Measuring Networks across Countries: an Empirical Exploration

Axel Marx and Jadir Soares

http://www.unido.org/fileadmin/user_media/News/2011/UN%20Report%20LowRes.pdf

“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

2.1 INTRODUCTION

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
	Republic
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  Guatemala	0.8120.538  0.493	0.2070.374  0.388	0.2230.324  0.348	0.414 0.412  0.410	4950  51
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
EGYAZE  MKD	EgyptAzerbaijan  Macedonia	0.5630.485  0.445	0.2970.294  0.201	0.2080.265  0.354	0.3560.348  0.333	5960  61
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.