THE EFFECTS ON POLICY OF THE COMPOSITION OF THE ICT PUBLIC POLICY NETWORK IN SWAZILAND

The composition, relationships, alliances, power structures, norms and bureaucracies in policy networks affect not only the policymaking process but also the policies that result. This article reports on a study which analysed the dynamics of the ICT policymaking network in a developing country, Swaziland. The study uses a policy network analysis (PNA) approach to analyse the Swaziland national ICT policy network. The findings of the study show that government recruited mainly conformist actors into the policy network so as to meet set deadlines, and that policymaking was dominated by political agendas and strong foreign intervention, while side-lining key local policy actors.

Using the case of the Swaziland national ICT policy network, this study sought to understand the role of policy actors in an ICT policymaking exercise, making explicit their effect on the process of policy formulation.It sought to understand the dominance, the alliances, the marginalisation and the power structures that existed in this policy network and how these characteristics affected the overall policy process.This article discusses two questions: • How does the composition of a policy network affect policy processes?
• How do relationships, alliances, power structures, interests, bureaucracies and norms in policy-making bodies affect policy processes?
Swaziland was chosen both as a response to calls for more ICT policy research in Africa and because it is one of the smallest African countries, often ignored in research (Delano, 2009;Kendall, Kendall & Kah, 2006).A policy network analysis (PNA) approach was used as a theoretical underpinning for the study because of its ability to understand and identify the key characteristics of policy networks.The Swaziland national ICT policy network was established in 2006.Of the 22 policy actors in the network, 16 of them were interviewed for this study.The study offers an analysis of an often ignored source of development project failures, namely the role of policy actors.

OVERVIEW OF PERSPECTIVES ON POLICY ACTORS AND POLICYMAKING
Public policy involves a statement of intent on any socio-economic matter that outlines the goals and aspirations of a country, and a set of principles that the government and the populace need to uphold to achieve the identified goals (Cloete, Wissink & De Coning, 2006).The policies contain the intentional actions of the government, are the authoritative allocation of values for the society and provide a projected programme of goals, values and practices (Dye, 1978).Dye's (1978) definition makes it clear that government is the primary agent in public policymaking and that government has the power to make decisions on behalf of the people.This, however, does not negate the role of non-governmental participants in the formulation and implementation of public policy (Dye, 1978).Public policies are typically a result of conciliation processes and are by definition techno-political processes that define and match goals among concerned social actors (Howlett, Ramesh & Perl, 2009).
Policies are interrelated decisions taken by political actors in terms of selecting goals and the means of achieving them (Howlett, Ramesh & Perl, 2009).Therefore, the capacity of government to implement its decisions is a crucial component of public policy; government's capacity therefore also determines its policy choices.
Sometimes policymaking processes do not involve active participation of the larger population.Governments may make critical decisions and policies through closed and non-transparent processes.In some cases public participation is only symbolic (Barnes, 2006;Kendall, Kendall & Kah, 2006;Mohamed, 2006).Policymakers in the developing world often ignore the idea that effective policies can only be achieved by outlining clear and possible goals through policy improvements and by appealing to societal interests.On the other hand, the lack of participation of civil society in the development of the policies may lead to resistance to those policies (Heeks, 1999).
Which information decision makers choose and how they do so may cause gaps between process and substance in public policymaking (James & Jorgensen, 2009).Currently, developing country policymakers face an array of economic, social and political choices, due to the challenges of a complex, fast-changing and uncertain environment.Osman (2008) attributes the lack of clarity, drive and vision among policymakers to the lack of democracy and good governance in many countries.
The involvement of foreign actors in the policy process, which is often the case in developing countries, brings additional dynamics into an already complex process.In most cases the foreign actors possess power due to financial strengths and access to economic and political elites that may result in their views, or views aligned with their agenda, being privileged.This may result in acceptance of minority views at the expense of representative and democratic views (Norton-Griffiths, 2010).Such a situation creates an environment where consensus and coalition become more important than inclusive policy processes, such that policy alternatives that are deemed to be compatible with existing policies and regulations are more likely to be selected (Liu, Lindquist, Vedlitz & Vincent, 2010).

POLICY NETWORK ANALYSIS (PNA) APPROACH
Policy network analysis (PNA) offers a way to study policymaking processes, structures, outcomes and relations between policy actors.PNA emanates from examining policy networks, in ways that describe meso-level relations (between government and interest groups) and micro-level (interpersonal) relations (Marsh & Rhodes, 1992;Moran, Rein & Goodin, 2006Rhodes, 1997).Networks can be seen as clusters of actors connected to each other by resource and interest dependencies (Borzel, 1998).These then become establishments that provide actors with platforms for certain courses of action (Blom-Hansen, 1999).PNA developed as a criticism of earlier policy analysis methods, which were seen to be too instrumental and mechanistic.PNA addresses this critique by considering policies primarily as a result of the collaboration of different sets of actors.

POLICY NETWORK SCHOOLS OF THOUGHT
The two schools of thought in the field of policy network analysis are (i) the interest intermediation and (ii) the governance schools.The former sees policy networks as varying forms of relationships between interest groups and the state, while the latter sees policy networks as a way of organising political resources that are scattered in both the private and public sectors (Borzel, 1998).This study adopts the interest intermediation paradigm because it defines and understands policy networks as a relationship between the government and interest groups.
In interest intermediation, policy networks are seen as power dependency relationships between government and interest groups where resources are exchanged (Borzel, 1998).Policy networks can become structures through which entities make routine decisions in a given area of policy.In most cases, privileged groups and entities have strong relations with governments and this may result in other interests being side-lined (Moran, Rein & Goodin, 2006).
The constitution and behaviour of networks affects the policy outcomes and illustrates the power variations among the different actors (Rhodes, 1997).Interest intermediation emphasises the need for negotiation between interest groups and government actors, and is often characterised by close relationships between particular groups and the government.During such negotiation processes some ideas may become more dominant and controlling than others.These are not necessarily the government's ideas (Rhodes, 1997).Generally, the dominance of particular actors is a result of their respective resources and skills.Over time, however, participants' interests may become institutionalised in government, while interactions and relationships become routine, due to constant consultations and private rather than public lobbying (Marsh & Rhodes, 1992;Moran, Rein & Goodin, 2006.

POLICY NETWORK ACTORS
Various actors come to the policy network to pursue their own interests; and these can either be facilitated or constrained.Governments need the different actors to promote their policy objectives.Though they may deem some of the policy participants as extremist and unrealistic in their demands, others are seen as responsible and acceptable.
Governments need the actors' expertise, information and resources for political support and for implementing policies (Moran, Rein & Goodin, 2006.Policy actors can be categorised as: • Topocrats (having local autonomy) -these are governmental entities that organise policy networks to enhance their interests.They seek to protect their independence and the particular wing of government to which they belong.
• Expenditure advocates (focused on sectoral policy goals) -these seek to incorporate sector-specific goals in the political/policy system.They want to see new public programmes in place, more fiscal support for new and existing programmes, and new public sector regulation.
• Expenditure guardians (focused on macroeconomic control) -these seek to control and restrain public expenditure and activity.
Actors assume and select their positions depending on the policy issue at hand (Blom-Hansen, 1999).
There must be a clear balance and proper interest intermediation to get to the ideal policy.Excluding any one category type may result in a suboptimal policy.

POLICY NETWORK CLASSIFICATIONS
There are a number of policy network classifications.This study uses the Marsh and Rhodes (1992) typology, which is well developed and well documented since it has been successfully used to analyse a variety of problems in policy analysis (Borzel, 1998;Rhodes, 1997).This classification enables a uniform characterisation of policy processes and comparisons.It adopts the following four dimensions: memberships, interdependence/integration, resources and power.The typology acknowledges that policy networks are sometimes characterised by dominating interests.It sees policy networks as relationships between interest groups and the government; and policy networks are categorised according to the degree of closeness of the relationships between the two groups.
The typology treats policy networks as either policy communities or issue networks (Daugbjerg & Pedersen, 2004;Rhodes, 1997).Policy communities involve close relationships between policy actors.A policy community is characterised by a tight, closed, well-integrated and highly institutionalised network where membership is difficult to access.Issue networks, on the other hand, involve loose relationships between policy actors and may include government authorities, legislators, business people, academics, lobbyists, and even journalists.Issue networks normally involve policy consultations (Rhodes, 1997).Access to issue networks is generally open and the degree of integration and institutionalisation is low, which in most cases results in the instability of the network.

OPERATIONALISING PNA
The approach to adopting PNA in this study is based on the works of Marsh and Rhodes (1992).This model is meant to facilitate the understanding of the elements and relationships of PNA (Erridge & Greer, 2000) and is applied to the Swaziland case.The model, illustrated in Figure 1, shows the relationships between elements in a policy network.The structural position of a participant in the policy process relates to its standing in the society.It is dependent on the reputation of the actor and the resources that it possesses.Policy network actors may exchange the following five resources: authority, money, legitimacy, information, organisation.Access to information is vital and promotes knowledgeable and productive policy actors (Rhodes, 1988).In policy networks, actor skills are used to gain advantage in the policy interactions (Erridge & Greer, 2000;Marsh & Rhodes, 1992;Rhodes, 1988).Skilful actors can produce winning strategies aligned to their particular interests, which will in turn impact on organisational values (Rhodes, 1988).
The network structure is the actual formation and setup of the policy network.It is the team that has been assembled and tasked with the formulation of the policy.
A structure has rules and regulations (rules of the game), as well as norms and bureaucracies (Erridge & Greer, 2000;Marsh & Rhodes, 1992;Rhodes, 1988).Marsh and Rhodes' classification posits that networks are political structures that influence policy outcomes and are responsible for facilitating or constraining actors in the network.Rhodes (1988) identified rules of the game that reflect the core values of the network and may be formal (statute-based) or informal.Rules prescribe the general perception of the policy and its application and each player's role.However, the rules of the game are affected by behavioural norms, which usually regulate the behaviour of the participants in the network, eg general bureaucratic norms.
Exogenous factors are influences that are not physically present in the policy network but affect the proceedings of the network.These may take the form of foreign assistance or pressure from powerful organisations not represented in the network.In some cases they may come from public opinion (Erridge & Greer, 2000;Marsh & Rhodes, 1992;Rhodes, 1988).A network interaction is a confluence of different factors that may or may not work well together.It is a meeting place for diverse skills, resources, exogenous factors, rules and bureaucracies (Erridge & Greer, 2000;Marsh & Rhodes, 1992;Rhodes, 1988).The more diverse the actors, the more complicated the interactions may be.

CASE DESCRIPTION: SWAZILAND NATIONAL ICT POLICY NETWORK
Swaziland remains as the last absolute monarchy in Africa.In Swaziland, the head of government (the Prime Minister), ministers, some members of the House of Assembly (the lower house), two-thirds of the Senate (the upper house), senior government officials and controlling officers, and other senior administrators are appointed by the King (GoSa, 2005).This presents a rare political climate.In most democratic states, heads of state and the top echelon of government change with elections.In some democracies, this also means a change of agenda, focus, policy and strategy.In Swaziland, the head of state is installed on a permanent basis, with other parts of government changing and political heads at times rotating from one office to another.
The Swaziland National ICT Policy Network was composed of 22 actors of which three were private companies, four were non-governmental and 15 were either government departments or parastatals (see Table 1).The policy, formulated in 2006, was completed with the assistance of the government of Finland and the United Nations Commission for Africa (UNECA).The Finnish government provided financial and expert assistance for the process while UNECA provided expert support (GoSb, 2005).MTEC was the department responsible for ICTs in Swaziland at that time.

METHODOLOGY
Policy network analysis is a qualitative methodological approach.Data for this study was obtained through review of documents and interviews with 16 of the 22 members of the policy network between November and December 2011.
The members of the policy network were identified with the assistance of MTEC.The interviews sought to establish the relationships, alliances, power structures, interests, bureaucracies and norms that existed among the policy actors.Interview questions were based on the constructs of PNA, namely actors' structural position, resources, skills, as well as network structure, exogenous factors and network interaction.
Secondary sources used were the publicly available documents of the various actors, for example annual reports.
Efforts to acquire the policy network documents such as minutes and action plans were unsuccessful, since they are categorised as classified documents.Documents from actors not interviewed were also reviewed to establish their degree of conformism, structural position, skills and resources.
Memory lapses were a limitation since some policy actors could not clearly recall some of the events of the policy process.The failure to obtain some of the policy network documents was also a limitation, because the documents could have been used for triangulation purposes.

STRUCTURAL POSITION OF ACTORS
The policy network was populated mainly by government departments and units that were mostly loyal and conformist to the government, but not necessarily aligned to the societal interests (recall Table 1).Of the 22 actors in the policy network, 15 were either government departments or government-controlled institutions, therefore, most of the actors were in good standing with government in terms of reputation, authority and legitimacy.The Ministry of Public Service and Information was one of the senior government departments and hence had been involved in all government policies and strategies, while the Ministry of Public Works and Transport prided itself on being one of the mobilisers of the policy process.
The non-government actors who were recruited into the network had mainly earned their inclusion through activities that made them important to the government.AISPs, though small, were well recognised and respected for introducing the Internet in Swaziland.MTN was the only mobile telecommunication company in the country.Most of the other non-governmental participants were involved in community work such as charitable activities.Their involvement in community activities boosted their structural position in relation to government, and partly in relation to the Swazi populace.Government perceived the community work as complementary to its own activities.Some of the semi-conformist and non-conformist actors in the network had a history of challenging the government's position on issues, for example, CANGO had challenged government policy positions through press statements.However, this potential source of conflict with government was counterbalanced by its country-wide poverty reduction exercises.CANGO claimed to have garnered its influence through its hands-on involvement in the community, especially in poor communities.It claimed that it was included in the network because it was "… very popular with the masses … came at a time where there was no link with the grassroots people, so they [government] saw us as their saviour …" The reputation of the policy actors was assessed from interviews and from secondary sources, appraising their stance and views over a period of time.The relative size of some of the actors also contributed to their influence and authority in the Swazi society and in government circles.All actors commanded influence and authority in their own way.Figure 2 illustrates the reputation (influence and authority) the actors had in government and in Swazi society.

ACTOR RESOURCES
In 2005 most of the policy actors were well resourced.UNISWA's good standing was evident in their increasing student intake.As a result, "money or any resource was not at all an issue" for UNISWA.The Consumers Association and CANGO were receiving financial support from international organisations.FSE&CC was receiving funding and technical support from member companies, payments for services provided to members, and donations from international bodies.However, TIBIYO and SIDC were experiencing hardships.SIDC "[was] stagnant because our focus was on FDI and it was slowly going down".Like most institutions in the country at that time, government departments in the policy network were financially stable.
Considering the small size of Swaziland most of the actors were relatively large.Although some organisations were small in terms of staff numbers, they had large memberships.However, contrary to expectation, the unit in the MTEC responsible for ICT policy formulation was the smallest of all the actors and had only two staff serving as the secretariat of the policy network.The unit was inadequately funded and did not have a budget to remunerate the policy actors.

NETWORK STRUCTURE
All the actors, except the chairman, were recruited by MTEC.The chairman was recruited by Cabinet because it believed that he was a user and advocate of ICTs and ranked higher than other principal secretaries in terms of ICT usage and literacy.Since the various policy actors represented different sectors and groups in the population, MTEC gave them varying reasons for their recruitment.They all believed that their recruitment was, to a large extent, connected to their structural position.
Professionalism was said to be an important value that actors were required to uphold at all times.Yet, either due to memory lapses or overt weakness of the policy process, only three participants recalled being given terms of reference.Most said all that was emphasised in the first meeting was professionalism and the importance of getting the task "over and done with".
The policy process was fraught with myriad challenges, with a clear divide between government and nongovernment actors.Private and government participants each accused the other of bureaucratic approaches (see Table 2).Some actors, however, denied the existence of bureaucracy.The chairman stated that he "[…] did not see any bureaucracy, in fact, I did not care if it existed.I did not care if we formed the quorum or not -I just started meetings".Non-governmental actors felt government was controlling the policy process through its many participants and that the policy network was too large to achieve any tangible outcomes.Other accusations against the government team related to hosting meetings with no set targets.Some actors, both government and private sector, blamed the network leadership and secretariat for being inefficient and for creating a poor feedback cycle unable to keep the policy actors aware and informed of the policy process.This led to some members being frustrated and leaving the policy network.MHSW postulated that the "team contributed and left".
Remuneration was also a concern for participants from the private sector who expected financial compensation for their involvement in the process.Some lost interest due to no remuneration.Government departments criticised the private sector for thinking that their time was more valuable than that of others and for expecting remuneration for an activity conducted in the "national interest".

ACTOR SKILLS
The actors in the network possessed a wide range of skills, strategies and values relevant to the policymaking process.Most institutions sent representatives who were either IT experts or at least computer literate.However, none of them had previously been involved in an ICT policymaking process.To address its skills shortage, MTEC frequently requested technical assistance from UNECA, which "came in to assist us and even appointed someone from Zimbabwe".Furthermore, MTEC's Principal Secretary, the most senior official in a government department, "was not good technically [not computer literate] so the Cabinet had to find someone to stand in his place".In a bid to meet deadlines and to complete the policy work, the Cabinet appointed the .Principal Secretary from the Department of Housing and Urban Development, who supposedly had the relevant expertise, to chair the committee.
The skills that the actors possessed were not fully utilised, and therefore did not benefit the network much.For example, skills for mobilising, coordination and sensitising, which MPWT, CANGO and FSE&CC possessed, were not used, possibly because the policy process was weakened by the domination of the government actors.This may have influenced some actors to quit before they could contribute to the process.Only MPWT was successful in mobilising and sensitising policy actors about the importance of the policy at the beginning of the policy process.Utilising these skills extensively would have complemented the attributes of the chairman and his ability to get the actors to work together.
The strategies employed by policy actors ranged from persuasion to professionalism.The Chairman, UNISWA, the Consumers Association and MPSI believed firmly in persuasion.The Chairman felt it was his job to persuade the policy actors and to remind them that the policymaking process, and eventually the policy, would benefit them and their institutions.On the other hand, AISPs, SIDC, CSD, SBIS and MHSW were advocates of professionalism.Persuasion was probably the main strategy that kept the policy process alive, given that some participants claimed that the process was weak and lacked vibrancy.

EXOGENOUS FACTORS
The influence of external factors on the policy process was assessed by investigating whether the actors interacted directly with, or received assistance from, any institution that was not part of the process but had interests in it.Six participants fell into this category (see Table 3).There were no obvious differences among the actors in terms of the exogenous factors, except that most external influences (external to the network actors), especially on government actors, came from foreign agencies and from the government itself, namely the Minister and the King.UNECA was the most dominant exogenous actor in the policy network.UNECA provided most of the technical assistance to the secretariat and frequently checked if standards stipulated in the templates were followed in the formulation of the policy.The Chairman admitted that UNECA "came in and gave us some guidelines to follow".
The Minister (MTEC) stressed that "had it not been for UNECA, we would not have the policy we are talking about today".However, some policy actors felt that the behind-the-scenes workings of UNECA resulted not only in the adoption of cut-and-paste policy pillars adapted from other countries, but also the urgency to deliver the policy.TIBIYO postulated that "my understanding was that we had a policy from some African country which we were supposed to use to sort of speed the process up".As UNECA's hand became more visible in the proceedings some actors, especially private participants, pulled out of the process as they felt that their contributions were not necessary.
MTEC was also receiving financial and technical assistance from the Finnish government.Finnish experts occasionally visited the secretariat to verify that they were doing the right thing.The experts also expressed their expectations to the Minister (MTEC).
Like MTEC and ME, CANGO constantly dealt with donors who wanted them to support their projects on any given platform, including the national ICT policy network.CANGO's feeling was that "the donor conditions twisted our focus.It was a win-win situation, they get what they want and you get what you want".CANGO, however, had limited influence in the policy network and non-extensive involvement to effectively incorporate some of its external influences.
Regional bodies like the SADC also had some influence in the policy process.The greatest source of outside pressure for MPWT was regional bodies who from time to time supplied them with information and benchmarks which they shared with other government departments.
As a non-governmental organisation that worked with the civil society and other civic organisations, FSE&CC engaged labour unions in their preparations and consultations for the policy process.FSE&CC posited that the unions "were not involved so we had discussions with them, just to put them on the loop".
Government constantly expressed the view that Swaziland was lagging behind in terms of ICT compared to other countries.This put pressure on some policy actors to complete the policy process.More importantly, pressure came from the King's direct interest in the process and the objective for the policy to be ready that same year, 2005, in readiness for the World Summit on the Information Society in Tunis.

DISCUSSION AND CONCLUSION
Research points to the low success rate of ICT policies in developing countries and the need for further research focusing on ICT policy process in those countries.As a contribution to the discourse, this study explored the dynamics of ICT policy networks in Swaziland.By direct involvement in policymaking, policy actors exert influence on public policies and their outcomes.The relationships, alliances, powers, interests, bureaucracies and norms of the policy actors depend on the composition and dynamics of policy networks.
This study illustrates how government actors may systematically weaken the voice of other actors in a policy network, pushing their own agendas while making the process appear democratic.The Swaziland ICT policy network was skewed in favour of government and the actors were predominantly conformist.This meant that there were: 1) strong relations and alliances among government actors; 2) retention of power and control by government; 3) pursuit of government interests; and 4) dominance of government bureaucracy and norms.This allowed the government to shape the discourse .This is particularly detrimental in a context where the government has a unified voice that may arise through monarchy or one-party dominant states, or weak opposition.In Swaziland, the dominance of the government actors weakened the process and silenced alternative voices.This implies that the apparently democratic approach of including participants from diverse backgrounds in policy processes may not allow the majority voices to permeate into policy.
Agenda-setting in the policy process could also have negatively affected the process.The agenda for the ICT policy was mainly driven by exogenous factors.This could be partly due to lack of appreciation and understanding of ICTs and their impacts (Chigona, Vergeer & Metfula, 2012).The need to have an ICT policy did not emanate from within the country, rather it was born from processes outside the country such as WSIS.In the case of Swaziland, the government desired to have the policy ready to catch up with the latest benchmark.This raises the important question of how interest in and appreciation of ICTs and their impact can be aroused in the developing country context, how local debates and local ownership of the policy processes can be facilitated.
The study points to the powerful role of foreign actors in ICT policy in developing countries.The policy was mainly supported and shaped by Finnish experts and UNECA.The two exogenous actors were powerful, if invisible, participants in the policymaking process.Foreign technical support in ICT policy formulation is likely to be a welcome and valuable contribution to developing countries, particularly due to lack of local expertise in ICT policy.In this case, though the local actors possessed technical ICT knowledge and some were experienced in formulating other policies, no one had experience in formulating ICT policies.The input of the international partners was therefore invaluable.However, the involvement of the foreign actors contributed to undermining local enthusiasm in the policy process.
This study focused on the policy formulation process, addressing itself to the composition and dynamics of the policy network, not to the outcome of the process.It is recommended that future studies look at how policy network composition and dynamics affect the policy content and its implementation, as well explore factors pertinent to effective foreign support for policy in developing countries.
FIGURE 1: ELEMENTS AND RELATIONSHIPS OF POLICY NETWORKS FIGURE 2: REPUTATION (INFLUENCE/AUTHORITY) OF POLICY ACTORS TO GOVERNMENT AND SOCIETY King's interest to see the policy delivered as soon as possible.Pressure to deliver the policy in a short space of time.The sense that Swaziland is lagging behind in ICT.

TABLE 2 :
NORMS AND BUREAUCRATIC APPROACHES IN THE POLICY NETWORK

TABLE 3 :
EXOGENOUS FACTORS AND ACTORS IN THE POLICY NETWORK