Local e-Government Bench-learning

152
Treball de recerca Doctorat “Informació i Documentació en la Societat del Coneixement” Facultat de Biblioteconomia i Documentació Universitat de Barcelona Local e-Government Bench-learning Towards a new methodological framework to benchmark electronic services provision and adoption in local public administration Autor: Joan Batlle i Montserrat Tutor: Prof. Dr. Cristobal Urbano Data Lliurament: 30 de Juny de 2010

Transcript of Local e-Government Bench-learning

Page 1: Local e-Government Bench-learning

Treball de recercaDoctorat “Informació i Documentació en la Societat del Coneixement”Facultat de Biblioteconomia i DocumentacióUniversitat de Barcelona

Local e-Government Bench-learning Towards a new methodological framework to

benchmark electronic services provision and adoption in local public administration

Autor: Joan Batlle i MontserratTutor: Prof. Dr. Cristobal UrbanoData Lliurament: 30 de Juny de 2010

Page 2: Local e-Government Bench-learning

Local e-Government Bench-learning

2

Page 3: Local e-Government Bench-learning

Local e-Government Bench-learning

Table of ContentsAbstract..........................................................................................................................5Acknowledgements .......................................................................................................7 1 Introduction................................................................................................................9 2 Background: e-government, its benchmarking and evolution models.......................13

2.1 What does e-government mean?........................................................................13 2.1.1 e-Government definition............................................................................13 2.1.2 e-Government components........................................................................15 2.1.3 Local e-government...................................................................................16

2.2 Benchmarking e-government............................................................................17 2.2.1 Evaluating, measuring, benchmarking.......................................................17 2.2.2 Different proposals of e-government benchmarking..................................18 2.2.3 Local e-government benchmarking............................................................21

2.3 e-Government models.......................................................................................21 2.3.1 The early models.......................................................................................22 2.3.2 The extended models.................................................................................25 3 The Local e-Government Bench-learning survey......................................................27

3.1 Overview..........................................................................................................27 3.2 Methodology.....................................................................................................29 3.3 The sample of cities and its analysis..................................................................30 3.4 Services maturity levels....................................................................................32 3.5 Service adoption levels.....................................................................................32 3.6 The global services catalogue............................................................................33 3.7 European standard coverage and diversity.........................................................35 3.8 European maps and city maps...........................................................................35 3.9 Best practices....................................................................................................37 3.10 Results discussion...........................................................................................37

4 E-services benchmarking and adoption literature revisited.......................................39 4.1 Benchmarking electronic services provision.....................................................39 4.1.1 Online Availability of Public Services benchmarking, the European de factostandard....................................................................................................40 4.1.2 MeGAP: a US local public service benchmarking.....................................42 4.1.3 E-Government at the American Grass-roots...............................................44 4.1.4 Evolution of Local Government E-services in Australia............................44 4.1.5 Digital Governance in Municipalities Worldwide: a biased perspective.....45 4.1.6 Summarizing the evolution of electronic service benchmarking................46 4.2 Measuring electronic services adoption.............................................................47 4.2.1 Adoption is not demand ............................................................................47 4.2.2 The importance of measuring adoption......................................................47 4.2.3 Measuring adoption...................................................................................48 4.2.4 e-Government services target....................................................................49

3

Page 4: Local e-Government Bench-learning

Local e-Government Bench-learning

4.2.5 How often citizens consume e-government services?................................49 4.2.6 Variables to measure adoption...................................................................50 4.2.7 Measurement methodologies.....................................................................51 4.2.8 Summarizing adoption...............................................................................52 4.3 Summary of the electronic service benchmarking.............................................52

5 Local e-Government Bench-learning improvements.................................................53 5.1 The Local e-Government Bench-learning most interesting innovations.............53 5.2 Some important limitations revealed by the study.............................................55 5.3 Improvements...................................................................................................55

5.3.1 Hypothesis formulation ............................................................................55 5.3.2 Statistical analysis.....................................................................................56 5.3.3 The sample................................................................................................57 5.3.4 The services catalogue...............................................................................57 5.3.5 Collection of the information.....................................................................58 5.3.6 Service maturity levels..............................................................................58 5.3.7 Measurement of adoption..........................................................................59 5.3.8 Scoring......................................................................................................59 5.3.9 Charts........................................................................................................60 6 Final remarks............................................................................................................61

7 References................................................................................................................63 8 Bibliography............................................................................................................69 8.1 Journal Articles.................................................................................................69 8.2 Reports.............................................................................................................73 8.3 Books................................................................................................................75 8.4 Conference proceedings....................................................................................76

9 Annex: The Local e-Government Bench-learning survey..........................................79

Index of TablesTable 1: Benchmark surveys from Kunstelj and Vintar (2004)......................................19Table 2: Cities characteristics.......................................................................................31Table 3: Cap Gemini basic public services....................................................................41Table 4: New Jersey results (Holzer, 2003)...................................................................45Illustration IndexIllustration 1: Baum and Di Maio four stages model.....................................................22Illustration 2: Layne and Lee stages model...................................................................23Illustration 3: European map of maturity and adoption.................................................36

4

Page 5: Local e-Government Bench-learning

Local e-Government Bench-learning

AbstractA new methodology to benchmark electronic service provision and adoption in localpublic administration is introduced in the framework of e-government research andcompared with existing ones. This methodology was tested in a first benchmarkingsurvey, led by the research report author, carried out in 2008 with a sample of 15European cities, whose results are also discussed. The main innovations of thismethodology are the addition of adoption, beyond provision; an extension of the Baumand Di Maio stages model of e-government with a “web 2.0” stage; a bottom-upservices categorisation more suitable to local administrations; and a results presentationbased on charts where cities are placed with respect to averages. The survey focus onlocal administration is novel, and thus the results themselves. The need of moreextensive surveys (of local e-government especially) is revealed, as they will providemore reliable data and allow deeper and more refined analyses. Measuring adoptionremains a research challenge; the hypothesis formulation, data analysis andvisualisation of the results should be improved for these more extensive surveys, andnew scores should be introduced to facilitate temporal comparisons in the future. En el marc de la recerca en e-government, es presenta una nova metodologia perrealitzar estudis comparatius de provisió i adopció de serveis electrònics al'administració pública local, i es compara amb les existents. Aquesta metodologia vaser provada en un primer estudi, liderat per l'autor d'aquest treball de recerca, que esdugué a terme l'any 2008 a 15 ciutats europees, i els resultats del qual es presententambé. La innovació més important d'aquesta metodologia és la inclusió de l'adopciódels serveis a part de la provisió; una extensió del models de nivells de Baum and DiMaio incloent un nivell per “web 2.0”; una categorització de serveis construïda de baixa dalt més adequada per l'estudi de l'administració local i la presentació dels resultatsbasada en grafics en els que les ciutats es poden comparar amb la mitjana. L'enfocespecific en administració local és també nou, i per tant, també ho són els resultats en simateixos. Es posa de manifest la necessitat de realitzar més estudis, especialment enl'àmbit de l'administració local, ja que aquests aportarien dades significatives quepermetrien anàlisis més en detall. Mesurar l'adopció és encara un repte; la formulació deles hipòtesis, l'anàlisi de les dades i la visualització dels resultats haurien de millorar enaquests properes edicions de l'estudi. Convindria també introduir un nou sistema basaten un indicador únic a fi de facilitar comparatives al llarg del temps. Se presenta en el marco de la investigación en e-government una nueva metodologíapara realizar estudios comparativos de provisión y adopción de servicios electrónicos enla administración pública local, y se compara con las existentes. Esta metodología fueprobada en un primer estudio, liderado por el autor de este trabajo de investigación,realizado el año 2008 en 15 ciudades europeas, y cuyos resultados se presentantambién. La innovación más importante de esta metodología es la inclusión de laadopción de los servicios a parte de la provisión, una extensión de los modelos deniveles de Baum and Di Maio incluyendo un nivel por "web 2.0", una categorización de

5

Page 6: Local e-Government Bench-learning

Local e-Government Bench-learning

servicios construida de abajo a arriba más adecuada para el estudio de la administraciónlocal y la presentación de los resultados basada en graficos en los que las ciudades sepueden comparar con la media. El enfoque específico en administración local estambién novedoso, por tanto, también lo son los resultados en sí mismos. Se pone demanifiesto la necesidad de realizar más estudios especialmente en el ámbito de laadministración local, ya que estos aportarían datos significativos que permitiríananálisis más en detalle. Medir la adopción es todavía un reto, la formulación de lashipótesis, el análisis de los datos y la visualización de los resultados deberían mejorar enestos próximas ediciones del estudio. Convendría también introducir un nuevo sistemabasado en un indicador único para facilitar comparativas a lo largo del tiempo.

6

Page 7: Local e-Government Bench-learning

Local e-Government Bench-learning

Acknowledgements This research report has been possible thanks to the patience and wise direction of Prof.Dr. Cristobal Urbano. Master lectures of Prof. Dr. Maite Bárrios (introducing statistical analysis) and Prof. Dr.Àngel Borrego (on methodology) have been source of inspiration to review the workdone in performing the first European Local e-Government Bench-learning carried outin 2008. Moreover, I would like to thank the smart advice of and the hard work done by Prof. Dr.Josep Blat, a friend of mine, who has been proofreading, reviewing the English textonce and again, and correcting and criticizing the discourse in order to make itscientifically rigorous and understandable. I would like to express my deepest appreciation and heart felt thanks to the colleaguesthat have been working with me in the first edition of the Local e-Government Bench-learning survey: Júlia López, Raquel Calderón and Prof. Dr. Mila Gascó. I would further like to thank all the cities representatives that collaborated in thebottom-up process to design the methodology and provided the data needed to performthe first benchmarking survey.Last but not least, I thank EUROCITIES for its full support and the publication of theresults obtained in the first edition; and especially the Barcelona City Council becausewithout its sponsorship all this work would not have been possible.

7

Page 8: Local e-Government Bench-learning

Local e-Government Bench-learning

8

Page 9: Local e-Government Bench-learning

Local e-Government Bench-learning

1 IntroductionThe research discussed in this report deals with benchmarking e-government servicesprovided by local public administration. Although local public administrations are themost active in public service provision to citizens and businesses, our literature reviewshows that there are no studies, and no benchmarks confronting supply and demand,that is, services supply and citizens' adoption at the level of local public administration. An innovative study, reported in the Local e-Government Bench-learning (Batlle,Calderón & López, 2009), based on an innovative methodology was launched in 2008to cover this lack of benchmarks. This study, where the author of this research reportplayed a directive role, is deeply analysed in this memoir, both in terms of themethodology adopted and the results obtained, and grounded in a literature review ofthe field. The results are interesting in themselves, but the pilot character of the study limits theirsignificance. On another hand, our analysis provides us with improvements that can beincluded before launching the second edition of the bench-learning study. Let us frame first local public administration in the current e-government trends. Generally, e-government is described as the use of information and communicationtechnologies (ICT) in government processes, and the high potential of ICT to transformgovernmental organizations into more efficient, accessible and accountable ones iswidely recognized. According to EUROSTAT, the public sector is responsible for almost40% of the European GDP. The combination of both statements points out the capitalimportance of the e-government transformation of the public sector for the Europeaneconomic growth. Surveys performed in Europe show that between 50% and 80% of the citizensinteraction with public bodies takes place at sub-national level (Heeks, 2006), i.e., withlocal and regional public administrations. While e-government takes place on all publicorganizations at state, regional and local level, such as autonomous communities,counties, regions..., it is in cities where it takes special relevance, because cities arecloser to citizens and deliver the largest number of services directly to people (Moon &Norris, 2005). Nowadays, all city governments and administrations find themselves in a transitionfrom the old model of local public organization to a new one, claimed to be moreefficient, as e-government improves efficiency, accuracy, timeliness, and effectiveness,among other things (Moon & Norris, 2005). Electronic government is no longer just anoption but a necessity for countries aiming at better governance (Gupta & Jana, 2003). The relationship of cities with their citizens is also changing from a scenario based on amultiplicity of specialized counters to a “one-stop-shop” model attended by multivalentcivil servants supported by ICT. At the same time, cities are evolving from situations in

9

Page 10: Local e-Government Bench-learning

Local e-Government Bench-learning

which citizens are asked time and again to submit this and that, to offer a moreintelligent organization capable to know who the citizen is and which the citizen's needsare in each case. In other words, public organizations are reforming themselves to beproactive and to efficiently provide services to citizens. This transition is creating thecurrent framework to understand the important role of e-government in themodernization of the public administration, and, most especially, the local one.Let us turn now towards the focus of our research: methodologies to benchmark locale-government.The e-government transformation touches the core business of the local publicadministrations. e-Government is, at the same time, a process and involves multiplestages or phases of development (Al-Hashmi & Darem, 2008). Therefore, cities mustcarefully manage the process to shorten it and ensure a successful outcome. Citymanagers can and should use the experience of previous innovators as a guide to makeinformed decisions (Kaylor, Deshazo & van Eck, 2001). City managers need tools to measure, evaluate and compare this process. They mustunderstand the current status of e-government development, work out the extent towhich their objectives have been reached, validate the effectiveness of strategies andaction plans, ascertain strengths and weaknesses or shape new guidelines (Kunstelj andVintar, 2004).Private sector can provide tools to better manage this transformation. Businesslikemeasurement models have been introduced in public sector management in order toaccount for public sector performance (Noordegraaf, 2003). Benchmarking is an usualbusiness management practice used mainly for marketing and sales purposes. In the e-government field, benchmarking also means reviewing comparatively the performanceof (e-government) organisations (Heeks, 2006). This comparative exercise contributesto a broader view of the transformation process in which public organizations areimmersed, allows to identify leaders and followers, to understand different stages ofgrowth and points out best practices (Heeks, 2006). In European context, benchmarking e-government is a key tool to measure the progressmade in achieving the objectives established by the European Council. The EuropeanCommission stressed the importance of benchmarking in the eEurope 2002 Action Plan(EC, 2000) because benchmarking “aims to ensure that actions are carried outefficiently, have the intended impact and achieve the required high profile in allMember States”. And more specifically highlights the importance of benchmarking e-government in the eEurope 2005 Action Plan (EC, 2002-1) as well as in the i2010 e-Government Action Plan: Accelerating e-Government in Europe for the Benefit of All”(EC, 2006). However, there is very little information at the local level. Up to now most of theavailable information, data, measures and comparisons of e-government evolution referto state level administration, the best example being the benchmark of the onlineavailability of public services commissioned by the European Commission and yearly

10

Page 11: Local e-Government Bench-learning

Local e-Government Bench-learning

performed by Cap Gemini from 2001 up to now (Cap Gemini, 2002, 2003, 2004, 2005,2006, 2007, 2009). Some information is also available at regional level but only fewstudies deal with the local level. Moreover, indicators and metrics defined at the statelevel (EC, 2002-2), are not applicable at the local level, as we will discuss latter on. As aconsequence, is not possible to perform sound comparisons between different city casesacross different countries. The few surveys aiming at benchmarking local e-government are focused on generalaspects of the official website of the local administration, in terms of content,accessibility, quality and usability. Moreover, they focus on the services offering anddeal neither with e-services quality nor with citizens’ adoption. A complete view of the e-government process in cities is not possible today. Researchpapers suggest also that current approaches to monitor evaluation and benchmarking e-government development do not support a comprehensive e-government assessment andneed to be improved in order to give policy makers elements for their decisions.Therefore, additional research is needed in the field of benchmarking online publicservices (Kunstelj and Vintar, 2004).To fill the gap in local e-government benchmarking the author of this research reportdesigned and launched a pioneering field study of several key European cities (entitledLocal e-Government Bench-learning) in 2008. This study tried to introduce severalinnovations in the services catalogue including a wider coverage of services adapted tolocal administrations, with a focus based on a new maturity stages model, on the qualityand services adoption, and with a classification identifying standard services as well asdiversity. On the other hand, the study adopted a novel city charts visualisation to offerdecision makers material suitable to inform future decisions on e-governmentdevelopment strategies. In this research work, I will review in depth the applicability and correctness of theLocal e-Government Bench-learning study comparing it with the state of the art in e-government benchmarking - at different administration levels -, discussing its mainfindings and contributions, especially the methodological ones, at the localadministration level, and proposing further improvements. After this introduction, the second chapter starts by a literature review about e-government definitions underlining how wide the term is, and the need to focus on somespecific issues, such as those raised above, and how these definitions are applicable tothe local public administration will be discussed.In a second section, I present the different existing e-government benchmarks, howsome of them focus on citizen's e-readiness, while others pay attention to internal ITadoption of process re-engineering. And it could be seen that the majority of existingbenchmarks are devoted to compare the electronic service provision while nobenchmarks can be found in electronic services adoption. The following section of the chapter is devoted to the analysis of different e-government

11

Page 12: Local e-Government Bench-learning

Local e-Government Bench-learning

maturity models. Most of these models were designed ten years ago (2000 and 2001) toexplain the general evolution of e-government and we discuss whether they still can beapplied to electronic services provision and which are the limitations almost 10 yearslatter. The third chapter presents the Local e-Government Bench-learning and explains itdeeply. How it was conceived and why it is named bench-learning. How cities wereselected and which are the main characteristics of the sample. The bottom-upmethodology used is described and justified. The main outputs, the European chart forservice provision and adoption, the city charts and the list of best practices, arepresented as well as the partial results and the different variables used on the process, asthe global services catalogue and the coverage charts for each category. The charts arediscussed along with the main findings. Chapter four revisits more thoroughly e-government benchmarking through a literaturereview, identifying and discussing the main benchmarks that have been proposed ononline services provision. From this analysis we see the need of refining themethodology reviewing the clusters of services, adding aggregate scores, reviewing thematurity models used to still adapt it more at the local level specificities. Anothersection highlights how scarce the academic material in the area of online public servicesadoption is. And not only shows the need of more research but also justifies taking amore pragmatical approach to service's adoption assessment. I devote the fifth chapter to analyse in depth the Local e-Government Bench-learningidentifying strengths and weaknesses. This chapter presents in fact the conclusions ofthe research work. Local e-Government Bench-learning survey covers an important lackof information in local public administration without forcing cities to compete. Theinformation provided allows city managers to balance the e-government plans accordingto general trends and facing local needs. But this study present some weaknesses as thesmall sample used and the self-assessment methodology used to measure maturity andadoption. Furthermore another weakness is a crude use of statistical tools. From thisanalysis I list and qualify recommendations to improve the quality and applicability ofthe methodology proposed in the study that should be taken into consideration whenpreparing a second edition of the Local e-Government Bench-learning survey. Finally, the sixth chapter presents the final remarks of the research report. The mainfinding is the singularity of the benchmark introduced in our study. There are no similarstudies in the European local e-government area, and only a similar survey performed inUS by Kaylor et al. in 2001 tailored to US local administrations can be found. Themethodology proved to be correct and it requires only some refinements, but the resultspresentation could be improved and the sample must be widened. However, furtherresearch, standards and new methods are clearly needed in the field of electronic serviceadoption measurement.

12

Page 13: Local e-Government Bench-learning

Local e-Government Bench-learning

2 Background: e-government, its benchmarking and evolutionmodelsThe aim of this chapter is to present the state-of-the-art of different aspects of e-government research. We begin by an analysis of the different definitions of e-government based on a review of the related literature is presented, discussing thedifferent meanings of the term e-government, and seeing that the e-governmentdefinition must be extended when focusing at local public administrations. In a second section, we discuss the main working lines and trends of the existingliterature about benchmarking e-government. This analysis shows how benchmarkingelectronic services provision is one of the most frequent benchmarks while publicadoption of those services is a missing theme of the current benchmarks. Finally, a third section is devoted to motivate the need of e-government models, and todescribe and analyse the different existing models, highlighting their limitations andconcluding with the need to update the current models to explain the latest trends in e-government service provision.

2.1 What does e-government mean?There has been a succession of terms used as synonyms for the same concept: electronicgovernment, government IT, e-governance, online government, one-stop government,digital government and, more recently, e-government, the latter being the most used inthis work. e-Government is a broad concept and at the same time divergent. In thissection we discuss several definitions that exist in the literature, the main sub-areas orcomponents and how they translate into the local public administration.

2.1.1 e-Government definitionKaylor et al. (2001) define e-government as the ability to communicate and/or interactwith government agencies via the Internet in any way more sophisticated than a simpleemail letter. Other authors define it as the electronic provision of information andservices by governments 24 hours per day, seven days per week (Moon & Norris, 2005).Both are narrow definitions of e-government because they focus only on citizensinteraction through Internet, although it is probably the most popular definition of e-government (Bannister, 2007).Usually, narrow definitions of e-government have been formulated in the context of aspecific study, as, for instance, Layne and Lee (2001) definition that puts the focus on aspecific technology, Internet and the World-Wide-Web, and its capability to enhance theaccess to and the delivery of, government information and services to citizens, businesspartners, employees, other agencies, and government entities, when they propose a

13

Page 14: Local e-Government Bench-learning

Local e-Government Bench-learning

framework to assess government transformation when going on-line. Similarly Kayloret al. (2001) definition above quickly followed by an interesting survey of online publicservices among US municipalities, and Moon and Norris (2005) by an analysis of twosurveys about local government websites in the US. Other authors, focus on the efficiency rather than the technology but narrowly again.Carter and Bélanger (2005) define e-government as the use of information technology toenable and improve the efficiency with which government services are provided tocitizens, employees, businesses and agencies. According to the World Bank (Lanvin andLewin, 2006) e-government refers to the use by government agencies of informationtechnologies (such as Wide Area Networks, the Internet, and mobile computing) that areable to transform relations with citizens, businesses, and other branches of government.Therefore, although they open to ICT in general and are not restricting themselves toInternet, they have a specific narrow focus. Authors use these definitions with a twofold purpose: to define the term and to focus theattention in a specific part of the public administration functions, and thus should beconsidered as incomplete definitions of the term. Those narrow definition that areinstrumented to justify some specific studies can only be accepted in the context of suchstudies.In fact, e-government is a much more substantial transformation than e-service delivery(Löfstedt, 2005), it is lot more than using Internet or providing services throughinternet, or even improving the services, is much more than gathering the information,downloading files or making online transactions (Sakowicz, 2003). e-Government refers to the use of ICT in public administration. Such a broad definition of the term allows the researchers to see e-government in all ofits extension, impacts and benefits. To be more specific, some authors (Janssen, 2003)provide a detailed definition of the term listing different components or applicabilityareas related with different public administration functions such as internal use, servicesprovision or citizens participation, to mention some of them. Other authors presents e-government through the potential benefits or impacts: “e-government utilizestechnology to accomplish reform by fostering transparency, eliminating distance andother divides, and empowering people to participate in the political processes that affecttheir lives” (Al-Hashmi and Darem, 2008). Both approaches are useful to emphasizethat e-government is not simply a matter of giving government officials computers orturn the current processes into digital ones, but the new aspects the emerge with theseprocesses. Although all these broad definitions of the term are valid, the most used is the oneprovided by OCDE (2003) that summarizes the broad scope by defining e-governmentas the use of ICT, and particularly the Internet, as a tool to achieve better governance.Or, in its long formulation, “e-government applies to the use of information andcommunications technology to public administration tasks and processes, and aiming atreshaping government to citizen relationship, government to government, public

14

Page 15: Local e-Government Bench-learning

Local e-Government Bench-learning

services provision, internal efficiency and citizens participation”. This broad definition of e-government is the one we share in this research report as itswide view gives room for different sub-disciplines.

2.1.2 e-Government componentsAs we see, the OCDE definition of e-government is not restricted to specifictechnologies or tasks. As technology evolves and is applied to more fields e-governmentwill extend inside other public administration functions not yet involved in e-government. Let us mention, as an example, the city infrastructure management whichis currently experiencing a deep transformation thanks to the new ICT solutions basedon sensor networks and command centres. This should be also considered as part of e-government. As e-government is embracing all the functions that a public administration is doing,different aspects of ICT have to be considered, and several themes or components areincluded in e-government research. One of these themes, for instance, is ICT internaladoption, which refers to the use of ICT in public administration to change structuresand processes of government organisations (Lambrinoudakis, Gritzalis, Dridi, & Pernul,2003) aiming at improving internal efficiency. Other themes are the studies aboutinternal business process re-engineering, IT management, IT project assessment,interoperability or the relation with other public administration levels, e-governmentevaluation, etc. Janssen (2003) takes a more systematic approach identifying five maine-government components: provision of ICT infrastructure and ICT skills to facilitatecitizen's inclusion named e-society, automation of internal processes, providing citizenswith better services (e-administration), including citizens in policy (e-participation), andchanging the relation between politics and administration.Both Janssen and Lambrinoudakis et al. identify a lot of common components:automating internal processes, providing citizens with better services, etc. But let mehighlight that in both cases there is a component directly related with the public servicesprovision, which means the interaction between citizens and public administration andcompanies and public administration it is what Janssen (2003) calls e-administration. e-Administration aims to provide citizens with effective and efficient service delivery.This component includes not only the services' provision but also other related aspect asthe online services take-up or adoption, the multichannel service provision, the servicequality or user satisfaction. Latter on in this research report we will focus our attention on e-administration, but nowwe turn to local e-government, discussing the importance of local e-government, itsspecific components and the applicability of the previous concepts to localadministrations.

15

Page 16: Local e-Government Bench-learning

Local e-Government Bench-learning

2.1.3 Local e-governmentWhile e-government takes place on all public organizations at state, regional and locallevel, such as autonomous communities, counties, regions, ..., it is on cities where ittakes special relevance because cities are closest to citizens and deliver the largestnumber of services directly to them (Moon & Norris, 2005). Local government is the front-end of government services. According to Heeks (2006)local governments are the main point of contact for delivery of services in developedcountries, who also states that surveys performed on Europe show that services tocitizens reach the highest importance in local public administrations since between 50%to 80% of the citizens interaction with public bodies take place at sub national level.The same happens in US where local governments are the governments closest to thepeople. According to Norris (2005), local governments are key player because what theydo directly impacts citizens, far more and more immediately than the actions of state orfederal governments, because local governments deliver the vast majority of servicesthat directly touch the lives of citizens. And this proximity to people makes it alsoimportant in e-participation and e-democracy (Sakowicz, 2003).The importance of local e-government is increasing day by day. According to Lanvinand Lewin (2006) specific e-government services are increasingly handled at the localrather than the national level. This is the case, for instance, for small and medium-sizedenterprise (SME) registration, vehicle and drivers’ licenses, enrolment at educationalinstitutions and vocational programs, furthering human resources skills, or professionalauthorizations and licenses (for example, for shops, pharmacies, and so on). A similartrend is detected in Australia where municipalities had to re-frame their processes toaccommodate the expectations of their communities (Shackleton, Fisher, & Dawson,2004). This importance is being reinforced by the increasing role played by cities as,according to the World Bank (Lanvin and Lewin, 2006), in the last 50 years theproportion of population living in cities had increased to one-half, and it is estimatedthat, by 2050, six billion people (that is, two-thirds of the world population) will live incities. In such a context, maintaining adequate levels of production and delivery of keypublic services is a major challenge for cities.e-Government at local level has its own specificities. All the definitions and componentspresented in the section before apply to local e-government, but in addition, cities andmunicipalities are developing specific functions that cannot be found at other levels.City services provision and city management (traffic management, public transport,environment, pollution, cleaning, garbage collection, public lighting, culture, education,social care, crime contention, floods control and disaster management...) are traditionalfunctions of the local public administrations that are increasing considerably the use ofICT, and are not included in other public agencies. In addition, in some countries, citieshave to deal with basic service provision as water supply, gas and electricity. All ofthese are specific functions that can not be found in other levels of publicadministration. Therefore they are giving a special flavour to local e-government byadding specific components.

16

Page 17: Local e-Government Bench-learning

Local e-Government Bench-learning

These two characteristics make local e-government an interesting research subject.However, despite the rising importance and own specificities of local e-government, itlacks information and research related. According to Wohlers (2007), “an increasingbody of research examines the breadth of e-government at the international and nationallevels, while a systematic analysis of e-government at the local level and acrossdifferent population sizes remains scant”. In the following section we will have theopportunity to see more this unbalanced situation.

2.2 Benchmarking e-governmentNow we turn towards the specific aspect of benchmarking e-government. Through anoverview of the literature analysing current benchmarks I will show that the majority ofthe existing benchmarks focus on service delivery and how little benchmarks has beendone combining provision and adoption of electronic services. In addition here again thelocal dimension is missing since the vast majority of benchmarks focus at internationalor national level.

2.2.1 Evaluating, measuring, benchmarkingNowadays, e-government “is a major paradigm shift in the way that government andpublic administration have to function” (Peristeras, Tsekos, & Tarabanis, 2002). In thisprocess e-government in all its extension is being deployed to transform publicorganizations into more efficient and effective organizations, delivering higher qualityservices, and increasing transparency, citizens' participation and accountability (EC,2007). And it is a process that carries out enormous investment by governments(Alshawy, Alahmary, & Alalwany, 2007). This transformation that started some years is no longer just an option but a necessityfor countries aiming for better governance (Gupta & Jana, 2003). As it is anevolutionary phenomenon, “e-government initiatives should be accordingly derived andimplemented” (Layne & Lee, 2001). Policy makers and city managers must carefullymanage the process to shorten it and ensure a successful end. They can and should usethe experience of previous innovators as a guide to make informed decisions (Kaylor etal., 2001), to answer some of the questions they have:

• How my is city doing in e-government?• Are the current e-government strategy/investments showing expected results?• Is the e-government strategy well-balanced?• Which are the most advanced areas of e-government service provisions?• Which are the best practices in Europe?• Where and how should my city improve the e-government agenda?

The answers to these questions can only come from evaluating, measuring and

17

Page 18: Local e-Government Bench-learning

Local e-Government Bench-learning

comparing e-government. According to Kunstelj and Vintar (2004), the evaluation andcomparison are the key tools to discover the current state of e-government development,work out the extent to which objectives have been reached, validate the effectiveness ofstrategies and action plans, ascertain strengths and weaknesses, shape new guidelines orlook for examples of best practice. And that is why the European Commission is alsopromoting e-government evaluation, measurement, benchmarking and case-basedimpact and benefit analysis based on common indicators (EC, 2006).But what is the difference among evaluation, measuring and benchmarking? Jones,Irani, Sharif, and Themistocleous (2006) state that evaluation aims to determine thevalue and benefit derived of the e-government investments. According to Jones, Irani,and Sharif (2007), this analysis should investigate various perspectives, not onlycitizens in general, but also require the inclusion of specific needs of the specific targetgroups of citizens that are using a particular e-government services (Janssen & vanVeenstra, 2005), identify and quantify benefits and consider social and technical contextof use (Alshawy et al., 2007). Benchmarking is a well known practice in business world used mainly for marketingand sales purposes. It referrers to perform systematic comparisons with competitors inorder to visualize leadership and position of a company into a list or ranking ofcompanies. According to Heeks (2006), in e-government field, benchmarking meansundertaking a review of comparative performance of e-government organisations at thenational and international levels. These comparative exercises contribute to a broader view of the reform process inwhich public organizations are, identify leaders and followers, understand differentstages of growth and identify best practices (Heeks, 2006). Hence benchmarkingbecomes time by time a more common practice in public sector management acting as a“reality check” for managers and policy-makers and allow them to measure the progress(Pacific Council, 2002). In a successful execution of an e-government strategy,benchmarking through indicators is critical component of the implementation process(Graafland-Essers & Ettedgui, 2003).Both, evaluation and benchmarking are private sector techniques introduced into publicsector management in order to measure public sector outcomes and performance(Noordegraaf, 2003; Cole & Parston, 2006). Although recognizing the importance of e-government evaluation, this research report deals with e-government benchmarking. After understanding the importance and the real dimension of e-government, in the nextsection an in-deep assessment of the state-of-the art in e-government benchmarking isperformed.

2.2.2 Different proposals of e-government benchmarkingThe different e-government dimensions that have been discussed in previous sectionsbring different possible benchmarking exercises. According to Bannister (2007) these

18

Page 19: Local e-Government Bench-learning

Local e-Government Bench-learning

multiple dimensions of e-government creates problems for would-be benchmarkers.Following in this section we will review the existing benchmarks and comment thegeneral characteristics to see that not only there is a clear lack of local benchmarks butalso there are no benchmarks facing supply and demand of electronic services.Kustelj and Vintar (2004) offers a view of the benchmarks carried out in e-governmentarea up to 2004 which is still interesting despite the years passed. In that study, eachbenchmark is characterized by the e-government dimension in which it is focused. Thee-government dimensions that Kustelj and Vintar take into account are: government e-readiness (which includes citizens e-readiness and business e-readiness), back-office

19

Table 1: Benchmark surveys from Kunstelj and Vintar (2004)

Page 20: Local e-Government Bench-learning

Local e-Government Bench-learning

(which includes organizational aspects), project management and internal process re-engineering, front-office (that includes official website and e-services, supply anddemand), and finally, effects and impacts as generated public value. It should be noticedthat a benchmark can be positioned in more than one dimension in this table (table 1). A total of forty-one benchmarking surveys were identified: fourteen of them dealingwith e-government readiness, fifteen are about citizens and business readiness, only twodeal with the back-office as the primary focus (although other four included it in someway); twenty-six had services provision as main topic, fourteen studied services'adoption in some way, but only three benchmarked both the supply and demand ofonline services1, with four focused on e-government impacts or effects as a main topic,while six more took this topic as supplementary. Codagnone, Caldarelli, Cilli, Galasso, and Zanchi (2006) in the eGEP project updates2the Kunstelj and Vintar (2004) list of e-government benchmarking. This new editionreports sixty-four benchmarks and studies about e-government measurement classifiedin four clusters or categories: e-readiness, supply side, demand side, impacts. The mainresults of this survey is that the overwhelming majority of the reports focuses onsupply-side indicators (number of services available online) and/or e-readiness(presence/absence of structural and institutional conditions for the development of e-government and more in general of the Information Society), while an increasing, butstill limited number, considers the demand side (i.e. take-up and satisfaction withservices). Despite the potentially broad scope of e-government, the majority of benchmarks arefocused on the front-office (Janssen, 2003; Kunstelj & Vintar, 2004), which is the arearelated with official web pages and citizen-related e-services. This is not surprisingsince it contains the most visible parts of e-government. In addition, it has to bementioned that there is a tradition of e-government models and standard indicators thatprovide a good basis to perform measurements and comparisons in that area, as we willsee later on in the next section.Another interesting dimension shown in the Kunstelj and Vintar table (table 1) is thegeographic area in with these benchmarks are performed. Only seven applied to EuropeUnion countries (EU-15 that time) and seven world wide. Unfortunately, the focus onlocal public administration is not analysed, but that is the point in the next section.To summarize, benchmarks have mainly evaluated only service delivery (without takinginto account demand or adoption) at the national government level (Löfstedt, 2005)despite the importance of local administration in relationship to citizens. Thischaracteristic will be deeply analysed in chapter four.1 These 3 studies are Cullen and Houghton (2000) in New Zealand, NOIE and DMR (2003) in

Australia, and REGIONAL-IST (2003) developed in Spain, Germany, Portugal, Italy and Hungary.2 Although Codagnone et al. (2006) do not mention Kunstelj and Vintar (2004) works, the coincidence

in the survey of the performed e-government benchmarks, the proposed classification and the labelsused in the resulting table give enough clues to understand that this group of researchers knew thework performed by Kunstelj and Vintar.

20

Page 21: Local e-Government Bench-learning

Local e-Government Bench-learning

2.2.3 Local e-government benchmarkingAs we have already commented, Kunstelj and Vintar provide information about thegeographic scope of the benchmarks (which is not the case of Codagnone et al. survey).In that analysis can be observed that there are no benchmarks focused at local level. Other authors and researchers highlighted this lack of local based comparisons.According to Sakowicz (2003), most of studies take into consideration the wholecountry, i.e administrations and government at all levels. Heeks (2006) concludes thatthe majority of benchmarks have focused on national e-government. Although indeveloping countries, it is local governments particularly that are the main point ofcontact for delivery of services. Lanvin and Lewin (2006) remarked that despite theanalytical efforts that have been made to describe local e-government initiatives andtheir good practices, remarkably “little attention has been granted to measuring the e-readiness of sub-national spaces, including cities”. And they point out at the studiesperformed by the World Bank (Kaufmann, Kraay, & Mastruzzi, 2005) mainly focussedon city governance (including indicators on state capture, informal money laundering,red tape, and trust in politicians, as well as bribery in affecting utilities, laws, andpermits), and the other by Rutgers University in collaboration with South Korea’sSungkyunkwan University in Seoul (analysed in chapter 4) as two of the moresystematic attempts to measure “urban performance” or competitiveness made.Why there are not benchmarks at local level? Is the opinion of United NationsDepartment of Economic and Social Affairs that the different role cities play in differentcountries makes comparison difficult. According Cole and Parston (2006), the diversityin socio-economic environments and context makes difficult compare performanceacross institutions. In addition, de diversity in services and functions makes even moretough data collection and comparison. Collecting internationally comparable data at thelocal level – where it even exists – is especially difficult due to differences in politicaland economic systems. A public function that is highly centralized in one country maybe highly decentralized in another (UN, 2010). This different role played by cities indifferent countries is one of the challenges of the Local e-Government Bench-learningthat will be presented in the following chapter.

2.3 e-Government modelsThe development of information systems architecture can evolve through a number ofphases or stages of growth (Janssen & van Veenstra, 2005). In this section, it will bepresented and discussed some of the existing models of growth applied to e-government. It will also exposed that despite their utility, e-government models do notreflect the real complexity of the e-government model and need to be updated.Stage models aims at de-constructing information systems architecture developmentinto a series of stages, development or evolution goes from one stage to another(Andersen & Henriksen, 2006). This general approach also applies to e-government as a

21

Page 22: Local e-Government Bench-learning

Local e-Government Bench-learning

process in which information systems architecture plays a key role. In that case, it takesalso the name of e-government models. Therefore, e-government models, stages ofgrowth, models of maturity or models of sophistication are all synonyms when talkingabout e-government.

2.3.1 The early modelsAlthough stages of growth were introduced years before, the main e-governmentmodels appeared in 2000 and 2001 (Baum & Di Maio, 2000; Hiller & Belanger, 2001;Layne & Lee, 2001; Ronaghan, 2002; Wescott, 2001). Several authors proposeddifferent e-government models. All of them illustrates the organizational stage in an e-government development process (Andersen & Henriksen, 2006) and outline thestructural transformations of governments as they progress toward electronically-enabled government (Layne & Lee, 2001). e-Government models are important because they are offering a basis to measure e-government development degree and to guide strategies (Andersen & Henriksen, 2006).According Norris (2009), models are providing a good framework to asses developmentlevel easy to use and shared by other practitioners. It is thanks to the existence of thesemodels that e-government benchmarking progressed. Therefore, e-government modelsproved to be a useful tool to assist policy makers in devising their own plans andinitiatives (Al-Hashmi & Darem, 2008).

22

Illustration 1: Baum and Di Maio four stages model

Page 23: Local e-Government Bench-learning

Local e-Government Bench-learning

However, the enormous diversity that is included in the e-government definition meansthat designing a general assessment model is very complicated and it is almostimpossible to include all possible variants (Kunstelj & Vintar, 2004). This multifacetedreality of e-government is also the reason why they appeared several e-governmentmodels. Some of them are focused in internal IT adoption and organizational changes (Layne &Lee, 2001), others3 have been designed to model development of web content, orservices provided through Internet (Ronaghan, 2002), others are oriented to classifyprojects in an overall evolution of and e-government Strategy (Baum & Di Maio, 2000)or focusing on technology infrastructures for e-government as the IBM model (Nguyen,Sansoni, & Le Noir, 2003) The Baum and Di Maio model (Illustration 1) is also known as Gartner model. It wasthe first published in 2000. It is a four stages model aiming to classify e-governmentprojects. The first stage is a mere web presence, in which governments provide basicinformation on line. Next is a stage in which citizens can interact with governments online. In the third stage this interaction is transactional. Citizens can conduct business online with governments. The final stage implies a transformation of the relationshipbetween citizens and governments. These changes due to e-government produce a muchmore citizen centric and responsive government.

Layne and Lee defined in 2001 a four stages model (Illustration 2) to describeorganizational changes towards a full enabled e-government. The first stage is3 United Nations Division for Public Economics and Public Administration

23

Illustration 2: Layne and Lee stages model

Page 24: Local e-Government Bench-learning

Local e-Government Bench-learning

“catalogue”, it is similar to the initial stage of the Baum and Di Maio model. It ischaracterized by a basic web presence providing static information. The second stagecalled “transaction” includes online forms and services supporting online databaseaccess. The third stage is characterized by the vertical integration with other publicagencies processing related functionalities and finally in the fourth stage appears thehorizontal integration of information and services which means the sharing online dataand information across departments within governments and among governments toprovide the final service. The Hiller and Belanger model, also introduced in 2001, has five stages. It starts with“information”, followed by “two-way communication”. The third is “integration”,followed by “transaction” and finally “participation”. This latter stage, what makes thedifference with previous models, is close to e-democracy and permits citizens toparticipate electronically in their governments. Despite the existence of several e-government models, the European Commission chosea new model (from now Cap Gemini model) based on the method developed by theDutch government. It is a four stages model to measure the level of onlinesophistication of the services. This model was adopted for Cap Gemini Ernst & Youngin 2001 (Cap Gemini, 2002, 2003, 2004, 2005, 2006) to perform the survey onelectronic public services commissioned by the European Commission in theframework of the eEurope 2002 Action Plan (EC, 2000), becoming a standard de facto.The four levels are defined below:

1. Information: online info about public services2. Interaction: downloading of forms3. Two-way interaction: processing of forms, includes authentication4. Transaction: case handling; decision and delivery (payment)

Another stage model designed for official website assessment that worth to becommented is the used by United Nations in its e-Government Survey performed in2003, 2005, 2008 and lately in 2010. Initially the proposed model defined by Rutgers University was based on four stagesbeing the first when does not exist on the website information at all, thus scoring zero.Therefore, to better compare this model with the previously exposed it can beconsidered as a three stages model. The stages are defined as follows:

1. Information about a given topic exists on the website (including links to otherinformation and e-mail addresses)

2. Downloadable items are available on the website (forms, audio, video, and otherone-way transactions, pop-up boxes)

3. Services, transactions, or interactions can take place completely online (creditcard transactions, applications for permits, searchable databases, use of cookies,digital signatures, restricted access)

These first models suggest a lineal, stepwise and progressive evolution increasing in24

Page 25: Local e-Government Bench-learning

Local e-Government Bench-learning

complexity and sophistication (Baum & Di Maio, 2000; Layne & Lee, 2001). Eachfurther step in the models’ predicted evolution of e-government produces more e-government and e-government qualitatively better (Norris, 2009). As it has been said,their strength is that they propose a practical and comprehensible framework (Baum &Di Maio, Layne & Lee) easy to use.In the models defined before 2006, there are some problems; one is regarding pro-activity and services simplification/integration which leads to a service suppression asthe highest stage of e-government maturity. According to Janssen, Rotthier and Snijkers(2004), that means “neglecting the more fundamental process of re-defining servicedelivery in an online environment”, a process that leads towards less but better services. When analysing advanced public services, those models are not working correctly.There is a problem with levels of interactivity and interoperability or integration. Thereis no way to reflect reach interaction and participative service provision by means ofweb 2.0 technologies.The purpose of the Cap Gemini model is services maturity assessment. It is using fourstages to reach transaction level while other models are using only two (Layne & Lee)or three (Baum & Di Maio) to reach the same sophistication level. This characteristicmakes the model more suitable to explain the e-services development, since it offer amore detailed and incremental stages of sophistication. On the other hand, it evaluatesneither organizational e-government action like the e-government re-designing of back-office procedures, nor service availability through other channels, nor the adoption andthe use of these services, nor the impact of the e-government programmes.But although focusing in service delivery, the model is well developed for singleservices evaluation but do not take into account service integration (Janssen, 2003;Kunstelj and Vintar 2004). And it is again a lineal model, although it is recognized sinceits inception that not all services can evolve through all sophistication levels.

2.3.2 The extended modelsJanssen (2003) observed that the existing models proved to be a useful tool for incipientand mid term e-services development but are not reflecting correctly the characteristicsof the most advanced actions in service provision. And stated that more research wasneeded in this area. In that sense the extended version of Cap Gemini and UN modelsare proposing some corrections to solve the aforementioned problems.Janssen, pointing out at what have to be the aim of e-government in service provision,noticed that existing models are not useful in high levels of sophistication. According tothis author, the highest levels of sophistication might actually be the proactivecompletion of the transaction within government or even its elimination. In that sense, the Cap Gemini model evolved in 2007 adding a 5th level ofsophistication named “personalisation” to reflect the concept of pro-active service

25

Page 26: Local e-Government Bench-learning

Local e-Government Bench-learning

delivery, (i.e. the government pro-actively performs actions to enhance the servicequality and the user friendliness), and the idea of automatic service delivery (i.e. thegovernment automatically provides specific services to citizens or business, so there isno need for the user to request the service)(Cap Gemini, 2007, 2009).Also, the UN model was further extended in order to take into account new trends inelectronic service provision based on the introduction of web 2.0 technologies.Therefore, the model was presented as a four levels model in the 2010 edition asfollows:

1. Emerging information services. Government websites provide information andlinks to ministries, departments and other branches of government.2. Enhanced information services. Government websites deliver enhanced one-way

or simple two-way e-communication between government and citizen, such asdownloadable forms for government services. Some limited e-services enablecitizens to submit requests which will be mailed to their house.

3. Transactional services. Government websites engage in two-way communicationwith their citizens. Electronic authentication of the citizen’s identity is required.

4. Connected services. Government websites have changed the way governmentscommunicate with citizens. They are proactive in requesting information andopinions from the citizens using Web2.0 and other interactive tools. E-servicesand e-solutions cut across the departments and ministries in a seamless manner.E-services are targeted to citizens.

But these extended models still do not explain the observed evolution of some servicesin terms of interactivity and interoperability. This problem was already pointed out byKunstelj and Vintar in 2004. This author noticed that lineal models do not allow toexplain the right levels of interactivity and interoperability or integration in electronicservices. Sophistication and integration should be seeing are two different variables, twodifferent dimensions. One, the sophistication of a service is a a front-office measure ofhow much can be accomplished online, the other one, integration of a service reflectshow mature and advanced is the back-office (Heeks, 2006). Therefore, according toKunstelj and Vintar, two-dimensional models should be developed.As a conclusion it should be noticed that according to Norris (2009), the models of e-government “were developed in a vacuum”. Which means that they were not based onresearch or on a careful reading of relevant bodies of literature. They were alsotechnologically deterministic. As such, “they were almost certainly destined to bewrong”. And that must be taken into account when evaluating a specific e-governmentarea. Models are providing a good framework to asses development level easy to useand shared by other practitioners, but at the end, they are an incomplete description ofthe reality.

26

Page 27: Local e-Government Bench-learning

Local e-Government Bench-learning

3 The Local e-Government Bench-learning survey

3.1 OverviewThe Local e-Government Bench-learning study, is a bottom-up grounding exercise thatintends to cover the lack of useful information helping to measure e-government at citylevel and providing a methodology for further measures. It is an ongoing study carriedout by a group of 15 European cities led by Barcelona, with the technical support ofPENTEO ICT Analyst company and the Pompeu Fabra University in Catalonia asScientific Advisor, within the Knowledge Society Forum of the Eurocities Network. The aim of this study has been to perform a benchmark on e-government services andapplications provided by several local administrations. All the benchmarks undertakenso far has been mainly focused at a national level, and the cities believed in the need topatch the lack of global indicators that can be applied at local level. The study shouldresult in a measurement framework and a common set of indicators that would helpdecision-makers in municipalities to better manage the e-government transformationprocess. Furthermore, it should provide the basis for the future assessment of theeffectiveness of e-government strategies.Up to now, benchmarks typically have been restricted to the services offering, i.e., e-services delivery or provision. The study intends to make a significant contribution in e-government measurement by extending it to the citizens' adoption (in European cities).In addition, this approach leads us to an improved learning process as it allows us toidentify best practices based on the deeper analysis of the citizens' adoption. This is whythe study was called a “bench-learning” project.The methodology used has been based on an information gathering process divided infour stages: data collection of sample cities common services, provision of ameasurement framework, surveys delivery to assess maturity and acceptance level and,finally, analysis and a final report covering the expected outputs. An innovative common measurement framework has been defined which should be thebase for further analysis. When comparing with existing benchmarking methodologiesseveral innovations can be identified. The first important differentiating characteristics is the corpus of common servicesshared by all the European Union cities despite different member state policies andpublic administration organization. That corpus, denoted as standard services is thebasis for all the measurements performed. A part from the standard services the studyhas take into account the additional services that are offered by city administrationaiming to solve local needs of their citizens or just trying to raise citizens awareness. The second innovation of the study is the Global Services Catalogue which groups

27

Page 28: Local e-Government Bench-learning

Local e-Government Bench-learning

services into nine thematic categories (i.e. transport, social care, education...) thusproviding a sound view of the services offered to citizens. Results are providedaccording to these nine thematic categories. Two variables called standard coverage anddiversity have been introduced to characterize the defined categories. Analysis of thesetwo variables gives interesting conclusions about policy impact. An innovative servicesmaturity stages framework has been defined, extending the Braun and Di Maio fourstages framework to reflect the web 2.0 characteristics of most advanced services. Whenaiming at benchmarking success of e-government services, provision should beconfronted with citizens adoption, that is why a measurement framework for citizensadoption has been introduced in the study. A third and most important innovation lies in the departure from the typical ranking listthat benchmarks use to offer as main output. The survey is neither providing a rankingof cities nor an indicator to condense the results of the data analysed for each city. TheLocal e-Government Bench-learning provides an innovative format to present theresults through the use of maps presenting at the same time maturity and adoption levelwith a strategy of comparing each city results with European averages, that provides away to represent the results in a comparative way without falling into fostering anegative competition among cities. To better characterise the e-government city profile,a new variable has been introduced (lambda) informing about the coherence in the e-government development across all nine categories and finally, a new variable (rho) hasbeen defined to measure the “distance” of a service regarding the European average.This variable allows both to identify in a numerical way which services can beconsidered as a best practice and also to qualify or score the good practice. The results obtained in this survey show, first of all, that there is an important corpus ofstandard services shared by all European cities. This corpus reaches a 53% of theservices analysed (43 services of a sample of 81) thus providing a good basis formeasurement and comparison. A second finding to be mentioned is that servicesprovision ranges between 2 a 3,5 in all the categories which means that, services mainlyare at interactivity and transactional levels with a few at the transformational one.According to the results obtained it can be derived that adoption of online services is notan issue as they are used, on average, more than expected.However, all these results are not significant at a European level, as the sample of citiesis very small. The main objective of the survey is to define a methodology to benchmarkonline public services provision and citizens adoption in European municipalities. As aconsequence of the methodology definition a first round of measures is performed andso the survey will also provide the first set maps describing the overall situation and thedetails per service category and per city. Finally, the survey aims to supply a selection ofbest practices.The project team was composed by the project director4 (belonging to the Municipality4 The author of this research work has been also the project director of the Local e-Government Bench-

learning survey and who conceptualized the bench-learning methodology as a bottom-up exerciseaiming to provide useful information for e-government managers and covers the existing gap in how

28

Page 29: Local e-Government Bench-learning

Local e-Government Bench-learning

of Barcelona, a support technician (from Barcelona as well) an IT analyst fromPENTEO IT Analyst and an academic advisor from Pompeu Fabra University.

3.2 MethodologyThe methodology was based on PENTEO METRICS methodology with some ad hocmodifications to ensure a bottom-up approach. PENTEO METRICS methodology isbased in a five steeps process: design, measuring, analysis, benchmarking and report. Inthe design steep, it take place the definition of participants, calendar, roles, definition ofindicators, questionnaires and data-gathering process. The second steep is “measuring”,is in this steep when data is gathered, usually using interviews, questionnaires anddocumentation. In analysis steep, data is processed and indicators are generated. Thebenchmarking phase is in which data from each participant is compared with a referencegroup and scoring is calculated performing the typical ranking list. Finally in the reportphase, the final results and recommendations are elaborated and presented.A bottom-up approach was needed to ensure that meaningful results are obtained at theend of the survey. Therefore some modifications were introduced in the methodology.First of all a project committee was created with the participation of all cities. Thecommittee, chaired by the project director, meet regularly a minimum of one time perphase to asses the progress and discuss about next steeps. Second, all data gathering wasbased on self-assessment using questionnaires to be fulfilled by each city e-servicesleader. Neither interviews were performed nor other documentation was used.Benchmarking phase was also modified in order to compare data only with theEuropean average. Finally, specific report was designed to make results more useful forcities to progress and guide further steeps in their e-government plans.Cities were involved from the beginning, participating in several meetings to providefeedback about which services should be benchmarked and how to group these servicesin a meaningful way which we describe latter on, how to assess online services maturityand also adoption; and finally how results should be presented for bench-learning toprovide useful information to city managers.A first questionnaire was designed to ask for the list of main services provided by eachcity. Data gathered was compiled and a first general list of services was created. Asecond questionnaire was circulated among cities showing the list of all compiledservices. This second questionnaire was used to ask cities about which of the servicesprovided by other cities are also provided by them.After this two rounds a list of 81 services was compiled. During the gathering processwas discovered that on each European country cities have slightly different role an thatgive some differences in the set of functions to accomplish. Therefore, different set ofservices are found. In some cases cities are not responsible for a service provision but

local public administrations are performing in e-government. It was, at the same time, who designedthe project team and recruited its members.

29

Page 30: Local e-Government Bench-learning

Local e-Government Bench-learning

they are offering information and even access to the service provider (usually a higherlevel public administration). Therefore, it was needed a third round of questionnaires topick up detailed information about each service provision.The assessment of the maturity and adoption was also discussed and prepared with theparticipant cities. The information was gathered using questionnaires which wereanswered by service's responsible of each city. The most mature services were doublechecked, and detailed information was asked to be provided to ensure that highest levelsof maturity were understood and applied correctly. There was no reference group to compare with. This is one of the differences of thisstudy versus a traditional benchmark, here each participant city has been comparedagainst the European average of the sample.

3.3 The sample of cities and its analysisThe participant cities were recruited through mailing in two European networks ofcities: EUROCITIES with more than 130 members and Major Cities of Europe. Thecities were asked to pay a fee of 3.000 Euros to participate in the survey thus allowingto partially cover the cost of the study. During the recruitment process, eighteen citiesjoined the study and started working. Throughout the process, three cities in the surveydropped: Malmo (Sweden), Sanliurfa (Turkey) and Lyon (France). The reasons that moved cities to joint the study were also diverse. An important part ofthe cities were at the middle of the e-government plan deployment process(Birmingham, Barcelona, Turin, The Hague, Vienna) and were checking there were onthe right way. Other cities were in the process to draw a e-government strategy (Murcia,Milan) therefore searching for information about possible services and good practices.Some cities were just released the e-government strategy and starting the deployment(Bergen) thus willing to stablish an initial international comparison to further assess theprogress. Other cities with a recognized leading position were searching to validate itsposition at European level (Helsinki, Enschede)The sample of cities was formed by a total of 15 European cities (Table 2). It is a ratherheterogeneous sample reflecting the diversity of urban areas in Europe. It includescapital cities as Tallinn, Helsinki or Vienna, second cities as Milan, Birmingham,Barcelona, Bergen or Rijeka. Cities over a million inhabitants as Birmingham, Vienna,Milan or Barcelona, Medium cities as Helsinki, and small cities with less of 250.000inhabitants. Out of the 15 cities, 20% are capital cities and 33% are second cities in their countries.In total, the fifteen sample cities represent:� 10.087.736 inhabitants which is the 2% of the EU27 population. Cities are

going from144.043 inhabitants from Rijeka, to 1.677.867 inhabitants fromVienna. The average population per city is 672.516 inhabitants.30

Page 31: Local e-Government Bench-learning

Local e-Government Bench-learning

� 214.334 civil servants (going from 455 civil servants of Rijeka to 57.000 ofBirmingham. The ratio of habitant per civil servant goes from 15 for Bergen andHelsinki to 241 for Tallinn or 317 for Rijeka.

The sample is quite representative in terms of GPD per capita since the sample averageis 26.744€ (going from 12.305€ per capita of Rijeka to 60.322€ per capita of Bergen)close to the EU27 average GDP that is 25.100 for 2008. In terms of GDP per capitavariation, our sample is slightly more homogeneous than the EU27 reality since theEuropean GDP varies from 1 to 6 across the EU27 (source: EUROSTAT) while in oursample it varies from 1 to 5.Regarding household internet access the sample is also representative of the Europeanreality since the sample household internet access is 61,4%, ragging from 39,40% inMurcia to 91% in The Hague, while the EU27 average is 60% in 2008 (source:EUROSTAT).Furthermore, it is also a representative sample in terms of population age, since thesample average is 40,2 when in Europe was 40,6 in 2007 (source: EUROSTAT).However, our sample has a higher share of people living in cities over a million, the55%, (49% including Lyon, Malmo and Sanliurfa) while EU27 has an average of 33%.Apart from the excessive weight of large cities, that could have been corrected includingmore small and medium cities, the sample is quite representative of the European urbanreality.

31

Table 2: Cities characteristics

464.038 7.949 12,90% 39,25 65,00% 30.000 58,4

BCN 1.595.110 13.393 21,30% 43,00 60,70% 25.651 119,1

250.985 17.123 31,00% 38,30 64,00% 60.322 14,7

Bilbao 354.180 2.504 17,00% 44,00 45,70% 19.648 141,4

1.000.600 57.000 22,40% 33,00 55,00% 23.213 17,6

372.256 4.926 26,20% 47,70 55,60% 35.156 75,6

155.000 1.700 6,00% 38,30 83,00% 25.200 91,2

560.994 38.623 28,30% 40,00 85,00% 27.258 14,5

1.304.263 17.640 17,00% 37,00 43,00% 39.442 73,9

422.861 2.058 11,10% 37,00 39,40% 20.312 205,5

144.043 455 18,49% 41,20 45,00% 12.305 316,6

401.372 1.664 43,40% 67,90% 16.576 241,2

475.904 7.850 34,00% 38,10 91,00% 12.400 60,6

908.263 12.000 9,00% 44,90 50,00% 19.604 75,7

1.677.867 29.449 33,00% 41,00 65,30% 29.449 57,0

672.516 14.289 22,07% 40,20 61,04% 26.436 104,2

InhabitantsCivil

Servants

Population with university

degree

Population average

age

Household Internet

Access(%)

Per Capita GDP

Inhabitants/civil servant

Antwerp

Bergen

Birm ingham

Bologna

Enschede

Hels inki

Milano

Murcia

Rijeka

Tallinn

The Hague

Torino

Vienna

Average

Page 32: Local e-Government Bench-learning

Local e-Government Bench-learning

3.4 Services maturity levelsService maturity has been measured since the beginning of the e-government. The mostaccepted framework is the four maturity levels proposed by Baum and Di Maio (2000):information, interactivity, transaction and transformation. Our study uses thisframework as the starting point extending it including a fifth level on top of them. Thislevel reflects a more advanced degree reflecting new features in user's interaction takinginto account the use of emerging web 2.0 technologies. This new trend means lettingand promoting citizens’ participation in service provision, quality evaluation anddiscussion about services’ improvement. This is why the new level is called the“participation level”. Therefore, the measurement framework used to assess e-service maturity was based inthe five levels defined below:

1. Information Level: Ability to offer relevant information in a basic provisionprocess, usually website based.

2. Interactivity Level: Users can generate basic content -email, template, queries-which is introduced in the government databases. It is a one-way interactionprocess in which the user can start but not complete a procedure on-line.Physical attention at the municipal counters is still required.

3. Transaction Level: The achievement of this two-way interaction level impliesthe possibility to completely perform a service –payments, certificates- throughthe use of electronic means.

4. Transformation Level: Full availability on-line, implying a full integration for alle-government services into a single portal. Users can access all the services froma single virtual office from any place. This level usually implies a businessprocess re-engineering both in the back office and the front office.

5. Participation Level: Let and promote citizens participation in all e-services sothat opinions are taken into account to enhance quality and effectiveness. Thisfifth level provides an indication of the extent to which on-line provision isbased on new models to use available information obtained from the citizenship,reaching a high degree of pro-activeness in services’ delivery. In short, this levelis based on the application of the web 2.0 strategies to the on-line municipalservices delivery.

3.5 Service adoption levelsMost studies so far did not face measuring the e-service adoption because of its inherentdifficulties that we will explain in deep in chapter four. Web site visits or number ofpages served, are indicators commonly accepted to assess the popularity of a website.But adoption or take up of an e-service can not be measured by access only. We need to find a suitable adoption indicator. Data on transactions performed trough an

32

Page 33: Local e-Government Bench-learning

Local e-Government Bench-learning

e-service application is a better indicator. However, it might not be applicable to thewhole range of services due to their different levels of maturity and the heterogeneity ofIT solutions used to provide them. Cities have found difficulties to share a singlemagnitude to measure adoption in a consistent way and the best magnitude could bedifferent for each e-service.There are other factors that must be taken into account to measure real adoptiondifferent. Each service is focused on a specific target group of citizens (youngsters,elderly people, entrepreneurs, immigrants…). As not all services are focused on samegroup of citizens, the characteristics of the target as literacy, digital literacy or ratios ofinternet access, for instance, define success in a different way. Furthermore, amultichannel perspective is needed to assess the success of e-services delivery. That iswhy only few benchmarks and studies have faced this issue up to now.In our study, we measured the perceived adoption instead of measuring the citizensadoption. Each service's responsible measures perceived adoption by comparing thecurrent use with the use initially expected. This variable can be measured by self-assessment using a measurement framework based on five levels. Each level defines adegree of the perceived adoption, ranging from a complete failure to a big success asdefined below:

1. No adoption. The service has been a complete failure and the municipality isconsidering its withdrawal.2. Low adoption. The service is being used by fewer citizens than previouslyexpected.3. Medium adoption. The service is being used by the average of citizens expected.4. High adoption. The service is being widely used among citizenship.5. Excellent adoption. The service has been a complete success and the

municipality is devoting more investment on it.

3.6 The global services catalogueThe Global Services Catalogue has been developed following a common agreementamong the participant cities, and integrating all the common and outstanding servicesthat they offer. Each city has drafted a list with a description of the most relevant e-services they provide. Amongst all the e-services gathered, a common set of serviceshas been selected on which the bench-learning exercise has been performed. In total, this first release of the Global Services Catalogue includes eighty-one servicesthat have been grouped according to nine main categories, listed below:

1. Channelling consists in 10 services (6 standard, 4 additional) related to enhancethe contact between the citizenship and its municipality. It includes city’s webpages, personal folders or bulletins.

33

Page 34: Local e-Government Bench-learning

Local e-Government Bench-learning

2. Citizens’ Engagement consists in 10 services (6 standard, 4 additional) devotedto satisfy citizens’ expectations, attract their attention and consolidate its bonds.It encourages the relationship between the citizens and their administrationspromoting or being on themselves a pre-stage of e-democracy.

3. Education consists in 13 services (8 standard, 5 additional) related to satisfycitizens’ educational needs. It includes services related to children, adult andvirtual learning services, cultural projects and libraries.

4. Employment & Business consists in 7 services (4 standard, 3 additional) devotedto help citizens obtain a work, facilitate business procedures, recruitment,tendering and small business promotion.

5. Environment consists in 7 services (3 standard, 4 additional) related toenvironmental care, including garbage collection and maintenance of thecommunal properties as well as encouragement of habits related to sustainableurban growth.

6. Lifecycle consists in 10 services (6 standard, 4 additional) devoted to ease thenecessary transactions among citizens and municipalities related to their dailylife, such as personal certificates, digital documents management systems orcensus.

7. Social Care consists in 14 services (5 standard, 9 additional) related to integratedservices that are available from public health and social care providers, includinghousing services and care for people with special needs (disabled, children oraged people).

8. Transport consists in 6 services (2 standard, 4 additional) devoted to satisfycitizen’s mobility needs.9. Urban Planning consists in 4 services (3 standard, 1 additional) related to land

use and property related issues like construction permits, building regulations,certificates or land acquisition.Once categorized, all the services have been divided according to their percentage ofcoverage between standard and additional services. European coverage is understood asthe percentage of cities that provide a service. Therefore, if a concrete service is offeredby all the sample cities its European coverage is 100%. If a service is offered by 50% ormore cities it is a Standard Service. On the contrary, it is an Additional Service. In total,43 services from 81 has been identified as standard services and 38 from 81 areclassified as additional.Standard services are the 53% of the identified services. They show common interests inEuropean municipalities and the route map that the sample cities have been recentlyfollowing in e-services deployment. On the other hand, additional services, 47% of theidentified services, show the wide variety of services provided by European cities.

34

Page 35: Local e-Government Bench-learning

Local e-Government Bench-learning

A chart showing the standard services list and which cities are offering them ispresented to summarize the main characteristics of each category. Therefore, the studyshows 9 charts for services provision. In each chart, services are characterized accordingto whether the service is provided by the city or not.

3.7 European standard coverage and diversityWhile standard services seems to identify general patterns, additional services shouldsurely complement them since the fact that a general pattern exist doesn’t mean thatthere aren’t important and interesting exceptions. These exceptions will be expressed inthe form of a diversity mark focusing in the variety of circumstances that exist whilestudying e-government in several cities. In short, the study of diversity avoids anexclusive focus on what is common or on dominant patterns, a defect commonly donein previous benchmarks at national level.Diversity compares the current coverage of all services (both standard and additional) tothe perfect scenario: the total provision of all services by all cities. Therefore, diversityof a category is defined by the following mathematical expression:

Diversity = 1 – (services provision/perfect coverage)Diversity takes always a value between zero and one. As it has been defined, diversity helps analysing the variations in government provisionin Europe. Specifically, in this study, it is useful to compare coherence in cities deliveryof services for each of the nine considered categories. A large diversity would imply abig difference in service provision among European cities. A category with a diversityof 0.67 will show low coherence, that is European cities would offer quite differentservices achieving only a 33% of coincidence in services delivery in a given category.On the other hand, it is also of great interest to restrict our attention in the delivery ofstandard services: those services offered by 50% or more of the participant cities. It isimportant to see how common interests have affected e-services deployment and havehelped to create a shared vision of e-services’ delivery. Hence, it will be helpful toconnect the coverage results with the diversity mark achieved by each one of the ninecategories considered in this study.

3.8 European maps and city mapsThe most important innovation of the survey lies in main results presentation. Thisstudy does not score all surveyed cities an provide the typical ranking list thatbenchmarks use to offer as main output. It is also my opinion that focusing in only onescore or indicator can lead to a behaviour in which organizations avoid experiments andinnovation (Janssen, 2003). The Local e-Government Bench-learning project innovatesin results presentation through the European and city maps (Figure 3). These maps are

35

Page 36: Local e-Government Bench-learning

Local e-Government Bench-learning

bar charts presenting at the same time maturity and adoption score for each category. Having defined the different levels of maturity and adoption and having gathered andqualified all the services according to maturity and adoption, the results are plotted percategory (illustration 3). This chart provides an interesting view of the Europeanaverage local e-government degree of development, and helps to find correlationsamong maturity and adoption.

As it can be seen (illustration 3), this chart not only show the variables' value for all thecategories, but also to grasp a general mark for European e-government. In addition itmakes easier to compare maturity among categories, thus showing the harmonisation inservices provision. Moreover, plotting together maturity and adoption, the chart allowsto visualize correlation among both variables, if it is the case. When plotting separately standard services and additional services, maturity, adoptionand possible correlation among variables are more clear.Regarding cities results, rather than providing a ranking the survey use the city maps.City maps show city's e-government profile against the European average values percategory in a similar format that the European map do. In that case, two charts are usedone for maturity and one for adoption. Using these charts each city can see in whichcategories are more developed and in which they need to improve.The strategy of comparing each city results with European averages, provides a way torepresent the results in a comparative way without falling into fostering a negativecompetition between cities. In addition, to better characterise each e-government cityprofile, a new variable has been introduce (lambda) informing about the coherence inthe e-government development across all nine categories. High marks on lambda wouldindicate that the e-government development is inhomogeneous thus having categorieswith a high levels of services' maturity and at the same time categories presenting low

36

Illustration 3: European map of maturity and adoption0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0

urban planning

transport

social care

lifecycle

environment

employment & business

education

channeling

citizens engagement

maturity

adoption

Page 37: Local e-Government Bench-learning

Local e-Government Bench-learning

marks on services maturity. And the other way around, a low mark on lambda wouldindicate that the city has a homogeneous and well balanced set of e-governmentservices.

3.9 Best practicesThe best practice list is one of the key outputs of the survey. In fact is what brings thereal learning dimension to the study. The proposed methodology brings to us a list ofselected best practice for each category that are related with the real maturity stage ofeach category. That means that it is not necessary to register a high mark on maturity tobe a best practice, what is important is how is the maturity of the best practice candidateregarding the average maturity of the category. To make easier this exercise a newvariable (rho) has been defined to measure the “distance” of a service regarding theEuropean average. This variable helps us not only to identify in a mathematics waywhich services can be considered as a best practice but also to qualify this goodpractice. However, there is a second dimension which is adoption. A best practice has tobe not only a highlighted service (behind the average) but also, should be characterizedby a good acceptance, therefore, rating higher than 3 in adoption. The higher the better. There is no need to tell what a best practice is, but it is important to draw your attentionto the current methods used to identify best practices which are rather arbitrary andsubjective. Here, in this study the innovation is that the best practice selection is aconsequence of the measurement framework and over all methodology defined tocompare different services.

3.10 Results discussionFirst of all, the findings show at first glance the existence of a set of standard servicesshared by all the participant cities. Despite the existing differences in publicadministration organization across European countries, an important number ofcommon services shared by all local public administration can be found. Second, that corpus of common services was distributed in 9 categories. Coverage anddiversity allow some reflections. It is quite remarkable that all the Global Cataloguecategories achieve percentages of coverage in a range between 67% and 83%. Ingeneral, if a majority of cities show a common services agreement, a category could beconsidered as consolidated. However, it is important to pay attention, at the same time,to how populated is a category regarding standard services. A consolidated category willshow: a well established set of standard services and a good European standardcoverage.When a well established category shows at the same time high diversity, there is thecase of Lifecycle or Citizens’ Engagement, means that it is an active category in whichcities are working and putting efforts due to the interest of citizens (demand) or

37

Page 38: Local e-Government Bench-learning

Local e-Government Bench-learning

politicians (strategic plans or even European recommendations). Urban Planning and Channelling show the lowest diversity marks, hence appearing tobe the most stable categories in which municipalities provide very similar services. But,in the case of Urban Planning this stability is also accompanied of a low activity inservice provision since it’s the second less populated category.On the other hand, high diversity scores with a low populated set of standard services –having big or low standard coverage- which is the case of Social Care, Transport andEnvironment, clearly means that the category is not well established, there is a lack ofcommon European understanding despite being an active working area. This could berelated with current EU-funded programs that intended to foster concrete key areas5.Currently, these categories are on the focus of current European policies. Third, local e-government development in Europe is still below transaction level, beingonly more advanced than level 3 in only two categories: Channelling and UrbanPlanning. Fourth, local e-government is in general, perceived as highly adopted by citizens. TheEuropean average of perceived adoption is always above 3 and even above 3.5 for fiveout of the nine categories, which means that in average e-services are being used bymore people than initially expected. Therefore, the provision of public services byelectronic means should be considered as well accepted by citizenry. Fifth, web 2.0 is being incorporated to local e-Services provision, since several citieshave yet started delivering level 5 services. Sixth, additional services are more advanced than standard services. When analysingmaturity it can bee seen that additional services show greater marks than standardservices. Seventh, there is no evidence of relation between service maturity and perception ofadoption. Although it could initially be thought that more advanced services will also bemore adopted, this survey, based on the information analysed, does not prove thatrelation. No correlation can be observed among maturity and adoption. Eight, when plotting standard services a part from additional services, it can beobserved that additional services show higher maturity than standard services but a quitesimilar adoption values, which seams to indicate that additional services are no moreadopted than standard services.

5 European Commission funding programmes are currently focussing in social care, mobility andenvironment among other areas.

38

Page 39: Local e-Government Bench-learning

Local e-Government Bench-learning

4 E-services benchmarking and adoption literature revisitedThe aim of this chapter is twofold. First the state-of-the-art in e-government servicesbenchmarking is reviewed, and the most important benchmarks identified are describedand discussed in terms of applicability to local level e-service benchmarking. Thesecond part of the chapter is devoted to understand the adoption measurement problem.It starts by analysing the state-of-the-art in e-government services adoption andhighlighting the importance of measuring adoption to have a complete view of the e-government service provision. Finally, we show the current difficulties to performadoption assessments.

4.1 Benchmarking electronic services provisionAs we discussed in chapter two, despite the potentially broad scope of e-government,most studies to date have concentrated on measuring maturity in electronic serviceprovision, either by proposing stages of growth models or by examining web-basedservice delivery (Janssen, 2003; Kunstelj & Vintar, 2004; Griffin, Foster, & Halpin,2005; Heeks, 2006), which is not surprising, as according to Peristeras et al. (2002), themajor changes brought by electronic government introduction “appear to have profoundeffects, forcing even a reconsideration of what types of outcomes public administrationsshould produce, and what kind of interrelations with civil society actors should maintainin order to fulfil their essential social functions”, and that directly impacts in the serviceprovision.Two important aspects are contributing to maintain e-services benchmarking on thefocus of practitioners and academics. One is the inherent importance of the front-end: inpublic administrations, the front-end includes web official web page and e-servicesprovision and both are the most visible parts of e-government (Janssen, 2003). Thismeans that, in practice, e-government is still perceived mainly as deliveringgovernmental information and services electronically 24/7 (Norris, 2009).The second is the existence of a well established standard in terms of models andindicators. e-Government models to measure service sophistication are widely known,comprehensible and simple to apply. For instance, the benchmark of online publicservices introduced in 2001 by the European Commission established an e-governmentmodel for services maturity assessment, and a set of basic services to be evaluated. TheEuropean Commission defined also the percentage of online public services as the mainindicator to measure the contribution of the European e-government to the objectives ofthe Lisbon agenda. This benchmark, commissioned to Cap Gemini from 2001, has beenyearly measuring the progress of European e-government and its methodology became ade facto standard in Europe. Before turning to the most relevant benchmarks so far, let us state some limitations of

39

Page 40: Local e-Government Bench-learning

Local e-Government Bench-learning

current benchmarks that we will discuss later in this chapter.Benchmarking online service provision should not be confused with benchmarkingofficial websites as it usually happens. Official websites benchmarking try to measure towhich extent public organizations have adopted the web tools as means to contact, andprovide services to, citizens and business. These benchmarks usually cover aspects suchcontent, navigation, public outreach and communication, accessibility, privacy andsecurity, online services, and citizen participation. (Panopoulou, Tambouris, &Tarabanis, 2008). Electronic services provision is one of the common components of theofficial websites benchmarking, and when benchmarking websites the focus is todetermine to which extent the service provision is a relevant component of websites.Benchmarking online service provision is not benchmarking only web serviceprovision. Usually, benchmarks focus on web services delivery thus considering onlythose services delivered through official websites. But benchmarking electronic serviceprovision should take into account all electronic channels used to delivery services tocitizens. It is not only web services provision, but also it includes the use of kiosks,digital TV and mobile devices when used to deliver services to citizens. Although theweb services provision is the most used delivery channel, taking the web as the onlymedium for e-government service delivery the multichannel efforts are missing(Janssen, 2003). Another limitation is that mots of the existing benchmarks focus on state level serviceswith no applicability to local level public administration. Most of them provide a scoreas a main result giving little information about how a city is doing in terms of serviceprovision. Finally, according to Panopoulou et al. (2008), current benchmarks only measure the“supply” side and not the actual use or take-up rate, the “demand” side. Is the opinion ofthis author that this gap encourage countries to get good grades for making lots ofapplications but it does not matter if these applications are actually used by citizens.And this should be corrected more and more by the use of usage indicators and byweighing them together with output indicators.In the following sections I analyse the most relevant benchmarks and comparativestudies in electronic service provision to demonstrate the aforementioned limitations.

4.1.1 Online Availability of Public Services benchmarking, the European de factostandardThe European Union has a systematic evaluation and follow-up of the e-governmentdevelopment in member countries since October 2001 when the European Commissiondefined and launched a benchmarking methodology to measure and compare electronicservices provision among member states6. The benchmark has been performed yearly by6 The Cap Gemini benchmark was defined by the European Commission in the framework of the e-

Europe Action Plans for Member States aiming to ensure a generalized electronic access to main basic

40

Page 41: Local e-Government Bench-learning

Local e-Government Bench-learning

Cap Gemini and now constitutes a valuable source of information to study the evolutionof the European e-government. However, it is not suitable for local level measurementsand has some limitations in terms of maturity stages.One of the main contributions of this benchmark is that it became a de facto standard tobenchmark online government in Europe. The standard indicators defined by theEuropean Commission were: percentage of basic public services available online, publicuse of government online services for information/to submit forms, and percentage ofpublic procurement that can be carried out online (EC, 2002-2), although the benchmarkonly measures the first one.The benchmark is based on the maturity assessment of a list of 20 basic public services(12 for citizens and 8 for business) using a four stages e-government model (see chaptertwo). The list of the 20 basic public services were grouped in 5 categories, allowing amore grounded analysis (see table 3).

Category ServicesTax collection 1. Tax declaration, notification of assessment

2. Corporation tax: declaration, notification 3. VAT: declaration, notification 4. Customs declaration 5. Social contribution for employees

Registration 6. Car registration (new, used and imported cars) 7. Certificates (birth, marriage): request and delivery 8. Announcement of moving (change of address) 9. Registration of a new company 10. Submission of data to statistical offices

Citizen services 11. Job search services by labour offices 12. Public libraries (availability of catalogues, search tools) 13. Declaration to the police 14. Health related services (e.g., interactive advice on availability services in

different hospitals; appointments for hospitals) Financial content 15. Social security contributions: unemployment benefits, family

allowances, medical costs (reimbursement or direct settlement studentgrants

16. Public procurement Permit and licencerequest

17. Personal documents (passport and driver’s licence) 18. Enrolment in higher education / university 19. Application for building permission 20. Environment-related permits (including reporting)

Table 3: Cap Gemini basic public services

In addition, the benchmark introduced the services categories to group similar servicespublic services (Cap Gemini, 2002).

41

Page 42: Local e-Government Bench-learning

Local e-Government Bench-learning

allowing a more grounded analysis. Despite its evolution and refinement over thealmost ten editions, the benchmark is not applicable to local public administrations. Aquick look at the services services shows that most of them are not provided at locallevel thus confirming the focus on state level.Even though the Cap Gemini benchmark contributed to the popularization of a fourlevels model of e-government services sophistication, it does not take into account theuse of web 2.0 technologies in the service provision. Therefore, there is no way toreflect rich interaction and participative service provision by means of web 2.0technologies that are nowadays emerging.

4.1.2 MeGAP: a US local public service benchmarkingIn 2001, Kaylor et al. conducted a benchmark of local public services provision,motivated by the initial request of the city of Ann Arbor (Michigan, US). Thebenchmarking study surveyed 38 cities whose population ranged between one and twohundred thousand inhabitants in the 1990 census, since Ann Arbor falls within thatrange. According to Kaylor et al., the set of functions that municipalities tend to offer issmall relative to the private sector, therefore, benchmarking the progress of e-government implementation should be a simpler matter in the public sector. With apragmatic approach Kaylor et al. selected a palette of services that are performed bylocal administrations in the US. A total of 51 services were identified and grouped in 12categories. Each service was evaluated using a four levels services sophisticationassessment framework. The results were consolidated per category. Then, a summarystatistic was defined to encapsulate all the results and was de base for a score used torank cities. Let us to discus this approach more deeply.A problem with presenting a score is that its meaning is by no means transparent: highscores could indicate both, a wide range of functions that are web-enabled, or a largedegree of technological sophistication -the latter happening in fact in most cases.Another issue is that the score provides little information about e-governmentdevelopment. The services sophistication levels assessment framework used by Kaylor et al. wasdefined as follows, for each giving topic:

1. information about the topic exists at the website2. a link to a relevant contact (either a phone number or email address) exists at thewebsite3. downloadable forms on the topic are available online4. transaction or other interaction can take place completely online

Although it does not differ substantially from other e-government models, Kaylor et al.

42

Page 43: Local e-Government Bench-learning

Local e-Government Bench-learning

model is not based on previously existing ones7 such as those presented and discussed inchapter 2.The definition of services categories made possible the discussion of different degreesof sophistication per category and helped to identify some patterns among cities usefulto make final recommendations. It is interesting to notice that during his research Kaylor et al. found a variety of servicesthat were not anticipated. It was observed that quite a few cities had developedinnovative and sophisticated functions online that were not reflected in the service'catalogue used in the benchmarking survey.This benchmark defined the methodology named MeGAP (The Municipal E-Government Assessment Project) as a benchmarking tool to assess the status ofcity/town e-government. The methodology was updated by the Public SphereInformation Group (PSI Group) in 2005 and it is currently at its third version (MeGAP-3). In this version, the catalogue contains 68 local services to be evaluated in USAcities. These services are grouped four categories.

1. Information dissemination (e.g., city codes, official minutes, traffic information,municipal government directory)2. Interactive functions (e.g., bidder applications, down-loadable forms, buildingpermit process, business license)3. E-Commerce functions (e.g., utilities payment, property tax look-up andpayment, code enforcement)4. E-Democracy (e.g., e-meetings, e-forums, user customization,volunteeropportunities)

The initial services sophistication assessment framework has been maintained. This methodology has been applied or partially applied in some other benchmarkingsurveys as (Flak, Olsen, & Wolcott, 2005) in Norway with a sample of 30 municipalitiesand a palette of 68 local services, and (Arslan, 2007) in Turkey with a sample of 3.228municipalities and a specific palette of 25 local services. Apart from that, it has beenperformed twice in United Sates cities by Public Sphere Information Group. One of thecriticism is that the categories are made mainly in terms of interactivity, therefore, theyare not reflecting thematic areas of services. Second, is that the provision of a score ishiding important details in how a city is deploying e-government. Finally, it does nothelp to identify best practices, although provides the ranking of cities according to the7 But, were they existing? Kaylor et al. benchmark was commissioned in June 2000 “On June 19th 2000,

the City Council of Ann Arbor (Michigan) passed a resolution authorizing the exploration of thepossible expansion of the role of the Internet in providing city services“ (Kaylor et al., 2001). Thebenchmark started in September 2000, which means that it is contemporary to the definition of first e-government models of chapter 2. The first one, the Baum and Di Mayo four phases e-governmentmodel, was published in November 21st of 2000, therefore Kaylor et al. actually started the benchmarkbefore the publication of the first e-government model.

43

Page 44: Local e-Government Bench-learning

Local e-Government Bench-learning

obtained score and a set of interesting conclusions based on the per category aggregatedscore.

4.1.3 E-Government at the American Grass-rootsNorris examined in 2005 the likely future trajectory of local e-government in the U.S.The Norris' research was based on the data provided by two surveys8 conducted in 2000and 2002 about local government adoption of e-government in the United States . The2000 survey was mailed to 3,749 local governments obtaining a 50.2% of responses.The 2002 survey included 7,844 local governments with 52.6% of responses. Norriscomplemented this information with a focus group with 40 local governments. Thestudy analysed the average maturity of the e-government at local level and the likelyfuture trends. Despite it was focused on local level, and mainly in electronic services provision, thestudy can not be considered a benchmarking exercise.

4.1.4 Evolution of Local Government E-services in AustraliaShackleton, Fisher, and Dawson carried out in 2004 a research work to determine theevolution of government e-services in Australian local public administration. Theresearch was based on a benchmarking exercise mixed with a case study. The MunicipalAssociation of Victoria identified 22 main functions or services undertaken byAustralian municipalities (particularly those based in Victoria).These services were grouped in four categories.

1. e-management basic information Web, site navigation, contact details 2. e-service: product and service details, product and service support 3. e-commerce: transaction handling 4. e-decision making/e-democracy: sense of community, links

The research started with a quantitative study of the characteristics of 20 localgovernment websites. This quantitative study identified common features and providedsome indication of maturity levels of those sites. As a second steep, was performed acase study of the approach of one municipal council and a resultant examination of itsmaturity level in the implementation of e-government services (Shackleton et al., 2004).Again here, although the research was focused on local level, the first part consisted8 The author used the data provided by the International City/County Management Association (ICMA)

in their Local Government Survey performed in 2000 and 2002. (http://icma.org/en/press/home).ICMA is a organization of professional local government leaders building sustainable communities toimprove lives worldwide. ICMA, provides member support; publications, data, and information. TheAssociation is an internationally recognized publisher of information resources ranging fromtextbooks and survey data to topic-specific newsletters and e-publications.

44

Page 45: Local e-Government Bench-learning

Local e-Government Bench-learning

more in a official website benchmark than in a benchmark of online services. As all themunicipalities surveyed were placed in the same state sharing same legal framework, iswas easy to list the set of services to be evaluated. This model is not applicable atinternational level. A second criticism is derived to the lack of a well referenced e-government model for services sophistication assessment.

4.1.5 Digital Governance in Municipalities Worldwide: a biased perspectiveUnited Nations has performed three editions of its Digital Governance in MunicipalitiesWorldwide. This prestigious benchmark is based on the Rutgers-SKKU Municipal E-Governance methodology. The survey is a benchmark of the city's official website. Thefirst of such survey was conducted in 2003 through a collaboration between the E-Governance Institute at Rutgers, the State University of New Jersey, Campus at Newarkand the Global e-Policy e-Government Institute at Sungkyunkwan University and co-sponsored by the Department of Economic and Social Affairs, United Nations. Thesurvey was conducted again in 2005 and finally in 2007.The methodology was defined in an previous Holzer's work (Holzer, 2003) as a reactionof the first survey performed by Kaylor et al. (2001). Holzer tested the methodologybenchmarking a sample of ten New Jersey largest cities. The resulting ranking score canbe seen in table 4.

The used methodology consist on analysing official websites established by citygovernments. The instrument for evaluating them consisted of five equally weighedcomponents: (1) Security and Privacy, (2) Usability, (3) Content, (4) Services, and (5)Citizen Participation. For each of the five components, 18–20 measures were applied,each coded on a scale of two to four points.The study provides a ranking of cities per component in addition of the overall ranking

45

Table 4: New Jersey results (Holzer, 2003)

City Population Ranking E-government RankingNewark 1 4Jersey City 2 10

3 9Elizabeth 4 8Edison 5 3

6 7Dover 7 5Hamilton 8 1Trenton 9 2Camden 10 6

Paterson

Woodbridge

Page 46: Local e-Government Bench-learning

Local e-Government Bench-learning

of cities. Since the same study was also performed in 2005 and 2007, the progress canbe visualized and analysed. (Holzer & Kim, 2003, 2005, 2007). The first and more important limitation of the survey is that how the sample of cities isselected. The selection was made taking from ITU data these cities with an onlinepopulation greater than 100,000. Researchers selected 98 countries with the highestpercentage of Internet users, and then “examined the largest9 city in each of thosecountries as a surrogate for all cities in the country” (Holzer & Kim, 2003). In thissample the larger city in each country has been taken as the representative city of thecountry no matter whether it is the most advanced in terms of e-government. This biasshould be taken into account when interpreting the results provided for the survey. A second observation is that the survey is benchmarking official websites, analysingdifferent aspects in which services provision is only one of them. Furthermore, even evaluating service provision, the survey only takes into account thosedelivered through website. That means that electronic service provision through otherchannels such as Digital TV or kiosks are neglected. Finally, the study does not providea selection of good practices.

4.1.6 Summarizing the evolution of electronic service benchmarkingIn this section we have been reviewing and discussing the main benchmarks relatedwith public service and local government. The European most well know benchmark ofelectronic services, the “Online Availability of Public Services” benchmark, has beendeeply analysed and discussed by many authors (Janssen, 2003; Bannister, 2007; Heeks,2006). We would like to highlight here that this benchmark only focus on serviceprovision and provides no data regarding adoption. In addition, it is performed at statelevel, therefore provides no data at local level. An interesting benchmarking methodology has been found in “The Municipal E-Government Assessment Project” performed by Kaylor et al. in 2001. This benchmarkis the most interesting since the methodology is based on a bottom-up approach,focused only in cities and what cities are providing in terms of electronic services.Nevertheless, it neither measure adoption nor provide a best practice selection.“E-Government at the American Grass-roots” is a secondary study based on twopreviously performed surveys. It analyse data to discover web adoption patrons in localpublic administration. It is not focused on electronic services. Therefore, it is not9 In the first study performed by Holzer benchmarking New Jersey municipalities (Holzer, 2003), he

uses a sample of cities formed by the most populated cities of the state. He supports his criterion withexisting literature that suggests a positive relationship between population and e-government capacityat the local level. However, from the results obtained in this first study it can be derived that thiscorrelation does not exist (see table 4). Despite this evidence, Holzer used the same argument tojustify a quite arbitrary sample of cities made by the most populated cities on each of the mostadvanced countries in terms of e-readiness.

46

Page 47: Local e-Government Bench-learning

Local e-Government Bench-learning

relevant in our research. The “Evolution of Local Government E-Services” performed by Shackleton, Fisher, andDawson in 2004 in Australia is again an interesting exercise since it is based on theservices cities are offering to citizens. Nevertheless, here again it lacks of adoptionmeasurement. Finally, the “Digital Governance in Municipalities Worldwide” designed by Holzer andKim in 2003 is probably the most important for capital cities since it has became thestandard methodology of the UN benchmarking of world capital cities performed in aby-annual basis although it is focused on official websites.From this revision it can be seen that no benchmarks can be found focusing both locallevel and measuring, at the same time, electronic service provision and adoption.

4.2 Measuring electronic services adoptionThe aim of this section is to analyse through a literature review the state of the art in e-government services adoption and highlight the importance of measuring adoption inorder to have a complete view of the e-government service provision. In spite of that,the section also show the current existing difficulties to perform adoption assessments.

4.2.1 Adoption is not demand Adoption is not demand. Warkentin, Gefen, Pavlou, and Rose (2002) describe adoptionas the intention of citizens to engage in e-government to receive information and requestservices from the government. Also Carter and Belanger (2005) define it as intent touse, and Gilbert, Balestrini, and Littleboy (2004) establish it as the willingness to use e-government services. The three definitions are pointing at the same concept, theintention or will to use electronic services. But intention is not real use. Although some authors have found that intention-to-use ane-government service is a strong predictor of actual usage (Belanger & Carter, 2008)there is a clear difference between intention to use and real usage of a service. Bothauthors redefine adoption as the real usage of e-government services and introduce anew concept, which is demand as the willingness of citizens to use e-governmentservices.

4.2.2 The importance of measuring adoptionThe service adoption measurement is important as it has been stated by differentauthors. According to Kumar, Mukerji, Butt, and Persaud, (2007), the main reason isthat the ultimate objective of e-government programs ought to be the frequent andrecurring use of online services by citizens. Heeks (2006) states that matching e-

47

Page 48: Local e-Government Bench-learning

Local e-Government Bench-learning

government supply to demand is of paramount importance to manage change in e-government programs. Moreover, when delivering services through different channels,which is the general trend, measuring adoption and comparing adoption of electronicservice delivery with other channels usage is an important instrument to manage thedelivery channel shift towards cheaper channels (Janssen & Wagenaar, 2004). In thethree cases measure adoption is identified as key to successful manage the e-government program. But this importance is not reflected into the existing benchmarks. As we saw in earlychapters, several authors have found that the majority of performed e-governmentbenchmarking focuses on supply-side indicators (# of services available online) whilean increasing, but still limited number, considers the demand-side (i.e. take-up andsatisfaction with services) (Janssen, 2003; Kunstelj & Vintar, 2004; Janssen et al., 2004;Heeks, 2006; Codagnone et al., 2006; Panopoulou et al., 2008). Nowadays, UnitedNations (2010) recognize that most of the statistics are derived from supply sideindicators and often by website assessments alone. Little information is yet available onthe demand side of e-government. Few surveys exist that would indicate ‘how’ citizensuse these services and ‘what’ they see as maximizing public value. (UN, 2010). The real objective of e-government services is misunderstood. This abundance ofsupply-side survey unbalance with a scarcity of demand-side studies leads to amisinterpretation of what is e-government service progress. The existing benchmarksare pushing countries to prioritize getting good grades for making lots of applicationswithout paying attention whether are actually used by citizens or not. (Janssen et al.,2004). This situation points out at an existing gap between the demand for and thesupply of e-government performance information (Panopoulou et al., 2008) that needsto be covered.

4.2.3 Measuring adoptionThere is no shared framework for adoption assessment. As the main studies about e-administration or service to citizens are only focusing in measure the “supply” of e-government services and say nothing about the actual use or take-up of online services,which is the “demand” side (Janssen, 2003), no framework for adoption measurementhas been discussed into the research community.The research on adoption is mainly focused on the overall adoption process descriptionor more specifically focussed on barriers and enablers. It can be found few studies aboutthe technology adoption life-cycle models and the diffusion theory for the adoption oracceptance of a new product or innovation, according to the demographic andpsychological characteristics. A more interesting group of academic work pays attentionto actual use of websites, portals, e-services, information content and other elements ofsupply focusing in the level of interest in using and reasons for not using these elementsas well as evaluations of the quality of them as perceived by the users and evaluation oftheir perceptions, requirements and needs (Kunstelj & Vintar, 2004; Kumar et al., 2007;

48

Page 49: Local e-Government Bench-learning

Local e-Government Bench-learning

Alsuwaidi, 2009).How to measure adoption remains unattended. Only few studies face the adoptionmeasurement of e-Service due to inherent difficulties (see chapter 2). It only can befound some clues in literature about web access analytic tools. There is a clear lack of asuitable framework for adoption assessment. Therefore, how to measure adoptionremains unattended in terms of research and Löfstedt recognize (2005) that there shouldbe more research into this factor, and more elaborate models and methods should bedeveloped to understand citizens adoption of e-services. Is United Nations opinion thatwhat’s needed is international consensus about how to assess e-governmentperformance (UN, 2010).

4.2.4 e-Government services targetIn a measurement exercise a reference, a target, should be established before performingany measurement. In the case of adoption of an electronic service, the reference shouldbe the group of citizens and companies identified as potential users of the service. Toknow which are the target users of e-government services is probably one of the mostdifficult questions to solve since many factors are influencing the real target of e-government services.Each service has its specific target. According to Janssen and van Veenstra (2005), whenevaluating e-government success, the first challenge is the investigation of variousperspectives which may not only include addressing and meeting the general needs of atarget group such as citizens, but also require the inclusion of specific needs of targetgroups of citizens that are using a particular e-government service. Although e-government services is often seen as a unit, actually it is composed by a multiplicity ofdifferent services. In general, each service is targeted to specific groups of citizens suchas the unemployed, families, pensioners, architects, lawyers, students, etc. (Alshawy etal., 2007). Therefore, a case by case analysis of the target should be performed. Social, cultural and technical context of use and age modify the real target. In thisanalysis social, cultural and technical context of use are also important factors to takeinto consideration in e-government evaluation (Alshawy et al., 2007) (Patel & Jacobson,2008) since they impact in the real target of users. Moreover, it is recognized theimportance of including the variable of age on the ability to use and subsequently theintention to adopt new technological advances (Gilbert et al. 2004).Therefore, the definition of the e-government service target of users is a complex issue.Several factors must be taken into account and require a case by case approach andneeds further research.

4.2.5 How often citizens consume e-government services?Frequency of use is a factor to determine the adoption threshold. How often citizens

49

Page 50: Local e-Government Bench-learning

Local e-Government Bench-learning

consume e-government services is an important questions that must be know andcomplement the existence of a well known target before measuring e-governmentservices adoption. Different studies determined that citizen contact with government is relatively rare.Surveys performed in the US shows that only half of respondents had contacted anylevel of government in the previous year. Two thirds of those, contacted any level ofgovernment less than every few months (Horrigan, 2005). That is not an estrange phenomena. Bannister (2007) notice that most transactions withgovernment (at least for the individual citizen) occur once a year (tax, car and televisionlicences), many documents are renewed even less frequently (passport, drivinglicence ...) and some documents, such as death or birth (and in some cases marriage)certificates, are once in a lifetime events. This low frequency of use will also apply to e-government services defining a maximum level for electronic services adoption whichshould be lower than total number of contacts.

4.2.6 Variables to measure adoptionSeveral variables are needed tot measure adoption. Although a short number ofvariables should be desired actually there are different views and proposals. In Janssenet al. (2004) study of the current benchmarks, the authors identify a set of variables thatare used to measure usage or adoption of electronic services:

• Number of individuals that have made use of electronic services offered• Number of businesses that have made use of electronic services offered• Percentage of citizens that has visited government websites to search for

information• Number of businesses that have made payments online• Percentage of internet traffic that pertains to electronic service delivery

In a similar survey performed by Codagnone et al. (2006), the author propose a set ofweb metrics to measure electronic services adoption:

• Number of hits or user contact sessions• Number of document downloads• Amount of time users spend on a site• Number of transactions completed• Web analytic (click streams, repeat use, cross-usage)

Different sophistications levels needs of different variables. The measurement ofelectronic service adoption is tightly related with the sophistication stage of the service.Therefore, in general the variable or indicator will be different according to the maturityor sophistication level of the services. When analysing official websites, visits or

50

Page 51: Local e-Government Bench-learning

Local e-Government Bench-learning

number of pages served are both indicators commonly accepted to assess the popularityof a website. But adoption or take up of e-services is not only access. These variablescan be suitable to measure adoption of those services being at stage 1 in the Baum andDi Maio e-government model (see chapter 2).When a service reaches stage 2 in Baum and Di Maio model, although number of pagescan partially explain how adoption is, the number of forms downloaded is a morerealistic indicator. Similarly happens in stage 3 where the key in that case, is datatransactions performed trough e-services applications. However, data transaction is notdirectly applicable in higher levels of sophistication where integration tends to reducethe number of transactions performed by citizens when using electronic services. Therefore, no indicator can be used in all range of electronic services due to theirdifferent level of maturity and the heterogeneity of IT solutions used to provide them.As we see, each e-service needs to define which is the best magnitude to measureadoption. To arrive at a nuanced view of usage, we should combine indicators forinformation seeking, information provision, and transactions (Janssen et al., 2004). Hereagain a case by case analysis should be performed thus adding a level more ofcomplexity in the adoption measurement.

4.2.7 Measurement methodologiesA variety of measurement methodologies are used to assess e-services adoption. A factthat is linked with the diversity of indicators we saw in previous section.Surveys are the preferred source of adoption information. Heeks (2006) found that ingeneral, use measurement is mainly performed by surveys. The lack of a singleindicator to measure real use of electronic services, the complex process of measuringthe real use of a palette of different services in different stages makes practical the useof surveys in which citizens are asked about their use of e-government. Many studiesare using this methodology. The same author also noticed that self-assessment is one of the most commonly useddespite its drawbacks. Is the opinion of Heeks that internal self-assessment works wellfor some things, such as reporting of lessons learned. It works less well for others wherethere can be a "public relations bias": the respondent is aware that their response will bepublicly reported and will thus produce a good or bad reflection. Here again the samereasons would move to perform adoption assessment using this technique, in additionthis methodology is the cheapest one. Codagnone et al. (2006) point out to the emerging use of web metrics/crawlers as aneven richer source to assess service use in certain situations if they can be objectivelyreported. This includes not merely usage indicators such as number of page hits orcompleted transactions but also the introduction of the e-commerce concept ofconversion rate to understand how effective are the e-government solutions whenevaluating specific e-government solutions.

51

Page 52: Local e-Government Bench-learning

Local e-Government Bench-learning

4.2.8 Summarizing adoptionThe inherent difficulties to define the e-government service target of users. The lack of aconsensus in a measurement framework, the variety of indicators and the existence ofdifferent practices to measure adoption undermine cross-country benchmarking.

4.3 Summary of the electronic service benchmarkingIn this chapter we have been revisiting literature to analyse the main benchmarks relatedwith public service and local government and the difficulties around the public adoptionmeasurement of public service. Although some benchmarks tailored at local government have been found, none of themmeasures both provision and adoption of electronic services at the same time.One of the reasons for the scarcity of local level surveys is that the differences existingin the role of cities across different countries undermine extensive comparisons inpublic services at local level. A second reason is the lack of an adoption measurementframework and the difficulties to stablish standard indicators for adoption, as it has beendiscussed in section 4.2. In such a context, some critical decisions must be taken to proceed with such exercises.The first is to reduce the geographical scope to European Union, which is an area withcultural similarities and under a convergent legal framework. A second pragmaticdecision is the measurement of the perceived adoption instead of measuring the realadoption. Both decision have been driving the Local e-Government Bench-learning.

52

Page 53: Local e-Government Bench-learning

Local e-Government Bench-learning

5 Local e-Government Bench-learning improvementsThe aim of this chapter is analyse the local e-government methodology while applyingthe knowledge brought by the previous chapters. An analysis of the strength andweaknesses of the study is performed. And finally, some improvements to the Local e-Government Bench-learning methodology are proposed.

5.1 The Local e-Government Bench-learning most interestinginnovationsThe Local e-Government Bench-learning has specific characteristics that makes it apioneering experience in its field. The literature review confirmed that there is a lack ofbenchmarking studies confronting provision versus adoption of e-services; and that noinformation is available comparing development of local e-government in Europe either.The literature review revealed that the Local e-Government Bench-learning is one of thefew surveys analysing and confronting both provision and public adoption of electronicservices, and the first European wide benchmarking study focused on electronic servicesprovision at local level.Moreover, it has been found that some authors (Janssen et al., 2004; Heeks, 2006;Bannister, 2007) are critics with benchmarking practices considering them not usefulenough because often the purpose is not clear then are not well designed and tend tocompare things not really comparable in addition they not provide the neededinformation for practitioners. The authentic aim of a benchmarking study is to learnfrom those who are really doing better. Therefore, these studies should clearly provide abest practices identification. However, actually they are only offering a best players(best scoring) identification sometimes complemented with some relevant practicesdescription. In consequence, these studies do not offer the kind of information as thekey to e-government success.Local e-Government Bench-learning intents to go far beyond academic orientedsurveys, offering more useful information to managers that it is usually provided,allowing them to learn from those cities which are really doing better as the key to e-government success (Janssen et al., 2004).The Bench-learning name comes from combining the words benchmarking andlearning. The aim of the survey is to go further than measuring and comparing Europeanwide electronic services provision in cities: it should provide more useful informationand in better ways for city managers to learn and improve e-government developmentstrategies. The typical benchmark provides as a main result a ranking of cities going from best toworst, while our bench-learning provides first of all, a set of charts showing theEuropean averages and comparing each city with the European average. Therefore it

53

Page 54: Local e-Government Bench-learning

Local e-Government Bench-learning

helps practitioners to easily identify areas in which the city is performing better orworst.As a second output it provides a list of best practices that are obtained as a consequenceof the measurement and comparison performed. This list is organized of services. In thisstudy, it is not necessary to register a high maturity mark to be a best practice, what isimportant is how is the maturity of a service regarding the average maturity of thecategory. In addition, for a service to be a best practice, not only it has to be far from theaverage maturity, but also, it must be characterized by a good acceptance, therefore,rating higher than 3 in adoption. The current methods used to identify best practices arequite arbitrary and subjective. Here, in this study the innovation is that the best practiceselection is a consequence of the measurement framework and over all methodologydefined to compare different services.Apart from the aforementioned innovations, other methodological innovations aim toadapt better to the specificities of both: the local government characteristics and theconfrontation of services provision versus adoption. In that sense, the first innovation introduced has been the definition of a servicescatalogue shared by all surveyed cities which is itself service-oriented and fits with theEuropean local administrations role. This catalogue was made through a bottom-upapproach and showed at the same time the diversity of innovation carried out by citiesin electronic services provision. It has been developed following a common agreementamong the sample cities and integrating all the common and outstanding services thatthey offer. Amongst all the e-services gathered, a common set of services has beenidentified and used a a basis to perform comparisons. In addition, the catalogue has beenused to show the diversity of innovation carried out by cities in electronic servicesprovision. These are the additional services which represent the 47% of the identifiedservices, and represent the wide variety of services considered by European citiesshowing the innovation leaded by cities in the field of services provision.A new maturity stages model designed to measure local level services sophistication andcurrent trends on services provision has been introduced. As it have already beencommented in chapter 2, existing e-government stage models are not reflecting newfeatures in service provision. In our study we took the four maturity levels proposed byBaum and Di Maio (2000) as the starting point, extending it to include a fifth level ontop of them. This level reflects a more advanced degree reflecting new features in user'sinteraction taking into account the use of emerging web 2.0 technologies.Although it not solve the lack of consensus in how to measure services adoption, thebench-learning propose an innovative approach to tackle the gap on adoptioninformation. Therefore it allows to confront provision and adoption at the same time.

54

Page 55: Local e-Government Bench-learning

Local e-Government Bench-learning

5.2 Some important limitations revealed by the studyUnfortunately, some of the innovations were difficult to introduce. The servicecatalogue has been made to stablish a set of common services in which performmeasurements and comparisons. The consensus have been reached with and importantnumber of services because only European cities have been analysed. This set ofservices must be considered an European set of common services and it can not beapplied to perform more extensive comparisons including different continents becauseexisting differences in the role cities play undermine possible comparisons. In the caseof willing to extend the methodology to a wider geographical area a new set of servicesshould be build.We had to take a pragmatic approach in adoption measurement choosing internal self-assessment to measure it, therefore, moving from measuring adoption to actuallymeasuring perception of adoption. A measurement framework was defined with theinvolvement of all participant cities, but as each service responsible applied it, anhomogeneous criterion cannot be assumed. Improving the adoption measurementframework should be a priority in future research.Regarding maturity assessment, even though providing common instructions, it is notpossible to assure a complete homogeneity among city’s results, along with the risk ofmisunderstanding maturity level 5 due to its novelty. Therefore, a general refinement atthe measurement process is needed to reduce data subjectivity. In addition, future works should include a larger sample of cities in order to increasestatistical reliability. Having a larger sample would allow a segmented analysis distillingreference types for comparison according to city characteristics such as size or GDP percapita.

5.3 ImprovementsSome of the future improvements can also be derived from the literature review. In thefollowing pages I discuss the main improvements in the methodology that should followa second edition. Of the bench-learning study.

5.3.1 Hypothesis formulation An explicit formulation of the hypotheses would clarify not only the purpose of thestudy but also the outputs. The first edition of the study lacks it. As a first step, now wemake the following hypotheses more explicit:

• H1: European cities are sharing an important corpus of common services. Thiscorpus is the basis to perform sound comparison at European level. Theimportance of this hypothesis is clear, should it be not true benchmarking cannot be performed. To verify this hypothesis it should be investigate service

55

Page 56: Local e-Government Bench-learning

Local e-Government Bench-learning

coverage with cluster analysis to see whether a cluster of services with highcoverage can be identified.

• H2: The most mature or sophisticated services are also the most widespread. Iseems at first thought that the more mature or sophisticated is a service the moreused by citizens it will be. For practitioners, this is an important hypothesis toverify or reject since usually the more sophisticated is a service, the moreexpensive is the development. To verify this hypothesis we have to investigatewhether exist a significant correlation among sophistication and adoption.

• H4: Cities are developing special sophisticated services to attract citizensattention. There is a trend for cities to create especial and sophisticated servicesto attract citizens interest and rise citizens awareness towards new electronicchannels thus pushing for a channel shift in service provision. To verify thishypothesis we should investigate whether exist a significant increase ofsophistication in additional services in comparison with standard services.

In addition, and after testing the methodology in a first survey and having analysed theresults, the methodology proved to be useful to identify whether European policy in e-government impacts in local government cohesion. Also, it helps to discover whethersmall cities are performing better than big cities. Although it is widely recognized theagility of small cities in services innovation, there is no clear basis to such statement. Toverify or reject it, city maps should be analysed and compared. However, it has to be said that through the literature review performed in 2.2 and 4.1 nobenchmark with explicit hypothesis formulation has been found.

5.3.2 Statistical analysisAs the study was exploratory, it did not go beyond using simple average calculations.However, even at this pilot stage, there are already more than 1.200 entries, the use ofstatistical tools would have been justified. Let us to mention some initial ideas: clusteranalysis is fundamental to identify groups and classify elements into these groups. Itmight allow us to identify and better justify the groups and categories of standard andadditional services. Here also analysis of variance would have helped to assesdifferences among different groups of services. When comparing provision and adoption of public services, computed correlations ofsophistication and adoption levels for standard and advanced services could be used toreject or confirm hypothesis. When introducing a score (see 5.3.8) then both correlation and cluster analysis can beapplied for cities results themselves, provided the sample is larger and segmented as it isdiscussed in next section. Also here cluster analysis and analysis of variance would helpto assess the existence of different groups of cities in terms of how they are developinge-government projects.

56

Page 57: Local e-Government Bench-learning

Local e-Government Bench-learning

5.3.3 The sampleTo lack of a good sample of cities can undermine the applicability of the resultsobtained (see 4.1.5). In our survey (see table 2), a bigger sample of cities is needed tohave statistical significance, and hence to offer a more rigorous knowledge of theEuropean results. In particular there would be better grounded average values, which isthe basis for cities to be compared with. With a larger sample of cities, maintaining the representativeness of the European localpublic administrations as it has been discussed in 3.3, it makes sense to segment thesample with cities by similar characteristics, such as GDP, social aspects, citizens e-readiness, etc. Performing a more detailed analysis within the segments themselves, andalso comparing one of each other, the benchmark would provide a richer view of theEuropean local e-government reality.

5.3.4 The services catalogueThe global services catalogue has been developed following as a bottom-up exercise.Compiling and analysing all the services provided by the sample cities, we reached acommon agreement and integrating all the common and outstanding services that theyoffer. The services catalogue is the basis to perform a sound comparison amongdifferent cities of different countries. Without this basis there is no room forcomparisons, that is why the most important benchmarks proceeded defining, first ofall, this basis of common services (see section 4.1). A part from the basic services otheradditional services have been found (see 3.6). Although this basis must remain stable toprovide a clear view of the evolution along different survey editions, a periodic revisionof those services and its coverage should be done in order to detect when an additionalservice is becoming a standard service. In the catalogue, the services have been grouped in nine thematic categories. CapGemini and Kaylor et al. proceeded in a similar way because it allow to better analysee-government development, not only in specific cites, but also globally detecting areasof raising interest as it has been discussed in chapter 3. Nevertheless, the categoriesdefinition can be improved to better reflect thematic areas of local governmentresponsibilities, and solve existing ambiguity and overlapping. In addition, a newdefinition of services category aligned with the clusters of services defined in the CapGemini benchmark could provide an extra basis to compare development of similarthematic areas (although with different services) across local and state level.Here, twofold complementary improvements are proposed. First of all, an analysis ofthe competencies or responsibilities of cities across Europe should be performed toidentify the different main areas that should map with the defined categories. A second improvement tackles service cataloguing to solve the existing ambiguity andoverlapping among categories. Semantic web technologies could be the tool to thatpurpose. According to Klischewski (2003), the use of Semantic web in e-government is

57

Page 58: Local e-Government Bench-learning

Local e-Government Bench-learning

especially challenging due to “differences of interpretation of e.g. law, regulations,citizen services, administrative processes, best-practices, and,not the least, manydifferent languages to be respected within and across regions”. Service category is only one of the possible e-service cataloguing mechanisms.According to Vassilakis and Lepouras (2006), the definition and development of apublic services ontology is the starting point to achieve a well stablish service category.A category definition based on semantic mapping would help to deal withconceptualization and contextualization of e-services increasing the clarity of thecategory definition and avoiding overlapping and miss classification. Here, as thesurvey is being developed in a wide area rich in cultural, historical and linguisticdiversity, semantic mapping can be used as a tool to discover the relationships betweenvocabulary words. This tool will facilitate the constructions and population of the publicservice ontology. The construction of a public service ontology that has already beenperformed between UK, Greece and Spain in the context of Ontogov project(Tambouris, Gorilas, Kavadias, Apostolou, Abecker, Stojanovic, & Mentzas, 2004;Apostolou, Stojanovic, Pariente-Lobo, Batlle, & Papadakis, 2005), could be a startingpoint to extended it covering all EU countries.

5.3.5 Collection of the informationInformation gathered from the sample cities has been based on self-assessment. Theparticipant cities have assessed themselves regarding maturity and adoption scoresthrough the use of several templates, and, a set of instructions were developed to fullyexplain which stages should be considered. In order to follow a common understanding,however, 15 evaluators with possibly 15 different criteria have worked on each citytemplates, hence making not possible to assure good homogeneity of results. Animproved strategy has to be envisaged. In the literature review (see 4.2.7) we learned that are different possibilities to collectdata. Self-assessment is the cheaper one but it can introduce a bias in reported data. Themost prestigious benchmarks (see 4.1.1, 4.1.2 and 4.1.5) proceeded using a team ofindependent experts that evaluated websites and electronic services offered. Thisguarantee a more systematic analysis although, increase notably the cost of thebenchmark.

5.3.6 Service maturity levelsThis study aimed at measuring e-government service provision and adoption inEuropean local public administration. As it was already mentioned in Batlle and López(2007) local administrations are introducing web 2.0 technologies into e-servicesprovision. This emerging trend is a plus in e-services sophistication that comes after theinternal processes transformation, and there are no e-government stages model that takeit into account. That is why in the Local e-Government Bench-learning we decided to

58

Page 59: Local e-Government Bench-learning

Local e-Government Bench-learning

extend the e-government four phases proposed by Gartner Group (Baum & Di Maio,2000) with a new fifth level, hence adding a not yet widely accepted stage to a well-known methodology. When analysing the results in depth, we observe that there are a lot services that havereached maturity level 5 in a wide variety of categories, according to the self-assessment. As web 2.0 technologies in e-government is a novelty, it is possible that thesurveyed cities have misunderstood the correct use of this new level. A second problem to be solved in the model is the raising of integrated services,personalized and proactive which leads to a number of services reduction is not wellreflected in the proposed model.We need to find a suitable model to reflect both these new characteristics and the use ofweb 2.0 technology. In Europe, the e-government maturity model most used in publicservice assessment is the model proposed by Cap Gemini (4.1.1). On the other hand,Kaylor et al. benchmark, the other interesting benchmark that worth to be taken intoconsideration, use a primitive e-government stages model as it has been alreadydiscussed in 4.1.2. In our opinion the bench-learning survey can be improved using anew stages model defined by extending the Cap Gemini model to include a new level ofsophistication to reflect the emerging use of web 2.0 technologies in e-servicesprovision.

5.3.7 Measurement of adoptionStatistics of adoption are not widely used among European cities, not even in thesurveyed cities. Generally, cities only have real take-up data about websites’ visits,telephone calls or face-to-face attention. This situation is due to a lack of appreciation ofthe benefits of measuring adoption but also to a lack of awareness of which services,interactions or channels to measure (see chapter 4). This is an important gap to beaddressed.Let us stress that the perception of adoption measured in the study is, and must be, verydifferent from one country to another, as anyway the perception of adoption is relativeto the local situation. An example can illustrate this: Murcia city council wouldprobably rate as a success achieving a 40% of adults’ courses subscription throughInternet, while in the more Internet developed Stockholm this might be rated as afailure.

5.3.8 ScoringWhat usually benchmarks provide is a ranking based on a score. A score or a ranking ofcities according its position give little information to practitioners. That is why we donot provide such score in our first edition.

59

Page 60: Local e-Government Bench-learning

Local e-Government Bench-learning

Nevertheless, Kaylor et al. (see 4.1.2) proved that the introduction of a general score aswell as a score per category, in our case should it be both for service/categorysophistication and adoption, would allow better discussion of the results in some way asit is performed in Kaylor et al. benchmark. Although benchmarks that have a strong focus on the scoring system are only useful tomeasure general and specific evolution from one edition to another (see 4.1.5), so wecan improve our survey having a scoring system to settle a basis to assess progress infuture survey editions, when adding a temporal dimension. The Local e-Government Bench-learning survey aims at setting a path for furtherexercises of this sort and, hence, generating a stable temporal series in the near future.

5.3.9 ChartsThe introduction of spider charts (also known as radar charts) instead of bar charts (seeillustration 3) would provide a more clear view of the general position of a citycompared with the European average. A spider chart is typically created to graphicallyshow the size of the gaps among three to eight organizational performance areas. Thechart would have a radii for each category. The score of each category is plotted as adata point in the corresponding radii, a line connects all the data points. A spider chartcan show the city performance and the European average at the same times allowingquick comparisons making visible strengths and weaknesses of the city. When plottingprovision and adoption per category, it easily can be derived whether a significantcorrelation exist among these two variables. Spider charts are being used in severaleditions of the Cap Gemini survey (see 4.1.1) becoming a de facto standard and creatinga culture of results presentation that is worth to be imitated. Beyond improving visualization, the new results provided by using more sophisticatedtools and by the segmented analysis of a larger and more significant sample will providebetter insights of the data.

60

Page 61: Local e-Government Bench-learning

Local e-Government Bench-learning

6 Final remarksTo fill the gap in local e-government benchmarking the author of this research reportdesigned and launched a pioneering field study of several key European cities entitledLocal e-Government Bench-learning in 2008. This study tried to introduce severalinnovations in the services catalogue, adapted to local administrations, with a widercoverage focus based on a new maturity stages model, on the quality and servicesadoption, with a classification identifying standard services as well as diversity. On theother hand,the study adopted an innovative city charts visualisation to offer decisionmakers material suitable to inform future decisions on e-government developmentstrategies. In this research report I reviewed in depth the applicability and correctness of the Locale-Government Bench-learning study, comparing it with the state-of-the-art in e-government benchmarking -at different administration levels-, reviewing its mainfindings and contributions, especially, the methodological ones at the localadministration level and proposing further improvements. Literature review shows that the methodology of this bench-learning survey isinnovative, well designed and tailored to local government. The first survey performedprovided feasible results that are presented in a useful format for practitioners, thusproving the validity of the methodology. The survey covers the initial objectives of providing: a general view of the Europeanlocal e-government development, detailed information for each of the sample city, and aextensive list of best practices. It provides answers to policy makers and city managersmain questions when managing the e-government process, to shorten it and ensure asuccessful end. And allow take profit of the experience of previous innovators as aguide to make informed decisions. At the same time it provides enough information toconfirm or disprove the implicit hypothesis (which are explicitly formulated in 5.3.1.).Therefore, the initial objectives have already been reached in the first edition.Nevertheless, due to its novelty, this study show some weaknesses that should becorrected in subsequent editions of the study. The new edition of the study Local e-Government Bench-learning is intended to consolidate a methodology to benchmark theprovision of electronic public services and adoption of same by the citizens in the areaof local e-government in Europe. While the methodology is to encourage theidentification of best practices that are prone to be used as accelerator in the process ofimplementing electronic services. The desirability of repeating the study is not justified only by the consolidation of themethodology and the increasing stringency of the measures but also for providing aperspective, a vision that allows more evolutionary view the efforts of European localauthorities to improve services offered to its citizens.

61

Page 62: Local e-Government Bench-learning

Local e-Government Bench-learning

62

Page 63: Local e-Government Bench-learning

Local e-Government Bench-learning

7 ReferencesAl-Hashmi, A., & Darem, A. B. (2008). Understanding phases of E-government

project.Emerging Technologies in E-Government.Alshawy, S., Alahmary, A., & Alalwany, H. (2007). eGovernment evaluation factor.citizens perspective. Paper presented at the European and Mediterranean

Conference on Information Systems 2007, Spain. Alsuwaidi, H. (2009). Evaluating the factors affecting for the adoption of E-governmentservices in the public sector (UAE).Andersen, K. V., & Henriksen, H. Z. (2006). E-government maturity models: Extension

of the layne and lee model. Government Information Quarterly, 23(2), 236. Apostolou, D., Stojanovic, L., Pariente-Lobo, T., Batlle, J., & Papadakis, A. E. (2005).

Configuring e-government services using ontologies. UPGRADE. the EuropeanJournal for the Informatics Professional, 6(6).

Arslan, A. (2007). Turkish local e-governments: A longitudinal study. The ElectronicJournal of e-Government, 5(2), 95-106.

Bannister, F. (2007). The curse of the benchmark: An assessment of the validity andvalue of e-government comparisons. International Review of AdministrativeSciences, 73(2), 171.

Batlle, J., Calderón, R., & López, J. (2009). Local e-government bench-learning survey.Brussels: EUROCITIES. Batlle, J., & López, J. (2007). Key success factors for the adoption of eServices.

Brussels: EUROCITIES. Baum, C., & Di Maio, A. (2000). Gartner’s four phases of e-government model.

Stamford, Ct., Gartner Group, 21, 12-6113. Belanger, F., & Carter, L. (2008). Trust and risk in e-government adoption. Journal of

Strategic Information Systems. Cap Gemini. (2002). Online availability of public services: How does europe progress.

Brussels: European Commission. Cap Gemini. (2003). Online availability of public services: How is europe progressing.Brussels: European Commission. Cap Gemini. (2004). Online availability of public services: How is europe progressing.

Brussels: European Commission. Cap Gemini. (2005). Online availability of public services: How is europe progressing.Brussels: European Commission.

63

Page 64: Local e-Government Bench-learning

Local e-Government Bench-learning

Cap Gemini. (2006). Online availability of public services: How is europeprogressing?. report of the 6th measurement. Brussels: European Commission.

Cap Gemini. (2007). Benchmarking the supply of online public services. 7thmeasurement. Brussels: European Commission.

Cap Gemini. (2009). Smarter, faster, better eGovernment. 8th eGovernment benchmarkmeasurement. Brussels: European Commission.

Carter, L., & Bélanger, F. (2005). The utilization of e-government services: Citizentrust, innovation and acceptance factors. Information Systems Journal, 15(1), 5-25. Codagnone, C., Caldarelli, L., Cilli, V., Galasso, G., & Zanchi, F. (2006). Compendium

to the measurement framework. eGovernment economics project. Rome: eGEPConsortium.

Cole, M. I., & Parston, G. (2006). Unlocking public value: A new model for achievinghigh performance in public service organizations Wiley.

Cullen, R., & Houghton, J. (2000). Democracy online: An assessment of new zealandgovernment web sites. Government Information Quarterly, 17(3), 243-67.

European Commission. (2000). eEurope 2002 an information society for all. actionplan. Brussels: Commission of the European Communities.

European Commission. (2002-1). eEurope 2005: An information society for al. actionplan No. COM(2002) 263). Brussels: Commission of the European Communities.

European Commission. (2002-2). eEurope 2005: Benchmarking indicators No.COM(2002) 655). Brussels: Commission of the European Communities. European Commission. (2006). i2010 eGovernment action plan: Accelerating

eGovernment in europe for the benefit of all No. COM(2006)173). Brussels:Commission of the European Communities.

European Commission. (2007). eGovernment progress in EU27+: Reaping the benefits.Brussels: Commission of the European Communities. Flak, L. S., Olsen, D. H., & Wolcott, P. (2005). Local E-government in norway.

Scandinavian Journal of Information Systems, 17(2), 41-84. Gilbert, D., Balestrini, P., & Littleboy, D. (2004). Barriers and benefits in the adoptionof e-government. International Journal of Public Sector Management, 17(4), 286-

301. Graafland-Essers, I., & Ettedgui, E. (2003). Benchmarking e-government in europe and

the US. statistical indicators benchmarking the information society project.Cambridge (UK): RAND Europe.

Griffin, D., Foster, A., & Halpin, E. (2004). Joined-up e-government: An exploratory

64

Page 65: Local e-Government Bench-learning

Local e-Government Bench-learning

study of UK local government progress. Journal of Information Science andTechnology, 2.

Gupta, M. P., & Jana, D. (2003). E-government evaluation: A framework and case study.Government Information Quarterly, 20(4), 365.

Hachigian, N. (2002). Roadmap for E-government in the developing world: 10 questionE-government leaders should ask themselves. Los Angeles, CA: Pacific Council onInternational Policy.

Heeks, R. (2006). Understanding and measuring egovernment: Internationalbenchmarking studies. Paper presented at the UNDESA Workshop E-Participationand E-Government: Understanding the Present and Creating the Future,Budapest, 27-28.

Hiller, J. S., & Belanger, F. (2001). Privacy strategies for electronic government. E-Government, , 162-198.

Holzer, M. (2003). A comparative e-government analysis of new Jersey’s 10 largestmunicipalities.Holzer, M., & Kim, S. (2003). Digital governance in municipalities worldwide. an

assessment of municipal web sites throughout the world. Newark (New Jersey):The E-Governance Institute. National Center for Public Productivity. RutgersUniversity.

Holzer, M., & Kim, S. T. (2005). Digital governance in municipalities worldwide(2005). A longitudinal assessment of municipal web sites throughout the world.Newark (New Jersey): The E-Governance Institute. National Center for PublicProductivity. Rutgers University.

Holzer, M., & Kim, S. T. (2007). Digital governance in municipalities worldwide(2007). A longitudinal assessment of municipal websites throughout the world.Newark (New Jersey): The E-Governance Institute. National Center for PublicPerformance. Rutgers University.

Horrigan, J. B. (2005). "On demand" citizens: E-government at high speed. Paperpresented at the FedWeb Conference at George Mason University. Janssen, D. (2003). Mine’s bigger than yours: Assessing international e-governmentbenchmarking. Paper presented at the 3rd European Conference on e-Government,

MCIL, Reading, MA, 209-218. Janssen, D., Rotthier, S., & Snijkers, K. (2004). If you measure it they will score: Anassessment of international eGovernment benchmarking. Information Polity, 9(3),

121-130. Janssen, M., & van Veenstra, A. (2005). Stages of growth in e-government: Anarchitectural approach.

65

Page 66: Local e-Government Bench-learning

Local e-Government Bench-learning

Janssen, M., & Wagenaar, R. (2004). An analysis of a shared services centre in e-government. Paper presented at the 37th Annual Hawaii International Conferenceon System Sciences, Hawaii.

Jones, S., Irani, Z., & Sharif, A. (2007). E-government evaluation: Reflections on threeorganisational case studies. Paper presented at the 40th Annual HawaiiInternational Conference on System Sciences, Hawaii. 105.

Jones, S., Irani, Z., Sharif, A., & Themistocleous, M. (2006). E-government evaluation:Reflections on two organisational studies. Paper presented at the 39th HawaiiInternational Conference on System Sciences, Hawaii.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2005). Governance matters IV: Governanceindicators for 1996-2004. Washington: Wold Bank Institute.

Kaylor, C., Deshazo, R., & Van Eck, D. (2001). Gauging e-government: A report onimplementing services among american cities. Government Information Quarterly,18(4), 293-307.

Klischewski, R. (2003). Semantic web for e-government. Electronic Government, ,1059-1060.

Kumar, V., Mukerji, B., Butt, I., & Persaud, A. (2007). Factors for successful e-government adoption: A conceptual framework. The Electronic Journal of e-Government, 5(1), 63–76.

Kunstelj, M., & Vintar, M. (2004). Evaluating the progress of e-governmentdevelopment: A critical analysis. Information Polity, 9(3), 131. Lambrinoudakis, C., Gritzalis, S., Dridi, F., & Pernul, G. (2003). Security requirements

for e-government services: A methodological approach for developing a commonPKI-based security policy. Computer Communications, 26(16), 1873-1883. Lanvin, B., & Lewin, A. (2006). The next frontier of E-government: Local governments

may hold the keys to global competition. Global Information Technology Report,2007, 51–63.

Layne, K., & Lee, J. (2001). Developing fully functional E-government: A four stagemodel. Government Information Quarterly, 18(2), 122-136. Löfstedt, U. (2005). E-government–assessment of current research and some proposals

for future directions. International Journal of Public Information Systems, 1(1),39-52. Nguyen, T. T., Sansoni, S., & Le Noir, M. (2003). How e-government are you? e-

government in france: State of play and perspectives. US: IBM BusinessConsulting Services. NOIE, & DMR. (2003). E-government Benefits study (Commonwealth of Australia ed.).

66

Page 67: Local e-Government Bench-learning

Local e-Government Bench-learning

Camberra (Australia): National Office for the Information Economy.Noordegraaf, M. (2003). Management by measurement? public management practices

amidst ambiguity. Public Administration, 81(4), 853. Norris, D. F. (2009). E-government among american local governments: Adoption,impacts, barriers and lessons learned. Paper presented at the The International

Conference for Administrative Development: Towards Excellence in Public SectorPerformance, Riyadh (Saudi Arabia).

Norris, D. F., & Moon, M. J. (2005). Advancing e-government at the grassroots:Tortoise or hare. Public Administration Review, 65(1), 64-75.

Norris, D. (2005). E-government at the american grassroots: Future trajectory. Paperpresented at the 38th Annual Hawaii International Conference on ystem Sciences,2005. HICSS'05, Hawaii. 125b.

OCDE. (2003). The E-government imperative: Main findings.Panopoulou, E., Tambouris, E., & Tarabanis, K. (2008). A framework for evaluating

web sites of public authorities. Paper presented at the Aslib Proceedings: NewInformation Perspectives, , 60(5) 517–546.

Patel, H., & Jacobson, D. (2008). Factors influencing citizen adoption of E-government:A review and critical assessment. Paper presented at the 16th EuropeanConference on Information Systems, Galway (Ireland). 1058-1069.

Peristeras, V., Tsekos, T., & Tarabanis, K. (2002). Analyzing e-government as aparadigm shift. UNTC Occasional Papers Series, 1, 2-14.

REGIONAL-IST. (2003). E-government and e-business indicators: Selection of keyindicators in accordance with eEurope.

Ronaghan, S. A. (2002). Benchmarking E-government: A global perspective. UnitedNations. Division for Public Economics and Public Adminsitration. Sakowicz, M. (2003). How to evaluate E-government? different methodlogies and

methodes. Paper presented at the Eleventh NISPAcee Annual Conference. Shackleton, P., Fisher, J., & Dawson, L. (2004). Evolution of local government E-services: The applicability of e-business maturity models. Paper presented at the

37th Annual Hawaii International Conference on System Sciences, Hawaii. 5. Tambouris, E., Gorilas, S., Kavadias, G., Apostolou, D., Abecker, A., Stojanovic, L., etal. (2004). Ontology-enabled e-gov service configuration: An overview of the

OntoGov project. Knowledge Management in Electronic Government, , 122-127. United Nations. (2010). E-government survey 2010. leveraging e-government at a time

of financial and economic crisis No. ST/ESA/PAD/SER.E/131). New York: UnitedNations.

67

Page 68: Local e-Government Bench-learning

Local e-Government Bench-learning

Vassilakis, C., & Lepouras, G. (2006). An ontology for e-government public services.Encyclopedia of E-Commerce, E-Government and Mobile Commerce, , 865–870.

Warkentin, M., Gefen, D., Pavlou, P. A., & Rose, G. M. (2002). Encouraging citizenadoption of e-government by building trust. Electronic Markets, 12(3), 157-162. Wescott, C. G. (2001). E-government in the asia-pacific region. Asian Journal of

Political Science, 9(2), 1-24. Wohlers, T. E. (2007). E-government: Trends and sophistication at the local level ofgovernment. Paper presented at the Annual Meeting of the Midwest Political

Science Association, Palmer House Hotel, Chicago, IL.

68

Page 69: Local e-Government Bench-learning

Local e-Government Bench-learning

8 Bibliography

8.1 Journal ArticlesAl-Hashmi, A., & Darem, A. B. (2008). Understanding phases of E-governmentproject.Emerging Technologies in E-GovernmentAlsuwaidi, H. (2009). Evaluating the factors affecting for the adoption of E-governmentservices in the public sector (UAE).Andersen, K. V., & Henriksen, H. Z. (2006). E-government maturity models: Extension

of the layne and lee model. Government Information Quarterly, 23(2), 236. Andersen, K. V. (2006). e-government: Five key challenges for management. Electronic

Journal of e-Government, 4(1), 1-8. Apostolou, D., Stojanovic, L., Pariente-Lobo, T., Batlle, J., & Papadakis, A. E. (2005).

Configuring e-government services using ontologies. UPGRADE. the EuropeanJournal for the Informatics Professional, 6(6)

Arslan, A. (2007). Turkish local e-governments: A longitudinal study. The ElectronicJournal of e-Government, 5(2), 95-106.

Bannister, F. (2007). The curse of the benchmark: An assessment of the validity andvalue of e-government comparisons. International Review of AdministrativeSciences, 73(2), 171.

Baum, C., & Di Maio, A. (2000). Gartner’s four phases of e-government model.Stamford, Ct., Gartner Group, 21, 12-6113.

Belanger, F., & Carter, L. (2008). Trust and risk in e-government adoption. Journal ofStrategic Information Systems,

Carter, L., & Bélanger, F. (2005). The utilization of e-government services: Citizentrust, innovation and acceptance factors*. Information Systems Journal, 15(1), 5-25.

Castelnovo, W., & Simonetta, M. (2007). The evaluation of e-government projects forsmall local government organisations. Electronic Journal of e-Government, 5(1),21-28.

Chen, Y. C., & Gant, J. (2001). Transforming local e-government services: The use ofapplication service providers. Government Information Quarterly, 18(4), 343-355.

Claver-Cortes, E., De Juana-Espinosa, S., & Tari, J. J. (2008). E-government maturity atspanish local levels. International Journal of Value Chain Management, 2(1), 4-17.

69

Page 70: Local e-Government Bench-learning

Local e-Government Bench-learning

Codagnone, C. (2007). Measuring eGovernment: Reflections from eGEP measurementframework experience. Fostering Public Sector Performance in Europe, , 88.

Codagnone, C. (2008). The future of eGovernment: Supply and demand trends andneeds. Delivered by RSO/Sineura within the Study “Future Technology Needs forFuture eGovernment Services” for the European Commission, DG InformationSociety, eGovernment Unit, Milan, Italy.[Online] Available at: Http://82.187, 13

Cook, M. E. (2000). What citizens want from e-government. Center for Technology inGovernment,

Cotterill, S., Researcher, P., Building, M. K., & King, S. (2006). The role of socialnetworks in the development of english local e-government.Coursey, D., & Norris, D. F. (2008). Models of E-government: Are they correct? anempirical assessment. Public Administration Review, 68(3), 523-536.

doi:10.1111/j.1540-6210.2008.00888.xCullen, R., & Houghton, J. (2000). Democracy online: An assessment of new zealandgovernment web sites. Government Information Quarterly, 17(3), 243-67. Flak, L. S., Olsen, D. H., & Wolcott, P. (2005). Local E-government in norway.

Scandinavian Journal of Information Systems, 17(2), 41-84. Gilbert, D., Balestrini, P., & Littleboy, D. (2004). Barriers and benefits in the adoptionof e-government. International Journal of Public Sector Management, 17(4), 286-

301. Griffin, D., Foster, A., & Halpin, E. (2004). Joined-up e-government: An exploratorystudy of UK local government progress. Journal of Information Science and

Technology, 2Griffin, D. (2005). An exploratory evaluation of UK local e-government from anaccountability perspective. Electronic Journal of eGovernment, 3(1)Guerrini, W. (2010). E-government and online government-citizen interaction. E-

Governance: Electronic Administration and Democracy.GADE, Gupta, M. P., & Jana, D. (2003). E-government evaluation: A framework and case study.

Government Information Quarterly, 20(4), 365. Hiller, J. S., & Belanger, F. (2001). Privacy strategies for electronic government. E-

Government, , 162-198. Holden, S. H., Norris, D. F., & Fletcher, P. D. (2003). Electronic government at the local

level: Progress to date and future issues. Public Performance & ManagementReview, , 325-344.

Holzer, M. (2003). A comparative e-government analysis of new Jersey’s 10 largestmunicipalities.

70

Page 71: Local e-Government Bench-learning

Local e-Government Bench-learning

Holzer, M., & Manoharan, A. (2006). Global trends in municipal E-government: Anonline assessment of worldwide municipal web portals. Foundations of E-Government,

Holzer, M., & Yang, K. (2004). Performance measurement and improvement: Anassessment of the state of the art. International Review of Administrative Sciences,70(1), 15.

Janssen, D., Rotthier, S., & Snijkers, K. (2004). If you measure it they will score: Anassessment of international eGovernment benchmarking. Information Polity, 9(3),121-130.

Janssen, M., & van Veenstra, A. (2005). Stages of growth in e-government: Anarchitectural approach.

Joseph, R. C., & Kitlan, D. P. (2007). Key issues in E-government and publicadministration. Handbook of Research on Public Information Technology,

Kaliontzoglou, A., Sklavos, P., Karantjias, T., & Polemi, D. (2005). A secure e-government platform architecture for small to medium sized public organizations.Electronic Commerce Research and Applications, 4(2), 174-186.

Kaufmann, D., & Kraay, A. (2008). Governance indicators: Where are we, where shouldwe be going? The World Bank Research Observer, 23(1), 1. Kaylor, C., Deshazo, R., & Van Eck, D. (2001). Gauging e-government: A report on

implementing services among american cities. Government Information Quarterly,18(4), 293-307.

Klischewski, R. (2003). Semantic web for e-government. Electronic Government, ,1059-1060.

Kumar, V., Mukerji, B., Butt, I., & Persaud, A. (2007). Factors for successful e-government adoption: A conceptual framework. The Electronic Journal of e-Government, 5(1), 63–76.

Kunstelj, M., & Vintar, M. (2004). Evaluating the progress of e-governmentdevelopment: A critical analysis. Information Polity, 9(3), 131.

Lambrinoudakis, C., Gritzalis, S., Dridi, F., & Pernul, G. (2003). Security requirementsfor e-government services: A methodological approach for developing a commonPKI-based security policy. Computer Communications, 26(16), 1873-1883.

Lanvin, B., & Lewin, A. (2006). The next frontier of E-government: Local governmentsmay hold the keys to global competition. Global Information Technology Report,2007, 51–63.

Layne, K., & Lee, J. (2001). Developing fully functional E-government: A four stagemodel. Government Information Quarterly, 18(2), 122-136.

71

Page 72: Local e-Government Bench-learning

Local e-Government Bench-learning

Löfstedt, U. (2005). E-government–assessment of current research and some proposalsfor future directions. International Journal of Public Information Systems, 1(1),39-52.

Löfstedt, U. (2007). E-government services in local governments-a study ofdevelopment in swedish municipalities. Journal of Organisational Transformationand Social Change, 4(2), 157.

Millard, J. (2008). eGovernment measurement for policy makers. European Journal ofePractice, 4

Moon, M. J. (2002). The evolution of e-government among municipalities: Rhetoric orreality? Public Administration Review, 62(4), 424-433. Noordegraaf, M. (2003). Management by measurement? public management practicesamidst ambiguity. Public Administration, 81(4), 853. Norris, D. F., & Moon, M. J. (2005). Advancing e-government at the grassroots:

Tortoise or hare. Public Administration Review, 65(1), 64-75. OCDE. (2003). The E-government imperative: Main findings.Osimo, D., Millard, J., Foley, P., Patrick, W., Lörincz, B., Strano, G., et al. (2008).Benchmarking eGovernment in the web 2.0 era: What to measure, and how.

Benchmarking, Peristeras, V., Tsekos, T., & Tarabanis, K. (2002). Analyzing e-government as aparadigm shift. UNTC Occasional Papers Series, 1, 2-14. Room, G. (2005). Policy benchmarking in the european union. Policy Studies, 26(2),

117-132. Siau, K. (2005). Synthesizing e-government stage models- a meta-synthesis based on

meta-ethnography approach. Industrial Management Data Systems, 105(3), 443. Sweeney, A. D. P. (2008). Electronic government-citizen relationships. Journal of

Information Technology & Politics, 4(2), 101-116. Tambouris, E., Gorilas, S., Kavadias, G., Apostolou, D., Abecker, A., Stojanovic, L., et

al. (2004). Ontology-enabled e-gov service configuration: An overview of theOntoGov project. Knowledge Management in Electronic Government, , 122-127. Van Dooren, W. (2006). Performance measurement in the flemish public sector: A

supply and demand approach. Faculté Des Sciences Sociales.Louvain, Belgique, Vassilakis, C., & Lepouras, G. (2006). An ontology for e-government public services.

Encyclopedia of E-Commerce, E-Government and Mobile Commerce, , 865–870. Warkentin, M., Gefen, D., Pavlou, P. A., & Rose, G. M. (2002). Encouraging citizen

adoption of e-government by building trust. Electronic Markets, 12(3), 157-162.

72

Page 73: Local e-Government Bench-learning

Local e-Government Bench-learning

Wescott, C. G. (2001). E-government in the asia-pacific region. Asian Journal ofPolitical Science, 9(2), 1-24.

8.2 ReportsBatlle, J., Calderón, R., & López, J. (2009). Local e-government bench-learning survey.Brussels: EUROCITIES. Batlle, J., & López, J. (2007). Key success factors for the adoption of eServices.

Brussels: EUROCITIES. Cap Gemini. (2002). Online availability of public services: How does europe progress.Brussels: European Commission. Cap Gemini. (2003). Online availability of public services: How is europe progressing.

Brussels: European Commission. Cap Gemini. (2004). Online availability of public services: How is europe progressing.Brussels: European Commission. Cap Gemini. (2005). Online availability of public services: How is europe progressing.Brussels: European Commission. Cap Gemini. (2006). Online availability of public services: How is europe

progressing?. report of the 6th measurement. Brussels: European Commission. Cap Gemini. (2007). Benchmarking the supply of online public services. 7th

measurement. Brussels: European Commission. Cap Gemini. (2009). Smarter, faster, better eGovernment. 8th eGovernment benchmark

measurement. Brussels: European Commission. Codagnone, C., Caldarelli, L., Cilli, V., Galasso, G., & Zanchi, F. (2006). Compendium

to the measurement framework. eGovernment economics project. Rome: eGEPConsortium.

Criado, J. I. (2009). Nuevos enfoques en el estudio de la administración electrónica.neoinstitucionalismo y tecnologías de información y comunicación e internet enlas administraciones públicas

Deloitte. (2005). e-citizenship for all benchmark report 2004. Brussels: EUROCITIES. Deloitte. (2005). eCitizenship for all benchmark report 2005. Brussels: EUROCITIES. European Commission. (2000). eEurope 2002 an information society for all. action

plan. Brussels: Commission of the European Communities. European Commission. (2002-1). eEurope 2005: An information society for al. action

plan No. COM(2002) 263). Brussels: Commission of the European Communities. 73

Page 74: Local e-Government Bench-learning

Local e-Government Bench-learning

European Commission. (2002-2). eEurope 2005: Benchmarking indicators No.COM(2002) 655). Brussels: Commission of the European Communities.

European Commission. (2002-3). eEurope benchmarking report No. COM(2002) 62).Brussels: Commission of the European Communities. European Commission. (2006). i2010 eGovernment action plan: Accelerating

eGovernment in europe for the benefit of all No. COM(2006)173). Brussels:Commission of the European Communities. European Commission. (2007). eGovernment progress in EU27+: Reaping the benefits.

Brussels: Commission of the European Communities. Graafland-Essers, I., & Ettedgui, E. (2003). Benchmarking e-government in europe and

the US. statistical indicators benchmarking the information society project.Cambridge (UK): RAND Europe. Hachigian, N. (2002). Roadmap for E-government in the developing world: 10 question

E-government leaders should ask themselves. Los Angeles, CA: Pacific Council onInternational Policy. Holzer, M., & Kim, S. (2003). Digital governance in municipalities worldwide. an

assessment of municipal web sites throughout the world. Newark (New Jersey):The E-Governance Institute. National Center for Public Productivity. RutgersUniversity.

Holzer, M., & Kim, S. T. (2005). Digital governance in municipalities worldwide(2005). A longitudinal assessment of municipal web sites throughout the world.Newark (New Jersey): The E-Governance Institute. National Center for PublicProductivity. Rutgers University.

Holzer, M., & Kim, S. T. (2007). Digital governance in municipalities worldwide(2007). A longitudinal assessment of municipal websites throughout the world.Newark (New Jersey): The E-Governance Institute. National Center for PublicPerformance. Rutgers University.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2005). Governance matters IV: Governanceindicators for 1996-2004. Washington: Wold Bank Institute.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2009). Governance matters VIII: Aggregateand individual governance indicators, 1996-2008. Washington: World BankInstitute.

Kaufmann, D., Leautier, F., & Mastruzzi, M. (2005). Governance and the city: Anempirical exploration into global determinants of urban performance No. 3712).Washington: World Bank Institute.

Kreizman, G. (2002). E-government: Hype and realityGartner Group.

74

Page 75: Local e-Government Bench-learning

Local e-Government Bench-learning

Nguyen, T. T., Sansoni, S., & Le Noir, M. (2003). How e-government are you? e-government in france: State of play and perspectives. US: IBM BusinessConsulting Services.

Nistal, G., & Castro, J. M. (2002). Rationalising eGovernment policies in the europeanunion. A review of the eGovernment working groups. Madrid: Ministerio deAdministraciones Públicas.

REGIONAL-IST. (2003). E-government and e-business indicators: Selection of keyindicators in accordance with eEurope

Ronaghan, S. A. (2002). Benchmarking E-government: A global perspectiveUnitedNations. Division for Public Economics and Public Adminsitration. Saha, P. (2008). Government e-service delivery. identification of success factors from

citizens' perspective (Doctoral Thesis. Lulea (Sweden): Business Administrationand Social Sciences. Lulea University of Technology.

Sawicka, A., & Snaprud, M. eGovernment monitor (eGovMon). Norway: The ResearchCouncil of Norway. Schellong, A. (2010). EU eGovernment benchmarking 2010+. Cambridge, MA:

Institute for Quantitative Social Science. Harvard University. United Nations. (2010). E-government survey 2010. leveraging e-government at a time

of financial and economic crisis No. ST/ESA/PAD/SER.E/131). New York: UnitedNations.

8.3 BooksBatlle, J., Esteban, D., & Rodriguez, J. R. (Eds.). (2007). E-government city models:

Cases from european cities. Barcelona: Ajuntament de Barcelona.Cole, M. I., & Parston, G. (2006). Unlocking public value: A new model for achieving

high performance in public service organizations Wiley.Dunleavy, P., & Margetts, H. (2006). Digital era governance: IT corporations, the state,

and E-government Oxford University Press, USA.MV Consultoria. (2010). IMUE Libro blanco para la medición del uso efectivo de la

administración electrónica MV Consultoria.NOIE, & DMR. (2003). E-government Benefits study (Commonwealth of Australia ed.).

Camberra (Australia): National Office for the Information Economy.OECD. (2003). The E-government imperative. Paris (France): OECD Publishing.Organisation for Economic Cooperation and Development.

75

Page 76: Local e-Government Bench-learning

Local e-Government Bench-learning

Ragin, C. (2000). Fuzzy-set social science. Chicago, IL: University of Chicago Press.

8.4 Conference proceedingsAichholzer, G., Schmutzer, R., & Hochgerner, J. (1998). Bringing public administration

closer to the citizens. Paper presented at the Conference of the Information SocietyForum of the European Commission. Wien.

Alshawy, S., Alahmary, A., & Alalwany, H. (2007). eGovernment evaluation factor.citizens perspective. Paper presented at the European and MediterraneanConference on Information Systems 2007, Spain.

Gefen, D., Pavlou, P., Warkentin, M., & Rose, G. (2002). Egovernment adoption. Paperpresented at the Eighth Americas Conference on Information Systems, 569-576.

Heeks, R. (2006). Understanding and measuring egovernment: Internationalbenchmarking studies. Paper presented at the UNDESA Workshop E-Participationand E-Government: Understanding the Present and Creating the Future,Budapest, 27-28.

Horrigan, J. B. (2005). "On demand" citizens: E-government at high speed. Paperpresented at the FedWeb Conference at George Mason University,

Janssen, D. (2003). Mine’s bigger than yours: Assessing international e-governmentbenchmarking. Paper presented at the 3rd European Conference on e-Government,MCIL, Reading, MA, 209-218.

Janssen, M., & Wagenaar, R. (2004). An analysis of a shared services centre in e-government. Paper presented at the 37th Annual Hawaii International Conferenceon System Sciences, Hawaii. 10.

Jones, S., Irani, Z., & Sharif, A. (2007). E-government evaluation: Reflections on threeorganisational case studies. Paper presented at the 40th Annual HawaiiInternational Conference on System Sciences. HICSS 2007, Hawaii. 105.

Jones, S., Irani, Z., Sharif, A., & Themistocleous, M. (2006). E-government evaluation:Reflections on two organisational studies. Paper presented at the 39th HawaiiInternational Conference on System Sciences, Hawaii.

Norris, D. F. (2009). E-government among american local governments: Adoption,impacts, barriers and lessons learned. Paper presented at the The InternationalConference for Administrative Development: Towards Excellence in Public SectorPerformance, Riyadh (Saudi Arabia).

Norris, D. (2005). E-government at the american grassroots: Future trajectory. Paperpresented at the 38th Annual Hawaii International Conference on System Sciences,2005. HICSS'05, Hawaii. 125b.

76

Page 77: Local e-Government Bench-learning

Local e-Government Bench-learning

Panopoulou, E., Tambouris, E., & Tarabanis, K. (2008). A framework for evaluatingweb sites of public authorities. Paper presented at the Aslib Proceedings: NewInformation Perspectives, , 60(5) 517–546.

Patel, H., & Jacobson, D. (2008). Factors influencing citizen adoption of E-government:A review and critical assessment. Paper presented at the 16th EuropeanConference on Information Systems, Galway (Ireland). 1058-1069.

Peters, R. M., Janssen, M., & van Engers, T. M. (2004). Measuring e-governmentimpact: Existing practices and shortcomings. Paper presented at the 6thInternational Conference on Electronic Commerce, 489.

Sakowicz, M. (2003). How to evaluate E-government? different methodlogies andmethodes. Paper presented at the Eleventh NISPAcee Annual Conference,

Shackleton, P., Fisher, J., & Dawson, L. (2004). Evolution of local government E-services: The applicability of e-business maturity models. Paper presented at the37th Annual Hawaii International Conference on System Sciences, 5-8.

Wauters, P. (2006). Benchmarking e-government policy within the e-europeprogramme. Paper presented at the Aslib Proceedings: New InformationPerspectives, , 58(5) 389-403.

Wimmer, M., Codagnone, C., & Janssen, M. (2008). Future e-government research: 13research themes identified in the eGovRTD2020 project. Paper presented at theHicss, 223.

Wohlers, T. E. (2007). E-government: Trends and sophistication at the local level ofgovernment. Paper presented at the Annual Meeting of the Midwest PoliticalScience Association, Palmer House Hotel, Chicago, IL.

77

Page 78: Local e-Government Bench-learning

Local e-Government Bench-learning

78

Page 79: Local e-Government Bench-learning

Local e-Government Bench-learning

9 Annex: The Local e-Government Bench-learning survey

79

Page 80: Local e-Government Bench-learning

Local e-Government Bench-learning

80

Page 81: Local e-Government Bench-learning

����������� ��� ����������� ������������ �������������� �� � ���� ����������� ������������� �������������� ������ !"#$%�#&�%'!��() *+,-,(.�/#$0��1�2$#3"����������� 45 6� 78984��() *+,-,(.��:�%'!��!%;#$0�#&�<�=#$�(3$#"!���>�%�!:?��#3�@!@����ABCDE�%'!��!%;#$0�F$��1:�%#1!%'!$�%'!��#>���1#G!$�<!�%:�#&�#G!$�AHI���$1!�>�%�!:����:#<!�HJ�(3$#"!���>#3�%$�!:?�() *+,-,(.�$!"$!:!�%:�%'!���%!$!:%:�#&��%:�<!<F!$:���@�!�1�1!:����@���#13!�;�%'�%'!�(3$#"!�����:%�%3%�#�:��>$#::���;�@!�$��1!�#&�"#��>K��$!�:��&&!>%��1�>�%�!:?�-'!:!���>�3@!L�!>#�#<�>�@!G!�#"<!�%E�%'!�!�G�$#�<!�%E�%$��:"#$%���@�<#F���%KE�:#>�����&&��$:E�>3�%3$!E�%'!���&#$<�%�#����@�0�#;�!@1!�:#>�!%KE���@�:!$G�>!:�#&�1!�!$�����%!$!:%?��() *+,-,(.�;!F:�%!L�MMMN���������ON���

Page 82: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ����� ! ��"����#��$%�& ��' "�����(��! � %��)*+,�-.�/0+1234� � 5678�960�:23�)37,;;20<�5678�=-0:282*3�>�?60;�@2,23;�)*+,�-.�@63A28-06� � B-60�@6+882�>�BC8*6�DE42F�)*+,�-.�@23G20�� � D*0:6�H,3;+6:<�I6JK@3*++�@3-L6MM2�>�D63;�=N2*+�)*+,�-.�@*8L6-�� � I63A-;�I73-<�@860A6�?233O0�>�/*0P-6�Q360A-�)*+,�-.�@*3R*0GP6R� � )P3*;�53*A2�>�)86*32�S6N*;�)*+,�-.�@-8-G06� � T*7;2442�5637-8-<�D2:6�T7*:*<�D7A6�U60288*�>�S60*282�=63-FF*�)*+,�-.�V0;AP2:2� � /0:3W�:2�X-;6�H4*23*0G;�>�X-0�N60�:23�Y-8M�)*+,�-.�?28;*0M*� � =7-R-�Y636M-34*<�Y63*�I*;M686�>�I23N*�Y7MM-020�)*+,�-.�I*860� I6736�T6RL63606<�/0+-02886�D-0G-<�/82;;60:3-�I7;7R2A*�>�H+2.60-�@63+-8-++6�)*+,�-.�I73A*6�� � X-;6�I63+Z02F�>�B-;W�I63+Z02F�)*+,�-.�X*J2M6�� � 986:-�Y-0�>�=6+J606�[\]\�)*+,�-.�=688*00�� � 9_*0 �a82N<�Y6*�Y-+M6�>�Y6+3*0�H6N-R_G*�)*+,�-.�=P2�?6G72� � I63+20�@7;APR60�)*+,�-.�=73*0� � � Q360A-�)63A*88-�)*+,�-.�9*2006�� � b0G3*:�T +F8�>�H-0J6�H6++28L23G23���!�c%d !��ce ����5-R427�Q6L36�f0*N23;*+,� S3g�I*86�T6;AE��(� "%�������"%��5Vh=Va�b)=�/068,;+� � X6i728�)68:23E0�d%%� "���=P2�?6G72<�aA+-L23��jjk� @37;;28;<�B60763,��jjl�� D,-0<�I63AP��jjl� ������@63A28-06<�B702��jjl���� H+-AMP-8R<�aA+-L23��jjl� @37;;28;<�B60763,��jjm��

Page 83: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ����������� !"#$%$ !&#'�()�#*+�,("-$&.�."(/ �000000000000000000000000000000000000000000000000000000000�1�$&#"(2/%#$(&�000000000000000000000000000000000000000000000000000000000000000000000000000000000000000�3�40�5+#*(2(6(.7�000000000000000000000000000000000000000000000000000000000000000000000000000000000000�8�9:;<=>?@=;ABCDEFBAG;%!#!6(./+�0000000000000000000000000000000000000000000000000000000000000000�44�104�%!#+.("$+'� 4��101�2$H+"'$#7� 14�10��.+&+"!6�!&!67'$'�()�%(H+"!.+�!&2�2$H+"'$#7� 11�103�'#!&2!"2�%(H+"!.+�!&2�2$H+"'$#7� +"�%!#+.("7� 13��0�5!#/"$#7�!&2� +"%+ #$(&�()�!2( #$(&�00000000000000000000000000000000000000000000000000��4��04�5!#/"$#7�6+H+6'� �4��01� +"%+ #$(&�()�!2( #$(&�6+H+6'� ����0��+/"( +!&�!H+"!.+'� +"�%!#+.("7� �3��03�'#!&2!"2�!&2�!22$#$(&!6�'+"H$%+'�!H+"!.+'� �I��0J�+/"( +!&�!H+"!.+'� +"�'+"H$%+� �8�30�%$#$+'�5! '�000000000000000000000000000000000000000000000000000000000000000000000000000000000000000�31�J0�*$.*6$.*#+2�'+"H$%+'�0000000000000000000000000000000000000000000000000000000000000000000000000�IJ�%(&%6/'$(&'�0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000�IK�"+)+"+&%+'�00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000�84��

Page 84: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ������ !"�����#$%&'#()*#)+�,-�+.#�/012,3�4567879:;4:7<8=9>:;?@>;@A<B?:7<8C>D:>E;78F9>64G<;>HI48F@9>78>+.#�JKL�+.#�KM*,),-+(K+,&)�J&(N-O�+.#�JKL�,+�M#2,'#(-�-#(',3#-�+&�3,+,P#)-�K)M�3&*/K),#-�K)M�+.#�JKL�3,+,P#)-�K(#�,)+#(K3+,)Q�J,+.�,+R�S+�,-�K�(KM,3K2�+(K)-T&(*,)Q�/(&3#--�+.K+�3K)�1#�3K((,#M�&0+�+.K)N-�+&�S)T&(*K+,&)�K)M�U&**0),3K+,&)-�V#3.)&2&Q,#-�#)K12,)Q�K�J,M#�;48F@><W>X<997E7B7:7@9>:<>;@:I78Y>:I@>ZI<B@>X?EB7H><;F487[4:7<8=9>Z4\><W>Z<;Y78FC>�]33&(M,)Q�+&� _`abV]VO�+.#�/012,3�-#3+&(�,-�(#-/&)-,12#�T&(�)#K(2L��cd�&T�+.#�%ef�,)�0(&/#R�V.#-#�T,Q0(#-�/&,)+�&0+�.&J�+.#�+(K)-T&(*K+,&)�&T�+.#�/012,3�-#3+&(�,-�&T�3K/,+K2�,*/&(+K)3#�T&(� 0(&/#K)�#3&)&*,3�Q(&J+.R�S)�KMM,+,&)O�,)� 0(&/#O�+.#�-01-,M,K(,+L�/(,)3,/2#�#)-0(#-�+.K+�M#3,-,&)-�K(#�+KN#)�K-�32&-#2L�K-�/&--,12#�+&�+.#�3,+,P#)R�V.K+�*#K)-�+.K+�+.#� 0(&/#K)�_),&)�M&#-�)&+�K3+�0)2#--�,+�,-�*&(#�#TT#3+,'#�+.K)�K3+,&)�+KN#)�K+�K�)K+,&)K2O�(#Q,&)K2�&(�2&3K2�2#'#2R�g&(#&'#(O�-0('#L-�3K((,#M�&0+�,)�g#*1#(�b+K+#-O�K(#�-.&J,)Q�+.K+�1#+J##)�hcd�K)M�icd�&T�+.#�3,+,P#)j-�,)+#(K3+,&)-�J,+.�/012,3�1&M,#-�K(#�M&)#�K+�-01$)K+,&)K2�2#'#2�kl##N-O�mccnoR�S)�-03.�K�3&)+#p+O� 0(&/#K)�3,+,#-�q�J.,3.�1#)#T,+�&T�K�)&+#J&(+.L�K0+&)&*L�J,+.�K)�,*/&(+K)+�)0*1#(�&T�3&*/#+#)3,#-�q�K(#�+.0-�/2KL,)Q�K)�,*/&(+K)+�(&2#�,)�+.#�&(QK),PK+,&)�&T�+.#� 0(&/#K)�b&3,#+LR�V.#(#T&(#O�K2+.&0Q.�#$%&'#()*#)+�,-�K�(#K2,+L�K+�K22�/012,3�&(QK),PK+,&)-�2#'#2-�k-+K+#O�(#Q,&)K2�K)M�2&3K2oO�,+�.K-�,+-�1,QQ#-+�,*/K3+�K+�K�2&3K2�2#'#2R�r&JKMKL-O�K22�3,+L�Q&'#()*#)+-�K)M�KM*,),-+(K+,&)-�T,)M�+.#*-#2'#-�,)�K�+(K)-,+,&)�T(&*�+(KM,+,&)K2�*&M#2-�&T�2&3K2�/012,3�&(QK),-K+,&)�+&�K�)#J�&)#�*&(#�#TT,3,#)+�K)M�*&(#�#TT#3+,'#�k#$%&'#()*#)+�f(&Q(#--�,)� _mstO�mccsoR�]+�+.#�-K*#�+,*#O�+.#�(#2K+,&)-.,/�J,+.�3,+,P#)-�,-�3.K)Q,)Q�T(&*�K�-3#)K(,&�1K-#M�&)�K�*02+,/2,3,+L�&T�-/#3,K2,P#M�3&0)+#(-�+&�4>u<8@$-+&/$9I<Xv>6<5@BC�w .K+�,-�1#.,)M�+.,-�/(&3#--�,-�+.#�+(K)-T&(*K+,&)�&T�+.#�J.&2#�&(QK),PK+,&)�/0++,)Q�3,+,P#)-�K+�+.#�3#)+(#�&T�+.#�K3+,&)O�+.K+�,-�3(#K+,)Q�K�(#K2�3,+,P#)$&(,#)+#M�&(QK),PK+,&)�3K/K12#�+&�0)M#(-+K)M�J.&�+.#�3,+,P#)�,-�K)M�J.K+�K(#�+.#�3,+,P#)�)##M-�,)�#K3.�3K-#R�S)�&+.#(�J&(M-O�+.#-#�3.K)Q#-�+(L�+&�2#KM�+&�/(&K3+,'#)#--�K)M�K)�,)3(#K-#�,)�/012,3�'K20#R�]-�+.,-�+(K)-T&(*K+,&)�+&03.#-�+.#�3&(#�10-,)#--�&T�+.#�2&3K2�/012,3�KM*,),-+(K+,&)O�3,+,#-�*0-+�3K(#T022L�*K)KQ#�+.#�/(&3#--�,)�&(M#(�+&�-.&(+#)�,+�K)M�#)-0(#�K�-K+,-TK3+&(L�&0+/0+R�U,+L�*K)KQ#(-�)##M�+&�.K'#�(#2#'K)+�,)T&(*K+,&)�K1&0+�J.K+�/(&3#--#-�+&�+K3N2#�T,(-+O�.&J�+&�(#*&M#2�+.#*O�K)M�.&J�+&�*#K-0(#�+.#�/(&Q(#--�ke0)2#K'LO�gK(Q#++-O�xK-+&JO�y�V,)N2#(O�mccnoR�z(&*�+.#�3,+,#-j�/&,)+�&T�',#JO�+J&�,*/&(+K)+�+.,)Q-�K(#�)##M#M�+&�2#KM�+.#�+(K)-T&(*K+,&)�/(&3#--�-033#--T022L{�+.#�#p,-+#)3#�&T�K)�#$%&'#()*#)+�3,+L�*&M#2�K)M�+.#�*#K-0(#*#)+�&T�+.#�3,+Lj-�#$%&'#()*#)+�M#'#2&/*#)+R�l&J�,-�+.#�3,+L�M&,)Q�+.#�|&0()#L}�w .#(#�,-�+.#�3,+L�Q&,)Q}�]+�J.,3.�-+KQ#�&T�+.,-�|&0()#L�,-�+.#�3,+L}�V.#-#�K(#�3K/,+K2�~0#-+,&)-�#K3.�U a�)##M-�+&�K)-J#(�+&�*KN#�+.#�+(K)-T&(*K+,&)�/(&3#--�K�-033#--R��V.#�/(#',&0-�-0('#L�K1&0+�#$�<A@;86@8:>�7:\>�<5@B9>@8:7:B@5>u@$%&'#()*#)+�U,+L��<5@B9�>H49@9>W;<6>�?;<X@48>H7:7@9v�k &M(�Q0#PO�xK+22#O�y� -+#1K)O�mccsoO�3&)M03+#M�M0(,)Q�mccn�K)M�mccs�,)�+.#�3&)+#p+�&T� _`aUSVSb�)#+J&(NO�/(&',M#-�K�'K20K12#�+&&2�T&(�3,+L�*K)KQ#(-�+&�,),+,K+#�K)M�Q0,M#�+.#�#$%&'#()*#)+�+(K)-T&(*K+,&)R�V.K+�-0('#LO�J.,3.�

Page 85: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ����� ! " #��$�%"$&! $ !'#�$(()�$�*� ��#+,�'#)�-#� ����.")�(#$��/! !#�0��*�1��*�1��#'#��&#$2!�3�.")�(#$���! !#��4$�#2��"�*� )$��4�)-$ !���$&��3� *#�($� � #��5#$)�0�1*!�*�1#)#� *#�6#5�2#�!�!��� $6#��$�2� *#�-$!��!��(!)$ !����4� *#!)�(�&!�!#�7��8& *�"3*�$�&� ��4�&#�������$��9#�&#$)� �4)�-� *$ ��")'#50�! �!����&5���#��4� *#��##2#2���� )!9" !���� ���"��#��4"&&5�-$�$3#� *#�&��$&�("9&!��$2-!�!� )$ !��� )$��4�)-$ !��7�:*#�� *#)�-!��!�3�2!-#��!���!�� *#�-#$�")#-#� ��4�*�1�$��! 5�!��()�3)#��!�3�!�� *!��!-(�) $� �()��#��7��:*#)#�!��$�&$�6��4�!�4�)-$ !���)#3$)2!�3�#+,�'#)�-#� ���� *#�&��$&�3�'#)�-#� �&#'#&7�;(� ����10� *#�-�� �)#&#'$� �� "2!#��$9�" �#+,�'#)�-#� �-#$�")#-#� ���-($)!�3�#+,�'#)�-#� �#'�&" !���*$��9##�� $!&�)#2�$ ���"� )5�&#'#&�</$(3#-!�!0�=>>?@�$�2�</$(3#-!�!0�=>>A@7�8& *�"3*���-#�!�4�)-$ !���!��$&���$'$!&$9&#�$ �)#3!��$&�&#'#&0���&5�&! &#�!�4�)-$ !���!��4��"�#2����&��$&�&#'#&�#+,�'#)�-#� 7��B") *#)-�)#0�1*!&#�!�2!�$ �)��$ ���"� )5�&#'#&�$)#�1!2#&5�$��#( #2�$�2���--��&5�"�#20� *#�()#�#��#��4��"�*�� $�2$)2��$ �)#3!��$&�&#'#&�2#�)#$�#�7�8�2�4!�$&&50�$ �&��$&�&#'#&0� *#)#�!���"�*�$�&�1�"�#��4�� $�2$)2�!�2!�$ �)�� *$ �! �!���� �(���!9&#� ��(#)4�)-�%"$&! 5���-($)!�����9# 1##��2!44#)#� ��! 5��$�#�7��:*#�/$(3#-!�!�C#��*-$)6�*$��$�&��3� )$2! !��0�$�2����� ! " #��$�'$&"$9&#���")�#��4�1#&&�-#$�")#2�2$ $�����#)�!�3�#+,�'#)�-#� 7�D�1#'#)0�! �1$��2#�!3�#2��#$)&5�E>�5#$)��$3��!�� *#�4)$-#1�)6��4� *#�#+.")�(#�8� !���F&$���4�)�G#-9#)�H $ #�� ��#��")#�$�3#�#)$&!I#2�#&#� )��!��$��#��� ��-$!��9$�!��("9&!���#)'!�#�7�8��$�����#%"#��#0�! �!��4��"�#2����#+,�'#)�-#� �$ �$�� $ #�&#'#&0�1*!�*��$��9#�#$�!&5�2#)!'#2�4)�-� *#�&!� ��4� *#�=>�9$�!���#)'!�#��$�$&5�#27�8�%"!�6�'!#1���� *#�#� 1#� 5��#)'!�#���*�1�� *$ � *#5�$)#��� �$((&!�$9&#� ��&��$&�#+,�'#)�-#� �-#$�")#-#� 7��B") *#)-�)#0� *#�/$(3#-!�!�C#��*-$)6�!� )�2"�#2�4�")�&#'#&���4�#+,�'#)�-#� ��#)'!�#����(*!� !�$ !��J�!�4�)-$ !���<:E@0���#+1$5�!� #)$� !���<:=@0� 1��1$5�!� #)$� !���<:K@�$�2� )$��$� !���<:L@7�M�1$2$5�0��! 5�#+,�'#)�-#� ��#)'!�#��$)#�#'�&'!�3�%"!�6&5�$�2�2##(&5�NOPQRSTOUVQWXYZ[RVQ\RR]X^OT_\RR\RXPQ XVQ_a[ VQWX_VNVb\QcRX_TQN)!9" !��7�8& *�"3*� *#�/$(3#-!�!�-�2#&�)#�#� &5�#'�&'#2�!��&"2!�3�$�� *�&#'#&��4���(*!� !�$ !��0�! �� !&&�2�#���� �4! �1! *�&��$&�#+,�'#)�-#� ��*$)$� #)!� !��7�:*$ �!�� *#�)#$����1*5� *!���")'#5���-#��9$�6� �� *#��)!3!�$&�-�2#&��4�4�")�(*$�#���4�#+,�'#)�-#� �!� )�2"�#2�95�,$) �#)�,)�"(�<C$"-�d�e!�G$!�0�=>>>@0�1*!�*�$)#J�()#�#��#0�!� #)$� !��0� )$��$� !���$�2� )$��4�)-$ !���$�2�$2$( �! � ��2#��)!9#�*!3*#)���(*!� !�$ !���&#'#&�� *$ �$)#��"))#� &5�$((#$)!�3�$ �&��$&�&#'#&7�G�)#�'#)0�)#3$)2!�3�&��$&�&#'#&0�9#��*-$)6!�3��")'#5��-$!�&5�4��"����� *#�3#�#)$&�$�(#� ���4� *#��44!�!$&�1#9�! #7�8& *�"3*0� *#)#�$)#���-#�#f�#( !�����"�*�$�� *#�Y\gVNVb\QRhV XSTOXPaaXZ\Q_hUPOiXO\ TON]Xjk\aTVNN\XlXmnopgqrqmstXuvvwtXuvvxXPQ XuvvyX#2! !���@���$))!#2��" �95�e#&�! #�$�2�()�-� #2�95� *#�.;z{/|:|.H��# 1�)6�1! *� *#��(#�!$&��"((�) ��4�:*#�D$3"#�/! 5�/�"��!&0�!��3#�#)$&0��")'#5��)$)#&5�4��"�����#+�#)'!�#��%"$&! 5��)��! !I#��$2�( !����4�#+�#)'!�#�7�:*#)#4�)#0�$���-(&# #�'!#1��4� *#�#+,�'#)�-#� �()��#���!���! !#��!���� �(���!9&#� �2$57��

Page 86: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ������ !"#�$%��&%$'(!)�$&�!*$+)�)�,�-'$.',���$%� $"! �,/0$1,'&(,&)��&�)�!)�1,�!&2�)�,�',! �&,,2�$%��!1�&.�)�����&%$'(!)�$&�)$�*,)),'�(!&!.,�)�,�,/0$1,'&(,&)�)'!&�%$'(!)�$&�-'$",���($)�1!),2�3!'", $&!�4�)5�4$+&"� �)$�-'$-$�,��&�6+&,�7889�)�,�"',!)�$&�$%�!�&,:�:$'#�&.�.'$+-;����,�:$'#�&.�.'$+-�:!��,1,&)+! 5�"',!),2��&�<")$*,'�7889�:�)��)�,�!�(�$%�2,%�&�&.�!�(,!�+',(,&)�%'!(,:$'#�%$'� $"! �,/0$1,'&(,&)�*!�,2�$&�!�*$))$(/+-�!--'$!"��!&2�%$"+�,2�$&�,/=,'1�",��(!)+'�)5�!&2�"�)�>,&?��!2$-)�$&�!&2@�!)�)�,��!(,�)�(,@�-,'%$'(�&.�)�,�%�'�)�(,!�+',(,&)�$1,'�AB�C+'$-,!&�"�)�,�;������',-$')����-',�,&)�&.�)�,�',�+ )��$%�)����:$'#�&.�.'$+-;�

Page 87: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ����� !"�#� ��$�%&'�(')*�+,-./�&�012-3&4�+,.35*&'6-./�,7,'3-8,�-.�95-35�05,�',8:408�85)9�).41�'&.6-./8�)(�3-0-,8;�05,�9)'6�2')2)8,<�&-*,<�0)�+,�&�+)00)*=:2�,7,'3-8,�80&'0-./�0)�4-80,.�3-0-,8�&.<�:.<,'80&.<-./�05,�6-.<�)(�8,'>-3,8�05,1�2')>-<,�&.<�05,�',&4�.,,<8�05,1�5&>,�95,.�*&.&/-./�05,�,=?)>,'.*,.0�0'&.8()'*&0-).@�A,>,4)2-./�&�*,&8:',*,.0�('&*,9)'6�<,8-/.,<�(')*�05,�83'&035�+1�05,�3-0-,8�&.<�()'�05,�3-0-,8@�B5,�8,4,30-).�)(�8,'>-3,8�9&8�CDEFGHIJHKLIEFHDMKNDOOPHQRISTGFGHCPHMTKTFEUHVLFHRFENOKEHDRFHQRFEFJKFGHKLRINWLHXYTKPHZDQE[H05&0�3)*2&',�,&35�3-01�9-05�05,�\:')2,&.�&>,'&/,@�B5-8�+)00)*=:2�&22')&35�4,<�:8�0)�&�34,&'�4,&'.-./�2')3,88�-.�95-35�&(0,'�&�<,0&-4,<�&.&418-8�)(�05,�2')>-<,<�8,'>-3,8;�+,80�2'&30-3,8�&',�9,44�-<,.0-(-,<;�4-80,<�&.<�<,83'-+,<�0)�(&3-4-0&0,�85&'-./�)(�6.)94,</,�&.<�<-88,*-.&0-).@�B5,8,�&',�05,�',&8).8�951�05,�8:'>,1�9&8�&48)�]:&4-(-,<�&8�&�^,.35=4,&'.-./�2')_,30@�B5,�0,&*�9&8�3)*2)8,<�+1� &'3,4).&� -01� ):.3-4�&30-./�&8�0,&*�4,&<,';�a,.0,)�b B�c.&4180�05&0�3&''-,<�):0�05,�B,35.-3&4� ))'<-.&0-).;�<&0&�3)44,30-).�&.<�&.&418-8;�A'@�d-4&�?&83e�)(�05,�a)*2,:�%&+'&�f.->,'8-01�&8�g3-,.0-(-3�c<>-8)'�&.<�hi�3-0-,8�+,4)./-./�0)�05,�\fjk`bB\g�l,09)'6@�B5,�*,05)<)4)/1�,*24)1,<�5&8�+,,.�+&8,<�).�a,.0,)m8�d\Bjb c�*,05)<)4)/1�-.34:<-./�8)*,�&<=5)3�&''&./,*,.08�95-35�9,',�.,3,88&'1�<:,�0)�05,�.&0:',�)(�05,�8:'>,1@��nopqrstuvtwsxysz{|t}~�����t�sy�z�z�zp���FJKFI�EHZ�V��Y�H�FKLIGIOIWPH�&8�35)8,.�8-.3,�).,�)(�05,�/)&48�)(�05,�8:'>,1�9&8�0)�2')>-<,� &.� &/',,<� *,&8:',*,.0� ('&*,9)'6� &224-3&+4,� 0)� 4)3&4� &<*-.-80'&0-).8@�B5,',()',;� -0� 8,,*,<� &� 8:-0&+4,� 8):'3,� 8-.3,� -0� -8� &� 3).0'&80,<� *,&8:',*,.0�*,05)<)4)/1@�k.�05,�)05,'�5&.<;�-0�2')>,<�0)�5&>,�&�9-<,�&*):.0�)(�,72,'-,.3,8�+,5-.<�95-35�5&<�&44)9,<�+,.35*&'6-./�-.�8)*,�3&8,8�&4)./�9-05�05,�<,0,30-).�)(�+,80=2'&30-3,8@��EHKLFH�TWNRFHELI�E�H�FJKFI�EHZ�V��Y�HTEHDH�TSF=80&/,�*,05)<)4)/1@�B5,�(-'80�80&/,�-8�',4&0,<�0)�05,�<,(-.-0-).�)(�8:-0&+4,�-.<-3&0)'8;�8:'>,18�&.<�<&0&=/&05,'-./�2')3,88@�B5-8�80&/,�-8�)(�3):'8,�3':3-&4�8-.3,�-0�9-44�<,0,'*-.,�05,�(-.&4�',8:408@�B5,�)+_,30->,�-8�0)�35))8,�-.<-3&0)'8�05&0�9-44�+)05�2')>-<,�05,�-.()'*&0-).�4))6,<�()'�&.<�8:-0&+4,�0)�+,������������������������ ¡� ¢�£�¤ ���������¥¦��§����¦� ����  ����¥¦��§����¦���������  ��¦���������������¦���������©�����¦��������ª ��� ���«��§�����������¦��¢¬���­��� ¢­� ���������ªª�������� �®©°±²®³© µ±°µ®¶·µ¹®±º»²³µ¼±°³±³½¾©©

Page 88: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ������� !"#�$%&! �%'(�)*+%�,-%!%��*'�+&,��!�-&.%�/%%*�+-�!%*0�,-%�1��+%!!�,���&,-%��,-%�'&,&�*%+%!!&�#�,��$%&! �%�,-%!%��*'�+&,��!��!�&"!��'%2�*%'(�)*�,-%�!%+�*'�!,&�%0�,-%�&+, &"�'&,&3�&,-%��*��1��+%!!�,&4%!�1"&+%!�2�""�5�*��5-&,�-&!�/%%*�'%!��*%'�' ��*��,-%�2��!,�!,&�%(�)*+%�,-%��*2��$&,��*��!��&,-%�%'0�,-%�*%6,�!,&�%�!,&�,!7�,-%��*2��$&,��*��!�1��+%!!%'�&*'�,-%��*'�+&,��!�&�%��%*%�&,%'(�8-%*0�,-%�2� �,-�!,&�%��2�,-%�1��+%!!�,&4%!�1"&+%7�5�,-�,-%��%*%�&,%'��*'�+&,��!0�,-%�!+���*���!�+&"+ "&,%'�&"�*��,-%�'�22%�%*+%!�,�5&�'!�,-%��%2%�%*+%���� 1(�9�*&""#0�' ��*��,-%�"&!,�!,&�%0�&��%1��,��!��%*%�&,%'�!,&,�*��,-%�/%*+-$&�4�*��!,&�%��%! ",!�:!+���*��&*'�+�$1&��!�*�5�,-��%2%�%*+%���� 1;0��*+" '�*��&*��*,%�1�%,&,��*��2�,-%��%! ",!�&++��'�*��,��%&+-�1&�,�+ "&��+&!%!(�<�5%.%�0�&!��,�-&!�/%%*�!,&,%'��*�,-%��*,��' +,��*0�5-%*�&11��&+-�*��%3=�.%�*$%*,�$%&! �%$%*,�&,�"�+&"�"%.%"0�,-%�%�&�%�*��! �,&/"%�!,&*'&�'!�'%2�*%'�#%,(�8-%�%2��%0�,-%�'%!��*�1-&!%��*+" '%'�,-%�+�*!%*! !��.%��&�+�$$�*��%2%�%*+%7�,-%��"�/&"�+&,&"�� %(�>!��,�5�""�/%�%61"&�*%'�2 �,-%��'�5*0�&""�!%�.�+%!��*+" '%'��*�,-%�+&,&"�� %�5%�%��&,-%�%'�?@AA@BCDEFGCHCIJKFLM@L@JNAJOFPAJ@QFHRIFGNHNA@ESIKJFGNHIE@MCIJFBIMIFTCJGSJJITFNDTFNEMIITF&$�*��1&�,�+�1&*,!(��PD@HRIMFTC??IMIDGIFNU@DEFVIDHI@KJFWXYZ[\PFUIHR@T@A@E]FNDTFHRIFMIJSAHCDEFUIHR@T@A@E]F�!�,-%�2&+,�,-&,�,-�!�! �.%#�&�$%'�,��&.��'�,-%��&*4�*�!��2�+�,�%!(�8-%�%2��%0�,-%�&*&"#!�!�&*'�/%*+-$&�4�*��!,&�%!�+�$/�*%'��*,���*%7��*!,%&'��2�&��%2%�%*+%���� 10�,-%��%2%�%*+%� !%'�2���+�$1&��!�*�-&!�/%%*�,-%��"�/&"��%! ",0�,-%��*'�+&,���/%�*��,-%�'�22%�%*+%!�/%,5%%*�%&+-�+�,#�&*'�,-%���� 1��%! ",(�8-�!�-&!�1��.%*�,��/%�&�.%�#�+�*!,� +,�.%�5&#��2�+�$1&��*�0�/%+& !%��,�-��-"��-,!�,-%�&�%&!�5-%�%�+�,�%!�%6+%"�&!�5%""�&!�,-%�&�%&!�5-%�%�,-%#�-&.%�,���$1��.%�5�,-� ,��%+ ���*��,���&*4�*�!(�*'%%'0��&*4�*�!�&�%�*�,�! �,&/"%�2���"�+&"�+&!%!�' %�,��,-%�5�'%�.&��%,#��2�+&!%!(�8-%��%! ",�*��$%,-�'�"��#�+�*!�!,%'�,-%�%2��%��2�2� ��1-&!%!�'%.%"�1%'�' ��*��_ �7�������� abcdefghiggjdekflmngefndopbqcgrfpstutotclmngvswnfnx

Page 89: Local e-Government Bench-learning

� ��������������� ������������� ������������������������� ����� !�"#$�%��&"�'#(&$)�"#$�&��*$+�,(&�-$&�! $-.�/ 0$�(11�"#$�'(�"�0�'( "�0�"�$&�#(-�23� $-�"#$�&��*$+)�"#$�03443 �352$0"�*$&�( -�(�,3�6�'1( �,$�$�(!�$$-.�7#$�&$03 -�'#(&$�3%�"#$�&��*$+�,(&�%30�&$-�3 �"#$�$1(53�("�3 �3%�"#$�8135(1�9$�*�0$&�:("(13!�$.�;(0#�0�"+�-�(%"$-�(�1�&"�,�"#�(�-$&0��'"�3 �3%�&34$�3%�"#$�$<9$�*�0$&�"#$+�'�3*�-$-.�=43 !&"�(11�"#$�$<9$�*�0$&�!("#$�$-)�(�03443 �&$"�3%�&$�*�0$&�,(&�&$1$0"$-�3 �,#�0#�"#$�5$ 0#<1$(� � !�$>$�0�&$�,(&�'$�%3�4$-.�7#�&�1�&"�3%�&$�*�0$&��&�3%�%� -(4$ "(1��4'3�"( 0$�%3��(�&�00$&&%�1�,3�6�( -��"�0( �5$�&$$ �(&�03 &$?�$ 0$�3%�"#$�033'$�("�3 �'#�13&3'#+�(-3'"$-�%3��"#�&�&��*$+.�=1&3)�(13 !�,�"#�"#$�&$�*�0$&�� "$!�("� !�"#$�!135(1�0("(13!�$)�"#$�&$�*�0$�0("$!3��$&�,$�$�(&�,$11�(!�$$-.��7#$�"#��-�'#(&$�03 &�&"$-�3 �"#$�!("#$�� !�3%�(11�� %3�4("�3 ��$1("$-�"3�$(0#�( -�3 $�3%�"#$�&$�*�0$&�� 01�-$-�� �"#$�!135(1�&$�*�0$&�0("(13!�$�� �3�-$��"3�4$(&��$�53"#�"#$���4("���"+�1$*$1�( -�"#$���'$�0$'"�3 �3%�(-3'"�3 �1$*$1�(&�-$"(�1$-�� �:#('"$��@.�=�&'$0�%�0�&$"�3%�"$4'1("$&�,(&�-$&�! $-�( -�(!�$$-�� �3�-$��"3�&("�&%+�"#�&� $$-.��/ 0$�(11�0�"�$&�&�54�""$-�"#$�"$4'1("$&�%�1%�11$-�%3��3 $�( -�$(0#�3%�"#$�&$�*�0$&�� "$!�("� !�"#$�!135(1�0("(13!�$)�"#$�( (1+&�&�'�30$&&�&"(�"$-A�(*$�(!$&)�-�%%$�$ 0$&�( -�-�&'$�&�3 &�,$�$�0(10�1("$-.�B �"#�"#�&)�"#$�;��3'$( �!135(1��$&�1"&�,$�$�35"(� $-�(&�,$11�(&�",3�4('&�&#3,� !�$(0#�0�"+�034'(�$-�,�"#�"#$��$%$�$ 0$.�C� (11+)�"#�&��$'3�"�!("#$�&�(11�"#$�� %3�4("�3 ��$1("$-�"3�"#$�&��*$+�( -�#3'$&�"3�&$"�"#$�'("#�%3��%��"#$��$>$�0�&$&�3%�"#�&�&3�".�=!(� )�(�-�(%"�3%�"#�&��$'3�"�,(&�'�$&$ "$-�"3�(11�'(�"�0�'( "&)�,#3�#(-�"#$���!#"�"3�'�$&$ "�03��$0"�3 �'�3'3&(1&)�( -�(�4$$"� !�"336�'1(0$�� �3�-$��"3�&#(�$�( -�-�&0�&&�"#$�%� (1��$&�1"&.�7#$�4(� �0#(�(0"$��&"�0�3%�"#$��$&�1"� !�4$"#3-313!+�#(&�5$$ ��"&�53""34�<�'�(''�3(0#)�DEFGEHFIHGJEKLKMNHDFNEHNEKHIOLPKQRIHSFTH3%�'�3*�-� !�(��&$%�1�"331�%3��0�"+�4( (!$�&.�73�"#("�(�4)�"#$�0311(53�("�3 �3%�0�"�$&�� �"#$�,#31$�'�30$&&�#(&�5$$ �3%�'(�(43� "��4'3�"( 0$.��;(0#�'#(&$�� 01�-$-�("�1$(&"�3 $�,3�6&#3'�,�"#�(11�'�32$0"�4$45$�&�( -�&$*$�(1�'�32$0"�033�-� ("�3 �4$$"� !&.�='(�"�%�34�(""$ -� !�( -�'(�"�0�'("� !�� �"#$�4$$"� !&)�$(0#�4$45$��%�34�(11�"#$�� *31*$-�0�"�$&�#(&�"(6$ �(0"�*$�'(�"�3 �$(0#�'#(&$�3%�"#$�&��*$+.�7#$+�#(*$�5$$ �(&6$-�"3�03 "��5�"$�"3�"#$�-$%� �"�3 �3%�� -�0("3�&)�?�$&"�3 (��$�'�$'(�("�3 )��$'3�"� !�&"��0"��$�3�"1� � !�( -�-("(�'�3*�&�3 .��=1&3)�(�0311(53�("�*$�,$5&�"$�,(&�&$"��'�&3�"#("�(11�&��*$+$-�0�"�$&�03�1-�&#(�$�(11�"#$�-30�4$ "&�!$ $�("$-�-��� !�"#$�'�30$&&�(&�,$11�(&�"3�6$$'��'�(�-�&0�&&�3 �%3��4�"3�03 "� �$�-$5("� !�"#$�4(� �"3'�0&�"�$("$-�-��� !�"#$�,3�6� !�!�3�'&�4$$"� !&.�C� (11+)�(�03 &"( "�0344� �0("�3 �#(&�5$$ �#$1-�5$",$$ �7$0# �0(1�:33�-� ("3��( -�'(�"�0�'( "&�"3�!��-$�"#$�,3�6�( -�&31*$�?�$&"�3 &�-��� !�"#$�'�06� !��'�'�30$&&$&.�

Page 90: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#$��%&�'�( � )*+�,�( �-%"#% �.�%&��/�0)*%#�!1�-('(� �2�(-��!*��$( '�.�(1�'���1�3'�'!4$�5� 61'2�*#�78�$9()":� 8!*-�$%1!�7;#!(1:� 8�*9�1�7<%*2!,:�8($4!%�7;#!(1:� 8(*"(19�!"�7=>:� 8%$%91!�7?'!$,:�01 -��.��7<�'��*$!1. :� @�$ (1A(�7B(1$!1.:� C($!1�7?'!$,:�C)*-(!�7;#!(1:� D(E�A!�7F*%!'(!:� �!$$(11�70 '%1(!:��)*(1�7?'!$,:� ����@!9)��7<�'��*$!1. :� G(�11!�76) '*(!:�HIJKLMNOMPIQRSTSUIVRMTSRSLWMMX ('��'�� ���/�#!*'(-(#!1'�-('(� Y����0)*%#�!1�-%)1'*(� �!*��#*� �1'�!'�'��� )*+�,Z����� !"#$��( �[)('��*�#*� �1'!'(+��%&�'���0)*%#�!1�)*4!1�!*�! Y�2('���!$&�%&�'���-('(� �4�$%19(19�'%�'���1%*'��*1�#!*'�%&�0)*%#��!1.�!1%'��*��!$&�$%-!'�.�(1�'��� %)'��*1�#!*'Z�C%*�%+�*Y�%)'�%&�'����/�-('(� Y�"%*��'�!1�/�\�!*���('��*�-!#('!$ �%*� �-%1.�-('(� �7]�-!#('!$ �!1.� � �-%1.�-('(� :�!1.�'���*� '�!*���('��*�"�.()"�%*� "!$$�-('(� Z����(*�_`aY�1)"4�*�%&�-(+($� �*+!1' Y�#�*-�1'!9��%&��%) ��%$.�?1'�*1�'�!--� �%*��.)-!'(%1!$�$�+�$�( �bcdefghijfklmnodlplnkqfroemlstfujnfdbgrcnvdfdhlubwlcluxfpeofujnfdhomnxyf�?1�'%'!$Y�'����/�-('(� �#!*'(-(#!'(19�(1�'�( � )*+�,�*�#*� �1'5���Y�z{Y{]|�(1�!4('!1' Y�9%(19�&*%"�'���� Y� ]�(1�!4('!1' �&*%"�D(E�A!Y�'%�'����Y|{{Yz|{�(1�!4('!1' �&*%"�G(�11!Z�61�!+�*!9��#%#)$!'(%1�#�*�-(',�2%)$.�4��%&�|{}Z/�|�(1�!4('!1' Z�}� Y]] �-(+($� �*+!1' Y�9%(19�&*%"�'��� //�-(+($� �*+!1' �%&�D(E�A!Y�%*�'����Y{���%&��!$$(11�!1.�01 -��.��'%�'���]zZ|}]�%&�@�$ (1A(�%*��+�1�'���/{Z����%&�8(*"(19�!"Z�~sfbmnobtnf���frnofibrlubfepf��q����qftelstfpoegfujnf��q�����#�*�-!#('!�%&�D(E�A!�'%�'���|�Y����frnofibrlubfepf�notnsy������

Page 91: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ������ !�"#$%&'($#)"(!*+ ��,-�./�012�03�452�367/4�/428/9�452�8:746;68:14�;6462/�<272�:/=2>�40�?6/4�452�@:61�/27A6;2/�452B�03327�4570CD5�2?2;47016;�@2:1/E�F5:4�61307@:4601�/50<2>�45:4�104�:??�;6462/�>2?6A27�07�5:A2�452�GHIJKLMINOKLOPKQRSISTPSUOIKOVSWMXSLOIJSOUGQSOUSLXMPSUOIKOPMIMYSTUZO[GPJOPKHTILN\UOJMUIKLNOGTVO2]827621;2�61�2 _0A271@214�>63327/�370@�64/�126D5 0C7/�61�@C?468?2�<:B/E�F527230729�45272�6/�:�122>�03�5:A61D�:�;0@@01�?6/4�03�aC7082:1�/27A6;2/�45:4�;:1�`2�C/2>�:/�:�@2:/C72@214�7232721;2E�b 645�45:4�:6@9�:�_?0 :?�c6462/�c:4:?0DC2�<:/�;72:42>�0C4�03�:??�452�61307@:4601�72;0??2;42>E��F52�_?0 :?�c6462/�c:4:?0DC2�61421>/�40� 761D�40D24527�:??�452�;0@@01�/27A6;2/�370@�:??�452�;6462/�:/�<2??�:/�452�/82;636;�?0;:?�/27A6;2/�45:4�<272�<0745�@21460161D9� 045� B�45267�/C;;2//�07�6110A:4601E�d1;2�:??�/27A6;2/�<272�;0??2;42>9�452B�<272�D70C82>�:;;07>61D�40�@C16;68:?�:;4601�:72:/E�F52�72/C?4�03�452�870;2//�<:/�:�?6/4�03�1612�;:42D0762/E�.4�452�/:@2�46@29�452/2�;:42D0762/�52?8�40�2A:?C:42�;6462/�/50<61D�/27A6;2/�>2?6A27B�>6332721;2/�:;70//�aC7082E�e1�404:?9�26D54B 012�/27A6;2/�5:A2� 221�6>2146362>�:1>�D70C82>�6140�1612�@:61�;:42D0762/9�?6/42>�`2?0<f� gh ijkjlmnopqmnrsrmtmnkf��u�/27A6;2/�:72�;01/6>272>�61�456/�;:42D07B�>2A042>�40�GIILGPIOPMIMYSTU\OGIISTIMKTOIKOSWSPILKTMPOPJGTTSWUOGTVOIKOvKUISLOPMIMYSTU\O8:746;68:4601E��F52�/27A6;2/�61;?C>2>�61�456/�;:42D07B�21;0C7:D2�452�72?:4601/568�`24<221�452�;646w21/�:1>�45267�:>@616/47:4601/�870@0461D�07�`261D�01�452@/2?A2/�:�8:45�40�2 x2@0;7:;B�gh yz{||}~~�|�f��u�/27A6;2/�:72�61;?C>2>�61�456/�47:1/A27/:?�;:42D07B�72?:42>�40�215:1;2�452�;014:;4� 24<221�452�;646w217B�:1>�64/�@C16;68:?64BE�F56/�D70C8�03�/27A6;2/�61;?C>2/�;64B�<2 �8:D2/9�827/01:?�30?>27/�07� C??2461/E�gh ����{���|f����/27A6;2/�:72�61;?C>2>�61�456/�;:42D07B�72?:42>�40�/:46/3B�PMIMYSTU\O[VHPGIMKTGWOTSSVUZO�JMUOvGQMWNOKvOUSLXMPSUOMTPWHVSUOUSLXMPSUOLSWGISVO40�;56?>7219�:>C?4�:1>�A674C:?�?2:7161D�/27A6;2/9�/27A6;2/�72?:42>�40�;C?4C7:?�870�2;4/�:1>�?6 7:762/E�gh ���~���}|��������|}��f���/27A6;2/�:72�;01/6>272>�61�456/�;:42D07B�>2A042>�40�52?8�;646w21/�0 4:61�:��0 �07�@:=2� C/612//�870;2>C72/�2:/627E��27A6;2/�>2A042>�40�72;7C64@2149�421>2761D�:1>�/@:??� C/612//�870@04601�:72�61;?C>2>�61�456/�>6A6/601E�gh �|����|�}|�f� �� /27A6;2/� :72� :>@6442>� 61� 456/� ;:42D07B� 72?:42>� 40�a1A6701@214:?�;:72E�F56/�D70C8�61;?C>2/�/27A6;2/�72?:42>�40�452�@:61421:1;2�03�452�;0@@C1:?�870827462/�:/�<2??�:/�40�452�21;0C7:D2@214�:1>�:;�C6/64601�03�5: 64/�/C8807461D�/C/4:61: ?2�C7 :1�D70<45E�gh ���}���~}f��u�/27A6;2/�:72�61;?C>2>�61�456/�;:42D07B�>2A042>�40�2:/2�452�12;2//:7B�47:1/:;4601/�:@01D�;646w21/�:1>�@C16;68:?6462/�72?:42>�40�452�?632�;B;?2�03�:�827/01�?6A61D�61�452�;64B9�/C;5�:/��27/01:?�c274636;:42/9�>6D64:?�>0;C@214/�@:1:D2@214�/B/42@/�07�;21/C/E�gh ����{~��{�}f����/27A6;2/�:72�;01/6>272>�61�456/�;:42D07B�72?:42>�40�6142D7:42>�/27A6;2/�45:4�:72�:A:6?: ?2�370@�8C ?6;�52:?45�:1>��0;6:?�c:72�870A6>27/E�F56/�D70C8�:?/0�61;?C>2/�50C/61D�/27A6;2/�:1>�;:72�307�8208?2�<645�/82;6:?�122>/�

Page 92: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#$�%�&'�#�($)�*(�!+$��,$*,#$-.�/0 1234567289�:� $(;�&$ �!($�!�<�==$���)�='� �&!=$+*(>��$;*=$��=*� !=� ?>�@ABACDEFGHIJKALABMHEDDNGOHPQAGHRSIALMHTUJVWHAE@LVNDGHGDUXA@DGH($#!=$��=*�$! $�='$�,'> �&!#�<*"�#�=>�='(*Y+'�='$�&�=>�">�;!(�*Y �<$!) �*?�Z(!) ,*(=.�/0 [2\34]6 344_4 9�a� $(;�&$ �!($��)&#Y�$���)�='� �&!=$+*(>�($#!=$��=*�#!)��Y $�!)��,(*,$(=>�($#!=$��� Y$ �! �,$(<�= �?*(�&*) =(Y&=�*)%�"Y�#��)+�($+Y#!=�*) %�&$(=�?�&!=$ �*(�#!)��!&bY� �=�*).�c*($*;$(%�*)&$�&!=$+*(�d$�%�!##�='$� $(;�&$ �'!;$�!# *�"$$)���;��$��!&&*(��)+�=*�='$�)Y<"$(�*?�&�=�$ �='!=�!($�*??$(�)+�='$<%�='Y �!)!#> �)+�='$�(�,$(&$)=!+$�*?�&*;$(!+$.��e$(;�&$�&*;$(!+$�� ��$?�)$��! �='$�,$(&$)=!+$�*?� Y(;$>$��&�=�$ �='!=�!($�,(*;���)+�!� $(;�&$.�Z'$($?*($%��?�!�&*)&($=$� $(;�&$�� �*??$($��">�!##�*?�='$�&�=�$ ��= �&*;$(!+$�f*Y#��"$��ggh.�Z'� �,!(!<$=$(�f�##�"$�?Y(='$(�Y $���)�='� � =Y�>�=*�! $ �'*f�&*<<*)�!� $(;�&$�� �!<*)+�='$� Y(;$>$�� &�=�$ .� c*($*;$(%� &*;$(!+$� !=� &!=$+*(>� #$;$#� !# *� �)?*(< � !"*Y=� &*<<*)�Y)�$( =!)��)+��)� Y&'�&!=$+*(>.��i&&*(��)+�=*�*Y(�<$='*�*#*+>%�&*;$(!+$�f! �! $ $��">� $#?j$;!#Y!=�*).�k)&$�='$�l#*"!#�e$(;�&$ �m!=!#*+Y$�f! �&($!=$�%�!�bY$ =�*))!�($�f�='�='$�$�+'=>j*)$� $(;�&$ �f! � $)=�=*�$;$(>�&�=>%�f'�&'�f! �! n$��=*�,*�)=�*Y=��?�='$>��$#�;$($��='!=� $(;�&$�*(�!� �<�#!(�*)$.��i&&*(��)+�=*�='$�,($;�*Y �&*) ��$(!=�*)%��?�='$� $(;�&$�� �*??$($��">�ogh�*(�<*($�*?�='$�,!(=�&�,!)=�&�=�$ ��=�� �#!"$##$��! �e=!)�!(��e$(;�&$.�p?�)*=%��=�� �&!##$��i���=�*)!#�e$(;�&$%�! � '*f)��)�=!"#$�a9]]e$(;�&$�m*;$(!+$�q�og�h� e=!)�!(��e$(;�&$� or� $(;�&$ �e$(;�&$�m*;$(!+$�s�og�h� i���=�*)!#�e$(;�&$� �t� $(;�&$ �uvwxyz{|z}y~���y��z��������]e=!)�!(�� $(;�&$ %�f'�&'�($,($ $)=�:oh�*?�='$���$)=�?�$�� $(;�&$ %� '*f�&*<<*)��)=$($ = ��)��VUJWDSEHIVEA@AWSLABADGFHGDUXA@DGHNDWLJMIDEBOH�BHAGHAIWJUBSEBHBJHUDISU�HBQSBHBQDHGBSENSUNH $(;�&$ �&*)&$,=��*$ �)*=�($,($ $)=�!�+(*Y,�*?� $(;�&$ �='!=�!##�&�=�$ �<Y =��$#�;$(.�p=�� �)*=�!�+(*Y,�*?�<!)�!=*(>��Y(*,$!)�$je$(;�&$ .�k)�='$�&*)=(!(>%��=�!�< �!=��$ &(�"�)+�+$)$(!#�$jl*;$()<$)=�"$'!;�*Y(�!<*)+�='$� Y(;$>$��&�=�$ �Y �)+�!�;$(>� �<,#$�&!=$+*(�d!=�*).���EHBQDHJBQDUHQSEN�HBQDHSNNABAJESLHGDUXA@DGFH@JE@DWB�H���HJRHBQDHANDEBARADNHGDUXA@DG�HQATQLATQBGH��??$($)&$ �*)�='$�&�=>� �(*#$�j&*<,$=$)&�$ j�!&(* ��Y(*,$!)�&*Y)=(�$ �!)�� '*f �='$��))*;!=�*)�&!,!"�#�=�$ �*?�='$�#*&!#�,Y"#�&� $&=*(.��

Page 93: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#$%&'#(�")*�+,-.�/*+01�022*3-�4�5*.4,+*5�6,*1�02�.)*�$+0/4+�(*36,7*-� 4.4+089*�:*3�74.*803;��'.�,<7+95*-�4�-)03.�5*-73,:.,0<�02�*47)�-*36,7*�70<-,5*3*5=�/0.)�-.4<5435�4<5�455,.,0<4+���>?>?�@ABACDEFGHDEIJIDKDEB�4��LMNOPNQPRSTMTUVOLWRVOXNXVYVOMRLVQZTSVL[�\� ] _ a]bcdefagfhdij akhllbdmlaneehllo�-*36,7*�.)4.�022*3-�.)*�:0--,/,+,.;�.0�20++01�.)*�p**.,<8-�02�.)*�7,.;�709<7,+�+,6*�0<�.)*�1*/-,.*��'.�p4;�4+-0�,<7+95*�.)*�:0--,/,+,.;�.0�70<-9+.�70pp9<4+�4<5�p9<,7,:4+�5*7,-,0<=�3*45�5079p*<.-�4<5�6,*1�:0+,.,74+�,--9*-�5*/4.*5�1,.),<�.)*� ,.;� 09<7,+��'.�p4;�,<7+95*� ,<203p4.,0<� 3*8435,<8� :3*-*<.4.,0<q464,+4/,+,.;� 02� 4++� :0+,.,74+�p*p/*3-��\� rstuvwvxsuvyz{rtyw|}}|}o�9-*�02�7+4--,74+�4-�1*++�4-�*+*7.30<,7�.00+-�,<�4�70++4/034.,6*�:4.)�.0�.)*�5*6*+0:p*<.�02�<*1�203p-�02�*~:43.,7,:4.,0<��'.�p4;� ,<7+95*� 5*6*+0:p*<.� 02� 43*4-� 5*60.*5� .0� 0<~+,<*� -.95,*-=��9*-.,0<<4,3*-=�:30�*7.-�4<5�0<~+,<*�2039p-��\� �vu�{�y�zwv�{�v}u|z}��0<~+,<*�-*36,7*�.0�p4�*�70p:+4,<.-�4<5�<0.,2,74.,0<-�.)3098)�.)*�5,8,.4+�5*-���")*�7,.,�*<�p4;=�203�*�4p:+*=�2,+*�4�70p:+4,<.�03�p4�*�4�<0.,2,74.,0<�4/09.�.)*�*<6,30<p*<.�03�*6*<�.)*�,++*84+�70<-.397.,0<�02�4�/9,+5,<8��")*3*�p4;�/*�4�-*36,7*�+*6*+�483**p*<.�,<�.*3p-�02�:306,5,<8�4<�4<-1*3�.0�.)*�7,.,�*<��\� �vu�{�tw�v�|}{�yz}��usuvyzo� ,.,�*<-�0<~+,<*�477*--=�9-94++;�.)3098)�.)*�437),6*�)0p*:48*=�.0�.)*�5,8,.4+,�*5�437),6*-�230p�.)*�7,.;�� ,.,�*<-�p4;�3*�9*-.�.0�6,*1�4<�437),6*��:3*6,09-+;�/301-*5�.)3098)�.)*�70<.*<.�+,-.��/;�2,++,<8�4�203p�0/.4,<4/+*�,<�.)*�3*45,<8�)4++�02�.)*�437),6*�/9,+5,<8�03�0<~+,<*���R���R��������WR����������\� �|}|t�suvyz{�yt{ �zvwvxs�{¡�v�¢vz£}o�4�7,.,�*<�1,-),<8�.0�3*-*36*�4�p9<,7,:4+�/9,+5,<8�203�4<;�-:*7,4+�*6*<.�p4;�6,-,.�.)*�p9<,7,:4+�1*/-,.*�4<5�70<.47.�.)*�-*36,7*�,<�7)438*�02�3*-*364.,0<-��"),-�-*36,7*�p4;�/*�7)438*5�¤�\� ¥vy}¦}�¤!9.0p4.,7�p47),<*-�.)3098)�1),7)�4<;�3*-,5*<.�p4;�477*--�7,.;�709<7,+�,<203p4.,0<�-97)�4-�*~(*36,7*-=�.)*�48*<54�03�),-q)*3�:*3-0<4+�20+5*3��")*3*�p4;�4+-0�/*�4�23**�*~p4,+�-*36,7*��")*�-*36,7*�p4;�,<7+95*�4+-0�4<�0:.,0<�.0�:937)4-*�.,7�*.-�203�4<;�*6*<.�.)4.�.4�*-�:+47*�,<�.)*�7,.;��")*��,0-�-�p4;�/*�+074.*5�,<�:9/+,7�43*4-�+,�*�+,/343,*-=�p43�*.-�03�p9<,7,:4+�/9,+5,<8-��/��NPPTMT§ON RSTMTUVOLWRVOXNXVYVOMRLVQZTSVL[�\� �y}u{sz¢{©y�z¢o�0<~+,<*�-*36,7*�.)3098)�1),7)�4�7,.,�*<�p4;�-9/p,.�4�3*�9*-.�.0�3*.3,*6*�4�+0-.�.),<8��")*��94+,2,*5�022,7*3�1,++�70<.47.�),pq)*3�203�p03*�,<203p4.,0<�4<5q03�.0�8,6*�/47��.)*�0/�*7.��\� ª|wyz¢{�v�|{«}�sz¢�¤���¬­����R�®R���R�­������¬�� W�R°��±R��R����6,3.94+�103+5�02�(*70<5�²,2*���

Page 94: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ��� � !"#$%&'()$*+,-./0123451/6735489/561/:45;<3/2=79>/56?5/43>/?/3451/:785?48489/@ABCDE@�AFG@HAI�JK�LJMMAFGE@NO�PAIL@HQGHJFI�JK�ARAFGIO�J@�JGSA@�MEGA@HED�ICLS�EI�B@EQSHLI�J@�RHPAJT�UFG@HAI�K@JM�LHGHVAFI�MEN�WA�QDELAP�EKGA@�E�LSALX�KJ@�JKKAFLAI�J@�HFEQQ@JQ@HEGA�DEFBCEBAT��� YZ[\]_ ab�PH@ALG�LJMMCFHLEGHJF�DHFA�GS@JCBS�cSHLS�E�LHGHVAF�LEF�dCA@N�E�MCFHLHQED�cJ@XA@�@ABE@PHFB�MCFHLHQED�IA@RHLAI�EFP�JWGEHF�EF�HMMAPHEGA�@AIQJFIAT���efe��ghijjk��lj��ET�mnopqorq�stoppUuuvpw�mUrxvsUmby�� � !"#$z{|}a~[�a��MEHF�cAWIHGA�JK�GSA�MCFHLHQEDHGNT�vG�HI�GSJCBSG�GJ�WA�GSA�IJDA�AFG@EFLA�GJ�EDD�GSA�JGSA@I�IHGAI�EFP�IA@RHLAI�JKKA@AP�WN�GSA�LHGN�LJCFLHDT�vG�HI�EDIJ�GSA�K@JFG�HMEBA�JK�GSA�LHGNTy�� � ! �-�$#% �-�$(�+'% �('�-�$b� HG� ISJCDP� HFLDCPA� QA@IJFEDHVAP� CIAKCD�HFKJ@MEGHJF�KJ@�GSA�LHGHVAFIT�nSA�PAGEHDI�HFLDCPAP�HF�GSA�KJDPA@I�MEN�RE@N�K@JM�JFA�MCFHLHQEDHGN�GJ�EFJGSA@T�mJMA�A�EMQDAI�JK�Q@JRHPAP�HFKJ@MEGHJF�cJCDP�WA�KHILED�HFKJ@MEGHJFO�JF�DHFA�QENMAFGI�J@�ELLAII�GJ�QEHP�G@HWCGAIT��� Y|`�� \[\_|`z�`]_ az�a}[_�z�Z[\�z~�(�-����GJJD�KJ@�LHGHVAFI�GJ�MEXA�@AdCAIGI�JF�EFN�IA@RHLA�J@�HFKJ@MEGHJF�LJFLA@FHFB�GSA�sHGN�tEDD�EFP�HGI�IA@RHLAIT�nSHI�LJFICDGEGHJF�LEF�WA�MEPA�GS@JCBS�IARA@ED�LSEFFADIT�oLLJ@PHFB�GJ�GSA�LSEFFAD�LSJIAF�GSA�EFIcA@�LEF�WA�HMMAPHEGA��LSEG�J@�QSJFA�T�nSA@A�MEN�WA�E�IA@RHLA�DARAD�EB@AAMAFG�HF�GA@MI�JK�Q@JRHPHFB�EF�EFIcA@�GJ�GSA�LHGHVAFT��� ��^a\_ ��a��a\\a���WCDDAGHF�PADHRA@AP�JF�DHFA�GJ�GSA�LHGHVAFISHQ�LJFGEHFHFB�EFN�XHFP�JK�IA@RHLA�J@�ELGHRHGN�JK�BAFA@ED�HFGA@AIGT�nSA�IA@RHLA�MEN�WA�EREHDEWDA�CFPA@�ICWIL@HQGHJFT�nSA�LHGHVAF�MEN�LSJJIA�GSA�LJFGAFGI�SA�ISA�cEFGI�GJ�@ALAHRA�EFP�HGI�@ABCDE@�@ALC@@AFLAT��� ���z[`�z�_�\�[ z� _�Z\��nSA�wvm��wAJB@EQSHL�vFKJ@MEGHJF�mNIGAMI��IA@RHLA�MEN�Q@JRHPA�HFKJ@MEGHJF�@ABE@PHFB�LHGN�MEQI�EFP�MEN�HFLDCPA�EF�EPP@AII�DJLEGJ@�IA@RHLAT�nSA�MCFHLHQEDHGN�MEN�Q@JRHPA�JGSA@�XHFPI�JK�HFKJ@MEGHJF�HFGAB@EGAP�cHGS�E�wvm�MEQO�KJ@�HFIGEFLAO�SJIQHGEDI�A�ELG�IHGCEGHJFO�ILSJJDI�E@JCFP�E�CIA@�SJCIAO�LSC@LSAI�J@�IG@AAG�ME@XAGIT�nSA�RH@GCED�KDHBSG�HI�EF�EPREFLAP�IHMCDEGJ@�GS@JCBS�cSHLS�GSA�LHGHVAF�LEF�QA@KJ@M�E� �qO�¡�q�J@�ARAF���¢/£4345/57/561/:45;<3/¤?48/?55¥?:54783/7¥/561/067=1/:45;¦y�� Y��\|}_§a�z¨ a©zª[�a�b�QJIIHWHDHGN�GJ�LCIGJMHVA�EFN�MCFHLHQED�cAWIHGA�?::7¥«489/57/:454¬183</4851¥1353/?8«/4¤­?4¥¤1853¦/®643/:78:1­5/:?8/21/£1¥;/cHPAT�mJMA�A�EMQDAI�JK�cAWQEBA�LCIGJMHVEGHJF�LJCDP�WAb�QJIIHWHDHGN�GJ�LSEFBA�GSA�KJFG�IHVA�JK�GSA�cAWQEBAO�QJIIHWHDHGN�GJ�LSJJIA�cSHLS�HFKJ@MEGHJF�HI�ISJcF�HF�GSA�MEHF�cAWIHGA�ELLJ@PHFB�GJ�GSA�CIA@��GS@JCBS�GSA�CIA�JK�LJJXHAI��J@�GSA�QJIIHWHDHGN�GJ�DHIGAF�GJ�GSA�HFKJ@MEGHJF�ISJcF�JF�GSA�IL@AAFTzy�� ª�|�a���a�z � |�}[\_|`z �_\Zz |\Za�z ª�© _�z °�} _ _�\�[\_|`�z �|�[ �z�a�_|`[ z[`�z`[\_|`[ ±��HFKJ@MEGHJF�EWJCG�Q@JLAPC@AI�cHGS�JGSA@�QCWDHL�?«¤48435¥?54783/36708/78/561/¤²84:4­?=45;<3/0123451¦/³5/¤?;/48:=²«1/561/QJIIHWHDHGN�GJ�PJcFDJEP�GSA�FAAPAP�KJ@MI�J@�@APH@ALG�GSA�MCFHLHQED�IHGA�GJ�GSA�PAIH@AP�IA@RHLATy

Page 95: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ��� � !"#$%&'"()*&+,-*./�0123451�6786�9:;<=�8<<:9�5464>1?0�6:�018257�8?@�A;B<45�B;4<=4?C�:2�0123451�855:2=4?C�6:�8�DEF�G8AH�E6�5:;<=�8<0:�B1�A:004B<1�6:�C431�I11=B85J�=42156<@�6:�8�512684?�A<851�:I�0123451H�BH�KLLEMENOKP�QRKOOSPPEOD�FSTUEQSFVW�� XY,&*Z*)%&%" [-\'&*%/W A120:?8<4>1=� ;01I;<� 4?I:2G864:?� I:2� 1?612A24010H�E?I:2G864:?�4?5<;=1=�4?�671�I:<=120�G8@�382@�I2:G�:?1�G;?454A8<46@�6:�8?:6712H�F:G1�1]8GA<10�:I�4?I:2G864:?�9:;<=�B1V�I4058<�4?I:2G864:? �:?_<4?1�A8@G1?60�:2�855100�6:�8<218=@�A84=�624B;610H��� ! !"a-*"b),)c&Y%/W5464>1?074A�5:?6856�672:;C7�671�;01�:I�671�F7:26�d1008C1�F123451�eFdFfH�M71�0123451�5:;<=�B1�;01=�6:�2151431�=4II121?6�J4?=0�:I�4?I:2G864:?�=42156<@�:?�671�G:B4<1�A7:?1H�M71�0123451�G8@�B1�;A:?�21g;106�8?=�G8@�21g;421�0:G1�J4?=�:I�4=1?64I45864:?H��hihjklmnopqr�s�8H�FMKOLKTL�SLtQKMENO�FSTUEQSFV��� u*&v!+w--\" bw)\'*&Y" X'x+$,)-YV� :?_<4?1� 0123451� 6786� G8@� 4?5<;=1�4?I:2G864:?�8B:;6�A21_yz{||}~y���yz�����|�������{���|yy���}�����|�����|�������A8@� 671� 4?0524A64:?� :?_<4?1H� M71� 0123451� G8@� 21g;106� 8?@� J4?=� :I�4=1?64I45864:?�I2:G�671�;012�8?=�G8@�B1�4?61C2861=�9467�8�DEF�G8AH��� bw)\'*&Y"X'x+$,)-YV�:?_<4?1�0123451�6786�G8@�4?5<;=1�4?I:2G864:?�8B:;6�yz{||}~y���yz�����|�������{���|yy���}�����|�����|������������{����yz���64:?�:?_<4?1H�M71�0123451�G8@�21g;106�8?@�J4?=�:I�4=1?64I45864:?�I2:G�671�;012�8?=�G8@�B1�4?61C2861=�9467�8�DEF�G8AH��� �'x\,"X'x+$,)-Y/�:?_<4?1�0123451�6786�G8@�4?5<;=1�4?I:2G864:?�8B:;6�8=;<6�5:;2010�8?=�8564346410�B:67�4?�A;B<45�8?=�A243861�51?6210�8?=�671�A:004B4<46@�6:�A12I:2G�8?=�A8@�671�4?0524A64:?�:?_<4?1H�M71�0123451�G8@�21g;106�8?@�J4?=�:I�4=1?64I45864:?�I2:G�671�;012�8?=�G8@�B1�4?61C2861=�9467�8�DEF�G8AH��� b--Z&*$,)-Y"�-�&v!+w--\/�:?_<4?1�0A851�I:2�A821?60�61857120�8?=�06;=1?60�6:�1]578?C1�4?I:2G864:?�8B:;6�057::<�13:<;64:?�e7:G19:2J �8564346410�:2�1]8G0fH�M71�0123451�G8@�21g;106�8?@�J4?=�:I�015;21�4=1?64I45864:?�I2:G�671�;0120H��� �)�*$*)&%/�:?_<4?1�=868B801�078A4?C�8�97:<1�4?61C2861=�5868<:C;1�:I�8<<�671�A;B<45�<4B282410�4?�671�546@H�M71�5464>1?�9:;<=�:?<@�7831�6:�G8J1�8�;?4g;1�018257�6:�I4?=�8�B::J_QL_LUL_G8C8>4?1�8?=�9:;<=�7831�671�578?51�6:�2101231�46�:?_<4?1�:2�1]61?=�<:8?�A124:=0H��� bx\,x*$\"�&*),$�&"!),&%/�:?_<4?1�4?I:2G864:?�8?=�B::J4?C�6:�512684?�G;01;G0�eG;?454A8<�:2�?:6f�:2�5;<6;28<�04610�8?=�8564346410H�M7121�G8@�B1�671�A:004B4<46@�6:�A8@�671�210123864:?�:?_<4?1H��� [xY'%"a-*"bx\,x*$\"u*-�&+,%V�A:004B4<46@�6:�I4?=�4?I:2G864:?�8?=�8AA<@�:?_<4?1�I:2�8?@�J4?=�:I�I;?=�I:2�5;<6;28<�A2:�1560H�M7121�G8@�B1�671�A:004B4<46@�6:�=4C468<4>1�671�8<218=@�0;BG4661=�4?I:2G864:?�:I�67:01�:2C8?4>864:?0�6786�@182<@�8AA<@�I:2�671�I;?=0�0:�6786�?:�1]628�1II:26�40�?11=1=H��� !x��&*"b$�Z%"-*"%)�)\$*/�:?_<4?1�0123451�21C82=4?C�4?I:2G864:?�8B:;6�

Page 96: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !�"#�$%�&'!�()*%�'!�+ ,#- !%.�/0 �% !1)2 ��#3�),24�* �+0 �$'%%)5)4)+3�+'��#( �#,�),%2!)$+)',�#,*�+'��#( �+0 �$#3� ,+�',64), .789 :;<=>?@A;<@BCDE=F@>EG@BH=<>IJ<<KIJ?><@LI=KCK=KDFM�',64), �N 5%)+ �+0#+��#3�),24�* �#44�+0 �#2+)1)+) %�$ !&'!� *�),�+0 �2)+3.�O,3� 1 ,+�'!-#,)P !��#3�5 �#54 �+'�! -)%+ !�+0 )!� 1 ,+%�),�+0 �$'!+#4.�/0 ! �2#,�5 �#�&)4+ ! *�% #!20��+)4)+3�+'�&),*�#,3� 1 ,+.�75.�OQQR/RSTOU�VQW"O/RST��VXYR"V�M789 ZDGK>@[;<\F];^_7% !1)2 �'&& ! *�),�+0 �, )-05'�!0''*%�'&�+0 ���,)2)$#4)+3.�/0 �2)+3��#3�%�%+#),�),)+)#+)1 %�&!'��2)+)P ,%�5�)4*),-�� *)#6$!' 2+%�#5'�+�+0 )!�'N,�, )-05'�!0''*�#,*�%0#! �)+�N)+0�4'2#4�! %)* ,+%.789 a<>?@bKF=;<c@:<DFD<C>=K;E_�',64), �%)+ �N0 ! �2)+)P ,��#3� d20#,- �),&'!�#+)',�#5'�+�+0 �$#%+�'&�+0 �2)+3.�e'!� d#�$4 f�#��% !��#3��$4'#*�#,�'4*�$)2+�! �#,*�'+0 !��% !%��#3��#( �2'�� ,+%�',�)+.�/0)%�% !1)2 ��#3�5 �#1#)4#54 ��,* !�%�5%2!)$+)',.789 g>hDF@>EG@F ;<=F@?DEGKEi_�! % !1#+)',�'&� j�)$� ,+�&'!�-#� %�'!�%$'!+%�',64), �1)#�+0 �*)-)+#4�* %(.�/0 ! �2'�4*�5 �+0 �$'%%)5)4)+3�+'�5'!!'N�+0 �%$'!+%��#+ !)#4�)&�, * *.789 DklD><EKEiM�+ 20,'4'-3�%�$$'!+ *� *�2#+)',�#+�#,3�4 1 4�N0 ! �+0 �� *)���'&� ),%+!�2+)',� )%� 2'�$�+ !� + 20,'4'-3f� $#!+)2�4#!43� ),1'41),-� *)-)+#4�+ 20,'4'-) %.�R,�%'� �),%+#,2 %f�,'�),6$ !%',�),+ !#2+)',�+#( %�$4#2 .�V4 2+!',)2�4 #!,),-�2'�4*�5 ��% *f�&'!�),%+#,2 f�+'�'&& !�#*�4+� *�2#+)',�2'�!% %�N)+0'�+�&#2 6+'6&#2 �#++ ,*#,2 .@�mnmom�pqr��sqptu�vtw��x�ytp���#.��/OTQOXQ�Vz{US|zVT/�OTQ�}W�RTV����VXYR"V�M�89 ~DI<JK=hDE=@��;�F@KE@ <KC>=D@FDI=;<��� '56% #!20��+)4)+3�$!'�'+ *�53�+0 �2)+3.�W% !%�N'�4*�0#1 �+0 �$'%%)5)4)+3�+'�*'N,4'#*�+0 �! 4 1#,+�&'!��'!�+'�#$$43�&'!�#�'5�',64), .�O4%'f�+0 �% !1)2 �2'�4*�#4%'�5 �),+ -!#+ *�N)+0�'+0 !� �$4'3� ,+�% !1)2 %�&!'��+0 �2)+3�'!�&!'��'+0 !�$�54)2�#*�),)%+!#+)',%.�89 :J�?KI@�DEGD<F@��;�F@KE@ J�?KI@FDI=;<���'56% #!20��+)4)+3�+'�),&'!��#5'�+�'$ ,�'!�3 +�+'�'$ ,�$�54)2�+ ,* !%.�/0 ! �2'�4*�5 �#1#)4#54 �&'!�%�+'�*'N,4'#*.�O4%'f�+0 ! �2'�4*�5 �+0 �$'%%)5)4)+3�+'�#$$43�',64), .�89 :;<=>?@ A;<@ �;h^>EKDF�� %$ 2)&)2� N 5%)+ � &'!� 2'�$#,) %� 2',+#),),-�),&'!�#+)',�',f�&'!�),%+#,2 f�+#d %f�& %�'!�$ !+), ,+�$!'2 *�! %.�" !+#),�% !1)2 %�2'�4*�5 �&�443�$ !&'!� *�',64), .�/0 ! �N'�4*�#4%'�5 �#�,'+)&)2#+)',�% !1)2 �&'!�2! *)+%f� *�2#+)',�$!'-!#�%�'!� 1 ,�$�54)2�+ ,* !%.�89 :;<=>?@A;<@BE=<D <DEDJ<F��%$ 2)&)2�N 5%)+ �&'!�5!#,*�, N�'!�3 +6+'65 �2'�$#,) %.�{!'2 *�! %�2'�4*�5 �&�443�$ !&'!� *�',64), f�),24�*),-�5�%), %%�! -)%+!#+)',.�/0 ! �N'�4*�#4%'�5 �#�,'+)&)2#+)',�% !1)2 �&'!�2! *)+%f� *�2#+)',�$!'-!#�%�'!�$�54)2�+ ,* !%.�89 aEk?KED@�JFKEDFF@~DiKF=<>=K;E_�% !1)2 �+'�$ !&'!��#�! -)%+!#+)',�&'!�#,3�+3$ �'&�2'�� !2)#4� ,+ !$!)% �2'�$4 + 43�',64), .�/0)%�% !1)2 ��#3�#4%'�),24�* �

Page 97: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !�" #�� �$%&���%'�(�!%)$�#�� #*+"#�,-.,�/0012134/5�6785397642�/40�:;<146<<�<6=>1?6<@-AB CDCEFGHEIJKLMNKK@�OP���<7<�(�PQ"R��� �"#O P$�P�S"(��P�T�+ R%+�.U("#�((�%. U��$U#"R"!%+�!P SP%$(V�P�R�#�+)� !�#�T�!U.+"R���#T�P(V�%Q%"+%."+"�)� O�RP�T"�(V��Q�#�(� P��TUR%�" #� !! P�U#"�"�(,-AB WXYZ[\] Z_ a�(U!! P�"Q��� +�O P������#�P�!P�#�UP(�b� �+%R&� O�&# b+�TS�� P��'!�P"�#R�(�"#����"P�.U("#�((�%P�%,�2�"(�RP�%�"Q��(!%R��(� U+T�S"Q��%#� !! P�U#"�)�� �O"#T�RP�%�"Q��( +U�" #(�� �!P .+�$(� P�O"#T�#�b�"T�%(�O P�T�Q�+ !"#S�#�b�!P TUR�(�%#T�(�PQ"R�(,--�cdcefghijk�hlghm�%,�<2/40/=0�64>1=347642�<6=>1?6<@-AB nNopoqLMrEstHuKa-"#O P$%�" #�%Q%"+%.+�� #*vwxyz{y|}{~wx|z{y���vwx|z�}{���z%P�%(�%#T��"$�(� O�R ++�R�" #�%#T�%#)� ���P�U(�OU+�"#O P$%�" #,�2���(�PQ"R��$%)�%+( �.��"#��SP%��T�b"���%��1<�$%!�%#T�$%)�P��U"P��"T�#�"O"R%�" #,-AB ������]����Y���]�Z�_�Y_Z_�Y]�[��YY�[�]��ZX[�]������]����X�_�[��a� #*+"#���y{�w�yz��z{y��{�z~}�}|yz�xz}z���vw�z�{��y{��z�|{}��w�w z���¡�vy�¢£¤z¥w�w¦yxzR U+T�P��U�(��O��T.%R&�# �"O"R%�" #,�/�(�PQ"R��+�Q�+�%SP��$�#��"#���P$(� O�!P Q"T"#S�%#�%#(b�P�� �����R"�"§�#�$%)�%+( �.��(�%��T,�AB IJqupE tK©NEªGqqNo©LGM@�(!�R"%+� #*+"#��("�����%��"#T"R%��(�����%P�%(�%#T��"$�(� O�R ++�R�" #�O P�.U+&)�b%(��,�?"�"§�#(�b U+T��%Q������ !�" #�� �P��U�(��R ++�R�" #�"#����"P�#�"S�. UP� T(,�AB «¬NHt©LGMtqE­MFGH®t©LGM@�2���R"�"§�#(�$%)�(U.(RP".��O P�R #�"#U U(�# �"R�(�%. U��3!�P%�" #%+�1#O P$%�" #� O�����R"�)�"#R+UT"#S�P %T�$%"#��#%#R��%#T�P�!%"P�("��(�%. U��P %T*$%&"#SV�R+ (�T b#�P %T(V�%RR"T�#�(V��%P��b P&� P��'R%Q%�" #V�%#T�"#��PPU!�" #(� O�S%(� P��+�R�P"R"�)°� #�����P��U�(��T��*$%"+�%TTP�((,�2���(�PQ"R��"(�"#��SP%��T�b"���%��1<�$%!,--AB ±LHE²JtqL©pa� #*+"#��%"P��U%+"�)�$ #"� P"#S�b�.("��,�2�"(�(�PQ"R��$%)���+!�����R"�"§�#(�"!�.��%b%P�� O�b�"R��"(������U%+"�)� O�����%"P�"#����"P�#�"S�. UP� TV�b P&!+%R�� P�%#)�T�("P�T�+ R%�" #�OP $�����R"�),�/+( V�"#O P$%�" #�P�S%PT"#S�/"P�³U%+"�)�$%#%S�$�#�V�R #R�#�P%�" #(� O�!%P�"R+�(V�+�%TV�(U+!�UP�T" '"T�V�R%P. #�$ # '"T�� P� ���P�"$!UP"�"�(,�.,�/0012134/5�64>1=347642�<6=>1?6<@-AB ¨ N EµqNtEDtHuN©Ka� #*+"#��(�PQ"R����P US��b�"R������R"�"§�#�R%#�U!+ %T�"��$(���¶(���b%#�(�� �(�++� P�S"Q��%b%)�"#R+UT"#S�!"R�UP�(V�R�%P%R��P"(�"R(�%#T�%#)� ���P�U(�OU+�"#O P$%�" #,�3���P�U(�P(�$%)�R #�%R���"$¶��P�� �.U)�����!P TUR��"O����)�%P��"#��P�(��T,�2���(�PQ"R��$%)�.�� OO�P�T�U! #�P�S"(�P%�" #,�/+( V�����(�PQ"R��R U+T� OO�P�"#O P$%�" #�%. U��!�)("R%+�b�.�O+�%�$%P&��(�"#��SP%��T�b"���%��1<�$%!,-AB sGqqJ©LGMEª·tHrNa-�%'�� �T"(R UP%S������U(�� O�! ++U�"#S�!P"Q%���Q��"R+�(�"#("T������R�#�P%+�!%P�� O�����R"�)�.)�%!!+)"#S�%#��#�P%#R��R�%PS��P�+%��T�� �����! ++U�"#S��$"((" #�+�Q�+(,�2���(�PQ"R��$%)�.��R $!+���+)�!�PO P$�T�

Page 98: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"����#�$!%&'&()*�+#,-&�#.���/0/1��234565�4�).�789:;9<;�=>?@ABA=@�7@<C>A@7D�EF GHIJKLMNOPHIQRSRTMQHJUV( --&,&*&�W�� ��#X!#-��)%W�Y&%Z� [�(#�- %)*�'#��&[&')�#�\,&���]�Z# _ abc ddefgbhibjkl*&%#.�8�#��#X!&�#Z�&%[ �$)�& %�-� !*Z�,#�-#%��Z&�#'�*W�� ���#�!-#��#&��#��,W�#l$)&*� ����� !"����#�Z#-&�#Z�'�)%%#*.�>[���#�-#�m&'#��#X!&�#-�)%W�()W$#%�]���&-�$)W�)*- �,#�(#�[ �$#Z�' $(*#�#*W� %l*&%#.VEF GKnoNMQRKLOpHqRJQIrUV( --&,&*&�W�� �(#�[ �$�)��#"&-��)�& %�&%���#�'#%-!-� %l*&%#.��8�#�'&�W�' !%'&*�$)W�(� m&Z#���#�!-#��+&���)�' %[&�$)�& %�#&��#�� %l*&%#� ����� !"����#��#X!#-�#Z�'�)%%#*.VEF pHJRsHLTHOpHqRJQIMQRKLUV( --&,&*&�W�� � [[&'&)**W��#"&-�#��)�'&�&t#%�'�)%"#� [�)ZZ�#--� %l*&%#.�8�&-�&%[ �$)�& %�$)W�,#�!-#Z�)-���#� [[&'&)*�)ZZ�#--���� !"��+�&'����#�' $$!%&')�& %�'&�&t#%l$!%&'&()*&�W�-�)**�,#�Z %#]�&%'*!Z&%"���#� [[&'&)*�m �&%"�)ZZ�#--.�8�#�'&�W�' !%'&*�$)W�(� m&Z#���#�!-#��+&���)�' %[&�$)�& %�#&��#�� %l*&%#� ����� !"��)%W� ��#��'�)%%#*.�EF uRqRQMNOuKTovHLQJOwMLMqHvHLQOxrJQHvUV-#�m&'#� [[#�#Z� %l*&%#�,W���#�'&�W�' !%'&*���� !"��+�&'����#�!-#��$)W��#X!#-��)%W�%##Z#Z�Z '!$#%���#")�Z&%"�eyz gdb{e|g}~}{gb�dg��gy_bj|bg�_d }_yab}j�egyhi�b�|b_ gbygd�e}gbdg��edgyb k~b()W$#%�]���&-�$)W�)*- �,#�(#�[ �$#Z�' $(*#�#*W� %l*&%#.VEF �L�NRLHO�RLHJOPKLJoNQMQRKLOMLsOGMrvHLQU� %l*&%#�-#�m&'#���)��+ !*Z� [[#����#�( --&,&*&�W�� �' %-!*��)%Z�()W�)%W��W(#� [�[&%#�$ !%�#Z�!(�,W���#�'&�&t#%.�8�#�-#�m&'#�$)W�,#�(� m&Z#Z�+&���)%�)*#���-W-�#$���)��+ !*Z�&%[ �$���#�'&�&t#%� [�)%W�'* -#�()W$#%��Z#)Z*&%#.VEF �L�NRLHO�M�HJOMLsO�HHJOGMrvHLQU� %l*&%#�-#�m&'#���)��+ !*Z� [[#����#�( --&,&*&�W�� �' %-!*��)%W��)�� ��[##�-�)�!-�)%Z�()W�&�.�8�#�-#�m&'#�$)W�,#�(� m&Z#Z�+&���)%�)*#���-W-�#$���)��+ !*Z�&%[ �$���#�'&�&t#%� [�)%W�'* -#�()W$#%��Z#)Z*&%#.VEF ���������������5������������������ %l*&%#�-#�m&'#���)��)&$-�� �#%�)%'#�#[[#'�&m#%#--�&%�)%W�(� '#Z!�#��#*)�#Z�� �$)��&)"#� ��'&m&*�!%& %D��#X!#-��[ ��)%�)m)&*),*#�Z)�#� ��)%�)m)&*),*#�$!%&'&()*�� $]�&%-'�&(�& %�&%���#� [[&'&)*��#"&-�#�� ��)��#*(�"!&Z#�+&���)**���#�-�#(-�� �,#��)Y#%.VEF ��������DV %l*&%#�-#�m&'#���)��+&**�)** +���#�!-#��Z +%* )Z�)%W�,&**��#*)�#Z�� �$!%&'&()*�-#�m&'#-�\")-]�+)�#�� ��+)-�#�.��)W$#%��-� !*Z�)*- �,#�(#�[ �$#Z� %l*&%#.VV,.�9;;>8>�:9=�=>?@ABA=@�7@<C>A@7DVEF �L�NRLHO�H�RTNHJOpHqRJQIMQRKLU�-#�m&'#���)��+ !*Z�)** +�)�'&�&t#%�� ��#"&-�#���&-��#��m#�&'*#� %l*&%#.�>[���#�-#�m&'#��#X!&�#-�)%W�()W$#%�]���&-�$)W�)*- �,#�(#�[ �$#Z�' $(*#�#*W� %l*&%#.VEF xHMIT�OSKIO�oIRMNOGNMTHJD� %l*&%#�-#�m&'#���)��)** +-�-#)�'��[ ��,!�&)*�

Page 99: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#�$%&'!(��!&")*"��*(�+'",)&' ���-)*%-!*,".��/+!&!�-�0�$!�'*,!1&�,')*�2',+���345�-��.���67686���9:;��9;<=��.�5/>?@>A@�5BC4>D�C>AE�5EAF4CE5G�HI JKLMNOPQRKSQTUKVWUQXNYZQ[KXQ\O]K^UMQKSQ_O`USQaZSUbYQKRQJK^UWUMMOUMMcd)*e�'*!�"!&f' !�,+&)%1+�2+' +��!)��!�2',+�! )*)-' �+�&("+'�"�-�0�����0�g)&��*0�+)%"'*1�+!���'*�,!&-"�)g�! )*)-' �g%*('*1�)&��&)f'('*1����&)f'"')*���� )--)(�,')*�',"!�g.��HI hNMbiWU QTbSjNOPQTUS NYMcd)*e�'*!�"!&f' !�,+&)%1+�2+' +�('"�$�!(��!)��!� �*�&!k%!",�����&l'*1��' !*"!�,+�,�2'������)2�,+!-���&l�'*�,+!�&!",&' ,!(��&!�"�)&��"l�g)&����!&")*�����&l'*1���� !�*!m,�,)�,+!'&�+)%"!".�HI hNM]KLOYMQRKSQTUKVWUQXNYZQ[KXQ\O]K^UMcd"!&f' !�,+&)%1+�2+' +��*0�&!"'(!*,�2',+�! )*)-' �+�&("+'�"�-�0�&! !'f!�+!���g&)-�,+!'&�-%*' '���',0�'*�g)&-�)g�('" )%*,"�g)&��*0�� ,'f',0�)&�"!&f' !.�HI TUS NYMQRKSQnSPbONoNOPQTLiWN]QpqUOYMG�)*e�'*!�"!&f' !�,+&)%1+�2+' +� 'f'��"!&f�*,"�&! !'f!�����' �,')*"�g&)-� ','r!*"#� ��'-�� )&(�* !��*(�'""%!��!&-',"�g)&��&&�*1'*1��%$�' �!f!*,".�HI \O`NqN LbWQbO`Qsb^NWtQubSUQvUSqN]UQwN x]bW]LWbYNKOG�)*e�'*!�"!&f' !�&!��,!(�yz{y|}{~���������y���{�z����{�z��{y|�y{~��{������}{��}����{����{�z�{y|z�}{g�-'�'!"�)&�'*('f'(%��"�2',+�"�! '���*!!("#�! )*)-' �)&�*),.�HI ��������������������c��*�)�!*� �,��)1%!�2+!&!��*0�%"!&�-�0�g'*(�,+!��((&!""�)g��*0�-!(' ���() ,)&�'*�,+!� ',0.�/+!�"!&f' !�-�0�$!�'*,!1&�,!(�2',+���345�-��.�$.�>@@4/4B?>D�5BC4>D�C>AE�5EAF4CE5G�HI �KONYKSNOPQuZNW SUOQ LSNOPQJKWN btQvUbMKOMQ�iURKSUQbO`QbRYUSQM]ZKKW�c�,))��,+�,���&!*,"� �*�%"!�,)��))l��g,!&�,+!'&�l'("��*0,'-!�,+!0� �**),�$!�,�l'*1� �&!�)g�,+!-.�/+!�,))��-�0�'* �%(!�-)*',)&'*1�(%&'*1�+)�'(�0�"!�")*"��*(�$!g)&!��*(�)&��g,!&�" +))��� ,'f','!"#�,+�,�'"#�,+!��)""'$'�',0�,)��!,�,+!� ','r!*�('&! ,�0�!*,!&�,+!�&!k%!","�g)&����0'*1�"k%�&!"��*(�,+!�-)*',)&'*1�$!g)&!��*(�)&��g,!&�" +))��+)%&".�HI  NSYZ`btQ UOURNYMcd)*e�'*!�"!&f' !�,+&)%1+�2+' +�l'("�%*(!&��� !&,�'*��1!�-�0�&! !'f!����&!"!*,�¡! )*)-' �)&�*),¢�g&)-�,+!'&�-%*' '���',0.�/+'"�"!&f' !�-�0�$!�)&'!*,!(�,)���&1!�g�-'�'!"�)*�0�)&�g�-'�'!"�2',+�! )*)-' �+�&("+'�".�£�*l�� )%*,�*%-$!&�)&�),+!&�'*g)&-�,')*�-�0�$!�&!k%'&!(�%*(!&�'(!*,'g' �,')*.�HI uZNW iNSYZQwWWKXbO]UMc�)*e�'*!�"!&f' !�,+&)%1+�2+' +�,+!�*!2���&!*,"�-�0�&!k%!",��*�! )*)-' �$!*!g',���'(�$0�,+!�-%*' '���',0�'*� �"!�)g�$'&,+.�£�*l�� )%*,�*%-$!&�)&�),+!&�'*g)&-�,')*�-�0�$!�&!k%'&!(�%*(!&�'(!*,'g' �,')*.�HI JUbWYZQvUSqN]UMQKRRUSU QYKQpW USWtQYZSKLPZQhK^KYN]bc��*0�l'*(�)g�()-),' �yzz�{��}�{yz{�}���z��}{y|}{���}{z�{z��{�}z��}��{|}��y|¤{¥|}{�}�¦��}{~��{§}{%�)*�&!k%!",��*(�-�0�&!k%'&!�")-!�,0�!�)g�&!1'",&�,')*.�

Page 100: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ��� � !"#$%"&$#'(")(*�+,%-(.%+%/0(")(�,1�&,2(%-&"/'-(3"4"%$5+67899:�;<=>�89�?:=@8�8A=�B?@=�B=C8@=�8A@9;DA�=:=B8@9CEB�F=?C<�EC�B?<=G�HA=C�F9CE89@ECD�=:>=@:I�J=9J:=�A=?:8AG�B=@8?EC�J?@?F=8=@<�?@=�=KB==>=>�9@�EC�B?<=�?C�?B8EL=�9@�J?<<EL=�?:?@F�D9=<�9MMN���� O4P/,+%"&2(Q+&�(Q""&1$#+%$"#(%-&"/'-(R3O067J@9LE<E9C�9M�A?C>�A=:>�B9FJ;8=@<�9@�STU<�89�?::� C;@<=<�<9�8A?8�8EF=VB9C<;FECD� 8?<W<�:EW=�J@=<B@EJ8E9C<�9@�A=?:8A�@=J9@8<�X=B9F=�<A9@8=@N�YAE<�<=@LEB=�F?I�A=:J�=CA?CBECD�=MM=B8EL=C=<<�?C>�=MMEBE=CBI�9M�8A=�F=>EB?:�J@9M=<<E9C?:<�8A?8G�EC�<9F=�B?<=<G�A?L=�89�LE<E8�J?8E=C8<�EC�8A=E@�A9;<=<N��� R",$5�(Z "&[0(%-&"/'-(R3O06�J@9LE<E9C�9M�STU<�89�J9:EB=F?C�EC�BA?@D=�9M�8@?MMEB�B9C8@9:N�YAE<�<=@LEB=�F?I�A=:J�=CA?CBECD�=MM=B8EL=C=<<�EC�MEC=<�?C>�98A=@�C98EMEB?8E9C<�8@=?8F=C8�<ECB=�E8�B?C�X=�<=C8�>E@=B8:I�89�8A=�B=C8@?:�9MMEB=N��� O#$4+,0(\�'$0%�&6�@=DE<8=@�M9@�>9F=<8EB�?CEF?:<�?C>�HE:>�?CEF?:<�:ELECD�EC�B?J8ELE8IN�]C<B@EJ8E9C�M9@�>?CD=@9;<�?CEF?:<�<A9;:>�X=�F?C>?89@IN�U:<9G�?�>9B;F=C8�J@9LECD�8A?8�8A=�?CEF?:�A?<�?C�E>=C8EMEB?8E9C�FEB@9BAEJ�F?I�X=�@= ;E@=>N��_ _abcdef�g�dc�?N�hYUiTUjT�YjUihSkjY�hljm]nlho��� R/P,$5(p&+#0q"&%(.�&r$5�0o�9CV:EC=�<=@LEB=�@=D?@>ECD�ECM9@F?8E9C�?X9;8�J;X:EB�Y@?C<J9@8�EC�D=C=@?:�?C>�F9@=�<J=BEMEB?::Io�s9;@C=IVJ:?CC=@<G�LE@8;?:�8EBW=8<�?B ;E<E8E9C�9@�89;@E<8EB�@9;8=<N�YA=@=�F?I�?:<9�X=�?C�?:=@8�<I<8=F�89�ECM9@F�;<=@<�?X9;8�?CI�ECBE>=CB=N��� Q+&(R+&[$#'o�9CV:EC=�<=@LEB=�8A?8�F?I�ECB:;>=�ECM9@F?8E9C�?X9;8�F;CEBEJ?:�B?@�J?@WECD�?@=?<�9@�9CV:EC=�?B ;E<E8E9C�9M�=:=B8@9CEB�J?@WECD�L9;BA=@<N�YA=�<=@LEB=�F?I�X=�EC8=D@?8=>�HE8A�?�t]h�F?JN�S?IF=C8�<A9;:>�?:<9�X=�J=@M9@F=>�8A@9;DA�F9XE:=�JA9C=N��� p&+))$5(Z �P(Q+4�&+0o�9CV:EC=�?BB=<<�89�8A=�ECM9@F?8E9C�@=B=EL=>�XI�8A=�8@?MMEB�B?F=@?<�EC�9@>=@�89�W==J�8A=�BE8Eu=C<AEJ�;J>?8=>�9C�8@?MMEB�s?F<G�?BBE>=C8<�9@�C=H<�@=:?8=>�89�8@?MMEB�EC�8A=�BE8IN�XN�UTT]Y]kiUv�YjUihSkjY�hljm]nlho��� !/#$5$q+,(w$525,�0(x�#1$#'o�9CV:EC=�<=@LEB=�8A?8�H9;:>�?::9H�8A=�BE8Eu=C�;<=�?CI�F;CEBEJ?:�XEBIB:=N�YA=<=�XEBIB:=<�H9;:>�X=�J:?B=>�EC�<=L=@?:�JEBW;J�?@=?<�?@9;C>�8A=�BE8IN�YA=�<=@LEB=�F?I�X=�;C>=@�<;X<B@EJ8E9C�?C>�F?I�@= ;E@=�9CV:EC=�J?IF=C8N��� Q+&(R"",$#'67<=@LEB=�8A?8�EC8=C><�89�J@9F98=�<;<8?EC?X:=�F9XE:E8I�XI�J@9F98ECD�8A=�<A?@=�9M�J@EL?8=�B?@<�8A?8�>?E:I�F?W=�8A=�<?F=�@9;8=N�]8�B9;:>�X=�X?<=>�EC�?�H=XVX?<=>�<9M8H?@=�J:?8M9@F�8A?8�H9;:>�?::9H�;<=@<�89�>=MEC=�J99:ECD�J@=M=@=CB=<�y9@EDECG�>=<8EC?8E9CG�C;FX=@�9M�<=?8<�?L?E:?X:=G�=8BNz�?C>�MEC>�?L?E:?X:=�B?@<N��� {# ,$#�(p$5[�%(%"(�#%�&(%-�(x$4$%�1(p&+))$5(O&�+6�9CV:EC=�?B ;E<E8E9C�9M�8A=�J=@FE8�89�=C8=@�8A=�AE<89@EB?:|:EFE8=>�8@?MMEB�u9C=N�]>=C8EMEB?8E9C�<A9;:>�X=�

Page 101: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !��"#��$%$ $�&'�()�*�())+)��,#�-./01/21�324/0�56/00708�-92:7;9-<=>? @ABCDEFGDHABIJKELHDCM�NOPQ!O��R��S!T��UV,U�W,X�!OTQ "��UV��YNRR!Z!Q!UX�UN�,R[�\N��,OX�[!O"�N\�TNORU� TU!NO�Y��W!U�UV�N ]V�UV��7OU��O�U#�7\�Y,XW�OU�RVN Q"�Z��"NO� �UV!R�W,X�,QRN�Z��Y��\N�W�"�TNWYQ�U�QX�NOPQ!O�#=>? _ BabFCKI Jc BCI `BaI dFHcaHBeI fKeFc DHAB<� NOPQ!O�� Y ZQ!T� !O\N�W,U!NO���],�"!O]�,TU ,Q�,O"�\ U ���Q,O"P R��YQ,OR�,O"�T,",RU�����] Q,U!NO#�.V��R��S!T��W,X�Z��!OU�]�,U�"�g!UV�,�87-�W,Y#�-NW��!O\N�W,U!NO�W,X���� !���!"�OU!\!T,U!NO#�>? hijklmlnokipqrnosopkjotqlumvjwoklvuxqkoyoklvuxqwvjkzoziqpkokiwiukp{|M=NOPQ!O��R��S!T��UV,U�gN Q"�Q�U�UV�� R���,R[R�\N��,OX� �Z,O�T��U!\!T,U��NOPQ!O�#�}~������������������������������������������������~��������������~������7\�Y,XW�OU�!R���� !��" �UV����TN Q"�,QRN�Z��UV��YNRR!Z!Q!UX�UN�Z��Y��\N�W�"�NOPQ!O�#=Z#�/117.7�0/6�324/0�56/00708�-92:7;9-<�>? �G�FHCHDHABIA�I_ BaI��I�AEKHeBKECM�NOPQ!O��!O\N�W,U!NO�\N��!WW!]�,OUR�NO�VNg�UN�NZU,!O�,�Y�NY��UX�!O�UV�!��VNRU�T!UX#�.V��Y �YNR��N\�UV!R�,YYQ!T,U!NO�W,X�Z��UN�,QQNg�\N��!]O��R�g!RV!O]�UN�Y �TV,R��,OX�[!O"�N\�Y�NY��UX�!O�UV��T!UX�UN�,YYQX�\N��UV����� !��"�,YY�NS,Q�S!,�UV��!OU��O�U#�.V!R�TN Q"�Z��"NO��ZX�,�QNT,Q�Q,gX���N��,�Y ZQ!T�ONU,�X#�I�#��17:92-7.��/R�RU,U�"��,�QX�!O�UV!R�TV,YU���R�Q�TU�"�R��S!T�R�V,S��Z��O�]�N Y�"�,TTN�"!O]�UN�UV�!���������������������������������������������������������������,O",�"�,O"�,""!U!NO,Q�R��S!T�R#�  V!Q��RU,O",�"�R��S!T�R�W,X�R��W�UN�!"�OU!\X�]�O��,Q�Y,UU��OR�,""!U!NO,Q�!O\N�W,U!NO�W RU�R ��QX�TNWYQ�W�OU�UV�W�R!OT��UV��\,TU�UV,U�,�]�O��,Q�Y,UU��O��¡!RU�"N�R�ONU�W�,O�UV,U�UV����,���ON�!WYN�U,OU�,O"�!OU���RU!O]��¡T�YU!NOR#�.V�R���¡T�YU!NOR�g!QQ�Z���¡Y��RR�"�!O�UV��\N�W�N\�,�"!S��R!UX�W,�[�\NT R!O]�!O�UV��S,�!�UX�N\�T!�T WRU,OT�R�UV,U��¡!RU�gV!Q��R �S�X!O]��P8NS��OW�OU�!O�R�S��,Q�T!U!�R#�.V!R�S,Q ,ZQ��!O\N�W,U!NO�NZU,!O�"�\�NW�UV��,""!U!NO,Q�R��S!T�R�RVN Q"�ONU�Z�� OONU!T�"�¢2,]!O ��££¤¥#�7O�RVN�U�UV��R �S�X�N\�"!S��R!UX�,SN!"R�,O��¡TQ R!S��\NT R�NO�gV,U�!R�TNWWNO�N��NO�"NW!O,OU�Y,UU��OR#�}~�������¦�§�������������������§��������§������������������ ������§��~����~��~����������~��\NQQNg!O]�TNWY,�,U!S��\,TUN�<�"!S��R!UX�©���ª�¢«�R��S!T�R�Y�NS!R!NO�¬�«�Y��\�TU�TNS��,]�¥�.V��"!S��R!UX�\,TUN��TNWY,��R�UV��T ���OU�TNS��,]��N\�,QQ�R��S!T�R�¢ZNUV�RU,O",�"�,O"�,""!U!NO,Q¥�UN�UV��Y��\�TU�RT�O,�!N<�UV��UNU,Q�Y�NS!R!NO�N\�,QQ�R��S!T�R�ZX�,QQ�T!U!�R#�.V���\N�� �"!S��R!UX�V�QYR�,O,QXR!O]�UV��S,�!,U!NOR�!O��P8NS��OW�OU�Y�NS!R!NO�!O�9 �NY�#��

Page 102: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ������� !"�#$�"%�� $&'�&�($)���*����+$,,�&$,�-��$&.�/�0�&�($)���*�0�1�23$ �4� !,,*'�&�($)���*�+�,,�#$�5�$&��"��6����5)($*��%� %-3!)$� %6$)$" $��"� ���$��&$,�($)*�%4��$)(� $��4%)�$! 6�%4��6$�"�"$� %"��&$)$&� !�$7%)�$�8���,!)7$�&�($)���*�"5-#$)�+%5,&��-3,*�!�#�7�&�44$)$" $��"��$)(� $�3)%(���%"�!-%"7�95)%3$!"� ���$�8��� !�$7%)*�+��6�!�&�($)���*�%4�/8:;�+�,,��6%+�!�,%+� %6$)$" $8�<"� %" )$�$'�95)%3$!"� ���$��+%5,&�%44$)�=5��$�&�44$)$"���$)(� $��! 6�$(�"7�%",*�!�>>?�%4� %�" �&$" $��"��$)(� $��&$,�($)*8�@"��6$� %"�)!)*'�!��-!,,�&�($)���*�-!)A�+%5,&�-$!"��6!��3!)�� �3!"����"�&�44$)$"�� ���$��&$,�($)��6$��!-$�A�"&�%4��$)(� $��4%)�!�3!)�� 5,!)� !�$7%)*8�B%)��"��!" $'�!� !�$7%)*�+��6�!�&�($)���*�-!)A�%4�/81C�+�,,��DEFGDHIDGJEDKLKMJKGHMGNKLOHJKNPGQLEOHNHEMGRKSMHMIGTDSTGUVLEQKSMGJHTHKNG+%5,&�%44$)��$)(� $��+��6�!"�WC?�%4� %�" �&$" $8��8>�X9Y9Z�[��Y�[\2<2�@B�]@^9Z�X9��Y_�_< 9Z2< \��"�!"!,*����%4��6$�X,%#!,�2$)(� $�]!�!,%75$�-5���#$�3$)4%)-$&��!A�"7��"�%�! %5"��#%�6�3!)!-$�$)�.���!"&!)&� %($)!7$�!"&�&�($)���*8�@"�%"$�6!"&'��6$�3$) $"�!7$�%4� %($)!7$� !"��$,,�+6� 6��6$� %�" �&$" $�%4���!"&!)&��$)(� $���"�95)%3$���8�<�� !"��"4%)-�!#%5�� %--%"��"�$)$����%)�-!*#$� %--%"�3%,��� !,�95)%3$!"�! ��%"��!44$ �$&�$a2$)(� $��&$3,%*-$"��!"&�DKbQKcGTEGJLKSTKGSGNDSLKcGOHNHEMGEdGNKLOHJKNPGcKbHOKLeGTDLEVIDGTDKGVNKGEdGKbKJTLEMHJGRKSMNf�@"��6$�%�6$)�6!"&'�&�($)���*��6%+��(!)�!��%"���"�$aX%($)"-$"�� %"��&$)�"7�#%�6�!&&���%"!,�!"&�NTSMcSLcGNKLOHJKNfGgDKLKdELKhGJERQSLHMIGNTSMcSLcGNKLOHJKNPGJEO$)!7$�!"&�&�($)���*� !"�6$,3�&�� %($)�"7�6%+�!&&���%"!,��$)(� $�� %($)!7$�-!*�&�44$)�4)%-�%"$� !�$7%)*��%�!"%�6$)8��Y$i�'��6$)$����!��!#,$� %-3!)�"7��6$�95)%3$!"�3$) $"�!7$�%4���!"&!)&��$)(� $�� %($)!7$�!"&��6$�&�($)���*�-!)A�3$)� !�$7%)*8��,�%'��"�%)&$)��%�4! �,��!�$����� %-3)$6$"��%"'��6$��%�!,�"5-#$)�%4��$)(� $��!"&�&$�!�,$&�"5-#$)�%4���!"&!)&�!"&�!&&���%"!,��$)(� $��3$)� !�$7%)*����3)%(�&$&8�������������� jjjjjjklmnojpqjrsltulvujrovwxyozj{|wovl}ojltuj~xwovzxs�j�ovjylso}|v�j�������� ��������������� ����������������� �������������������� ����������������� ���������{xsx����������������j 1/� :� �� W�'��?� /'�C�{�lttonnxt}j 1/� W� �� W/'W>?� /'�:��u�ylsx|tj 1>� �� �� ;;';W?� /'>�����n|��otsj�j��zxtozzj ;� C� �� ;:'//?� /'>C��twxv|t�otsj ;� C� �� W/'//?� /'�>��x�oy�ynoj 1/� W� �� :;'C/?� /'>��r|yxlnjylvoj 1�� :� W� W/'//?� /'C��kvltz�|vsj :� >� >� ;C'C:?� /'�:��vmltj�nlttxt}j �� >� 1� W�'��?� /'>��

Page 103: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� �!"��#�$#%&$'&()#��*&��&))�+&�#,-$�# �&+*�#.#�/#$+#0�&,# �-1� �&02&$2�+-.#$&,#�&)3&4 ��0�&�$&0,#�(#�3##0�567�&02�8�7�3��*�0-�+-$$#)&��-0�3��*��*#�0"%(#$�-1� �&02&$2� #$.�+# 9�:#0+#;��*� �$# ")��+-01�$% �0-��-0)4��*#� "��&(�)��4�-1��*#� #$.�+# �+*- #0�3*�)#�+-0 �$"+��0,��*#�<)-(&)�=#$.�+# �>&�&)-,"#� "(%���#2��-� �"24;�("��&) -��*#�*�,*�2#,$##�-1� �&02&$2� #$.�+# �+-0 -)�2&��-0��0�?"$-/#9�@-�#A/)&�0��*#�%#&0�0,�-1��*# #�1�,"$# ����%" ��(#�&) -�-( #$.#2�*-3�/-/")&�#2�&$#��*#� #��-1� �&02&$2� #$.�+# �-0�#&+*�+&�#,-$49����+-")2�(#� "//- #2��*&��&�+&�#,-$4�3��*�&�*�,*� �&02&$2�+-.#$&,#�&02�&�3#))�/-/")&�#2� #��-1� �&02&$2� #$.�+# �� �&�3#))�# �&()� *#2�+&�#,-$49��0�-�*#$��#$% ;��*#�%&B-$��4�-1�+���# � *-3�&�+-%%-0�"02#$ �&02�0,�-0��*#�(& �+� #$.�+# ��-�(#�2#)�.#$#2�-$�#.#0�&�*�,*�2#,$##�-1�&)�,0%#0��&+$- �?"$-/#&0�+-"0�$�# �$#,&$2�0,�+��4�+-%/#�#0+# ��0��*&��+&�#,-$49��C *#0�)--'�0,� /#+�1�+&))4�&���*#� �&02&$2�+-.#$&,#� +-$# ��0��&()#�D;�+&�#,-$�# �%-$#�$#)&�#2�EFGHIJKFLMGNHEHOMPQRGSEEMPEHFPGSPTGMPUSPNMGEUMGIVPHNHJSWHEXY�-Y+���Z#0�+-%%"0�+&��-0�)�'#�[\]\ _ abc_ ded_f _ ]�-$�ghijjkllmjn;�-$�opqmilrqisk�&//#&$��-�(#��*#�%- ��3#))�# �&()� *#2�+&�#,-$�# ;�$#&+*�0,��*#�*�,*# ��+-.#$&,#� +-$# �3��*�&�3#))�/-/")&�#2� #��-1� �&02&$2� #$.�+# 9��t) -�3#�1�02�usvijrwlijjmjn;�&02�xjymspjzkj{�3*�+*�/$# #0��&0��%/-$�&0�� �&02&$2�+-.#$&,#�("��3��*�&�.#$4�)-3�/-/")&��-0�-1��*#� �&02&$2� #��-1� #$.�+# ;�(-�*�3��*�-0)4��� �&02&$2� #$.�+# �3*�+*�+-")2�+-0�$�("�#��-��*#�*�,*�%&$'�-0� �&02&$2�+-.#$&,#9��|0��*#�-�*#$�*&02;�}m~kq�qlk�&)-0,�3��*��sij�wps{�rxzwlp�zkj{r�rv��mjk���&02�x��qi{mpj�&//#&$��-�(#��*#�+&�#,-$�# �3��*��*#�)-3# �� �&02&$2�+-.#$&,#�&%-0,�?"$-/#&0�+���# ;�3��*�/#$+#0�&,# �(#�3##0�567�&02�6879�}m~kq�qlk�&02�x��qi{mpj�+&�#,-$�# �&$#��3-�-1��*#�%- ��/-/")&�#2�+&�#,-$�# ��0��*� � "$.#4;�3*�) ���sij�wps{�&02�xzwlp�zkj{r�rv��mjk���&$#�+&�#,-$�# ��*&��/$# #0��&� *-$��)� ��-1� #$.�+# 9��#,&$2�0,�2�.#$ ��4� +-$# �/#$�+&�#,-$4�1-"02��0�@&()#�D;��*#�&0&)4 � �*#$#�*& ��-��&'#��0�-�+-0 �2#$&��-0�*-3�/-/")&�#�� ��*#� #��-1�&22���-0&)� #$.�+# �&02�*-3�� ����$#)&�#2��-��*#��-�&)�0"%(#$�-1� #$.�+# 9�C *#0�&�+&�#,-$4� *-3 �&�*�,*�2�.#$ ��4�%&$'����%#&0 ��*&���*#$#�� �&0��%/-$�&0��0"%(#$�-1�2�11#$#0��-11#$ �-1� #$.�+# �/$-.�2#2�(4�+���# 9��1;�&���*#� &%#���%#��*#� #��-1�&22���-0&)� #$.�+# �� �3#))�/-/")&�#2;��*� �+&�#,-$4�+&0�(#�+-0 �2#$#2�& �&0�&+��.#�+&�#,-$4��0�3*�+*�+���# �&$#�+"$$#0�)4�3-$'�0,�&02�/"���0,�#11-$� 9�@*#�+&�#,-$4;�&�/&$��-1�(#�0,�3#))�# �&()� *#2�-$�0-�;�� �#.-).�0,9�@*� �/$-+# �+-")2�(#�2"#��-�&�$#&+��-0��-�+���Z#0� �2#%&02;�/-)���+&)�/$�-$���# �-$�#.#0�?"$-/#&0�$#+-%%#02&��-0 9��opqmilrqisk�2# /��#�(#�0,�&�3#))�# �&()� *#2�+&�#,-$4� *-3 ��*#�*�,*# ��2�.#$ ��4�%&$'�&02�/$# #0� ��*#�%- ��/-/")&�#2� #��-1�&22���-0&)� #$.�+# 9����� ��*#�(# ��#A&%/)#�-1�&0�&+��.#�+&�#,-$4��0�3*�+*�?"$-/#&0�+���# �&$#�2#)�.#$�0,�&�$�+*�.&$�#�4�-1� #$.�+# 9��sij�wps{�&02�xjymspjzkj{;�&$#�&) -�,--2�#A&%/)# �-1�&+��.#�+&�#,-$�# 9��gm{m�kjabc_ ded_f _ ]��&)�*-",*� *-3 �&�,--2�%&$'�-0�2�.#$ ��4;�*& �&���*#� &%#���%#�&�)-3�0"%(#$�-1�&22���-0&)� #$.�+# �3*�+*�%#&0 ��*&���*#�+&�#,-$4�� �3#))�# �&()� *#2�&02�+-0 -)�2&�#29�|0��*#�-�*#$�*&02;�usvijrwlijjmjn;�x��qi{mpj�&02�ghijjkllmjn� *-3��*#�)-3# ��2�.#$ ��4�

Page 104: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"�#$%&$��''$��(%)�*+�,$�*#$��+!*�(%�&*(-$�&�*$)+�($!�(%�.#(&#��/%(&('�0(*($!�'�+-(1$�-$�2�!(�(0���!$�-(&$!��%1�0+.�'$�&$%*�)$�+3��11(*(+%�0�!$�-(&$!4��5(%�002"�6789:;7<=>?@?ABCD=<CC��%1�EDF<G;G9<"�!#+.!��(1�1(-$�!(*2���� !�.(*#��0!+���%+%�!()%(3(&�%*�!$*�+3��1-�%&$1�!$�-(&$!4�H#+!$�&�*$)+�($!�!#+.��0!+�+%$!�+3�*#$�!��00$!*���� !�+%�!*�%1��1�&+-$��)$4�I+*#�!$��!�*+�,$�J/(*$�.$00�$!*�,0(!#$1�&�*$)+�($!4��4��KHLMNLON�PQRSOLTS�LMN�NURSOKUHV�WSO�PLHSTQOV�X+�$�!'$&(3(&�002"���!'$&(�0�3+&/!�+%�$�&YZ[\Z]Y Z_ \\ a b]Zcd] e[a fgZf ah c fZia[h f[bZfZ'�$!$%*$14�H#$�3+00+.(%)�%(%$���'!�)��'#(&�002��$'�$!$%*�!*�%1��1�!$�-(&$!�'�+-(1$1�'$��'��*(&('�%*�&(*2�3+��$�&#�+3�*#$�%(%$�&�*$)+�($!4�K(%&$�'�+-(!(+%�+3�!$�-(&$�(!�-$�2��/&#��$0�*$1�*+�*#$�&(*2�&+�'$*$%&$!�+%�$�&#�+3�*#$�&�*$)+�($!"�(*�(!�(%1(&�*$1��!�.$00�.#$%�*#$�!$�-(&$�(!�%+*���&+�'$*$%&$�+3�*#$�&(*2�,$&�/!$�(*�(!�'�+-(1$1�,2��%+*#$���1�(%(!*��*(+%�j!+�(*�(!�,$2+%1�*#$�&(*2�&+�'$*$%&$!"�,/*�*#$�&(*2�1+$!�+33$��(%3+���*(+%�j+��*��%!�&*(+%�0�0(% !k��$0�*$�*+�*#(!�!$�-(&$4�5(%�002"�(%�+�1$��*+�&+�'0$�$%*�*#$�-(!(+%�+3�*#$�!*�%1��1�!$�-(&$!�+33$�$1�,2�$�&#�&(*2"���1$*�(0�+3�*#$�%/�,$��+3��11(*(+%�0�!$�-(&$!�(!�+33$�$14��lmnmopqrstpquvupw pqnt vxyz{|}t ~�|�{��x�t ~{|�{xt ~�����t ~�|��x����t ~����x�t px���{�{t �{���x��t w���xt w�|���t ���{��t n����xxt n�{t����{t n�|�xt ��{xx�tl�y�tl��x���t��{x�|�tr{����x�tv��{��t tt tt tt tt tt tt tt tt tt tt tt tt tt tt tt��|y���}�y��xt�|��{��{�t tt tt tt tt tt tt tt tt tt tt tt tt tt tt ttl�y�tl��x���t���y{x�t tt tt tt tt tt tt tt tt tt tt tt tt tt tt ttl�y�tv|����{�tl�x���y�y��xt tt tt tt tt tt tt tt tt tt tt tt tt tt tt tt�{�{|��y��xt��|tw �x���}��t~�����x��t tt tt tt tt tt tt tt tt tt tt tt tt tt tt tt�����t tt tt tt tt tt tt tt tt tt tt tt tt tt tt ttv���y��x��tr{|���{�t tt �t tt tt tt �t tt tt �t tt �t tt �t �t �t� tt �|����{�t �� �|����{�t��t�x�y�{|t����x��|�y��x�tttt q�yt}|����{�t tt �|����{�t��t�x�y�{|t����x��|�y��x�tq�t�x��|��y��xt���{|{�t�������� ¡¢£¤��¥¤ ¦¢� ¡§�§¤�©ª ¤�¥�¥¤«¤�� ©¤�£§¬¤©?¡§�§¤�ª© ¤�¥�¥¤«¤���(!�+%$�+3�*#$�*.+��+!*�&+-$�$1�S/�+'$�%�&�*$)+�($!�.(*#���*+*�0�%/�,$��+3�­�!*�%1��1�!$�-(&$!��%1���*+*�0�!*�%1��1�&+-$��)$�+3�®�4�� 4��N$!'(*$�+3�%+�!$�-(&$�.(*#�°±± �&+-$��)$"�*#$�$�(!���#()#�&+-$��)$�+3��00�*#$�!*�%1��1�!$�-(&$!4�K+�$�+3�*#$��!*�%1�+/*�0( $�²D>;?G:B=GD9?9DC><=C"���!$�-(&$�*#�*�(%*$%1!�*+�!*�$%)*#$%�*#$�0(%$�,$*.$$%�*#$�&(*(³$%��%1�(*!��/%(&('�0(*2��%1��0!+�+33$����!*�+%)��00k'/�'+!$�&+��/%(&�*(+%�0(%$"�*#$�$3+�$�&�$�*(%)���-(-(1�3$$0(%)�+3�!/''+�*4�P(*(³$%�'��*(&('�*(+%��$0�*$1�!$�-(&$!�0( $�µ¶>DGD8µ>D:=?8¶:G<CC<C?+��²D>;?²:B=GD9?89<=µ¶;?C<CCD:=C��$�&#���#()#�&+-$��)$�'$�&$%*�)$��$�%(%)�*#�*��/%(&('�0(*($!�!*��*$)($!�#�-$�,$$%�3+&/!$1�+%�$k·^ [cadc¹gfZf]ad] e fº�

Page 105: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"��#�!$%�&'($%�")��'(�"�*�++���"&#+!�)�,�-&"�%(.!��/�!"��)!%)�,!0�.�!"1�'&.2�3456�7��)(*��")&"�")�.��!��$(�+�0�+� +&1!$%�8!�+,�8(.�9:.( �&$�':$!-! &+!"!���!$�")!��&.�&5�;<=>>?@@A>BC =DEFGHIC JKHLGMNDKC JGHOGDC JPMQKNC JPHRPDOSKRC JNMNODKC ?DTLSGUGC <GMTPDVPC WPMKDC WXHLPKC YPZGVKC [KMMPDDC [SGC<KOXGC [XHPDC \PGDDKC;PE]TC<NRGIKOGC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CC;PEP_GDTCGHTNDKMCaNMUGHTC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CC;NDTXMEKEPNDCbDcMPDGC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CCJXMMGEPDd>GFTMGEEGHC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CCBAeCKDUC\PHEXKMCaMPOSEC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CC;XTENRP_GUCf GQIKOGTC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CCBAeCJKTGUCgPHGLENH]C CC CC CC CC CC CC CC CC CC CC CC CC CC CC CC?DEGHIHPTGCaNMUGHTC CC CC CC CC CC CC CC CC CC CC CC CC CC CC CC=UUPEPNDKMCTGHhPLGTC CC iC jC jC jC iC jC iC jC CC jC CC jC jC CCkkkCC HNhPUGUC ll HNhPUGUCQ]CKDNESGHCKURPDPTHKEPNDmCCCC >NECIHNhPUGUCCC HNhPUGUCQ]CKDNESGHCKURPDPTHKEPNDmC>NCPDnNHRKEPNDCNnnGHGUCopqrsqtskuvwxtqyxkzvtku{qrrx||}ryk~xtw}�x~k������$�#�8(.�/�u{qrrx||}ry�!��")��")!.,�'(�"�-(0�.�,�-&"�%(.1�!$�9:.( ��*!")�&��"&$,&.,�-(0�.&%��(8��45���/�&"�")���&'��"!'�/�!"�)&��&�)!%)�$:'#�.�(8��"&$,&.,���.0!-���3�75��������������������������������������������������������+!$��(.��:++�"!$���*�+�""�.��)(*�&�0�.1�)!%)� �.-�$"&%��(8� .(0!�!($/�!$�")��8!.�"�-&���.�&-)!$%�&�"("&+�-(0�.&%�5��(*�0�./��('��*(.2�.�%&.,!$%�-:�"('!��,�*�#� &%���(.�9$"�. .!����(+,�.��)&���"!++�"(�#��,($�/�#(")���.0!-���.�+&"�,�"(�")��-($-� "�(8�  �#��545��¡(.�(0�./�u{qrrx||}ry�,!0�.�!"1�'&.2�345�¢7/�*)!-)��)(*��!$�")!��-&"�%(.1�!"��+(*��"��-(.�/�)&��&+�(� .(0�,�")&"�")��*(.2�-&..!�,�(:"�#1�")�� &."!-! &$"�-!"!����)(*��&�)!%)�:$,�.�"&$,!$%/�&$,�")�.��!��$("��(�':-)�&-"!0!"1�!$�,� +(1!$%�,!88�.�$"!&+���.0!-��5�£)�.�8(.�/�")!��-&"�%(.1�-&$�#��-($�!,�.�,�&��($��(8�")��'(�"�-($�(+!,&"�,�-&"�%(.!��5�����

Page 106: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#$%& !'()*+,& -.+/*01'.& -*+2*'& -304.1& -3+53'26.5& -1012'.& �'7/6*8*& 9*073':3& ;30.'& ;<+/3.& =3>*:.& ".003''& "6*&9.2<*& "<+3'& ?3*''.&@+*A7/6110& 6308+*'&�8</.(31'& && && && && && && && && && && && && && && && 6308+*'&�8</.(31'& && && && && && && && && && && && && && && &&!8<0(&�8</.(31'& && && && && && && && && && && && && && && && 11,*+.(31'&915*AB/6110& && && && && && && && && && && && && && && &&C34+.+3*7& && && && && && && && && && && && && && && && <0(<+.0&9*+3(.2*&B3(*7& && && && && && && && && && && && && && && &&D<'87&E1+& <0(<+.0&@+1>*/(7& && && && && && && && && && && && && && && &&B<55*+& .5,7& && && && && && && && && && && && && && && &&@1+(.0&E1+&�F*'(7&.'8&�G(+./<++3/<0.+&!/(3F3(3*7& && && && && && && && && && && && && && && &&!883(31'.0&7*+F3/*7& && && H& && I& J& J& J& && H& && H& J& K& J&� && @+1F38*8& LL @+1F38*8&4M&.'1(6*+&.853'37+.(31'N&&&& %1(&,+1F38*8&&& @+1F38*8&4M&.'1(6*+&.853'37+.(31'N&%1&3'E1+5.(31'&1EE*+*8&OPQRSQTSUVWXYTQZYU[WTU\S] QP_WRUYTX_Y U\S] QP_WR�ab�c�deafg�haigjbakagh�ljmeno�bpmqarl�fpg�palpgbf�restgj�mk�bfcrhcjh�bgjiaugb�vwx�csmrl�cyy�fpg�rarg�umrbahgjgh�ucfglmjagb�ar�fpg�bejigz�crh�c�bfcrhcjh�umigjclg�mk�{{|{}~o�tgymq�fpg�cigjclg|���gjfcar�bgjiaugb�jgcup�deafg�palp�umigjclgo�beup�cb�fpg�V]�P]TQ�U�YT_PQZYU_PY o�c�qahg�mr�yarg�bgjiaug�mkkgjarl�arkmjscfamr�ctmef�crz�ueyfejcy�bafg�ar�fpg�uafzo�mj��TY� �WW�UV�_�STYRU\S] QP_WR|��mqgigjo�mfpgj�bgjiaugb�ya�g�VWW�YTQP_WRU�W�Y� �WW��crh��WTPQ�U[WTUYXYRP Umryz�jgcup�c���~�njmiabamr|��gbahgbo�af�ab�asnmjfcrf�fm�jgscj��fpcf�cyy�uafagb�cjg�mkkgjarl�c�qahg�upcjfgj�mk�ueyfejg�jgycfgh�bgjiaugb�beup�cb��S]�PU\S] QP_WRo��_�TQT_Y �mj��]RS U[WTU ]�P]TQ�U�TW�Y P �������������������� �¡¢£���¢���������¤¥|��x�cymrl�qafp�afb�smhgjcfg�umigjclg�crh�palpgbf�restgj�mk�bfcrhcjh�bgjiaugb�ab�njmtctyz�c�umrbgdegrug�mk�fpg�g¦abfgrug�mk�asnmjfcrf�hakkgjgrugb�ar�uafagb�umsngfgrugb�ar�fpab�cjgc|�����

Page 107: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ��� � !"� �#$%&%'()*#�))% +,-./01% '203/45,2% '/06/,% '74825% '7097,6:29% '5456,2% �,;3:/</% =/4;7,>7% �742,% �?0372% @7A/>2% $2447,,% $:/%=26?/% $?07,% B7/,,2%@/30?7-9/,-%CD50%A58;%7,%107E2-/%;/3-50F% %% %% %% %% %% %% %% GG GG GG %% %% %% %% %%�?8473%$/,</0;%CA58;%7,%1?8473%;/3-50F% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%�50-24%D50%H5912,7/;% %% %% %% %% %% %% %% %% %% %% GG %% %% %% %%�50-24%D50%�,-0/10/,/?0;% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%!,I47,/%'?;7,/;;%@/67;-02-75,% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%+<<7-75,24%;/0E73/;% %% J% %% %% J% K% %% %% %% %% K% K% %% K% K%� %% �05E7</<% GG �05E7</<%8L%2,5-:/0%2<97,7;02-75,M%%%% #5-%105E7</<%%% �05E7</<%8L%2,5-:/0%2<97,7;02-75,M%#5%7,D5092-75,%5DD/0/<%NOPQRPSRTUVWXSPYXTZVST[\] V_\XQOT TabcdQXccTcXSWdeXcT[\] V_\XQOT TabcdQXcc�fg�h�ijfkl�mfnlogfpflm�qhklrsotu�vfkw�sxyt�z�fmlxkfpflm�gkhxmhom�glonfqlg�hxm�h�kskhy�gkhxmhom�qsnlohrl�sp��{|�}lysv�kwl�hnlohrl~�� fkw�kwl�l�ql�kfsx�sp��j}yfq�klxmlog���s}g�fx��j}yfq�glqkso��kwhk�hol�spplolm�}t�hyy�qfkflgu�hyy�skwlo�glonfqlg�whnl�h�ysv�gkhxmhom�qsnlohrl��ho�g~���ykwsjrw�kwl���|�sp�kwl�qfkflg�hol�qsxgfmlofxr�h�ml�yst�lxk�sp�h��sokhy�glonfql�pso�qs��hxflg�hxm�lxkol�olxljog�fx�somlo�ks�lxwhxql�hxm�phgklx�}jgfxlgg�olyhkfsx�vfkw�kwl��jxfqf�hyfktu�[\] V_\XQOT`TabcdQXcc�olhqwlg�h��lmfj��mfnlogfkt�gqsol���~�z��kwhku�hqqs��hxflm�}t�h��lmfj��xj�}lo�sp�gkhxmhom�glonfqlgu�gwsvg�kwhk��jos�lhx�qfkflg�ms�xsk�vfmlyt�gwhol�h�qs��sx�hroll�lxk�fx�klo�g�sp��osnfmfxr�hx�hxgvlo�ks�l��yst�lxk�hxm�}jgfxlgg�ks�fqg~���#B*@!#� �#$% +,-./01% '203/45,2% '/06/,% '74825% '7097,6:29% '5456,2% �,;3:/</% =/4;7,>7% �742,% �?0372% @7A/>2% $2447,,% $:/%=26?/% $?07,% B7/,,2%@/3L347,6%120>;% %% %% %% %% %% %% %% %% GG %% %% %% %% %% %%�?8473%�051/0-L%� 27,-/,2,3/% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%'?4>L%.2;-/%3544/3-75,% %% %% %% %% %% %% %% %% GG %% %% %% %% %% %%!1/02-75,24%7,D5092-75,% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%+70%�?247-L% %% %% %% %% %% %% %% %% %% %% %% %% %% %% %%+<<7-75,24%;/0E73/;% %% K% %% %% J% K% %% %% K% K% %% K% %% K% K%T%% �05E7</<% GG �05E7</<%8L%2,5-:/0%2<97,7;02-75,M%%%% #5-%105E7</<% %% �05E7</<%8L%2,5-:/0%2<97,7;02-75,M%#5%7,D5092-75,%5DD/0/<%NOPQRPSRTUVWXSPYXTZVST[QWdSVQ\XQOTcXSWdeXcT

Page 108: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� �� �� �!"#$%& '�" ()�& �*���+$,"-�. �-,"/&%&" �"!�%*. 0.,0�%�,/&#�%1�2345673893:��.%�.�'""0�%*. 0.,0�#"/�,.'��"!��;<1�=$*�&*��.%�" ()�>�%*. 0.,0�%�,/&#�%?�@�/�,*��(�%%1�0&/�,%&*)�%�"A%�.��&'��#.*�'",)�B.,C�D;?EFG�-,"/& '�*�.*�&*�&%�.�/�,)�.#*&/��#.*�'",)?�H##",0& '�*"�*�.*1�#&*&�%�.--�.,�*"�"!!�,�/�,)�A&0��. 0�0&!!�,� *�.00&*&" .(�%�,/&#�%�*"�*��&,�#&*&I� ,)1�" ()�#" /�,'& '�& �,�#)#(& '�%�,/&#�%1�-,"-�,*)�B.& *� . #��. 0�A.%*��#"((�#*&" ?�J�&%�%&*$.*&" �&%�(&C�()�*"�#�. '��& �*���!"(("A& '�)�.,%�%& #��+$,"-�. �!$ 0& '�(& �%��./��&0� *&!&�0�2345673893:�%�,/&#�%�.%�.�-,&",&*)�!",��KL"/�, B� *��/"($*&" ?�J��,�!",�1�*���*"*.(� $B=�,�"!�%*. 0.,0�%�,/&#�%�&%�(&C�()�*"�',"A�. 0�*���0&/�,%&*)�B.,C�&%��M-�#*�0�*"�0�#,�.%��& �*���!"(("A& '�)�.,%?�NOPQRSRNQT UVWXYZ[T \]ZY_V]T \YZaYVT \b_c]T \bZdbVae]dT \_aV]T QVfeYgYT hY_fbVibT jb_]VT jkZb]T lbmYi]T n]__bVVT noh]akYT nkZbVT pbYVV]TqYZf V]_TRYZWbrb]WYfT TT TT TT TT TT TT TT TT TT TT ss ss TT TT TTq [k_]WbVTlYabfWZtTuYVfkfvT TT TT TT TT TT TT TT TT TT TT ss ss TT TT TTlYfbgYV YTlYabfWZ]WbVT TT TT TT TT TT TT TT TT TT TT ss TT TT TT TTwbabW]_Tg kdYVWfTj ]V]aYdYVWTftfWYdTT TT TT TT TT TT TT TT TT TT TT ss TT TT TT TTxVy_bVYTPbVYfTR Vfk_WT]VgTq]tdYVWT TT TT TT TT TT TT TT TT TT TT ss TT TT TT TTxVy_bVYTn]zYfT]VgTPYYfTq]tdYVWfT TT TT TT TT TT TT TT TT TT TT TT TT TT TT TTxVy_bVYTj ]ZZb]aY{Rb|b_T}VbVTlYabfWYZT TT TT TT TT TT TT TT TT TT TT TT ss TT TT TT}Wb_bWbYfT TT TT TT TT TT TT TT TT TT TT TT ss TT TT TTUggbWbV]_TfYZ|bYfT TT TT TT ~T ~T ~T ~T TT ~T TT ~T ~T ~T TT �T� TT qZ |bgYgT ss qZ |bgYgTctT]V WeYZT]gdbVbfZ]WbVoTTTT � WT[Z |bgYgTTT qZ |bgYgTctT]V WeYZT]gdbVbfZ]WbVoT� TbVr Zd]WbVTrrYZYgT�:�3��6���7496��9��76��5�9����9��9645�9���5�9����9�&%�*���(�%%�#"/�,�0�+$,"-�. �#.*�'",)�"$*�"!�*��� & ��#.*�'",&�%�#" %&0�,�0�& �*�&%�%$,/�)�A&*��.�%*. 0.,0�#"/�,.'��"!���?>;<1�.(*�"$'��&*��.%�" ��"!�*����&'��%*� $B=�,�"!����������������������������� ¡����������������������¢�£� �&%&" �/.,&�%�!,"B�.�B& &B$B�"!�>F<�*"�.�B.M&B$B�"!��;<1�%�"A& '�*�.*�*��,��&%� "�.',��B� *�=)�*���-.,*&#&-. *�#&*&�%�& �*�,B%�"!�(&!��#)#(��%�,/&#�%?��J���B"%*�#"/�,�0�%�,/&#�%�.,��¤9�5�93�9�69�5�:6�:5731�.�%�,/&#��*�.*���(-%�,�'$(.*& '�#&*&I� �#" *.#*�. 0�-,"/&0�%�. �"!!&#&.(�A,&**� �,�#",0�"!�#&*&I� %�. 0�¥96�73����96:5�5��:9�1�.�%�,/&#��*�.*�"!!�,%�.�#�,*.& � $B=�,�"!�$%�!$(��5�9����9�#�,*&!&#.*�%�%$#��.%�B.,,&.'�1�0�.*��",�=&,*�?�¦�%&0�%1�0&/�,%&*)�,�.#��0�. �& *�,B�0&.*��B.,C�D;?F§G�%�"A& '�*�.*�+$,"-���.%�.�¨$&*��-"-$(.*�0�#�.,*�,�"!�(&!��#)#(��%�,/&#�%�*�.*�/.,)�!,"B�" ��#"$ *,)�*"�. "*��,?��

Page 109: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#�!$%# !&'()*+# ,-*.)/0&-# ,)*1)&# ,2/3-0# ,2*42&15-4# ,0/01&-# %&6.5)7)# 8)/62&92# :2/-&# :;*.2-# $2<)9-# =-//2&&# =5)#8-1;)# =;*2&# >2)&&-#80;62&1#?0*#"0(# &.04)6# ## ## ## ## ## ## ## ## ## ## ## ## @@ ## ##A26-3/)#B-*92&1#B)*42'6## ## ## ## ## @@ ## ## ## ## ## ## ## ## ## ##A26.0;&'6#?0*#"0(# &.04)6# ## ## ## ## ## ## ## @@ ## ## ## ## ## ## ##�*1-&2C2&1#B;3/2.#%D)&'6# ## ## ## ## @@ ## ## ## ## ## ## ## ## ## ##E-42/F#!27G�-/.;/-'20&# ## ## ## ## ## ## ## @@ ## ## @@ ## ## ## ##B5F62.2-&6#A-'-3-6)# ## ## @@ ## @@ ## ## ## @@ ## ## ## ## ## ##!772'20&-/#�)*D2.)6# H# I# J# ## ## K# ## H# L# J# J# M# J# M# I#NNNN## B*0D27)7# @@ B*0D27)7#3F#-&0'5)*#-742&26*-'20&O#### P0'#+*0D27)7### B*0D27)7#3F#-&0'5)*#-742&26*-'20&O#P0#2&?0*4-'20&#0??)*)7#QRSTUSVUNWXYZVS[ZN\XVNQX]S_N]SVZNZVY]Z NQX]S_N]SVZ�ab�c�defgh�iajk�l�amnojapanm�bjcomcem�bneqarnb�com�c�bjcomcem�rfqnecdn�fp�stuv�wknenpfenx�aj�rfgym�zn�bjcjnm�jkcj�jkab�rcjndfe{�ab�inyy�nbjczyabknmv�|j�kcb�jf�zn�}nojafonm�jkcj�cyy�bjcomcem�bneqarnb�cen�enycjnm�jf�jkn�rcen�fp�onnm{�hnfhynx�}c{�aj�zn�ao�jne}b�fp�kfgbaodx�mabrfgojb�fe�pc}ay{�cyyficornbv�wkn�}fbj�rfqnenm�bjcomcem�bneqarnb�ab�~X�`T[N\XVN_X�NT]X� Z �c�bneqarn�jkcj�knyhb�paomaod�kfgbaod�pcrayajanb�pfe�hnfhyn�iajk�nrfof}ar�onnmbv��o�jkn�fjkne�kcomx�QX]S_N]SVZNaorygmnb�s�mappnenoj�cmmajafocy�bneqarnb�com�znaod�bf�pce�jkn�}fbj�hfhgycjnm�rcjndfe{�endcemaod�bneqarnb�ofj�encrkaod�jkn�}aoa}g}��tu�fp�rfqnecdnv�|o�cmmajafox�maqnebaj{�encrknb�jkn�kadknbj�brfen��tv����ao�jkab�rcjndfe{�bkfiaod�c�kadk�mabrfemcorn�ao�QX]S_N]SVZ�enycjnm�cmmajafocy�bneqarnb�ikark�}ncob�c�dencj�crjaqaj{�ao�jkab�cencv�=$!P�B�$=# !&'()*+# ,-*.)/0&-# ,)*1)&# ,2/3-0# ,2*42&15-4# ,0/01&-# %&6.5)7)# 8)/62&92# :2/-&# :;*.2-# $2<)9-# =-//2&&# =5)#8-1;)# =;*2&# >2)&&-#B;3/2.#=*-&6+0*'#�)*D2.)6# ## ## ## ## ## ## @@ ## ## @@ ## ## @@ ## ##�-*#B-*92&1# ## ## ## ## ## ## ## ## ## ## ## ## @@ ## ##=*-??2.#� )3#�-4)*-6# ## ## ## ## ## ## ## ## ## ## ## ## @@ ## ##!772'20&-/#6)*D2.)6# ## I# ## I# J# H# J# J# H# ## ## J# ## ## J#N## B*0D27)7# @@ B*0D27)7#3F#-&0'5)*#-742&26*-'20&O#### P0'#+*0D27)7### B*0D27)7#3F#-&0'5)*#-742&26*-'20&O#P0#2&?0*4-'20&#0??)*)7#NQRSTUSVUNWXYZVS[ZN\XVN�VST �XVRNZVY]Z N

Page 110: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#�$�%&�'�&()*%'+�*',)-./0�&%1*)�%,�2'&�,2)�&3'++)&,�&),�.4�&,'15'/5�&)/6%*)&7�.1+0��7�'+,2.8-2�9�.4�,2)3�:%,2�'1�);*)++)1,�*.6)/'-)�.6)/�<�=>�?2)/)4./)7�,2%&�(/.6)&�,2',�@8/.()'1�*%,%)&�'/)�':'/)�.ABCDEBEFCGEHEBIHJKGCLMNEBKABICOBNICIPEMOQBHKRISICTUB�VMBLWWICIKMXBWIYEGOICTBDLOBCDEBOENKMWBDIZDEOCBNLCEZKGIEOQBHLG[B\]U _ BODKaIMZBCDLCBCDEGEBLGEB5%&('/',)���� !"#�$�&)/6%*)&�,2',�6'/0�4/.3�.1)�*%,0�,.�'1.,2)/�2%-2+%-2,%1-�%,�'&�'1�)3)/-%1-�'*,%6)�*',)-./0>��b-'%17�'&�*.33)1,)5�%1�,2)�+'&,�&)*,%.17�,2%&�81)c8'+�&%,8',%.1�%&�);()*,)5�,.�*2'1-)�&%1*)���� !"#�$�/)+',)5�.1d+%1)�&)/6%*)&�2'6)�e))1�%5)1,%4%)5�'&�.1)�.4�,2)�f)0�&8**)&&�4'*,./&�4./�,2)�)6.+8,%.1�.4�)dg.6)/13)1,�&)/6%*)&�58/%1-�,2)�1);,�0)'/&>�?2)/)4./)7�,2)�183e)/�.4�&,'15'/5�&)/6%*)&�%&�+%f)+0�,.�-/.:�'15�5%6)/&%,0�4'*,./�%&�);()*,)5�,.�5)*/)'&)�%1�,2)�4.++.:%1-�0)'/&>�hijklmnokllplqm krstuvwm jxvyuz{rxm juv|urm j}z~x{m j}v�}r|�x�m j{z{|rxm �r�y�u�um �uz�}r�}m �}zxrm ��vy}xm i}�u�xm �xzz}rrm ��um�x|�um ��v}rm �}urrxm�{r�sv�ys}{rmnuv�}s�m mm mm mm mm mm mm mm mm mm mm mm mm mm mm mmoxr����umnzxr�mxr�mj�}z�}r|miu|�zxs}{rm mm mm mm mm mm mm mm mm mm mm mm mm mm mm mm�uvs}�}yxsu�mm mm �� mm mm mm mm mm mm mm mm mm mm mm mm mmk��}s}{rxzm�uv�}yu�m mm mm mm mm mm mm �m mm �m mm mm mm mm �m �m� mm nv{�}�u�m �� nv{�}�u�m~�mxr{s�uvmx��}r}�vxs}{r�mmmm l{smwv{�}�u�mmm nv{�}�u�m~�mxr{s�uvmx��}r}�vxs}{r�ml{m}r�{v�xs}{rm{��uvu�m�$� ������#��������#������ �"�� � ��!������!����� �"�� � �7�*.1&%5)/%1-�%,&�+.:�183e)/�.4�&,'15'/5�&)/6%*)&� ¡8&,��¢7�%&�.1)�.4�,2)�,:.�3.&,�*.6)/)5�*',)-./%)&>�b�,.,'+�.4�£�=�.4�,2)�*%,%)&�'53%,�(/.6%5)�'�f%15�.4�.1d+%1)�*.1&,/8*,%.1�()/3%,�./�.44)/�+'15�8&)�(+'1&�'15�e8%+5%1-�,.�%,&�*%,%¤)1/0>�¥','�&2.:&�6)/0�+%,,+)�WIOJEGOIKMBIMBOCLMWLGWBOEGYINEOQBJGKYIOIKMBLMWBLBNKHHKMBLJJGKLNDBCKB���� �"�� � ��,.(%*&�e0�'+3.&,�'++�,2)�('/,%*%('1,�*%,%)&>��?2)�+.:�5%6)/&%,0�4'*,./� �>�9¢�'+.1-�:%,2�,2)�&3'++)&,�&),�.4�&)/6%*)&�&2.:&�,2',�,2)�381%*%('+%,%)&�*.6)/)5�'�&3'++�&),�.4�&,'15'/5�&)/6%*)&�'15�'/)�5.%1-�1.�3./)�'*,%6%,0�%1�,2',�'/)'7�(/.6%1-�,2',�*%,%)&�*.1&%5)/�4):�&%3%+'/�'15�'+/)'50�)&,'e+%&2)5�(+'1&�%1����� �"�� � �d/)+',)5�'55%,%.1'+�&)/6%*)&>��

Page 111: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#$!%� &'�()#*)(!$�&��+� '�(!$�&���,��-./012/3�456547�/89�:9;<=>9?:�@ABC;=BD�8A:�E99F�@9A:C;9G�:=F>9�B89�E9H=FF=FH�IJ�9KLI<9;F@9FBM�E9=FH�B89�@I:B�A>>9NB9G�G9J=F=B=IF�OPQRSTUVRWXOUVYOZR[Q\Q[]R VT T]Q_R ZRaXUWRbRcYRdXYTRYeROPQRfghijklimnopqrstuvwxyxtqzty{|q}ys~�y~�t�q�y�����AC@����=�-A=IM������,��C;�:C;<9D�C:9:�B8=:�J;A@9�I;��A:�B89�:BA;B=FH�NI=FB�9�B9FG=FH�B89�;AFH9�IJ�@ABC;=BD�BI�=F>�CG9�A�F9��J=JB8��9<9��IF�B89�BIN�BI�;9J�9>B�A�@I;9�AG<AF>9G�G9H;99�IJ�@ABC;=BD�A:�=:�A�;9AGD�;9>IHF=�9G�=F�B89��A:B�;9�9A:9:�IJ�B89��ANH9@=F=�E9F>8@A;��;9NI;B���ANH9@=F=M������,�2F�IC;�>A:9M�B8=:�F9���9<9��;9J�9>B:�B89�F9��J9ABC;9:�=F�C:9;�=FB9;A>B=IF�BA�=FH�=FBI�A>>ICFB�B89�C:9�IJ�9@9;H=FH�B9>8FI�IH=9:�@A=F�D�EA:9G�TeR� Q R����R�PY]ReQ�ROVQe_]RY]R VTWTOYe�R�YOY�Qe]�R XVOY�Y XOYTeRYeR]QV\Y�QR VT\Y]YTe�R�UX[YOZRQ\X[UXOYTeRXe_R_Y]�U]]YTeRX TUOR]QV\Y�Q]�RYW^VT\QWQeO�RPQe�QROPQReQ�R[Q\Q[RY]R�X[[Q_Rf XVOY�Y XOYTeR[Q\Q[��R�/89�@9A:C;9@9FB�J;A@9�I;��C:9G�=F�B8=:�:C;<9D�=:�EA:9G�=F�B89�J=<9��9<9�:�G9J=F9G�E9�I� �A,�2FJI;@AB=IF�49<9� �.E=�=BD�BI�IJJ9;�;9�9<AFB�=FJI;@AB=IF�=F�A�IF9�EA:=>�N;I<=:=IF�N;I>9::M�C:CA��D��9E:=B9�EA:9G,�0:9;:��=���IF�D�E9�AE�9�BI�>IF:C�B�I;�;9>9=<9�=FJI;@AB=IF�AFG�FIB�BI�GI�F�IAG�AFD��=FG�IJ�JI;@�I;�H=<9�J99GEA>�,�2B�=:�9::9FB=A��D�B89�J=;:B�:B9N��89F�9FB9;=FH�B89��9E��I;�G,�¡¢w�£�y¤�GA=�D�>C�BC;A��AH9FGA,�E,�2FB9;A>B=<=BD�49<9��0:9;:�>AF�H9F9;AB9�EA:=>�>IFB9FB�K9@A=�M�B9@N�AB9M�¥C9;=9:K��8=>8�=:�=FB;IGC>9G�=F�B89�HI<9;F@9FB�GABAEA:9:,�2B�=:�A�:I�A�IF9K�AD�=FB9;A>B=IF�N;I>9::�=F��8=>8�B89�C:9;�>AFM�JI;�=F:BAF>9M�GI�F�IAG�I;�:CE@=B�JI;@:�IFK�=F9,�¦I�9<9;M�B89�N;I>9::�>AFFIB�E9�>I@N�9B9G�IFK�=F9�AFG�:I@9B=@9:�N8D:=>A��ABB9FB=IF�AB�B89�@CF=>=NA��IJJ=>9:�=:�;9¥C=;9G,�¡¢w�£�y¤�IFK�=F9�GI�F�IAG�IJ�B89�;9¥C=;9G�JI;@�BI�N9;JI;@�A�:>8II��:CE:>;=NB=IF,��F>9�=B�=:�J=��9GM�B89�C:9;��=���:9FB�=B�EA>��IFK�=F9,�>,�/;AF:A>B=IF�49<9��/;AF:A>B=IF:�KNAD@9FB:M�>9;B=J=>AB9:K�>AF�E9�>I@N�9B9G�B8;ICH8ICB�AF�9�9>B;IF=>�>8AFF9�,�/89�A>8=9<9@9FB�IJ�B8=:��9<9��=@N�=9:�B89�NI::=E=�=BD�BI�>I@N�9B9�D�N9;JI;@�A�:9;<=>9�IFK�=F9�B8;ICH8�B89�C:9�IJ�9�9>B;IF=>�@9AF:,�/89;9JI;9M�B8=:�=:�A�B�IK�AD�=FB9;A>B=IF��=B8�B89�C:9;�=F��8=>8�FIB�IF�D�B89�JI;@:�>AF�E9�GI�F�IAG9G�IFK�=F9�ECB�A�:I�:9FB�EA>��IFK�=F9,�¡¢w�£�y¤tIFK�=F9�A>¥C=:=B=IF�AFG�NAD@9FB�IJ�@CF=>=NA��@C:9C@:�B=>�9B:,�

Page 112: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#$% &!'(%"�)*+*,�-.,,�!+!(,!/(,('0�%"1,("*2�'3!'�(#2�$.,,�("'*4 !'(%"�$% �!,,�*15%+* "&*"'�#* +(6*#�("'%�!�#("4,*�7% '!,��8#* #�6!"�!66*##�!,,�#* +(6*#�$ %&�!�."(9.*�+( '.!,�%$$(6*�$ %&�!"0�7,!6*���3(#�,*+*,�:(,,�.#.!,,0�(&7,0�!�/.#("*##� **"4("** ("4�7 %6*##�/%'3�("�'3*�/!6;�%$$(6*�!"��<=>?@ABC<?B@@DE>?B@?F?ED<GHI?FJKDCDI<AF<DBCL?MBA>BN>AO?D<?PBQRJ?FRIB?DKSRG?FC?DC<>ACFR?E=FCT>?DC?BAJ>A?<B?DC<>TAF<>?FI?KQE=?QI>AHI?DC@BAKF<DBC?FI?SBIIDUR>?DC?<3*�$ %"'�%$$(6*�#%�'3!'�('#��*,(+* 0�:%.,��."�* 4%�!�6%&7,*'*�63!"4*�("�% �* �'%�&!;*�('�!#�*!#0�!"��6,*! �!#�7%##(/,*���3(#�:%.,��&*!"�("�&!"0�6!#*#�'3!'�7* #%"!,�("$% &!'(%"�$ %&�.#* #�:(,,�/*�7 *+(%.#,0� *&*&/* *��#%�'3!'�:3*"�7* $% &("4�!"0�;("��%$�#* +(6*�%",0�(�*"'($(6!'(%"�(#�"**�*���VWXYZ[\] *9.( *&*"'�%$�(�*"'($(6!'(%"�_(�*��*,*6' %"(6�ab�:3*"�$(,,("4�'3*�$% &�'%�!#;�@BA?F?EBCI<AQE<DBC?S>AKD<L?c=>?A>KFDCDCT?QI>AHI?DC@BAKF<DBC?PDRR?U>?FQ<BKF<DEFRRG? *&*&/* *��!"��("' %�.6*��("�'3*�$% &�/0�'3*�#0#'*&��*��d! '(6(7!'(%"�)*+*,�e"!/,*�!"��7 %&%'*�6('(f*"#�7! '(6(7!'(%"�("�!,,�*1g>ANDE>HI?R>N>RI?IB?<=F<?BSDCDBCI?FA>?'!;*"�("'%�!66%."'�'%�*"3!"6*�9.!,('0�!"��*$$*6'(+*"*##��h,#%�('�(#�*i7*6'*��'3!'�'3*�.#* � *6*(+*�#%&*�#% '�%$�$**�/!6;�$% �'3%#*�6%"' (/.'(%"#���3(#�j'3�,*+*,�7 %+(�*#�!"�("�(6!'(%"�%$�'3*�*i'*"'�'%�:3(63�%"1,("*�7 %+(#(%"�(#�/!#*��%"�"*:�&%�*,#�'%�.#*�!+!(,!/,*�("$% &!'(%"�%/'!("*��$ %&�'3*�6('(f*" 02� *!63("4�!�3(43��*4 **�%$�7 %1FE<DN>C>II?DC?I>ANDE>IH?J>RDN>AGL?kC?I=BA<O?<=DI?R>N>R?DI?UF#*��%"�'3*�!77,(6!'(%"�%$�'3*�l */���m�7 *&(#*#�'%�'3*�%"1,("*�&."(6(7!,�#* +(6*#��*,(+* 0��VWXYZ[\n�7%##(/(,('(*#�'%�&!;*�6%&&*"'#�%"�!"0�*+*"'�!""%."6*��("�'3*�6.,'. !,�FT>CJF?FCJ?I>>?B<=>A?QI>AHI?EBC<ADUQ<DBCIL�!��"$% &!'(%"�� h/(,('0�'%�%$$* � *,*+!"'�("$% &!'(%"�("�!�%"*�/!#(6�7 %+(#(%"�7 %6*##2�.#.!,,0�:*/#('*�/!#*���/��"'* !6'(+('0�� 8#* #�6!"�4*"* !'*�/!#(6�6%"'*"'�1*&!(,2�'*&7,!'*2�9.* (*#1�:3(63�(#�("' %�.6*��("�'3*�4%+* "&*"'��!'!/!#*#��6��� !"#!6'(%"�� � !"#!6'(%"#�17!0&*"'#2�6* '($(6!'*#1�6!"�/*�6%&7,*'*��'3 %.43%.'�!"�*,*6' %"(6�63!""*,������ !"#$% &!'(%"�� -.,,�!+!(,!/(,('0�!"��("'*4 !'(%"�$% �!,,�*15%+* "&*"'�#* +(6*#��8#* #�6!"�!66*##�!,,�#* +(6*#�$ %&�!�."(9.*�+( '.!,�%$$(6*�$ %&�!"0�7,!6*��*��d! '(6(7!'(%"�� e"!/,*�!"��7 %&%'*�6('(f*"#�7! '(6(7!'(%"�("�#* +(6*�7 %+(#(%"2�9.!,('0�>NFRQF<DBC?FCJ?JDIEQIIDBC?FUBQ<?I>ANDE>HI?DKSABN>K>C<L�opqrstuvtwsxyz{s|}t~p��xz��trsysr|c=>?F@BA>K>C<DBC>J?R>N>RI?=FN>?U>>C?QI>J?DC?<=DI?IQAN>G?<B?FII>II?ED<GHI?KF<QAD<G?@BA?>FE=?BC>?%$�'3*�#* +(6*#�("6,.�*��("�'3*�5,%/!,��('(*#��!'!,%4.*��h66% �("4�'%�%. �&*'3%�%,%402��!'!�!##*##&*"'�3!#�/**"�7* $% &*��'3 %.43�#*,$1*+!,.!'(%"���3*��!'!�4!'3* ("4�7 %6*##�:!#�7* $% &*��'3 %.43�'3*�.#*�%$�!�9.*#'(%""!( *�'3!'�:!#�#*"'�'%�!,,�'3*�#. +*0*��6('(*#��"�'3(#�9.*#'(%""!( *2�6('(*#�:* *�!#;*��'%�#*,$1*+!,.!'*�'3*�&!'. ('0�%$�'3*( �#* +(6*#�!66% �("4�'%��!/,*������

Page 113: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ����� �!"�#!���$%�!&"%'%�"!&(�"���$�%')%�"*%�+!,,%'%&-%�$%"#%%&�"*%�./0!.1.��%)%��/&+�"*%�.!&!.1.��%)%��/-*!%)%+�$2�/�-!"2�,�'�/���"*%�-/"%(�'!%���1$.!""%+�"���"1+23�4��+���� �#%�+%,!&%�"*%�,����#!&(�+%�"/�,/-"�'5�67879:;<9=979:>=?<>@7ABCBA7D�.!&!.1.�./"1'!"2��%)%��67E7F�G%�"/�.1�"�$%�H��!"!)%�$2�+%,!&!"!�&3�4*!��,/-"�'�#!����*�#�*�#�*�.�(%&%�1��"*%�-!"2�%IJ�)%'&.%&"��,,%'�!�3��K&��&%�*/&+ �!,�+%�"/�!��*!(* �/���-/"%(�'!%��#!���*/)%�/�*%"%'�(%&%�1��./"1'!"2�+%)%��H. %&"�H'�)!&(�"*/"�"*%�-!"2�!��+%)�"!&(�!"��%,,�'"��"��%)��)%�,!'�"�-%'"/!&��%�%-"%+�-/"%(�'!%�3�K&�"*%��"*%'�*/&+ �/���#�+%�"/�,/-"�'�#!����*�#�*�#�/���-/"%(�'!%���%%.�"��*/)%��!.!�/'�./"1'!"2�+%)%��H.%&"3�4*%'%,�'% �!"�#!���H'�)%�"*/"�"*%�-!"2�!��%)��)!&(�/���!"���%')!-%��*�.�(%&%�1��23���3L�MNOPNM4QKR�KS��GKM4QKR�TNUNTV�4*%��&I�!&%��%')!-%�/+�H"!�&�.%/�1'%��"*%��*/'%��,�-!"!W%&��"*/"�/'%�1�!&(�"*%�%IV%')!-%�,'�.�"*%�"�"/��H�H1�/"!�&�"��#*!-*�"*%�%IXB?C<YB7<Z7>:?[B>B\]7 <><_B aZ7:\bc><b 7bd7>eB7bddB?B\7BIV%')!-%��!��/�f%2�,/-"�'�"��/��%���"*%�'%/���1--%����,�/&�%IJ�)%'&.%&"�"'/&�,�'./"!�&3�Q"�H'�)!+%��)/�1/$�%�!&,�'./"!�&�"���1--%��,1��2�./&/(%���-/��%IJ�)%'&.%&"�H��!-!%��/&+�"*%�c?bgBY>aZ7cb?>dbA<b]7hZ7:7Yb ZBi=B YBj7>eB7= \B?Z>: \< [7bd7ebk7>b7< Y?B:ZB7Y<>!W%&�/#/'%&%���/&+�"/f%�1H��,��&I�!&%��%')!-%��!��-'!"!-/��"��%&*/&-%�"*%�H%',�'./&-%��,�/&2�N1'�H%/&�.1&!-!H/�!"23�l�#%)%' ��&�2�,%#��"1+!%��,/-%�"*%�/+�H"!�&�.%/�1'%.%&"��,�%IV%')!-%�+1%�"��"*%�!&*%'%&"�+!,,!-1�"!%��m%IP!"!W%&�M'�n%-" �Loopq�mN�+�4���f!" �Loorq3��4��.%/�1'%�/+�H"!�&��,�%IV%')!-%� �,!'�"��,�/�� �"*%'%�!��"*%�&%%+�"��,!&+�/��1!"/$�%��%"��,�!&+!-/"�'�3�4*%�.%/�1'%���,�)!�!"��!&�/��H%-!,!-�#%$��!"%��'�%)%&�"*%�&1.$%'��,�H/(%���%')%+�/'%�$�"*�!&+!-/"�'��-�..�&�2�/--%H"%+�"��/��%���"*%�H�H1�/'!"2��,�/�#%$��!"%3�s1"�/+�H"!�&��,��&I�!&%��%')!-%��!��&�"��&�2�/--%��3�Q&�"*!��-/�% �+/"/�"'/&�/-"!�&��H%',�'.%+�"'�1(*�%IV%')!-%��/HH�!-/"!�&��#�1�+�$%�/�$%""%'�!&+!-/"�'3�l�#%)%' �!"�!��&�"�/HH�!-/$�%�"��/���'/&(%��,��%')!-%��+1%�"��"*%!'�+!,,%'%&"��%)%���,�./"1'!"23����/�-�&�%t1%&-% �"*%'%�!��&��1&!t1%�./(&!"1+% �"��.%/�1'%�/+�H"!�&�!&�/�-�&�!�"%&"�#/23�N/-*�%IV%')!-%�&%%+��"��+%,!&%�#*!-*�"*%�$%�"�./(&!"1+%�"��.%/�1'%�"/f%�1H�!�3�V%-�&+�2 �/�'!(�'�1��/+�H"!�&�.%/�1'%.%&"�.1�"�"/f%�!&"��/--�1&"�"*%�'%/��"/'(%"��,�%/-*��%')!-%�/&+�!"��!&*%'%&"�-*/'/-"%'!�"!-�3�N/-*��%')!-%�!��,�-1�%+��&�/��H%-!,!-�"/'(%"��,�-!"!W%&��u@b= [Z>B?Zj7BA\B?A@7cBbcABj7c:?B >Zj7 B >?Bc?B B=?Zj7<99<[?: >Zvw7ke<Ye7e:Z7\<ddB?B >7-*/'/-"%'!�"!-��m�!"%'/-2 �+!(!"/���!"%'/-2 �'/"!����,�! >B? B>7:YYBZZvw]7xeB?Bdb?Bj7>eB79:;<9=97/+�H"!�&��%)%��!��+!,,%'%&"�,�'�%/-*��%')!-%3�4*!���1')%2 �1&+%'�"/&+!&(�"*%�!.H�'"/&-%��,�.%/�1'!&(�/+�H"!�&�$1"�/����'%-�(&!W!&(�!"��!&*%'%&"�+!,,!-1�"!%��"*/"��1'H/��%��"*%��!.!"/"!�&���,�"*%�H'�n%-" �*/��+%,!&%+�/&�/�"%'&/"%�./(&!"1+%5�"*%�H%'-%H"!�&��,�/+�H"!�&3�4*%�H%'-%H"!�&��,�/+�H"!�&�!��+%,!&%+�/��"*%�/+�H"!�&�

Page 114: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !��"#$%"!�&!'(%)!�*&�%��%&�+!')!%(!,�-"���.!*&/'!,0�12�� !�&!'(%)!�+'�(%,!'3�%"��� !'�4�',&3�� !�$�)*$�+/1$%)�*,.%"%&�'*�%�"�,!+*'�.!"��'!&+�"&%1$!����� !�&!'(%)!�+'�(%&%�"5��6��&!��/+�*� �.�7!"!�/&�)'%�!'%�"3�*�.!*&/'!.!"���'*.!4�'8�1*&!,��"��%(!�$!(!$&� *&�1!!"�,!�%"!,5�9*) �$!(!$�,!�%"!&�*�,!7'!!����+!')!+�%�"����*,�+�%�"�!*&2����1!�'!)�7"%:!,�12�� !�&!'(%)!��4"!'��"�!*) �)%�23�*",�'*"7%"7��'�. �*�)�.+$!�!��*%$/'!����*�1%7�&/))!&&�*&�& �4"�%"��*1$!���1!$�4;���<5�=��*,�+�%�"� 6 !�&!'(%)!� *&�1!!"�*�)�.+$!�!��*%$/'!�*",�� !�./"%)%+*$%�2�%&�)�"&%,!'%"7�%�&�4%� ,'*4*$5�>5�?�4�*,�+�%�"� 6 !�&!'(%)!�%&�1!%"7�/&!,�12��!4!'�)%�%:!"&�� *"�+'!(%�/&$2�!@+!)�!,5��5�A!,%/.�*,�+�%�"� 6 !�&!'(%)!�%&�1!%"7�/&!,�12�� !�*(!'*7!����)%�%:!"&�!@+!)�!,5��5�B%7 �*,�+�%�"� 6 !�&!'(%)!�%&�1!%"7�4%,!$2�/&!,�*.�"7�)%�%:!"& %+5�C5�9@)!$$!"��*,�+�%�"� 6 !�&!'(%)!� *&�1!!"�*�)�.+$!�!�&/))!&&�*",�� !�./"%)%+*$%�2�%&�,!(��%"7�.�'!�B/.*"�D!&�/')!&�*",�%"(!&�.!"���"�%�5��EFGHIJKLJMINOPQIRSJTINQITUPVWJVXJFYVTUPVWJHIOIHRZ[\] _ ab]c]defad]g h]i]hj\_i] k]]d lj]g fd e\fjjlbi]mea _jj]jjnfemojp]bn]pefad a _gapefad^��'�!*) ��"!����� !�&!'(%)!&�%")$/,!,�%"�� !�q$�1*$�r%�%!&�)*�*$�7/!5�s7*%"3�� !�,*�*�7*� !'%"7�+'�)!&&�4*&�+!'��'.!,�� '�/7 �� !�/&!����*�t/!&�%�""*%'!�� *��4*&�&!"�����*$$�� !�&/'(!2!,�)%�%!&5�u"�� %&�t/!&�%�""*%'!3�)%�%!&�4!'!�*&8!,����&!$�#!(*$/*�!�� !�+!')!+�%�"����*,�+�%�"����� !�&!'(%)!&�� !2�+'�(%,!�*))�',%"7����6*1$!��5�6 %&�.!*&/'!3�*$� �/7 �%&�1*&!,��"�&!$�#*&&!&&.!"�3�%"��'.&�*1�/��� !�&/))!&&����� !�!#v!'(%)!&�+'�(%,!,�12�!*) �./"%)%+*$%�2�*",�.*8!&�+�&&%1$!�� !�*"*$2&%&����!@%&�%"7�'!$*�%�"&�wfe\ j]bifn]ojc_elbfem_dg _gapefadx��5��9yDz{9s=�s|9Dsq9v�{9D�rs69qzD}�6 !�"!@���%7/'!&�& �4�� !�9/'�+!*"�*(!'*7!&���'�!*) �)*�!7�'2����� !�q$�1*$�v!'(%)!&�r*�*$�7/!5�6 !&!�*(!'*7!&�%")$/,!�*$$�&!'(%)!&�&/1.%��!,����&�/,2�12�*$$�)%�%!&3�4 !� !'�&!'(%)!&�*'!�&�*",*',��'�*,,%�%�"*$5�~/'� !'�,�4"3�� !�'!&/$�&�*'!�+'!&!"�!,�*))�',%"7����� %&�)$*&&%�%)*�%�"5�6 !�9/'�+!*"�*(!'*7!�.*�/'%�2�$!(!$�%&�>5�3�4 %) �.!*"&�� *��� !�.*�/'%�2�*(!'*7!����$�)*$�!#�ai]bdc]dej]bifn]j a e\] j_cph] ga]jdoejefhhb]_n\ eb_dj_nefad_hh]i]h�x�dhm ewa ale a ^� !�"%"!�)*�!7�'%!&��������������*",������Z��������Z��*'!�'!*) %"7��'�&$%7 �$2�&/'+*&&%"7�$!(!$��5��6 !�9/'�+!*"�*(!'*7!�+!')!+�%�"����*,�+�%�"�%&��5C�4%� �*�.*@%./.�����5�5�6 %&�& �4&�� *��$�)*$�!#q�(!'".!"��&!'(%)!&�*'!�1!%"7�/&!,�12�*"�*(!'*7!����/&!'&� %7 !'�� *"�%"%�%*$$2�!@+!)�!,5�s$� �/7 3�*$$�)*�!7�'%!&�*) %!(!�+!')!+�%�"����*,�+�%�"�&)�'!&�*$4*2&�*1�(!��3�� /&�+'�(%"7�� *��� !�+!')!+�%�"����*,�+�%�"���'�*$$�� !�)*�!7�'%!&�%&�+�&%�%(!3�%��)*"�"���1!�)�"&%,!'!,�*&� %7 $2�*,�+�!,5�

Page 115: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� ��� �� �� !�� !�� "�� "�� #�� #�� ���$%&'()*+'((,(-.%'(/*0%./01,'+)1'%2+,32141+2)2(5,%0(62(.26*+0462(.)7)&$/,(2//28$1'.,0(19'((2++,(-1,.,:2(/)2(-'-262(. 6'.$%,.4*2%12*.,0()03)'80*.,0( ;<=>?@ABC;BDA>BEA;@A>FA@GH?C;?I;BJ?@GH?C;BCJ;KBG=>HGL;MADAMN;OPQRSTU�VQPVWXYZQP�RS[P�V[P�\T�T]RU[VRT �_UQ �RST�_ZaXUT�ZY�RS[R�ZR�YTT Y�PQR�RQ�\T�VQUUTW[RZQP�bcdeccfghcijklcmhgnodpikdqgofrgscilcsdktfgtugortsdktfv�wTa[U ZPa�YxTVZ_ZV�V[RTaQUZTYy�zS[PPTWWZPa�ZY�RST�V[RTaQU{�|ZRS�\TRRTU�`[RXUT�YTU}ZVTYy�UT[VSZPa�[��~�� [U��|SZVS�ZY�YQ T|STUT�\TR|TTP��U[PY[VRZQP[W�[P ��U[PY_QU [RZQP�WT}TWY~�OWRSQXaS�\TZPa�RST� QYR� [RXUT�V[RTaQU{�ZR�UT[VSTY�QPW{�[� T ZX �WT}TW�Q_�xTUVTxRZQP�Q_�[ QxRZQP~���ST�V[RTaQU{�|ZRS�RST�|QUYR�VQ \ZP[RZQP�Q_�_ZaXUTY�ZY��QVZ[W�V[UT~��ST�[}TU[aT� [RXUZR{�WT}TW�Q_�ZRY�YTU}ZVTY�ZY�RST�WQ|TYR��|ZRS�[��~��UTYXWR�RS[R� QTY�PQR�T}TP�UT[VS�RU[PY[VRZQP[W�WT}TW��RST�[}TU[aT�xTUVTxRZQP�Q_�[ QxRZQP�ZY�[Y�|TWW�RST�YTVQP �WQ|TYR~��_�VQXUYT�RSZY�S[Y�RQ� Q�|ZRS�RST�VQ xWT]ZR{�Q_�RST�V[RTaQU{�[Y�|TWW�[Y�|ZRS�RST�_[VR�RS[R�RSTUT�[UT�}TU{� Z__TUTPR�VQ xTRTPVTY�[P �YQVZ[W�xQWZVZTY~�<CDH>?CKACG�YSQ|Y�RST�YTVQP �|QUYT� [RXUZR{�[}TU[aT�[P �[�xTUVTxRZQP�Q_�[ QxRZQP�YWZaSRW{�\TWQ|�RST�[}TU[aT~��>BCN@?>G�[P ��>�BC;@MBCCHCE�[UT�_QWWQ|ZPa���BCCAMMHCE;V[RTaQU{�ZP�\TYR� [RXUZR{~��Q|T}TUy�ZP�RS[R�V[YTy�\QRS�UT[VS�RST�SZaSTYR� [U�Y�Q_�xTUVTxRZQP�Q_�[ QxRZQP~���������������������������������������� ���G;¡;�=NHCANN;[P ;¢HIAFLFMA;[UT�V[RTaQUZTY�RS[R�YSQ|� Z � [RXUZR{� [U�Y�[WQPa�|ZRS� Z �xTUVTxRZQP�Q_�[ QxRZQP�YVQUTY~�

Page 116: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������ !"#!$#�!"#�!##% %&"!'��($)%*(��!)($!+(�� ,-�./-0.1-2�.34/-�50-2-67-8�96:;<8-�.;;�2-0/9:-2�2<3=977-8�74�27<8>�3>�.;;�:979-2?�@,-7,-0�2-0/9:-2�.0-�27.68.08�40�.8897946.;��%6�7,-�6-A7�B91<0-2?�7,-�0-2<;72�.0-�50-2-67-8�.::408961�74�7,92�:;.229B9:.7946?�96�408-0�74�4BB-0�.�89BB-0-67�5-025-:79/-��CDC CDE FDC FDE GDC GDE HDC HDE IDC IDE EDCJKLMNOPQMNNRNSTKMNUPVKTUVWRMQOWMKXQRYXWZWQXOXN[RKVN\XNTX\PQVZ\XNTO]OLJURNXUUX JWMTRVNW_MNNXQQRNSWRTRXNUOXNSMSX\XNT UTMN MK OUXK[RWXUO\MTJKRTZUTMN MK OUXK[RWXUOPXKWXPTRVNOVYOM VPTRVN �abcdefcghcdichjklmglngeolhjklmgemogpejqdkjrgscbcstgnldgtjemoedogtcdbkictgCDC CDE FDC FDE GDC GDE HDC HDE IDC IDE EDCJKLMNOPQMNNRNSTKMNUPVKTUVWRMQOWMKXQRYXWZWQXOXN[RKVN\XNTX\PQVZ\XNTO]OLJURNXUUX JWMTRVNW_MNNXQQRNSWRTRXNUOXNSMSX\XNT M RTRVNMQOUXK[RWXUO\MTJKRTZM RTRVNMQOUXK[RWXUOPXKWXPTRVNOVYOM VPTRVN �abcdefcghcdichjklmglngeolhjklmgemogpejqdkjrgscbcstgnldgeookjklmesgtcdbkictg ,-�=.7<097>�;-/-;2�B40�.8897946.;�2-0/9:-2�.0-�.;@.>2�2;91,7;>�.34/-�7,-�;-/-;2�0-.:,-8�B40�27.68.08�2-0/9:-2?�@97,�7,-�-A:-57946�4B�udvemghsemmkmf�.68�wkncirisc��xlikesgiedc�.68�ymbkdlmpcmj�.0-�7,-�46;>�7@4�:.7-1409-2�B40�@,9:,�7,-�8927.6:-�.=461�27.68.08�.68�.8897946.;�=.7<097>�92�.34/-�,.;B�.�54967�z{�|�.68�{���0-25-:79/-;>}���'44~961�.7�7,-�5-0:-57946�4B�.8457946�;-/-;2?�7,->�.0-�.2�@-;;�/-0>�29=9;.0�.=461�27.68.08�.68�.8897946.;�2-0/9:-2?�@97,�7,-�-A:-57946�4B�.1.96�7@4�:.7-1409-2��ymbkdlmpcmj��@,-0-�7,-�5-0:-57946�4B�.8457946�4B�.8897946.;�2-0/9:-2�92�{���.34/-�7,-�5-0:-57946�4B�.8457946�4B�27.68.08�2-0/9:-2��.68���emmcsskmf��@,-0-�7,-�5-0:-57946�4B�.8457946�4B�.8897946.;�2-0/9:-2�92���54967�3-;4@�7,-�5-0:-57946�4B�.8457946�4B�27.68.08�2-0/9:-2��

Page 117: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� �� �� �!"#$%& '�" �(���)�*+(&" �,�(-�� �.+($)&(/�+ 0�1�)#�1(&" �"!�+0"1(&" 2�(-"�'� �)+*�#" #*$%&" %�#"$*0�,��& !�))�03�4�'+)0& '�%(+ 0+)0�%�)5&#�%�+5�)+'�%2�(���,+)�#�+)(�%�"-%�+�#"))�*+(&" �,�(-�� �.+($)&(/�+ 0�1�)#�1(&" �"!�+0"1(&" �%#")�%3�6�+(�&%2�(���.")��%"1�&%(&#+(�0�+�%�)5&#��&%2�(����&'��)�&(%�1�)#�1(&" �"!�+0"1(&" �&%3�7 �(���#" ()+)/2�& �(���+00&(&" +*�%�)5&#�%�,+)�#�+)(� "�#*�+)�#"))�*+(&" �,�(-�� �.+($)&(/�+ 0�1�)#�1(&" �"!�+0"1(&" �%#")�%�#+ �,��",%�)5�03��38�9:47;9<=�<>94<?9@�;94�@94>AB9�6��� �C(�!&'$)�%�+**"-�#".1+)&%" �+." '�.+($)&(/�+ 0�1�)#�1(&" �"!�+0"1(&" �*�5�*�1�)�%�)5&#�2�#*+%%&!&�0�+##")0& '�("�(��&)�#+(�'")&�%3�<*%"2�&(��+%�,�� �& 0&#+(�0�-�&#��%�)5&#�%�+)��%(+ 0+)0�+ 0�+00&(&" +*2�%& #��(���#"5�)+'��!&'$)�%�+)��5�)/�0&!!�)� (3�DED DEF GED GEF HED HEF IED IEF JED JEF FEDKLMNOPQRSTLQSKRUOPQVUPPWXYOSMSZURQ[YSRPSO\KQ]LKL]U_SNOZU]_SMSZU_SNOQZ]NMLKKLKRNK_QUOPQTNYOPQVUPPWSNKMaU_RSOLQVUPPWMNYOMSRQZRLOU]bQKLKKSNOKQUMMLKKMS_bQMNYOMSRQRSK_LOKMS_bQU]MaSLKQMNOKYR_U_SNO[RN\KZUMLQVUPPW XU_Y]S_bZL]MLZ_SNOQNTQUPNZ_SNO �cdefgheijefkejlmnoinpigqnjlmnoigoqirglsfmltiuedeuvipnfiwmlmxeovyieohghereolivefdmkevzwmlmxeovyieohghereoli%�"-%�%&C�%(+ 0+)0�%�)5&#�%�+ 0�!"$)�+00&(&" +*�%�)5&#�%3�6-"�"!�(���%�)5&#�%�"!!�)�0�{�|}~�������+ 0�����}������)�+#��+�.+($)&(/�*�5�*�"!�8�+ 0�" ��"!�(��.�)�+#��%�+%�-�**�+�1�)#�1(&" �"!�+0"1(&" �*�5�*�"!�83�<1+)(�!)".�(�"%��%�)5&#�%2�&(�&%�)�.+)�+,*��(�+(�.+($)&(/�+5�)+'�%�)�.+& �,�*"-�()+ %+#(&" +*�*�5�*��3�4�'+)0& '�1�)#�1(&" �"!�+0"1(&" 2�(-"�'� �)+*�#"..� (%�%�"$*0�,��)�.+)��0��(����&'��1�)#�&5�0�+0"1(&" �"!�(���|}~������2�-�&#��#"$*0�,��#" %&0�)�0�+�%$##�%%!$*�%�)5&#�2�+ 0�(���*"-�.+)��"!����~��z����3�

Page 118: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� ���� ��� ��� !��� !��� "��� "��� #��� #��� ����$% $&'()&*+,+-./0&123345)(*.36).0&7+,8&(,8.)&9:;<$&=20.3&3+).*,()>;<$&2/3&?+),62@&'@+A8,./,.)5)+0.0&'(@3.)0&12334*60,(B+-.3&7.=52A.0*(/06@,2,+(/&(/C@+/.*+,>D0&8(B.52A.*+,+-./0&5.)0(/2@&'(@3.)0=6@@.,+/E/.70@.,,.) B2,6)+,>5.)*.5,+(/&('&23(5,+(/ �FGHIJKHLMHINHMOPQRLQSLJTQMOPQRLJRTLUJOVIPOWLXHGHXYLSQILZ[JRRHXXPRKLYHIGPNHYL\]�_� a]�bccd�]_a_ce�abfgch�Z[JRRHXXPRKL a]�_ c� i c]_�ja_kl_m�agclaic]n�o_�] fp]�c i _�]_adeale�]clgqc]�ade�_pf�aee _fdar�]clgqc]�ade�cgcd�]fh�arr�ja_kl_m�agclaic]�lcj ad�bc_pccd�_lad]aq_fdar�ade�_lad]sflja_fdar�rcgcrh�p` q �] fp]�a�jkq �qfd]frea_ce�qa_ciflmn��tciale di�uclqcu_fd�fs�aefu_fdh�_�jk]_�bc�jcd_fdce�_ c� i �jalv]�fs�wxyz{|}~���}�����ade�ZQRYVXOJOPQRLQR�XPRHh�bf_ �ralicrm�fgcl�_ c�agclaicn�� c�_pf�]clgqc]�p _ �rc]]�uclqc gce�aefu_fd�alc�a_�_ c�]ajc�_jc�_ c�dcpc]_��ZVYOQUP�HTL�H�LMJKHY�ade�ZPOP�HRYLMHIYQRJXLSQXTHIYn� ��� ��� �� �� !�� !�� "�� "�� #�� #�� ���06BB.)&*2B505).C0*8((@&*8+@3)./&.36*2,+(/5(),2@&'()&.?./,0�2*,+?+,+.0()2@&8+0,()>&5).0.)?2,+(/&12334B.3+2&7()�08(50&12334@+=)2)+.0A2B.0&2/3&05(),0&@./3+/A&12334'6/30&'()&*6@,6)2@&5)(�.*,[email protected])/+/A&12334*8+@3)./&.36*2,+(/*6@,6)2@&8.)+,2A.&0+,.0*((5.)2,+(/&8(B.C0*8((@236@,&.36*2,+(/ B2,6)+,>5.)*.5,+(/&('&23(5,+(/ �FGHIJKHLMHINHMOPQRLQSLJTQMOPQRLJRTLUJOVIPOWLXHGHXYLSQIL�TVNJOPQRLYHIGPNHYL�ekqa_fd�]�a�qa_ciflm�p _ �gclm�e ]uala_c�siklc]h�ar_ fki �_�] fp]�j e�rcgcr]�fs�ja_kl_m�ade�uclqcu_fd�fs�aefu_fdn��a_kl_m�rcgcr]�_cde�_f�bc�ajfdi�rcgcr�_pf�ade�_ lcc�p _ �gclm�lalc�c�qcu_fd]�]kq �a]��c�rcald dih��blalc]�ade��ajc]�ade�]ufl_�rcde dih�_ ]�ra]_�fdc�

Page 119: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#�$�%$&'(!&)�*+,+*��-�./�012343156�$*������ ��&�+��!#�+�&�7+(8+7&!�"��-�$9�7&!�"�*+,+*�:�$;�,+�*+,+*�./� �<=>534?5<@53A5@B1CD<CE<4FC@B1CD<4DF<G4BH31BI<J5>5J6<EC3<KG@JCIG5DB<L<2H61D566<653>1A56<M$&'(!&)�*+,+*���-�KG@JCIG5DB<L<2H61D566<$(+�%��&*)�$%�"#�*+,+*�N�$"9��O� �+(+$��&�+�7+(8+7&!�"��-�$9�7&!�"�*+,+*��!��$%�"#���$"9�./�P� +,+(O�QH2J1A<B5DF536�!��*$(#+*)�7+(8+!,+9�$��&�+�%��&�$9�7&+9��+(,!8+�!"�&�+�8$&+#�()O��'(7$��!"#�*+,+*�./��RSR RST USR UST VSR VST WSR WST XSR XST TSRYZ[\]Z_\ _abZcd\e_ffgaZhihjkl\m_abdmn[jh\maomZaci\ _jkcZk_khZomZa_cjok_\jk]oa _cjokmo ncjok\hp_alZ\e_ffg[nbi\Y_dcZ\ho Zhcjok_ja\qn_jci `_cnajcimZahZmcjok\o]\_fomcjok �=>534?5<4FC@B1CD<4DF<G4BH31BI<J5>5J6<EC3<KD>13CDG5DB<653>1A56<r#$!"�-�(�KD>13CDG5DB��+(,!8+�O�&�+�%$&'(!&)�*+,+*��$(+�%��&*)�$%�"#�*+,+*�N�$"9��O� �+(+$��&�+�7+(8+7&!�"��-�$9�7&!�"�*+,+*��!��$%�"#���$"9�./�s�+(+�!���"*)��"+�$99!&!�"$*��+(,!8+�tu 52<EJ54<G43v5B6w���� !"#�$�7+(8+7&!�"��-�$9�7&!�"��-�*+,+*�./�����RSR RST USR UST VSR VST WSR WST XSR XST TSRxy x\]oa\[ndjkZdd\e_ffgaZhanjc Zkc\emajz_cZ\dZhcoagmn[jh\cZkfZad\emn[jh\dZhcoagmoac_\]oa\ZkcaZmaZkZnadmoac_\]oa\ho m_kjZdjfZ_d\[_kb\e_ffg[ndjkZdd\aZljdca_cjok `_cnajcimZahZmcjok\o]\_fomcjok

Page 120: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"�#$� %�$&'()#(*#!+($&'()#!)+#,!&- '&.#/���/0#*( #1'*�%.%/�#0� �'%�0#23�45�67889:9;�<456�=>!))�//')"�7:;�?+-%!&'()@�<69:�ABBC4:D�75�1'*�%.%/��39EF4G93@�<9�H4:;�7�G759DBEI�<456�7:�4J8BE57:5�7JBK:5�BH�357:;7E;�39EF4G93�L�94D65�L�7:;�5<B�7;;454B:7A�39EF4G93M�N69E9HBE9@�<9�GBKA;�564:C�BH�7�JKG6�GB:3BA4;759;�G759DBEIM�OB<9F9E@�<69:�ABBC4:D�75�569�J75KE45I�A9F9A3@�569I�7E9�7JB:D�P�7:;�Q@�<456�B:AI�5<B�9RG9854B:3�ST)U/')�#*')�0#%()0-/&�!)+#$!.,�)&�7:;�T)U/')�#&!V�0#!)+#*��0#$!.,�)&0WM�X9EG9854B:�BH�7;B854B:�A9F9A3�357I�8E955I�GAB39�7:;�7E9�7JB:D�PMY�7:;�QMY�<456�F9EI�E7E9�9RG9854B:3M�Z[Z Z[\ ][Z ][\ [Z [\ _[Z _[\ [Z [\ \[Zabcdefghijkglmnnoapqkdedmrkgnmsmtmkfbuvmrdwdrvgaxtcdegfyfrskpbxkdrvgzbugcb{gdreb|fkpfmcspgksmsxkgf}|brdsbudrvglmnnonb|bsdekgpfmcspgkfuydefkgglmnnozm|dcqgmdn}emcexcmsdbrndkebxrskgzbugcb{gdreb|fkndkmtcfgamu~drvgafu|dskgepdcnufrg|brdsbudrvglmnnoepdcntduspgmccb{mrefkglmnnotduspnmqgtfrfzdskglmnnomrd|mckgufvdksfuglmnnom|txcmsbuqgrxukfkghijkglmnno |msxudsqafuefasdbrgbzgmnbasdbr ���� !"�#$� %�$&'()#(*#!+($&'()#!)+#,!&- '&.#/���/0#*( #�(%'!/#%! �#0� �'%�0#Z[Z Z[\ ][Z ][\ [Z [\ _[Z _[\ [Z [\ \[Zbr}cdrfgyfpdecfkgufvdksumsdbrglmnnoxsdcdsdfkg}ufmndrvkgmrngtdcck}br}cdrfgsm�fkgmrngzffkgamq|frskkfmuepgzbugtxudmcgacmefkglmnnoufkdnfrefgufvdksumsdbrabaxcmsdbrgufvdksuqglefrkxkoafukbrmcgefusdzdemsfkbr}cdrfg|muudmvf�edydcgxrdbrbr}cdrfgzdrfkgebrkxcsgmrngamq|frsndvdsmcgnbex|frskg|mrmvf|frs |msxudsqafuefasdbrgbzgmnbasdbr

Page 121: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"��!#$� %�!&�'($#�) %*!#!'$��!'$+�#,-�. '(�'($�( +($%'�) %*!# ' $%�/$'.$$&�0!'1# ',�!&)�*$#�$*' �&��2�!)�*' �&3�4%� '�(!%�/$$&�%'!'$)�/$2�#$-��2���1#%$�'( %�(!%�'��)��. '(�'($���0*"$5 ',��2�'($��!'$+�#,�!%�.$""�!%�. '(�'($�2!�'�'(!'�'($#$�!#$�6$#,�) 22$#$&'���0*$'$&�$%�!&)�%�� !"�*�" � $%3�7 $��!&�#$0!#8�'(!'� '� %�'($��!'$+�#,�. '(�'($�( +($%'�&10/$#��2�!)) ' �&!"�%$#6 �$%3�9&",�'(#$$�%$#6 �$%�#$!�(�'#!&%!�' �&!"�"$6$"�:�40/1"!'�#,�&1#%$%�;<4%-�=$!"'(�%'!'1%�$:0�& '�# &+�!&)�;�" �$�;<4%�:�!&)�'($,�(!**$&�'��/$�!)) ' �&!"�%$#6 �$%3�9&�'($��'($#�(!&)-�'($#$�!#$��&",�'.��%$#6 �$%�%(�. &+�!�0!'1# ',�/$"�.�"$6$"�>�?�< %��1&'%�2�#�"�.� &��0$%�!&)�@!0 ",�! )��!"�1"!' �&3��A$+!#) &+�*$#�$*' �&��2�!)�*' �&�%��#$%-��&",��&$�!)) ' �&!"�%$#6 �$�?�;�" �$�;<4%�:��#$!�($%�"$6$"��2��-�'(1%� '� %���&% )$#$)�( +(",�!)�*'$)3�BCB BCD ECB ECD FCB FCD GCB GCD HCB HCD DCBIJKLLMNOPQRONKSQJKTIJKLLMNUVMSMIQWOKJQKOXKWWYZ[RVMNOIJK\TZ]JIOTQJMNQTS[\MNMZKVORM_QTOVQ\WM\ OXKWWYNKJOZ]]VM\ OXKWWYNKJOZKJ_M\ SKI[JMIaZQJNQZIM]\O]LOKW]ZIM]\ �bcdefgdhidejdiklmnhmohfpmiklmnhfnphqfkrelkshtdcdtuhomehvefnuimekhudecljduh4%� '�(!%�/$$&�%'!'$)�/$2�#$-�vefnuimek�%(�.%�!�0 )�0!'1# ',�"$6$"�/1'�'($�( +($%'�*$#�$*' �&��2�!)�*' �&3��4""�%$#6 �$%�:�. '(�'($�$5�$*' �&��2�vefooljhwdxhjfqdefu�!&)�vefooljhtlqlkdphfedfhkljydkuh:�%(�.�!�6$#,�% 0 "!#�0!'1# ',-�. '(�6!"1$%�!0�&+�>�!&)�z3��A$+!#) &+�*$#�$*' �&��2�!)�*' �&-�!""�%$#6 �$%�!#$�( +(�!)�*'$)�. '(�!�%��#$�!/�6$�"$6$"�z3{-�. '(�'($�$5�$*' �&��2�|feh}mmtlng~��BCB BCD ECB ECD FCB FCD GCB GCD HCB HCD DCBVK\WU[TQOZVK\TO�OJQ [VKIM]\L]JQM\QJTOVK\WOKN�[MTMIM]\OXKWWYN]\TIJ[NIM]\OZQJSMITNQJIMLMNKIQTO SKI[JMIaZQJNQZIM]\O]LOKW]ZIM]\ �bcdefgdhidejdiklmnhmohfpmiklmnhfnphqfkrelkshtdcdtuhomeh�exfnhitfnnlnghudecljduh

Page 122: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !�"# !!$!%�&'()&� &�)*##� �%((+�,(-�$! .$(!�(/�/$%0�*&� !+�"�*&*!.&�$.&*#/� &� �10$.*�0!$/(�-�, .*%(�2�)$.'�(!#2�(!*� ++$.$(! #�&*�3$,*� % $!&.�.'�**�&. !+ �+�&*�3$,*&4�5'*�- .0�$.2�#*3*#&� �*�,#* �#2� �(0!+�.� !& ,.$(! #�#*3*#� !+�.'*�"*�,*".$(!�(/� +(".$(!�#*3*#&� �*� &�)*##�3*�2�&$-$# �� !+�,#(&*6�)$.'� �,*!.� #�3 #0*�(/�7484�9:�;<=<>��?@A��BCDE FGHI?�������� �A�J������K��L�J����� �

Page 123: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#�$%�&'"(�)*��+,�,-./!""�0�"$�,"-0#�,�)1! �,�!'�,�1�'�$-"�$%�"���'!'�� +"�2$)!�,�$%�"���23$.+3� +"+3$2-�4�,!' ��"���/-'! !*+3!"#�0$�,�'$"��+1�� $/*�"�' �,�%$)�,�)1! �,�)�3+"�0�"$�56789:;6<�'$)�=6>78?:@788A8BC�&'"(�)*�,�$(,�/+"-)!"#�3�1�3�1+3-�,�D-!"��,!/!3+)�"$�"���E-)$*�+'�+1�)+2�,4�(!"��"����F �*"!$'�$%�GHIHJKLMNOKLPQPKRKLI�+'0�ST788U@@A8B�"�+"�+)��/$)��"�+'��+3%�*$!'"�.�3$(�"���+1�)+2��+'0�V;WA7@?W76U�"�+"�!,��+3%�*$!'"�+.$1��"���+1�)+2�C�X!2'!%! +'"3#4�+33�/+"-)!"#�3�1�3,�)�/+!'�.�"(��'�3�1�3�Y�+'0�3�1�3���(!"�$-"�.!2�0!,"+' �,4�,��/!'2�"$��+1��+�,!/!3+)�/+"-)!"#�0�1�3$*/�'"C�����0!%%�)�' ��.�"(��'�"��� +"�2$)!�,�(!"��"���3$(�,"�+'0��!2��,"�3�1�3,�!,�3�,,�"�+'�$'��*$!'"4�"���3$(�,"�0!%%�)�' ��+/$'2�+33�"���,-)1�#�0� !"!�,4� Z[\[] _����)�%$)�4�"�!,�)�,-3"�*)$1�,�"�+"�"��� !"#�$%�&'"(�)*�!,��1$31!'2�+33�!",�,�)1! �,�%$33$(!'2�+��$/$2�'�$-,�+ "!$'�*3+'C�'�"�)/,�$%�*�) �*"!$'�$%�+0$*"!$'4�"���3�1�3,�+)��/+!'3#�.�3$(�"���E-)$*�+'�+1�)+2�4�(!"��"����F �*"!$'�$%�V;WA7@?W76U?(�! ��!,�,3!2�"3#�+.$1��"���+1�)+2�C?&33�*�) �*"!$'�$%�+0$*"!$'�3�1�3,�)�/+!'�.�"(��'�3�1�3�YCa�+'0��4�"��)�%$)��,3!2�"3#�.�3$(�"����F*� "+"!$',C������F �*"!$'�!,�+2+!'�$%�V;WA7@?W76U�"�+"�,-)*+,,�,�3�1�3��Ca4�/�+'!'2�"�+"�"���+0$*"!$'�$%�"��,��,�)1! �,�!,��,"!/+"�0��!2��)�"�+'�"���!'!"!+33#��F*� "�0C��bc[de[fghci[jklcdgldlm[glk[gn[olcfkhpqe[V;WA7@?S76U�,�)1! �,�"�+"��+,�,-)�3#���3*�0�"��� !"#�+ �!�1��"���+%$)�/�'"!$'�0�)�,-3",C�����r!' +�s)$t� "�0�1�3$*�0�+�su&�+,,!,"+'"�%$)�+/.-3+"$)#�'-),�,���3*!'2�"��/�"�)$-2��+�'-/.�)�$%�"��!)�"+,v,C�&/.-3+"$)#�'-),�,� +'�)�2!,"�)�"��!)�+ "!1!"!�,�"�)$-2��"���-,��$%�+'��3� ")$'! �!0�'"!%! +"!$'�w�xuy�+'0�,-.,�D-�'"3#� )�+"��!'1$! �,�%$)�"��!)�*+"!�'",4�(�$�+)��+3,$�)�2!,"�)�0�!'�"���,#,"�/C���!,�*)$t� "��+,�.��'�+'�+',(�)�"$�"���'��0�$%��+,!3#� +*"-)��!'"�)�,"!'2�+'0�'� �,,+)#�!'%$)/+"!$'�%$)�"���*-)*$,��$%�+���+3"� +)��t$.�(��)�.#�"���/$.!3!"#�$%�"��� +)��*)$1!0�)�!,� )- !+3C�z$)�$1�)4�!"��+,���3*�0�+1$!0!'2�/!,"+v�,�+'0�)�,")! "!$',�.$)'�%)$/�"���/+'#�/+'0+"$)#�%$)/+3!"!�,�'-),�,�+)��0+!3#� $'%)$'"�0�(!"��0-��"$�"���)+"��)�3$'2�!'"�)1+3,�.�"(��'�"���+ "!$'�0+"��+'0�"���/$/�'"�"���!'%$)/+"!$'�!,��'"�)�0�!'�"���,#,"�/C�

Page 124: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ����� !"#�$%�������� �&���'������(��)�'����� �*+,�-./012/3�4,5,46�781�9.1:,48;.�.1,�.<80/�.;�.5,1.=,�87�+.47�.�>82;/�.<85,�?018>,.;�1,604/6@�A2B�:./,=812,6�:4,.143�6/.;C�80/D�EFGFHIJKLMIJNONIPIJGQ�RSTUUVWWXUYZ[\]^W_ ]VUa[b[cdeXUVeeZ[\UfXg_U]VUaZ�hgTUe _ga[TUi[jgcTU[ WTUUXUYQ�62;:,�/+,�-./012/3�4,5,4�781�/+,6,�:./,=812,6�26�.<85,�?018>,.;�.5,1.=,6�.;C�<,38;C�/1.;6.:/28;.4�4,5,4Q�.4-86/�1,.:+2;=�/1.;6781-./28;.4�4,5,4�2;�68-,�:.6,6@�k,=.1C2;=�/+,�1,-.2;2;=�:./,=812,6Q�/+,3�.1,�.44�5,13�62-24.1�/8�/+,�?018>,.;�.5,1.=,6�.;C�-86/43�1,4./,C�/8�.�-./012/3�4,5,4�.180;C�/+,�/1.;6.:/28;.4�4,5,4�l@�

Page 125: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#�$%�&'( �)$*'�!+��',!*-�'�./!"���$0$-�*�$/+�1�,�)$20�*"�!*�'))�+/(,�#�1� '"�-$(!�+3�����1!%%�(�* ��4�"5��*�"��� '"�-$(!�+�5!"��"���)$5�+"�'*1��!-��+"�)�,�)+�!+�+)!-�")#�4�)$5�$*��2$!*"�'*1�'��')%6�� 7898:;<�����2�( �2"!$*�$%�'1$2"!$*�)�,�)+�'(��'+�5�))�'4$,��=/($2�'*�',�('-�+�!*�')0$+"�'))� '"�-$(!�+3�>!,�� '+�+�+"'*1�$/"�?�@ABCDEAFGHIJIKLMNGFMMOI@GPNQDGAFGHOIRNSFTETCFOIUQVGMBDQHI'*1�WQKVGIBCVGGNGXIY6�+!* ��"���1!+"'* �+�"$�"���',�('-�+�'(��(�0'(Z'4)��'*1�"���%!*')�(�+/)"�!+�'4$,��)�,�)��6�"��(�%$(�6�5!"��'�2�( �2"!$*�$%�'1$2"!$*�(�0'(Z'4)#��!-��(�"�'*��[2� "�13�\/"�$%�"���'%$(�0�*"!$*�1�%!,�� '"�-$(!�+6�"5$� '+�+�]�WQKVGIBCVGGNGX�'*1�UQVGMBDQH�]�'(��"���4�+"�'1$2"�1�4#�&'( �)$*'+� !"!_�*+6�')0$+"�(�' �!*-�)�,�)��3�\*�"���$"��(��'*16� DTNVCITVQFI!+�"���$*)#� '"�-$(#�+)!-�")#�4�)$5�"���=/($2�'*�',�('-�6��,�*�"�$/-��+/(2'++!*-�"���+'"!+%' "$(#�)�,�)�a3�\*��$%�"���(�'+$*+�"�'"�%��1�!*"$�"���*$"! �'4)��+/ �++�$%�WQKVGIBCVGGNGX�+�(,! �+�!+�"���\b\c�d($e� "6�'�*�5�5'#�"$�0'*'-��4/!)1!*-�2�(0!"+3�\b\c�!+�'�0/)"!] �'**�)�!*"�-('"�1�+$)/"!$*�"$�0'*'-��2�(0!"+�%$(�4/!)1!*-�5$(Z+�!*�"��� !"#3�f"�(�2(�+�*"+�'� $02)�"��(�]ghijhggkjhi8gllmkn8jhopqrjhi8osthigu8jh8pgitpv8mkithjutnjmhtpv8untlljhi8thr8uwungxuy8juuqgu;8z���*�5�+#+"�0�(�1/ �+�"���"!0��"$�$4"'!*�)! �* �+�%$(0�0$*"�+�"$�0!*/"�+�!*�0$+"� '+�+�'*1�'))$5+�"���"('*+%�(�"$�"� �*! ')�(�+$/( �+�%($0�4/(�'/ ('"! �"'+Z+�"$�,')/�]'11�1�' "!,!"!�+3�f"�(�./!(�1�'*�'22($' ��"$�2'("*�(+�!2�'*1�"� �*$)$-#�1�,�)$20�*"�!*�$(1�(�"$�' �!�,��%/))�!*"�-('"!$*�$%�!*"�(*')�'*1��["�(*')�'-�* !�+�!*�'�+�'0)�++�+!02)��2($ �++3�{ |}~�}�{��������� ��

Page 126: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#�$%&'�()�*#+,#-�."(/.�#0$#11#-&�23&4+%&'�+#.41&.5�/%&"�6314#.�37(6#�84+(9#3-�36#+3,#.�)(+�2(.&�()�&"#�$3&#,(+%#.:�; %&"%-�&"#�<=>?@A=BCDEFEGHIJCBII�3-K�31.(�&"#�<CLJM@C=BCD�3-K�NOPCCB??JCQE$3&#,(+%#.�&"#�$%&'�3$"%#6#.�.$(+#.�"31)�9(%-&�37(6#�&"#�84+(9#3-�36#+3,#:�R(/#6#+5�&"#+#�3+#�&"+##�$3.#.5�STUTVWXYZ[WX\]\W WXU5�_JBaAa?B�3-K�b@aJP?EaPMB5�/"#+#�&"#�23&4+%&'�()�&"#�.#+6%$#.�."(/.�3�K%.&3-$#�()�37(4&�(-#�9(%-&�7#1(/�84+(9#3-�36#+3,#.5�9+#.#-&%-,�&"#+#)(+#�3�23&4+%&'�1#6#1�$(++#.9(-K%-,�/%&"�%-&#+3$&%6#�.#+6%$#.�+3&"#+�&"3-�/%&"�&+3-.3$&%(-31�.#+6%$#.:��NOPCCB??JCQ�%.�&"#�(-1'�$3&#,(+'�.4+93..%-,�&"#�&+3-.3$&%(-31�1#6#1�c�+#,3+K%-,�23&4+%&'�1#6#1.5�defghijklgjfgdehkjmnjklgjoepqrqsthqkiudjgnnmfkjkmjsedljgvw(6#+-2#-&�&(/3+K.�3�-#/�)411�%-&#,+3&%(-�2(K#1�()�K#1%6#+%-,�.#+6%$#.�&(�$%&%x#-.:�y-#�()�&"#�+#3.(-.�()�&"%.�.4$$#..�$(41K�7#�&"#�.#+6%$#�#vz%31(,5�&"+(4,"�/"%$"�3-'�$%&%x#-�$3-�$(-&3$&�&"#�24-%$%931%&'�&"+(4,"�&"#�4.#�()�3�$"3&�1%-#:�!"#�3991%$3&%(-�,%6#.�&"#�%-"37%&3-&.�&"#�9(..%7%1%&'�&(�aOPD�/%&"�3-�())%$#+�3&�&"#�%-)(+23&%(-�K#.{�3-K�+#$#%6#�3-./#+.�%22#K%3&#1'�(-�|4#.&%(-.�+#,3+K%-,�9471%$�.#+6%$#.�())#+#K�7'�&"#�24-%$%931%&':�}(&371'5�311�$3&#,(+%#.�.##2�&(�"36#�|4%&#�3�"#&#+(,#-#(4.�23&4+%&'�#6(14&%(-:�!"#�K%))#+#-$#�7#&/##-�&"#�$3&#,(+%#.�/%&"�&"#�1(/#.&�3-K�"%,"#.&�1#6#1.�%.�#03$&1'�()�&/(�9(%-&.�3-K�3�"31)5�7#%-,�&"#�"%,"#.&�K%))#+#-$#�32(-,�311�&"#�93+&%$%93-&�$%&%#.�%-�&"%.�.4+6#'5��~j�j����!"%.�+#.41&�9+(6#.�&"3&�&"#�$%&'�()�*#+,#-�"3.�K#6(&#K�%&.�#))(+&.�&(�#6(16#�$#+&3%-�.#1#$&#K�$3&#,(+%#.�%-�)%+.&�913$#5�1#36%-,�(&"#+�$3&#,(+%#.�#6(14&%(-�)(+�)4&4+#�+#1#3.#.�()�%&.�#v�#+6%$#.:��#,3+K%-,�9#+$#9&%(-�()�t�mskqmpjhg�ghd�jt�tqpj�gf�gpudj�thegdjtfgjt�m�gj�efmsgtpjt�gft�gdj/%&"�&/(�#0$#9&%(-.5�3,3%-�_JBaAa?B�3-K�b@aJP?EaPMB:��&�%.�3.�/#11�+#23+{371#�&"#�+#.41&�+#3$"#K�/%&"�&"#�$3&#,(+'�<=>?@A=BCDEFEGHIJCBII5�/"%$"�9+#.#-&.�3�&(9�9#+$#9&%(-�()�3K(9&%(-�1#6#1�6314#5�9+(6%-,�&"3&�.#+6%$#.�%-$14K#K�%-�&"%.�$3&#,(+'�"36#�7#$(2#�3�$(291#&#�.4$$#..5�3$$(+K%-,�&(�&"(.#�+#.9(-.%71#�)(+�%&.�K#1%6#+':�

Page 127: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#�������� �$�%������&��'�%����� �()*+,-�./010231�4,35/)36�*070*�7,*501�85)30�9*-10�)2�:020/,*�3-�3;0�<5/-.0,2�,70/,:01=�,*3;-5:;�,*>,61�1*):;3*6�+0*->�3;0�<5/-.0,2�/015*31?�@;0/0�)1�3;0�1-*0�0A90.3)-2�-B�3;0�CDEDFGHIJKLMNONLPLMQ�9,30:-/6�>;0/0�3;0�4,35/)36�*070*�/015*3�)1�,+-53�;,*B�,�.-)23�15.0/)-/�3-�3;0�,70/,:0�,2R�9*0,/*6�15/.,1101�3;0�3/,21,93)-2,*�*070*�S?�T-/�3;0�/013�-B�9,30:-/)01=�3;06�,**�UVWXYZ[\V] VVZ[WX]_UY] [aVbVa[cde[XZf[cdgh[]iVUVjkUV[lik YZm[]iX][nYa\Xkol[lVUbYpVl[l]Yaa[/04,)2�)2�3;0�)230/,93)70�*070*?��

Page 128: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !�"� !#� $�%&��'"�"� !#�($�!)"#*!+&�,-.//0112/3�$�%�45./67859&�:#!+!���;$�<#*�=�>!?!>+�+*)�*@*($��>=�A!>"'�� !�$?!#$)!�B:$#�*(<>$#>=�*��� *+�>$+��($+!�' !#!�� !�%*@@!#!�(!�*+�"�!�:"*��&�:#!+!��*�)�"�>=�*��!#$(�*?!�+!#?*(!+�*�+�!$%�"@��#$�+$(�*"�$>�+!#?*(!+�$+�;"+��C<#":!$��(*�*!+�:#"?*%!D�E*)�*@*($��>=&�$>>�($�!)"#*!+�+!!;��"� $?!�$�+>*) �>=� !�!#")!�!"<+�;$�<#*�=�!?"><�*"�D�F !�%*@@!#!�(!�A!�'!!��� !�($�!)"#*!+�'*� �� !�>"'!+��$�%� *) !+��>!?!>+�*+�"?!#�"�!�:"*���$�%�$� $>@&�� GHIHJKL�MNOHPQHRSTUVPWXHXOYZS[OXHYOTV\O]H\PH __abcdefbcghgbibc � $+�+<#!>=� !>:!%�� !�(*�=�$( *!?!�� !�PU\VSNO]HYOXjT\XKHk lV\WXHMNHRSTUVPHSXHVH[PN[YO\OHmVY\HPQH\lOHnjNS[SmVTHoOUXS\OH]OZP\O]H\PH%!>*?!#�*�@"#;$�*"��#!)$#%*�)�(<>�<#$>�$(�*?*�*!+�A"� ��"�(*�*p!�+ *:�$�%��"<#*+�+D�F !�+*�!�"@@!#+�$�+!$#( �!�)*�!�@"#�$%?$�(!%�+!$#( !+�A=��=:!�"@�!?!��&�%$�!�"#�:>$(!�$�%�� !�%"'�>"$%$A>!�$)!�%$�@"#�� !��!q��;"�� D�r>+"&�*��"@@!#+�� !�:"++*A*>*�=�"@�+!$#( *�)�� !�;"?*!+�$�%�@*>;+�"��RSTUVPWXH[SNOnVXHV\H\lOHnPnON\K�s"#!"?!#&�$�%�' $��;$t!+�� *+�+!#?*(!�:$#�*(<>$#�*+�� !�:"++*A*>*�=��"�@*�%�*�@"#;$�*"��:"+�!%�A=�"� !#�$++"(*$�*"�+�@#";�� !�(*�=D�F !�:!#(!:�*"��"@�$%":�*"��>!?!>�"@�u*>A$"�*+�?!#=�+*;*>$#��"�� !�C<#":!$��$?!#$)!+�!?!��*@�+>*) �>=�A!>"'&�$>� "<) �'*� �%*+�$�(!+�"@�"�>=� $>@�$�:"*��D�F !�($+!�"@�,-.//0112/3�+�$�%+�"<��+*)�*@*($��>=&�+*�(!�� !�:!#(!:�*"��"@�$%":�*"��>!?!>�*+�$A"?!�C<#":!$��$?!#$)!�$�%�$A"?!�>!?!>��Dv�$+�'!>>&�$>� "<) �*�+�;$�<#*�=�+("#!�*+�@$#�A!>"'�@#";�� !�C<#":!$��$?!#$)!D�w xyz{y|}~�{w��������� ��

Page 129: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#�$%&'�()�*%+,%-."/,�%0�1#2(3�45+(6#/-�/7#+/.#�3"#-�&/28%-.�/1(5&�,/&5+%&'�/-9�02%."&2'�1#2(3�&"#�/7#+/.#�3"#-�&/28%-.�/1(5&�6#+$#6&%(-�()�/9(6&%(-:�3%&"�&"#�#;$#6&%(-�()�<=>??@AAB?C�%-�3"%$"�6#+$#6&%(-�()�/9(6&%(-�%0�02%."&2'�/1(7#�&"#�45+(6#/-�6/&&#+-D��EFGH>IBJ?K/-9�E?LBMJ?N@?IK/66#/+�&(�1#�&"#�$/&#.(+%#0�%-�3"%$"�"5.#�#))(+&0�"/7#�1##-�9#7(&#9�0%-$#�&"#'�/+#�&"#�(-2'�$/&#.(+%#0�&"/&�05+6/00�2#7#2�O�/-9�#7(27#�&(3/+90�&"#�&+/-0/$&%(-/2�2#7#2D�P#/-3"%2#:�QJHB>AKH>M@�%0�&"#�$/&#.(+'�/$"%#7%-.�&"#�2(3#0&�,/&5+%&'�2#7#2�(-2'�())#+%-.�%-)(+,/&%(-�&"+(5."�#2#$&+(-%$�,#/-0D�R-�&"#�(&"#+�"/-9:�&"#�+#0&�()�$/&#.(+%#0�+#,/%-�%-�2#7#2�O�(+�02%."&2'�/1(7#:�,#/-%-.�&"/&�&"#�,/S(+%&'�()�0#+7%$#0�+#,/%-�0&%22�%-�&"#�%-&#+/$&%(-�2#7#2�&"+(5."�3"%$"�50#+0�$/-�.#-#+/&#�1/0%$�$(-&#-&�%-�/�(-#�3/'�$(,,5-%$/&%(-�6+($#00�3#1T0%&#�1/0#9D�U%.-%)%$/-&2':�/22�$/&#.(+%#0�0##,�&(�"/7#�0%,%2/+�,/&5+%&'�#7(25&%(-D�!"#�9%))#+#-$#�1#&3##-�&"#�$/&#.(+%#0�3%&"�&"#�2(3#0&�/-9�"%."#0&�2#7#20�%0�02%."&2'�(7#+�(-#�6(%-&�/-9�/�"/2):�VWXWYZ[�&"#+#)(+#�6+(7%-.�&"/&�&"#�$%&'�()�*%+,%-."/,�%0�"/7%-.�/�"(,(.#-#(50�#7(25&%(-�()�/22�%&0�0#+7%$#0D�R-#�()�&"#�0#+7%$#0�&"/&�,%."&�"/7#�"#26#9�&"#�$%&'�()�*%+,%-."/,�/$"%#7#�%&0�E?LBMJ?N@?I>AK+#052&0�%0�*528'�\ /0&#�](22#$&%(-D�!"#�$%&%#-0"%6�%0�())#+#9�&"#�6(00%1%2%&'�&(�/662'�)(+�&"#�+#,(7/2�()�1528'�3/0&#�%&#,0�/&�/-'�/99+#00D�!"#�0#+7%$#�%0�/7/%2/12#�(-�3#10%&#�/-9�%-&#+/$&%7#�&#2#7%0%(-D�!"#�#/0%-#00�()�&"%0�0#+7%$#�&"/&�"#260�8##6%-.�&"#�$%&'�$2#/-�/-9�.+##-�/-9�&"#�/7/%2/1%2%&'�&"+(5."�,52&%62#�$"/--#20�%0�+#,/+8/12#D�_#./+9%-.�6#+$#6&%(-�()�/9(6&%(-�2#7#20:�/2,(0&�/22�$/&#.(+%#0�+#,/%-�/1(7#�2#7#2�OD D�!"/&�abcdeWfgcfWfgbWahdijikclifmneWkbojbkfipdWpqWbTrbosijbeWjifitbdneWhebWieWceWbukbjfbvWpoW05+6/00%-.�%-%&%/2�#;6#$&/&%(-0D��

Page 130: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !"#$�������� �%�&������'��(�&����� �)*+*,-.�/0121-32�45631�.�7*-2*+68.31�*9910�*9�1:;10<6712�9*0�=.+9�*9�3=1�7.31,*0612>�?63=�<.+512�[email protected]+C�.B*<1�D50*/1.-�.<10.,1>�B53�45631�+*?�96,5012�9*0�3=1�*3=10�=.+9�*9�3=1�7.31,*0612>�?63=�<.+512�[email protected]+C�B1+*?�3=1�D50*/1.-�/.3310-E�F6<1�7.31,*0612�23.-8�*53G�HIJKKLMMNKO>�PQRSJTNUKVWD@/+*C@1-3�X�B526-122>�YNZLS[SMLW.-8�\] JKW_MJKKNKO�26-71�3=1C�2=*?�0125+32�3=.3�,*�B1C*-8�30.-2.736*-.+�+1<1+� >�?=101�@*23�D50*/1.-�763612�.01E�a+2*>�9*0�3=1�7.31,*0C�*9�HIJKKLMMNKO>�)*+*,-.�/0121-32�.�3*/�<.+51�*9�bEc>�1<*+<6-,�3*?.082�-1?�5210:71-3018�@*81+2�*9�/5B+67�210<6712�81+6<10CE�d=1�,**8�0125+32�9*0�3=1�*3=10�9*50�7.31,*0612�.01�.2�?1++�[email protected]+1�26-71�3=1�D50*/1.-�.<10.,12�9*0�3=121�7.31,*0612�/0121-3�0.3=10�+*?�0125+32E�e*?1<10>�9*0�3=1�7.31,*0612�*9�fghgijklmnjkopojqjkh>�PKrN]UKsLKTVWtUSNJMWHJ]L�.-8�u]JKv_U]T�63�=.//1-2�1w.73+C�

Page 131: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !! "#��$���%��#"$����#&�"' &�"�&�(%#)�'*�%&*+�,�* -�./& !�%)�%0�&%1�"��0�)��� /1��#��#"�0�&+�&�(%&2%,*����%��%**�'%��1 &#�"�"/&!%""�����#)��&%'�#0��*�0�*�34��5���6#77�&�)'��,��-��)�����'%��1 &#�"�-#�������* -�"��%)6��#1��"��*�0�*"�#"��#1��&���%)��- �! #)�"$�����"�' )6�,#11�"��6#77�&�)'��%( )1�%**�����"/&0�+�6�'#�#�"$�89:9;<=�5�#"�&�"/*��!& 0�"���%������'#�+� 7�> * 1)%��%6�%�����& 1�)� /"��?@�&0#'�"��0 */�# )�-�#'��#"�' ��&�)��-#�������'#�+A"�(%�/&#�+�(%!4�B)����� ���&��%)6$�-��)�* 2#)1�%������!�&'�!�# )� 7�%6 !�# )�*�0�*"�7 &�> * 1)%$����+�%&��0�&+�"#(#*%&�� �./& !�%)�%0�&%1�"$�-#�� /��,#1�6#"�%)'�"4��C��#"�&�(%&2%,*��� �(�)�# )�����(%�/&#�+�&�"/*�"�%'�#�0�6�,+�����DEFGGHIIJGK�'%��1 &+�#)�����'#�+�"/&!%""#)1������&%)"7 &(%�# )�,%&&#�&�%)6��)��&#)1�*�0�*��$���/"�) �� )*+�7/**�#)��1&%�#)1��?LMNOPQROQS9TOPNUVOT9WXS9YZTM9[PMRMSUQ\9VUSU]OQT9[YPSUVU[YSUMQ9UQ9TOPNUVO9[PMNUTUMQ9YQ_9 XYZUSa9�0%*/%�# )4�B)�� 7�����&�%" )"� 7���#"�"/''�""�' /*6�,������"�&0#'��bFcdJFcefHghiIH$�%�)�-�-�,�! &�%*�'&�%��6�%'' &6#)1�� �����-�,�34j�!&#)'#!*�"4�C)7 &(%�# )�%)6�"�&0#'�"�%&��"�&/'�/&�6�#)�( 6/*�"�%** -#)1�/"�&"�� � &1%)#k�����#&� -)�� (��!%1�"4�5���( 6/*�"�'%)�,��(%)%1�6�0�&+��%"#*+l�#��#"�! ""#,*��� �"�#7�$�'* "�$�' (!&�""�m��n!%)6�%)6�!�&" )%*#k�����(4�opqrstpupov�������� ��

Page 132: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#$%&'&�()&#&"*#�+,-*&�"-$&�.-/,)&#�0.�12*,)-*34�2506&�!,)0(&2"�26&)2/&�.0)�277�$2#&#�5,*�*80�9:;<=>?@AB�2"'�:BC@DABEFB?GH�I-6&�$2#&#�$7&2)73�#*2"'�0,*J�:EKLAMEFB?NONP<Q@BFQQ4�R@SF=M=LFTNUA=@>LN=>DFTNVD>BQKAD?N2"'�WDP>BNKL>BB@BX4�#-"$&�*%&�12*,)-*3�7&6&7�0.�*%&�#&)6-$&#�0..&)&'�-"�*%&#&�$2*&/0)-&#�-#�10#*73�*)2"#2$*-0"274�)&12)Y2573�2506&�!,)0(&2"�26&)2/&#4�2"'�$7&2)73�&6076-"/�*082)'#�*)2"#.0)12*-0"27�7&6&7�ZH�[7#04�*%&�$2*&/0)3�0.�R@SF=M=LFN()&#&"*#�0"&�0.�*%&�*0(�627,&#�8-*%�2�)&#,7*�0.�\H]H� %-#�-#�2�6&)3�#-/"-.-$2"*�)&#,7*�#-"$&�10#*�!,)0(&2"�$-*-&#�#%08�)&#,7*#�,"'&)"&2*%�*)2"#2$*-0"27�7&6&7�8%&)&2#�!"#$%&'&�()&#&"*#�10)&�&6076&'�&G_&)6-$&#H�^%&�'-..&)&"$&�5&*8&&"�*%&�$2*&/0)-&#�8-*%�*%&�708&#*�2"'�%-/%&#*�7&6&7#�-#�#7-/%*73�5&708�*80�(0-"*#4� `abacde�^%&)&.0)&�()06-"/�*%2*4�&6&"�*%0,/%�$&)*2-"�80)Y�%2#�5&&"�'&60*&'�*0�%26&�2�%010/&"&0,#�'&6&70(1&"*4�#01&�$2*&/0)-&#�"&&'�*0�5&�-1()06&'H�^%2*�-#�*%&�$2#&�-"�:;<=>?@AB�2"'�:BC@DABEFB?TN*%&�0"73�*80�$2*&/0)-&#�)&12-"-"/�-"�-"*&)2$*-6&�7&6&7��H�f&/2)'-"/�(&)$&(*-0"�0.�2'0(*-0"4�!"#$%&'&�#%08#�(&)$&(*-0"�0.�2'0(*-0"�(2**&)"#�2506&�!,)0(&2"�26&)2/&#�.0)�277�$2#&#�8-*%0,*�&g$&(*-0"4�()&#&"*-"/�*%)&&�*0(�627,&#�2#�8&77�.0)�:EKLAMEFB?NONP<Q@BFQQ4�R@SF=M=LF�2"'�WDP>BNKL>BB@BX4�$2*&/0)-&#�8-*%�2�12*,)-*3�7&6&7�2506&�26&)2/&�2#�8&77H�h"�*%&�0*%&)�%2"'4�277�(&)$&(*-0"�0.�2'0(*-0"�)&#,7*#�)&12-"�2506&�7&6&7�\H�4�()06-"/�*%2*�277�#&)6-$&#�%26&�5&&"�,#&'�53�2�",15&)�0.�$-*-i&"#�%-/%&)�*%2"�-"-*-2773�&g(&$*&'4�2$$0)'-"/�*0�*%0#&�)&#(0"#-57&�.0)�-*#�'&7-6&)3H�

Page 133: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#!$%�������� �&�'������(��)�'����� �*+,-./,012314567/89/:7175.;/<571/71<65+.601+=2;51>-.2?5+81+;5.+@51/81+661<+,5@2./5715ABCDE�FGHI�JKLMNONPMQ1+175.;/<57:1<+,[email protected]/81RS/<S1,S51T-8/</?+6/,01S+71821<2T?5,58<5U1V661<+,5@2./571WBXYCZC�[CZC[\�W]HZC�EXC�EXYG �EGW_\WBEYH_W[�[CZC[�aXCGCFHGCb�\CGZYBC\�IWEcGYEd�XW\�eH_C�]CdH_ �EXC�Y_ECGWBEYZYEd�[CZC[�]WGGYCG�DGHIHEY_e�BHID[CEC�EGW_\WBEYH_\�EXGHceX�EXC�\CZCGW[�WZWY[W][C�C[CBEGH_YB�BXW__C[\�fHGCHZCGb�EXC�BYEd�HF�gC[\Y_hY�Y\�\EWGEY_e�EH�^CZC[HD�IHGC�<2T?65A1+8i1+i;+8<5i175.;/<571/81RS/<S1</,/j587:1?+.,/</?+,/281/717,+.,/8@1,2158S+8<51/,71YIDHGEW_BC�[CZC[�klmPnOlMopqrqstuKoMuubqkovKwnolMopxPqW_ �ynNKxPqNxwM�WGC�EXC�EXGCC�BWECeHGYC\�EXWE�B[CWG[d�\EW_ �HcE�FGHI�EXC�GC\E�WBXYCZY_e�\BHGC\�IHGC�EXW_�H_C�DHY_E�W]HZC�EXC�

Page 134: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#$�#%"�#&"'���())�*#+"& �,"-�-"".�+ �/#%"�-,.,)#��.#+��,+0�)"%")-'�1/"�2,33"�"$*"�4"+5""$�+/"�*#+"& �,"-�5,+/�+/"�) 5"-+�#$2�/,&/"-+�)"%")-�,-�"6#*+)0� $"�! ,$+7�89:9;�+/"�"3 �"�!� %,$&�+/#+�+/"�*,+0� 3�<")-,$=,�/#-�4""$�5 �=,$&� $�,+-�">?"�%,*"-�2"),%"�0�3 )) 5,$&�#�/ . &"$" �-�5 �=�!)#$'��@A9BCD9EBCDF9CGAHI9JB9KGA9LD9MBGBDH9BCGB9NDOMJAPJQM9MDFRJKDMQ9SDFKDSBJEA9ET9GHESBJE$�,-�-,.,)#��+ �BCD9UVFESDGA9GRDFGWD9VS9BE9XY9ZCJM9TGKB9HD[EAMBFGBDM9BCGB9KJBJ\DAQM9SDFKDSBJEA9ET9GHESBJEA9ET9-"�%,*"-�!� %,2"2�+/� �&/�")"*+� $,*�."#$-�,-�+/ �&/+�+ �4"�-),&/+)0�/,&/"��+/#$�,$,+,#))0�D]SDKBDHI9BCDFDTEFD9SFERJAW9BCGB9NDOMJAPJQM9KJBJ\DAM9GF"��"#20�#$2�5,)),$&�+ ��"*",%"�#$2�.#--,%")0��-"�")"*+� $,*�-"�%,*"-'�15 � +/"��*#-"-�#�"�"-!"*,#))0��".#�=#4)" �_ abcdefg�,-�+/"�*#+"& �0�5,+/�+/"�/,&/"-+�!"�*"!+, $� 3�#2 !+, $�.#�=�h/,&/"��+/#$��i7�5/"�"#-�jklcgmnocggegp�,-�+/"�)"--�#2 !+"2�*#+"& �0�"%"$�-#+,-3#*+ �0��"#*/"2�+/"�"6!"*+"2�+#�&"+�h)"%")�qi'��� rstuvwx�������� ��

Page 135: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ����������������������� � !"#$%&'()*+,-./+0�1,+0+23�4536,.37�8+-+80�96.3+�/8:0+�3:�3;+�<6,:1+52�5-+,5=+0>�?3�.0�+01+/.5887�,+45,@5A8+�3;+�/50+�:B�3;+�B:88:C.2=�B:6,�/53+=:,.+0D�EFGHIJKLMNJKOPOJQJKRS�TUVWXGFYZ[\]F_W Wa_�52b�cdXZefYdGS�588�C.3;�8+-+80�53�:,�06,1500.2=�3;+�3,5205/3.:258�8+-+8�g>�h;+7�;5-+�,+5/;+b�3;+�3,5205/3.:258�8+-+8�.2�C;./;�5�,+96+03�/52�A+�/:418+3+b�3;,:6=;:63�52�+8+/3,:2./�/;522+8�52b�5,+�+-:8-.2=�3:C5,b0�5�2+C�3,520B:,453.:2�8+-+8�i>�j2�3;+�:3;+,�;52bS�3;+,+�5,+�3C:�/53+=:,.+0�,+45.2.2=�A+8:C�4536,.37�8+-+8�k�52b�/8+5,87�A+8:C�<6,:1+52�5-+,5=+D�TlfaY l_ZG\m\nVeFZ_ee\52b\oYWFXa\WXd_>�?2�3;+0+�/50+0�:287�.2B:,453.:2�:,�b:C28:5b5A8+�B:,40�/52�A+�:A35.2+b�3;,:6=;�+8+/3,:2./�4+520>�p+-+,3;+8+00S�.2B:,453.:2�,+=5,b.2=�.30�1+,/+13.:2�:B�5b:13.:2�0;:C0�3;53S�b+01.3+�A+.2=�96.3+�.44536,+�&(qr!s(&t's!u!v($%&'#wxyu!x$'z#&'{(($'z!|z(q'uz#$'!$!u!#""}'(~y(su(w��h;+�b.BB+,+2/+�A+3C++2�3;+�/53+=:,.+0�C.3;�3;+�8:C+03�52b�;.=;+03�8+-+80�.0�+�5/387�3C:�1:.230S��'�'��h;+�5B:,+4+23.:2+b�,+0683�1,:-+0�3;53�3;+�/.37�:B��.852�;50�A++2�C:,@.2=�:2�.30�+)*+,-./+0�+-:863.:2�B:88:C.2=�5�96.3+�;+3+,:=+2+:60�C:,@.2=�1852>�h;+,+B:,+S�3;.0�;50�8+b�3:�5�0.3653.:2�.2�C;./;�0:4+�/53+=:,.+0�;5-+�A++2�8+B3�A+;.2b>���(|#qw!$|'y(qs(yu!x$'x�'#wxyu!x$t'!u'z#&'ux'{('�($u!x$(w'uz#u' !"#$%&'s!u!v($&z!y'z#&'�!w("}'5//+13+b�+8+/3,:2./�0+,-./+0�1,:-.b+b�A7�3;+�462./.158.37�50�0++2�.2�3;+�1+,/+13.:2�:B�5b:13.:2�B.=6,+>��88�/53+=:,.+0�0;:C�8+-+80�:-+,�3;+�3;.,b�035=+�.2b./53.2=�5�A.==+,�60+�3;52�.2.3.5887�B:,+0++2>�TUVWXGFYZ\52b�]F_W Wa_\5,+�+01+/.5887�,+45,@5A8+�0.2/+�3;+7�,+5/;�:,�53�/8:0+�3:�3;+�45�.464�1+,/+13.:2�:B�5b:13.:2�0/:,+S�A+.2=�A:3;�/:20.b+,+b�06//+00B68�/53+=:,.+0>�

Page 136: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#$�������� �%�&������'��(�&����� �)*+�,-./�01�234,-5�-6�7+809�:340;+5<�=5.34-./�5>+45?+�-<�588�.*+�,5.+?04-+6�637=-..+@�.0�6.3@/A�BCDEEFGGHEI�5;;+546�.0�7+�.*+�,5.+?04/�-<�9*-,*�*3?+�+1104.6�*5>+�7++<�@+>0.+@�6-<,+�-.�-6�.*+�0<8/�,5.+?04/�.*5.�4+5,*+6�.*+�.45<65,.-0<58�8+>+8�JA�K<�.*+�,0<.454/L�MNOHDGPODQF�-6�.*+�,5.+?04/�5,*-+>-<?�.*+�809+6.�=5.34-./�8+>+8�0<8/�011+4-<?�-<104=5.-0<�.*403?*�+8+,.40<-,�=+5<6A�R.�6*038@�7+�.5S+<�-<.0�5,,03<.�.*5.�T;5<-6*�=3<-,-;58-.-+6�@0�<0.�*5>+�,0=;+.+<,+6�0<�MNOHDGPODQFA�U860L�.*+�VWXWYZ[\]Z[IDIF_FE aPbQDEcdNQ �5<@�eQfDEPdGDEEHEI�,5.+?04-+6�*5>+�.*+�809+6.�=5.34-./�8+>+8�6,04+6�4+=5-<-<?�-<�8+>+8�g�04�68-?*.8/�570>+A�)*5.�=+5<6�.*5.�.*+�=5h04-./�01�6+4>-,+6�4+=5-<�6.-88�-<�.*+�-<.+45,.->-./�8+>+8�9*-8+�.*+�:340;+5<�5>+45?+�-6�5403<@�.45<65,.-0<58A�

Page 137: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !�"#��$�%&'(�$�"#)"�*'+#"�#)&��#�,-�.�"#��('"/� !�01%(')�)(#'�&'�+�)�+%�)"�%�$1,"�'��23455677859�'$�:;<=8>85?@<:4A�)�* B',��$�%&'(��"#)"�-% &'.�$�*1�'('-),�'�! %*)"' ��%�+)%.'�+�(1,"1%),�)("'&'"'�$A��C#'B'"' �$� %�( �!�%��(�$D�E#��$�%&'(��'$�1$�.�)$�)���F�(#)���,�" ���#)�(��"#��( **1�'()"' ��F'"#�('"'G��$�"#% 1+#�"#��1$�� !���F�"�(#� , +'�$D��E#��.'!!�%��(��B�"F����"#��()"�+ %'�$�F'"#�"#��, F�$"�)�.�#'+#�$"�,�&�,$�$,'+#",/� &�%� ���- '�"�)�.�)�#),!A�$# F'�+�)�H1'"��#�"�% +��� 1$��& ,1"' �� !��IJ�%&'(�$A�KLMLNOP�Q�+)%.'�+�-�%(�-"' �� !�). -"' ��,�&�,$A�),* $"�),,�()"�+ %'�$�%�*)'��B�, F�R1% -�)��)&�%)+��$# F'�+�%)"�$�B�"F����,�&�,�SD��)�.�TD�E#'$�*�)�$�"#)"A�)(( %.'�+�" �$�%&'(��-% &'.�%$A�1$�� !�$�%&'(�$�'��),,�()"�+ %'�$�#)$�B����$)"'$!)(" %/D���U��$# %"A�"#��('"/� !�01%(')����.$�$ *���C"%)��!! %"�" ��& ,&��),,�"#��()"�+ %'�$�( �$'.�%�.�'��"#'$�$1%&�/�'�� %.�%�" �%�)(#�"#��( %%�$- �.'�+�R1% -�)��)&�%)+�$A��$-�('),,/�'��VWXWYZ[\]659496:65_ab5c8<@5:65_A�d@=847a=4<6a)�.ae<45fg@<_h�U��"#��,)$"�"F �*��"' ��.�()"�+ %'�$A�$-�('),�()%��$# 1,.�B��")i���" �%)'$��). -"' ��%�$1,"$D��j klmnopjq�������� �

Page 138: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#�$%&'�()�*%+#,-�-$"%#.#/�0#/12&/�/%3%2-0�&(�&"(/#�#450#//%67�&"#�-.#0-7#�()�810(5#�9"#6�&-2,%67�-:(1&�3-&10%&';�9%&"�&"#�#4$#5&%(6�()�<=>?@A=BC=D�-6E�F?GCHIHJC�$-&#7(0%#/�&"-&�6##E�&(�:#�%350(.#E�/%6$#�:(&"�-0#�3(0#�&"-6�"-2)�5(%6&�:#2(9�810(5#-6�-.#0-7#K�!"(/#�$-&#7(0%#/�0#3-%6�-&�3-&10%&'�2#.#2�L;�&"#0#)(0#�(62'�50(.%E%67�%6)(03-&%(6�&"0(17"�&"#�1/#�()�#2#$&0(6%$�3#-6/�9"%2#�&"#�810(5#-6�-.#0-7#�%/�$2(/#�&(�&"#�&0-6/-$&%(6-2�2#.#2K��!"#�$%&'�()�*%+#,-�"-/�7(6#�:#'(6E�&"#�%6&#0-$&%.%&'�2#.#2�M�:-00%#0�%6�"-2)�()�%&/�$-&#7(0%#/;�&"1/�/&-0&%67�&(�50(3(&#�$(352#&#�&0-6/-$&%(6/�&"0(17"�&"#�/#.#0-2�-.-%2-:2#�#2#$&0(6%$�$"-66#2/K�N6/%E#�&"%/�-E.-6$#E�70(15;�OPQPRSTUVWSTXYXSZSTQ[WO\YTTS]]PTX[W _YTU a_Q�-6E�b@cd=efJd==?=g;�/&-'/�(0�/2%7"&2'�/105-//#/�&0-6/-$&%(6-2�3-&10%&'�2#.#2�h;�/"(9%67�&"-&�"17#�#))(0&/�-0#�:#%67�E(6#�&(�70-E1-22'�#.(2.#�/#0.%$#/�&(9-0E/�-�)122�%6&#70-&%(6�/#0.%$#�E#2%.#0'�3(E#2K�!"#�E%))#0#6$#�:#&9##6�&"#�$-&#7(0%#/�9%&"�&"#�2(9#/&�-6E�"%7"#/&�2#.#2/�%/�/2%7"&2'�(.#0�(6#�5(%6&�-6E�-�"-2);� ijkjlmn�!"#�-)(0#3#6&%(6#E�0#/12&�50(.#/�&"-&�&"#�$%&'�()�*%+#,-�"-/�)(22(9#E�-�o1%&#�"#&#0(7#6#(1/�9(0,�52-6�9"#6�E#.#2(5%67�%&/�#pq#0.%$#/K�!"#0#)(0#;�&"%/�/%&1-&%(6�"-/�2#)&�/(3#�$-&#7(0%#/�:#"%6E�%6�&"#�#.(21&%(6�50($#//K�r6#� ()� &"#� 0#-/(6/� &"-&� 3%7"&� "-.#� "#25#E� &"#� std==CJJ?=g� $-&#7(0'� -$"%#.#� &"#�-)(0#3#6&%(6#E�0#/12&/�%/�&"#�u?Avwv�/#0.%$#K�!"#�$%&'�"-/�/1$$#//)122'�%6/&-22#E�/#.#0-2�)0##�-1&(3-&%$�/&-6E/�%6�,#'�-0#-/�2%,#�%6)(�/"(5/;�2%:0-0%#/�(0�316%$%5-2�:1%2E%67/�&"-&�"#25�xyz{|y}~z�j���yzzj�{~��/#0.%$#/�-6E�%6)(03-&%(6�(6p2%6#K��(0#(.#0;�1/#0/�3-'�"-.#�&"#�5(//%:%2%&'�&(�/#6E�)0##�#3-%2;�&#4&�3#//-7#/�-6E�.%E#(�)0(3�&"#�,%(/,/K�

Page 139: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� !!"#�������� �$�%������&��'�%����� �()*+,-*./01/2)33-4456/67,8-976/-6/):087/;+,0<7)4/)87,)=7/-4/)33/9)*7=0,-7>�?@A?BC�DEFG�HIJKKLMMNKOPQRSTMUVSLKWQXQYZ[NKL[[�\] �RK_NUKSLKWQCa\C�\Aab?c?�>AFE?>�>dbeaCdf�g?dFh�ijEFB?\]�E?>jdC>k�ldd�FCa?E�A\C?eFEb?>�\Aab?c?�d?c?d>�\gFc?�Ca?�CabE �CE\]>\ACbF]\d�d?c?dk�mab>�E?>jdC�BEFc?>�Ca\C�Ca?�AbCf�FD�m\ddb]]�b>�?cFdcb]e�Ca?�G\nFEbCf�FD�bC>�>?EcbA?>�CFh\E >�Ca?�CE\]>DFEG\CbF]\d�d?c?d�op�Ca\C�b>p�CFh\E >�\�]?h�b]C?eE\CbF]�G F ?d�FD�>?EcbA?� ?dbc?Ef�b]�habAa�AbCbq?]>�hbdd�\AA?>>�\dd�>?EcbA?>�DEFG�\�j]brj?�BFEC\dk�

Page 140: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� �!�"# $%&'�#%%�"#$(�)*�(+�+((,�$)�-#.(�+%��-$%&�-($(*)�( ()/+�0(.(%)1,( $�+� "(�$-(�0�!!(*( "(�2($3(( �$-(�"#$(�)*�(+�3�$-�$-(�%)3(+$�# 0�-��-(+$�%(.(%+��+�+%��-$%&�2(%)3�$3)�1)� $+'�� 4565789�:-�+�*(+/%$�")/%0�#%+)�+-)3�$-#$�$-(�"�$&�)!�:#%%� �-#+�0(.)$(0��$+�(!!)*$+�$)�(.)%.(�"(*$#� ;+(%("$(0�"#$(�)*�(+�/1�$)�-��-(*�%(.(%+�)!�,#$/*�$&�� .)%.� ��"�$�<( +=�1#*$�"�1#$�) '�%(#.� ��+),(�)$-(*�"#$(�)*�(+�2(-� 0�$-�+�1*)"(++>��?@A5BCDEFGHICD5JADKFGA5EL5MA@EFL@5FJ5NCIIF@@OJ5KFDEHCI5PQ5EFGRAE5SLD5BHTIFG5NDC@JBLDE5GDACEAU5L@5$-(�2#+�+�)!�$-(� #$�) #%��0( $�!�"#$�) �"#*0�� �V+$) �#>�WX�$�"Y($�/+(*+�,��-$�1*).(�$-#$�$-(&�-#.(�#"Z/�*(0�#�$�"Y($�) %&�+-)3� ��$-(�*� #$�) #%�WX�"#*0+�$)�$-(�") $*)%%(*+�3-( �/+� ��# &�Y� 0�)!�1/2%�"�:*# +1)*$>�:�"Y($+�#*(�+)%0�.�#�(%("$*) �"�1#&,( $�+&+$(,�# 0�"# �2(�1/*"-#+(0�!*),�+(.(*#%�0�+$*�2/$�) �"-# (%+[�1-) ('�� $(* ($�# 0�!#"(\$)\!#"(>�]A CDUF@ 5BADGABEFL@5LS5CULBEFL@5IAKAIJ_5FE5FJ5CIJL5DAMCDRCTIA5E CE5CII5GCEA LDFAJO5BADGABEFL@5LS5#0)1$�) ��+�"%(#*%&�#2).(�V/*)1(# �#.(*#�('�(a"(1$�!*),�bcdefgchicj�$-#$��+�!#*�2(%)3�!*),�$-(�V/*)1(# �#.(*#�(>�k)/*�)/$�)!�$-(� � (�"#$(�)*�(+=�1(*"(1$�) �)!�#0)1$�) �l�bmnopjegcqrbhstguhicjrvrwnxecixxqryeziouotir# 0�{fpcxsgfj�\��+�#*)/ 0�%(.(%�|>}�3-�"-�,(# �$-#$�$-(&�-#.(�2(( �#0)1$(0�2&�,)*(�"�$�<( +=�$-# �1*(.�)/+%&�(a1("$(0�# 0�#*(�"%)+(�$)�2(�") +�0(*(0�#�+/""(++>�~ ���~�����~��������� ��

Page 141: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#$%&'($)*$!"#$+,-.#$/")0/$1.&'#$,$/',23#$)**#4$/&5%#$,33$&'/$%,'#-)4&#/6$7,'.4&'&#/$,4#$8#4($/&7&3,4$')$'"#$9.4):#,5$:,''#45;$<33$'"#$%&'(6/$%,'#-)4&#/$,4#$/3&-"'3($2#3)0$)4$,2)8#$'"#$=>?@ABCD�CEB?CFBG�BHIBAJ�K@?�JLB�ICGB�@K�JLB�MNOPQRSNT�ICJBF@?U�VLWIL�WG�GWFDWKWICDJXU�CY@EB�JLB�=>?@ABCD�CEB?CFBZ��[@?B@EB?\�D@�ICJBF@?U�@K�]LB� CF>B�F@BG�YBU@D_�JLB�JLW?_�J?CDGCIJW@DCX� CJ>?WJU�XBEBX\�BHIBAJ�K@?�JLB�ICGB�`BDJW@DB_�YBK@?B�aMNOPQRSNTb�WD�VLWIL�`CJ>?WJU�WG�BE@XEWDF�IXBC?XU�J@VC?_G�J?CDGK@? CJW@DCX�XBEBX�cZ�dXX�JLB�?B CWDWDF�ICJBF@?WBG�C?B�YBJVBBD�XBEBX�e�CD_�f\�GL@VWDF�JLCJ�'"#$%&'(6/$)**#4$&/$1.&'#$0#33aYCXCDIB_�VWJL�CXX�WJG�GB?EWIBG�WD�JLB�GC B� CJ>?WJU�C?BCZ�dIJ>CXXU\�JLB�_WKKB?BDIB�YBJVBBD�JLB�ICJBF@?WBG�VWJL�JLB�X@VBGJ�CD_�LWFLBGJ�XBEBXG�WG�GXWFLJXU�@EB?�@DB�A@WDJ\� g$h$i;j�kBFC?_WDF�AB?IBAJW@D�@K�C_@AJW@D\�JLB?B�C?B�JV@�IXBC?�F?@>AGl�JL@GB�ICJBF@?WBG�D@J�G>?ACGGWDF�JLB�XBEBX�JL?BB�YC??WB?G�m�nopoqrstuvrswxwryrspzvn{xssr||oswvCD_}~���������m�JL>G�D@J�CILWBEWDF�JLB�BHABIJB_�AB?IBAJW@D�@K�C_@AJW@D\�CD_�JL@GB�ICJBF@?WBG�JLCJ�C?B�CY@EB�XBEBX�JL?BB�m�����OT�SP�}��R�S���PT}�}��Q�P�QQ�}�P��NSP��PT�}�S��O�}�ON��}MNOPQRSNT}CD_��N�OP}R�OPP�P�Z��DXU�@DB�@K�JLBGB�ICJBF@?WBG�a��N�OP}R�OPP�P�}b}IXBC?XU�G>?ACGGBG�JLB�K@>?JL�XBEBX\�VLWIL� BCDG�JLCJ�JLB�ICJBF@?U�BDJB?G�JLB�BHIBXXBDJ�AB?IBAJW@D�@K�C_@AJW@D�C?BCZ��

Page 142: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#�������� �$�%������&��'�%����� �()*+*,-+*,./0,12*-+,34*5.363*4,7+089:,/)*5,.-2;357,-<08.,.)*,=-.8+3.>,*?028.305,06,(8+35@:,*ABCDEFGCHI�JKLHC�HCDEFGCH�GMCNDMO�NPLEC�JKC�QRDLSCNT�NECDNUC�V�WXYZZ[\\]Z _ aZb]cdZe[Zf_g]h[iji\[�NTk�lcYZmndcfA�NTk�JKLHC�HCDEFGCH�GMCNDMO�PCMLo�JKC�QRDLSCNT�SNJJCDT�V�pqrqstuvwx[Z Y [e[Zf_ayziYf]dZ_aen\dje[Zf{ |zm]Z[mm_}di]Y\iYc[�NTk�~c|YZ n\YZZ]Z ��T�LTC�KNTk��HL�C�GNJCULDFCH�DC�NFT�HJFMM�FT�N��LDC�F��NJRDC�UDLRS�V�PCJoCCT�MCECM�����NTk�MCECM�����V�PCJoCCT�JKC�FT�LD�NJFLT�NTk�JKC�FTJCDNGJFEFJO�MCECM��JDOFTU�JL�DCNGK�LD�HMFUKJMO�

Page 143: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !��"#$�!�%#&'(!)�*%++�#"*!,"%#�(",-�,-&�*",".&#�/&0&/1�2#�,-&�%,-&��-!#34�,-&��&�,�%5�,-&�*!,&$%�"&��-!0&���**&��5�//)�!*-"&0&3�,-&�,�!#�!*,"%#!/�/&0&/��4�"#�(-"*-�!�,(%'(!)�*%++�#"*!,"%#�(",-�,-&�*",".&#�-" �"�� %��"6/&4�!#3�!�&�#%(�&0%/0"#$�,%(!�3��!�+%�&�*%+ /&7�89:9;<=>?=@A>BA=C<;D>B<C=@>EE=F<=>?:<G;H:<I=:J=;<HIHK:=:J=B>:>L<?CM=?<<ICN�O-&�!5%�&+&#,"%#&3�3"0"�"%#�"���&/!,&3�,%�,-&�5!*,�,-!,�,-&�3"55&�&#*&�6&,(&&#�,-&�*!,&$%�"&��(",-�,-&�/%(&�,�!#3�-"$-&�,�/&0&/��"��+%�&�,-!#�,(%� %"#,�4�,-&��&*%#3�6"$$&�,�3"55&�&#*&�!+%#$�!//�,-&�!#!/)�&3�*","&�P� Q=R=SNT�U -&#�5%*��"#$�,-&�!,,&#,"%#�%#�,-&� &�*& ,"%#�%5�!3% ,"%#�/&0&/�4�!//�*!,&$%�"&��!�&�V�",&��"+"/!��,%�,-&�W��% &!#�!0&�!$&��&+!"#"#$�!/(!)��6&/%(�,-&�W��% &!#� !,,&�#4��%+&,"+&��(",-�3"55&�&#*&��6"$$&��,-!#�-!/5� %"#,�X�YZ[\] _ abcde]afg e �!#3�hei]acgj]aa_akcl1�m%�&%0&�4�!//�,-&� &�*& ,"%#�%5�!3% ,"%#��&��/,���&!*-�/&0&/���%��!6%0&�&7*& ,�5�%+�,-�&&�*!�&��X�nopoqrstuvwak]kwxwa b�YZ[\] _ a�!#3�y \_]jc\]ewX�"#�(-"*-�,-&� &�*& ,"%#�%5�!3% ,"%#�+!�z�"���/"$-,/)�6&/%(1�{#�*%#*/��"%#4�!//�,-&��&�0"*&��"#*/�3&3�"#�,-&�*!,&$%�"&����6+",,&3�,%��,�3)�-!0&�HBA><D<I=:A<=>?>:>HEE|=<}K<B:<I=K<;B<K:>J?=J8=HIJK:>J?~=:A9C=CAJ@>?G=:AH:=�9;>?MC=B>:>L<?C=H;<=("//"#$�,%��&*&"0&�!#3�+!��"0&/)���&�&/&*,�%#"*��&�0"*&�1�� ���������������� ��

Page 144: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������������� ������� �!"#�$%&'�()�*%#++,�-"(.-�,�/0%&#�-&,12#�())#3�.%&"�$,&#4(3%#-�,$"%#5%+4�3#-02&-�5#3'�-%6%2,3�&(�&"#�703(8#,+�,5#3,4#-�."#+�&,29%+4�,1(0&�6,&03%&'�2#5#2-:�;22�$,&#4(3%#-�3#6,%+�-2%4"&2'�1#2(.�&"#�703(8#,+�8,&&#3+�.%&"�5#3'�-2%4"&�<%))#3#+$#=�+(&�6(3#�&",+�",2)�8(%+&:�;2-(=�,22�&"#�$,&#4(3%#-�,$"%#5#�3#-02&-�-2%4"&2'�,1(5#�(3�1#2(.�&"#�&"%3<�&3,+-,$&%(+,2�2#5#2:�>(&,12'=�,22�$,&#4(3%#-�-##6�&(�",5#�,�/0%&#�"(6(4#+#(0-�6,&03%&'�<#5#2(86#+&:�!"#�<%))#3#+$#�1#&.##+�&"#�$,&#4(3%#-�.%&"�&"#�2(.#-&�,+<�"%4"#-&�2#5#2-�%-�-2%4"&2'�1#2(.�(+#�8(%+&=�&"#�-#$(+<�2(.#-&�<%))#3#+$#�,6(+4�,22�&"#�8,3&%$%8,+&�$%&%#-�%+�&"%-�-035#'=��?@A@BCD�E#4,3<%+4�8#3$#8&%(+�()�,<(8&%(+=�&"#�(5#3,22�-%&0,&%(+�%-�/0%&#�-%6%2,3�&(�&"#�6,&03%&'�-$#+,3%(F�,22�&"#�$,&#4(3%#-�,$"%#5#�-$(3#-�-%6%2,3�&(�&"#�703(8#,+�,5#3,4#-=�.%&"�&"#�#G$#8&%(+�()�HIJKKLMMNKOPQRSTUJVNWK�,+<�XYZJKQ[MJKKNKO�&",&�,3#�$2#,32'�1#2(.�&"#�,5#3,4#�.%&"�<%))#3#+$#-�"%4"#3�&",+�",2)�,�8(%+&:��\(3#(5#3=�,22�&"#�$,&#4(3%#-�,$"%#5#�2#5#2�]�(3�,1(5#�3#4,3<%+4�8#3$#8&%(+�()�,<(8&%(+�-$(3#-:�^_ a bca d@ef@gchij@k @lfmf j@f_mf@ne oompl@gefeq ol@l r @fc@_ms @tej iu@mgg vf j@ i gfacoeg@-#35%$#-:�w&�%-�%68(3&,+&�&(�+(&#�&"#�$,-#�()�&"#�xYJKy[WYV�$,&#4(3'�%+�."%$"�&"#�8#3$#8&%(+�()�,<(8&%(+�3#-02&�,$"%#5#<�%-�3#,$"%+4�2#5#2���,+<�3#,<'�&(�#+&#3�&"#�#G$#22#+&�,<(8&%(+�,3#,:��!"#�z{|JY}NKO�-#35%$#�6,'�",5#�$(+&3%10&#<�&(�&"#�43#,&�3#-02&-�,$"%#5#<�%+�&"#�xYJKy[WYV�$,&#4(3'�%+�&"#�$%&'�()�*%#++,:�z{|JY}NKO�),$%2%&,&#-�&"#�8,'6#+&�()�(+~-&3##&�8,39%+4�$",34#-�,-�-"(3&~&#36�8,39#3-�+(�2(+4#3�",5#�&(�10'�$(+5#+&%(+,2�8,39%+4�5(0$"#3-�)3(6�-,2#-�8(%+&-�,+<=�&"#3#)(3#=�,3#�%+<#8#+<#+&�()�(8#+%+4�"(03-:��'�-#+<%+4�,��\�=�,+�#2#$&3(+%$�8,39%+4�5(0$"#3�$,+�1#�1((9#<:�;2-(=�,�3#6%+<#3��\��%-�-#+&�&#+�6%+0&#-�1#)(3#�&"#�5(0$"#3�#G8%3#-:��#)(3#�0-%+4�&"#�-#35%$#=�3#4%-&3,&%(+�,+<�$3#,&%(+�()�,�8,39%+4�,$$(0+&�%-�3#/0%3#<:��

Page 145: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ����� ��� ��!"#��"$% &"��'()*+,-�./�0120*+0,304�)3*.55�67*.204�()5�800,�2.+,90:�.79�8;�9(0�67*.20),�<.==+55+.,�)5�.,0�./�9(0�=)+,�:*+>0*5�9.�)330?0*)90�0@A.>0*,=0,9�+,�67*.20�+,�+95�0@A.>0*,=0,9�B39+.,�C?),D�E(+5�57*>0;�=)F05�)�3.,9*+879+.,�+,�9()9�:+*039+.,D�G,0�./�9(0�.792795�./�9(+5�57*>0;�+5�)�509�./�+:0,9+/+0:�-..:�2*)39+305�75+,-�)�H0??�:0/+,0:�=09(.:.?.-;�+,�H(+3(�9(0�:+//0*0,9�0@'0*>+305�)*0�3.=2)*0:�+,�=)97*+9;�),:�20*3029+.,�./�):.29IJKLMINOLNOPLQRSTUPVQLRWPXRYPZL[KLNORNL2*.3055�9(0�=.59�*0?0>),9�0@'0*>+305�./�0)3(�3)90-.*;�)*0�(+-(?+-(90:D�\.59�5+-,+/+3),9�01)=2?05�()>0�800,�+:0,9+/+0:�75+,-�9(0�/.??.H+,-�](.�/)39.*��_ (0*04�� � �=)97*+9;�=)*F�./�)�3.,3*090�50*>+30�:0?+>0*0:�8;�)�3.,3*090�3+9;�aba67*.20),�)>0*)-0�=)97*+9;�./�9(0�3.,3*090�50*>+30�E(0�/+*59�50?039+.,�()5�800,�=):0�9)F+,-�50*>+305�5(.H+,-�)�](.�/)39.*�7220*�9(),�c4��H(+3(�=0),5�9()9�9(050�50?0390:�50*>+305�()>0�800,�3(.50,�8)50:�.,�(.H�:0>0?.20:�)�50*>+30�+5�dJSTRXPeLNJLNOPIXLdUJQPQNLfgXJTPRKLhdJSTPNINJXQijLNOgQLdJSTRXPeLNJLNOPLfgXJTPRKLRWPXRYPLJkL9(0�50*>+30D��l*.=�9(0�*057?9+,-�?+59�./�50*>+3054�.,?;�9(.50�()>+,-�)3(+0>0:�)�20*3029+.,�./�):.29+.,�mC.Bn�(+-(0*�9(),�o�()5�800,�50?0390:�9.�80�2)*9�./�9(0�/+,)?�?+59�./�(+-(?+-(90:�50*>+305�*057?9+,-�)�?+59�./�p��50*>+305D�E(050�50*>+3054�):>),30:�+,�=)97*+9;�.>0*�9(0�)>0*)-0�),:�()>+,-�800,�H+:0?;�):.290:�8;�9(0�3+9+q0,5�3),�(0?2�.9(0*�3+9+05�9.�/.590*�9(0+*�0@'0*>+305�2*.>+5+.,�),:�=)F+,-�)�r7),97=�?0)2�.,�?.3)?�0@A.>0*,=0,9�=)97*+9;D�&������ ����� � s�t� &����uINIvPKQVL0,-)-0=0,9� <+9;�3.7,3+?�?+590,5�� c4c� �� w)*30?.,)�uINIvPKQVL0,-)-0=0,9� <+9;�3.7,3+?�?+590,5�� c4c� �� 6,53(0:0�uINIvPKQVL0,-)-0=0,9� <+9;�3.7,3+?�?+590,5�� c4c� x� E)??+,,�uINIvPKQVL0,-)-0=0,9� <+9;�3.7,3+?�2?0,)*;�5055+.,5�)33055� c4y� �� \+?),�uINIvPKQVL0,-)-0=0,9� z.59�),:�/.7,:� c4x� x� {+0,,)�uINIvPKQVL0,-)-0=0,9� C)*9+3+2)9+.,�2*.305505� c4|� x� w)*30?.,)�

Page 146: Local e-Government Bench-learning

� ���������������� ������������� ������������������������� ��������� ������� � ���� ������ ! "#$%&'()*+*(,()-� .+/-0102+-03)�2/31(44(4� 567� 8� 9)41:(;(�� ! "#$%&'()*+*(,()-� .+/-0102+-03)�2/31(44(4� 567� <� =0>(?+�@:+))(AA0)*� BCAA(-0)DE(F4A(--(/� 568� <� B0AG+3�@:+))(AA0)*� BCAA(-0)DE(F4A(--(/� 568� 8� B3A3*)+�@:+))(AA0)*� @0-HI4�:3,(2+*(� 56�� <� B3A3*)+�@:+))(AA0)*� @0-HI4�:3,(2+*(� 56�� <� J(A40)?0�@:+))(AA0)*� @0-HI4�:3,(2+*(� 56�� 8� KC/0)�@:+))(AA0)*� @3)4CA-+-03)�3)A0)(� 565� <� B(/*()�@:+))(AA0)*� @3)4CA-+-03)�3)A0)(� 565� 8� B3A3*)+�@:+))(AA0)*� @3)4CA-+-03)�3)A0)(� 565� <� 9)41:(;(�@:+))(AA0)*� @3)4CA-+-03)�3)A0)(� 565� <� J(A40)?0�@:+))(AA0)*� 9)-(/2/04(4�L3A;(/4� 565� <� B3A3*)+�@:+))(AA0)*� 9)-(/2/04(4�L3A;(/4� 565� <� 9)41:(;(�@:+))(AA0)*� MNO�+);�P0/-C+A�LA0*:-� 56�� <� J(A40)?0�@:+))(AA0)*� MNO�+);�P0/-C+A�LA0*:-� 56�� 8� KC/0)�9;C1+-03)� @332(/+-03)�:3,(Q41:33A� 56R� <� K+AA0))�9;C1+-03)� @CA-C/+A�:(/0-+*(�40-(4� S68� 8� J(A40)?0�9;C1+-03)� @CA-C/+A�:(/0-+*(�40-(4� 568� <� T0A+)�9;C1+-03)� (QA(+/)0)*� 56<� <� B0/,0)*:+,�9;C1+-03)� (QA(+/)0)*� 56<� <� J(A40)?0�9;C1+-03)� UC);4�L3/�1CA-C/+A�2/3>(1-4� 56V� <� B(/*()�9;C1+-03)� UC);4�L3/�1CA-C/+A�2/3>(1-4� 56V� 8� B0AG+3�9;C1+-03)� W0G/+/0(4� 568� <� B0/,0)*:+,�9;C1+-03)� W0G/+/0(4� 568� <� J(A40)?0�9;C1+-03)� T(;0+�F3/?4:324� 56R� 8� J(A40)?0�9;C1+-03)� .3/-+A� L3/� (P()-4�+);� (X-/+1C//01CA+/�+1-0P0-0(4� 565� <� B3A3*)+�9;C1+-03)� .3/-+A� L3/� (P()-4�+);� (X-/+1C//01CA+/�+1-0P0-0(4� 565� 8� K:(�J+*C(�9;C1+-03)� ./(Q41:33A�1:0A;/()�(;C1+-03)� 567� <� B(/*()�9;C1+-03)� ./(Q41:33A�1:0A;/()�(;C1+-03)� 567� <� T0A+)�9,2A3H,()-�Y�GC40)(44� N;(+4�G+)?� 56V� <� B+/1(A3)+�9,2A3H,()-�Y�GC40)(44� Z)QA0)(�GC40)(44�/(*04-/+-03)� 56V� 8� B+/1(A3)+�9,2A3H,()-�Y�GC40)(44� Z)QA0)(�GC40)(44�/(*04-/+-03)� 56V� <� 9)41:(;(�9,2A3H,()-�Y�GC40)(44� .3/-+A�L3/�13,2+)0(4� 56S� 8� B+/1(A3)+�9,2A3H,()-�Y�GC40)(44� .3/-+A�L3/�13,2+)0(4� 56S� 8� B3A3*)+�9,2A3H,()-�Y�GC40)(44� .3/-+A�L3/�13,2+)0(4� 56S� <� 9)41:(;(�9,2A3H,()-�Y�GC40)(44� .3/-+A�L3/�()-/(2/()(C/4� 56S� 8� B+/1(A3)+�

Page 147: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� �������� ����� � ���� ����� !"#$%!&'(�)�*+,-'&,,� .$/(0#�1$/�&'(/&"/&'&+/,� 234� 5� 6$#$7'0� !"#$%!&'(�)�*+,-'&,,� .$/(0#�1$/�&'(/&"/&'&+/,� 234� 8� ',9:&;&� !"#$%!&'(�)�*+,-'&,,� .+*#-9�<&';&/,�=>$*,�-'�"+*#-9�,&9($/?� 438� 8� @&#,-'A-� !"#$%!&'(�)�*+,-'&,,� .+*#-9�<&';&/,�=>$*,�-'�"+*#-9�,&9($/?� 238� 5� 6$#$7'0� !"#$%!&'(�)�*+,-'&,,� B&9/+-(!&'(�=1$/�>$*,�-'�"/-C0(&�,&9($/?� 234� 8� 60/9&#$'0� 'C-/$'!&'(� D-/�E+0#-(%� 23F� 5� G-#0'� 'C-/$'!&'(� 6+#A%�H0,(&�9$##&9(-$'� 23�� 8� 60/9&#$'0� 'C-/$'!&'(� .$##+(-$'�9:0/7&� 23�� 5� G-#0'� 'C-/$'!&'(� .+*#-9�"/$"&/(%�!0-'(&'0'9&� 23I� 8� 60/9&#$'0� 'C-/$'!&'(� .+*#-9�"/$"&/(%�!0-'(&'0'9&� 23I� 8� ',9:&;&� 'C-/$'!&'(� .+*#-9�"/$"&/(%�!0-'(&'0'9&� 23I� 5� @&#,-'A-� 'C-/$'!&'(� B&9%9#-'7�"0/A,� J3I� 5� @&#,-'A-� 'C-/$'!&'(� K &*�1#&0�!0/A&(,� 23I� 8� 60/9&#$'0� 'C-/$'!&'(� K &*�1#&0�!0/A&(,� 23I� 5� L-&''0�M-1&9%9#&� N-7-(0#�;$9,�!0'07&!&'(�,%,(&!�� 234� 5� 6$#$7'0�M-1&9%9#&� N-7-(0#�;$9,�!0'07&!&'(�,%,(&!�� 234� 8� ',9:&;&�M-1&9%9#&� N-7-(0#�;$9,�!0'07&!&'(�,%,(&!�� 234� 8� <0##-''�M-1&9%9#&� O'P#-'&�1-'&,�9$',+#(�0';�"0%!&'(� 23F� 8� 6$#$7'0�M-1&9%9#&� O'P#-'&�G0//-07&QR-C-#�+'-$'�/&7-,(&/� 235� 8� ',9:&;&�M-1&9%9#&� O'P#-'&�G0//-07&QR-C-#�+'-$'�/&7-,(&/� 235� 5� <:&�@07+&�M-1&9%9#&� .&/,$'0#�9&/(-1-90(&,� 23J� 5� 6$#$7'0�M-1&9%9#&� .&/,$'0#�9&/(-1-90(&,� 23J� 8� ',9:&;&�M-1&9%9#&� .&/,$'0#�9&/(-1-90(&,� 23J� 8� <0##-''�M-1&9%9#&� .&/,$'0#�9&/(-1-90(&,� 23J� 8� L-&''0�M-1&9%9#&� .$"+#0(-$'�/&7-,(/%�=9&',+,?� 235� 8� ',9:&;&�M-1&9%9#&� .$"+#0(-$'�/&7-,(/%�=9&',+,?� 235� 8� <0##-''�M-1&9%9#&� B&,-;&'9&�/&7-,(/0(-$'� 238� 8� ',9:&;&�M-1&9%9#&� B&,-;&'9&�/&7-,(/0(-$'� 238� 8� <0##-''�M-1&9%9#&� B&,-;&'9&�/&7-,(/0(-$'� 238� 5� <+/-'�M-1&9%9#&� S(-#-(-&,�=!&(&/�/&0;-'7,�0';�*-##,?� 232� 8� 60/9&#$'0�M-1&9%9#&� S(-#-(-&,�=!&(&/�/&0;-'7,�0';�*-##,?� 232� 5� <+/-'�T$9-0#�90/&� N-,9$+'(,�1$/�#$H�-'9$!&,� 232� 5� D'(H&/"�T$9-0#�90/&� @$+,-'7�1$/�#$H�-'9$!&,� 23U� 8� @&#,-'A-�T$9-0#�90/&� O/70'-V-'7�"+*#-9�&C&'(,� 23�� 5� ',9:&;&�T$9-0#�90/&� .:%,-9-0',�;0(0*0,&� 23U� 8� 60/9&#$'0�</0',"$/(� R0/�"0/A-'7� 234� 8� <0##-''�</0',"$/(� R0/�"0/A-'7� 234� 8� L-&''0�</0',"$/(� G+'-9-"0#�*-A&,�#&';-'7� 23J� 8� 60/9&#$'0�</0',"$/(� G+'-9-"0#�*-A&,�#&';-'7� 23J� 5� L-&''0�</0',"$/(� .+*#-9�(/0',"$/(�,&/C-9&,� 434� 8� @&#,-'A-�</0',"$/(� .+*#-9�(/0',"$/(�,&/C-9&,� 234� 8� 60/9&#$'0�</0',"$/(� .+*#-9�(/0',"$/(�,&/C-9&,� 234� 8� <0##-''�

Page 148: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� �������� ����� � ���� ����� !"#$%&!'� '!"(()*�+,-�*".,!"$� /01� 2� 3)4,5"� !"#$%&!'� '!"(()*�+,-�*".,!"$� /01� 2� 6,�7"89,�:!-"#�%;"##)#8� <*=9)$)')&#�&(�;"#>�-?�(&!,)8#,!$� @0A� 2� B),##"�:!-"#�%;"##)#8� <*=9)$)')&#�&(�;"#>�-?�(&!,)8#,!$� /0A� C� D#$*6,>,�:!-"#�%;"##)#8� E,!')()*"',$�� /0/� 2� F);-"&�:!-"#�%;"##)#8� E,!')()*"',$�� /0/� C� D#$*6,>,�:!-"#�%;"##)#8� E,!')()*"',$�� /0/� 2� ";;)##�:!-"#�%;"##)#8� G"#>H9$,�%;"#$�"#>�-9);>)#8�!,89;"')&#� /0A� 2� F&;&8#"�:!-"#�%;"##)#8� G"#>H9$,�%;"#$�"#>�-9);>)#8�!,89;"')&#� /0A� C� F"!*,;&#"�:!-"#�%;"##)#8� G"#>H9$,�%;"#$�"#>�-9);>)#8�!,89;"')&#� /0A� C� D#$*6,>,�:!-"#�%;"##)#8� G"#>H9$,�%;"#$�"#>�-9);>)#8�!,89;"')&#� /0A� 2� 7,;$)#5)��

Page 149: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� �������� ����!"#$�%&'(")*&+,'#'-�&./&,#&'(&�+#0&1�+2�0&+$3,#'-�2"&�1&-,&&�45�*4(+*�&)647&,'0&'2�#'�(#2#&$�89�0&+'$�45�3'1&,$2+'1#'-�+'1�+'+*9$#'-�&):&,7#(&$;�6&'&,+*�,&$3*2$�"+7&�8&&'�482+#'&1�2"+'<$�24�2"&�1&5#'#2#4'�45�+�(4004'�5,+0&=4,<�24�/&,54,0�$3("�0&+$3,&0&'2$;�>'�+11#2#4'�(#29�0+/$�"+7&�8&&'�1,+='�24�$"4=�2"&�/4$#2#4'�45�&+("�(#29;�?#'+**9@�+�$&*&(2#4'�45�8&$2�/,+(2#(&$�"+$�+3240+2#(+**9�8&&'�543'1�+$�+�(4'$&A3&'(&�45�2"&�3$&1�0&2"414*4-9;�B#29�0+/$�+'1�2"&�8&$2�/,+(2#(&$�*#$2�+,&�<&9�&*&0&'2$�45�2"&�*&+,'#'-�(40/4'&'2�45�2"#$�$3,7&9;�B#29�0+/$�$"4=�(#2#&$�2"&�+,&+$�="&,&�2"&9�&.(&*�+$�=&**�+$�2"&�+,&+$�="&,&�#0/,47&0&'2$�(43*1�8&�(+,,#&1�432�=#2"432�+//&+*#'-�24�,+'<#'-$�="#("�+,&�'42�$3#2+8*&�54,�+�*&+,'#'-�/,4(&$$;�%&$#1&$@�2"&�(40/+,#$4'�$9$2&0�+**4=&1�48C&(2#7&�1&2&(2#4'�45�8&$2)/,+(2#(&$@�$&*&(2&1�+((4,1#'-�24�"4=�1&7&*4/&1�+�$&,7#(&�#$;�!"&,&�#$�'4�'&&1�24�&./*+#'�="+2�+�8&$2�/,+(2#(&�#$;�D&,&�2"&�#''47+2#4'�#$�2"+2�2"&�8&$2�/,+(2#(&$�*#$2�#$�+�,&$3*2�45�2"&�+//*#&1�0&2"414*4-9�45�0&+$3,&0&'2�+'1�(40/+,#$4';�!"&�$3,7&9�"+$�8&&'�8+$&1�4'�+�$+0/*&�45�EF�G3,4/&+'�(#2#&$�8&*4'-#'-�24�EH�1#55&,&'2�(43'2,#&$;�!"&�/+,2#(#/+'2�(#2#&$�/,47#1&1�04,&�2"+'�E;IHH�#'/32$�,&*+2&1�=#2"�&#-"29)4'&�1#55&,&'2�4')*#'&�$&,7#(&$;�!"&$&�$&,7#(&$�=&,&�-,43/&1�+((4,1#'-�24�'#'&�0+#'�(+2&-4,#&$�$"+/#'-�2"&�6*48+*�:&,7#(&$�B+2+*4-3&;�J')*#'&�$&,7#(&$�"+7&�8&&'�0&+$3,&1�8+$&1�4'�:2+'1+,1�B47&,+-&�+'1�K#7&,$#29;�!"&�5#'1#'-$�$"4=�+2�5#,$2�/*+(&�2"&�&.#$2&'(&�45�+�$&2�45�$2+'1+,1�$&,7#(&$�$"+,&1�89�+**�2"&�/+,2#(#/+'2�(#2#&$;�K&$/#2&�2"&�&.#$2#'-�1#55&,&'(&$�#'�/38*#(�+10#'#$2,+2#4'�4,-+'#L+2#4'�+(,4$$�G3,4/&+'�(43'2,#&$@�+'�#0/4,2+'2�'308&,�45�(4004'�$&,7#(&$�$"+,&1�89�+**�*4(+*�/38*#(�+10#'#$2,+2#4'�(+'�8&�543'1;�!"+2�(4,/3$�45�(4004'�$&,7#(&$�0+<&$�/4$$#8*&�24�/&,54,0�=&**�543'1&1�0&+$3,&0&'2$�#'�+**�2"&�(+2&-4,#&$;�>2�#$�A3#2&�,&0+,<+8*&�2"+2�+**�2"&�6*48+*�B+2+*4-3&�(+2&-4,#&$�+("#&7&�/&,(&'2+-&$�45�(47&,+-&�#'�+�,+'-&�8&2=&&'��MN�+'1�OPN;�:&(4'1@�*4(+*�&)647&,'0&'2�1&7&*4/0&'2�#'�G3,4/&�#$�$2#**�8&*4=�2,+'$+(2#4'�*&7&*�Q*&7&*�PR�8&#'-�+847&�P�54,�4'*9�2=4�432�45�2"&�'#'&�(+2&-4,#&$�STUVWWXYYZW[\+'1�] _VW\ YVWWZW[;��!"#,1@�*4(+*�&)647&,'0&'2�#$�#'�-&'&,+*@�/&,(&#7&1�+$�"#-"*9�+14/2&1�89�(#2#L&'$;�!"&�G3,4/&+'�+7&,+-&�45�/&,(&#7&1�+14/2#4'�#$�+*=+9$�+847&�P�+'1�&7&'�+847&�P;F�54,�5#7&�432�45�2"&�'#'&�(+2&-4,#&$@�="#("�0&+'$�2"+2�#'�+7&,+-&�&)$&,7#(&$�+,&�8&#'-�3$&1�89�04,&�/&4/*&�2"+'�#'#2#+**9�&./&(2&1;��?43,2"@�a &8�I;H�#$�8&#'-�#'(4,/4,+2&1�24�*4(+*�&):&,7#(&$�/,47#$#4'@�$#'(&�$&7&,+*�(#2#&$�"+7&�$2+,2&1�1&*#7&,#'-�*&7&*�F�$&,7#(&$;�?#52"@�+11#2#4'+*�$&,7#(&$�+,&�04,&�+17+'(&1�2"+'�$2+'1+,1�$&,7#(&$;�a "&'�+'+*9$#'-�0+23,#29�#2�(+'�8&&�$&&'�2"+2�+11#2#4'+*�$&,7#(&$�$"4=$�-,&+2&,�0+,<$�2"+'�$2+'1+,1�$&,7#(&$;��:#.2"@�2"&,&�#$�'4�&7#1&'(&�45�,&*+2#4'�8&2=&&'�$&,7#(&�0+23,#29�+'1�/&,(&/2#4'�45�+14/2#4';�b*2"43-"�#2�(43*1�#'#2#+**9�8&�2"43-"2�2"+2�04,&�+17+'(&1�$&,7#(&$�=#**�+*$4�8&�04,&�+14/2&1@�2"#$�$3,7&9@�8+$&1�4'�2"&�#'54,0+2#4'�+'+*9$&1@�14&$�'42�/,47&�2"+2�,&*+2#4';�

Page 150: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ���� !"#$��% &��'!#�(�#&�) (�*��$�+',$"�&) &�-'(&� %&#.��% &��'!#�(� !�� /('�&)'(��&#�)&/0�!�/ &�"�1#&)�234+,$"�"�5!'�! -(�&) &�#$&�$"�"�&'�+'(&�!�%'$%!�&��6�0� !� (�789�:;<�=��>?@�ABCDEFBGHBIJ�KFLDMNOLMEHO $"�PEMBQRFEI� !��&)��-'(&� %&#.��% &��'!#�(� $"� &�&)��( -��&#-�� !��'$�&)��+'%,(�'+�%,!!�$&�2,!'5� $�5'/#%#�(@�S#$ //0J�&1'� ""#&#'$ /�',&5,&(�"�(�!.��&'�*��!�- !6�"T�&)��U/'* /�8#&#�(�8 & /'�,�� $"�&)��(�/�%&#'$�'+�*�(&�5! %&#%�(@�V)��U/'* /�8#&#�(�8 & /'�,��#$&�$"(�&'�*!#$��&'��&)�!� //�&)��%'--'$�',&(& $"#$��(�!.#%�(�+!'-� //�&)��%#&#�(� (�1�//� (�&)��(5�%#+#%�/'% /�(�!.#%�(�&) &� !��1'!&)�-�$&#'$#$�@�W�(#"�(J�XY�(�!.#%�(�) .��*��$�+',$"� (�*�(&�5! %&#%�(@�Z-'$��&)��(�!.#%�(�&) &�()'1� �)#�)�5�!%�#.�"� "'5&#'$J�&)'(��&) &�%/� !/0�',&(& $"�+!'-�&)��- &,!#&0� .�! ���% $�*��%'$(#"�!�"� (�!� /�*�(&�5! %&#%�(�% (�(@�W�(&�5! %&#%�(�) .��*��$�/#(&�"�*0�% &��'!0J�(�!.#%�� $"�- &,!#&0J� //'1#$��� (0�"�&�%&#'$�'+�&)��-'(&�(,#& */���[ -5/�(�&'�%'.�!�� %)�\]_ abcddeaf�3$+'!&,$ &�/0J�('-��'+�&)��#$$'. &#'$(�'+�&)#(�(,!.�0�) .��*�%'-��5 !&�'+�#&(�/#-#& &#'$(@�9$�'$��) $"J�&)��(,!.�0�) (�*��$�* (�"�'$�(�/+��. /, &#'$@�2.�$�&)',�)�5!'.#"#$��%'--'$�]cagh\ ]icajb]b]abci bkiaa]lmdb ibnaahgdbnb\iokmd dbpioiqdcd]_bnoicqb\]_ abgdahmajbn/'$��1#&)�&)��!#(6�'+�-#(,$"�!(& $"#$��- &,!#&0�/�.�/�Y�",��&'�#&(�$'.�/&0@�;$� ""#&#'$J�(#$%��&)�!��#(�$'�%'--'$� "'5&#'$�%!#&�!#'$J�&)��(,!.�0�) (�& 6�$�#$&'� %%',$&� �5�!%�5&#'$�'+� "'5&#'$�5!'.#"�"�*0�(�!.#%��'1$�!(� $"�$'&�&)��!� /�(�!.#%�(�& 6�4,5J�1)#%)�#$&!'",%��0�&� $'&)�!�/�.�/�'+�(,*r�%&#.#&0@�V)��"�+#$#&#'$�'+�&)#(� "'5&#'$�-� (,!�-�$&�+! -�1'!6�()',/"�*�� �5!#'!#&0�&'�*��('/.�"�#$�+,&,!��!�(� !%)@�;$� ""#&#'$J�+,&,!��1'!6(�()',/"�#$%/,"�� �/ !��!�( -5/��'+�%#&#�(�#$�'!"�!�&'�#$%!� (��(& &#(&#% /�!�/# *#/#&0s�&)#(�1',/"� (�1�//� //'1� �(��-�$&�"� $ /0(#(�"#(&#//#$��!�+�!�$%��&05�(�+'!�%'-5 !#('$� %%'!"#$��&'�%#&0�%) ! %&�!#(&#%(�(,%)� (�(#t��'!�U:u�5�!�% 5#& @�S,!&)�!-'!�J� �!�+#$�-�$&� &�&)��-� (,!�-�$&�5!'%�((�#(�$��"�"�&'�!�",%��" & �(,*r�%&#.#&0@��

Page 151: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������� !���"#$%&'()#�*#+#�,-&&+�.//01�234)+#5&)+�6&�"#$%&'()#1�"#457�817�9�*:�;#:(7�21�<.///=1�>#$+)&$?@�A(4$�B-#@&@�(C�DE>(F&$)5&)+�;(6&'1�G&H($+�IJ�8K;E�.EL��M1�>#$+)&$�>$(4H1�8#HN&5:):1�<.//L=1�K)E':)&�2F#:'#O:':+P�(C�B4O':%�,&$F:%&@Q�R(S�:@�D4$(H&�H$(N$&@@:)NT�L+-�5&#@4$&5&)+1�8#HN&5:):�C($�*:$&%+($#+&�>&)&$#'�C($�U)C($5#+:()�,(%:&+P�#)6�;&6:#Q�D4$(H&#)�8(55:@@:()1�8#HN&5:):1�<.//�=1�V-&�W@&$�8-#''&)N&Q�"&)%-5#$X:)N�+-&�,4HH'P�(C�K)E':)&�B4O':%�,&$F:%&@1��+-�5&#@4$&5&)+1�8#HN&5:):�C($�*:$&%+($#+&�>&)&$#'�C($�U)C($5#+:()�,(%:&+P�#)6�;&6:#Q�D4$(H&#)�8(55:@@:()1�8KG*U,1�,&$F:%&7�8(554):+P�G&@&#$%-�9�*&F&'(H5&)+�U)C($5#+:()7�D4$(H&#)�8(55:@@:()1�-++HQYY%($6:@1&4$(H#1&4�<*&%&5O&$�.//0=1�*&'(:++&�9�D4$(%:+:&@�<.//M7�.//Z�#)6�.//[�&6:+:()@=1�&8:+:\&)@-:H�C($�2''Q�D4$(H&#)�"&)%-5#$X�G&H($+1�D4$(%:+:&@1�*4)'&#FP7�B17�;#$N&++@7�R17�"#@+(S7�,17�9�V:)X'&$7�]1�<.//L=1�*:N:+#'�D$#�>(F&$)#)%&1�UV�8($H($#+:()@7�+-&�,+#+&�#)6�&E>(F&$)5&)+1�K C($6�W):F&$@:+P�B$&@@1�&E8:+:\&)�B$(3&%+1�-++HQYYSSS1&E%:+:\&)1N(F14X�<]#)4#$P�.//_=1�&E>(F&$)5&)+�B$(N$&@@�:)�DW.� Q�G&#H:)N�+-&�"&)&C:+@�<.//�=1�*:$&%+($#+&�>&)&$#'�C($�U)C($5#+:()�,(%:&+P�#)6�;&6:#Q�D4$(H&#)�8(55:@@:()1�A:)#'� G&H($+� K)� V#X&� WH� (C� a(%#'� &E,&$F:%&@� .//LE/�� <.//�=1� D@6� V(('X:+1�-++HQYYSSS1&@61($N14XY&@6+(('X:+Y+#X&4H�<]#)4#$P�.//_=1��R&&X@7�G1�<.//L=1�"&)%-5#$X:)N�&>(F&$)5&)+Q�U5H$(F:)N�+-&�I#+:()#'�#)6�U)+&$)#+:()#'�;&#@4$&5&)+7� DF#'4#+:()� #)6� 8(5H#$:@()� (C� &>(F&$)5&)+1��-++HQYYSSS1(&%61($NY6#+#(&%6YZ[Y�_YM0Z/Z/�L1H6C�<]#)4#$P�.//_=�B&$C($5#)%&� ;&#@4$&5&)+� C($� +-&� >(F&$)5&)+� K)Ea:)&� U):+:#+:F&1� -++HQYYSSS1+O@E@%+1N%1%#Y@:#@YH&$C($5#)%&YH&$C($5#)%&.ZE&)N1#@H�<]#)4#$P�.//_=1�G#N:)7�81�81�<�__Z=1�8()@+$4%+:)N�,(%:#'�G&@&#$%-1�B:)&�A($N&�B$&@@1�G(6$bN4&\7�]1�G17�"#+''&7�]17�9�D@+&O#)7�*1�<.//�=1�DE>(F&$)5&)+�8:+P�;(6&'@Q�%#@&@�C$(5�D4$(H&#)�%:+:&@1�234)+#5&)+�6&�"#$%&'()#1���

Page 152: Local e-Government Bench-learning

� ������������� �� ������������ ��������������������� ������ !��"#$��%&'(�)' **!+�)',-!*&('�.� #�.&/(-�*��0 ,� !(�"#$��%+�)' **!+�)',-!*&('�.� #�.&/(-�*�1+�.'*�!,2(+�3�45�6�7.5�8('*#9 �%+��:2;!<+�)',-!*&('�.� #�.&/(-�*��