Organising collective reputation: an Ostromian perspective

: What do collective reputation and communal pastures have in common? Collective reputation is an important type of collective good produced by many business networks. We argue that it has the structure of a common-pool resource, which points to the relevance of Elinor Ostrom’s theory about the community governance of natural common-pool resources. After adapting the Ostromian framework to the phenomenon of collective reputation, we explore the experience of two groups of winemaking enterprises in Hungary who set up systems of quality assurance in order to protect and improve their joint reputation. We examine if the conditions identified by Ostrom as favourable for the self-gov-ernance of commons are also conducive to the governance of collective reputation. Our findings validate our conjecture that research on goal-oriented business networks may use insights from the mature theory of ‘governing the commons’. Potential pathways for further research are outlined. we illustrate the relevance of for the of We explore the experience of two who set up self-governing to and as immaterial We examine indeed a prerequisite the emergence of self-governance in


Introduction
What factors are conducive to the self-governance of collective reputation by business groups? In our paper, we attempt to identify these factors using a case study conducted in two Hungarian winemaking communities. We take the accumulated insights about the governance of natural common-pool resources (CPRs) as the basis for our investigation. In particular, Elinor Ostrom and her co-authors were able to identify those resource and group attributes that were conducive to the emergence of successful self-governance for natural CPRs (Ostrom 1990(Ostrom , 2005Dietz et al. 2003).
Like communities managing natural resources, many business networks are also stable multilateral cooperative relationships that aim at providing certain types of collective goods (Kilduff and Tsai 2003;Provan and Kenis 2008). How relevant is Ostrom's CPR-theory for such groups or networks of business organisations? We show that at least one important type of collective good produced by business networks -collective reputation -has the structure of a CPR. For such cases, Ostrom's theory is directly relevant. We show further that the conditions that were found to be favourable for the self-governance of natural resources can easily be reinterpreted for collective reputation. Aspects of collective reputation are analogous to attributes of a natural CPR; characteristics of a business community are analogous to those of a user group of the commons.
With a case study, we illustrate the relevance of Ostromian conditions for the production of collective reputation. We explore the experience of two local communities of winemaking enterprises in Hungary who set up self-governing quality assurance systems to protect and improve their collective reputation as an immaterial CPR. We examine whether the presence of Ostromian resource and user attributes was indeed a prerequisite of the emergence of self-governance in the two communities.
The case study illustrates the potential importance of collective reputation for stable groups of private organisations as well as the role of multilateral cooperative efforts in producing it as a collective good. It also shows the fruitfulness of applying Ostrom's theoretical framework to purposeful, goal-oriented networks of organisations. 1 Space does not allow us to dig into the institutional details of governance in the two communities. In an accompanying article (Mike and Megyesi 2016), we analyse if the principles of institutional design which Ostrom identified for the governance of natural CPRs are also present in the self-governance of collective reputation.
First, we elaborate the concept of collective reputation as an immaterial common pool resource (Part 2). We then adapt Ostrom's taxonomy of favourable resource and user attributes of natural CPRs to the phenomenon of collective reputation (Part 3). In Part 4, we present the case study and examine the validity of our conjecture that the presence of Ostromian attributes facilitates the selforganisation of producer communities. The final section evaluates the results and concludes.

Collective reputation as a common-pool resource
CPRs are goods from which it is difficult to exclude or limit users, whose consumption of the good subtracts from the units available to others. They present problems that 'are among the core social dilemmas facing all peoples' (Ostrom 2005, 219). Although their most studied examples are natural or man-made agricultural resources, such as common pastures, fishing grounds or irrigation systems, a great many other phenomena can be successfully analysed as CPRs, ranging from technological resources [e.g. frequency spectrums (Wormbs 2011)] to services [e.g. healthcare (Jecker and Jonsen 1995)] and immaterial goods [e.g. scientific information (Hess and Ostrom 2006)]. We study one type of immaterial CPR: the collective market reputation of producers in a geographic area. The reputations of 'Swiss made' watches, 'Silicon Valley start-ups', 'Parma ham' or 'Bordeaux wine' are maintained by groups of producers and may become 'depleted' just like common pastures, fisheries or water basins (King et al. 2002;Patchell 2008).
Reputation is a signal that helps overcome the information asymmetry between a producer and its customer. Reputation only works if customers receive sufficient and credible information at low cost about the quality of a product. For small firms, conveying reliable and easy-to-digest information may be prohibitively costly. They may opt for creating a 'joint brand' with similar producers (Tirole 1996;Fishman et al. 2010). In other words, they can economise on the scale of investing in reputation. The development of a 'local image' (Ray 1998) or the use of geographical indication for similar products or services in a given region or locality (Van Ittersum et al. 2007) represent important instances of such a collective brand.
Collective reputation has been interpreted both as a pure public good (Cox and McCubbins 1993) and a CPR (Patchell 2008). Both concepts entail the difficulty of exclusion. If a region is famous for its good wine, no local winemaker can be excluded from the benefits of its good reputation, unless there are special rules to exclude him. The difference between a pure public good and a CPR consists in subtractability: does the use of reputation by an individual subtract from the benefit of reputation for others, or not? We believe that, for a group of producers, a fundamental collective action dilemma is that they may individually abuse their collective reputation by producing low quality at a cost advantage and obtaining high benefit in the short run. Although they do not literally appropriate or consume 'units' of reputation, they do subtract from its stock of value. Solving this problem requires the institutionalisation of quality control. Like for the natural commons, this may take the form of community self-governance. 2

The relevance of Ostromian conditions
As noted above, Ostrom (1990Ostrom ( , 2005 identified resource attributes as well as the attributes of resource users which had been found to increase the likelihood that self-governance for a natural CPR will emerge. These attributes can be interpreted in a relatively straightforward way for the collective reputation of local producers as an immaterial CPR. Table 1 summarises the original Ostromian conditions and their adaptations to collective reputation. By analogy, we conjecture that the presence of all of the above attributes at least in a weak form enables the community to form institutions of selfgovernance with an aim to preserve and improve collective reputation. No substantial institution-building can be expected in a group of producers as long as it is seriously lacking in any of these attributes. 4. Case study: Producers' efforts to improve their collective reputation in two Hungarian winemaking communities, 1990-2014 Most studies are sceptical of the viability of creating self-organised groups of producers in post-communist countries. The successful self-governance of CPRs is rarely observed. Their absence is usually explained as being due to a relative weakness of social capital in producer communities and the lack of supportive political and administrative context (Sikor 2002;Theesfeld 2004;Upton 2008;Schlüter et al. 2010;Schmidt and Theesfeld 2012). Yet, there are important exceptions, proving that such cooperative efforts can emerge even in a relatively hostile socio-political environment (Mearns 1996;Gorton et al. 2009;Sutcliffe et al. 2013). Our examples fall in the latter category.
Two communities of winemaking enterprises, Csopak and Tihany, are located on the northern shore of Lake Balaton in the Midwest of Hungary. Both belong to a wine region with great traditions. Csopak is the region's most famous locality for producing white wine, especially Riesling. 3 Tihany is well-known for producing high quality red wine in the special climate of a peninsula, in a region otherwise known for white wines. Starting from a virtual tabula rasa after the end of communism, producers in both localities struggled long to establish systems of quality assurance to protect and possibly improve their collective reputation as an immaterial CPR. Tihany succeeded in 2008, with Csopak succeeding in 2012. Although both initiatives are relatively recent and more time needs to pass until Natural CPRs (Ostrom 2005, 244-5) Collective reputation of local producers their robustness can be judged, in the post-communist context their very existence and functioning are significant achievements.
Based on in-depth fieldwork, we identified to what extent the resource and user attributes defined by Ostrom characterised the two winemaking communities after the end of communism and how these conditions changed over time. As we shall show, self-organisation indeed presupposed the presence of Ostromian resource and user attributes. Each community established its system of collective quality assurance only after these attributes were sufficiently available. This also explains why the two communities set up their respective system at different times. The late-coming group was thwarted for a time by the absence of some important attributes.
Fieldwork was conducted in both places in March-May, 2012. After studying each community's history and geography, statistical data on wine production as well as official and press documents, 18 semi-structured interviews were conducted with local winemakers, officials of the wine communes and representatives of local governments. In December 2014, follow-up information on the self-governing initiatives observed in 2012 was collected through the websites of interviewed winemakers, their associations and professional wine-related media.

Resource attributes
Consumers' perception of the quality and character of wine is influenced by the place of its production or 'terroir'. Although the true significance of geography is continuously debated, geographical indications on labels are used more and more frequently to signal uniqueness and high quality (Johnson and Robinson 2007). Hungary is a country with a rich history of long-standing local identities in winemaking. In what is perhaps the most famous literary treatment of Hungarian wine, The Philosophy of Wine (1998), Hamvas strongly emphasises the distinct characters of regions and localities. One highly praised region is the district of Lake Balaton, among whose outstanding and idiosyncratic localities he mentions both Csopak and Tihany. The association of narrow locality and quality is now clearly present in local thinking, too. As an interviewee put it: 'It is obvious that Lake Balaton or Transdanubia [the even larger region] are in general not known for high-quality wines. It would be against common sense to market high-quality Riesling produced in Csopak under such general names'.
The continuing presence of collective reputation was indirectly confirmed when winemakers reflected on the years around the end of communism: 'the wine produced was uniform and mediocre; very often wine produced outside the region was bottled as 'Tihany' or 'Csopak' wine'. The reputation of both communities was at a long-time low in 1990, reflecting the deteriorated state of Hungarian viniculture in general. Creating a system of quality assurance became a key challenge.
Organising collective reputation: an Ostromian perspective 7 The Hungarian market for high quality wines emerged slowly but steadily. The first professional wine guide to orient consumers appeared in 1995. Nineteen years later, a national survey of consumer preferences (Research Institute of Agricultural Economics 2014) identified 16% of consumers as 'wine experts', 38% as 'demanding consumers' and 46% as 'non-distinguishing consumers'. 'Wine experts' said that terroir was most important for them, followed by grape variety, price and winery. For 'demanding consumers', terroir ranked second after grape variety, while for 'non-distinguishing' buyers it ranked third. So terroir was clearly important in the upper segment of the market. The survey also showed willingness to pay for quality: 40% of respondents stated that they were willing to pay more than 3.5 euros (the price level where any quality begins), and 10% were willing to pay more than 6.5 euros (where the top range begins). 4 Although there are no direct indicators of the market value of collective reputations, several market actors emerged over time to provide useful feedback. 'Hungarian Wine-Grower of the Year' awards are given annually; numerous prize contests take place and several professional wine merchants with reputation organise markets for demanding customers (Tóth 2010).
From the perspective of the present, one can clearly see that the market value of collective reputations could be improved (R1), there are some indicators of this value (R2), and the slowly increasing demand for high-quality wine gives some predictability to revenues (R3) due to terroir reputation. However, these resource attributes strengthened very slowly as market development was protracted and uncertain throughout the 1990s and early 2000s. Although we could detect no difference in the general market environment for the two communities, they do differ in one respect which affects the feasibility of improving the value of reputation: size. Tihany has roughly 70 hectares of vineyards, while Csopak is considerably larger with approximately 250 hectares. Both are small territories with wine-growers who work in proximity (R4) but Tihany producers can achieve less economy of scale in building a collective reputation.

User attributes
The basic structure of producers' communities is similar throughout the whole statistical wine region officially entitled 'Balatonfüred-Csopak', to which Tihany and Csopak belong. In 2005, 5325 registered wine-growers cultivated 2270 hectares, i.e. 0.43 hectare per capita. The vast majority (95%) of these people were part-time producers who drank their own wine or sold it (informally) to a local pub. They had no real stakes in either individual or collective market reputation. As the president of a wine commune put it: 'Smaller producers are not worth talking about… because they do not belong to the future, sad as it may be'. There are also a few large-scale wineries 5 with more than 70 hectares, based outside of the localities, which focus on mass products and show no special interest in terroir reputations. In between, there is a group of medium-sized private wineries which build their own brands and target quality. This is the group, consisting of 10-15 cellars/community at most, whose members take an interest and are active in improving collective reputation. They can be divided into two subgroups which differ in many respects: (1) estates of local families and (2) non-local investors.
The former category consists mainly of local 'self-made men', who started winemaking in the 1980s as employees at a cooperative or state farm, engaged in individual production on small household plots, and acquired estates of 5-25 hectares after de-collectivisation. They nurture close and trustful personal ties (U4) and share professional experience and views that go back to the era of communism and often much longer. They view their enterprises as being embedded in a long chain of generations: 'Professional know-how is transmitted from fathers to sons, from grandfathers to grandsons'. Their heirs work at the estate and plan to continue the family business. Their livelihood depends entirely on winemaking and complementary services, such as running a restaurant, a shop or a bed-andbreakfast. The reputation of the winemaking community is clearly salient for them in terms of both identity and income (U1), and they plan for the long-term (U3).
Non-local investors arrived in the area mostly in the 1990s and early 2000s. They collected financial capital in other activities, and bought a cellar and land with the aim to produce high quality wine. Although the relations with the first group are mutually respectful, they are somewhat less embedded in local personal networks (U4). Their personal and family identity is less tied to the local winemaking community, and several of them draw income from non-wine-related activities. Nonetheless, the reputation of the terroir is economically salient for them as they target the higher end of the market (U1). Moreover, they do invest and plan for the long-term (U3). Their understanding of what is valuable in the terroir and how it should be developed (U2) differs from that of the typical local self-made-man. In interviews and marketing materials, they tend refer to ancient traditions that were broken during the communist era and the need to experiment and learn from the outside world. The winemaker who is perhaps the most respected member of this group in Csopak emphasises on his website '50 years of lack of stewardship by the state' and the 'inheritance of mere ruins'; makes a pledge to remain 'humble before traditions that were often broken but always revived', take 'the great wines of the world as one's measure' and 'to experiment with everything'. 6 However, the most marked difference of vision is not between local selfmade-men and non-local investors but between winemakers aspiring to different levels of quality. A cellar's own brand strategy influences its understanding of how the reputation should be improved. This is a more or less general feature of any wine-growing area, due to the heterogeneity of market demand (Patchell 2008).
Obviously, those who target the higher end of the market tend to prefer stricter quality prescriptions for the community. Moreover, there is a growing international trend, in which top producers tend to believe, that each community should focus on grape varieties that are historically typical of the area (Johnson and Robinson 2006). In this respect, Tihany and Csopak differ significantly. In Tihany, no grape variety has a particular historical distinction so producers are satisfied to agree that Tihany wine should be made of red grapes, chosen from a broad selection of grape varieties. By contrast, the role of Riesling in Csopak's reputation is a matter of dispute. Some are quite liberal-minded: 'If a producer thinks he can sell wine made of a certain grape variety, let him try'. Others are intransigent: 'When I saw a new plantation of Chardonnay, I almost cried; good Chardonnay could be produced anywhere in the world, a good Csopak Riesling only here.' If we compare the two communities, one important difference is that local family estates clearly dominate the group of medium-sized producers in Tihany, while non-local investors have a strong presence in Csopak. This implies, according to our explanation above, that in terms of the salience of collective reputation for producers (U1), Tihany is in a somewhat more advantageous situation than Csopak. At the same time, they are not markedly different in terms of the producers' discount rate (U3) or trust and reciprocity among producers (U4). As far as the common understanding of collective reputation (U2) is concerned, we detected no sharp divergence of views in Tihany but a clear lack of consensus in Csopak before 2010. This difference was partly due to the strong presence of non-locals in Csopak and the lack thereof in Tihany. More important was the unresolved dilemma of 'Riesling only' versus 'any grape varieties' in Csopak, while no similar debate plagued the community in Tihany. After 2010, the market success of Riesling advocates and professional discourse eventually led to the gradual convergence of views among quality-minded medium-sized wineries that Csopak's collective reputation should be based on high-quality Riesling wine. However, producers who target lower echelons of the market have not become part of this consensus.
Another favourable circumstance in Tihany has been the presence of a respected local winemaker who has been accepted as a leading personality of the community (U6) since the 1990s. He had worked as an engineer for the state farm and carried his reputation over to a successful family estate. No such 'natural leader' was found in Csopak up to the early 2010s. Only then did a young person appear on the scene and start to organise collective efforts. He is the heir of a medium-sized family estate, whose owner is best described as having 'one foot' in the locals' group (due to local ancestry) and 'another foot' in the investors' group (having accumulated his wealth from non-wine-related activities outside Csopak). Thus, he was able to act as a trust broker (Upton 2008) between locals and non-locals. His efforts became meaningful also thanks to the convergence of views on the role of Riesling among quality-minded wineries.
After 1990, various groups of local winemakers engaged in several civic and market initiatives with the aim of promoting collaboration and improving their collective reputation. As early as 1984, a 'Wine Knighthood' was established. Four 'wine route' associations, a 'circle of friends of wine' as well as some initiatives to promote local culture (including wine) for tourism followed. However, these organisations did not aim directly at quality assurance and tended to focus more loosely on the broader winemaking and touristic region of Balatonfüred-Csopak. They are best interpreted as small steps in the process of developing organisational experience in the local communities (U6) that could be partly transferred to building self-governing systems of quality assurance.
As the mushrooming of initiatives attest, winemakers enjoy freedom and autonomy (U5) in establishing civic associations. However, a national legislation established official wine communes (hegyközség) (by Act CII of 1994) and designated them as official representative bodies responsible for wine administration as well as quality assurance. In every wine region, each village or town was obliged to form a wine commune on its own or together with neighbouring villages (depending on their size). Although wine communes were conceived as self-governing bodies which could freely set their internal rules, they enjoyed only limited autonomy in drawing their own boundaries 7 and no autonomy in determining membership, which was obligatory for all wine-growers and merchants with vineyards above 1000 m 2 . Table 2 summarises the resource and user attributes that have characterised the two communities since 1990, highlighting the most important differences as well as changes in the attributes (italicised). Throughout the 1990s, the uncertainty of market conditions in the aftermath of the collapse of communism implied unfavourable resource attributes. It was unclear whether and to what extent the market value of reputation could be improved and how it could affect the flow of individual sales revenues. Reliable indicators of the reputation's market value were also largely missing. According to Ostrom's logic, these factors formed a serious impediment to self-organisation in this period. This suggests that no successful attempt at institution-building could be expected before the early 2000s.
Wine markets and individual market positions consolidated by the early 2000s so the crucial question became if the producer communities had characteristics that made them fit for self-governance. As we explained, Tihany was not lacking entirely in any of the Ostromian user attributes in this period. Despite some limitations, its circumstances were relatively promising. Its only enduring disadvantage compared to Csopak was its smaller size and the resulting more limited potential for increasing the market value of its collective reputation (R1). At the same time, Csopak was characterised by somewhat lower salience of collective 7 Wine-growers in the territory of each municipality could decide whether to form a separate wine community on their own or join producers in adjacent municipalities. However, the law prescribes a minimum wine-growing land size for a wine commune. Initially, this minimum was very low and virtually non-binding (50 hectares). In 2007, it was modified by legislation (Act CLV of 2007) to 300 hectares, which led to involuntary mergers of wine communites throughout the country. Tihany was obliged to join its larger neighbour Balatonfüred-Szőlős at this time.

Presence of favourable resource attributes
Feasible improvement of the market value of reputation (R1)

Reputation downgraded but survived under collectivisation.
Slowly emerging demand for high-quality wine, identified by terroir. Very small territory -limited potential supply.

Intermediate
Reputation downgraded but survived under collectivisation. Slowly emerging demand for high-quality wine, identified by terroir.
Indicators on present and future market value of reputation (e.g. predictions of demand, survey results, prizes, quality signals) (R2)

Intermediate
Gradual emergence of indirect signals, e.g. prizes, contests, national forums of wine evaluation.

Intermediate
Initial uncertainty in long-term changes in consumption trends but persistence of demand.

Yes
Communities in small geographic areas.

Presence of favourable user attributes
Salience of collective reputation for producers (in terms of income or identity) (U1)

Yes
Producers with strong local identity and economic reliance on winemaking.

Intermediate
Heterogeneous producers, incl. outside investors with no initial local identity and partial economic reliance on winemaking.  Low discount rate of producers for future revenues thanks to collective reputation (U3)

Yes
Producers with long-term goals of local production.

Yes
Very small, closely knit local community.

Yes
Atmosphere of trust despite different visions.

Intermediate
Freedom to organise civic associations. Limited autonomy for official wine communes. reputation for its producers, especially in terms of identity (U1) and clearly lacked in two critical attributes in the 2000s: common understanding (U2) and local leadership (U6). While the salience of collective reputation remained largely constant, the latter two factors changed around 2010 thanks to the convergence of views on the role of Riesling and the appearance of a young leader on the local scene. This suggests that Tihany already had the necessary qualities for setting up a system of quality assurance in the 2000s, while Csopak became capable of such action only in the early 2010s.

Institutionalising the management of collective reputation
The challenge of reviving the historical reputation of the two communities was already perceived by the newly founded private winemaking enterprises in the 1990s. However, no serious effort was undertaken in either group until the middle of the decade of the 2000s. This finding corresponds to our expectation that selfgovernance would not commence in this period because important preconditions of successful collective action were missing at the time. Most importantly, it took considerable time until the market potential of improved collective reputation became sufficiently predictable. Also in accordance with our expectation, it was in Tihany where a selfgoverning system of collective quality assurance was first set up in 2008. In that year, the winemaking community successfully applied for the right to use a 'Protected Designation of Origin' (PDO) label for Tihany wine and set up a governance system to regulate its use. A PDO is the name of an area or specific place used as a designation for an agricultural product or a foodstuff. If a name is legally defined as a PDO, it may only be attached to the product if the production process is located in the specified area and follows specific methods (Van Ittersum et al. 2007). In 2004, the Hungarian government adopted a new wine law (Act XVIII of 2004), which gave wine communes the opportunity to define a PDO for their produce and propose detailed rules for its enforcement. Tihany was one of the very few communities which took the opportunity. 8 The adopted regulation defines detailed conditions under which a winemaker may attach the name Tihany and the acronym DHC (Districtus Hungaricus Controllatus) to his wine. Grapes must come from plantations that are officially classified as first class within the cadastral land register of Tihany. Eligible grape varieties, a ceiling yield, a minimum sugar content at harvest as well as cultivation and processing technologies are defined. The enforcement of rules is delegated to the wine commune. An elected official of the commune (hegybíró) monitors the plantations, inspects the grapes before the harvest as well as their processing in the cellars. The official is accountable to the board of the commune, whose members are elected by and accountable to the general assembly.
The Csopak community did not seize this opportunity to set up its own system of quality assurance. There was no leader who would make the initiative, and there was no common understanding about the ways of improving collective reputation. The legal limits on the autonomy of the wine commune were more consequential in Csopak than in Tihany. The law required that all producers in the designated geographical area participate in crafting the self-governing system. However, the heterogeneity of views and interests as well as the lack of leadership precluded any meaningful consensus in Csopak. In 2009, the national government introduced PDO regulations by decree for all wine regions and localities that had not set up a self-organised solution by that time. However, these administrative rules did not aim at quality assurance. They were far too permissive to protect or improve collective reputation.
It took further four years until a subgroup of producers in Csopak embarked upon building a quality assurance system called Csopak Codex in 2012. Unlike in Tihany, it was set up outside the national framework that regulated PDOs. It still seemed impossible to convince all or even the majority of the wine commune to agree to a concrete set of quality prescriptions necessary for the improvement of collective reputation. Yet, a young winemaker could now act as an institutional entrepreneur relying on an emerging consensus among high-quality producers that Csopak should specialise in Riesling rather than a potpourri of grape varieties. The Codex was adopted in 2012 and first applied to the wines of 2013. Eleven wineries participated in the process and gained the right to use the label. This means that basically all cellars that aim at high quality were willing to subject themselves to the process of certification and bear its non-negligible costs.
Quality assurance is based upon the registered trademark of Csopak Codex. A bottle of wine may carry this sign if it has undergone a strict multi-stage process of control and testing. Grapes, solely of the Riesling variety, must come entirely from first-class vineyards that belong historically to the Csopak wine commune. A Producers' Committee is responsible for monitoring the entire production process and carrying out controls at four points. Its elected secretary and a representative of the municipality consensually select a Testing Committee of five independent wine experts. The municipality is also the formal owner of the trademark. Its involvement is mainly due to three reasons: (i) the use of the village name needs official acceptance (outside the PDO system); (ii) the municipality's involvement makes it less likely that a rival group of producers receives similar support; and (iii) the producers' group can use the municipality's resources in its 'fights' with national authorities.
The robustness of these recently crafted systems of governance remains to be seen. 9 However, their establishment clearly tracked the occurrence of conditions analogous to the resource and user attributes identified by Ostrom as favourable for the governance of natural CPRs. No institution-building effort was undertaken in either community until it acquired all of these attributes at least in a weak form. Tihany reached such a fortunate state sooner than Csopak. Therefore, it could embark upon successful institution-building earlier. Csopak was able to follow suit only when two gaps in the community's attributes -the lack of common understanding and the absence of leadership -were finally overcome.
Despite important differences, both systems of quality assurance became institutionalised as 'community governance' (Ostrom 1994). The self-organisation of the producer groups is backed by the broad legal framework of PDO regulation and trademark law, respectively, but its rules are defined, applied and enforced by group members and office-bearers accountable to them. It is worth noting that Provan and Kenis (2008) distinguish three governance forms of goal-oriented networks: those governed by (i) participants, (ii) a lead organisation, and (iii) a network administrative organisation. Our examples fall in the last category, in which a (rather simple) separate administrative entity is set up specifically to govern the network and its activities.

Conclusion
Collective reputation is an important type of collective good that requires the formation of a stable multilateral cooperative relationship among several autonomous organisations. We examined the efforts of two communities of private enterprises to maintain and improve their collective reputation by setting up selfgoverning systems of quality assurance. The underlying analytical structure of their challenge closely resembled the structure experienced by users of natural CPRs. Collective reputation can be viewed as an immaterial CPR, which requires tending and investment as well as rules to contain free-riding and depletion. Ostrom (1990Ostrom ( , 2005 identified a set of resource and user attributes generally conducive to the self-governance of natural CPRs. We conjectured and showed in our case study that the same attributes, properly reinterpreted, are also relevant for collective reputation. As long as they were not present, self-organisation could not commence. Once a community attained all necessary resource and user attributes, the institutions governing collective reputation were set up relatively quickly. The different timing of institution-building reflected the fact that the favourable conditions materialised at different dates in the two communities. The 'late-coming' group was clearly held back for a time by the absence of two attributes: (i) a common understanding of how collective reputation is established and actions affect it; and (ii) leadership. Once a leader appeared on the scene and common understanding emerged, institution building was jump-started with startling speed. The quality and robustness of the specific institutions warrants further research. Here, we focussed on the preconditions and process of their establishment.
The validity of our findings needs to be explored further. Additional case studies and their subsequent quantitative meta-analysis (Poteete and Ostrom 2008) can corroborate or modify the lessons learnt here. It is also worth considering that collective reputation can take many other forms. In our examples, individual enterprises (wineries) retained their individual brands but also introduced a common label. One can think of cases where independent producers market their products fully under a collective brand name. Alternatively, they may rely solely on individual brands while developing a non-branded form of collective reputation of trustworthiness or high quality. Do our findings also apply to such cases? Our theoretical framework suggests a positive answer but empirical verification is needed. Another possible line of inquiry would tackle other types of CPRs that may represent value for goal-oriented networks of businesses (Kilduff and Tsai 2003). A physical infrastructure or an information pool (Hess and Ostrom 2006) may in effect be a CPR, too. Again, does the analogy implied by theory actually show up in similar empirical patterns? If our study makes the claim more plausible that the links between the important but relatively uncharted territory of community governance among business organisations and the mature theory of natural CPRs are worth exploring further, we have not written in vain.