The current issue and full text archive of this journal is available on Emerald Insight at:

www.emeraldinsight.com/0144-3585.htm

Digital politics
Digital politics: internet and
democracy in Africa
Olaniyi Evans
School of Management and Social Sciences,
Pan-Atlantic University, Lagos, Nigeria
169
Received 20 August 2017
Abstract Revised 12 January 2018
20 April 2018
Purpose – The purpose of this paper is to investigate the relationship between internet use and democracy in Accepted 22 April 2018
Africa. It examines the non-linearities and causality between the two variables in the short and long run for
38 countries in Africa.
Design/methodology/approach – The study is empirical. It uses pooled mean group and causality tests
for the sample of 38 African countries.
Findings – The panel long-run and short-run estimates show evidence of significant non-linear relationship
between internet usage and democracy. While internet usage is significantly and negatively related to
democracy, squared internet usage is significantly but positively related. This suggests that internet usage
increases with the decrease of democracy, but after a certain level of internet usage which is the turning point,
democracy starts to increase. Additionally, there is uni-directional causality from internet usage to
democracy. However, a bi-directional causality exists between squared internet usage and democracy.
Research limitations/implications – The empirical evidence from this study suggests that internet usage
and democracy are highly interrelated to each other in Africa. The findings support that at the macro level,
Africa is moving toward a new stage, where internet will lead to improved levels of democracy and
digital politics.
Practical implications – Remarkably, the paper shows that democracy displays a quadratic relationship
with internet usage. As a whole, the findings indicate a U-shaped pattern: democracy decreases with internet
usage, stabilizes, and then increases. In other words, internet usage increases with the decrease of democracy,
but after a certain level of internet usage which is the turning point, democracy starts to increase.
Social implications – Many African Governments that have frequently imposed restrictions on internet
and social media need to stop. The decline in democracy as internet usage increases may be explained by
more severity of these restrictions. However, the findings support that at the macro level, Africa is moving
toward a new stage, where internet will lead to improved levels of democracy and digital politics.
Originality/value – Contrary to previous conceptual papers, the current study empirically investigates the
causality between internet and democracy in 38 African countries. The findings indicate a U-shaped pattern:
democracy decreases with internet usage, stabilizes, and then increases. In other words, internet usage
increases with the decrease of democracy but after a certain level of internet usage which is the turning point,
democracy starts to increase.
Keywords Internet, Democracy, Causality, Pooled mean group, Digital politics
Paper type Research paper

1. Introduction
Digital politics is shorthand for how internet technologies have heightened the complex
interactions between political actors and their constituencies (Vaccari, 2013; Miller, 2016).
It describes how the digital revolution is disrupting politics and transforming political
institutions and modes of political communication (Ferdinand, 2000; Vaccari, 2013;
Miller, 2016). In Africa, the digital revolution has gathered momentum. Average citizens are
increasingly and actively participating in online public discourses. In Uganda, a website
called Yogera, or “speak out,” provides a platform for the citizens to complain about poor
service, scrutinize the government or blow the whistle on corruption. In Kenya, a website
called Mzalendo or the “Eye on the Kenyan Parliament”, profiles politicians, scrutinizes
public expenditure and highlights citizens’ rights (Ridgwell, 2017). Journal of Economic Studies
Vol. 46 No. 1, 2019
pp. 169-191
The author of this paper has not made their research data set openly available. Any enquiries © Emerald Publishing Limited
0144-3585
regarding the data set can be directed to the corresponding author. DOI 10.1108/JES-08-2017-0234
JES As obtainable in established democracies such as the USA, the UK and Germany, the
46,1 internet is becoming a force for increased democracy and for the spread of human rights in
Africa (Ferdinand, 2000; Evans, 2018a). However, the internet as a platform for active
participation in public discourse is challenged in the continent. Many African Governments
have frequently imposed restrictions on social media and internet access during elections,
supposedly to prevent “hate speech” and violent elements stoking unrest. However, this has
170 been often interpreted as an attempt by autocratic politicians to silence pro-democracy
movements and control the flow of information.
A puzzle, therefore, ensues: while the internet aids active political engagement, many
African Governments nevertheless continue to impose restrictions that undermine
democracy. This puzzle has given rise to a stream of studies investigating the
relationship between internet and democracy. While these studies have increased in
number, there are a limited number of empirical attempts to verify the relationship (Best and
Wade, 2009). Indeed, the majority of studies in the literature are case studies and theoretical
analyses (e.g. Adibe et al., 2017; Mustapha, 2017; Ruijgrok, 2017). Many of the few previous
studies that approached the issue of internet and democracy with empirical analysis are
from developed countries (e.g. Best and Wade, 2009; Grofman et al., 2014; Nam, 2017;
Pirannejad, 2017; Ruijgrok, 2017; Shen, 2017).
Even though there is growing interest in empirical analysis to pinpoint possible
consistent causal relations between internet and democracy, the exact nature of the link has
not been clarified so far for Africa. Contrary to the previous conceptual papers that
examined internet and democracy in Africa, the current study empirically investigates the
causality between internet and democracy using the Economist Intelligence Unit (EIU)
democracy index and individuals using the internet (percent of population) as proxies of
democracy and internet respectively. The study uses pooled mean group (PMG) and
Granger causality for a panel of 38 African countries. Methodologically, studying a sample
of 38 countries is superior to a single-country analysis, particularly considering the high
degree of structural features and socio-political parallels between African countries.
To the best knowledge of the author, no empirical study has investigated the relationship
between internet and democracy in Africa, especially one covering as many as a panel of
38 African countries. This study, therefore fills the gap, by investigating the causality between
internet usage and democracy using PMG and Granger causality. There is no doubt that
rigorous empirical analysis of this nature will benefit scholars and policymakers in Africa.
The outline of this paper is as follows. Section 2 briefly reviews relevant conceptual,
theoretical and empirical studies on the internet and democracy relationship in both
developed and developing countries. Section 3 dwells on the evolution of internet and
democracy in Africa. Section 4 introduces the empirical methodology. Section 5 presents the
empirical results. Section 5 concludes with policy implications.

2. Literature review
There is a burgeoning literature on the effects of information and communication
technologies on democratic processes. The scope and span of the debate is impressive,
largely due to the historic implications of the information age and the complexity of
democratic governance (Weare, 2002). The study of the relationship between internet and
democracy has engaged many researchers using different approaches from conceptual and
theoretical (Weare, 2002; Margetts, 2013; Grofman et al., 2014; Adibe et al., 2017; Mustapha,
2017; Ruijgrok, 2017) to empirical (Best and Wade, 2009; Nisbet et al., 2012; Salgado, 2012;
Lee, 2017; Nam, 2017; Pirannejad, 2017; Shen, 2017; Evans, 2018a).
Many important studies in the literature have emanated from developed countries. For
example, Best and Wade (2009) examined the global effect of the internet on democracy by
observing the relationships between measures related to internet prevalence and democracy
over the period of 1992–2002. Their findings suggest that internet usage is an insignificant Digital politics
predictor of democracy for the 1992–2002 period, but a stronger predictor for the 2001–2002
period. The authors hypothesized that the jump in the ability of internet penetration to
explain variation in democratization in 2000s suggests that the internet is increasingly
becoming a positive force for democratization. However, some regions have insignificant
positive internet/democracy correlation. In the same vein, Pirannejad (2017), using a panel of
122 countries for the period 2000–2014, examined the impact of internet extension on 171
democracy promotion employing a dynamic panel data model. The author found that
internet extension has a significantly positive impact on democracy promotion.
Ruijgrok (2017) investigated the relationship between internet usage and protests in
democracies and authoritarian states. Developing a theoretical rationale substantiated with
anecdotal evidence from the Tunisian revolution (2010–2011), the author argued that unlike
in democracies, internet usage facilitates protests in authoritarian regimes. Internet usage
increases the number of protests in authoritarian states and this effect remains robust
across a number of model specifications. While information already flows freely in
democracies, internet usage has increased information flow in authoritarian regimes in spite
of authoritarian endeavors to gag cyberspace. Also, Ruijgrok (2017) suggested that the
increased information flow positively affects protests in authoritarian states through four
complementary causal pathways: “(1) by reducing the communication costs for oppositional
movements; (2) by instigating attitudinal change; (3) decreasing the informational
uncertainty for potential protesters; and (4) through the mobilizing effect of the spread of
dramatic videos and images” (p. 498).
In a related study, Nam (2017) examined whether the internet is a tool for oppression or
democracy. Analyzing the relations between national level on a global scale (national
economy, human capital, civil liberties, electoral rights, ICT deployment and online filtering),
the author found that “internet-enabled participation did not significantly mediate political
evolution from electoral democracy to liberal democracy. On the contrary, online filtering or
censorship significantly mediated the relationship between electoral democracy and liberal
democracy” (p. 538). Moreover, the study identified the countries that have experienced
inconsistencies between their degree of civil oppression and inclination for internet-enabled
participation by categorizing their degrees of liberal democracy and internet-enabled
participation. Similarly, Shen (2017), using the Internet Society’s Global Internet Usage
Survey data, examined public demand for internet freedom and control along with their
macro-societal and micro-individual predictors. The author found that internet usage is a
positive predictor of demand for internet freedom, but not of demand for internet control.
The amount of internet freedom in a given country, along with individual perception of
freedom supply, was found to be negatively related to people’s demand for both internet
control and internet freedom. The study also suggested that the effect of internet usage on
demand for internet control and freedom is dependent on perceived freedom supply in
individual countries.
Another stream of the literature is conducted in emerging economies. For example,
Salgado (2012) examined whether and how the internet is being used as a democratization
tool in emergent democracies. The author examined the media’s impact on democratization
and the potentials of the internet in Mozambique and Angola. The author showed that,
despite the limited access, the internet, through online newspapers, information websites,
blogs, etc., has strengthened civil society in various contexts: promotion of participation and
discussion, motivation of different actors in the public sphere, dissemination of media
contents, the creation of new independent media and influencing journalists, who employ
the internet to look for new opinions. Further, Lee (2017) examined the effects of internet
usage on political participation in East and Southeast Asia using the Asian Barometer
Survey for the two periods 2005–2007 and 2010–2011. The author showed that internet
JES usage is positively related to unconventional political participation, but unrelated to
46,1 institutionalized action. Also, internet usage is found to provide mobilizing capacities for
organizational membership and political action.
Recently, there are a growing number of studies from Africa. For example, Ott (1998)
analyzed the role of electronic media in promoting the formation of democratic political
regimes in Africa. The author argued that access to electronic information can have a
172 positive impact on democracy in Africa, by providing civil society with greater leverage
vis-à-vis the state and political elites. In a similar study, Nisbet et al. (2012), using a multilevel
model for public opinion data from Africa and Asia, examined the relation between
individual internet usage, national internet usage and citizen demand for democracy for
28 countries. The authors found that internet usage, rather than national internet usage, is
positively related to higher citizen commitment to democratic governance. Additionally,
democratization and internet usage is found to moderate the relation between internet usage
and demand for democracy. Adibe et al. (2017) assessed the Cybercrime Act 2015 and its
effects on online press freedom in the Nigerian liberal authoritarian state. Examining how
Nigerian political leadership wrongfully applied the act to undermine press independence,
the authors found that Nigeria’s online press freedom has steadily worsened since the 2015
Cybercrime Act.
Overall, studies on internet and democracy have increased in number. Yet there are a
limited number of empirical attempts to verify the relationship (Best and Wade, 2009).
Indeed, the majority of studies in the literature are case studies and theoretical analyses
(e.g. Adibe et al., 2017; Mustapha, 2017; Ruijgrok, 2017) and the few empirical studies on
internet and democracy are from developed countries (e.g. Best and Wade, 2009; Grofman
et al., 2014; Nam, 2017; Pirannejad, 2017; Ruijgrok, 2017; Shen, 2017). Contrary to the
previous conceptual papers that examined internet and democracy in Africa, the current
study empirically investigates the causality between internet and democracy using PMG
and Granger causality for a panel of 38 African countries.

3. Stylized facts
3.1 Internet usage
Internet usage had increased dramatically in many parts of Africa (Evans, 2018b, c). The
number of unique mobile phone subscribers in the continent had reached half a billion in
2015. In total, 25 percent of the half a billion subscribers had access to a mobile broadband
(3 G/4 G) connection. This percentage had grown rapidly as a result of network rollouts and
the increasing access to affordable mobile broadband devices and tariffs.
Particularly in West Africa, internet usage increased rapidly over the past years,
as demand for faster data networks became increasingly higher (Evans, 2018d).
Figure 1 juxtaposes the levels of internet usage in West Africa in 2006 and 2015. Between
2006 and 2015, internet usage had increased dramatically in the region. In 2015, Cote
D’ivoire, Gambia, Ghana, Senegal and Nigeria were the best-performing West African
countries. Especially in Nigeria, almost one out of two people were connected.
Ghana already had more mobile broadband subscribers per 100 inhabitants than Belgium
or Canada.
Like West Africa, internet usage in East Africa also increased strongly during the
ten-year period (Figure 2), reflecting increases in both individuals using internet and
internet usage in the region. The countries that enjoyed the highest internet usage in this
region are Kenya, Mauritius and Seychelles. Particularly in Kenya, over 46 percent of the
population have access to the internet. This is allegedly as a result of the Kenyan
Government’s efforts in ensuring internet access to all 47 counties of Kenya. The 46 percent
penetration rate places Kenya at par with Russia and far ahead of high- and middle-income
countries such as Italy, Portugal and China.
 (a) Digital politics
25.00 23.48
21.00
20.00
17.12

15.00
173
10.00
6.79
5.24 5.90
4.70
5.00
2.72
1.54 1.52
0.64 0.57
0.00
Benin Benin Côte Côte Gambia Gambia Ghana Ghana Guinea Guinea Liberia Liberia
d’Ivoire d’Ivoire
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

(b)
50.00 47.44
45.00
40.00
35.00
30.00
25.00 21.69
20.00
15.20
15.00 10.34
10.00 5.55 5.61 7.12
5.00 0.98 2.22 2.50 2.00
0.73 0.29 0.23
0.00
Mali

Mali

Mauritania

Mauritania

Nigeria

Nigeria

Senegal

Senegal

Sierra Leone

Sierra Leone

Togo

Togo
Niger

Niger

Figure 1.
Internet usage in
West Africa (percent
of the population)
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

Similarly, in Central Africa, internet usage increased intensely over the ten-year period
(Figure 3). The best-performing countries are Gabon, Sao Tome, Equatorial Guinea, and
Cameroon. The worst-performing countries in the region are Central African Republic and
Democratic Republic of Congo. In some of these countries, poor internet connectivity and the
cost of the internet kept most people disconnected.
In Southern Africa, internet usage had been on the rise too (Figure 4). Countries such as
Botswana and South Africa enjoyed the highest internet usage in the region. However, poor
internet connectivity in the rural areas and the exorbitant prices of internet-enabled cellular
phones kept the poor disconnected. In fact, most schools and offices in the rural areas, where
the majority of people lived, had no internet provisions.
Table I shows the best-performing and the worst-performing countries across Africa in
terms of internet usage in 2015. Overall, the countries with the lowest score (worst) in Africa
include Benin, Liberia, Tanzania, Guinea and Central African Republic (Table I). The countries
with the highest score (best) include Seychelles, South Africa, Mauritius, Nigeria and Kenya.
In between these two extremes, there are a wide range of countries doing well, and not so well.
JES (a)
46,1 60.00
50.14
50.00 45.62

40.00

30.00
174
20.00 16.70
11.60
7.53 9.30 9.00
10.00 4.17
0.31 0.61 0.43 0.84
0.00
Ethiopia

Ethiopia

Kenya

Kenya

Madagascar

Madagascar

Malawi

Malawi

Mauritius

Mauritius

Mozambique

Mozambique
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

(b)
70.00
58.12
60.00

50.00

40.00 34.95

30.00 26.61
19.22 21.00
18.00 16.36
20.00

10.00 5.36 4.16

2.12 2.14 1.30 2.53 2.40
0.00
Rwanda

Rwanda

Seychelles

Seychelles

Sudan

Sudan

Tanzania

Tanzania

Uganda

Uganda

Zambia

Zambia

Zimbabwe

Zimbabwe
Figure 2.
Internet usage in East
Africa (percent of the
population)
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

30.00

25.00

20.00

15.00

10.00

5.00

0.00
Angola

Angola

Cameroon

Cameroon

Central African Republic

Central African Republic

Chad

Chad

Congo Democratic

Congo Democratic

Congo Republic

Congo Republic

Equitorial Guinea

Equitorial Guinea

Gabon

Gabon

Sao Tome

Sao Tome

Figure 3.
Internet usage in
Central Africa (percent
of the population)
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015
 60.00 Digital politics
50.00

40.00

30.00
175
20.00

10.00 Figure 4.
Internet usage in
0.00 Southern Africa
Botswana Botswana Namibia Namibia South Africa South Africa (percent of the
population)
2006 2015 2006 2015 2006 2015

Overall, internet usage in Africa has been gathering pace over the last decade, as the
continent continues to close the digital gap with the western world. Despite the huge
progress in internet usage across Africa, there is still a huge gap to be filled. For example,
Africa remains the least-connected continent in the world. Africa’s broadband household
penetration was 15.4 percent in 2015, far behind the world average of 52.3 percent. Asia and
the Pacific is 46.4 percent. The Americas is 64.4 percent. Europe is 84 percent. However,
several African countries (e.g. Nigeria, South Africa and Kenya) are on the right path, with
heavy investment in infrastructure in the ICT sector improving the internet.

3.2 Democracy
Even though Africa’s standard of governance is almost unimaginably better than it was at
the end of the Cold War, democracy is still struggling to take root in many parts of the
continent. Overall, Africa’s democracy shows steady progress, despite hiccups. Figure 5
shows the levels of democracy in West African countries for 2006 and 2015. In 2015, the
best-performing countries include Benin, Liberia, Mali and Senegal, while Togo, Nigeria,
Mauritania and Guinea were the worst-performing. Once the center of Africa’s gory tyrants,
self-serving military juntas, anarchic one-party systems, brutal dictatorships and hideous
civil wars, West Africa is changing as multi-party elections are sweeping across the once
politically volatile region. Despite West Africa’s murky political history, elections and
peaceful changes of government were becoming the norm. Some of the region’s enduring
wars, like those in Liberia and Sierra Leone, had ended, and democratic elections brought in

Best-performing countries Worst-performing countries

Seychelles 58.12 Benin 6.79

South Africa 51.92 Liberia 5.90
Mauritius 50.14 Tanzania 5.36
Nigeria 47.44 Guinea 4.70
Kenya 45.62 Central African Republic 4.56
Botswana 27.50 Madagascar 4.17
Sudan 26.61 Congo Democratic 3.80 Table I.
Sao Tome 25.82 Chad 2.70 Internet usage in
Gabon 23.50 Sierra Leone 2.50 Africa (percent of
Ghana 23.48 Niger 2.22 the population)
JES (a)
46,1 0.60 0.56 0.56
0.52
0.50
0.49
0.50

0.40 0.38 0.37

0.35

176 0.30
0.29
0.25
0.21 0.23
0.20

0.10

0.00
Benin Benin Côte Côte Gambia Gambia Ghana Ghana Guinea Guinea Liberia Liberia
d’Ivoire d’Ivoire

2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

(b)
0.70
0.60
0.60
0.52
0.47 0.49
0.50

0.40 0.35 0.35 0.36 0.35

0.31 0.32
0.30
0.30 0.26 0.24
0.21
0.20

0.10

0.00
Mali

Mali

Mauritania

Mauritania

Niger

Niger

Nigeria

Nigeria

Senegal

Senegal

Sierra Leone

Sierra Leone

Togo

Togo
Figure 5.
Democracy in
West Africa
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

capable democratic governments. For example, in Nigeria, 2015 was the first year an
opposition party would gain power through elections.
However, countries such as Guinea-Bissau were an exception to the overall trend of
democratization in the region: the country underwent five military coups in recent decade,
as no elected president completed his term. Moreover, violent extremist organizations, such
as Boko Haram, are still displacing millions and reversing social progress in broad parts of
the Lake Chad Basin (Evans and Kelikume, 2018). Political competition is always the major
flashpoint for conflict, with elections often igniting strife and violence. Finally, economic and
social exclusion, scarce natural resources, ethnic divisions, herder/farmer differences are all
drivers of conflict to various degrees (Evans and Alenoghena, 2015; Adeola et al., 2017;
Evans and Kelikume, 2019).
In particular, East Africa is ranked as the most democratic region on the continent.
Figure 6 shows the levels of democracy in 2006 and 2015 in East Africa. In 2015, the
best-performing countries were Mauritius, Seychelles, Tanzania, Uganda and Zambia.
The worst-performing were Ethiopia, Kenya, Sudan, Rwanda and Zimbabwe. However,
all East African countries are democracies; they hold regular elections, whether
marred by violence, exclusion of candidates or voters and allegations of vote-rigging.
 (a) Digital politics
0.80 0.75

0.70 0.66

0.60 0.55
0.51
0.50 0.46 0.47
0.42 0.44 0.42
0.36 0.38
0.40
0.30
0.32
177
0.20
0.10
0.00
Ethiopia

Ethiopia

Kenya

Kenya

Madagascar

Madagascar

Malawi

Malawi

Mauritius

Mauritius

Mozambique

Mozambique
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

(b)
0.70
0.63
0.60 0.56
0.50
0.47 0.48
0.50 0.45 0.44 0.45

0.40
0.32
0.30 0.25
0.22
0.18 0.16
0.20
0.12
0.10

0.00
Rwanda

Rwanda

Seychelles

Seychelles

Sudan

Sudan

Tanzania

Tanzania

Uganda

Uganda

Zambia

Zambia

Zimbabwe

Zimbabwe

Figure 6.
Democracy in
Eastern Africa
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

Citizens eagerly await and participate in these elections, even if often disappointed from a
credibility angle. In Uganda, the ruling National Resistance Movement has been in power
since 1986. The leader of the opposition has often been subject to frequent arrests and
assaults by the police. In Tanzania, the Independence party Chama Cha Mapinduzi
continues with its half a century-long dominance, even though majority of citizens reject a
single-party system.
Figure 7 shows the levels of democracy in 2006 and 2015 in Central Africa. In 2015,
countries such as Sao Tome and Cameroon were the best-performing in the region. Unlike
the neighboring West Africa where many countries can reasonably be described as
democracies, Central Africa has experienced little progress in terms of human rights or
democracy. Central Africa has some of the continent’s longest-serving presidents, including
Equatorial Guinea’s Teodoro Obiang Nguema Mbasogo and Cameroon’s Paul Biya who
have both been in power for more than 35 years; and Chad’s Idriss Derby, for 25 years
(Smith, 2016). In most of the countries, political opposition is mainly fragmented, civil
society weak, and news media fragile.
JES 0.60
0.50
46,1 0.50
0.38
0.44
0.40 0.34
0.29 0.29 0.28
0.30 0.25 0.24 0.26
0.21 0.21
0.20 0.16 0.17 0.17 0.17
0.10 0.09
0.10

178 0.00
Angola

Angola

Cameroon

Cameroon

Central African Republic

Central African Republic

Chad

Chad

Congo Democratic

Congo Democratic

Congo Republic

Congo Republic

Equitorial Guinea

Equitorial Guinea

Gabon

Gabon

Sao Tome

Sao Tome
Figure 7.
Democracy in
Central Africa
2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015 2006 2015

Figure 8 shows the levels of democracy in 2006 and 2015 in Southern Africa. Over the
ten-year period, Botswana and South Africa are the best-performing. In the region, former
liberation movements are the dominant political parties, a reminder of people’s struggle
with imperialist oppressors. South Africa had been embroiled in domestic policy tensions
since it became a democracy. Robert Mugabe, reigning in Zimbabwe since independence
in 1980, shows no intention of retiring. Zambia’s main Opposition Leader, Hakainde
Hichilema of the United Party for National Development, has been on trial for high
treason. Beyond multi-party systems and regular elections, the countries resemble very
little of true democracies.
Table II shows the best-performing and the worst-performing countries in Africa as a
whole in terms of democracy in 2015. The best-performing countries include Botswana,
Mauritius, South Africa and Seychelles. The worst-performing countries include Central
African Republic, Gambia, Zimbabwe, Congo Republic, Rwanda and Guinea. In between
these two extremes, there are a wide range of countries doing well, and not so well.
In conclusion, traditional democratic institutions have not, without problems, taken root
in the African soil. Often, attempts at nurturing democratic norms have floundered, leaving
authoritarianism or anarchy. Many authoritarian leaders continue to maintain an iron grip

0.70

0.68

0.66

0.64

0.62

0.60

0.58

0.56

Figure 8. 0.54
Democracy in Botswana Botswana Namibia Namibia South Africa South Africa
Southern Africa
2006 2015 2006 2015 2006 2015
Best-performing countries Score Worst-performing countries Score
Digital politics
Botswana 0.67 Guinea 0.25
Mauritius 0.66 Rwanda 0.25
South Africa 0.66 Congo Republic 0.24
Namibia 0.60 Zimbabwe 0.22
Seychelles 0.56 Gambia 0.21
Ghana 0.56 Central African Republic 0.17 179
Senegal 0.52 Chad 0.17
Benin 0.50 Congo Democratic 0.17
Zambia 0.50 Sudan 0.12 Table II.
Sao Tome 0.50 Equatorial Guinea 0.09 Democracy in Africa

on power in many parts of the continent, either amending laws to extend their terms of
office, repressing opposition and civil society or hosting rubber-stamp elections. As a result,
most countries in the continent continue to grapple with democratic deficits.

4. Methodology
4.1 Data
Annual data for 38 African countries for the period of 2006–2015 are chosen on the basis of
availability. Data are collected for the following African countries: Angola, Benin, Botswana,
Cameroon, Central African Republic, Chad, Congo Democratic, Congo Republic, Cote
D’Ivoire, Equatorial Guinea, Ethiopia, Gabon, Gambia, Ghana, Guinea, Kenya, Liberia,
Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria,
Rwanda, Sao Tome, Senegal, Seychelles, Sierra Leone, South Africa, Sudan, Tanzania, Togo,
Uganda, Zambia and Zimbabwe. Other African countries do not have the full set of the two
data series and thus are not selected for the study.
In line with studies such as Guo and Ladner (2016) and Berg-Schlosser (2017), the proxy
for democracy is the democracy index (Democracy) and collected from the EIU. The EUU’s
democracy index is a measure of democracy that reflects the state of political freedoms
and civil liberties and is designed to achieve more in-depth comparisons of the quality of
democracy (EIU, 2013). The index for any given country is the product of the aggregation
of five categories: electoral process and pluralism; functioning of government;
political participation; political culture; and civil liberties. The EIU democracy index
exhibits important strengths compared with other measures of democracy in the literature
(e.g. democracy indices by Freedom House and Polity). Particularly, it takes into
consideration not only experts’ views but other sources of data as well, such as public
opinion data and electoral turnout from comparative surveys (Fernandes et al., 2015).
In line with studies such as Evans and Adeoye (2016), Salahuddin and Gow (2016), Adeola and
Evans (2018), and Evans et al. (2018), the proxy for internet usage is individuals using the internet
percent of population) (Internet) and collected from World Development Indicators online database.
This indicator is defined as the number of individuals using the internet per 100 inhabitants, and these
are individuals who used the internet in the last three months from any location.
Table III gives the descriptive statistics of the two variables used in the analysis.

4.2 The model

In line with studies such as Cham (2016), Pirannejad (2017) and Stier (2017), this study
empirically investigates the following model:

Democracyit ¼ ai þbit I nternet it þcit I nternet 2it þEit ; (1)

JES Democracy Internet
46,1
Mean 0.382056 9.734281
Median 0.362500 5.342074
Maximum 0.754250 58.11870
Minimum 0.083000 0.227669
SD 0.159613 11.50306
180 Skewness 0.248410 1.966244
Kurtosis 2.340405 6.599231
Jarque–Bera 10.79668 449.9663
Probability 0.004524 0.000000
Sum 145.1811 3699.027
Table III. Sum sq. dev. 9.655535 50149.42
Descriptive statistics Observations 380 380

the coefficients b and c represent the long-run elasticity estimates of Democracy with respect
to Internet usage and squared internet usage, respectively. The relationship between
Democracy and Internet is non-linear. Thus there is a U-shaped curve where at first an
increase in Internet leads to lower Democracy (b o0). Eventually, a point is reached when
an increase in Internet leads to an increase in Democracy (cW 0).
The estimation follows these steps: testing for unit roots of the variables; examining the
existence of long-run co-integration relationships and their magnitude; estimation of panel
error correction models for short- and long-run dynamics; and Granger causality tests.

4.3 Panel unit root tests

Recent literature has suggested that panel unit root tests have higher power than individual
unit root tests (Im et al., 2003). The four types of panel unit root tests used in this study are
Levin et al. (2002), Im et al. (2003) and Fisher-type tests using ADF and PP tests (Maddala
and Wu, 1999).
Theoretically, the tests are multiple-series unit root tests applied to panel data structures.
The following AR(1) process is considered for panel data:

yit ¼ ri yit1 þX it di þeit (2)

where i ¼ 1, 2, …, n cross-sections, that are observed over t ¼ 1, 2, …, Ti periods.

Xit represents the exogenous variables in the model, including any individual trends or
fixed effects; ρi are the autoregressive coefficients; and the errors εit are assumed to be
mutually independent idiosyncratic disturbance (iid). If |ρi|o 1, yi is said to be weakly
stationary. On the contrary, if |ρi| ¼ 1 then yi has a unit root.
For the purposes of panel unit root testing, two natural assumptions can be made about
ρi. First, it is assumed that the persistence parameters are common across cross-sections
such that ρi ¼ ρ for all i. Levin, Lin and Chu test employs this assumption. Alternatively,
ρi can be allowed to vary freely across cross-sections. Im, Pesaran and Shin; Fisher–ADF
and Fisher–PP tests are of this form.
The four tests, Levin, Lin, and Chu; Im, Pesaran and Shin; Fisher–ADF; and Fisher–PP
tests, are, therefore, used in this study to determine the panel unit roots of the data series.

4.4 Johansen–Fisher panel cointegration test

Fisher (1932) derives a combined individual test (Fisher/Johansen) that uses the results of
the individual independent tests. Using Fisher’s result, Maddala and Wu (1999) propose an
alternative approach to testing for panel cointegration by combining tests from individual Digital politics
cross-sections to derive a test statistic for the full panel.
If πi is the p-value from individual cointegration tests for cross-section i, then under the
null hypothesis for the panel:

X
N
2 log ðpi Þ-w2 2N ; (3)
i¼1 181
the value of χ2 is based on MacKinnon et al. (1999) p-values for Johansen’s cointegration
trace test and maximum eigenvalue test.
The Johansen–Fisher panel cointegration test is, therefore, used in this study to
determine the panel cointegration of the variables.
Pooled mean group. Having established that there is a long-run relationship for all
countries in the panel of countries, it is appropriate to turn to the estimation of the complete
panel error-correction model:
X
p X
p X
p
DDemocracyit ¼ bj DDemocracyitj þ yj DI nternet itj þ dj DI nternetitj2
j¼1 j¼0 j¼0
 
þEct t1 Democarcyit1 ag I nternet it1 lj I nternetit12 þeit ; (4)

Pesaran et al.’s (1999) PMG approach is used to estimate the equation in order to obtain the
estimates of the error correction parameters, λi (or speed of adjustment to the equilibrium
values), and of the short-run parameters βj and θj for each country in the panel. Thus, the
error correction parameters and short-run coefficients can differ across countries. The lag
length p is selected using the Schwarz (SC) selection criteria.

4.5 Dumitrescu–Hurlin panel causality tests

The Dumitrescu–Hurlin panel causality test is used to test for causality among the
variables. The Dumitrescu and Hurlin (2012) allows all coefficients to be different across
cross-sections (heterogeneous across cross-section):
a0;i aa0; j ; . . .; a1;i aa1; j ;

b0;i ab0; j ; . . .; b1;i ab1; j ; (5)

α stands for coefficients of internet across different cross-sections, and β represents
coefficients of democracy across the different cross-sections. The test is computed by
running standard Granger causality regressions for individual cross-sections and taking the
average. This approach takes into account two different statistics. The first statistics, the
Wbar statistic, is the average of the individual test statistics. Zbar is the standardized
version of this test statistic, which follows a standard normal distribution.

5. Empirical analysis
5.1 The panel unit root test
Since most macroeconomic series have a unit root (Nelson and Plosser, 1982), it is
appropriate to check the order of integration of the variables. The panel unit root tests for
Internet, Internet2 and Democracy for the 38 African countries for the years 2006–2015 are
as shown in Table IV. The summary panel unit root test is computed using individual fixed
effects as regressors, and automatic lag difference term and bandwidth selection (using the
JES Level First difference
46,1 Method Statistic Prob. Statistic Prob.

Democracy
Levin, Lin and Chu t −34.9532*** 0.0000 −45.0964*** 0.0000
Im, Pesaran and Shin W-stat −9.13782*** 0.0000 −9.27240*** 0.0000
ADF–Fisher χ2 113.734*** 0.0013 164.942*** 0.0000
182 PP–Fisher χ2 99.0827** 0.0189 181.624*** 0.0000
Internet
Levin, Lin and Chu t −4.63771*** 0.0000 −28.7412*** 0.0000
Im, Pesaran and Shin W-stat 3.11431 0.9991 −7.53538*** 0.0000
ADF–Fisher χ2 73.3003 0.5665 191.173*** 0.0000
PP–Fisher χ2 130.207*** 0.0001 172.390*** 0.0000
Internet2
Levin, Lin and Chu t −0.31146 0.3777 −22.6002*** 0.0000
Im, Pesaran and Shin W-stat 6.54403 1.0000 −4.53478*** 0.0000
ADF–Fisher χ 2
44.9090 0.9983 160.879*** 0.0000
PP–Fisher χ2 48.7052 0.9938 123.679*** 0.0005
Table IV. Notes: Probabilities for Fisher tests are computed using an asymptotic χ2 distribution. All other tests
Summary of the assume asymptotic normality. Newey–West automatic bandwidth selection and Bartlett kernel; balanced
panel unit root test observations for each test. *,**,***Represent 10, 5 and 1 percent levels of significance, respectively

Schwarz criterion for the lag differences, and the Newey–West method and the Bartlett
kernel for the bandwidth).
For Democracy, the result indicates the absence of a unit root, as the Levin, Lin and Chu;
Im, Pesaran and Shin; and both Fisher tests reject the null of a unit root at 1 percent
significance level. For Internet, the result indicates the presence of a unit root, as Im, Pesaran
and Shin W-stat and ADF–Fisher χ2 reject the null of a unit root at first difference at
1 percent significance level. For Internet2, the result indicates the absence of a unit root, as
both Fisher tests reject the null of a unit root at 1 percent significance level. Thus,
Democracy is I(0) while Internet and Internet2 are I(1). With this combination of I(1) and I(0),
PMG is the most appropriate method to estimate the model.

5.2 Johansen–Fisher panel cointegration test

To apply PMG for estimation, a cointegrating relation must first be established among the
set of variables (Adeola and Evans, 2017). For that reason, the presence of a cointegrating
relation is tested using the Johansen–Fisher panel cointegration test as shown in Table V.
The pooled results show evidence of cointegrating relationship between the variables in the
model. Even the individual cross-section results indicate cointegrating relationship between
the variables for most of the countries (Table AI in the Appendix).

5.3 PMG estimation

Table VI presents the panel long- and short-run results from the PMG estimation. The
results show that internet usage (Internet) has a significant negative impact on democracy
while squared internet usage (Internet2) has a significant positive impact, both in the short
and long run. This finding provides evidence of a non-linear relationship between internet
usage and democracy. This suggests that as internet usage increases, democracy
decreases, but after a certain level of internet usage which is the turning point, democracy
starts to increase.
 To further appreciate the nature of the relationship between internet and democracy in Digital politics
Africa, the study looks at the short-run estimates for each country. As shown in Table AII in
the Appendix, internet usage (Internet) has a significant negative impact on democracy
while squared internet usage (Internet2) has significant positive impact on democracy in all
the countries (except in Chad, Gabon, Mauritius, Sierra Leone and Togo). This additionally
reinforces the evidence of the non-linear relationship between internet usage and democracy
as shown above. However, there are different significance levels regarding the impacts of 183
internet usage on democracy in the different countries.

5.4 Causality tests

Additionally, as shown in Table AII in the Appendix, the significance of the Ectt−1 in
most of the countries indicate a long-run convergence to equilibrium of the variables.
The existence of this long-run relationship between Internet and Democracy suggests
that there must be causality at least in one direction. Table VII therefore
reports the Wbar statistic and Zbar statistics for the Dumitrescu–Hurlin panel causality
tests. The empirical results support uni-directional causality from Internet to Democracy.

Pooled results
Hypothesized no. of CE(s) Fisher Stat.a (from trace test) Prob. Fisher Stat.a (from max-eigen test) Prob.**

None 441.5*** 0.0000 408.4*** 0.0000

At most 1 180.9*** 0.0000 180.9*** 0.0000 Table V.
Notes: aProbabilities are computed using asymptotic χ2 distribution; **MacKinnon et al. (1999) p-values. Johansen–Fisher panel
***Represent 1 percent levels of significance cointegration test

Coefficient SE t-stat Prob.

Panel A: long run estimates

Internet −0.489*** 0.017 −28.381 0.000
Internet2 0.308*** 0.010 29.490 0.000
Panel B: short run estimates
Δ (Internet) −2.444* 1.470 −1.662 0.099
Δ (Internet (−1)) −0.322 1.421 −0.227 0.821
Δ (Internet2) 1.078* 0.620 1.738 0.084
Δ (Internet (−1)2) 0.820 0.767 1.070 0.287
Ectt−1 −0.436*** 0.116 −3.769 0.000 Table VI.
Notes: *,**,***Represent 10, 5 and 1 percent levels of significance, respectively Panel results

W-Stat. Z bar–Stat. Prob.

Internet→Democracy 4.49167*** 5.49831*** 0.0000

Democracy→Internet 2.23720 1.35488 0.1755
Internet2→Democracy 3.07575*** 2.89603*** 0.0038
Democracy→Internet2 2.91015*** 2.59167*** 0.0096
Internet2→Internet 1.68594 0.34173 0.7326 Table VII.
Internet→Internet2 16.8553*** 28.2210*** 0.0000*** Dumitrescu–Hurlin
Notes: *,**,***Represent 10, 5 and 1 percent levels of significance, respectively panel causality tests
JES Additionally, there is bi-directional causality between Internet2 and Democracy. This
46,1 indicates that internet usage stimulates democracy which, in turn, boosts internet usage
even further in these countries.
Overall, the empirical evidence from this study suggests that internet usage and
democracy are highly interrelated to each other in Africa. The findings are broadly
consistent with earlier studies on the relationship between internet usage and democracy
184 (e.g. Salgado, 2012; Lee, 2017; Pirannejad, 2017). According to Salgado (2012), the internet,
through online newspapers, information websites, blogs, etc., has strengthened civil
society in various contexts. According to Lee (2017), internet usage is positively related to
unconventional political participation and can provide mobilizing capacities for
organizational membership and political action. Also, Pirannejad (2017) showed that
internet extension has a significantly positive impact on democracy promotion. Similarly,
the findings are consistent with some studies from Africa such as Nisbet et al. (2012) who
showed that internet usage is positively related to higher citizen commitment to
democratic governance.

6. Summary and implications

6.1 Summary
This study has investigated the relationship between internet usage and levels of
democracy in 38 African countries, using PMG and causality tests for the period 2006–2015.
The empirical results have shown that internet usage has a significant negative impact on
democracy while squared internet usage has a significant positive impact on democracy,
both in the short and long run. Particularly, in the short run, this evidence is true for most of
the countries except Chad, Gabon, Mauritius, Sierra Leone and Togo. Also, there are
different significance levels regarding the impacts of internet usage on democracy in
individual countries. Moreover, the empirical results support uni-directional causality from
internet usage to democracy and bi-directional causality between squared internet usage
and democracy.

6.2 Theoretical implications

While much of the literature has suggested a positive impact of internet usage on democracy
(Nisbet et al., 2012; Salgado, 2012; Lee, 2017; Pirannejad, 2017), this study has expanded the
literature on the relationship between internet usage and democracy in Africa and has
offered new insights into the relations between the two variables both in the short and long
run. The empirical results indicate that internet usage has a significant negative impact on
democracy while squared internet usage has a significant positive impact on democracy,
both in the short and long run. This finding provides evidence of a non-linear relationship
between internet usage and democracy. Democracy displays a quadratic relationship with
internet usage. This suggests a U-shaped pattern: democracy decreases with internet usage,
stabilizes, and then increases (Figure 9). This suggests that as internet usage increases,
democracy decreases, but after a certain level of internet usage which is the turning point,
democracy starts to increase.
The results also show uni-directional causality from internet usage to democracy. There
is also bi-directional causality between squared internet usage and democracy. This
indicates that internet usage stimulates democracy which, in turn, boosts internet usage
even further in these countries. This suggests that internet usage and democracy are highly
interrelated to each other. That is, the evidence underscores the feedback effects between
internet usage and democracy. This study, therefore, contributes to the literature in
highlighting that internet expansion leads to democracy while democracy also leads to
expansion of the internet both in the short and long run.
6.3 Practical implications Digital politics
The empirical results show that internet usage has a significant negative impact on
democracy. This is not surprising because the internet as a platform for active participation
in public discourse in Africa is subjected to restrictions that undermine democracy. Many
African Governments continue to impose restrictions on internet and social media access,
especially during elections. In many instances, some even employ social media to stimulate
hate speech against opposition parties. 185
Nonetheless, squared internet usage has a significant impact on democracy. First, most
African countries have made heavy investment in infrastructure in the ICT sector
improving access to the internet. For example, there are increasing number of submarine
cables connecting Africa to America, Asia and Europe and international connectivity is no
longer a problematic issue. This has aided information sharing, both within the continent
and to the world. Second, many authoritarian regimes in Africa are increasingly confronted
with the internet as a force for increased democracy. The rise in democracy as internet usage
increases may explain the change to increased internet freedom in those countries and the
move toward more democratic norms and digital politics.
While the empirical results show that the internet plays a significant role in democracy in
both the short and long run, this evidence is true for most of the countries except Chad,
Gabon, Mauritius, Sierra Leone and Togo in the short run. This implies that while the
internet plays a significant role in democracy in the other countries, its impact on democracy
is not significant for Chad, Gabon, Mauritius, Sierra Leone and Togo in the short run. The
absence of short-run significant impact of internet on democracy in these countries implies
that internet is not the proper solution to increase the levels of democracy in the short run in
these countries.
The study found a U-shaped pattern between democracy and internet usage. As
explained earlier, this suggests that as internet usage increases, democracy decreases, but
after a certain level of internet usage which is the turning point, democracy starts to
increase. This might be explained that for low levels of the penetration rate, the internet
service is prioritized for other sectors of the economy such as e-commerce and education,
leaving the variables associated with democracy minimized. As the penetration rate
increases, some social elements linked to democracy arise.
The empirical results also support uni-directional causality from internet usage to
democracy. Additionally, there is bi-directional causality between squared internet usage
and democracy. This indicates that internet usage stimulates democracy which, in turn,
boosts internet usage even further in these countries. However, there are different
significance levels regarding the impacts of internet usage on democracy in individual
countries. The different significance levels regarding the existence of non-linear relationship
Democracy

Turning point

Figure 9.
U-shaped pattern
between internet and
democracy
Internet
JES between internet usage and democracy is expected as internet usage and levels of
46,1 democracy are not evenly distributed in Africa. That is, the different levels of internet usage
and democracy have different levels of impacts across the continent.
The findings of this study have important policy implications. The implication is that, as
many African countries begin to tread the path of increased democracy and the spread of
human rights, the internet should be seen as laying the foundation for Africa that is
186 democratic in the short and long run. In this regard, policymakers need to pay more
attention to the relationship between internet and democracy in Africa in order to ensure
that potential gains are fully maximized. Different governments and institutions in Africa
need to encourage and invest in internet infrastructure for increased democracy. It is,
therefore, the responsibility of all stakeholders to collaborate to develop policies, tools, and
applications that will maximize the benefits of digital politics, advancing access to the
internet, and thereby promoting democracy and human rights.

6.4 Limitations and future research

The analysis of the relationship between internet and democracy in this study is limited in
ways that future studies can perhaps investigate. Future research could extend the findings
of this study by incorporating variables that can influence the relationship between internet
and democracy. More causality tests would also show how the presence of these variables
could influence the results and the related interpretations. Further, this study used the
democracy index collected from EIU as a proxy for democracy. Comparable studies could
extend this study using other proxies of democracy. This issue needs to be re-examined in
future theoretical and empirical works.

References
Adeola, O. and Evans, O. (2017), “Financial inclusion, financial development, and economic
diversification in Nigeria”, The Journal of Developing Areas, Vol. 51 No. 3, pp. 1-15.
Adeola, O. and Evans, O. (2018), “Digital health: ICT and health in Africa”, Actual Problems of
Economics, Vol. 8.
Adeola, O., Boso, N. and Evans, O. (2017), “Drivers of international tourism demand in Africa”, Business
Economics, Vol. 53 No. 1, pp. 25-36.
Adibe, R., Ike, C.C. and Udeogu, C.U. (2017), “Press freedom and Nigeria’s Cybercrime Act of 2015: an
assessment”, Africa Spectrum, Vol. 52 No. 2, pp. 117-127.
Berg-Schlosser, D. (2017), “Contemporary problems and future perspectives of empirical research on
democracy”, Parties, Governments and Elites, Springer VS, Wiesbaden, pp. 303-320.
Best, M.L. and Wade, K.W. (2009), “The internet and democracy: global catalyst or democratic dud?”,
Bulletin of Science, Technology & Society, Vol. 29 No. 4, pp. 255-271.
Cham, J. (2016), “The internet and autocracy: unpacking the effects of Internet freedom”, Proceedings of
the 17th International Digital Government Research Conference on Digital Government Research,
ACM, pp. 164-173.
Dumitrescu, E.I. and Hurlin, C. (2012), “Testing for Granger non-causality in heterogeneous panels”,
Economic Modelling, Vol. 29 No. 4, pp. 1450-1460.
Economist Intelligence Unit (2013), Democracy Index 2012: Democracy is at a Standstill, The Economist
Intelligence Unit Limited, London.
Evans, O. (2018a), “Digital government: ICT & public sector management in Africa”, in Włodzimierz, S.
(Ed.), New Trends in Management: Regional, Cross-Border and Global Perspectives, London
Scientific Publishing, London.
Evans, O. (2018b), “Connecting the poor: the internet, mobile phones and financial inclusion in Africa”,
Digital Policy, Regulation and Governance, Vol. 20 No. 6, pp. 568-581.
Evans, O. (2018c), “Digital agriculture: mobile phones, internet & agricultural development in Africa”, Digital politics
Actual Problems of Economics, Vols 7/8 Nos 205/206, pp. 76-90.
Evans, O. (2018d), “Repositioning for increased digital dividends: internet usage and economic
wellbeing in Sub-Saharan Africa”, Journal of Global Information Technology Management.
Evans, O. and Adeoye, B. (2016), “The determinants of financial inclusion in Africa: a dynamic panel
data approach”, University of Mauritius Research Journal, Vol. 22, pp. 310-336.
Evans, O. and Alenoghena, R. (2015), “Corruption effects on Nigeria: aggregate and sectoral estimates 187
using VAR”, Journal of Economic & Financial Studies, Vol. 3 No. 2, pp. 18-26.
Evans, O. and Kelikume, I. (2018), “The effects of foreign direct investment, trade, aid, remittances and
tourism on welfare under terrorism and militancy”, International Journal of Management,
Economics and Social Sciences, Vol. 7 No. 3, pp. 206-232.
Evans, O. and Kelikume, I. (2019), “The impact of poverty, unemployment, inequality, corruption and
poor governance on Niger Delta Militancy, Boko Haram Terrorism and Fulani Herdsmen
Attacks in Nigeria”, International Journal of Management, Economics and Social Sciences.
Evans, O., Adeniji, S., Nwaogwugwu, I., Kelikume, I., Dakare, O. and Oke, O. (2018), “The relative effect
of monetary and fiscal policy on economic development in Africa: a GMM approach to the
St Louis equation”, Business and Economic Quarterly, Vol. 2, pp. 3-23.
Ferdinand, P. (2000), “The internet, democracy and democratization”, Democratization, Vol. 7 No. 1, pp. 1-17.
Fernandes, T., Cancela, J., Coppedge, M., Lindberg, S.I. and Hicken, A. (2015), “Measuring political
participation in Southern Europe: the varieties of democracy approach”, V-Dem Institute,
Department of Political Science University of Gothenburg, available at: www.v-dem.net
Fisher, R (1932), Statistical Methods for Research Workers, Oliver and Boyd, Edinburgh.
Grofman, B., Trechsel, A.H. and Franklin, M. (Eds) (2014), The Internet and Democracy in Global Perspective:
Voters, Candidates, Parties, and Social Movements, Vol. 31, Studies in Public Choice, Springer.
Guo, Y. and Ladner, A. (2016), “Public administration in Singapore and Switzerland”, in Guo, Y. and
Woo, J. (Eds), Singapore and Switzerland: Secrets to Small State Success, World Scientific
Pub Co Inc., pp. 1-27.
Im, K.S., Pesaran, M.H. and Shin, Y. (2003), “Testing for unit roots in heterogeneous panels”, Journal of
Econometrics, Vol. 115 No. 1, pp. 53-74.
Lee, S.H. (2017), “Digital democracy in Asia: the impact of the Asian internet on political participation”,
Journal of Information Technology & Politics, Vol. 14 No. 1, pp. 62-82.
Levin, A., Lin, C.F. and Chu, C.S.J. (2002), “Unit root tests in panel data: asymptotic and finite-sample
properties”, Journal of Econometrics, Vol. 108 No. 1, pp. 1-24.
MacKinnon, J.G., Haug, A.A. and Michelis, L. (1999), “Numerical distribution functions of likelihood
ratio tests for cointegration”, Journal of Applied Econometrics, Vol. 14 No. 5, pp. 563-577.
Maddala, G.S. and Wu, S. (1999), “A comparative study of unit root tests with panel data and a new
simple test”, Oxford Bulletin of Economics and Statistics, Vol. 61 No. S1, pp. 631-652.
Margetts, H. (2013), “The internet and democracy”, The Oxford Handbook of Internet Studies,
doi: 10.1093/oxfordhb/9780199589074.013.0020.
Miller, C. (2016), The Rise of Digital Politics, Demos, London.
Mustapha, M. (2017), “The 2015 general elections in Nigeria: new media, party politics and the political
economy of voting”, Review of African Political Economy, Vol. 44 No. 152, pp. 312-321.
Nam, T. (2017), “A tool for liberty or oppression? A cross-national study of the internet’s influence on
democracy”, Telematics and Informatics, Vol. 34 No. 5, pp. 538-549.
Nelson, C.R. and Plosser, C.R. (1982), “Trends and random walks in macroeconomic time series: some
evidence and implications”, Journal of Monetary Economics, Vol. 10 No. 2, pp. 139-162.
Nisbet, E.C., Stoycheff, E. and Pearce, K.E. (2012), “Internet use and democratic demands: a
multinational, multilevel model of internet use and citizen attitudes about democracy”, Journal of
Communication, Vol. 62 No. 2, pp. 249-265.
JES Ott, D. (1998), “Power to the people: role of electronic media in promoting democracy in Africa”, First
46,1 Monday, Vol. 3 No. 4.
Pesaran, M.H., Shin, Y. and Smith, R.P. (1999), “Pooled mean group estimation of dynamic heterogeneous
panels”, Journal of the American Statistical Association, Vol. 94 No. 446, pp. 621-634.
Pirannejad, A. (2017), “Can the Internet promote democracy? A cross-country study based on dynamic
panel data models”, Information Technology for Development, Vol. 3 No. 2, pp. 1-15.
188 Ridgwell, H. (2017), “‘Digital democracy’ turns average citizens into influencers in Africa”, available at:
www.voanews.com/a/with-digital-democracy-average-citizens-find-their-voice-in-africa/391330
9.html (accessed August 9, 2017).
Ruijgrok, K. (2017), “From the web to the streets: internet and protests under authoritarian regimes”,
Democratization, Vol. 24 No. 3, pp. 498-520.
Salahuddin, M. and Gow, J. (2016), “The effects of internet usage, financial development and trade
openness on economic growth in South Africa: a time series analysis”, Telematics and
Informatics, Vol. 33 No. 4, pp. 1141-1154.
Salgado, S. (2012), “The web in African countries: exploring the possible influences of the internet in the
democratization processes”, Information, Communication & Society, Vol. 15 No. 9, pp. 1373-1389.
Shen, F. (2017), “Internet use, freedom supply, and demand for internet freedom: a cross-national study
of 20 countries”, International Journal of Communication, Vol. 11, pp. 2093-2114.
Smith, E. (2016), “In central Africa, citizens are using social media to build democracy. Here’s how”,
The Washington Post, available at: www.washingtonpost.com/news/monkey-cage/wp/
2016/04/06/in-central-africa-citizens-are-using-social-media-to-build-democracy-heres-how/
(accessed August 11, 2017).
Stier, S. (2017), “Internet diffusion and regime type: temporal patterns in technology adoption”,
Telecommunications Policy, Vol. 41 No. 1, pp. 25-34.
Vaccari, C. (2013), Digital Politics in Western Democracies: A Comparative Study, JHU Press, Baltimore.
Weare, C. (2002), “The internet and democracy: the causal links between technology and politics”,
International Journal of Public Administration, Vol. 25 No. 5, pp. 659-691.

Further reading
Economist Intelligence Unit (EIU) (2016), “Democracy index”, available at: www.eiu.com (accessed
August 11, 2017).
WDI (2017), “Individuals using the Internet (% of population)”, available at: http://databank.worldbank.
org/data/reports.aspx?source=world-development-indicators (accessed August 11, 2017).
Appendix Digital politics

Trace test Max-eigen test

Cross-section Statistics Prob.a Statistics Prob.a
189
Hypothesis of no cointegration
Angola 51.0586 0.0000 50.9764 0.0000
Benin 26.2820 0.0008 22.5574 0.0020
Botswana 22.4777 0.0038 21.2162 0.0034
Cameroon 16.4486 0.0358 15.4582 0.0322
CAR 29.8654 0.0002 22.9823 0.0017
Chad 23.8380 0.0022 23.0785 0.0016
Congo D 32.9812 0.0001 26.3056 0.0004
Congo R 33.7928 0.0000 33.7696 0.0000
Cote D 11.8397 0.1648 11.2313 0.1430
Equatorial G 23.4506 0.0026 18.7739 0.0090
Ethiopia 34.4732 0.0000 34.0456 0.0000
Gabon 24.0258 0.0021 20.3745 0.0048
Gambia 11.2371 0.1973 10.1331 0.2033
Ghana 16.1785 0.0394 16.0564 0.0258
Guinea 24.2047 0.0019 22.8007 0.0018
Kenya 16.8839 0.0307 15.2644 0.0346
Liberia 6.0324 0.6917 4.5469 0.7976
Madagascar 33.3260 0.0000 29.7958 0.0001
Malawi 16.1907 0.0392 15.8909 0.0274
Mali 5.7188 0.7285 5.7187 0.6495
Mauritania 12.4430 0.1369 12.0805 0.1077
Mauritius 3.1993 0.9571 3.1940 0.9330
Mozambique 28.7059 0.0003 21.2356 0.0034
Namibia 15.4136 0.0514 11.4592 0.1326
Niger 35.9251 0.0000 28.9410 0.0001
Nigeria 43.0788 0.0000 36.2975 0.0000
Rwanda 17.1388 0.0280 13.3360 0.0697
Sao Tome 305.8533 0.0001 285.1055 0.0001
Senegal 23.0370 0.0030 20.0919 0.0054
Seychelles na 0.5000 na 0.5000
Sierra Leone 29.8275 0.0002 25.5248 0.0006
South Africa 14.2530 0.0763 12.3830 0.0971
Sudan 31.6124 0.0001 28.1327 0.0002
Tanzania 6.9477 0.5837 6.8404 0.5082
Togo 17.5574 0.0241 15.5660 0.0310
Uganda 51.8129 0.0000 44.0309 0.0000
Zambia 6.6959 0.6132 6.2569 0.5805
Zimbabwe 15.7726 0.0454 9.7048 0.2320
Hypothesis of at most 1 cointegration relationship
Angola 0.0823 0.7742 0.0823 0.7742
Benin 3.7246 0.0536 3.7246 0.0536
Botswana 1.2614 0.2614 1.2614 0.2614
Cameroon 0.9904 0.3196 0.9904 0.3196
CAR 6.8831 0.0087 6.8831 0.0087
Chad 0.7595 0.3835 0.7595 0.3835
Congo D 6.6756 0.0098 6.6756 0.0098
Table AI.
Individual cross-
(continued ) section results
JES
Trace test Max-eigen test
46,1 Cross-section Statistics Prob.a Statistics Prob.a

Congo R 0.0233 0.8787 0.0233 0.8787

Cote D 0.6085 0.4354 0.6085 0.4354
Equatorial G 4.6767 0.0306 4.6767 0.0306
Ethiopia 0.4276 0.5132 0.4276 0.5132
190 Gabon 3.6513 0.0560 3.6513 0.0560
Gambia 1.1041 0.2934 1.1041 0.2934
Ghana 0.1221 0.7267 0.1221 0.7267
Guinea 1.4040 0.2360 1.4040 0.2360
Kenya 1.6194 0.2032 1.6194 0.2032
Liberia 1.4855 0.2229 1.4855 0.2229
Madagascar 3.5302 0.0603 3.5302 0.0603
Malawi 0.2999 0.5840 0.2999 0.5840
Mali 0.0001 0.9933 0.0001 0.9933
Mauritania 0.3625 0.5471 0.3625 0.5471
Mauritius 0.0053 0.9416 0.0053 0.9416
Mozambique 7.4703 0.0063 7.4703 0.0063
Namibia 3.9544 0.0467 3.9544 0.0467
Niger 6.9840 0.0082 6.9840 0.0082
Nigeria 6.7813 0.0092 6.7813 0.0092
Rwanda 3.8028 0.0512 3.8028 0.0512
Sao Tome 20.7478 0.0000 20.7478 0.0000
Senegal 2.9452 0.0861 2.9452 0.0861
Seychelles 0.6273 0.4284 0.6273 0.4284
Sierra Leone 4.3027 0.0380 4.3027 0.0380
South Africa 1.8700 0.1715 1.8700 0.1715
Sudan 3.4796 0.0621 3.4796 0.0621
Tanzania 0.1073 0.7432 0.1073 0.7432
Togo 1.9914 0.1582 1.9914 0.1582
Uganda 7.7820 0.0053 7.7820 0.0053
Zambia 0.4390 0.5076 0.4390 0.5076
Zimbabwe 6.0678 0.0138 6.0678 0.0138
Notes: Probabilities are computed using asymptotic χ2 distribution. aMacKinnon–Haug–Michelis (1999)
Table AI. p-values
 Ectt−1 Δ (Internet) Δ (Internet (−1)) Δ (Internet2) Δ (Internet (−1)2)
Digital politics
Angola −1.084* −0.003 −0.088 0.111** 0.114**
Benin 0.066*** 1.126* −0.994* −0.769* 0.976*
Botswana 0.724* −0.837* −0.071* 0.483* 0.029*
Cameroon −1.360* 1.939* −0.117* −1.109* 0.084*
CAR −1.612** 0.095 −0.112** 0.757** 0.039
Chad −1.294* 14.948 −4.036 −7.957 17.704 191
Congo D −0.784* 0.461* 0.290** 0.435 0.689***
Congo R −1.009* 0.166** −0.241* −0.103** 0.240*
Cote D −0.195* −0.567* 0.108* 0.453* −0.322*
Equitorial G −0.526* 0.408* 0.259* −0.417* −0.031*
Ethiopia −0.213* −0.070 0.180* 0.293* −0.180**
Gabon −1.247* −14.657 −2.551 6.244 0.624
Gambia −1.313* −12.870 16.022*** 4.718** −9.334**
Ghana −0.051 0.474* −0.113* −0.185* 0.125*
Guinea −0.553* −0.120*** −0.474* 0.422** 3.120
Kenya 0.006 0.490** −2.006 −0.113** 0.771**
Liberia −0.090* −0.188* 0.018* 0.136* 0.187*
Madagascar −1.164* 0.310* 0.114* −1.194* 1.681*
Malawi −0.010*** 0.093* 0.051* −0.034* −0.058*
Mali −1.406* 0.231* −0.044* 0.010* 0.086*
Mauritania −1.906* −0.762* 0.566* 0.402* −0.687*
Mauritius 0.201 1.003 −0.760 −0.328 0.290
Mozambique −0.298* 0.045* 0.059* 0.115* −0.203*
Namibia 0.619* −0.529* −1.321* 0.297* 0.707*
Niger −0.681* 0.653* 0.172* 0.805* 1.073*
Nigeria −0.191* 6.088 8.538 −3.973*** −4.567
Rwanda −0.066 0.007 −0.056 −0.021* −0.061
Sao Tome 0.356* 41.036 −43.185 −16.291** 17.668
Senegal 0.516** 2.095* −0.537*** −0.912* 0.461*
Seychelles −1.564* 22.487 24.033 −7.164** −7.490***
Sierra Leone −0.220 −0.371 1.290 −0.895 1.162
South Africa 0.041* −0.044 −0.209* 0.020*** 0.091*
Sudan −0.505* 14.252 −1.385*** −6.133 0.742*
Tanzania 0.847* −0.443* 1.681* 0.604* −1.784*
Togo −0.669* 14.677 −5.828 −9.068 6.241
Uganda 0.098* −0.018* −0.053* −0.005* 0.030*
Zambia 0.394* 1.274* −0.788* −0.597* 0.545*
Zimbabwe −0.429** −0.025 −0.668** 0.001* 0.397* Table AII.
Notes: The estimations are obtained from the pooled mean group approach. *,**,***Represent 1, 5 and 10 Individual country
percent levels of significance, respectively short-run estimates

Corresponding author
Olaniyi Evans can be contacted at: [email protected]

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]