Measuring Networks across Countries: an Empirical Exploration

Axel Marx and Jadir Soares

“A world in which horizontal and vertical government networks comprise different types of government institutions (regulatory, judicial, legislative), perform different functions (information exchange, enforcement cooperation, technical assistance and training), have different members, have different degrees of formality and coexists in different ways with international organizations is a messy world indeed. It may seem impossibly complex.” Anne-Marie Slaughteriii


Notwithstanding the importance of networks in the academic literature, little solid empirical data is available to measure networks on the country level and the effect on relevant outcome parameters such as government effectiveness, industrial development and/or GDP/per capita on country level. There are many excellent network studies available which quantitatively analyse network effects on organizational level, and case studies which describe the importance of networks. However, few studies are available which aim to capture the degree to which a country is ‘networked’ or connected taking into account that networks develop and are influential on distinct levels (intra-organizational networks, inter-organizational networks and international networks) (for an exception see Maoz, 2010). No overall network index currently exists which enables a comparison between countries and which substantiates the importance of networks for relevant outcomes. So, is it “impossibly complex” to measure these networks?

For the purpose of this report, we explore the possibilities for constructing such an index. Given the limited scope of the study, in terms of both duration and budget, we were able to collect little new data on the basis of surveys, expert interviews, data-mining in raw datasets of existing databases, etc. Figure 2.1, based on the work of Adcock and Collier (2001) presents the ideal-typical process of constructing and validating new concepts and indicators. Starting from the distinction between intra-organizational networks, inter-organizational networks and international networks, we therefore used an

inductive approach to construct an index of connectedness and hence to work our way from level 4 to level 1 in the concept development process. As a result, more than 70 databases (see Annex 1) containing country level data for a significant number of countries were screened for indicators which can be related to international, inter-organizational and intra-organizational networks. In total more than 7000 existing indicators were considered (Annex 2 contains more information on the variable selection

process). Some indicators were identified as being potentially relevant, i.e. proxies for indicators on the levels identified in the report, intra-organizational, inter organizational and international.

This approach has several disadvantages. First, we have to work with available data. The report develops new indicators on the basis of existing data which is not the same as gathering new data on the basis of the concept development framework outlined in figure 2.1. Chapter 1 stressed the importance of institutionalized/embedded ties for knowledge management and knowledge creation. Such a finegrained assessment is not possible if data is not specifically collected from that theoretical perspective. Working with existing data makes it difficult to differentiate between arm-length and embedded networks. Chapter 1 also argued that the ‘ecosystem’ of private sector development consists of many different actors, potentially creating a wide diversity of networks which in all likelihood is not captured in existing datasets. Secondly, while many existing variables were screened there are likely to be many more relevant databases. Future work could focus on identifying these. Nevertheless, the present report does succeed in developing an initial indicator for connectedness. Its use will shed some light on the fruitfulness of continuing the effort to develop more fine-grained measures of connectedness. In subsequent chapters, several proposals for the development of new indicators are made.

What is the result of screening more than 7000 variables with the purpose of identifying network indicators? Surprisingly few indicators are available. Figure 2.2 presents the seven variables which were selected for the purpose of the connectedness index. For international networks we aimed to identify indicators that capture the flows of information and policy diffusion between public authorities, as well as the information flows between economic actors (Slaughter, 2004; Martínez-Diaz &Woods, 2009).

Two indicators were selected to capture this degree of international connectedness, namely the KOF (Swiss Economic Institute) political globalization indicator and the KOF economic networks indicator. The political globalization index captures inter alia the membership in international inter-governmental organizations and the number of international treaties which are signed and ratified by a country.

The economic networks indicator measures the actual economic and financial flows between countries (trade, FDI, portfolio investments). Several other economic indicators capture economic flows, but the KOF is the most comprehensive and suitable one for the purpose of this report.

Three variables were selected to capture the degree of inter-organizational interconnectedness within a country, namely university-industry collaboration, networks and supporting industries and the degree to which individuals are members of professional organizations which are often established for networking purposes. The first two indicators are drawn from the Global Competitiveness Report.

University industry collaboration measures the extent to which business and research collaborate on research and development. It captures the networks between business and universities, when working together pursuing innovations. Networks and supporting industries captures the number and quality of local suppliers and the extent of their interaction (i.e. clusters, or the concentration of interconnected businesses). Both are in the literature on inter-organizational networks and economic geography recognized as important indicators to capture the degree of connectedness between these

organizations. (Podolny & Page, 1998; Powell &Smith-Doerr, 1994; Saxenian et al. 2001; European Commission, 2008) The third indicator is drawn from the World Values Survey and aims to capture networks of professionals that collaborate each other for specific purposes. Networking in the context of professional association can be regarded as a relevant networking strategy in the context of information exchange (see Burt, 1995; Baker, 2000; Putnam, 2000 for a more general argument on the importance of association).

Intra-organizational networks are hard to capture. To measure intra-organizational networks we identified two proxies based on the degree to which firms offer training (Cross & Parker, 2004). The idea is that training enhances internal networks and learning resulting from increased interaction between people within an organization. One measure comes from the World Bank Enterprise Surveys and measures the percentages of firms offering formal training. A second measure is based on the Global

Competitiveness report and focuses on-the-job training which is in turn based on the local availability of specialized research and training services in a country and the extent to which companies in a country invest in training and employee development.

The indicators will be discussed more extensively in the following sections. Figure 2.2 presents the different components of the connectedness index.

Connectedness Index

International networks

Economic Globalization (KOF)

Political Globalization (KOF)

Inter-organizational networks

University Industry Collaboration (GCR)

Networks and Supporting Industries (GCR)

Professional Association (WVS)

Intra-organizational networks

Firms Offering Training (WB-ES)

On the Job Training (GCR)

To analyse the relationship with relevant outcome variables, the report focuses on four variables, namely two policy-related variables (government effectiveness and regulatory quality) and two economy-related variables (industrial development and GDP per capita). Government effectiveness and regulatory quality are chosen since networks are assumed to contribute to better policy formulation and implementation (see discussion in Part 1).

Government effectiveness and regulatory quality in turn are important for better private sector development and economic development, the ultimate parameters in which we are interested (see also Altenburg (2011, pp. 35-36)). Government effectiveness, from the World Bank governance indicators series, captures different aspects of policymaking and implementation, including the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such

policies. The link with private sector development is specifically made in the concept of regulatory quality, from the World Bank governance indicators series, which refers to the ability of governments to formulate and implement sound policies and regulations that permit and promote private sector development (Kaufman et al. 2009). The UNIDO Competitive Industrial Performance (CIP) Index benchmarks competitive industrial activity at the country level and is an indicator for industrial development. GDP per capita from the World Development Indicators is included as a second general measure for economic development.

The analysis will focus on the one hand on analysing the variation in the connectedness index and its subindices, and on the other hand on the relationship with other relevant parameters such as policy effectiveness, industrial development and economic development, without implying any causal relationship. Data on the latter indicators is drawn from the World Bank governance indicators, the UNIDO Competitive Industrial Performance Index and World Bank development indicators on GDP per capita PPP. Table 2.1 presents the variables which are used to compose the connectedness index, the sources from which they are drawn and the name of the variable in the source database.

Table 2.1: Variables and Sources of the Connextedness Index
Variable Source Source variable
Political Networks KOF Index of Globalization Political Globalization
Economic Networks KOF Index of Globalization Actual flows in economic terms
University-Firm Networks Global Competitiveness Report University-industry collaboration in R&D
Inter-firm Networks Global Competitiveness Report Networks and supporting industries
Personal Networks World Values Survey A072: Member of professional associations or
A104: Active/inactive membership of professional organization
Formal Training Enterprise Surveys L.10: Over fiscal year … [last complete fiscal year], did this establishment have formal training programs for its permanent, full-time employees?
On-the-job Training Global Competitiveness Report On-the-job training
Government Effectiveness Worldwide Governance Indicators Government effectiveness
Regulatory Quality Worldwide Governance Indicators Regulatory quality
Competitive Industrial Performance (CIP) Industrial Development Report Competitive industrial performance
GDP per capita World Development Indicators GDP per capita, PPP (current international $) (NY.GDP.PCAP.PP.CD)


The discussion of the available indicators makes clear that several potential relevant networks are currently not captured in the datasets which were screened in the context of this report. The private sector development ecosystem is such that many types of actors can form relevant knowledge networks. Most importantly no indicators are available, to our knowledge, which capture the degree to which governmental structures are connected to the ‘private’ sector in a country, neither on the level of interportfolio organizational networks nor on the intraorganizational level (for example number of bureaucrats with significant private sector experience).

The following sections discuss the different subindices, the connectedness index and the relationship with relevant other variables, government effectiveness, CIP and GDP/per capita.

2.2. The international networks sub-index

The International Networks sub-index is based on two indicators from the KOF Index of Globalization, political and economic globalization. Political globalization is a proxy for the degree to which states are networked on an international level.

This indicator is based on the number of embassies in a country, the number of international organizations of which the country is a member, the number of UN peace missions in which a country participated, and the number of international treaties a country signed (Dreher, 2006). The proxy for economic globalization (networks) is based on the flows of goods and services (KOF actual flows). This indicator takes into

account the exports and imports of goods and services, foreign direct investments (FDI stocks), the interportfolio of investments of a country, and the income payments to foreign nationals.

After the selection of the indicators, the International Networks Sub-index was created based on the arithmetic mean of political and economic networks, transformed on a scale from 0-1.The sub-index of International Networks covering 121 countries is presented in table 2.2.

Table 2.2: International Networks Sub-index
ISO code Country International Network Index International Network Rank ISO code Country International Network Index International Network Rank
BEL Belgium 1.000 1 CYP Cyprus 0.837 17
NLD Netherlands 0.963 2 CHL Chile 0.833 18
HUN Hungary 0.940 3 NOR Norway 0.831 19
IRL Ireland 0.935 4 ESP Spain 0.829 20
CHE Switzerland 0.934 5 BGR Bulgaria 0.820 21
AUT Austria 0.929 6 ETH Ethiopia 0.812 22
SWE Sweden 0.920 7 SVK Slovakia 0.788 23
LUX Luxembourg 0.906 8 CAN Canada 0.787 24
DNK Denmark 0.904 9 EST Estonia 0.787 25
PRT Portugal 0.862 10 ITA Italy 0.787 26
CZE Czech Republic 0.852 11 SVN Slovenia 0.775 27
FIN Finland 0.851 12 ISL Iceland 0.768 28
SGP Singapore 0.849 13 TUN Tunisia 0.757 29
MYS Malaysia 0.844 14 JOR Jordan 0.753 30
FRA France 0.840 15 AUS Australia 0.736 31
DEU Germany 0.837 16 HRV Croatia 0.735 32


POL Poland 0.730 33 PNG Papua New Guinea 0.520 79
ZAF South Africa 0.730 34 BIH and Herzegovina 0.519 80
GRC Greece 0.728 35
NZL New Zealand 0.726 36 TUR Turkey 0.514 81
PAN Panama 0.725 37 SEN Senegal 0.508 82
THA Thailand 0.719 38 KGZ Kyrgyzstan 0.506 83
ISR Israel 0.718 39 IND India 0.498 84
NGA Nigeria 0.714 40 JPN Japan 0.498 85
GBR United Kingdom 0.696 41 GTM Guatemala 0.493 86
MLT Malta 0.690 42 MEX Mexico 0.487 87
ZMB Zambia 0.687 43 AZE Azerbaijan 0.485 88
JAM Jamaica 0.686 44 CRI Costa Rica 0.475 89
KAZ Kazakhstan 0.681 45 MDA Moldova 0.472 90
LTU Lithuania 0.675 46 PRY Paraguay 0.468 91
USA United States 0.673 47 ALB Albania 0.464 92
PER Peru 0.666 48 CHN China 0.460 93
ZWE Zimbabwe 0.657 49 BWA Botswana 0.454 94
URY Uruguay 0.654 50 BRB Barbados 0.446 95
BHR Bahrain 0.651 51 MDG Madagascar 0.446 96
ROU Romania 0.647 52 MKD Macedonia 0.445 97
UKR Ukraine 0.646 53 PAK Pakistan 0.445 98
KOR Korea, 0.639 54 GEO Georgia 0.443 99
NAM Namibia 0.626 55 MLI Mali 0.442 100
MAR Morocco 0.610 56 TCD Chad 0.425 101
RUS Russian Fed 0.604 57 DOM Dominican 0.420 102
ARG Argentina 0.602 58
BOL Bolivia, 0.590 59 OMN Oman 0.418 103
BRA Brazil 0.583 60 Sri Lanka 0.408 104
PHL Philippines 0.580 61 Kuwait 0.400 105
MRT Mauritania 0.577 62 CMR Cameroon 0.392 106
BLZ Belize 0.566 63 NIC Nicaragua 0.384 107
SLV El Salvador 0.565 64 VEN Venezuela, 0.377 108
EGY Egypt 0.563 65 Bolivarian Republic of
ECU Ecuador 0.560 66 LSO Lesotho 0.358 109
IDN Indonesia 0.556 67 KEN Kenya 0.352 110
COL Colombia 0.554 68 UGA Uganda 0.339 111
KHM Cambodia 0.552 69 BGD Bangladesh 0.305 112
CIV Côte d’Ivoire 0.551 70 ARM Armenia 0.281 113
MUS Mauritius 0.549 71 BEN Benin 0.278 114
HND Honduras 0.539 72 MWI Malawi 0.277 115
DZA Algeria 0.539 73 CAF CAR 0.262 116
LVA Latvia 0.538 74 SYR Syrian Arab Republic 0.260 117
TTO Trinidad 0.538 75 BFA Burkina Faso 0.255 118
SRB Serbia 0.530 76 BDI Burundi 0.119 119
MOZ Mozambique 0.529 77 TZA Tanzania 0.091 120
GUY Guyana 0.528 78 HTI Haiti 0.000 121


The international sub-index shows significant variation in the degree to which countries are linked to each other on the international level, both politically as well as economically. The comparison with the median indicates that a significant proportion (more than 50 per cent) of the countries achieve relatively high scores on the sub index. However, several countries also receive lower scores and are outside the international dynamics between countries. It should be noted that a score of zero does not imply that a country is totally unconnected, but that – taken the variation between countries into account and due to the re-scaling of the variables, which is necessary for index-creation (see annex 2) – a country has a score of zero, indicating that in comparison to other countries the international connectedness is very low.

2.3 The interorganizational networks sub-index

The Inter-organizational Networks Sub-index was created based on following three indicators. First, the indicator on networks and supporting industries is taken from the Global Competitiveness Report 2008.

This indicator is based on an Executive Opinion Survey, and takes into account the quality and quantity of local suppliers, and the state of cluster development. The University-Industry Collaboration indicator is also taken from the Global Competitiveness Report that measures to what extent business and universities collaborate on research and development (R&D) in a country. Finally, the professional association indicator, which captures the degree to which individuals are involved in professional associations, was taken from the World Values Survey.

The Inter-organizational Networks Sub-index was created by the arithmetic mean the three indicators, transformed on a scale from 0-1. The Interorganizational Networks sub-index, covering 81 countries, is presented in table 2.3.

Table 2.3: Inter-organizational Networks Index
ISO code Country International Network Index International Network Rank ISO code Country International Network Index International Network Rank
USA United States 1.000 1 NOR Norway 0.798 9
CHE Switzerland 0.976 2 IND India 0.795 10
SWE Sweden 0.874 3 NLD Netherlands 0.784 11
DEU Germany 0.865 4 GBR United Kingdom 0.781 12
FIN Finland 0.845 5 SGP Singapore 0.760 13
CAN Canada 0.823 6 AUS Australia 0.749 14
TWN Taiwan, Province of China 0.817 7 KOR MYS Korea, Republic of Malaysia 0.730 0.688 15 16
JPN Japan 0.807 8 HKG Hong Kong SAR, China 0.658 17


NZL New Zealand 0.629 18
FRA France 0.616 19
ZAF South Africa 0.607 20
CHN China 0.601 21
CZE Czech Republic 0.593 22
PRI Puerto Rico 0.585 23
ISR Israel 0.584 24
THA Thailand 0.577 25
ARM Armenia 0.567 26
IDN Indonesia 0.550 27
ITA Italy 0.534 28
SVN Slovenia 0.513 29
BRA Brazil 0.508 30
CHL Chile 0.500 31
ESP Spain 0.494 32
HUN Hungary 0.464 33
EST Estonia 0.457 34
CYP Cyprus 0.452 35
SAU Saudi Arabia 0.436 36
COL Colombia 0.413 37
DOM Dominican 0.408 38
LTU Lithuania 0.403 39
SVK Slovakia 0.401 40
MEX Mexico 0.396 41
GTM Guatemala 0.388 42
JOR Jordan 0.385 43
VNM Viet Nam 0.383 44
TUR Turkey 0.381 45
TTO Trinidad 0.374 46
and Tobago
HRV Croatia 0.364 47
ZMB Zambia 0.356 48
PHL Philippines 0.344 49
UKR Ukraine 0.344 50


ARG Argentina 0.335 51
POL Poland 0.323 52
UGA Uganda 0.322 53
NGA Nigeria 0.313 54
MLI Mali 0.312 55
RUS Russian Federation 0.306 56
EGY Egypt 0.297 57
PER Peru 0.295 58
AZE Azerbaijan 0.294 59
ROU Romania 0.279 60
MAR Morocco 0.276 61
TZA Tanzania 0.273 62
LVA Latvia 0.257 63
PAK Pakistan 0.255 64
SRB Serbia 0.247 65
BGR Bulgaria 0.241 66
BFA Burkina Faso 0.236 67
URY Uruguay 0.221 68
BGD Bangladesh 0.215 69
GHA Ghana 0.209 70
ETH Ethiopia 0.207 71
MKD Macedonia 0.201 72
SLV El Salvador 0.198 73
VEN Venezuela, 0.152 74
ZWE Zimbabwe 0.113 75
DZA Algeria 0.075 76
KGZ Kyrgyzstan 0.069 77
GEO Georgia 0.064 78
BIH Bosnia and Herz 0.062 79
ALB Albania 0.026 80
MDA Moldova 0.000 81


The inter-organizational sub-index also varies significantly between countries. Inter-firm networks (clusters), firm-university networks and personal networks are very highly developed in some countries but underdeveloped in a large number of countries. The median indicates that overall the degree of interorganizational interconnectedness is below 0.5, indicating that a significant number of countries have less developed inter-organizational networks as operationalized in the inter-organizational network sub-index. In our sample, it is partly a consequence of the low level of personal networks measures by the professional association indicator. It should be stressed that this is only a very partial operationalization on the basis of available data, which does not take into account several other elements which could be important in terms of interorganizational networks, most importantly the links between other actors of the private sector development eco-system which are not included in the sub-index. Again, the zero score does not indicate a complete absence of inter-organizational networks, but is a result of the re-scaling method, indicating a comparatively low level of inter-organizational connectedness.

2.4 The intraorganizational network sub-index

The Intra-organizational Networks Sub-index was created based on two indicators. The Percentage of Firms Offering Formal Training comes from the World Bank Enterprise Surveys, most specifically from the question L10 which assessed whether an establishment offered formal training programs for its permanent, full-time employees.

The On-the-job Training indicator from the Global Competitiveness Report 2008-2009 is based on the local availability of specialized research and training services and the extent to which companies invest in training and employee development.

The Intra-organizational Networks sub-index was created by the arithmetic mean of the two training indicators. The index, covering 163 countries, is presented in table 2.4.

Table 2.4: Intra-organizational Networks Index
ISO code Country International Network Index International Network Rank
CHE Switzerland 1.000 1
DNK Denmark 0.975 2
USA United States 0.972 3
SWE Sweden 0.940 4
NLD Netherlands 0.908 5
SGP Singapore 0.893 6
WSM Samoa 0.890 7
FIN Finland 0.886 8
JPN Japan 0.880 9
BEL Belgium 0.833 10
CAN Canada 0.817 11
GBR United Kingdom 0.817 12
FRA France 0.804 13
NOR Norway 0.801 14
ISL Iceland 0.789 15
AUS Australia 0.766 16
IRL Ireland 0.759 17
AUT Austria 0.757 18
CHN China 0.751 19
LBN Lebanon 0.747 20
SVK Slovakia 0.736 21
TWN Taiwan, 0.725 22
ISR Israel 0.716 23
SVN Slovenia 0.700 24
NZL New Zealand 0.678 25
EST Estonia 0.666 26
CZE Czech Republic 0.662 27
FJI Fiji 0.660 28
LUX Luxembourg 0.656 29
HKG Hong Kong SAR, China 0.646 30
PRI Puerto Rico 0.646 31
TUN Tunisia 0.646 32
THA Thailand 0.640 33
FSM Micronesia, 0.626 34


MYS Malaysia 0.608 35
DEU Germany 0.608 36
KOR Korea, Republic of 0.573 37
BRA Brazil 0.572 38
QAT Qatar 0.552 39
ARE United Arab Emirates 0.542 40
CRI Costa Rica 0.540 41
LTU Lithuania 0.536 42
SWZ Swaziland 0.533 43
KEN Kenya 0.523 44
ZAF South Africa 0.517 45
ESP Spain 0.506 46
POL Poland 0.503 47
VUT Vanuatu 0.489 48
SAU Saudi Arabia 0.489 49
BRB Barbados 0.485 50
CHL Chile 0.485 51
GRD Grenada 0.473 52
JAM Jamaica 0.464 53
LVA Latvia 0.458 54
CYP Cyprus 0.451 55
ARG Argentina 0.450 56
BLR Belarus 0.450 57
PER Peru 0.449 58
DOM Dominican Republic 0.433 59
SLV El Salvador 0.430 60
CPV Cape Verde 0.426 61
PAN Panama 0.422 62
PHL Philippines 0.407 63
KWT Kuwait 0.403 64
MWI Malawi 0.403 65
ITA Italy 0.394 66
ECU Ecuador 0.394 67
VNM Viet Nam 0.393 68
LKA Sri Lanka 0.388 69
PRT Portugal 0.387 70
BHR Bahrain 0.378 71
IDN Indonesia 0.378 72
MLT Malta 0.366 73
ROU Romania 0.364 74
COL Colombia 0.364 75
HUN Hungary 0.362 76
COD Congo, Democratic 0.362 77
BHS Bahamas 0.357 78
MKD Macedonia 0.354 79
SRB Serbia 0.354 80
GTM Guatemala 0.348 81
IND India 0.345 82


NAM Namibia 0.342 83
RUS Russian Federation 0.340 84
HRV Croatia 0.338 85
LSO Lesotho 0.334 86
KHM Cambodia 0.329 87
CMR Cameroon 0.325 88
VEN Venezuela, 0.325 89
Bolivarian Republic of
TTO Trinidad 0.324 90
and Tobago
NER Niger 0.323 91
JOR Jordan 0.322 92
MUS Mauritius 0.321 93
UGA Uganda 0.321 94
MNG Mongolia 0.320 95
BOL Bolivia, 0.316 96
Plurinational State of
HND Honduras 0.315 97
BWA Botswana 0.313 98
BFA Burkina Faso 0.306 99
MNE Montenegro 0.297 100
KGZ Kyrgyzstan 0.293 101
KAZ Kazakhstan 0.293 102
NGA Nigeria 0.292 103
BGR Bulgaria 0.291 104
TUR Turkey 0.286 105
TLS Timor-Leste 0.283 106
MEX Mexico 0.282 107
BIH Bosnia and 0.280 108
TGO Togo 0.279 109
GMB Gambia 0.279 110
GAB Gabon 0.278 111
OMN Oman 0.276 112
TZA Tanzania 0.275 113
MAR Morocco 0.268 114
AZE Azerbaijan 0.265 115
BEN Benin 0.255 116
UKR Ukraine 0.255 117
GHA Ghana 0.253 118
SEN Senegal 0.250 119
LAO Lao People’s 0.243 120
PRY Paraguay 0.243 121
URY Uruguay 0.241 122
RWA Rwanda 0.236 123
CIV Côte d’Ivoire 0.234 124
MDG Madagascar 0.228 125


ARM Armenia 0.224 126
GRC Greece 0.224 127
ETH Ethiopia 0.223 128
WBG West Bank 0.222 129
and Gaza Strip
ERI Eritrea 0.217 130
ZMB Zambia 0.215 131
EGY Egypt 0.208 132
MDA Moldova 0.208 133
ZWE Zimbabwe 0.208 134
TCD Chad 0.204 135
NIC Nicaragua 0.203 136
KOS Kosovo 0.198 137
GUY Guyana 0.195 138
MOZ Mozambique 0.195 139
SYR Syrian Arab Republic 0.182 140
BTN Bhutan 0.182 141
MLI Mali 0.169 142
LBY Libyan 0.167 143
ALB Albania 0.165 144
GIN Guinea 0.154 145
GEO Georgia 0.142 146
AGO Angola 0.132 147
TJK Tajikistan 0.124 148
SLE Sierra Leone 0.122 149
SUR Suriname 0.119 150
BDI Burundi 0.108 151
MRT Mauritania 0.106 152
BGD Bangladesh 0.104 153
LBR Liberia 0.101 154
DZA Algeria 0.093 155
AFG Afghanistan 0.071 156
PAK Pakistan 0.056 157
YEM Yemen 0.050 158
GNB Guinea-Bissau 0.044 159
COG Congo 0.031 160
TON Tonga 0.027 161
UZB Uzbekistan 0.008 162
NPL Nepal 0.000 163


The intra-organizational sub-index varies significantly between countries. The low median score indicates that these instruments to strengthen internal networks are less widespread among countries. A limited number of countries achieve high scores, while a large group of countries receive lower scores, as is indicated by the median. Again, the zero score does not indicate a complete absence of intraorganizational networks, but is a result of the rescaling method, indicating a comparatively low level of intra-organizational connectedness.

2.5 The Connectedness Index

The Connectedness Index is the average of three sub-indices (International, Inter-organizational, and Intra-organizational Networks). It is presented in table 2.5.

Table 2.5: Connectedness Index
ISO code Country International Inter-org Intra-org Connectedness Connectedness
Network Index Network Index Network Index Index Rank
CHE Switzerland 0.934 0.976 1.000 0.970 1
SWE Sweden 0.920 0.874 0.940 0.911 2
NLD Netherlands 0.963 0.784 0.908 0.885 3
USA United States 0.673 1.000 0.972 0.881 4
FIN Finland 0.851 0.845 0.886 0.861 5
SGP NOR CAN SingaporeNorwayCanada 0.8490.8310.787 0.7600.7980.823 0.8930.8010.817 0.834 0.810 0.809 6 7 8
DEU GBR FRA GermanyUnited KingdomFrance 0.837 0.6960.840 0.8650.7810.616 0.6080.8170.804 0.770 0.7650.754 9 10 11
AUS Australia 0.736 0.749 0.766 0.750 12
JPN MYS CZE NZL JapanMalaysiaCzech RepublicNew Zealand 0.8980.844 0.852 0.726 0.8070.6880.5930.629 0.880 0.6080.6620.678 0.728 0.713 0.7020.678 13 14 15 16
ISR Israel 0.718 0.584 0.716 0.673 17
SVN Slovenia 0.775 0.513 0.700 0.662 18
KOR THA Korea, Republic of Thailand 0.6390.719 0.730 0.577 0.573 0.640 0.648 0.646 19 20
SVK Slovakia 0.788 0.401 0.736 0.642 21
EST Estonia 0.787 0.457 0.666 0.637 22
ZAF South Africa 0.730 0.607 0.517 0.618 23
ESP CHL SpainChile 0.829 0.833 0.494 0.500 0.5060.485 0.6100.606 2425


CHN China 0.460 0.601 0.751 0.604 26
HUN CYP ITA BRA HungaryCyprus ItalyBrazil 0.940 0.837 0.787 0.583 0.4640.4520.5340.508 0.362 0.4510.3940.572 0.5890.5800.5720.554 27 282930
IND India 0.498 0.795 0.345 0.546 31
LTU Lithuania 0.675 0.403 0.536 0.538 32
POL Poland 0.730 0.323 0.503 0.519 33
IDN Indonesia 0.556 0.550 0.378 0.494 34
JOR Jordan 0.753 0.385 0.322 0.487 35
HRV Croatia 0.735 0.364 0.338 0.479 36
PER Peru 0.666 0.295 0.449 0.470 37
ARG BGR COL Argentina BulgariaColombia 0.602 0.820 0.554 0.3350.2410.413 0.4500.2910.364 0.4630.451 0.444 383940
PHL NGA ROU PhilippinesNigeriaRomania 0.580 0.714 0.647 0.344 0.313 0.279 0.4070.2920.364 0.4440.440 0.430 414243
DOM ZMB DominicanRepublicZambia 0.4200.687 0.408 0.356 0.4330.215 0.420 0.419 44 45
LVA Latvia 0.538 0.257 0.458 0.417 46
RUS Russian Federation 0.604 0.306 0.340 0.417 47
UKR Ukraine 0.646 0.344 0.255 0.415 48
ETH TTO GTM EthiopiaTrinidad and Tobago 








0.414 0.412 




SLV El Salvador 0.565 0.198 0.430 0.398 52
TUR MEX TurkeyMexico 0.5140.487 0.381 0.396 0.2860.282 0.3940.388 5354
MAR Morocco 0.610 0.276 0.268 0.385 55
SRB Serbia 0.530 0.247 0.354 0.377 56
URY ARM UruguayArmenia 0.654 0.281 0.2210.567 0.2410.224 0.3720.357 5758














UGA ZWE UgandaZimbabwe 0.3390.657 0.3220.113 0.3210.208 0.3270.326 6263
MLI Mali 0.442 0.312 0.169 0.308 64
KGZ BIH VEN KyrgyzstanBosnia and Herzegovina Venezuela,Bolivarian Republic of 0.506 0.519 0.377 0.0690.062 0.152 0.2930.2800.325 0.2890.2870.285 6566 67
BFA Burkina Faso 0.255 0.236 0.306 0.266 68
PAK Pakistan 0.445 0.255 0.056 0.252 69
DZA MDA AlgeriaMoldova 0.5390.472 0.0750.000 0.093 0.208 0.236 0.227 70 71
ALB Albania 0.464 0.026 0.165 0.218 72
GEOTZA GeorgiaUnited Republic of Tanzania 0.4430.091 0.0640.273 0.142 0.275 0.216 0.213 7374
BGD Bangladesh 0.305 0.215 0.104 0.208 75


The connectedness index clearly shows the overall variation in the degree to which countries are networked, both internally as well as internationally (for a discussion on using the median for comparison purposes see annex 1). Some countries obtain consistently high scores across the various network indicators and hence on the connectedness index, whereas other receive consistently lower scores. Also, it is interesting to note that similar connectedness scores were reached following very distinct paths. For example, Hungary (0.589) and Brazil (0.554) occupy the 27th and 30th ranking positions, respectively.

However, while Brazil is very consistent in the three components of connectedness (0.583 for International Networks, 0.508 for Inter-organizational Networks, and 0.572 for Intra-organizational Networks), the scores of Hungary vary significantly: a very high score is achieved (0.940) in the International Networks Subindex, a mean score in the case of the Inter-organi -zational Networks Sub-index, and a low score (0.362) in the Intra-organizational Networks Sub-index. The similar result in the Connectedness index is, in part, a consequence of our choice of the aggre gation procedure (equal weighting) that uses a full compensability system, i.e., a low score in one indicator is equally compensated by a high score in other.

More generally, the differences on country level between indices are interesting. Some Asian countries, such as Japan and China, score below median on international networks but (very) highly on interorganizational and intra-organizational networks. Others, including some European countries such as Poland and Hungary, score highly on international networks but show only median scores on interorganizational and intra-organizational networks. Still others, such as India, score very highly on one indicator, in casu inter-organizational networks, but below median on the other two indices. This variation, both across countries and within countries, and across types of networks, reveals that very different dynamics are unfolding with regard to the development of networks.

Graphs 2.1-2.3 present the scatter plots between the three sub-indices: international, inter-organization and intra-organization networks. The X and Y-axis present the median scores. The graphs help us to visualize the different scores of countries and between countries on the different network subindices. For example, on the top left of graph 2.2 one can observe that Bulgaria scores very highly in the international sub-index but below the median in the intraorganizational networks sub-index. Another example of the disparity between the sub-indices is the case of India (top of graph 2.3), whose score is very high on inter-organizational networks, but only median on intra-organizational networks.

2.6 The relationship between connectedness and government, industrial and economic performance

In order to analyse the relationship between connectedness and government effectiveness, regulatory quality, competitive industrial performance, and GDP per capita PPP a correlation matrix was constructed. The graphs clearly show a strong positive linear relationship between on the one hand connectedness and on the other hand different performance indicators.

Given the linear relationship between the variables (see graphs 2.4-2.7) the Pearson Product-Moment Correlation Coefficient was used to calculate the correlation between the different indicators (see annex 2).

The correlations are presented in table 2.6.

The analysis clearly shows the strong relationship between connectedness and government effectiveness, regulatory quality, industrial competitiveness and economic development. This is further supported by the high correlations which are all highly significant (see table 2.6). Both the overall connectedness index as well as the subindices on international networks, inter-organizational networks and intra-organizational networks are highly and significantly correlated to the performance indicators.

There are two interesting exceptions. First, regulatory quality, an indicator which is directly related to private sector development, is still highly and statistically significantly correlated with connectedness but the correlation is much less strong than in case of government effectiveness (also compare graphs 2.4 and 2.5). This is an interesting finding which needs to be analysed more in depth, especially since regulatory quality and government effectiveness are highly correlated. An in-depth comparison of countries which score very differently on regulatory quality and government effectiveness in its relationship with connectedness should be further pursued. Secondly, on the level of personal networks measured as membership in professional associations, the table indicates that this network measure is not significantly correlated to any of the performance measures. This can be the result of different methodological and substantial reasons which need to be further explored.

On the one hand, this high correlation is of course an interesting and relevant finding. No correlation would indicate that networks are ‘much ado about nothing’ and that we would not be able to find empirical evidence to support the increased attention for networks. This is clearly not the case. Networks do play an important role. However, the high

correlations also show that much more work is necessary to further understand the concept of networks and assess the impact of networks. The correlations are simply too high to draw many definite conclusions. Several methodological and substantial points are at stake.

First of all, we have to ask ourselves whether the results are spurious, i.e. whether there are any latent variables that drive connectedness and/or its subindices as well as the other variables. With regard to connectedness (especially intra-organizational networks as measured by training) and economic development, it might for example be the case that both are influenced by the development of human capital. Other theoretical reasons might probably be identified which could hypothesize why high correlations occur. Further theoretical development is necessary in this respect.

Secondly, the results might indicate that several of the indicators used are correlated proxies for the same phenomenon and that they are influenced by a same underlying dynamic. The latter can be explored a bit further by a closer inspection of the ranking of the connectedness index. The top 30 consists mostly of OECD Member States with the exception of Brazil, China, Cyprus, Malaysia, Singapore, South Africa and Thailand. Some of these exceptions score (very) highly on indicators such as the Human Development Index (UNDP) or economic development indicators. Hence, connectedness is very high in highly developed or rapidly developing countries. This indicates that networks are highly correlated to the development level. Whether they are a cause, consequence or both cannot be disentangled on the basis of the present analysis.

The latter is related to a third and obvious point that correlation is not causation since we do not know the direction of the cause; a third variable might be involved which is responsible for the covariance between X and Y. Hence the correlations and identified relationships should definitely not be considered causally relevant. International political networks for example can be a consequence of economic development as highly developed economies are more likely to have more embassies because they can afford it. The presence of such a large and highly educated diplomatic corps is also likely to affect the number of agreements a country can initiate, which is another element in the international political networks indicator. Similarly, the degree of university-industry interactions is affected by the presence of an elaborated tertiary educational tier, which in turn is partially a result of the development level of a country.

Although these arguments might reverse causality it should also be noted that the analysis on the subindex level shows that there are several cases where the level of economic development (as measured by CIP or GDP per capita) or policy effectiveness (as measured by government effectiveness and regulatory quality) is the same but the variation in networks very substantial (see graphs in annex 3), indicating that if a reverse causal argument would hold other factors contribute to network development. Taking it a step further, it might be the case that network dynamics emerge which further in time have an effect on the other variables. Much more theoretically informed empirical research is needed to figure out how networks causally play out in the dynamics of increased policy effectiveness, private sector development and economic development. In addition, we need more refined data and time-series to get grip on the issue of causality.

2.8 Conclusions

This chapter explored the possibility of constructing an index to capture the degree to which a country is networked on different levels.

The exploration was carried out on the basis of an inductive, data-searching approach. Many datasets and variables were screened. Very few contain data on networks. In addition, the data displays limitations:

 Insufficient time series are available for a better causal analysis.

 The connectedness index could only be calculated for 75 countries because data is lacking.

 The data only very partially captures the idea of networks, both in terms of their structures (the many potential networks which might arise out of the eco-system of private sector development) and of their nature (embedded versus arm-length networks).

 The remaining indicators which are included in the index and which are considered as a proxy for networks, such as the intra-organizational ones on training, also capture other aspects such as human capabilities development.

 So far, general indicators capturing network effects were considered. One good way forward to capture more precise networks and network effects, especially on the international level, would begin by making use of social network analysis tools and develop indicators on the basis of dyadic relations between countries. Zeev Maoz

(2010) in a very recent publication explored this further and makes convincing arguments for a better exploitation of network tools in the context of international relations and international political economy research.

Finally, the available data only allows for an indirect link to the nexus of networks and knowledge management. Data on knowledge networks is limited and more conceptualization is needed to guide empirical research in this area.

Notwithstanding the limitations of the data, especially from a theoretical and conceptual perspective, it was possible to create a connectedness index to further substantiate the relevance of examining networks. The results show that there is significant variation in networks across countries and also within countries across levels of networks. This is an interesting finding which triggers many questions on how to explain this variation. The variation correlates highly with other outcome variables such as government effectiveness, industrial development and economic development. As such this finding is highly interesting, but not definite causal arguments can be drawn from this link at this stage. Networks are probably cause and consequence and influence other parameters in causality loops. In general, concept development with the aim of developing indicators which capture the ‘network effect’ would best follow the process outlined in figure 2.1. More conceptual and empirical refinement is required. Given the rise of the importance of networks this might be further explored by bringing together experts on international relations, on economic clusters and inter-organizational networks, intra-organizational networks, international and national datasets and social network analysis in order to explore further existing datasets, identify opportunities to create more data and further conceptualize the concept of connectedness as a measurable indicator to capture the degree of network formation.

Оставите одговор

Ваша адреса е-поште неће бити објављена. Неопходна поља су означена *