Vous êtes sur la page 1sur 6

eGOVERNMENT UTILIZATION IN EUROPEAN LANDSCAPE:

HIEARCHICAL CLUSTER ANALYSIS


Mirjana Peji Bach
Faculty of Economics & Business Zagreb, University of Zagreb, Department of Informatics
Trg J.F. Kennedyja 6, HR-10000 Zagreb, Croatia
mpejic@efzg.hr

Jovana Zoroja
Faculty of Economics & Business Zagreb, University of Zagreb, Department of Informatics
Trg J.F. Kennedyja 6, HR-10000 Zagreb, Croatia
jzoroja@efzg.hr

Abstract: eGovernment applications have become widely used in recent two decades with the goal to
improve communication and cooperation between private and public entities. Goal of the paper is to
explore the recent trends of eGovernment utilization by the enterprises in European countries with the
usage of hierarchical cluster analysis. Cluster analysis revealed that European countries can be
partitioned into homogenous groups regarding eGovernment usage. However, results indicate that the
level of eGovernment utilization is not directly related to the economic development and
competitiveness of a country, although it is related to the perceived barriers towards eGovernment.
Keywords: eGovernment, Competitiveness, Cluster Analysis, Hierarchical Cluster Analysis, Global
Competitiveness Report, Barriers

1 INTRODUCTION

eGovernment provides better services for its citizens and enterprises using Internet and
information and communication technologies (ICTs) [7; 15]. eGovernment offers many
improvements, e.g. transformation of bureaucratic mechanisms and administration,
increasing participation, openness, transparency and communication with users [20]. Stier
(2015) indicate that eGovernment increases efficiency and transparency of government
operations, strengthening democracy and better services to citizens and business [20].
Development of eGovernment started in OECD countries during 1990s with the
increasing development and usage of Internet [5, 2]. Estonia is one of the leading countries
regarding ICTs usage in public administration, which is the result of the investments in this
area. In the last 10 years Estonia spend 1% of national budget on development and usage of
ICTs in public services [8].
Number of authors has indicated in their research that factors which influence the most
on the eGovernment usage are: economic and technical readiness of the countries, and
understanding of public authorities to citizens need and lower costs [12; 18; 11]. In order to
clarify the factors that could influence eGovernment usage, our work is focused on two
research goals. First, we aim to shade some light into the similarities and differences among
European countries according to their recent usage of eGovernment. Second, we investigate
if the European countries that are similar to the level of usage of eGovernment are, in the
same period, similar according to their level of economic development and competitiveness,
as well as according to the perceptions of their companies regarding the barriers in
eGovnernment. In the first stage, the cluster hierarchical analysis is used to organize
European countries into sensible groupings using data from Eurostat for 2013, according to
their usage of eGovernment [3]. In the second stage, defined clusters are compared by the
means of Anova analysis according to: (i) Global Competitiveness Indeks (GCI) and GDP
per capita, and (ii) perceptions regarding the barriers in eGovernment [17].
The paper is organized as follows. The first section provides an introduction. In the
second section data and the model specification are explained. Results of the cluster and the
Anova analysis are described in the third section. At the end, a discussion and conclusion
close the paper.

2 METHODOLOGY

2.1 Quantifying eGovernment utilisation

Data on eGovernment utilisation and data on perceived barriers towards the utilization of
eGovernment were collected for 32 European countries (EU countries, Iceland, Norway,
Macedonia and Turkey). Other European countries were not used in the analysis because of
missing data for 2013 for the selected variables. The percentage of enterprises in each
country that are using specific form of eGovernment and that encountered specific barrier is
outlined in the Table 1.

Table 1: Utilization and perceived barriers of eGovernment in selected European countries in 2013
Utilization of eGovernment Perceived barriers towards the utilization of eGovernment
Area/ country Complicated Requiring Not Any of
eForms- eForms- Confidentiality
eTendering eVAT eSocial and time paper of being the
obtaining returning and security
consuming person aware (d) reasons
EU-28 13 77 74 59 55 21 29 33 24 54
EU-27 13 77 74 59 55 21 29 32 24 54
EU-15 12 76 73 61 54 20 29 29 25 51
Belgium 12 72 74 65 39 26 31 24 27 56
Bulgaria 10 77 79 75 73 25 15 20 16 44
Czech
Republic 19 90 81 41 30 35 46 53 26 77
Germany 7 66 61 46 48 29 19 29 25 52
Estonia 25 81 80 73 73 8 13 33 5 40
Ireland 30 88 95 84 77 18 14 25 10 39
Greece 10 92 81 78 74 24 14 22 13 44
France 19 91 87 83 80 21 46 30 41 61
Croatia 19 84 81 71 57 33 36 58 15 75
Italy 9 73 58 27 27 27 53 51 40 79
Cyprus 16 82 47 9 12 16 21 39 14 60
Latvia 18 79 88 83 84 27 20 47 26 61
Lithuania 30 99 99 95 96 8 10 26 6 33
Luxembourg 10 84 65 46 35 19 28 36 33 61
Hungary 12 82 81 73 71 8 8 33 10 39
Malta 18 79 58 15 25 20 27 36 12 57
Netherlands 13 83 85 71 57 2 2 2 1 3
Austria 15 84 77 60 51 30 20 38 17 59
Poland 24 81 86 29 70 32 36 62 34 77
Portugal 19 81 85 78 79 25 33 37 25 61
Romania 15 57 52 48 49 19 23 31 13 56
Slovenia 1 86 81 79 78 13 18 29 33 49
Slovakia 22 86 71 54 59 42 57 74 16 89
Finland 6 92 89 75 75 8 19 29 25 46
Sweden 19 92 87 64 64 7 18 23 28 43
United
Kingdom 14 80 87 84 65 6 12 12 11 24
Norway 22 85 89 75 54 8 20 30 15 43
FRJ
Macedonia 29 64 56 50 35 16 15 37 8 53
Source: Authors research based on Eurostat data (2014)

2.2 Hierarchical cluster analysis and Anova analysis

In this paper we use following methodological approach. Firstly, we use hierarchical cluster
analysis in order to organize European countries into sensible groups for the year 2013,
regarding their usage of eGovernment. Secondly, we used Anova analysis to compare
identified clusters.
One of the best ways to analyse a large amount of data is to classify them into groups
which members are similar among each other and are different compared to members of
other groups [13]. The cluster analysis is one of a statistical classification method which is
used to examine hidden data structure and to group objects into homogeneous groups based
on their similar characteristics [9; 14]. There are four main steps of cluster analysis: feature
selection, algorithm selection, cluster validation and results. It is important to highlight that
cluster analysis is not a one-step process, but a process of several iterations [21].
Cluster analysis can be divided into two main groups according to the algorithm usage:
partitioned clustering and hierarchical clustering [4]. In this paper we used hierarchical
divisive clustering, in which data can be classified with a sequence of nested partitions,
which refers to the classification starting a cluster including all individuals towards the
smaller clusters and backwards, in other words using a top down approach [10]. After the
cluster analysis is conducted, the Anova analysis is often used with the goal to test
differences among clusters according to the selected differences [6].
In the first stage, data on the usage of eGovernment in 32 European countries was used
for cluster determination. Using the hierarchical cluster analysis, countries were grouped
regarding indicators of eGovernment usage by enterprises presented in Table 1 (left
columns). In this paper we used the Ward-method as the clustering method. In the second
stage, we used Anova analysis to compare identified clusters according to data presented in
Table 1 (right columns): (i) their competitiveness level and GDP per capita, and (ii) their
perceptions regarding the barriers in eGovernment.

3 RESULTS

3.1 Hierarchical cluster analysis results

According to eGovernment usage, selected countries are divided into clusters (Figure 1). The
results showed that 28 selected European countries are grouped into three clusters regarding
the rank of a particular cluster.

Figure 1: Selected European countries grouped into specific clusters based on different characteristics and
forms of eGovernment usage; 2013
Cluster A Bulgaria, Finland,
Greece, Hungary,
(N=7) Netherlands, Slovenia,
United Kingdom

Cluster B Austria, Croatia,


Czech Republic, Estonia,
(N=13) France,
Ireland, Latvia,
Lithuania, Norway,
Poland, Portugal,
Slovakia, Sweden

Cluster C Belgium, Cyprus,


FRY Macedonia,
(N=8) Germany, Italy,
Luxembourg,
Malta, Romania

Source: Authors research based on Eurostat data (2014)


In Cluster A there are seven countries: Bulgaria, Finland, Greece, Hungary,
Netherlands, Slovenia and United Kingdom. It can be noted that the Cluster A includes some
of the most developed European countries, e.g. Finland, as well as developing countries, e.g.
Bulgaria. Cluster B consists of 13 countries: Austria, Croatia, Czech Republic, Estonia,
France, Ireland, Latvia, Lithuania, Norway, Poland, Portugal, Slovakia and Sweden. This
cluster also consists of the most developed northern European countries, e.g. Norway, and
developing countries, e.g. Croatia. Cluster C consists of 8 countries: Belgium, Cyprus, FRY
Macedonia, Germany, Italy, Luxembourg, Malta and Romania. This cluster again consists on
countries that are different according to their economic development. It can be concluded that
countries within Cluster A, Cluster B or Cluster C are not similar regarding socio-economic
development neither regarding geographic position. In all three clusters developed and
developing European countries can be found. However, they are similar to their level of
eGovernment usage by the enterprises.
Table 2 represents descriptive statistics of the eGovernment usage indicators.

Table 2: Descriptive statistics of the eGovernment usage indicators across clusters (% of enterprises)
eForms- eForms-
Ward Method eTendering obtaining returning eVAT eSocial
9,429 83,143 83,286 76,429 70,429
Cluster A (n=7) (4,541) (4,776) (3,729) (4,315) (7,161)
Cluster B 21,615 86,231 85,077 68,462 67,231
(n=13) (4,556) (5,644) (7,342) (18,577) (17,210)
Cluster C 14,500 72,125 58,875 38,250 33,750
(n=8) (6,949) (9,403) (8,184) (19,241) (12,314)
Total 16,536 81,429 77,143 61,821 58,464
(n=28) (7,285) (8,925) (13,542) (22,241) (20,915)
Source: Authors research based on Eurostat data (2014)
Note: Means and standard deviations (in parentheses) are provided; bold letters indicate the highest average
value

In Cluster A there is the highest percentage of enterprises according to the usage of


eVAT (76,429% of enterprises) and eSocial (70,429% of enterprises). In Cluster B there is
the highest percentage of enterprises according to the usage of eTendering (21,615% of
enterprises), eForms obtaining (86,231% of enterprises) and eForms returning (85,077% of
enterprises). Countries in Cluster C are those with the lowest average values of all
eGovernment indicators.

3.2 Anova analysis results

In this section results of the Anova analysis are presented (Table 3).

Table 3: Descriptive statistics of Global Competitiveness Index, GDP per capita and perceptions regarding the
barriers in eGoverment and the ANOVA analysis; 2013
Cluster A Cluster B Cluster C F-value
Mean St.Dev Mean St.Dev Mean St.Dev
GCI 4,753 0,717 4,692 0,469 4,635 0,548 0,083
GDP per capita (000 EUR) 452,000 1131,062 129,408 197,004 61,513 75,701 1,009
Perceptions regarding the barriers in eGovernment
Confidentiality and security 12,286 8,958 22,615 11,948 21,500 5,099 2,741*
Complicated and time consuming 12,571 5,940 28,385 14,936 27,125 11,667 4,042**
Requiring paper of person 21,000 10,924 41,231 16,161 35,375 8,017 5,438
Not being aware about eGovernment services 15,571 10,518 20,308 10,773 21,500 11,452 0,624**
Any of the reasons 35,571 16,521 58,308 17,698 59,250 8,548 5,934***
Source: Authors research based on Eurostat data (2014) and Schwab (2014)
Note: *** statistically significant at 1%, ** statistically significant at 5%, * statistically significant at 10%
Comparison according to Global Competitiveness Index
In order to estimate the relationship of eGovernment usage and competitiveness of
selected European countries in 2013, calculated average values of GCI of each cluster are
presented in Table3. It can be seen that average values of GCI are the highest in Cluster A.
Cluster B (4,692) and Cluster C (4,635) have almost the same average values but less than
Cluster A. Anova analysis indicates that found differences are not statistically significant.

Comparison according to GDP per capita


In order to estimate the relationship of eGovernment usage and competitiveness of
selected European countries in 2013, calculated average values of GDP per capita of each
cluster are presented in Table 3. It can be seen that average values of GDP per capita are the
highest in Cluster A (452,000) where are grouped one of the most developed European
countries (Finland, Netherlands and United Kingdom). Cluster C (61,513) has the lowest
average value for GDP per capita where the less developed European countries can be found
(FRY Macedonia, Malta and Romania). Anova analysis indicates that found differences are
not statistically significant.

Comparison according to perceptions regarding the barriers in eGovernment


In order to estimate the relationship of eGovernment usage and competitiveness of
selected European countries in 2013, calculated average values for the barriers in
eGovernment usage of each cluster are presented in Table 3. It can be seen that the highest
average values are in the Cluster B for following indicators: Confidentiality and security,
Complicated and time consuming and Requiring paper of person. While, for the other two
barriers the average values are the highest in Cluster C: Not being aware about eGovernment
services and Any of the reasons). The lowest average values for all mentioned barriers in
eGovernment are in Cluster A where are grouped one of the most developed European
countries (Finland, Netherlands and United Kingdom). The conducted analysis showed that
calculated average values for barriers in eGovernment are statistically significant for all
indicators, except for the Requiring paper of person. Anova analysis revealed that differences
among clusters are statistically significant for the following barriers: (i) Confidentiality and
security is statistically significant at 10%, (ii) Complicated and time consuming and Not
being aware about eGovernment services are statistically significant at 5%, (iii) and indicator
Any of the reasons is statistically significant at 1%.

4 Conclusion

Results of our study indicated that European countries could be grouped into three clusters
using hierarchical cluster analysis according to their usage of eGovernment which was our
first goal of the paper. The countries grouped together in the clusters differ a lot to each other
according both to the level of their economic development as well as to their geographical
position. However, European countries grouped together had similar percentage of
enterprises that encountered barriers regarding the utilization of the eGovernment. The
conducted Anova analysis showed that calculated average values are not statistically
significant for GCI, GDP per capita and for one barrier in eGovernment: Requiring paper of
person. Nonetheless, Anova analysis revealed that differences among clusters are statistically
significant for other four barriers in eGovernment. Therefore, the conclusion for our second
goal could be that the level of eGovernment utilization in particular country is more related to
the perceived barriers then to the level of the economic development of the country.
In this paper we collect data for the year 2013 and for the European countries. In
order to expand this research, the analysis of data for the year 2015 should be conducted with
the goal to compare results from the year 2013 and 2015. Expanded analysis would also show
if there is any progress in eGoverment usage during the last two years, especially in the area
of eTendering. Also, it will be useful to compare other non-European countries according to
their level of eGovernment usage. In this paper, we have not corroborated former researchers
regarding impact of socio-economic development on eGovernment usage [12; 18; 11], but we
indicate that the perceived barriers have the stronger influence. Therefore, in further research
influence of national policies regarding eGovernment usage should be examined.

References

1. Candiello, A., Albarelli, A., Cortesi, A. (2012). Quality and Impact Monitoring for Local eGovernment
Services. Transforming Government: People, Process and Policy, 6(1): 112-125.
2. Elliman, T., Irani, Z., Jackson, P. (2007). Establishing a Framework for eGovernment Research: Project
VIEGO. Transforming Government: People, Process and Policy, 1(4): 364-376.
3. European Commission Database - Eurostat (2014). http://ec.europa.eu/eurostat/web/information-
society/data/main-tables [Accessed 22/05/2015]
4. Everitt, B., Landau, S., Leese, M. (2001). Cluster Analysis, 4th Edition. London: Arnold.
5. Guijarro, L. (2007). Interoperability Frameworks and Enterprise Architectures in E-Government
Initiatives in Europe and the United States. Government Information Quarterly,24(1): 89-101.
6. Hair, J.F., Black B., Babin, B., Anderson, R.E., Tatham, R.L. (2005). Multivariate Data Analysis, 6th
Edition. New Jersey: Prentice Hall.
7. Holliday, I., Yep, R. (2005). E-Government in China. Public Administration and Development, 25(3):
239249.
8. International Telecommunication Union (2012). Measuring the Information Society.
http://www.itu.int/en/ITU-D/Statistics/Documents/publications/mis2012/MIS2012_without_Annex_4.pdf
[Accessed 25/05/2015]
9. Jain, A.K. (2010). Data clustering: 50 years beyond K-means. Pattern Recognition Letters, 31(8): 651-
666.
10. Jain, A., Dubes, R. (1988). Algorithms for Clustering Data. Englewood Cliffs, NJ: Prentice-Hall.
11. Janssen, M., Kuk, G., Wagenaar, R.W. (2008). A Survey of Web-based Business Models for
eGovernement in The Netherlands. Government Information Quarterly, 25(2): 202-220.
12. Khalil, O.E.M. (2011). E-Government Readiness: Does National Culture Matter? Government
Information Quarterly, 28(3): 388-399.
13. Parr Rud, O. (2000). Data Mining Cookbook: Modeling Data for Marketing, Risk and Customer
Relationship Management. New York: John Wiley & Sons Ltd.
14. Rodriguez-Deniz, H., Voltes-Dorta, A. (2014). A Frontier-based Hierarchical Clustering for Airport
Efficiency Benchmarking. Benchmarking: An International Journal, 21(4): 486-508.
15. Rorissa, A., Demissie, D., Pardo, T. (2011). Benchmarking eGovernment: A Comparison of
Frameworks for Computing eGovernment Indeks and Ranking. Government Information Quarterly,
28(3): 354-362.
16. Sarrayrih, M. A., Sriram, B. (2015). Major Challenges in Developing a Successful eGovernment: A
Review on the Sultanate of Oman. Journal of King Saud University-Computer and Information
Sciences, 27: 230-235.
17. Schwab, K. (2013). The Global Competitiveness Report. Geneva: World Economic Forum.
http://www3.weforum.org/docs/WEF_GlobalCompetitivenessReport_2012-13.pdf[Accessed
22/05/2015]
18. Stier, S. (2015). Political determinants of e-government performance revisited: Comparing democracies
and autocracies. Government Information Quarterly, ISSN 0740-624X,
http://dx.doi.org/10.1016/j.giq.2015.05.004. [Accessed 10/06/2015]
19. Verdegem, P., Verleye, G. (2009). User-Centred E-Government in Practice: A Comprehensive Model
for Measuring User Satisfaction. Government Information Quarterly, 26(3): 487-497.
20. Wong, K., Fearon, C., Philip, G. (2007). Understanding eGovernment and eGovernnance:
Stakeholders, Partnerships and CSR. International Journal of Quality and Reliability Management,
24(9): 927-943.
21. Xu, R., Wunsch, D.C. II (2008). Recent Advances in Cluster Analysis. International Journal of
Intelligent Computing and Cybernetics, 1(4): 484-508.

Vous aimerez peut-être aussi