Spatial Inequality and the Internet Divide in Indonesia
-
Upload
muhammad-mitsaq-zamir -
Category
Documents
-
view
214 -
download
0
Transcript of Spatial Inequality and the Internet Divide in Indonesia
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
1/15
Spatial inequality and the Internet divide in Indonesia2010–2012
Sujarwoto Sujarwoto a,n, Gindo Tampubolon b
a Brawijaya University, Malang, Indonesiab University of Manchester, United Kingdom
a r t i c l e i n f o
Keywords:
Spatial inequality
Internet divide
Indonesia
a b s t r a c t
Spatial inequality has been one of the key development characteristics considered across
developing countries. However, relatively few studies examine the mechanisms by which
spatial inequality explains the existing digital divide in a developing country. Applying the
normalisation and stratication thesis in diffusion theory, this study examines the ways in
which spatial inequality is related to the Internet divide in Indonesia, a developing
country that is currently growing in its use of Information and Communication Technol-
ogy (ICT), but that has experienced unequal regional development in the last three dec-
ades. Data comes from the Indonesian national socio-economic survey (Susenas) 2010–
2012, which comprises 3.3 million individuals, 750,000 households and 292 districts. Far
from moving towards convergence, the Internet divide expanded during this period; the
inequality of Internet access by age, gender, income, and education deepens and widens
across urban–rural, city–countryside, and remote island–mainland island areas. The
results of analyses using both stratied and multilevel models indicate that supply factors
across districts – particularly district disparities in telecommunications infrastructures,
human capital and education services – are associated with the Internet divide. The
results are robust against individual, household and district socio-economic character-
istics associated with the Internet divide. Enlarging the distribution of telecommunication
infrastructures and education facilities, particularly across districts in rural, countryside
and remote islands, may thus help to bridge the Internet divide in Indonesia.
& 2015 Elsevier Ltd. All rights reserved.
1. Introduction
The debate about the impact of the rise of the information society has produced deeply contested visions predicting thefuture direction of trends (Norris, 2001; Van Dijk & Hacker, 2003; Hargittai, 2002; Warschauer, 2003; Dutta & Mia, 2007).
Optimists hope that the development of the Internet will have the capacity to reduce, although not wholly eradicate, tra-
ditional inequalities between the information-rich and the information-poor both between and within societies ( Norris,
2001; Van Dijk & Hacker, 2003; Hargittai, 2002; Warschauer, 2003). In contrast, pessimists believe that ICTs will reinforce
and exacerbate existing disparities. Sceptics suggest that both the fears and hopes are exaggerated, with technologies
adapting to the social and political status quo, rather than vice-versa (Norris, 2001; Dutta & Mia, 2007).
Contents lists available at ScienceDirect
URL: www.elsevier.com/locate/telpol
Telecommunications Policy
http://dx.doi.org/10.1016/j.telpol.2015.08.008
0308-5961/& 2015 Elsevier Ltd. All rights reserved.
n Corresponding author.
E-mail addresses: [email protected] (S. Sujarwoto), [email protected] (G. Tampubolon).
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010–2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
http://www.sciencedirect.com/science/journal/03085961http://www.elsevier.com/locate/telpolhttp://dx.doi.org/10.1016/j.telpol.2015.08.008mailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008mailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://www.elsevier.com/locate/telpolhttp://www.sciencedirect.com/science/journal/03085961
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
2/15
Norris (2001), further makes a distinction between normalisation and stratication models of ICT diffusion. Normal-
isation thesis suggests that over time, access to the Internet will become widespread, overcoming social and other
boundaries, to make its day-to-day use appear normal. This thesis suggests that the prole of the online community will
come to reect society as a whole, given the wider availability of simpler and cheaper plug-and-play technologies and faster
broadband services, facilitating delivery of popular mass entertainment (Norris, 2001; Dutta & Mia, 2007). It is presupposed
that the differences between groups increase only in the early stages of adoption, and that those differences disappear with
saturation in the last stages. Certainly, the ubiquity of ready devices such as Wi-Fi networks, tablets and smartphones
enabling access to the Internet supports this notion.In contrast, the stratication thesis draws on experience with older technologies (such as the telegraph, automobiles, or
telephones in the twentieth century) to suggest that if Internet technology traces the same path, then the notion of a rise in
social inequality in terms of access cannot be easily dismissed ( Norris, 2001; Hargittai, 2002). Although social stratication is
not inevitable, depending as it does on, state intervention to provide enabling infrastructure and ensure equitable dis-
tribution and the nature of skills required to use the technology, digital access may persist in dividing groups in society. This
outcome is far from inevitable, because the conditions under which innovations implemented are also determined, in part,
by their social consequences. The existing social structure may thus, also play a role; as Rogers (2003) pointed out, inno-
vation in highly stratied societies usually reinforces existing socioeconomic inequalities. Norris (2001) provides empirical
evidence that, despite the high rate of penetration of ICTs in Europe and the United States, the digital divide between and
within countries is still perceptible.
This study attempts to make a distinction between normalisation and stratication theses in the context of a developing
country. A more nuanced understanding of the nature of ICT diffusion can be gained through examining the links between
widening spatial inequality within a developing country and the digital divide. Amidst growing concern about increasinginequality, the spatial dimensions of inequality have begun to attract considerable policy interest (Lessmann, 2014; Kanbur,
Rhee & Zhuang 2014; Tan & Zeng, 2014). In China, Russia, India, Mexico and South Africa, as well as most other developing
and transition economies, there is a sense that spatial and regional disparities in economic activity, incomes and social
indicators have been on the increase in the last two decades (Kanbur et al., 2014; Kanbur & Venables, 2005). For developing
countries that experience deep disparities across space, information and communication technology offers a sliver of hope
for bridging these disparities. However, spatial inequality with regard to ICT access may itself become a development
challenge, given the growth of Internet use. The gaps in physical access continue to grow in these developing countries; the
question is when and to what extent they will close again, equalising access for every social category (according to either
normalisation or stratication). A deeper analysis of these types of spatial inequality and the mechanisms explaining
unequal ICT access would prove a signicant contribution to developing countries’ efforts to address the digital divide.
This study aims to answer some of the questions raised by examining the mechanism linking determinants producing
spatial inequality with respect to the Internet divide in Indonesia. It is also often cited as an emerging economic success
(World Bank, 2008). However, its economic development is characterised by an endemic problem of spatial inequalities(Akita & Lukman, 1995; Hill, 1996; Resosudarmo & Vidyattama, 2006; Hill, Resosudarmo & Vidyattama, 2008; Yusuf, Sumer
& Rum, 2014). We thus, also consider in some depth whether these spatial inequalities reinforce the social inequalities in
Internet access. In this study, spatial inequality means a disparity in resources and services due to discrepancies in social and
economic factors across geography (Kanbur & Venables, 2005). Four measures of spatial inequality related to Internet access
are used: economy, human capital, telecommunication infrastructure, and education services. In order to achieve the aim of
this study, annual data from Susenas 2010–2012 were studied using multilevel models to account for the effect of spatial
inequality across districts on unequal Internet access among individuals. The next section presents a synthesis of the lit-
erature on spatial inequality and the digital divide.
2. Spatial inequality and the digital divide
The notion that the Internet could reduce the economic importance of geographic distance has been discussed in theliterature (Negroponte, 1995; Kelly, 1998; Cairncross, 2001). Cairncross (2001) describes the narrative of the Internet as the
death of distance. The Internet, many believe, will level the playing eld for people both near and far from the centre. It
allows people to communicate over distance and thus, lifts the constraints of geography. Some believe that this will not only
change the social world but will also effectively eliminate distance as a cost factor (Grimes, 2000). According to this view,
the economy would work in a space, rather than a place; the cost of transport would be drastically reduced, distance would
become less important, and peripheral regions would benet from opportunities that were not available in the economy
based on the manufacturing industry (Negroponte, 1995; Kelly, 1998). Since ICTs are mainly based on immaterial and human
capital investments, regions or areas that have historically suffered from isolation, high transportation costs, or a lack of
physical private and public infrastructure might nd new paths for growth. Consequently, according to this view, the
concentration of income opportunities and wealth should decrease over time (Compaine, 2001). Although other predictions
were also present in the debate on the impact of the digital economy ( Norris, 2001), this view was largely dominant.
However, in the same way that normalisation is not only the thesis used to organise evidence regarding Internet access,
the death of distance is not the only narrative by which to explain the nature of ICT access across and within countries. Infact, this relationship is not merely about geographic proximity but also reects spatial inequalities. Indeed, spatial
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎2
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
3/15
inequalities themselves are an endemic feature of regional development in most developing and transition countries
(Kanbur et al., 2014; Kanbur & Venables, 2005). In these countries, differences in economic development still shape the rate
of the diffusion of technologies at the rm, regional and country levels. The reasons behind these stylised facts have been
investigated at length in recent times. The geo-spatial digital divide could emerge as a consequence of the rise of ICTs
(Rallet & Rochelandet, 2007). In addition, the rise of Internet use may also have the potential to exacerbate any existing
spatial inequality in developing and transition countries (Norris, 2001).
The overall empirical reality is one of large geographic differences in the rate of diffusion of ICTs, with the result that
disparities and inequalities seem to be reinforced, rather than reduced by these technologies. Cross-country studies haverevealed the disparities in ICT between North America and Europe, on the one hand, and African and Asian countries on the
other (Chinn & Fairlie, 2007; Oyelaran-Oyeyinka & Lal, 2005; Pohjola, 2003). These studies have identied some mechan-
isms by which determinants producing spatial inequality relate to the digital divide. First, these large digital disparities have
been explained by differences in economic capital. Disparities of economic capital distribution such as income and gross
domestic product are positively associated with the Internet divide ( Norris, 2001; Van Dijk, 2012). As the Internet has
become increasingly central to life, work and play – providing job opportunities, strengthening community networks and
facilitating educational advancement – the systematic exclusion of certain groups and areas, such as poorer regions and
communities, becomes even more important (Norris, 2001).
Second, inequality in human capital increases unequal access to the Internet. The unequal development of human capital –
meaning investment in digital skills and capacities through education, training, and lifelong learning – represents one of the
most important factors that facilitate Internet access. Education is one of the most signicant forms of social development,
producing the skills and experiences that are most likely to contribute to the use of ICTs. Academic institutions may also play
an important role in spreading ICTs because they are often among the rst institutions in a nation to become wired. Unequaldistribution and unequal access to education may therefore lead to a digital divide. Wilson, Wallin, and Reiser (2003),
explained that the numbers of Internet users are greatly affected by whether access is offered in schools, community centres,
cybercafés, and/or post-of ces, especially in poor countries where computer access at work and home is highly limited.
Third, disparities in the distribution of telecommunications infrastructures result in a digital divide ( Rao, 2005; Mariscal,
2005). Individuals need access to computers, landlines, mobile phones and networks in order to access the Internet.
Landline networks and satellite facilities cannot be deployed as extensively as they can in developed countries. In many poor
countries such as Indonesia, basic telecommunications services are still unavailable to some people on remote islands.
Unequal access to telecommunications infrastructures and services constitutes one of the challenges to bridging the digital
divide, and remedying it should be an objective of all stakeholders in developing countries.
3. Internet growth and regional development in Indonesia
Indonesia provides an interesting case for the examination of spatial inequality and the Internet divide in developing
countries. Internet growth in Indonesia has shown promising trends in the last two decades. Nielsen’s Southeast Asian
Digital Consumer (2013), reports that Indonesia’s Internet penetration rate is at 21%, growing at 20% annually since 2003.
Semiocast (2013), reported that the country was one of the biggest users of Twitter and Facebook after the United States,
India and Brazil. In 2012, there were 71.19 million Internet users in Indonesia or about 28% of Indonesia ’s population; these
numbers are in line with world Internet growth. The country is also often cited as an emerging economic success in
Southeast Asia (World Bank, 2008). Indonesia’s economic growth has been robust since the Asian nancial crisis in 1998,
and it appears well positioned with an average annual growth of 4–6% since 2002 (World Bank, 2008). The poverty
headcount ratio at $1.25 (PPP) decreased sharply from 47.7% in 1999 to 16.2% in 2011. The Human Development Index also
increased sharply from 0.479 in 1990 to 0.624 in 2011 ( World Bank, 2013).
Despite its impressive economic growth, Indonesia’s socio-economic development has been characterised by deep
spatial inequalities across regions. Biro Pusat Statistik (2014), recently reported that the overall Gini Index increased from
0.33 in 2002 to 0.41 point in 2013. Inequality is greater in urban areas with patterns closely aligned to total trends, whilerural inequality is consistently lower than urban inequality by approximately 7 points between 2002 and 2013 (Biro Pusat
Statistik, 2014). As Indonesia is characterised by many remote and isolated areas due to the archipelagic nature of the
country, geographic location, along with the relative openness of trade and market operations, can restrict growth and the
development process, allowing some areas to develop faster than others (Vidyattama, 2010). Gaps in human capital and
infrastructure endowments, which are often themselves a product of differing development experiences, hinder current and
future development and thereby exacerbate regional disparities (Vidyattama, 2010). Equality in access to education is still an
issue in Indonesia (Suryadarma, Suryahadi, Sumarto & Rogers, 2006). Although national data show a net school enrolment
ratio of 98% for primary schools in 2010, enrolment is still only 86.2% for junior high school and 56% for senior high school.
Gaps in telecommunications infrastructure across regions also appear, with 80% of the country’s Internet activities taking
place in Java while 60% of the country’s population occupies the island. Indonesia has six major islands and thousands of
smaller ones, and the majority of the central and eastern parts of the country have yet to see consistent electricity con-
nections, let alone Internet (Biro Pusat Statistik, 2012).
The socio-economic history of each of Indonesia’s regions plays a role in the disparities seen among them, particularly
with regard to the presence or lack of political and governing institutions to facilitate development (Hill et al., 2008). The
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010–2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
4/15
presence of natural resources also plays a signicant role in regional disparities in Indonesia, with regions that are heavily
endowed with abundant natural resources such as oil and gas more likely to have higher GDP per capita than those without.
This study aims to understand whether and to what extent these determinants producing spatial inequality are associated
with the Internet divide in Indonesia.
Table 1
Sample characteristics.
2010 2011 2012
Mean7SD or % Correlation with
Internet divide
Mean7SD or % Correlation with
Internet divide
Mean7SD or % Correlation with
Internet divide
Internet divide
Access Internet 9% 10% 12%Not access Internet 91% 90% 88%
District
GDP (in Trillion rupiah) 18.8730.2 0.320* 19.9731.2 0.411* 21.2732.4 0.121*
Gini index 0.3870.042 0.216* 0.4170.046 0.221* 0.4170.048 0.190*
Disparity in human capital 0.2670.06 0.210* 0.2570.07 0.310* 0.2570.06 0.198*
Disparity in telecommunication
infrastructures
Electricity access 0.567 0.12 0.561* 0.420* 0.217 0.29 0.431*
Landline networks 0.687 0.35 0.213* 0.210* 0.677 0.34 0.180*
Mobile phone access 0.237 0.15 0.242* 0.210* 0.167 0.15 0.210*
Cybercafé access 0.897 0.19 0.104* 0.105* 0.727 0.30 0.113*
Base transceiver station 0.897 0.14 0.113* 0.141* 0.827 0.13 0.128*
Mobile phone signal coverage 0.357 0.27 0.321* 0.220* 0.117 0.21 0.230*
Disparity in education services 0.927 0.12 0.210* 0.231* 0.907 0.13 0.214*
Spending for telecommunicationservices and infrastructures (in
Billion rupiah)
140722.5 0.109* 130722.1 0.105* 142729.7 0.110*
Spending for education services (in
Billion rupiah)
2087149 0.218* 2087153 0.320* 2847215 0.211*
Household
Household expenditure (in Million
rupiah)
1.0570.97 0.340* 1.0770.98 0.217* 1.0970.97 0.210*
Residential status
Urban areas 45% 0.210* 41% 0.311* 49% 0.231*
Rural areas 55% 0.210* 59% 0.311* 57% 0.231*
Landline telephone
Yes 8% 0.320* 7% 0.221* 6% 0.120*
No 92% 0.320* 93% 0.221* 94% 0.120*
Computer ownership
Yes 11% 0.421* 12% 0.320* 15% 0.312*
No 89% 0.421* 88% 0.320* 85% 0.321*Mobile phone ownership
Yes 75% 0.310* 78% 0.230* 84% 0.321*
No 25% 0.310* 22% 0.230* 16% 0.321*
Individual
Age 28719 0.210* 28719 0.312* 29719 0.221*
Gender
Female 50% 0.101* 50% 0.123* 52% 0.108*
Male 50% 0.101* 50% 0.123* 48% 0.108*
Education
University 13% 0.312* 13% 0.362* 14% 0.381*
High school 35% 0.142* 36% 0.160* 37% 0.132*
Secondary school and below 52% 0.210* 51% 0.120* 49% 0.231*
Employment status
Employed 93.9% 0.111* 93.8% 0.101* 93.7% 0.104*
Unemployed 6.1%
0.111* 6.2%
0.101* 6.3%
0.104*Poverty
oUS 2$ per day 46.1% 0.310* 43.3% 0.245* 41.1% 0.110*
4US 2$ per day 53.9% 0.310* 56.7% 0.245* 58.9% 0.110*
Reported * po0.05.
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎4
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
5/15
4. Data and method
4.1. Susenas 2010– 2012 and of cial statistics
The individual data used in the analysis are taken from the National Socioeconomic Survey (Susenas) 2010–2012. Susenas
is one of the oldest and the best-regarded national representative surveys in developing countries ( Friedman & Levinsohn,
2002; Ravallion & Lokshin, 2007); it is also the only one in Indonesia that covers the whole archipelago (Pradhan, Suryahadi,
Sumarto, & Pritchett, 2001; Biro Pusat Statistik, 2009). Conducted by the government’s Central Bureau of Statistics, it has
been elded yearly since 1993 and is representative at the district level. The annual sample size is about 250,000 households
(close to 1.2 million individuals) (Biro Pusat Statistik, 2009), in all districts in the country. During the three years analysed,
more than one hundred new districts emerged as a result of district splits. In such cases, the data from the split districts
were aggregated and assigned to the original district denition. Variables to track the year and the number of ‘child’ districts
for each split were used. The 1998, pre-decentralisation, district denition frame, which comprised 292 districts were
applied. This practise follows Dreze and Sen (2002), and Kruse, Pradhan, and Sparrow (2012).
The survey instrument contains a core questionnaire, which collects information regarding the socio-demographic char-
acteristics of individuals and households, their education, labour market activities, and access to various ICTs including
landlines, mobile phones, personal computers/laptops, and the Internet. The Susenas data is linked with the Indonesian Village
Potential Census (PODES) and of cial statistics. First, PODES data provides information about the distribution of tele-
communications infrastructures such as electricity, landline networks, internet cafés, and mobile phone signal networks across
districts. PODES data also provides information on the distribution of education facilities across districts. Second, the district
GDP and Gini index data 2010–
2012 were used, all retrieved from the government’s Central Bureau of Statistics. Third, district
spending on telecommunications and education services was included in the model to determine whether more district
spending on human capital and telecommunications development relates to Internet access. Data on district spending on
telecommunications services and infrastructures as well as education services were retrieved from the Ministry of Finance.
4.2. Measures of the Internet divide
The Internet divide is measured by a dummy variable indicating individual access to the Internet. In the survey,
respondents were asked whether they had accessed the Internet in the last three months with the question “have you
accessed the Internet within the last three months?”) (Biro Pusat Statistik, 2010). The survey denes access to the Internet as
a connection made by respondents to an Internet enabling system such as a computer terminal, laptop, PC/computer, and
mobile device (Biro Pusat Statistik, 2010). The Internet divide is described using the socio-demographic characteristics of
individuals who either had or did not have access to the Internet. Differential access to the Internet is related to individuals
and their characteristics, including income level, education, employment, age and gender (Van Dijk, 2012). Table 1 presentssample characteristics in this study.
4.3. Measures of spatial inequality
Spatial inequality is dened as a disparity in resources and services due to discrepancies in social and economic factors
across geography (Kanbur & Venables, 2005). In this study, “across geography” means districts within a country. Four
measures of spatial inequality related to Internet access are used in this study. First, the Gini Index is a standard measure of
spatial inequality of economic capital (Kanbur & Venables, 2005). Studies found that a higher Gini Index related to a digital
divide (Kiiski & Matti, 2002). Second, to examine whether spatial inequality in human capital relates to a digital divide,
district disparities in education outcomes were used as a proxy for spatial inequality in human capital Kanbur and Venables
(2005). Third, spatial inequality in telecommunication infrastructure is measured by district disparities in the availability of
telecommunication infrastructures such as electricity, landline networks, mobile phones, cybercafés, mobile phone signal
networks, and base transceiver stations. Prior studies found that the Internet access between and within countries arerelated to the availability of these telecommunication infrastructures (Norris, 2001; Quibria, Ahmed, Tschang & Reyes-
Macasaquit, 2003). Fourth, spatial inequality in education services is measured by district disparities in higher education
facilities (colleges and universities). Studies suggest that the presence of colleges and universities is important support for
Internet penetration (Norris, 2001; Wilson et al., 2003; Mossberger, Tolbert & Stansbury, 2003).
The Palma ratio is adopted to measure district disparities in human capital, telecommunications services and education
services (Palma, 2011). In this study, district disparity in human capital is measured by the ratio of the largest 10% of
university graduates divided by the smallest 40%’s share within a district. Likewise, district disparity in electricity services is
measured by the ratio of the largest 10% of people who have access to electricity divided by the smallest 40% ’s share within a
district.
4.4. Individual and district control variables
Control variables include individual and district variables related to the Internet divide. Studies have identied thataccess to the Internet has been linked to a number of individual demographic and socio-economic characteristics, among
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010–2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
6/15
them age, income, education, gender, and geographic location (i.e. urban–rural, big cities-small cities, and mainland-remote-
island) (Wilson et al., 2003; Warschauer, 2003; Dutton, Shepherd & di-Gennaro, 2007; Mosberger et al., 2003). Multiple
regression analysis across countries has shown that income levels and educational attainment are identied as providing
the most powerful explanatory variables for digital media access (Quibria et al., 2003; Hilbert, 2010). As for geographic
location, people living in urban centres have more access to computer services than those in rural areas. Gender was
previously thought to provide an explanation for the digital divide, with many believing that digital media use male
dominated; however, controlled statistical analysis has shown that income, education and employment act as confounding
variables and that females actually tend to embrace digital media to a greater degree than do males with the same level of income, education and employment (Ono & Zavodny, 2003; Hilbert, 2010; Brannstrom, 2012).
Household expenditure is used as proxy of income, this information is biased and dif cult to assess in many developing
countries, particularly in subsistence farming households. Income data is typically prone to under-reporting and mea-
surement error, with the contribution of individual production and in-kind transfers often overlooked. Household expen-
diture is, thus a more accurate measure of household economic resources, both in developing and developed countries
(Deaton & Zaidi, 2002; Jorgensen, 2002). Since price levels of consumer goods and services in Indonesia vary across the
country (Strauss et al., 2004), the amount of household expenditure has been deated with the consumer price index for
urban and rural regions. Rural ination is taken to be 5% higher than urban ination (Resosudarmo & Jotzo, 2009). This
calculation produces real spending adjusted for urban and rural ination. The Consumer Prices Index data were retrieved
from the government’s Central Bureau of Statistics (Biro Pusat Statistik, 2009).
District gross domestic product is used to determine whether district economic development may relate to the Internet
divide. District spending on education and telecommunications services and infrastructures was used to examine whether
district spending capacity on telecommunications services and infrastructures, and human capital development relates to
the Internet divide. These variables are particularly important given the decentralisation process that Indonesia has been
undergoing since 2001. Decentralisation is one of the well-known features in contemporary Indonesia that cannot be
ignored, especially when discussing spatial disparity among districts in the country (World Bank, 2007). Appendix A pro-
vides detailed information about each variable.
Table 2
Internet access across social groups and geography 2010–2012.
Variables 2010 (%) 2011 (%) 2012 (%)
Age
Young (o25 ) 11 16 18
Middle (25–50) 8 8 10
Old (450) 1 1 1 Δ young –middle age 3 8 8
Δ young –old 8 15 15
Gender
Male 9 11 14
Female 8 9 11
Δmale– female 1 2 3
Education
University 36 39 44
High school 17 19 21
Secondary school and below 2 3 4
Δuniversity–secondary school 34 36 40
Δhigh school–secondary school 15 16 17
Poverty
Non poor (42US$ per day) 9 14 17
Poor (o2US$ per day) 1 4 5
Δnon-poor – poor 8 11 12 Job status
Employed 13 15 17
Unemployed 11 11 11
Δemployed–unemployed 2 4 6
City areas
Cities 21 27 31
Country side 7 7 9
Δcities–countryside 133 20 22
Main–remote islands
Main islands 9 11 13
Remote islands 1 1 3
Δmainland–remote islands 8 10 10
Urban–rural areas
Urban areas 15 19 22
Rural areas 4 4 5
ΔUrban–rural areas 11 15 17
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎6
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
7/15
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
8/15
The regression equation of the models can be written as follows. Considering an individual i nested in district j, the
model is:
β β β * = + ∑ + + µ + ϵE W X ij o j j ij ij j ij
With E ij*¼ logit (P (E ij*¼1)), W j is a set of district characteristics (i.e. Gini index, GDP, district disparity in electricity, landlinenetworks, internet cafés, etc.), X ij is a set of individual characteristics (i.e. age, gender, job status, education and household
expenditure), m j is a random intercept varying over districts with mean zero and variance sm2, ϵij is normally distributed with
zero and variance sϵ
2.
Multilevel models were carried out using Generalised Linear Latent and Mixed Models (GLLAMM) commands using Stata
13. Rabe-Hesketh and Skrondal (2012) explained that GLLAMMs are a class of multilevel latent variable models for (mul-
tivariate) responses of mixed type, including continuous responses, counts, duration/survival data, dichotomous, ordered
and unordered categorical responses and rankings. In this analysis, GLLAMM is used with logit link as the dependant
variable (Internet divide) which is binary.
The multilevel models were carried out in several steps. First, multilevel stratied models were carried out to examine
whether the effect of various types of spatial inequality on the Internet divide differs across urban–rural, city–countryside,
and remote islands–mainland islands. Second, multilevel models for pooled data between 2010 and 2012 were carried out
to estimate the effect of various types of spatial inequality on the Internet divide across years and the sample population.Third, to determine whether spatial inequalities and socio-economic groups substitute or reinforce each other, we estimated
Table 3
Results of stratied analysis.
Variables Urban Rural Cities Countryside Remote islands Main islands
Coef se Coef se Coef se Coef se Coef se Coef se
District
GDP 1.251* 0.004 1.501* 0.004 1.101* 0.002 1.411* 0.011 1.611* 0.012 1.200* 0.009
Index Gini 0.245* 0.002 0.345* 0.002 0.115* 0.001 0.211* 0.012 0.361* 0.002 0.210* 0.006Disparity in human capital 0.367* 0.004 0.470* 0.004 0.210* 0.002 0.350* 0.011 0.521* 0.011 0.350* 0.007
Disparity in telecommunication
infrastructures
Electricity access 0.251* 0.011 0.250* 0.010 0.130 0.110 0.171* 0.013 0.311* 0.011 0.250* 0.015
Landline networks 0.267* 0.012 0.361* 0.013 0.221 0.111 0.261* 0.010 0.333* 0.011 0.261* 0.016
Mobile phone access 0.354* 0.014 0.611* 0.012 0.211* 0.010 0.351* 0.011 0.612* 0.010 0.330* 0.017
Cybercafe access 0.123* 0.002 0.213* 0.002 0.121* 0.002 0.212* 0.010 0.421* 0.005 0.110* 0.008
Base transceiver station 0.156* 0.003 0.341* 0.003 0.126* 0.003 0.341* 0.009 0.456* 0.002 0.121* 0.003
Mobile phone signal coverage 0.021* 0.002 0.221* 0.001 0.022* 0.002 0.120* 0.009 0.321* 0.004 0.020* 0.004
Disparity in education services 0.132* 0.007 0.234* 0.002 0.023* 0.005 0.091* 0.006 0.422* 0.007 0.130* 0.008
Spending for telecommunication
services and infrastructures
0.111 0.108 0.201 0.101 0.100 0.101 0.110 0.209 0.311 0.136 0.120 0.108
Spending for education services 0.412* 0.006 0.512* 0.003 0.410* 0.003 0.510* 0.012 0.531* 0.013 0.410* 0.002
Household
Household expenditure 0.616* 0.0 08 0.723* 0.013 0.603* 0.014 0.812* 0.003 0.788* 0.041 0.708* 0.007Connected to landline telephone 0.483* 0.015 0.326* 0.039 0.367* 0.022 0.681* 0.011 0.659* 0.108 0.549* 0.013
Have PC/laptop 1.412* 0.011 1.468* 0.018 1.428* 0.019 1.7012* 0.010 1.459* 0.056 1.541* 0.010
Have mobile phone 1.280* 0.051 2.146* 0.058 1.096* 0.097 1.611* 0.012 1.869* 0.142 1.895* 0.039
Individual
Age 0.106* 0.000 0.095* 0.000 0.110* 0.000 0.201* 0.001 0.087* 0.002 0.101* 0.000
Female 0.462* 0.014 0.231* 0.014 0.642* 0.017 0.260* 0.011 0.298* 0.048 0.382* 0.008
University 3.480* 0.019 4.071* 0.031 3.471* 0.032 4.481* 0.012 4.379* 0.127 3.678* 0.016
High school 2.466* 0.019 2.990* 0.031 2.416* 0.031 3.463* 0.021 3.253* 0.125 2.638* 0.016
Secondary school 1.663* 0.201 2.060* 0.032 1.543* 0.032 1.762* 0.023 1.979* 0.048 1.767* 0.017
Employed 0.020* 0.001 0.037 0.017 0.058* 0.019 0.020 0.012 0.083 0.057 0.022* 0.001
Poverty 0.201* 0.003 0.011 0.013 0.204* 0.002 0.202 0.105 0.022 0.021 0.211* 0.003
Years
2011 0.010* 0.001 0.017* 0.002 0.009* 0.002 0.015* 0.002 0.014* 0.002 0.013* 0.002
2012 0.012* 0.001 0.010* 0.001 0.010* 0.002 0.010* 0.001 0.011* 0.001 0.011* 0.001
Constants 2.716* 0.053 5.407* 0.064 1.852* 0.100 2.818* 0.041 6.133 0.181 4.055*
Variance between districts 0.18 0.16 0.15 0.19 0.12 0.20
N 1144809 1450941 375141 1103421 225501 2370249
Log likelihood 123782 65752 46093 123782 6606 186987
Reported * po0.05.
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎8
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
9/15
the association of the interaction between various forms of spatial inequality and individual socio-economic characteristics
on the Internet divide. For each of the models, the estimated coef cient, standard errors, district variances, and log like-
lihood as an indicator of model t are reported. All models were estimated using maximum likelihood estimation.
5. Results
5.1. Descriptive analysis
The descriptive statistics show that the dataset is relatively balanced across the study period. The average district GDPremained relatively similar, at about IDR 18.8–21.2 trillion (1.9–2.2 billion US$). The percentage of poor people across
Table 4
Results of multilevel analysis with cross-level interaction.
Model 1 Model 2 Model 3
Coef se Coef se Coef se
District
GDP 1.121* 0.005 1.101* 0.004 1.100* 0.003
Index Gini 0.201* 0.002 0.190* 0.002 0.189* 0.001Disparity in human capital 0.211* 0.001 0.200* 0.002 0.202* 0.003
Disparity in telecommunication infrastructures
Electricity access 0.232* 0.010 0.210* 0.009 0.212* 0.012
Landline networks 0.211* 0.020 0.201* 0.021 0.209* 0.020
Mobile phone access 0.311* 0.010 0.312* 0.011 0.314* 0.010
Cybercafé access 0.124* 0.006 0.115* 0.005 0.116* 0.004
Base transceiver station 0.256* 0.003 0.143* 0.002 0.165* 0.002
Mobile phone signal coverage 0.021* 0.004 0.019* 0.003 0.018* 0.004
Disparity in education services 0.141* 0.006 0.132* 0.004 0.134* 0.002
Spending for telecommunication services and infrastructures 0.131 0.112 0.141 0.110 0.143 0.111
Spending for education services 0.211* 0.001 0.215* 0.002 0.217* 0.003
Household
Household expenditure 0.617* 0.007 0.620* 0.006 0.531* 0.005
Rural areas 0.679* 0.009 0.681* 0.006 0.682* 0.007
Remote islands
0.800* 0.023
0.811* 0.022
0.781* 0.023Cities 0.555* 0.010 0.561* 0.011 0.573* 0.012
Connected to landline telephone 0.353* 0.014 0.358* 0.015 0.342* 0.014
Have PC/laptop 1.424* 0.010 1.428* 0.011 1.425* 0.012
Have mobile phone 1.718* 0.038 1.721* 0.038 1.722* 0.037
Individual
Age 0.101* 0.000 0.111* 0.000 0.123* 0.000
Female 0.378* 0.008 0.371* 0.007 0.322* 0.006
University 3.690* 0.016 3.691* 0.014 3.692* 0.013
High school 2.635* 0.015 2.641* 0.012 2.542* 0.010
Secondary school 1.781* 0.016 1.791* 0.013 1.692* 0.011
Employed 0.032* 0.001 0.037* 0.001 0.039* 0.002
Poverty 0.211* 0.001 0.200* 0.002 0.201* 0.003
Years 0.011* 0.003 0.012* 0.004 0.012* 0.004
2011 0.013* 0.002 0.011* 0.003 0.013* 0.003
2012
First quartile of household expenditureBottom 10 electricity access 0.017* 0.001
First quartile of household expenditureBottom 10 landline networks 0.019* 0.009
First quartile of household expenditureBottom 10 mobile phone access 0.010 0.013
First quartile of household expenditureBottom 10 cybercafe access 0.101* 0.002
First quartile of household expenditureBottom 10 mobile phone signal coverage 0.050* 0.002
First quartile of household expenditureBottom 10 base transceiver station access 0.045* 0.001
High school and aboveBottom 10 electricity access 0.011* 0.001
High school and aboveBottom 10 landline networks 0.025* 0.002
High school and aboveBottom 10 mobile phone access 0.132* 0.000
High school and aboveBottom 10 cybercafé access 0.010* 0.001
High school and aboveBottom 10 base transceiver station access 0.019* 0.001
High school and aboveBottom 10 mobile phone signal coverage 0.112* 0.004
Constants 3.522* 0.040 3.522* 0.040 3.522* 0.040
Variance between districts 0.22 0.22 0.22
N 3334533 3334533 3334533
Log likelihood
187933
189910
189910
Reported *po0.05.
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010–2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 9
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
10/15
districts remained large, with about 41–46% of people consuming less than US $2 a day. Averages of age, gender, and
employment status in the survey did not change signicantly during the three years studied. Household expenditure is
shown to have increased slightly. Only a small proportion of individuals had attained university education (13–14%); the
education level of the majority was secondary school or below (49–52%). About 15% and 7% of respondents lived in big cities
and remote islands respectively.
The Gini index increases by 0.3 points during the three-year study period. The gap in the Gini index across districts is
quite large at SD 0.42–0.48. District disparities in human capital, landline networks and cybercafé presence remained large.
However, the spatial gap in electricity, mobile phones, and mobile phone signal access decreased. District spending oneducation and on telecommunications services and infrastructures varied across districts with a range between IDR 208–
284 billion (21–29 million US$) and IDR 140–142 billion (14–15 million US$) respectively.
Table 2 shows the distribution of Internet access across years, socio-economic groups and geography. Internet access is
unequal, showing divergent trends among education and poverty levels as well as across generations. The deepest divide
between highly educated and less educated individuals in Internet access in 2012 is at 40%. Inequality in the Internet access
of poor and non-poor people also deepened substantially (from 8% in 2010 to 12% in 2012). A deepening gap in Internet
access between the younger and older generations also appears, from 8% in 2010 to 15% in 2012. Gaps in Internet access also
increased between females and males (from 1% in 2010 to 3% in 2012).
5.2. Spatial distribution of Internet access
A sense of the importance of area variations in Internet access can be gained from the map in Fig. 1, which highlights
geographical disparities across districts. Most districts in urban areas in Central Java, East Kalimantan, and North Sumatrahave more widespread Internet access than other regions in Indonesia, in particular Papua, Sulawesi and small islands across
Maluku and Ambon.
Fig. 2 shows disparities in telecommunications infrastructures across districts. It indicates that districts in Papua, Kali-
mantan, South Sumatra, and Central Sulawesi have greater disparities in telecommunications infrastructures than do dis-
tricts across Java and Bali.
The next section presents the results of multilevel analyses that reveal whether spatial disparities in telecommunications
infrastructures are related to the Internet divide in Indonesia.
5.3. Strati ed analysis
Table 3 shows the results of stratied analysis. The estimates for the stratied models suggest a substantial urban–rural,
city–countryside and remote island–main island difference in Internet access. The coef cient of spatial inequality indicators
for rural, countryside and remote island areas is higher than the coef cient of urban, city and main island areas.In all models, the association of GDP on the Internet divide is signicant at 5%. Likewise, higher district spending for
education services is likely to increase individual access to the Internet with statistical signi cance at 5%. However, district
spending on telecommunications services and infrastructures appears to have no signicant effect on the Internet divide.
Having a mobile phone signicantly increases an individual’s likelihood of having Internet access. The substantial difference
in the degree to which mobile phone ownership affects Internet access is also shown between urban–rural, city–countryside
and remote island–main island areas. However, signicant associations of poverty and unemployment with the Internet
divide are only shown in urban, city and main island areas. Poverty and unemployment appear not to be statistically
signicant inuences on Internet access for rural, countryside and remote island samples.
5.4. Pooled analysis
Table 4 shows the results of multilevel analysis. Model 1 presents results from multilevel logit regression before cross
variable interaction. The portrait of the digital divide across social and economic groups is explicit. The digital divide acrossgeneration and gender is shown from the negative and signicant associations of increased age and being female. Across the
models, the older generation and females are less likely to access the Internet than the younger generation and males.
Human capital in the form of education is signicantly associated with digital access; those who have graduated at the
university level have substantially wider access to the Internet. Economic capital is also strongly and signi cantly related to
digital access. Households with higher monthly expenditures are more likely have Internet access than those with less.
Likewise, having a mobile phone, personal computer/telephone networks at home increases the likelihood of having
Internet access. The geographic divide in digital access is also found in all models. Those living in rural, countryside, and
remote areas are less likely to have access to the Internet.
5.5. Interaction analysis
Interaction terms between indicators of telecommunications infrastructure inequalities and indicators of individual
socio-economic groups enable us to examine whether spatial inequalities and socio-economic groups reinforce each other
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎10
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
11/15
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
12/15
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
13/15
Appendix A. Variables denition
Variables name Questions/explanation Sources
Internet divide Apakah pernah mengakses internet dalam 3 bulan terakhir?” (“Have you accessed the Internet
within the last three months?”). The survey denes access to the Internet as a connection made
by respondents toward Internet enabling system such as computer terminals, laptop, PC/com-
puters, and mobile devices (Biro Pusat Statistik, 2010).
Susenas
2010/12
Coding 1¼Yes 0¼No
GDP District Gross Domestic Product (in Trillion rupiah) BPS 2010/12
Gini Index Gini ratio BPS 2010/12
Disparity in human capital The ratio of the largest 10% of the people graduated from university divided by the smallest
40%’s share within a district.
Susenas
2010/12
Disparity in electricity access The ratio of the largest 10% of the people who having access to electricity divided by the
smallest 40%’s share within a district.
PODES 2010/
12
Disparity in landline networks The ratio of the largest 10% of the people who having access to landline networks divided by the
smallest 40%’s share within a district.
Susenas
2010/12
Disparity in mobile phone access The ratio of the largest 10% of the people who having access to mobile phone divided by the
smallest 40%’s share within a district.
Susenas
2010/12
Disparity in cybercafé access The ratio of the largest 10% of the villages that having access to cybercafé divided by the
smallest 40%’s share within a district.
PODES 2010/
12
Disparity in base transceiver station
distribution
The ratio of the largest 10% of the villages that having access to BTS divided by the smallest
40%’s share within a district.
PODES 2010/
12
Disparity in mobile phone signalcoverage
The ratio of the largest 10% of the villages that having been coverage by mobile phone signaldivided by the smallest 40%’s share within a district.
PODES 2010/12
Disparity in education services The ratios of the largest 10% of the villages have university divided by the smallest 40%’s share
within a district.
PODES 2010/
12
Spending for telecommunication
services and infrastructures
Total amount of budget spend for telecommunication services and infrastructures (in Billion
rupiah).
SIKD 2009/
11
Spending for education services Total amount of budget spend for education services and infrastructures (in Billion rupiah). SIKD 2009/
11
Household expenditure Monthly household expenditure (in Million rupiah) Susenas
2010/12
Rural areas An urban area is dened as the areas that have a major non-agricultural activity and function as
the urban settlements, concentration and distribution of government services, social services,
and economic activities. A rural areas is dened as the areas that have a major agricultural
activity, including the management of natural resources in the region, and function as rural
settlements, government services, social services, and economic services (Biro Pusat Statistik,
2009). Respondent lives in rural areas, coding 1¼rural 0¼urban
Susenas
2010/12
Landline telephone Respondents have landline telephone, coding 1¼Yes 0¼No Susenas2010/12
Computer ownership Respondents have computer or laptop, coding 1¼Yes 0¼No Susenas
2010/12
Mobile phone ownership Respondents have mobile phone, coding 1¼Yes 0¼No Susenas
2010/12
Age Age of respondents Susenas
2010/12
Female Respondent is female, coding 1¼female 0¼male Susenas
2010/12
University Respondents have university education, coding 1¼University 0¼others Susenas
2010/12
High school Respondents only have high school education, coding 1¼High school 0¼others Susenas
2010/12
Secondary school and below Respondents only have secondary school/primary school education, coding 1¼secondary
school/primary school education 0¼others
Susenas
2010/12
Employed Respondents are being employed, coding 1¼employed 0¼employed Susenas2010/12
Poverty Household consumes less than US $2 per day Susenas
2010/12
Cities Biro Pusat Statistik (2009) classies 60 districts in Indonesia, which is categorised as big cities.
Big cities refer to districts within a region which have function as a centre of population, gov-
ernment, commerce, and culture. Population within big cities is above 2 million people.
Countryside is district with have population 10.000 and less.
BPS 2010
district code
Main islands Mainland is districts located at ve big islands in Indonesia (Java-Bali, Sumatra, Kalimantan,
Sulawesi, and Papua). Remote island is districts, which located at small islands outside ve big
islands in Indonesia. District in mainlands in Indonesia is generally more developed than district
within remote islands due to they have better access of infrastructures, facilities and services
(Biro Pusat Statistik, 2009).
BPS 2010
district code
Remote islands (1) Districts at small islands across Sumatra, Kalimantan, Sulawesi, Maluku, Nusa Tenggara, and
Papua (e.g. Kepulauan Nias, Raja Ampat, etc.)
(2) Districts at remote areas in Papua, Kalimantan, Sulawesi, Sumatra and Nusa Tenggara main
islands (i.e. Kabupaten Puncak, Yahukimo, etc.)
BPS 2010
district code
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010–2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 13
http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008http://dx.doi.org/10.1016/j.telpol.2015.08.008
-
8/15/2019 Spatial Inequality and the Internet Divide in Indonesia
14/15
References
Akita, T., & Lukman, R. A. (1995). Interregional inequalities in Indonesia: a sectoral decomposition analysis for 1975 –92. Bulletin of Indonesian Economic Studies, 31(2), 61–68.
Ballas, D., & Tranmer, M. (2012). Happy people or happy places? A multilevel modeling approach to the analysis of happiness and well-being. International
Regional Science Review, 35(1), 70–102.Birdsell, D., Mizzio, D., Krane, D., & Cottreau, A. (1998). Web users are looking more like America. Public Perspective, 3(1), 33–35.Biro Pusat Statistik. (2009). Susenas guide book 2009. Indonesia, Jakarta: BPS.
Biro Pusat Statistik. (2010). Susenas guide book 2010. Indonesia, Jakarta: BPS.Biro Pusat Statistik. (2012). Statistic years book 2012. Indonesia, Jakarta: BPS.Biro Pusat Statistik. (2014). Statistic years book 2014. Jakarta: BPS Indonesia.Brannstrom, I. (2012). Gender and digital divide 2000–2008 in two low-income economies in Sub-Saharan Africa: Kenya and Somalia in of cial statistics.
Government Information Quarterly, 29, 60–67.Cairncross, F. (2001). The death of distance: how the communications revolution is changing our lives. Boston: Harvard Business School Press.Castells, M. (2010). The rise of the network society. West Sussex: Wiley-Blackwell Publishing.Chinn, M. D., & Fairlie, R. W. (2007). The determinants of the global digital divide: a cross-country analysis of computer and internet penetration, 59. UK: Oxford
Economic Papers, Oxford University Press, 16–44.Compaine, B. M. (2001). The digital divide: facing a crisis or creating a myth?. Massachusetts: MIT Press.Deaton, A., & Zaidi, S. (2002). Guidelines for constructing consumption aggregates for welfare analysis . Washington D.C.: World Bank.Dutta, S., & Mia, I. (2007). The global information technology report 2006 – 2007: connecting to the networked economy. New York: Palgrave Macmillan.Dutton, W. H., Shepherd, A., & di-Gennaro, C. (2007). Digital divides and choices reconguring access. National and cross-national patterns of Internet
diffusion and use. In A. Anderson, M. Brynin, J. Gershuny, & Y. Raban (Eds.), Information and communication technologies in society: e-living in a digitalEurope. London: Routledge.
Dreze, J., & Sen, A. (2002). India: development and participation. New York: Oxford University Press.Friedman, J., & Levinsohn, J. (2002). The distributional impacts of Indonesia’s nancial crisis on household welfare: a “rapid response” methodology. The
World Bank Economic Review, 16 (3), 397–
423.Galliano, D. (2005). The evolution of the spatial digital divide: from internet adoption to internet use by French industrial rms. France: Université Montesquieu-Bordeaux & LEREPS Université des Sciences Sociales Toulouse Working paper Groupement de Recherches Economiques .
Grimes, S. (2000). Rural areas in the information society: diminishing distance or increasing learning capacity?. Journal of Rural Studies, 16 (1), 13–21.Hargittai, E. (2002). Second-level digital divide: differences in people’s online skills. First Monday, 7 (4), 1–8.Hill, H. (1996). The Indonesian economy since 1966: Southeast Asia’ s emerging giant . Cambridge: Cambridge University Press.Hill, H., Resosudarmo, B., & Vidyattama, Y. (2008). Indonesia’s changing economic geography. Bulletin of Indonesian Economic Studies, 44(3), 407–435.Hilbert, M. (2010). Digital gender divide or technologically empowered women in developing countries? A typical case of lies, damned lies, and statistics.
Women' s Studies International Forum, 34(6), 479–489.Hilbert, M. (2014). Technological information inequality as an incessantly moving target: the redistribution of information and communication capacities
between 1986 and 2010. Journal of the Association for Information Science and Technology , 65(4), 821–835.Hohlfeld, T. N., Ritzhaupt, A. D., Barrona, A. E., & Kemker, K. (2008). Examining the digital divide in K-12 public schools: four-year trends for supporting ICT
literacy in Florida. Computers & Education, 51(4), 1648–1663. Jorgensen, D. (2002). Did we lose the war on poverty?. In D. Jorgensen (Ed.), Econometrics volume 3: economic growth in the information age . Cambridge,
Massachusetts: The MIT Press.Kanbur, R., & Venables, A. J. (2005). Spatial inequality and development . New York: Oxford University Press.Kanbur, R., Rhee, C., & Zhuang, J. (2014). Inequality in Asia and the Paci c: trends, drivers, and policy implications. New York: Routledge.
Kelly, K. (1998). New rules for the new economy. Middlesex: Penguin Books.Kiiski, S., & Matti, P. (2002). Cross-country diffusion of the Internet. Information Economics and Policy, 14(2), 297–310.Korupp, S. E., & Szydlik, M. (2005). Causes and trends of the digital divide. European Sociological Review, 21(4), 409–422.Kruse, I., Pradhan, M., & Sparrow, R. (2012). Marginal benet incidence of public health spending: evidence from Indonesian sub-national data. Journal of
Health Economics, 31(1), 147–157.Lessmann, V. (2014). Spatial inequality and development – is there an inverted-U relationship?. Journal of Development Economics, 106 , 35–51.Mariscal, J. (2005). Digital divide in a developing country. Telecommunications Policy, 29(6), 409–428.Mossberger, K., Tolbert, J., & Stansbury, M. (2003). Virtual reality: beyond the digital divide. Washington: Georgetown University Press.Negroponte, N. (1995). Being digital. New York: Alfred A. Knopf. Inc.Nie, N. H., & Erbring, L. (2000). Internet and society: a preliminary report . Stanford, C.A: Stanford Institute for the Quantitative Study of Society.Nielsen’s Southeast Asian Digital Consumer. (2013). Southeast Asian digital consumer habits. Singapore: Nielsen Company.Nisbet, E. C., Stoycheff, E., & Pearce, E. (2012). Internet use and democratic demands: a multinational, multilevel model of Internet use and citizen attitudes
about democracy. Journal of Communication, 62(2), 249–265.Norris, P. (2001). Digital divide: civic engagement, information poverty, and the Internet worldwide. New York: Cambridge University Press.Notten, N., Peter, J., Kraaykamp, G., & Valkenburg, P. M. (2009). Research note: digital divide across borders-a cross-national study of adolescents ’ use of
digital technologies. European Sociology Review, 25(5), 551–560.Nugroho, Y., Putri, D. A., & Laksmi, S. (2012). Mapping the landscape of the media industry in contemporary Indonesia . Jakarta: CIPG.
Ono, H., & Zavodny, M. (2003). Gender and the Internet. Social Science Quarterly, 84(1), 111–121.Oyelaran-Oyeyinka, B., & Lal, K. (2005). Internet diffusion in sub-Saharan Africa: a cross-country analysis. Telecommunications Policy, 29(7), 507–527.Palma, J. G. (2011) ‘Homogeneous middles vs. heterogeneous tails, and the end of the ‘Inverted-U’: the share of the rich is what it’s all about’, Cambridge
working papers in economics 1111 (pp. 87–153). University of Cambridge Department of Economics Cambridge. Later published in Development and
Change, 42.Pohjola, M. (2003). The adoption and diffusion of ICT across countries: patterns and determinants. In D. C. Jones (Ed.), The new economy handbook. San
Diego: Academic Press.Pradhan, M., Suryahadi, A., Sumarto, S., & Pritchett, L. (2001). Eating like which “ Joneses?” An iterative solution to the choice of a poverty line “reference
group”. Review of Income and Wealth, 47 (4), 473–487.Quibria, M. G., Ahmed, S. N., Tschang, T., & Reyes-Macasaquit, M. L. (20 03). Digital divide: determinants and policies with special reference to Asia. Journal of
Asian Economics, 13, 811–825.Rao, S. S. (2005). Bridging digital divide: efforts in India. Telematics and Informatics, 22, 361–375.Rabe-Hesketh, S., & Skrondal, A. (2012). Multilevel and longitudinal modeling using stata Third ed.). College Station, TX: Stata Press.Rallet, A., & Rochelandet, F. (2007). ICTs and inequalities: the digital divide. In E. Brousseau, & N. Curien (Eds.), Internet and digital economics: principles,
methods and applications. New York: Cambridge University Press.Ravallion, M., & Lokshin, M. (2007). Lasting impacts of Indonesia ’s nancial crisis. Economic Development and Cultural Change, 56 (1), 27–56.Resosudarmo, B. P., & Vidyattama, Y. (2006). Regional income disparity in Indonesia: a panel data analysis. ASEAN Economic Bulletin, 23(1), 31–44.
Resosudarmo, B., & Jotzo, F. (2009). Working with nature against poverty: development, resources, and the environment in Eastern Indonesia. Singapore: ISEAS.Rogers, E. M. (2003). Diffusion of innovations 5th ed.). New York: Free Press.
Please cite this article as: Sujarwoto, S., & Tampubolon, G. Spatial inequality and the Internet divide in Indonesia 2010 –2012. Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2015.08.008i
S. Sujarwoto, G. Tampubolon / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎14
http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref1http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref2http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref3http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref4http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref4http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref4http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref4http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref5http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref5http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref5http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref5http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref7http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref7http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref7http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref7http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref8http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref8http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref8http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref8http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref9http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref10http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref10http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref10http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref10http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref11http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref11http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref11http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref11http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref12http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref13http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref13http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref13http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref13http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref14http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref14http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref14http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref14http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref15http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref16http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref17http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref17http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref17http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref17http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref19http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref20http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref21http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref22http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref24http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref25http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref26http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref27http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref28http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref29http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref29http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref29http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref29http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref29http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref30http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref30http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref30http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref30http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref31http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref32http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref32http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref32http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref32http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref33http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref34http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref35http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref36http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref38http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref39http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref39http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref39http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref39http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref40http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref40http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref40http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref41http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref41http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref41http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref41http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref42http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref43http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref44http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref44http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref44http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref44http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref45http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref46http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref46http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref46http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref46http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref47http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref48http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref49http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref49http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref49http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref49http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref49http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref50http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref51http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref52http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref53http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref53http://refhub.elsevier.com/S0308-5961(15)00128-7/sbref53