Sustaining business networks: Understanding the benefit bundles sought by members of local business associations

Local business associations can be important mechanisms for stimulating inter-firm cooperation leading to economic growth and development. However, previous research suggests that the unfulfilled expectations of their members can lead to low participation, high membership churn and network instability over time. As a departure from studies that have explored why local associations supply certain benefits and services, this paper draws on an original, demand side membership survey of local business associations to identify for the first time the bundles of benefits sought by members. Two bundles of benefits (instrumental and info-social) relating to thin and thick models of rational choice, respectively, are identified in explaining why firms join and remain part of associations. The relevance of these bundles to members was found to vary with business profile and length of membership, with the value of instrumental benefits reducing over time, whereas the demand for info-social benefits remained relatively stable. The findings have important implications for local strategies for sustaining business networks.


Introduction
Business associations have been described as powerful tools for fostering inter-firm co-operation and, as a result, economic development (Besser and Miller, 2010;Huggins, 2000). They come in many shapes and sizes. Some are sector/profession specific, others are more cross cutting in focus, and they range from formal chambers of commerce to informal business clubs. While some nationally organised business associations exist (such as the Federation of Small Businesses in the UK), most operate on a far smaller geographical scale and are locally oriented (Bennett and Ramsden, 2007). These Local Business Associations (LBAs) may provide a range of collective and individual benefits to members, facilitating business growth and strengthening social relationships between participants (Bennett and Ramsden, 2007;Huggins, 2000).
Previous research considers in detail the process of network formation (Lockett et al., 2013), the potential facilitating role of external actors or brokers (Huggins, 2000), the determinants of network success (Newbery et al., 2013), and network governance (Davies and Spicer, 2015) and evolution (Jack et al., 2008). The literature demonstrates that there are a range of motivations for joining small business networks (Lawton Smith and Romeo, 2013;Lockett et al., 2013). However, while attention has been paid to how benefits are provided (supply side), little consideration has been given to understanding the nature of, and reasons for variation in, the benefits sought by members. Such an understanding is critical if LBAs are to meet the demands of new and established members and, in so doing, survive and flourish. This is particularly important for LBAs as they often experience low levels of participation and high membership churn, with many networks suffering from instability over time (Curran and Blackburn, 1994;Phillipson et al., 2006). Evidence to date suggests that the expectations of members of LBAs often go unfulfilled, leading to them either becoming non-active or leaving the association altogether (Bennett, 2011;Laschewski et al., 2002). This paper analyses the benefits sought by members, drawing on a large and original demand side survey of local business associations. It begins by discussing the ways in which membership benefits have been theorised to date, focusing on 'thin' and 'thick' rational choice approaches, and the nature of 'benefit bundles'. The paper then details the approach to data collection and analysis. The latter draws on factor and cluster analysis to classify the main benefits sought, the composition of 'benefit bundles' and variations in the benefits sought by members. We then consider the significance of the research for the leadership of LBAs in meeting members' aspirations as a key ingredient for network sustainability.

Rational choice explanations for LBA membership
Rational choice perspectives suggest that a business owner/manager's decision to become and remain a member of a LBA is contingent upon a comparison of benefits versus costs (Becker, 1965;Olson, 1971). In other words, action follows from a calculation of consequences. There are two main versions of rational choice theory: thin and thick. In thin models, actors are self-interested and hold preferences which are stable and transitive (Hechter and Kanazawa, 1997). In pursuing their own interests, it is assumed that individuals possess perfect information and utilise this to calculate the best course of action to maximise their utility or in the case of small-business owners, profit (Goldthorpe, 1998). Thus, in thin rational choice models, pecuniary rewards drive entrepreneurial actions. Olson (1971) draws on the thin perspective to explain individuals' decisions to join groups. He argues that an individual's choice to join or remain a member of a group depends on his or her personal returns in relation to personally incurred costs, rather than any concern for the group's collective interest. Groups may therefore be vulnerable to free-riding -where individuals take the benefit of membership, but let others bear the cost. Olson (1971) argues that free-riding behaviour emerges after group-size crosses a size threshold, enabling a lack of contribution to be hidden from social censure or organisational sanction. Free-riding is most likely to arise in circumstances where the benefits of membership are not exclusive. Groups can provide a mixture of public and private goods or benefits. Public goods are non-exclusive and non-rivalrous. In other words if they are provided for one actor, all others can gain as well. For example, if a LBA successfully lobbies for improved car parking facilities in a town, all local businesses benefit regardless of whether they are members of the LBA or not. They are sometimes referred to as collective benefits (Bennett and Robson, 2001). In contrast, private goods are exclusive and rivalrous in nature. For instance, if an LBA provided accountancy training to members, it would be possible to exclude non-members from attending and therefore prevent free-riding. This is an example of a selective benefit (Bennett and Robson, 2001). Following Olson (1971) and thin models of rational choice, larger LBAs will only be able to exist if they provide selective benefits to members. Collective benefits, therefore, are only offered as a simultaneous by-product in a bundle of provision of selective benefits to members (Bennett and Robson, 2001).
Thin models of rational choice have been criticised as 'essentialist', suffering from an adherence to methodological individualism, inappropriate conceptualisation of human emotions and naı¨ve assumptions regarding perfect information and stability of preferences (Archer and Tritter, 2000;Zey, 1998). This has led to alternative conceptualisations, collectively labelled 'thick' models of rational choice (Goldthorpe, 1998). For instance, whereas thin models treat an actor's preferences as stable, so that a decision based on preferences made at one point in time will be repeated at another point unless the rewards and costs they face alter (Archer and Tritter, 2000), in thick models preferences are modified by interaction with social structures, which are dynamic, fluctuating over time and possess fuzzy boundaries (Johannisson et al., 2002). Evidence from Ostrom and Gardner (1993) indicates how the level of co-operation between individuals with a common interest exceeds that predicted by 'thin' approaches. Thus, Ostrom (1998) argues that bounded rationality and learnt norms of behaviour, or 'rules of thumb', contribute more to successful action than 'thin' models allow. In other words, behaviourally 'thick' models allow for core norms, such as reciprocity, trust and the collective interest, to become part of the explanation (Ostrom, 1998).
An alternative angle within thick rational choice theory focuses on ideas of limited rationality, which considers the difficulties of anticipating outcomes and implications of imperfect information (March, 1978). This recognises that individuals often find costs and benefits difficult to identify and isolate (Gill and Gill, 2012). Group members may make decisions and seek benefits that are based on imperfect understandings of the cost/benefit payoff. As such they may not aim to be optimisers, seeking the best payoff, but rather 'satisficers', seeking a satisfactory solution (Simon, 1991). In this case, rationality is bounded by imperfect information and individuals' capacity to calculate payoffs, so that they may well seek courses of action which are 'good enough'.
As a consequence, in the LBA context, whether through bounded rationality or the pursuit of non-pecuniary goals, thick rational choice models recognise that members may seek benefits that are not concerned primarily with making calculative profit maximisation decisions. In particular, thick models identify the importance of the varied and complex social motivations for individual behaviour and how decision making is also affected by limited information (Goldthorpe, 1998;March, 1978). Thus, while thin rational choice approaches conceptualise arms-length transactions, they may be a poor predictor of the outcomes of socially embedded exchanges (Uzzi, 1996).
Of relevance within this context of thick rational choice is the notion of strong and weak ties, referring to the nature of relationships that affect members' decision-making processes (Granovetter, 1973). While strong relational ties within a community may lead to greater solidarity and sense of local community belonging and support, they may also insulate the community from external influences leading to atrophy. If there is too much emphasis on strong ties, 'lock-in' effects are more likely that may be detrimental to innovation, adaptation and learning (Uzzi, 1996) and isolate businesses from regional, national and global opportunities (Oinas, 1997). Thus, 'strong' relational ties between members of a LBA may form the basis of a relatively stable and active membership core. While 'weak' relational ties may be characteristic of more transient relations involving a circulation of business owners moving in or out of membership or who participate infrequently in the life of the LBA (Cummings and Higgins, 2006). Over time some ties may strengthen or stabilise as interactions between members are repeated, while other ties may weaken and dissolve. Hence, a degree of membership stability may well be conducive to value creation over time, but too much stability may potentially be detrimental to knowledge creation, network dynamism and organisational adaptability (Dhanaraj and Parkhe, 2006). As such, the rationality of choice may change over time as a result of a dynamic and ever-changing membership make-up (Huggins and Johnston, 2009).

Benefits sought and benefit bundles
Individual entrepreneurs may therefore join LBAs with an intention to access certain benefits. In keeping with thick models of rational choice these benefits may be pecuniary or non-pecuniary, public or private, and selective or collective in nature, and are thus potentially manifold. They include access to new, or enhanced cooperation with existing, customers, boosting sales and profits. Other members of the association may be seen as future clients, sources of information about potential customers (Phillipson et al., 2006), or as a network of available services to enrich their own personal networks (Curran and Blackburn, 1994). Associations can offer access to promotional channels, for example, through collective marketing via web or print directories (Reilly and Szabo, 2005) or membership may enhance access to resources or lower their costs through relieving businesses of the need to source specialised services independently (Gulati et al., 2000). Knowledge acquisition and creation can also be vital (Van Cauwenberge et al., 2013). Membership may improve an owner-manager's awareness and understanding across business functions, for example relating to recruitment of personnel (Lowik et al., 2012). Other members can act as a sounding board for testing out new ideas. Social benefits include the opportunity to meet and mix with other business owners (Phillipson et al., 2006) and the creation of 'a sense of belonging' in a local community (Curran and Blackburn, 1994: 169). Lobbying and representation can occur where associations may act as a vehicle for engagement with local councils or public bodies (Bennett, 2011). Finally, associations may provide and maintain public goods, the benefits of which are available to non-members. Examples include improvements to the local physical environment, provision of car parks, the preservation of historic and public assets, crime prevention, and the attraction of funding for local initiatives (Medway et al., 2000;Reilly and Szabo, 2005).
Multiple benefits may be supplied by LBAs simultaneously as part of a service package or 'bundle' of benefits (Bennett and Robson, 2001). This may reflect an inherent indivisibility between benefits, for example, where customer networking benefits also bring reduced isolation and an opportunity to socialise, so that they cannot be delivered as completely independent benefits. They may, however, be deliverable as distinct bundles of benefits. Depending on the association, the relative importance of these benefit bundles may vary. Bennett (2011) found that while the dominant bundle in Chambers of Commerce related to service provision, in sector/trade associations it related to representation. What is less clear is whether and how the perceived relevance of sought-after benefit bundles varies between groups of members within LBAs.
The provision of benefits to members of LBAs has therefore been explained using two primary approaches. One follows Olson's (1971) logic of collective action and 'thin' rational choice models, which explore group behaviour through instrumental cost-benefit rationality. In contrast, 'thick' models incorporate bounded rationality (Simon, 1991) and learnt behavioural norms (Ostrom, 1998) and offer an alternative framework for understanding the benefits sought by members. The remainder of this paper draws on the findings of a membership survey to explore the degree to which these competing perspectives are able to account for the nature of the benefits sought by LBA members.

Methodology
Empirical research focused on 15 LBAs and their members, located in the north of England. This region's sectoral composition of firms resembles the UK as a whole (Atterton and Affleck, 2010). However, most economic and business enterprise statistics categorise the north of England (north east region plus Cumbria) as lagging the rest of the UK. For instance, the north east in 2013 had the highest rate of unemployment and the lowest gross disposable household income of any UK region (ONS, 2015). Cumbria also has above national average levels of unemployment and below average incomes. Birth rates of new firms (expressed as a percentage of active enterprises) consistently lag the national average in both the north east and Cumbria, by as much as 62% (ONS, 2015).
Specific LBAs were identified through an earlier, large-scale rural business survey undertaken by the Centre for Rural Economy (Atterton and Affleck, 2010) where a question asked respondents whether they were members of a local business association or network. Out of 830 usable responses, 46.5% reported they were members of general business associations, such as the Federation for Small Businesses, and 10% identified themselves as members of named LBAs. The research included all the LBAs identified at the time, although two were excluded as they were considered inactive by their leadership. The research also excluded larger Chambers of Commerce which, in terms of membership base and focus, had a regional rather than local remit. Those LBAs surveyed were attached to small towns, villages, and rural districts. By their nature these were small to medium in size, with an average of 58 members (max. 159).
LBAs were also classified as either having a homogenous membership, in the case of those dominated by a single sector such as tourism or retail, or heterogeneous, involving a mixture of sectors. While some LBAs may have been founded by local public bodies, all, with the exception of the Barney Guild, were supported by member subscription at the time of the research. Two tourism associations also raised revenue through additional fee-based services (Alnwick Tourism Association; Rothbury & Coquetdale Tourism Association). Table 1 describes the LBAs by homogeneity of sector, funding type and survey response rate. Environment and Planning C: Government and Policy 34 (7) A preparatory phase for the membership survey included observation of 20 meetings of LBAs and interviews with LBA leaders to gain an insight into the associations' history, objectives, and experiences. This also served to enhance support for the survey among associations and their members. Leaders were formally interviewed in person and via telephone and discussions held with them informally during membership meetings. In general, they regarded the survey as a positive initiative which had the potential to improve their association and raise its profile within the wider community.
The survey questionnaire was pre-tested through academic focus groups and a pilot survey of two associations. This received a response rate of 50%, a favourable figure compared against other comparable surveys (Atterton and Affleck, 2010). The pilot led to minor modifications to the survey design. All the members of the selected LBAs were surveyed with a postal questionnaire. To maximise response rates, a modified version of the 'tailored design method' (Dillman, 2007) was used. This thorough method is based on a theory of social exchange and advocates maximising contacts with potential respondents. As such, as well as the participant observation of meetings and leadership interviews, an advance letter, questionnaire pack, thank you/reminder postcard and, finally, a non-respondent follow-up pack were sent out in sequence. Where this strategy was not followed, due to association rules, the response rate was lower. For instance, the Berwick Chamber of Trade did not provide access to its membership database and sent out only the questionnaire, with no advance letters or reminders; consequently, the response rate was low (9%).
The questionnaire included a list of 11 potential benefits derived from the literature and validated by the interviews with LBA leaders. These included a collective voice; a contact network; new customer contacts; access to local knowledge; peer support; reduction of isolation; exchange of general information; an opportunity to socialise; a sounding board for ideas, issues, and problems; an improved reputation; and greater visibility to customers. LBA leaders regarded the list of benefits as encompassing the range of services sought by themselves and their members. Respondents were asked to rate the relevance of each potential benefit to them on a 5-point Likert scale where 1 ¼ not relevant to 5 ¼ very relevant. The survey also included questions designed to capture relevant personal perceptions relating to membership, business characteristics, engagement with the local economy (sales and supplies), and association performance.
The final response rate was 37% (n ¼ 313 respondents). Table 1 details the response from each participating association. Response rates varied from 9% to 57%. In part, this variation can be ascribed to differences between associations in the available strategies for contacting participants. There is also a small negative correlation between business association level response rate and individual years of membership (À.236 at 99% level), demonstrating that newer members were more likely to respond to the survey.
The final distribution of sampled members was skewed slightly towards more homogenous associations. Considering the sample as a whole, the median membership was 3 years and a mean of 6.8 years (spread between 0 and 60 years), with 43% of businesses describing their lifestage as growing, 48% as steady state, and 8% as in decline (with 1% unsure). Respondents were almost equally divided between males and females. Benchmark performance and specific feedback was provided to each of the 15 LBAs.

Survey findings
What benefits and benefit bundles do LBA members seek?
Members were presented with a list of benefits potentially provided by LBAs and asked how relevant they were to their business needs. Table 2 shows the mean relevance for each potential benefit. On average, all potential benefits listed were seen as relevant, with averages ranging from 3.01 to 3.85 The option 'exchange of general information' had the highest overall mean score for relevance (3.85), followed by 'access to local knowledge' (3.82). The least relevant potential benefit was perceived to be 'an opportunity to socialise' (3.01). Table 3 confirms that benefits are highly indivisible (correlated at the 99% level of significance). No benefit was entirely separate from another. The high degree of correlation suggested that the benefits were a candidate for exploratory factor analysis to identify the existence of structures or dimensions within the data. In addition to identifying data structure, a key aim of principal component analysis is a reduction in the number of variables (Hair et al., 2010). Using principal component analysis, two distinct bundles of benefits were identified, with a total variance explained of 66% and a data reduction rate of 80% (Table 4). The benefit of LBAs in providing a 'Contact Network' loaded equally across both factors and its removal achieved exclusivity of loadings with no decrease in significance measured by Cronbach's alpha. The two bundles are reliable measures of the

1274
Environment and Planning C: Government and Policy 34 (7) underlying structure of benefits, with high values for Cronbach's alpha .887 and .849, respectively (Table 5). Table 5 details the allocation each type of potential benefit between the two bundles with which they are highly correlated. All the benefits have a correlation of more than .350, which is the minimum magnitude of factor loading for the sample size (Hair et al., 2010). The sought-after benefits with high loadings for Bundle 1 relate to opportunities to socialise, reduced isolation and peer support, a collective voice, and access to local knowledge, information exchange and ideas. These benefits are sought by those members that want to be informed, supported and integrated socially. Members of this group also identified the  role of the associations in providing a wider collective voice within the local economy, as one business owner explained: All. . .activities are directed towards the overall improvement of trade through the promotion and improvement of [the town] as a trading environment. It is towards this wider goal that active members direct their efforts, frequently at their own expense and to the detriment of the amount of attention paid to their own business. (Member, LBA) We label this bundle Info-Social Benefits. Following thick rational choice theory, the benefits do not relate directly to an immediate commercial outcome, but rather to non-pecuniary outcomes and goals.
Bundle 2 relates to the benefits of membership for more direct pecuniary advantage, including visibility to customers, new customer contacts, and improved reputation. Following thin rational choice theory, these benefits are based on members' calculative reasoning and are sought by those that want to utilise LBA membership to generate financial returns to the business. We label these Instrumental Benefits.
Bundles 1 and 2 are strongly negatively correlated to one another 1 (À.504). This implies that an increase in one corresponds with a decrease in the relevance of the other bundle. For the sample overall, therefore, increases in the relevance of bundle 1 benefits are accompanied by a decreased interest in bundle 2 benefits and vice versa.
There was no significant difference between businesses at different stages of their life cycle and the relevance of either benefit bundle. However, length of membership had a significant impact on the relevance of instrumental benefits (bundle 2). In both heterogeneous and homogenous LBAs the relevance of bundle 2 benefits declines significantly with length of membership. In contrast, there are no significant differences in the relevance of info-social benefits (bundle1) by years of membership.

How do benefit bundles relate to business profile?
It is expected that businesses vary in terms of the benefits sought from LBA membership. To understand this further, cluster analysis was applied to define groups with the maximum homogeneity within the groups and maximum heterogeneity between the groups (Hair et al., 2010). Factor analysis preceded the cluster analysis since multicollinearity between the variables selected for clustering would bias the results. The factors formed the basis of the cluster analysis. The latter followed a two-stage approach. First, a hierarchical approach was used to discern the number of clusters and profile the cluster centres. Then, the observations were clustered utilising a non-hierarchical method (k-means) with the cluster centres from the hierarchical approach used as the initial seed points. This combined procedure allows one to take maximum benefit of the advantages associated with hierarchical and non-hierarchical methods, while at the same time minimising the drawbacks (Punj and Stewart, 1983). The clusters are validated based on a set of additional continuous and categorical variables. A three cluster solution was obtained: members seeking both instrumental and info-social benefit bundles (Dualbundle); those seeking mainly instrumental benefits (Instrumental); and those seeking neither benefit bundle but continuing to remain within membership (Habitual). Table 6 details the personal and business characteristics of each cluster. Statistical differences between clusters are explored using Chi-square and ANOVA tests. It is notable that every LBA contains a mix of these member profiles and that there is significant variance between them that is not explained by LBA homogeneity or funding type. Table 1 describes this variance. (1) Report lowest association performance (2) Lowest gains of customers and contacts (3) Most likely to gain nothing financially (4) More likely to be male (5) Less likely to run a home based business (6) Highest turnover (7) Least likely to purchase locally (8) Are the long-term members a A post-usage measure of member satisfaction, scored on a 5-point Likert scale, 1 ¼ very dissatisfied, 5 ¼ very satisfied. b Members estimated how much more they were willing to pay to save the association from a fictional imminent closure.

Dual-bundle members
Dual-bundle members account for 41.2% of the sample. Both instrumental and info-social benefits are highly relevant to this group. Firms in this cluster record the highest mean score for satisfaction with membership and are willing to pay the most to save the association from closure. They have gained the most new customers and useful contacts since joining. While they report only a moderate financial gain over the last 12 months as a result of association membership (mean ¼ »381, circa $596/445), this group is most able to attribute a financial gain to their membership of the LBA (61%). Since joining they are also the most likely to report an improvement in their businesses' performance and personal development, as well as perceiving that the LBA makes an economic and social contribution to the community. They are more likely to be female (accounting for 48% of all members of the cluster), previously employed in the public sector (51%) and running a home-based business (49%). The importance of info-social benefits is reflected in their commitment to the local economy, with 62% purchasing some products from within 30 miles, and in their higher levels of trust regarding local businesses and other members, and their view that the association is open to new ideas.

Instrumental members
Instrumental members account for 23.5% of the sample, for whom the instrumental benefits bundle is dominant. Its members report moderate levels for association performance and gains of customers and useful contacts. They are the group least likely to view membership as having led to no financial gain and report the highest gain over the past 12 months (mean ¼ »1195, circa »1868/1396). They are most likely to report a highly competitive business environment and their business premises are located further away from their homes compared to the other groups.
We identified no significant groupings of members who were interested solely in infosocial benefits. This suggests that while providing non-pecuniary benefits is necessary for associations in meeting the sought-after needs of their members, associations will be unlikely to sustain membership if they just seek to provide info-social benefits alone. In other words, while social benefits are valued by members in the Dual-bundle group, LBAs need to couple their delivery with more instrumental opportunities.

Habitual members
Habitual members, finding neither bundle relevant, accounted for 35.4% of the members 2 . The survey had included a comprehensive listing of benefits identified from the literature and in turn validated by the LBA leaders as being broadly encompassing. It is therefore unlikely that this finding is a result of not capturing another set of benefits that this group of firms was interested in. Members of this group tend to perceive weak association performance, low levels of new customers and useful contacts as a result of membership, the highest likelihood of gaining nothing financially (64%), and when they do the lowest level of financial gain over the last 12 months (mean ¼ »3, circa $4.7/3.5). They do not believe the association has improved their personal development or business performance since joining and barely agree that the association makes an economic contribution to the community. They are more likely to be male, less likely to run a home based business and this cluster registers the highest mean turnover. They are the least likely to purchase inputs locally and they are likely to have been members for the longest time (nine years on average). Overall, this cluster contains relatively large and mature businesses which have outgrown the local economy and association, perceive few benefits as being relevant, but which nevertheless remain members.
Further research is needed to explain the reasoning behind the continued and apparently habitual membership of these firms. One explanation, consistent with 'thick' rational choice models, is that exiting an association may engender social costs. Thus individuals may remain members out of a sense of duty (Hirschman, 1970), loyalty (Bennett, 2011), or local social norms (Atterton, 2007). They may wish to avoid the reputational risk of being seen by others as free riding on association benefits. Where members have made friendships and engaged together in collective action, leaving an association may cause emotional strain (Rothenberg, 1988). It is also conceivable that these firms may remain a member in the hope that eventually they will realise a return on their investment of time and resources in taking part in a network (Kahneman et al., 1990). This may be in expectation of some specific future benefits (Renko et al., 2012;Sarasvathy, 2001) or as a contingency in helping deal with future uncertainties (Fisher, 2012;Perry et al., 2012).

Implications for LBA leadership
The membership and benefit profiles identified suggest strategies that LBA leaders should consider in order to create the conditions that will sustain their associations. This is a major challenge facing LBAs, which typically experience phases of membership stagnation and inactivity. The findings suggest that a primary challenge lies in identifying and responding to the particular benefit bundles that are sought by member firms. In particular, meeting the needs of members who are interested in dual-benefits (both info-social and instrumental) is key. It is these firms that tend to be the most satisfied with, and also most likely to commit in the longer term to sustaining, the association.
In contrast, members who are solely interested in instrumental benefits appear to see diminishing value over time in their association membership, as pecuniary-based networking opportunities are exhausted, meaning that they may be the first to leave the association or become habitual members. To sustain instrumental benefits for these and dual-benefit members, a flow of new members is therefore necessary. While leaders believed their association to be welcoming of new members, few had an active recruitment plan. Leaders' time was largely spent on trying to persuade existing members to engage more, rather than on recruiting new members. Other than a web presence, the LBAs in the research did little to promote their services to the wider business community. However, the recruitment of new members has the spillover effect of delivering instrumental benefits to existing members who wish to build new business relationships. The survey findings also suggest that the targeting of female-led and home-based firms would be especially valuable priorities for recruitment, since they are more likely to be dual-bundle benefit seeking members and therefore more likely to commit for the longer term to the association.
On one level habitual members may appear problematic as they register the lowest levels of satisfaction and appear to derive few benefits from LBA membership. Understanding their needs and recognising their potential value to the association should be an important focus for LBA leadership. This group includes the largest firms, which typically have 'out-grown' the local economy, serving largely extra-local markets. Such firms may represent an untapped resource for LBAs in acting as a pool of mentors, and source of advice, knowledge and contacts for newer and smaller businesses. Small business owners have a higher regard for advice received from other, successful businesses compared to public sector agencies (Bennett and Robson, 2001).
Thus, while gaining few instrumental or info-social benefits themselves, their presence as members may be of significant value to other firms within the LBAs. They may therefore be an important resource for recruiting and retaining other members, as well as bringing reputational credibility and membership weight to an association's external profile and influence.
Finally, from observing these associations over a period of time, it is apparent that they lack formal feedback mechanisms from their memberships. The research suggests that the sourcing of regular formal feedback would benefit associations, by equipping them with an understanding of the bundles of benefits that are sought after by their members, how these may be differentiated across their membership base, their current levels of satisfaction, and how the LBA could be improved.

Conclusion
This paper demonstrates the need for a differentiated view of the sought-after benefits by members of LBAs. Such an understanding is vital in satisfying the aspirations of members and for inter-firm cooperation within networks to be sustained. The findings therefore have important implications for the future strategies of LBA leaders and the stability and adaptation of business networks in general, which can often struggle to maintain a supportive and active membership base.
For the first time, we have identified the benefit bundles sought by members of LBAs. From a large membership survey, two distinct benefit bundles have been ascertained (instrumental and info-social), linked to different rational choice approaches. Given a negative correlation between these bundles, there may be trade-offs in their delivery. Moreover, their relevance varies across LBA membership profiles and according to length of membership. Around two fifths of members (Cluster 1: Dual-bundle members) regarded both instrumental and info-social benefits as important. Info-social benefits are especially relevant for longstanding members. Other members prioritise instrumenta benefits (Cluster 2: Instrumental members). These benefits are likely to become less relevant over time, as members exhaust opportunities to meet and cultivate new customers. A final set of members seek few instrumental or infosocial benefits, but nevertheless remain within the LBA and may represent an important, but often untapped, resource for the associations (Cluster 3: Habitual members).
While provision of info-social ties plays a vital role in sustaining members' longer term satisfaction with LBAs, the research highlights the importance of dual provision (info-social alongside instrumental benefits). To deliver instrumental benefits and thus for this to be a sustainable approach LBAs require a continuous supply of new members. Without this, delivering instrumental benefits will be more problematic as opportunities for expansion of the customer base are exhausted. LBAs should therefore prioritise strengthening their external profile and recruitment of new entrants.
The paper has shown that while thin rational choice theory readily fits with the pecuniary motivations of Instrumental members, it does not describe Dual-bundle or Habitual members, where it is necessary to turn to insights from thick rational choice theory. In particular the profile of habitual members suggests that many find it difficult to identify specific benefits received but nonetheless remain longstanding LBA members. To understand the motives of this cluster, it is necessary to move away from the notion that pecuniary rewards to the individual, relative to personally occurred costs, drive all behaviour. The analysis points to a need for future research into how LBAs can effectively achieve an optimum balance of the membership profiles and benefits, appropriate to their local socio-economic circumstances, which can best enable them to be sustained and adapt over time.