The appearance of a deliberative model based on the relationship between public and private actors who collaborate to cope with common problems is not recent (Kooiman, 1993). New governing approaches relying on consensus and participation of stakeholders, in fact, have been internationally proposed since the late 1980s as possible solutions to remedy some criticalities shown by public   sector     reforms     inspired     by     New    Public Management (NPM) (Walsh, 1995). In this regard, public governance theories were aimed at attenuating some negative effects of the modernization of the public sector, by recognizing the importance of external orientation and balancing powers between insiders and outsiders in public management (Kickert, 1997). The diffusion of such a view was sped up in the 1990s by the increasing federalist and decentralization trends.  Since the first studies the network was seen as the most effective structure to represent the new scenario, as well as being strongly in line with the concept of governance, so as being defined as a 'governance network' when concerned with the articulation, resolution and realization of public values in society (Klijn and Skelcher, 2007). Therefore, governance networks refer to networks of actors, interdependent but autonomous, developed to collectively debate, decide and carry out public policies (Klijn and Koppenjan, 2000).

While first generation research was incredibly effective at justifying the rise of a governance network as a paradigm in line with new societal trends, new generation research has been advocated to fill the gap related to the development of operational tools and rules for improving network management. Multi-level relationships and network contextualization, under geographical and sectorial points of view, are also commonly seen as unresolved issues to be addressed to facilitate collective decision-making (Torfing, 2005).

This paper contributes to understanding development and functioning of governance networks within well defined space-time contexts, as well as the strategies and tools which permit public sector, at different levels, to exercise meta-governance prerogatives in order to collect contributions and steer actors' behaviour towards common goals (Sørensen and Torfing, 2009). The first aim is to highlight features of governance networks in tourism so as to deduce general principles from a significant case study. In particular, the empirical investigation focuses on network performance, evaluated under the dimensions of the integration among participants and the network managerial advancement. These dimensions are considered as basic conditions to improve outcomes towards both community and individual stakeholders and achieve network goals. The choice of tourism is due to the fact that this is a strictly local phenomenon and network approach, in destination research, has a very long tradition (Scott et al., 2008). Then to the nature of 'public good' which affects tourism, which implies the primary role of public sector in territory development. Furthermore in Italy, from 2001, a reform inspired by public governance theories and aimed to enhance tourism sector, is going on.

The second aim of this paper is to analyse the evolution of tourist networks, under the perspectives of integration and managerialization, through the comparison between performances measured at two moments in time over a three-year period. In particular, a personal interview approach is used to measure integration and managerialization dimensions after three years from the creation (first time interval: 2008-2011), and two and half years from the renewal of the partnership (second time interval: 2012-2014). Development trends are interpreted in light of external variables and in a specific way of the recent socioeconomic crisis, which have severely affected Eurozone   countries   since   2009,   and   even   harsher Mediterranean regions.

In 2013, except for France, Malta and Slovenia, all European countries overlooking Mediterranean Sea reported decreasing GDP rates, with Italy ranked in the third worst position (-1.9%) after Greece (-3.9%) and Cyprus (-5.4%). Employment crisis was also massive, as in December 2013 Italy recorded an unemployment rate of 12.7%, preceded only by Greece (27.4%), Spain (25.6%), Croatia (17.2%), Cyprus (16.5%), Portugal (15.2%), Slovakia (14%) and Bulgaria (12.8%) (Eurostat, 2014). As regards the region of Sardinia, field of this investigation, after a slight recovery in 2010 and 2011, all conjunctural indicators, such as production, turnovers and firm investments, dropped off dramatically in 2012, for a total decrease of GDP equal to 5.9% in the two-year period 2012-2013. Employment rate, in particular, reached in 2013 the worst value of the last 10 years (-7.3%), well beyond the National average (Bank of Italy 2014). In North Sardinia, also the quality of life, as shown by the deprivation index, dropped off in the wake of crisis, as well as the tourist flows and presences.  

In such contexts, more than elsewhere, it seems to be essential to intensify relationships between public sector, private sector and citizens in order to address social capital loss, social fragmentation and distrust of institutions.

The deep changes, from an organizational, accounting and management point of view, which had taken place in the public sector of Western countries, inspired by NPM and carried out through different steps and ways, gradually led to revise the role of public subject inside society and its relationship with community (Loffler, 2003). Criticisms about the short-sighted reproduction of business-like principles in the public sector and the excessive focus of reform on the operational side and the objectives of efficiency and effectiveness of public organizations, provoked an intense debate about legitimacy of political power in democracy and the actual value of NPM approach (Walsh, 1995). Public governance theories, since the 1990s, followed the concepts of decentralization and devolution which, in this period, gained increasing consensus.  Furthermore, they were based on a new awareness of the indefiniteness of the borders among public, private and non-profit sectors, and the necessity to better coordinate relationships between socioeconomic and supranational bodies (Rhodes, 1996).

Gradually, openness towards the external environment, social approval and participation have become essential elements for public government and developing network relationships a concrete way to reduce the presence and the weight of Western governments in economy (Denhardt and Denhardt, 2000).

Network configuration, since the beginning, has been adopted by different branches of research into the public sector. It reflects the dynamism, complexity and fragmentation that distinguish public policy issues (Marin and Mayentz, 1991). Networks are societal organizations halfway between the market and the State, characterized by a nexus of relationships between public and private actors who collaborate to achieve common goals on complex settings, and based on coordination and co-evolution (Jackson and Stainsby, 2000). In particular, governance networks are horizontal structures comprising interdependent but operationally autonomous actors that interact through negotiation. A common regulatory, normative, and cognitive framework is required to enable network self-regulation and propensity to generate public value (Kooiman and Jentoft, 2009).

The concept of performance has a broader meaning in relation to networks, as it coincides not only with the achievement of effectiveness, efficiency and profitability objectives, but necessarily surrounds members' involvement and satisfaction, as well as the possibility of changing traditional decision-making models (Voets et al., 2008; Verweij et al., 2013). In brief, public-private networks must be evaluated at least in three levels: in terms of outcomes towards community, improvement of participants' performance and achievement of network objectives (Provan and Milward, 2001).

The concept of a governance network is strictly associated to that of accountability, as the different categories of stakeholders must be enabled to effectively monitor and evaluate network results, in line with democratic control principle. Actually, the awkward relationship between democracy and deliberative model proposed by public governance theories has been one of the most debated issues in the last years (Arts and Tatenhove, 2004). Furthermore, some studies found that, in spite of the theoretical recognition of the value of collaboration, in many contexts typical 'government' approach is still used in public management (Loffler, 2003). This is partly due to the delay in research evolution from a first generation focused on network purposes, characteristics and connection with social trends (Kooiman, 1993), to a more contextualized and operative generation (Torfing, 2005).

Recent studies are mainly focused on the design of operational rules aimed at defining cooperation mechanisms and reward systems, as well as managerial tools to concretely promote collective decision-making and effective governance network management. Such an approach was also promoted by the European Commission in the green paper published in 2004[1], which identified as essential requirements of a PPP, that is a specific  type  of  governance  network, the project nature and middle-long orientation of the cooperation, the financial contribution of privates and the meta-governance role of the public sector. An emerging issue in the public governance research is multi-level governance, which focuses on the interplay between various government levels at several territorial tiers: supranational, national, regional and local (Marks, 1993) and has had little or no empirical grounding up till now (Petersen, 2011).

Evaluating governance networks' preconditions of performance

Despite the vagueness that often affected the first studies on networks and did not permit the building of a well-defined and univocal model of performance evaluation (Kenis and Provan, 2009; Singh and Prakash, 2010), a number of contributions have been able to identify critical factors influencing network outcomes (Menahem and Stein, 2013). These are considered to be essential references in the following attempt to define an evaluation frame of network performance. In the literature, the difficulty of quantifying specific indicators of performance (McGuire and Agranoff, 2011) suggested the pre-identification of some general conditions useful for facilitating the achievement of good performance under the three levels of community, network and participants (Provan and Milward, 2001). The level of integration among participants and the managerialization of a network stand out among the most significant pre-conditions of performance. The first one is viewed as a powerful tool to improve the life conditions of a community and single members' performance, while the second as an essential system to achieve the common goals of the partnership (Waring et al., 2013). The assumption of such an approach is that choices about a network's structural and operational aspects have a strong influence on performance (Goldsmith and Eggers, 2004). In the next sections, a set of variables to measure and evaluate both pre-conditions is identified on the basis of the literature. This theoretical framework is then employed in a case study, which permits the consideration of important external variables such as political orientation, regulation and multi-level relationships (Yang et al., 2013).

Integration among participants

With reference to integration among network participants, which guarantees the development of collective decision-making and social cohesion, a set of significant elements have emerged in the international debate. In the start-up phase, for instance, a variable to be monitored is the identification of the promoter, in order to verify the respect of the bottom-up principle (Agranoff and McGuire, 2001; Van Raaija, 2006). The criterion of selection of the relevant  interests  should  also  be evaluated as it should

be directed, following an inclusion perspective, to involve most of the population (Hanf and Scharpf, 1978; Vangen and Huxham, 2003; Menahem and Stein, 2013). Another two key elements are related to process management (Koppenjan and Klijn, 2004; Klijn et al., 2010): the concrete contribution, in terms of ideas and motivation, given by network members in the beginning phase, and the fair financial participation of privates (Commission of the European Communities, 2004; Herranz, 2010).

It is then important to identify the purpose of the partnership because the different options - promotion, coordination, collective planning, and collective implementation - translate into increasing degrees of collaboration intensity (Keast et al., 2007; Mandell and Keast, 2008). The project nature of a network is widely believed as an essential element for strengthening collaboration (Peters, 1998; Fischer, 2003), while particular importance is given to the fair length of the project, in order to facilitate actors and objective renewal (Provan and Milward, 1995; Kenis and Provan, 2007). Collaboration intensity, for instance in terms of amount and frequency of information and resource exchange, must be also evaluated during strategy implementation (Head, 2008; Edelenbos et al., 2010). Meanwhile, attention has to be paid to the fair contribution of participants in the operational costs (Lowndes and Skelcher, 1998).

The setting up of a meeting schedule makes a difference from an accountability perspective as it fosters regular comparison and discussion about objectives and results with all stakeholders (Mandell and Keast, 2008; Reeves, 2013). In this regard, information and communication technology (ICT) can be successfully exploited to promote participation, interactivity and social control through the development of online platforms and channels to communicate activities and results, and to gather feedback (Head, 2008; Moynihan et al., 2011).

Finally, among the variables that account for the good functioning of governance networks is the traditionally mentioned authoritative leadership. Though the public sector is often viewed as the most appropriate subject to meta-manage the network through catalysing, stimulating and feeding participation (Marin and Mayentz, 1991), the balancing of power and responsibilities between members needs to be constantly evaluated (Klijn and Koppenjan, 2000; Agranoff et al., 2003).

Managerialization of the network

Network maturity under a managerial perspective is an essential pre-condition to achieve common goals. Nevertheless, defining an advanced management system proved to be a very hard task in the past, due to the fact that traditional planning, controlling, reporting activities must be rethought to reflect network complexity (Verweij et al., 2013). In this context,  for  instance,  resources  are held by independent actors and management has to be strongly externally oriented (Klijn et al., 2010).

The criteria followed to select relevant interests deserves as much attention as for the integration profile, but to assess whether members holding the best resources, other than having a primary interest, have been included (Hanf and Scharpf, 1978; Vangen and Huxham, 2003; Menahem and Stein, 2013). Participants' contribution, in the starting phase, should then be judged in relation to negotiation intensity (Lowndes and Skelcher, 1998; Keast et al., 2004). The financial participation of privates, which influences the profiles of fairness and motivation, also deserves the same consideration it had with reference to integration (Herrantz, 2010).  

The organizational form of the partnership is a remarkable point, since different choices - no entities but informal agreements, negotiation table meetings, new entities (company, consortium, association) - imply very different involvements and coercion power (Keast et al., 2004; Mandell and Keast, 2008). Then, a powerful management tool is the setting of behavioural rules in order to anchor relationships to a common normative framework (Koppenjan and Klijn, 2004).

From a managerial point of view, the project nature of a network must be verified through the quantification, both technical and financial, of objectives, which are an essential steering and controlling mechanism (Edelenbos and Klijn, 2006). In order to safeguard the multidimensionality of network performance, the nature of objectives, in terms of input, output and outcomes, must be also assessed. Finally, the connection between objectives and a reward system is crucial for enhancing single contributions (Klijn and Koppenjan, 2000).

A remarkable variable to effectively manage networks, of course, is the creation of authoritative and independent control units to measure performance (Marin and Mayentz, 1991; Provan and Milward, 2001). Stability and continuity of internal control benefit from quick and transparent information channels, and setting meeting schedules to compare and debate results among members is strongly recommended in order to stabilize negotiation and give operational flexibility (Machado and Burns, 1998). The adoption of ICT to share information and data among participants is seen, also in this case, as a powerful mechanism for strengthening collaboration (Herranz, 2010; Klijn et al., 2010). Furthermore, the horizontal structure of networks, and their features of flexibility and lifelong learning, highlight the importance of establishing vocational training courses for participants (Klijn and Koppenjan, 2000; Keast et al., 2004).

Finally, even if participants have equal responsibility, a network manager assumes a primary role, and his/her functions imply a different coordinating power of members. He/she could just represent a simple information centre or have supervision responsibility, and he/she could act as a super parts authority in charge of settling controversies, rather than directly stimulating participation (Mandell and Keast, 2008).

The purpose of highlighting governance network peculiarities in tourism, and in particular investigating the profiles of integration and managerialization and their evolution at two moments in time, led to the adoption of a case study approach, both exploratory and explanatory (Yin, 1994). This was to gather information on a contemporary phenomenon characterized by complex, but not always evident, relationships with external variables such as political orientation, regulation and multi-level relationships. This choice was also in line with the purpose of evaluating the impact of the internationally ongoing crisis on the evolution of local networks. The significant case of the local tourist system 'Sistema turistico locale (STL) Sardegna Nord-Ovest', characterized by unique internal and external features, was selected and an empirical investigation was carried out through different phases and tools.

Several reasons make tourism a privileged field for network research, as testified by the long tradition of this kind of study in this sector (Pavlovich, 2003; Scott et al., 2008). Firstly it is a local phenomenon, and is comparable to a 'public good', so it is particularly appropriate to analyse public sector roles and participatory decision-making in the development of territories following a public governance perspective (Shaw and Williams, 2004; Klijn et al., 2010). Furthermore, in Italy, an innovative tourism development model, based on the principle of the cooperation between public and private actors, has been proposed. The reform Law No. 135/2001, in particular, promoted the development of tourist networks, called STLs, defined as 'homogeneous or integrated contexts which can comprise territories of different regions and are characterized by an integrated offering of cultural goods, environmental goods, tourist attractions, local handicrafts and agricultural goods, or by the presence of single or associated tourist companies'.

A lot of doubts have emerged during the implementation of the model, mainly referring to the effectiveness of promoting resource exchange and participation among stakeholders. The result was that the model has been scarcely enforced at country level also because regions were encouraged but not obliged to adopt it. In particular, an empirical investigation was carried out, using a personal interview approach, in 2011, on a sample of eight STLs, with the aim to measure and evaluate network effectiveness under integration and managerialization dimensions (Rotondo, 2013). After more than two and a half years, this evaluation is now being repeated on the STL Sardegna Nord-Ovest, which in the previous study was shown to be among the worst performers in terms of both integration and managerialization. This evaluation is being implemented using the same theoretical framework, direct observation of the different development phases, document analysis and new semi-structured interviews to the same STL's representatives: the president of the board of directors - acting also as the Chief Executive Officer (CEO) - and a member of the board of statutory auditors. In brief the longitudinal empirical investigation covers about a 5-year time horizon (that is also the length of the first activity programme of the STL) shared at two moments in time. The evaluation of integration and managerialization pre-conditions, in fact, is first performed after three years from the creation (first time interval: 2008-2011), and then repeated two and half years from the renewal of the partnership (second time interval: 2012-2014), in order to interpret network evolution in light of the external variables and especially of the recent, and still ongoing, crisis situation.

Among the main reasons driving case selection, is that in spite of showing low performance levels, the STL Sardegna Nord-Ovest decided to  reply  to  the  cut  of  regional  resources  from  2011  by renewing commitment and re-launching the collaboration. The new impetus given to the partnership in a situation of progressive abandonment of the model at national level makes the STL highly significant to interpret crisis effects on network development. In Sardinia, furthermore, the international financial crisis has affected real economy since 2008, provoking strong slowdowns in production and an increase in unemployment rates after a long expansive economic cycle. In 2012 the main conjunctural indicators, such as production, turnovers and investments of firms, dropped off again after a slight recovery, with a decrease in the employment rate of 1.1% compared to 2011 (twice and four times the average of the other Italian southern regions and all Italian regions, respectively). After falling, 3.4% in 2012, regional constant-price GDP decreased by 2.5% in 2013, when the regional employment rate reached the lowest level for the last ten years (-7.3%), considerably higher than the Italian average (Bank of Italy, 2014). In particular, North Sardinia showed an unemployment rate in 2012 equal to 14.65%, lower than the average of the Italian Southern regions (17.2%), but remarkably higher than the Italian average (10%) and higher than single averages of North-western (8%), North-eastern (6.7%) and central regions (9.5%) (Istat, 2012). From a social point of view, the crisis impacted on the quality of life of the community, as revealed by the decrease in the deprivation index in 2012 when the province of Sassari was in the third worst position in the regional ranking, with extremely bad results in education and services.

The tourism sector, which is outstandingly important for the Sardinian economy as it weighs on GDP for 10% (lower than the only percentages of Trentino Alto Adige and Valle d'Aosta at National level), has suffered from crisis effects after some time, only showing a weak recovery, after years of contraction, in 2013. With regards to North Sardinia, in the province of Olbia-Tempio, where 40% of regional tourist flows are concentrated, tourist presence dropped by 18.3%, while in the province of Sassari tourist presence decreased by 24.7% (Sardegna Statistiche Turismo, 2012).

From constitution to renewal

Law no. 135/2001 replaced the normative framework set with Law no. 217/1983, with the aim of promoting a new participated development model in tourism in contrast with the political and bureaucratic approach of the previous reform. Since regions have autonomy in tourism organizations, as decided by the reform of Title V of the Italian constitution, the law was just a proposal that the regional government could choose whether or not to acknowledge. Sardinia is among the regions to have acknowledged the normative framework, with DGR no. 34/11 of 2005, and consequently constituted the STLs. Nevertheless, the innovative impact of reform was limited as the regional government decided to establish STLs in the same number and with the same political borders of provinces, and to give provinces a preeminent power in their management.

In this situation, in 2008 the constitutive act of STL Sardegna Nord-Ovest was signed. The network involved 83 participants, among which were 64 local governments of the Province of Sassari, the Province of Sassari as network promoter and leader, and 17 private actors, mainly associations representing the different tourist sectors. The geographical extension of the network was major, as it covered more or less the whole 4,281 km² of the province. A board of directors was appointed, made up of six members of whom four represented the public sector, and it was headed by the president of the Province of Sassari. The STL, since the beginning, could rely on €2,180,993 of regional funding, the highest amount among the regional STLs.

After a complex start-up phase, testified by low levels of integration and managerialization compared to other similar experiences (Rotondo, 2013), the partnership suffered from strong disaffection in 2011, due to the uncertainty provoked by central governmental spending review and regional governmental decision to cut funding. Nevertheless, in July 2012, the majority of participants, and new entrants, reinvigorated collaboration by establishing co-financing shares and renovating government bodies. The board of directors is now made up of five members, four from the public sector and equally representing local governments from the coast and the internal territory, again headed by the province through the Assessor of Financial Affairs, also acting as CEO.

The network comprises 80 members,  15  of  which  are private, and extends across an area 220 km² larger than before, since some local governments from the province of Olbia-Tempio are now included. For this reason, it has been decided to change the official name of the partnership to 'STL Nord Sardegna' (Figure 1). Tourism has a central role in this area with strong relevance for the whole regional economy, since recent official data revealed that 41.6% of all accommodations are located in north Sardinia, and that the weight of passengers flying on Alghero and Olbia airports on total passengers has been constantly increasing for the last five years, up to 49.47% in 2013 (Assaeroporti, 2014). Three out of four of the main regional ports are located in north Sardinia (Olbia, Golfo Aranci and Porto Torres), and the recovery of maritime traffic recorded in 2013 after permanent decreases are mainly ascribable to them (Assoporti, 2014). The weight of arrivals and presences in north Sardinia accommodation to the total regional ones have been increasing since 2008, despite the fact that they have been slightly falling in absolute terms.

The activity of the STL in the last two years has been affected by strong financial pressure to have hit local governments, which, other than the tight restrictions of the stability pact, had to face major cuts through the Unique Fund for Financing Local Governments, established by Regional Law no. 2/2007. Difficulties were then amplified by uncertainty about the role and responsibilities of provinces following the referendum of May 2012, which declared the popular will of removing four Sardinian provinces (those most recently established), and National Law no. 56/2014 aimed at reshaping borders and competencies of local governments, which is still waiting the reform of the constitution to become effective. The STL is therefore in troubled waters again, and its possibilities of survival are unclear at present.

The evolution of the STL Nord Sardegna, under the profiles of cohesion among participants and managerial advancement, is assessed by comparing a set of 28 operational variables, 14 for each dimension, measured in the two time intervals and selected in accordance to the theoretical framework. This means that each variable mentioned in the theoretical section has been opera-tionalized into an indicator. Tables 1 and 2, other than describing variables and ways of measurement, permit the comparison of the scores, both for single variables and total values of the pre-conditions, measured in the two time intervals. Single variables have the same weights, while total scores have been standardized into a 0/1 scale in order to increase comprehensibility. Values have been attributed through personal interviews with STL's representatives, and results must necessarily be supported by qualitative analysis, in order to shed light on the context, causes and motivations behind network institutional and process strategies.

It is worth noting that total scores, for both dimensions, are around the average in each time interval. This confirms the generally limited attention paid to participation promotion and management system development. Otherwise, network evolution can be assessed by observing which variables increase or decrease over time. In this regard, from a temporal perspective a trend change can be noticed, as in the second time interval the profile of network integration improves while that of managerialization worsens more markedly.

With reference to the evaluation of integration (Table 1), no change appears in relation to the promoter, which is still the province, to the criteria used to select relevant interests, which is still representativeness, and to the intensity of members' contribution in the early stage of network renewal. 'At the beginning, socioeconomic actors’, say the president of the board, ‘were all aware of the importance of collaborating and the value of partnership’. During this phase the STL is perceived as 'a steering committee, a mind able to control the limbs, represented by participants, in tourist strategy formulation and implementation’.

The cooperation is strengthened in the face of crisis, and this is testified by both the increasing functions assigned to the STL (variable 5), which becomes 'a platform to debate, decide and implement the future of the territory', and the entrance of new members. The financial commitment of privates remains constant and  is in the bracket between 15 and 35% of company capital, while the cutting of regional resources seems to promote local governmental commitment through the sharing of operational costs (variable 10). Consequently, the profile of equity inside the network improves. Nevertheless, as time passes, local governments become more and more reticent to cooperate, and participation drops off. 'The public sector - the president argues - slowly drew back, as local governments stopped taking part in meetings and paying their own co-financing shares’. This change is basically due to the financial pressure burdening local governments, as resources available from the Unique Fund are significantly reduced by the region, the new provinces are put under temporary receivership after the referendum, and tourist competencies of the old provinces are limited by national legislation.

In this regard, the view of the statutory auditor highlights a major point for discussion: 'crisis, among the main causes of system disintegration, hit all participants, but it was more intensely perceived by local governments’. The belief is that the crisis effect had been amplified by what is defined as 'a lack of governance culture in the public sector', which is the inability to really understand that network advantages lie in the concepts of sharing, future-oriented vision, trust and common good. 'The public sector, and in particular local governments’, says the president, ‘have always been used to gain rather than share resources’.

In fact, a gradual weakening of the common vision of the future emerges, and this will be confirmed by managerialization results. With reference to integration, this circumstance can be observed in the replacement of a coherent and long-term development project by a set of short-term objectives, in large part gradually abandoned. Among these is the development of a multimedia platform to report the results of the network and stimulate interactivity with citizens. In general, community involvement does not improve, and citizens’ participation is basically ignored instead of being considered as a powerful tool to drive and stabilize the partnership. This is in line with the findings of the study performed by Mitchell and Reid on community integration on Taquile Island, Peru. They found that a direct and high level of community involvement in the tourism planning and management process is conducive to greater socioeconomic benefits for a majority of residents (Mitchell and Reid, 2001).

Formally the responsibility, compatibly with co-financing shares, is equal among members, but in practice the crisis scenario enhances the supremacy of the public sector and especially of the leader. In fact, network management prerogatives are gradually centralized in the figure of the president of the board who, as claimed by the  statutory  auditor,  'was  able  to   successfully coordinate relationships in a troubled time through commitment and personal, relational and reputation skills coming from the double role of assessor of the province and tourist entrepreneur’. However, this probably led to the arrested growth of the network as a whole. In fact, it seems that the enlargement and strengthening of the collaboration was not accompanied by an adequate development of planning and control systems at network level. Definitively, while integration seemingly increases, as testified by the new network mission, the entrance of new members  and  the  operational  costs   shared  by participants, when a crisis occurs some structural problems of the partnership seem to prevail over participant motivation and commitment. In particular, the representation as the main criterion to select relevant interests, the low level of resource exchange among participants and citizen involvement and the predominant role of the leader are the main causes of the structural incongruence.

With reference to managerialization (Table 2), behavioural rules and reward systems to stimulate long-term participation and promote meritocracy are not established even in the second time interval (variable 5 and 8). This issue is related to the gradual loss of the project  nature  of  the  STL and to the appearing of a new relationship model. Firstly, objectives are less in number and only qualitatively expressed (variable 6). Secondly, they only aim to show immediate and recognizable effects to participants, for instance the develop-ment of new tourist products such as sea-mountain packages together with other Italian regions, rather than websites to sell destination products, or the opening of a tourist office at Alghero airport.

These kinds of objectives testify a new approach in network relationships, based on the principle of 'activating and motivating participants’, as claimed by the president, ‘through "strong strategic actions"  having  short-term  effects’. The strategies used to launch the partnership at the two time intervals illustrate such a change. The first strategy taken to relaunch the partnership illustrates such a change. In fact, at the first time interval a camper was sent around the territory of local governments to gather tourist data, while at the second one an educational tour for tourist reporters and operators, with great media prominence, was organized. The importance was similar, but the impact more visible and immediate.

As the lack of resources continues, not even the stronger short-term orientation is able to improve programme implementation. Participants pro-gressively retire under the leader's protective umbrella  and  this frustrates the need, recognized even in the first time interval, for adopting communication channels and tools to exchange information (variable 10 and 12) and setting discussion meetings among participants (variable 11). Consequently, the accountability profile, both internally and externally, worsens. Particularly, the constant value of variable 11 at the two moments in time means there is no relationship between it and the managerialization level, but it also highlights a negative feature of the partnership. This is the passive attitude and involvement of participants and community. This is due firstly to initial choices not adequately promoting operational responsibility of participants. Secondly, to what emerges as a central point in the discussion, that is the value of the social capital of the territory. The social capital can be defined as a community awareness, ability and tradition in developing public-private partnerships. Bad or limited experiences, as in the case under investigation, seem to have an impact on the balance of power and the clarity of roles within a network. The importance of information sharing, social cohesion and mutual goals that constitute social capital is even higher for short-term governance networks' success, as shown by Borg et al. (2015) in their study of government funded collaborative networks for forest biodiversity.

Representatives' views about the impact of external variables support the interpretation of network evolution. The territory reveals an insufficient tradition of partnerships, as these had been usually limited to some political-administrative agreements, and so the STL was 'a brand new experience of an area’, the president notes, ‘where relationship capital is still to be built. (...) We are just beginning the process of establishing a long-term collaboration between public and private actors’. The lack of experience of participants is shown by what he defines as an 'attitude of resignation which suddenly emerges among participants when facing the first difficulties and conflicts’.

As financial crisis increases, the public sector is seen more and more as a lifeline, partly because, for both the interviewees ‘it has always been the driver of the tourism sector in the area’. From an inter-institutional point of view, while local governments have gradually referred their responsibilities up to exclusively rely on the leader, the latter in turn has complained about the lack of strategic and operational support by the region, which is seen as the main cause of network failure. The role of the region, the president argues, ‘was more counter-productive than insignificant, as it firstly created and considerably funded the STL, and then let it alone. (...) No kind of assistance or support has been given to accompany partnership development and maturity. The best metaphor to explain what really happened is that of the mother who gives birth to a child and then let him/her alone in the world’. Interestingly, excessive centralism and short-sighted projectuality that affected the STL, as revealed by the investigation at the second  time  interval, are now attributed to the regional approach to tourist network organization. Regional government, in the opinion of the STL president, ‘lacks of a strategic vision and awareness about the importance of collective decision-making in line with public governance view, and this is more critical than the lack of resources’. Another interesting point of view about the decision of the region to cut funding is that of the statutory auditor, who thinks that 'the region, even if appreciating the job of the STL Nord Sardegna, considered it as a "white fly". As all the other experiences of Sardinian STLs failed, regional government had to take note of the general failure and branded partnerships as ineffective’. From a multi-level governance perspective, the results show that the achievement of good network outcomes is strongly influenced by the support offered by higher hierarchical levels, which confirms the findings of the study of Ayres and Stafford on regional governance networks in England (2014), and that of Johnston on local partnerships (2015). Otherwise, centralism can negatively affect the process of diversification and spatial spread of the tourist supply, as reported by Baidal in his study on the evolution of the Spanish tourist model (2003).

Actually, this study shows that the initial choices of the regional government, such as replicating the political borders of provinces and giving provinces the highest power among participants, raise some concerns about the real purpose of promoting equal, independent and local partnerships.

In general, some structural problems of the partnership frustrate good intentions of participants (Klijn et al., 2010), whose role becomes more and more marginal with the first difficulties. Furthermore, the low level of social capital, shown by the passive attitude of network participants, is enhanced by an instability factor which persistently affects partnerships: the underestimation of advanced management systems at network level (Verweij et al., 2013). These, in fact, are essential tools to manage relationships and promote network self-regulation (Torfing, 2005). Finally, in such a context, the lack of resources following a crisis is not the cause of the problem while a hindrance in efforts to improve participant collaboration and negotiation.

Some recommendations for local governance can be made based on these findings. Firstly, public-private partnerships are an important way, in a time of crisis, to share resources and promote the growth of territory from a socio-economic point of view. Wicked problems related to crisis, in fact, can be view as sources of innovation (Kickert, 1993) and, as pointed out by Lowndes and Skelcher in their study of UK urban regeneration partnerships (1998), "the innovation has come in the form of strategies to develop interrelationships, trust and collaboration in an environment of resource scarcity". In this regards, Considine (2013) argues that governance networks can effectively be used to bring about systematic change, as they can be viewed as "a viable carrying mechanism, a unique form of inter-institutional capacity, able to achieve certain forms of learning that are not locked into an existing trajectory, as well as embedding what is learned in a format that is available to future actors who may be located in an adjacent domain". Secondly, crisis propagation can undermine partnerships affected by pre-existing structural problems, such as centralism or low community and participant involvement, rather than those built on instable ground. This finding is consistent with previous studies that have demonstrated that some structural preconditions can facilitate agreement among actors about goals and effective coordination of action (Robins et al., 2011). Thereby, local governments should not only promote partnerships but also provide partnership education. Community members should be made aware of the importance of partnerships for opening up local decision-making processes, as well as of the power and control they have in their development and management. In order to institutionalize partnerships local governments need to build what Machado and Burns (1998) define as a 'social rule system' , made up of a language, symbol systems, cultural codes and forms, and values. Evidently, such a process needs sufficient time to produce results. Furthermore, specific procedures and mechanisms are also required to build relationships in early network development and then promote self-regulation. In this regard, a 'management approach', despite being often neglected among other conditions, significantly influences network performance, as shown by Verweij et al. (2013) in their comparative analysis of 14 Dutch spatial planning projects. Several studies indicate that network management is an important determinant of outcomes, such as the research of Waring et al. (2013) on two partnerships in the English National Health Service, which also highlights the importance of considering the contingencies in how networks are operationalized on the ground.

This paper has focused on the profiles of the integration and managerialization of governance networks, viewed as essential pre-conditions for achieving good performance in relation to community, participants, and the network as a whole (Provan and Milward, 2001; Klijn et al., 2010). The significant case of STL Sardegna Nord Ovest was studied through the development and implementation of a theoretical framework during two distinct phases of the life of the partnership. The study revealed, first of all, a quite moderate attention paid to both the cohesion among members and managerial advancement in tourist networks. It also highlighted the risks behind the development of local partnerships in areas lacking similar experiences, which are the persistence of the bureaucratic model of government based on a top-down approach and the little room left for private initiative and self-regulation.

The evolution of the network was interpreted in light of key external variables, and especially the major socioeconomic crisis that has affected Sardinia, similarly to other Eurozone areas, since 2009. The experience contributes to the understanding of the impact of crisis on governance networks and the identification of the strategies, both institutional and operational, that can attenuate negative effects and promote network stability and effectiveness.

In comparison with the first phase of life, the second time interval of the partnership showed a slightly higher integration among participants, mainly due to the decision to strengthen collaboration to cope with funding cuts, but with a marked decline of managerialization. In contexts that have already appreciated partnerships' advantages, crisis advent can lead to the enlargement and stabilization of the network, while crisis propagation seems to mainly affect projectuality and future orientation, which are key elements from the public governance perspective (Rhodes, 1996; Stainsby, 2000). Uncertainties associated with crisis, in other words, facilitate the resurfacing of the typical elements of the bureaucratic approach to managing relationships between public and private actors, such as centralism and failed devolution, or undertaking of responsibilities. Short-termism, exclusive reliability on the leader, and possibility of immediate benefits prevail, in this case, over the concepts of future orientation and sharing. The same risk seems to undermine multi-level relationships, both between local governments and the leader, which become a sort of lifeline to the detriment of participation and programme implementation, and between the leader and the regional government, increasingly asked for resources, assistance and support.

Nevertheless, more important than the lack of resources for network disintegration is the lack of a 'governance culture’, that is the inability to understand and exploit the benefits of deliberative and operational processes based on flexibility and participation. In this regard, crisis appears to be a triggering event rather than the cause of partnership dissolution, due to intrinsic weaknesses instead. Trust, common good, subsidiarity and autonomy are some of the concepts whose meaning, in opposition to what had been stated at the beginning, quickly fades with the first problems.

Instability factors are sometimes not completely removable, at least in the short term, such as poor experience in partnerships on the territory. The 'relationship capital' around which to build up a network, in fact, needs sufficient time to be developed. Other times, however, the choices of institutional design and process management are the main causes of network success (Klijn, 2004; Klijn et al., 2010). Conservative choices or choices scarcely in line with public governance perspective by policy makers can undermine, from the start-up phase, self-regulation of a network and fairness among   participants.   These  problems  seem  to  impact more negatively on public-private networks developed in sectors, such as tourism, with a strong local propensity. Finally, the working phase is mainly affected by a persistent underestimation of developing advanced network management systems (Verweij et al., 2013).

The authors have not declared any conflict of interests.