Empirical Issues and Challenges for Multilevel Governance: The Case of the 2010 Vancouver Olympic Winter Games

How did a large network of over 600 actors successfully organize itself to serve a mega project dominated by three levels of government, even as control rested with a non-profit entity, included other sectors, and the governments involved did not normally work well together? The purpose of this paper is to examine how the three levels of government in Canada established a network to coordinate efforts for hosting the 2010 Vancouver Olympic Winter Games. This case study was built by means of documents and interviews, and supported by participant observations. The network was not found to be dense, but did include a multiplexity of ties (e.g., transactions, communications, collaborations, and coordinating bridges) by actors serving diverse strategic goals and scopes of work. The case was compared to data collected for the 2012 London Olympic Games to draw out key network governance coordination themes. Nine governance themes emerged associated with governance structure, processes, and evaluation: coordination mechanisms; internal engagement, momentum, and motivation; external transparency; formalization; balancing autonomy and interdependence; co-location; readiness exercises; political alignment; and time. The findings provide a framework for examining the governance of multi-level, multi-sectorial networks created to undertake a mega project and indicate how a network’s public and non-profit organizations’ activities and procedures can be influenced, modified, and impacted by the other actors (i.e., other public or non-profit organizations).


Introduction: Multi-Level, Multi-Sectoral Network Governance
How can three levels of government that do not easily trust each other work together in a mega project?Moreover, how can they be successful when the organizing network has hundreds of public, for-profit, and non-profit actors?This situation is increasingly pervasive in our globalized society.We find such instances with World Expositions, G8/G20 gatherings, and responses to international issues (e.g., major virus outbreaks, recessions, global warming), Olympic Games, and the FIFA World Cup, to name but a few.Such major instances usually involve governments as key actors, but these governments do not necessarily have ultimate decision-making control.In certain instances, national or international non-governmental bodies (e.g., United Nations, International Olympic Committee) act as the lead organizations, network administration organizations (NAO) (Provan & Kenis, 2007), or transnational advocacy networks (Price 1997).
Examining governance is important because it demonstrates governments' shift from state/hierarchy to markets/networks, with a concurrent increase in non-governmental actors' power in directing societal policy decision, as well as expanding international market and financial flows (Bevir, 2011).Governance is more than management; it comprises economic, social, political, and cultural aspects (Bellina, 2009).Yet, most studies, for better or for worse, focus on just one of these aspects, sticking to the technical, prescriptive/dogmatic level, such as corporate governance and its strong focus on board structure and composition of for-profit organizations (Bellina, 2009).Although providing valuable information, such studies have excluded important contextual and institutional aspects (e.g., cultural, social, and political elements).Current trends in networking, deverticalization, and business process outsourcing are argued to increase efficiency, and, employing the principles of good governance, are seen as "essential underpinnings for financial stability and economic growth" (Clarke & Branson, 2012, p. 1).Governance is thus dynamic, evolutionary, and multidimensional, thereby highlighting its appropriateness for examining multi-level and multi-sectoral networks.
The purpose of this paper, then, is to understand how three Canadian governments (federal, provincial, and municipal) successfully managed a network to coordinate their efforts within a larger network for hosting the 2010 Olympic Winter Games in Vancouver, British Columbia.Involving both non-governmental and governmental actors, including federal, provincial, and municipal actors and stakeholders, this case not only makes it possible to better understand the governance of large sporting events on the international scene, but, first and foremost, it enables us to better grasp the complexity of multilevel governance in Canada.The case of the 2010 Olympic Winter Games also allows better understanding of how a private international regime such as the Olympic system manages to strongly influence established governments at multiple levels (Cutler, 1999;Chappelet & Kuebler, 2008).The sheer magnitude of the Olympics, its international nature, the heavy involvement of governments, and the level of public funding all compound these issues.
Nine themes emerged surrounding the governance of a massive network centred on balancing autonomy and interdependence.The nine themes were then integrated into a framework for examining the governance of multi-level, multi-sectoral networks created to undertake a mega project and indicate how a network's public and non-profit organizations' activities and procedures can be influenced, modified, and impacted by other actors (i.e., other public or non-profit organizations).Beyond the salient issues of efficacy and efficiency, the democratic dimension is also a key concern here.The nature and extent of private influence over public institutions raises fundamental questions of transparency and accountability for all actors and/or stakeholders involved.This research will also help to better understand the empirical nature of contemporary democratic governance and its inherent reconfiguration of the porous boundaries between public and private spheres, as well as between actors within the public sphere itself.This paper is structured as follows.We first provide an overview of relevant literature to our theoretical framework.This is followed by details of the context, and then a description of the data collection and analysis techniques.The results are then presented and discussed.The paper concludes with future directions.As with all other Olympic Games organizing committees, VANOC had only a short time (less than seven years) to prepare.As such, outsourcing to various stakeholders was required.Of interest for this study, four governments assisted VANOC: the Canadian federal government, the British Columbia provincial government, and the Vancouver (host of city sports) and Whistler (host of mountain sports) municipal governments.Each government had a number of departments involved, coordinated by a set of intra-and inter-governmental committees.The governments signed a Multi-Party Agreement (MPA) with VANOC to outline each partner's responsibilities.The IOC now requires candidate cities to generate similar agreements, demonstrating the 2010 Winter Games' approach to be a best practice for organizing committee-government relationships.

Research Context
Each government had its own goals for the Games.For example, the federal government focused on a whole-of-government approach (horizontal management) to pursue pan-Canadian objectives such as athletic excellence, spreading the spirit of the Games across Canada, and indigenous people's participation (Government of Canada, 2010).The provincial government was focused on province-wide commitments to economic development, sport development, and cultural legacies throughout the province (British Columbia Olympic and Paralympic Winter Games Secretariat, 2007).Both municipalities had not only legacies as objectives, but focused on being effective and strong partners, service providers, regulators, and venue developers (Office of Olympic Operations, 2006;Resort Municipality of Whistler, 2007).
Each of the four governments created a Games secretariat or office to coordinate their respective departments.A cross-government and cross-partner governance structure (including other Games partners, namely Canadian Olympic and Paralympic committees and Four Host First Nations Society) was also created via the setup of various committees (e.g., MPA Partners' coordinating committee, Olympic and Paralympic Transportation Team).VANOC also created committees on which the partners and other stakeholders sat (e.g., master planning teams) (2010 Olympic andParalympic Winter Games Federal Secretariat, 2008).Thus, if we take the governments as the central focus or egos of the coordination network created for the planning and hosting of the 2010 Games, we find multiple levels (from local to international) and multiple sectors (public, non-profit, private) that needed to be effectively governed.

Theoretical Framework
The theoretical framework for this study is composed of (democratic) governance, network theory/analysis, and sports event management concepts.Each is described below.

Governance and Democratic Governance
Governance analysis typically deals with a perceived shift from hierarchy and bureaucracy to markets and networks-that is, from a state-centric (government-led) phenomenon to a society-centric (interdependent actors) one.If governance often appears to be an umbrella notion, covering many meanings (multi-level governance, shared governance, good governance, corporate governance), as well as numerous elements (networks, institutions, communities, processes) (Klijn & Skelcher, 2007), the same does not hold true for democratic governance.Most public administration literature defines it along the twin lines of performance and accountability, with a complementary (and varying) emphasis on democratic participation (e.g., Callahan, 2007).
In accordance to the perceived shift from hierarchy to networks, recent studies in public administration argue that state-centric approaches need to be replaced by more societycentric approaches in order to better reflect and understand developments in the policy process.Encouraging citizens to be more active and engaged in policy processes can, in turn, increase transparency and accountability of public organizations.In this perspective, the rise of inter-dependent actors in the private sector, civil society, and other governments and legislatures cannot help but diminish the state's autonomy and capacity.In this scenario of relative marginalization of the state, performance and accountability issues become increasingly complex, but remain as crucial as ever to ensure the democratic nature of politico-administrative regimes (Bevir, 2006).Yet, the passage from state-centric to society-centric approaches only make sense inasmuch as the state/civil society dichotomy holds true.Arguably, a more satisfactory approach to democratic governance will question this dichotomy and the traditional assumptions made between performance, accountability, and transparency.Such an approach, emphasizing patterns or configurations of stakeholders' governing capacities, appears to be tailor-made for the dynamic and complex nature of these multidirectional interactions, as well as the often-neglected resilient nature of hierarchy (Damgaard, 2006).As Bevir (2011) notes, "Practices of governance characteristically blur the boundaries between public and private, blending features of state, market and community; and they blur the boundaries between levels of government and between states, forging multijurisdictional and transnational patterns" (p.16).
Regarding governments and their policy networks, Enroth (2011) suggests that policy network research should see governance as territorially and institutionally unbound.Issues of interdependence, coordination, pluralism, accountability, participation, democracy, and autonomy therefore come into play.As such, there are links between governance and network literatures, as well as with issues found in organization theory (e.g., interdependence/coordination) (Lawrence & Lorsch, 1967).Citizens, civil society, private organizations, and governments at all levels are partners operating in a network, exchanging and redistributing tasks and responsibilities.

Network Theory and Analysis
While the municipal, provincial, and federal governments have certain similar needs, they also differ in interests and scope, yet they must work closely together (Parent, 2008).These structural and jurisdictional diversities, coupled with the varying interrelationships within the government stakeholder group, makes this grouping a good starting point to understand VANOC's stakeholder environment using network theory.Stakeholder (e.g., Rowley, 1997), public administration (e.g., Berry et al., 2004), and sport management (e.g., Quatman & Chelladurai, 2008) researchers call for a network analysis approach to understand the complex environment of an organization.Network theory (e.g., Gummesson, 2006;Iacobucci, 2008;Jurian & Erik-Hans, 2006) therefore helps move beyond dyadic ties to understand in greater depth one key stakeholder groupgovernment-and the network it built to coordinate efforts for the Olympic Games mega project.
A network is a set of organizations connected to each other in varying degrees based on prior contact and exchange.A network approach allows for multiple levels of analysis, where a description of both the node (e.g., government committee) and the relationship between nodes is possible and desired (Iacobucci, 2008).This is in perfect concordance with the present study as "governance refers to theories and issues of social coordination and the nature of all patterns of rule" (Bevir, 2011, p. 1).Networks can be characterized in terms of their density, hierarchy/centrality, types of relationships, quality of relationships, and multiplexity (Burt, 2000;Knoke & Yang, 2008;Provan, Veazie, Teufel-Shone, & Huddleston, 2004).For example, relationships between network actors-the ties-can be characterized as strong or weak and of different types, such as transaction, communication, boundary penetration, instrumental, sentiment, authority/power, and kinship/descent (Knoke & Yang, 2008).

Sports Event Management
Understanding the complex nature of preparing and hosting a major sports event like the Olympic Games is only beginning.Parent (2008) indicates that sports event organizing committees evolve over three operational modes: planning (bid phase, business plan, operational plan, divisional work packages); implementation (venue plans, Games-time); and wrap-up (final report, legacy management).She also identifies thirteen issue categories that an organizing committee must manage, such as politics, visibility, financial, organizing, relationships, human resources, interdependence, and participation.Theodoraki (2007) suggests that Olympic Games organizing committees are a hybrid of divisionalized form and missionary organizations (also see Mintzberg, 1979).Organizing committees move from work process standardization during the planning mode to output and skills standardization during implementation.Research in sporting events has generally focused on specific issues such as marketing, branding/image/sponsorship, tourism, economic impact, and socio-political and urban regeneration issues (e.g., Burbank, Andranovich, & Heying, 2001;Daniels & Norman, 2003;Séguin, Richelieu, & O'Reilly, 2008).This study moves the understanding of sports event management from an issue-specific focus to an overarching issue, that of governance (mechanisms and processes), and from general stakeholder identification to broader stakeholder network characterization.

Methods
Case studies have proven to be especially valuable for providing in-depth knowledge of complex events as they unfold over time, particularly for under-examined issues and when researchers seek to develop new theoretical models (Yin, 2009).Case studies "emphasize the rich, real-world context" (Eisenhardt & Graebner, 2007, p. 25) of the phenomenon under study, thus making such an approach ideal for studying VANOC and its stakeholders (particularly, the municipal, provincial, and federal governments).

Sampling and Data Collection
Based on discussions with the study's government partner, the federal government, and the sheer complexity of the situation (over 312,000 civil servants spread across roughly one hundred departments in four governments), we focused our sampling unit attention on the inter-departmental committees and issues clusters that were created as the main coordination mechanisms due to the complex, "higher-level system within which lowerlevel entities comprise the actors" (i.e., governments and civil servants) (Knoke & Yang, 2008, p. 10).
Data collection stemmed from two sources: archival material and interviews.For this paper, we focused our attention on materials specific to government.This included committees' terms of reference, meeting minutes, annual and final reports, and other Games stakeholders' documents that mentioned government, such as organizing committee internal documents (e.g., meeting minutes, memos, letters, final reports), stakeholder documents (e.g., annual reports, newspaper, and other media articles), and commemorative material (e.g., commemorative books).These documents provided information on mandates/objectives, issues dealt with, frequency of meetings, and representative departments on each committee.These were also necessary for identifying potential interviewees.Some documents were publicly available, but most governancerelated government documents (e.g., committee meeting minutes, terms of reference) were obtained through a Freedom of Information Act request.In total, 4,915 documents were analyzed.Interviewees were chosen on the basis of their high degree of knowledge of/involvement with the 2010 Games.This allowed us to obtain "unbiased data about long-term repeated patterns [and] to produce consensus answers to questions, which indicates greater validity" (Knoke & Yang, 2008, pp. 37-38).Snowball sampling was used when participants or the data indicated a need for more information to be gathered (i.e., until saturation was reached) (Miles & Huberman, 1994).We conducted 55 semi-structured interviews with 45 individuals after gaining our ethics certificate.Thirty-five interviews were conducted pre-Games (2008)(2009) with civil servants from all four governments to get a sense of the coordination network and its operations.

Data Analysis
Archival material and interviews from the 2010 Vancouver Games were combined for analysis.First, network analysis was undertaken.To do so, all Vancouver documents and interviews were inductively coded to identify all relationships.Each relationship was then described in an Excel table.The types of ties (Knoke & Yang, 2008) were used as a starting point, although emerging types were also found and considered.In all, 610 actors and 3,900 ties were identified and then transposed into an Excel matrix.Due to the large number of actors, the traditional 1-0 matrix could not be created for the whole network, as Excel does not allow sufficient columns.Such a matrix could only be done for the municipal and provincial levels.For the whole network, a basic matrix identifying relationships was created.Degree (number of ties), betweenness (ability to control information), and eigenvector (degree of importance) normalized centrality measures and were computed using the UCINET 6.0 network analysis software (Borgatti, Everett, & Freeman, 2002), as was density and in-and out-reach (see Bonacich, 1972;Wasserman & Faust, 1994).For the eigenvector measure to be robust, Borgatti, Everett, & Freeman (2002) recommend that the ratio of the highest eigenvalue to the next highest must be at least 1.5; our eigenvalue ratio was calculated to be 1.865, which was therefore robust.The NetDraw 2 (Borgatti, 2002) network analysis software program was used to visually represent the network using the software's algorithm for the eigenvector centrality measurement.
The interviews were then inductively coded using a constant comparison technique (Glaser & Strauss, 1967) for recurring patterns (themes, ideas, concepts) associated with the governments' network and their efforts to plan, implement, and coordinate their respective responsibilities.We engaged in first-and second-order coding to lead us to higher-order themes (Corley & Gioia, 2004).The first order coding included, for example, a variety of issues to handle: strategies, use of agreements, interdependence, performance, accountability, transparency, participation, hierarchical versus business approaches to the project, autonomy, and logistical/operational aspects.As noted earlier, an opportunity occurred to gather data from the London 2012 Games.We reviewed the London 2012 data for the same types of patterns.Those that could be found in both were noted and aggregated into categories based on relationships between concepts (e.g., freedom of information requests and engaging/gaining support from the community through being transparent/answering questions were aggregated into the external transparency category).Nine themes emerged as being consistently regarded as critical by the interviewees for the governance of such mega projects.Relationships between the categories were then determined, which resulted in our higher-order themes (see Figure 3 below).

Results
I think complexity is the right word.I think it's not complicated, but it's complex.I think that when you're talking about government involvement, specifically this type of concept, and even the evolution of what I would call this concept of partnership (P43).
The above characterizes the complex network that the host governments created to help plan and execute the Games, and the need to understand how such networks work for government operations.We divide the results section into our network analyses of the government's coordination network followed by our eight emerging themes for multilevel, multi-sectoral mega project network governance.

Network Analysis
Table 1 provides details of the number of actors for each government ego-network, as well as the entire network of 610 actors.Figure 1 provides a visual illustration of each government's network, analyzed using the eigenvector measure and strength of ties (the bigger the node and label, the higher the eigenvector value; likewise, tie thickness indicates strength).Table 1 also indicates the centrality analyses (betweenness, eigenvector, in-and out-reach).We found that, unsurprisingly, both the network and each government-level network was not dense given the high number of actors.Table 1 also demonstrates that the key actors for controlling information (betweenness centrality), for being generally important (eigenvector centrality), and for reaching out to other actors or being on the receiving end of requests (in-and out-reach) include, most notably, the Games organizing committee (VANOC), followed by federal-level departments and coordinating committees (Games secretariat and others), and other governments' secretariats and committees created for coordinating efforts related to the Games.Thus, the coordinating structures created by the various levels of government to help meet their responsibilities for hosting the Games were effective.The dominance of federal-level actors in the whole network is also noted and will be discussed in the next section.The whole network is depicted in Figure 2, analyzed in the same manner as Figure 1, but excluding actors with only one tie to make the network a little easier to read.We also analyzed each of the 3,900 ties found in the data.Although we started with Knoke and Yang's (2008) types of ties, we could only use certain categories, while other categories also emerged.We therefore suggest a new classification of types of ties for mega project coordination networks: • Transactional: an exchange of resources (e.g., a one-time grant from a government department to a non-profit organization); • Communication: a sharing of information (oral or written, through face-to-face, email, mail, or tele/video-conferencing, such as the Whistler Games secretariat conducting a presentation to residents); • Instrumental communication: pre-transaction, informal, and exploratory, but purposeful (e.g., Industry Canada exploring potential opportunities with Bell Canada, a corporate sponsor of the Games); • Regulatory: a form of normative power (e.g., the provincial secretariat having an overseeing role on VANOC regarding its spending); • Legal: network actors that are legally bound (e.g., the MPA signed by all governments, as the partners and "owners" of VANOC, and the NAO they created to oversee the preparations and coordination of the work for the Games); • Collaboration: a formal or informal partnership without legal attributes as identified by actors (e.g., the province's 2010 Speakers' Bureau collaborating with VANOC to conduct a speakers' series); • Coordinating bridge: committee members of internal-internal (internal to a department), internal-external (internal to a government but between departments), and external-external (between governments or with other actors) links, such as a federal department acting as chair of a cross-government Games committee; and • Generic internal or external link: a link found in the data but for which no additional information was found in order to classify it.
The two dominant types of ties were collaboration, preferred by most, and coordinating bridge, preferred by the federal government.Table 2 presents the frequencies for each type of tie.Collaboration was also preferred for an event-level and inter-government relationship.

Governance Themes
Nine governance themes emerged associated with governance structure, processes, and evaluation: coordination mechanisms; internal engagement, momentum, and motivation; external transparency; formalization; balancing autonomy and interdependence; colocation; readiness exercises; political alignment; and time.Each is described below.
These themes are not mutually exclusive but interrelated factors.The interrelationships, noted in the descriptions below, are illustrated in Figure 3. Coordination mechanisms.Supporting the network analysis, we found that coordination mechanisms/frameworks were important and needed to be established during the bid phase (i.e., before a city had won the right to host the event), which is to say, before planning/preparing for the Games commenced.Despite being created just after being awarding the Games, one federal interviewee noted that secretariats should have been created even earlier: "So in terms of coordination, I think that staffing the Federal Secretariat here was excellent.I think that should have been done earlier, and maybe that would have had more of an impact" (P16).These mechanisms/frameworks included forums/committees allowing for stakeholder/interest group participation and secretariat-like structures.Such mechanisms should therefore be established at strategic, intermediary, and operational levels.Such a structure makes decision making for mega, time-limited projects more efficient, as Participant 25 noted: So because it's a big project and because it's also a very time-limited project, we need the capacity to make decisions very quickly.I think this sort of committee structure helps us to do that, in some ways, because we can get decisions quite quickly when needed.know, the actions that we would have to take, for example the actions that Parks might need to take would not necessarily be so focused (P30).
Participant 15 supported this and compared the success of the coordination mechanisms to difficulties experienced during a G8 summit, which was linked to the degree of engagement (a whole-of-government approach) of the various departments: I think it's really good in the sense that I have previous experience at the G8 Summit in Kananaskis.… And what was lacking there was a whole-ofgovernment approach in the sense that departments were really left to themselves to go get money, resources, policy decisions.
Associated with engagement are momentum, moving files forward, and motivation.
Representative turnover hindered momentum, as the following comment highlights: It's a long period of time.… So often there was movement within the departments on the work.… They had no idea what their department was doing.They had just started.So there was a period of time we felt like we were constantly updating everybody on what we were doing (P13).
Department motivation was also part of this process: I think the biggest challenge for government is finding a way to motivate government departments to take action.I think the committees are great for information sharing, but they don't always result in anything other than really long meetings, right, with minutes (P16).
External transparency.While engagement, momentum, and motivation seem to be an internal process, transparency was used as an external process to gain support from the general public and media, and to help with these stakeholders' degree of engagement.But we get letters from the members of the public.We have regular monthly question times with the mayor where the mayor is questioned in public by the members of the London Assembly.… So, we're always answering questions of one kind or another, either directly or in supporting the mayor in terms of kind of accounting to the public (P46).
On the one hand, transparency can be a facilitator of helping to garner trust and decisionmaking acceptance, as the following participants pointed out: "[T]o do all that we do in a way that is transparent, that we can openly defend" (P1); and "I think transparency, absolutely.I guess it links to trust as well, and transparency about how you do your business, who you do your business with" (P51).On the other hand, transparency processes, such as freedom of information laws, can be perceived as hindering momentum: Freedom of Information.That's been a big, big issue because the Ministry in particular is like, no, no, we're not going to send you that file.Come to the meeting, you see the presentation, but because if we get a copy then it is Freedom of Information within the City.… VANOC tried to come up with some kind of rules of engagement that would protect government partners, maybe, from having certain information shared, but it would have been way more work and it's just been frustrating (P32).
Formalization.To be successful, the network needed to formalize relationships and responsibilities, again before preparations began (i.e., during the bid phase).A key document was the MPA, which did just this: [The MPA] talks about obligations, benefits.It talks about who's contributing what.It talks about who's receiving what.And so we all have commitments to each other in the development and execution of the Games.… Ultimately the MPA fits in because that's where the starting obligation is, from the MPA (P39).
Such formalized documents provided clear guidelines and helped reduce potential conflicts: "And we're putting that all down very clearly in black and white so there's no ambiguity between us and the other stakeholders" (P49).
Balancing autonomy and interdependence.As the earlier observation about the G8 stated, while departments and governments are technically autonomous, they are interdependent in such time-limited mega projects.This is particularly important for managing cross-sectoral relationships, such as between an Olympic Games Organizing Committee (OCOG) like VANOC or LOCOG for the London 2012 Games (independent, non-profit organizations) and government partners.Participant 52 supported this assertion: [T]he IOC requires there to be a sort of independent organizing committee out of LOCOG who weren't a direct public body.So that's quite an uncomfortable place for government to be in sometimes in that they are very much dependent and reliant upon a body over which they've got influence, but not direct control.And yet if the OCOG is not performing that has massive implications how the country is viewed internationally.
Thus, coordinating mechanisms and formalizing relationships and procedures helped balance organizational autonomy and interdependence.
Co-location.Co-location or physical proximity in the same building was identified not only for the Vancouver 2010 Games-Vancouver is three time zones away from Ottawa, Canada's capital-but also for London, and was noted by Participant 53 as an issue in past Games: One of the things I'm very glad we did do was to put these two organizations [LOCOG & Olympic Delivery Authority] in the same building, which sounds like a trivial, superficial thing.But actually, the difference being that when there was a problem, that you could walk to someone else's office and rather than sending them an email, a letter or demanding a meeting you could just walk over to their office and start shouting at them.It helped to bring those two organizations together and to make them work as a team in some ways.

Readiness exercises.
To ensure that the coordination mechanisms, lines of communication, and actors' responsibilities are ready, readiness exercises or test events were conducted.This was noted by Participant 14: Like the provincial government, we work with them on the emergency side of things to try to make sure things are running smoothly on the communication provincially.And from a municipal perspective too, a lot of the municipal governments that are involved, we have interactions with them, especially through the exercises themselves and all of the meetings we have for coordination of the exercises.
Such exercises served as evaluations for the degree of preparedness by the actors and the network (e.g., testing lines of communication between actors in the network).Any issues found through the exercises resulted in modifications to the coordination network and its processes.
Political alignment.Political alignment, continuity, or unity helped support the overall goal of the network, ensuring a more effective and efficient process.The following remark from Participant 49 illustrates the point well: It [the Olympic Board] still embraces all of the political parties.So, I have to say from the top end, so that's done very well.… There was continuity and support.I have to say there's been great support for this project.I mean, one of the things we worked out was that whatever happened in terms of the middle years of this project, which could have been the most difficult, we realized that whatever government was in power in 2012, this was going to be their party, their event.
Political alignment, then, was a contextual factor affecting the overall process.
Time.The unmovable deadline of the Games' opening ceremonies helped create momentum and push actors to work together towards a common goal: "I didn't think there was any doubt that working to an absolutely immovable fixed date creates a momentum and the sense of common purpose to get things sorted out" (P47).However, time was "a double-edged sword" (P15).A lack of time could result a crisis response: "The more they can delay making a decision, the more time is against us because at some point, it becomes a crisis" (P17).

Network Implications
The network analysis undertaken demonstrated the complexity of hosting a mega project like the Olympic Games.The presence of multiple types of ties and aspects to govern the network (e.g., coordination mechanism, co-location, readiness exercises) highlights the need for "flexibility and responsiveness for service provision" (Bogason & Musso, 2006, p. 3).In examining the types of ties, we argue that the choice of tie seems to depend on the actor's goal(s) within the network.For example, the Canadian federal government focused on a whole-of-government approach, which was broader and more strategic, whereas the municipalities focused on operational aspects (e.g., being effective service providers and venue developers) and the partnerships.As well, we found a possible link between the type of tie choice and the perception of participation.Whereas the federal government focused on horizontal management (i.e., public service participation), municipalities focused on their residents (i.e., public participation).These reasons may explain the differing emphasis on collaboration versus coordinating bridge ties.Moreover, we contribute to the literature by offering different types of ties for our large network context.There were logically no sentiment or kinship types of ties (cf., Knoke & Yang, 2008).Boundary penetration relations were observed as coordinating bridges and authority/power relations were modified to regulatory relations to highlight the normative forms of power present.
Despite different organizational objectives, the governments-or, at least, the provincial and municipal governments-had very similar structures, in that the strongest link was between VANOC and the government's coordinating secretariat.Each of these two actors could be seen as brokers for their respective network cliques (VANOC and its event network, the secretariat and its government's actors).This placed these two actors in powerful positions.Thus, the network coordination structures of these governments seemed to follow similar structural characteristics as indicated by past research with other industries (cf., Burt, 2000).Looking at the whole network, we found that VANOC dominated as expected, but so, too, did federal-level actors.This would suggest that although multi-level governance structures can assist local/regional and supra-national actors in gaining prominence (Zürn, Wälti, & Enderlein, 2010)-the implication being that the importance of national governments decreases-our findings do not support this statement outside the policy development network area, as most research on multi-level governance has indicated.In the case of multi-level, multi-sectoral mega-projects, we found that the national-level government dominated the network along with the NAO.

Governance Implications
To govern a large multi-level, multi-sectoral network, our findings indicated a need for coordination mechanisms that helped to balance actors' autonomy and interdependence (cf., Lawrence & Lorsch, 1967), internal and external support of the network through engagement/motivation/momentum (cf., Latham, 2012), mutual transparency (cf., Grigorescu, 2003) to build trust (cf.Lane & Bachmann, 1998) between actors, formalizing (cf.Hall, Haas, & Johnson, 1967) the coordination mechanisms and co-locating the actors to reduce potential conflicts, and readiness exercises to test the network's effectiveness and efficiency.These processes were impacted by the degree of political alignment throughout the project's lifecycle and that unmovable deadline: time.As such, proper governance of large networks involving multi-level and multi-sectorial actors requires a combination of strategic management, organization theory, organizational behaviour, and public and international management concepts.This suggests that examining such networks requires a more holistic, integrative view of the situation.While examining a particular aspect of the network's governance (e.g., structure or process or evaluation) is appropriate and needed, researchers must be aware of where their study fits within the larger framework of theoretical concepts impacting such networks.
With G8, world summits, and other global gatherings seemingly increasing in frequency, it becomes important to consider multi-level and multi-sectoral networks and how they are coordinated and managed.The themes presented in this study provide a basic framework for examining the governance of multi-level, multi-sectoral networks created to undertake a mega project.Within the coordinating mechanisms, actors have at their disposal a variety of ties (chief among them collaboration and coordinating bridges), as found in this study.Moreover, this framework indicates how a network's public and nonprofit organizations' activities procedures can be influenced, modified, and impacted by the other actors (other public or non-profit organizations).
The network's high number of actors and strong multiplexity also create significant challenges for accountability and performance management, both key parts of democratic governance (Callahan, 2007).Accountability rests on clear roles and responsibilities, clear expectations, and the possibility to know and assess who has done what, how, and when.Formalization and coordination mechanisms, amongst other themes, become critical to help foster accountability.At their core, they need to make it clear to all stakeholders who is accountable to whom and for what.However, the sheer size of the network makes it very difficult to not only build but maintain successful network-level accountability, especially regarding evaluation.In fact, the joint federally and provincially funded independent evaluation report commissioned with PricewaterhouseCoopers was cancelled due to funding issues post-Games and before the end of the contract (i.e., before the longer-term impacts could be evaluated).Moreover, network-level accountability may not be desirable.Trying to be accountable to everyone in the network may result in dilution of accountability.Participant 52 acknowledged this point: "I mean in some ways you wouldn't want everybody being accountable because there's a danger that no one's accountable."Likewise, the complex nature of the network (a high number of actors and tie multiplexity) makes performance management rather elusive.What part of the performance is to be managed (e.g., all linkages, all actors), who should manage the performance (i.e., the NAO, the rights holder) and how?Performance and accountability can perhaps be determined for individual actors, but aggregating this information to a 600-plus actor network becomes difficult, if not impossible.Furthermore, nothing suggests that organizational performance and accountability is a simple or complex aggregation of individual performance.Even the concept of transparency, a pillar of democratic governance, becomes difficult.However, our study demonstrates that externally focused transparency, as elusive as it may be, remains an important issue for gaining public support.In the end, the blurring of boundaries affects not only the dynamic between (international) private and (federal, provincial, local) public organizations, but clearly between actors within the public sphere itself.Inasmuch as accountability is a democratic requirement, and that it can only begin with transparency, it appears that network governance comes with its own issues and problems, not only in terms of efficacy and efficiency (managerial concerns), but, most prominently, in terms of democratic imperatives.More theoretically, our research questions the often-assumed complementary nature of network and democratic governance, and leads us to wonder if they are not, in fact, mutually exclusive, or trapped in a zero-sum game.

Conclusions
In summary, we sought to examine how the three levels of government in Canada established a network of over 600 actors to coordinate efforts for hosting a mega project, the 2010 Vancouver Olympic Winter Games.They did so by using a variety of types of ties, building coordinating mechanisms, including using a NAO, formalization, and readiness exercises.These were assisted by internal engagement, motivation, and momentum, as well as external transparency and co-location, in order to balance the autonomy and interdependence of the network's various actors.Political alignment and time formed a contextual background to these processes.
This study, of course, has limitations, which are in fact opportunities for future research.First, the Vancouver and London settings are relatively similar, which aided comparison.However, transferability of information to non-G8, developing countries and BRIC (Brazil, Russia, India, China) countries cannot be assumed.What would be the impact of these countries' political context on multi-level, multi-sectoral mega-project network coordination?Second, the large number of actors posed limits on our ability to analyze the network, even if we limited inquiry to committee, department, and organizationallevel actors, thereby excluding the individual level.We did find that we were able to garner the same general findings from a more restricted network (i.e., London 2012), which would allow future research to eschew such levels of detail.Space limitations also prevented additional analyses of smaller government networks.Future research could therefore consider one government level only and go into greater analytical detail, linking findings with network administration structures, multi-level governance (Piattoni, 2009), and quasi-federalism (Romney, 1999) to foster further discussion on this topic.Third, we presented nine key themes that emerged from the data, and noted their interrelationships.Testing the interrelationships would be required, as would further examination of other possible themes in a variety of other large network and/or megaproject settings.Finally, our findings lead us to ask: With such large networks, can we realistically speak of network or democratic governance principles?Re-conceptualizing "network-level" accountability, performance, participation, and transparency seem to be required.The blurring of boundaries does not only affect the dynamic between (international) private and (federal, provincial, local) public organizations, but also between actors within the public sphere itself.Are network governance and democratic governance antithetical to one another?Or is it more a function of size and complexity

Revue Gouvernance
Volume 15, numéro 2, 2018 23 than a consequence of the network form?We hope to see future research examine these questions.

On 2
July 2003, the International Olympic Committee (IOC) announced that the City of Vancouver would host the 2010 Olympic Winter Games.Created on 30 September 2003, the Vancouver Organizing Committee for the 2010 Olympic and Paralympic Winter Games (VANOC) had eight main divisions working towards hosting the Games.It was dissolved on 31 December 2010.The 2010 Olympic Winter Games took place 12-28February 2010 and included over 5,500 athletes and team officials from 82 nations/states, 10,000 media representatives, 25,000 paid staff and volunteers, and 3.5 billion television viewers (International Olympic Committee, 2010).VANOC's operating budget was CAN$1.884 billion and the Games' capital budget, to which the federal and provincial governments contributed most of the funding, was CAN$603 million (The Vancouver Organizing Committee for the 2010 Olympic and Paralympic Winter Games, 2010).

Figure
Figure 1.Individual Government Ego Networks

Figure
Figure 3. Interrelationship of Factors Associated with the Governance of a Multi-Level, Multi-Sectoral Mega-Project

Internal engagement, momentum, and motivation.
The success of the relationships created through the coordination mechanisms depended largely on the individuals representing the organizations, as one federal government employee noted: "The relationships are what made the coordination work, and especially at VANOC.That's what I found" (P7).Thus, individual and departmental engagement through leadership is essential.I mean they're [coordination mechanisms] adequate at say the degree to which they do what we need them to do depends on the leadership and the leadership certainly now is very much focused on we need to do what we need to do corporately to make this a successful game.If that leadership wasn't there, you