors in parentheses. ***p<0.01, **p<0.05, *p<0.1.

41

Appendix

A1. Matching communities between 1996 and 2011 census

As explained in Section 4, the geographical definition of communities changed

between the 1996 and 2011 census. In this section we explore how different strategies

of matching communities over time affects our estimates. In our main analysis, we

match 2011 communities to 1996 by nearest centroid. By doing this type of match-

ing, it is possible that many communities from 2011 match to the same community

in 1996. When this is the case, the group of 2011 communities are aggregated into

one unit. Additionally, it is possible that a number of communities from 1996 find

no matches, in which case they are dropped from the sample.

A second matching strategy employs the use of cartographic data for both census.

We calculate the areas of community polygons from 2011 inside 1996 community

polygons. If over 75% of the area of a 2011 community is contained in a single

1996 community, then it is matched to that community. As before, it is possible

that several 2011 communities match to a single 1996 community. Additionally, it

is possible that communities from both years find no match and are dropped. The

third strategy employed is similar to the first, but changing the direction of the

matches (i.e. matching 1996 to 2011 communities).

We estimate baseline model with no controls, and with 1996 controls and province

fixed effects for each community matching strategy. Additionally, for each strategy,

we also use the entire sample and a reduced sample, restricting to unique matches

between communities. The results are presented in Table A.1 below. The first,

second and third strategy described above are denoted Centroid, Area>75% and

Centroid R, respectively. The first two columns correspond to the initial estimates

presented in columns 1 and 4 of Table 2. Noticeably, point estimates do not change

significantly throughout community matching strategies. In addition, they remain

quite stable even when restricting the sample to unique matches between censuses.

This holds for all of South Africa, as well for the restricted sample of communities

within 30 kilometers of the former homeland border.

42

Tab

leA

.1:

OL

Ses

tim

ates

,ro

bust

ness

todi

ffere

ntco

mm

unit

ym

atch

ing

defin

itio

ns

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Dep.

Var.

:A

vera

ge

per

capit

ain

com

egro

wth

PanelA

:All

South

Afr

ica

Hom

ela

nd

(=1)

0.0

41***

0.2

19***

0.0

42***

0.2

25***

0.0

39***

0.2

59***

0.0

42***

0.2

57***

0.0

37***

0.2

15***

0.0

41***

0.2

43***

[0.0

01]

[0.0

23]

[0.0

02]

[0.0

22]

[0.0

02]

[0.0

28]

[0.0

02]

[0.0

29]

[0.0

02]

[0.0

22]

[0.0

02]

[0.0

28]

%E

mplo

yed

min

ing

1996

-0.0

15***

-0.0

36***

-0.0

17***

-0.0

30***

-0.0

34***

-0.0

57***

-0.0

20**

-0.0

42***

-0.0

23***

-0.0

37***

-0.0

24***

-0.0

49***

[0.0

05]

[0.0

09]

[0.0

05]

[0.0

09]

[0.0

11]

[0.0

12]

[0.0

09]

[0.0

12]

[0.0

06]

[0.0

09]

[0.0

07]

[0.0

12]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.1

03***

0.0

62***

0.1

29***

0.0

46***

0.1

33***

0.0

78***

0.1

29***

0.0

61***

0.1

11***

0.0

62***

0.1

16***

0.0

73***

[0.0

15]

[0.0

10]

[0.0

11]

[0.0

11]

[0.0

16]

[0.0

14]

[0.0

14]

[0.0

14]

[0.0

17]

[0.0

11]

[0.0

13]

[0.0

13]

Obse

rvati

ons

15,8

83

15,8

52

12,5

31

12,5

08

11,5

36

11,5

22

9,8

13

9,8

00

15,9

03

15,8

80

12,6

60

12,6

38

R-s

quare

d0.2

00

0.5

76

0.2

18

0.5

84

0.2

01

0.5

41

0.2

06

0.5

67

0.1

74

0.5

25

0.1

97

0.5

49

PanelB:W

ithin

30km

ofH

om

ela

nd

bord

er

Hom

ela

nd

(=1)

0.0

32***

0.2

08***

0.0

32***

0.2

06***

0.0

30***

0.2

50***

0.0

34***

0.1

83***

0.0

26***

0.2

13***

0.0

29***

0.2

63***

[0.0

02]

[0.0

31]

[0.0

02]

[0.0

33]

[0.0

02]

[0.0

36]

[0.0

02]

[0.0

39]

[0.0

02]

[0.0

34]

[0.0

02]

[0.0

30]

%E

mplo

yed

min

ing

1996

-0.0

16

-0.0

45***

-0.0

21*

-0.0

42**

-0.0

33**

-0.0

60***

-0.0

29

-0.0

47**

-0.0

27**

-0.0

53***

-0.0

34**

-0.0

88***

[0.0

13]

[0.0

16]

[0.0

13]

[0.0

17]

[0.0

15]

[0.0

20]

[0.0

18]

[0.0

21]

[0.0

12]

[0.0

18]

[0.0

14]

[0.0

20]

Hom

ela

nd*%

Em

plo

yed

min

ing

0.0

67***

0.0

76***

0.0

99***

0.0

62***

0.0

98***

0.0

79***

0.1

08***

0.0

69***

0.0

74***

0.0

83***

0.0

87***

0.1

16***

[0.0

21]

[0.0

17]

[0.0

17]

[0.0

19]

[0.0

20]

[0.0

22]

[0.0

23]

[0.0

23]

[0.0

21]

[0.0

19]

[0.0

20]

[0.0

21]

Obse

rvati

ons

9,5

94

9,5

82

7,6

18

7,6

07

7,0

99

7,0

94

6,0

77

6,0

72

9,6

16

9,6

07

7,7

12

7,7

03

R-s

quare

d0.0

83

0.5

72

0.0

93

0.5

75

0.0

87

0.5

39

0.0

90

0.5

56

0.0

58

0.5

22

0.0

70

0.5

38

1996

Contr

ols

-Y

es

-Y

es

-Y

es

-Y

es

-Y

es

-Y

es

Pro

vin

ce

FE

-Y

es

-Y

es

-Y

es

-Y

es

-Y

es

-Y

es

Geogra

phic

al

matc

hty

pe

Centr

oid

Centr

oid

Centr

oid

Centr

oid

Are

a>

75%

Are

a>

75%

Are

a>

75%

Are

a>

75%

Centr

oid

RC

entr

oid

RC

entr

oid

RC

entr

oid

R

Oth

er

rest

ricti

ons

--

Uniq

ue

Uniq

ue

--

Uniq

ue

Uniq

ue

--

Uniq

ue

Uniq

ue

Note

s:P

anel

Bre

stri

ct

sam

ple

tocom

munit

ies

wit

hin

30km

of

the

Hom

ela

nd

bord

er,

both

insi

de

and

outs

ide

the

Hom

ela

nds.

1996

contr

ols

inclu

de:

dis

tance

tohom

ela

nd

bord

er

(in

km

),p

er

capit

ain

com

e,

perc

enta

ge

of

whit

eand

perc

enta

ge

of

india

np

opula

tion,

avera

ge

age,

perc

enta

ge

wit

hno

educati

on,

perc

enta

ge

of

male

popula

tion,

perc

enta

ge

of

house

hold

sw

ith

ele

ctr

icit

yconnecti

on,

land

or

cell

phone,

no

refu

se

and

pip

ed

wate

rconnecti

on

(eit

her

at

hom

eor

publi

c).

All

contr

ols

are

inte

racte

dw

ith

Hom

ela

nd

indic

ato

r.R

egre

ssio

ns

are

weig

hte

dby

com

munit

yto

tal

popula

tion

in1996.

Defi

nit

ions

for

Geogra

phic

al

matc

h

typ

e:

(1)

Centr

oid

:each

com

munit

yfr

om

2011

ism

atc

hed

toth

eneare

st1996

com

munit

ycentr

oid

.(2

)A

rea>

75%

:If

over

75%

of

the

are

aof

a2011

com

munit

yis

conta

ined

ina

single

1996

com

munit

y,

these

are

matc

hed

(note

that

severa

lcom

munit

ies

from

2011

can

matc

ha

single

1996

com

munit

y).

(3)

Centr

oid

R:

like

(1)

but

changin

gdir

ecti

on

of

the

matc

h,

that

is,

matc

hin

g1996

to2011.

Uniq

ue:

rest

rict

sam

ple

to

one-t

o-o

ne

matc

hes

only

.R

obust

standard

err

ors

inpare

nth

ese

s.***p

<0.0

1,

**p

<0.0

5,

*p

<0.1

.

43