Grown Local: Community Attachment and Market Entries in the Franconian Beer Industry

Defining communities

Communities have often been defined as geographic areas where organizations interact, gain access to material and symbolic resources, and contribute to the establishment of norms and traditions (e.g., Audia, Freeman, & Reynolds, 2006; Barnett & Carroll, 1987; Baum & Mezias, 1992; Freeman & Audia, 2011; Marquis & Lounsbury, 2007; Thornton & Flynn, 2003). As such, geographic communities ¹ make legitimate and replicable templates of action available to newcomers that facilitate market entry (Freeman & Audia, 2011; Marquis & Lounsbury, 2007; Wezel, 2005).

In this paper, we question the enabling function of geographic communities and argue that local communities may also exhibit obstructive effects for newcomers in the presence of incumbent organizations deeply rooted in their local community. Franconian communities fit this description, as incumbent breweries are traditional and closely connected to community members. An example from this context may substantiate our claim. Brauerei Seelmann was established more than 400 years ago in Upper Franconia. With a long history in the industry, Seelmann knows the most ancient traditions of producing beer. Age also gives Seelmann the chance to secure enduring connections with consumers and with suppliers of hops and other raw materials, who may support them in various ways for their long-lasting ties (i.e., consumers could be more forgiving when quality blips occur, and suppliers can generously offer price-related benefits). Seelmann is not only long-serving but also a rather small brewery that produces 1000 hectoliters yearly. With this output, Seelmann’s products are exclusive to a limited number of consumers and seasons of the year. Being small also gives them the opportunity to establish one-to-one relationships with their consumers. These unique ties with consumers gain special meaning, particularly when coupled with Seelmann’s ancient character. In addition, Seelmann’s products are distributed within a radius of 25 kilometers, which makes them accessible only to local audiences. This local character allows them to interact with inhabitants of the community where their production facilities are based and reach proximate towns. Seelmann’s localness coupled with their historic and small character has led to unique and tight relationships with consumers who possess idiosyncratic knowledge of Seelmann’s history.

This example suggests a few reflections about the conditions under which communities sustain entrepreneurial action. Particularly, it stresses the need to (i) focus on communities as composed of both organizations and residents and (ii) consider the extent of connectedness between these two groups as theoretically relevant. Building on these considerations, we conceive communities as a ‘local level of analysis corresponding to the populations, organizations, and markets located in a geographic territory and sharing, and that result from common location, elements of culture, norms, identity, and laws’ (Marquis & Battilana, 2009, p. 286). This definition highlights the degree of attachment of incumbent organizations to the local community as critical to the entry decisions of would-be entrepreneurs. Moreover, it allows elaboration on the different ways (i.e., positive and negative) both organizations and inhabitants may influence entrepreneurial action (Dorado, 2013).

The consequences of producers’ local attachment on market entries

In different realms, from agriculture to energy supply to urban or rural cohabitation, communities benefit from establishing an identity tied to the specificities of their locale (e.g., Tak, 1990; Knox, 1997; Sussman 1999; Wiesmann & Liechti, 2004; Newman, 2007). Establishing such an identity allows leveraging the local pride of community members and reinforcing their identification with the producers (see Knox, 1997). Incumbent organizations endowed with a local identity interactively reinforce this process and contribute to the establishment of the standards community members expect from new entrants (Porac, Thomas, & Baden-Fuller, 1989; Rosa, Porac, Runser-Spanjol, & Saxon, 1999).

Recent studies on communities have clarified that incumbent organizations provide ready-to-use templates of action for newcomers (Marquis, 2003). In the American banking industry, entry rates of local banks were found to be contingent on the local history of incumbent banks in the community (Marquis & Lounsbury, 2007). Communities in which local banks were acquired by large non-local banks triggered the emergence of new local banks. In this case, the replication of local characteristics by new entrants was a successful strategy potentially because of the risk of the community losing its local character.

Nonetheless, the local attachment of incumbents to their community may also represent a competitive constraint for newcomers. The more locally attached incumbent organizations are, the more difficult it becomes for newcomers to attract customers because of the lack of trust and awareness among local consumers. Note that our arguments about the competitive pressures exerted by locally attached incumbents are independent of the number of producers active in the focal locale. If a community hosts relatively few local organizations, even these few instances can lead to the formation of a robust attachment to the community. Our claim, in fact, is that the degree of local attachment of incumbents to the community hinders new entries because newcomers may have a more difficult time entering a market characterized by even a few local incumbents with deep attachment to the community than when facing a set of large organizations that display limited local attachment.

Being locally attached means being close to community members in historic and idiosyncratic ways by, for example, producing at a small scale and serving a spatially proximate portion of the market. Such a small scale of local production allows the development of personalized relationships with community members. These ties also gain special meaning (Bourdieu, 1985) when coupled with the traditional character of a focal firm. Repeated interactions between producers and the local community (see Rosa et al., 1999) are required for incumbents to become locally attached. Being small but lacking tradition may not suffice to reproduce a tight relationship between producers and community members, thereby making the process of local attachment difficult to attribute to non-traditional producers. In a similar vein, having a long pedigree is not enough for an organization to be locally connected to the community when such a traditional organization has developed into a large production and has served several geographic markets.

Because of historic and ongoing interactions between producers and community members (see Rosa et al., 1999), producers’ local attachment is also likely to influence the expectations of members of a community regarding the characteristics to be exhibited by legitimate organizations. In particular, community members exposed to locally attached producers will anchor their expectations to these producers and their activities (Barsalou, 1985; Hannan, Pólos, & Carroll, 2007; Rosa et al., 1999). Because of the long-lasting character of the interactions between producers and community members, these expectations are enriched by idiosyncratic and collective stories (e.g., stories about their founders, families, and the moment of founding), which augment their meaning (Bourdieu, 1985), create an emotional bond with community members (Greve & Rao, 2012; Oertel, Thommes, & Walgenbach, 2016), and feed their institutionalization (Marquis, 2003; Marquis & Lounsbury, 2007). Sentiments of regional patriotism and emotional engagement with those producers that embody the identity of the community may also emerge.

The reproduction of such relationships between incumbents and consumers is difficult for newcomers, since they lack tradition and enduring interactions with community members. Even when trying to replicate incumbents’ characteristics such as small production or local distribution, new entrants lack historical rootedness in the community. Transgressions to the tradition that incumbents represent can trigger social sanctions, as in the case of the clash between the botti and barriques in Piedmont (Negro, Hannan, & Rao, 2011), the higher prices Champagne houses had to pay to grape sellers when violating some of the industry codes (Ody-Brasier & Vermeulen, 2014), or the lower appeal that US microbreweries experience when not using traditional words in their labels (Verhaal, Khessina, & Dobrev, 2015). Entrepreneurs considering these communities as potential destinations for their business will soon realize that their lack of tradition and local attachment represents a liability. As a result, they will refrain from locating their business in these communities. Thus, the more incumbents are perceived as attached to their communities, the lower the likelihood of entry of newcomers will be because of the competitive pressures exerted by incumbents’ local attachment. We propose the following hypothesis:

The effect of residential stability varies with the degree of producers’ local attachment

So far we have focused on producers and their attachment to the local community to understand the hindering effect of geographic communities on market entries. However, our view of communities as geographic territories where residents and organizations interact (Marquis & Battilana, 2009) suggests that community inhabitants too have an important effect on entrepreneurship (Cobb, 2007). Residential stability in particular has been found to reinforce the bonds between community members and producers, while it also provides local organizations with a stable reference point for local values, norms, and traditions (see Carroll & Torfason, 2011; Sampson, 2012). Residential stability also influences the emergence and maintenance of local institutions by facilitating the construction of personal ties, interaction with neighbors, sharing of information, and the diffusion of values, norms, and information within the community (Sampson, 2012). For instance, they are likely to meet innkeepers of local businesses in social circles such as the church, schools, and other local events (Carroll & Torfason, 2011). Residentially stable communities do not necessarily imply that residents embrace the same values and preferences, but they are certainly more likely to exhibit similar expectations and appreciation for local producers (Sampson, 2012). Conversely, in communities with a high turnover of residents, such an appreciation of local offerings is likely to be much lower. For instance, people who have just changed their residence are often ‘fearful of the dingy looking “mom-and-pop” place’ (Carroll & Torfason, 2011, p. 9) and may prefer less-locally engrained options.

We advance however that these effects of residential stability on market entries are contingent on the level of incumbents’ local attachment to a community. In communities that are marked by a high degree of producers’ local attachment, residential stability should have a detrimental influence on newcomers’ entries. This is because in stable communities, the preferences of consumers for local offerings have been met by locally attached producers with whom enduring interactions have been developed (Rosa et al., 1999). Thanks to extensive opportunities for stable residents to come into contact with local organizations, they are likely to develop extensive knowledge of and shared experiences with local businesses (Carroll & Torfason, 2011). Since new organizations are not endowed with thick and enduring connections to community members by default, newcomers tend to shy away from communities characterized by high levels of residential stability and local attachment. We therefore advance:

A different prediction, however, emerges when considering communities that enjoy high levels of residential stability but host incumbents with limited local attachment. Because of residential stability, community values and traditions are maintained and community members may ease the mobilization of the required resources to support local entrepreneurial attempts. The absence of locally attached producers leaves community interests and preferences unmet. Newcomers who aim to establish businesses aligned with the local community are thus perceived as reviving local traditions (Carroll & Torfason, 2011; Wherry, 2006) and feeding the regional identity of the community (Paasi, 1986, 1991). Because of the unmet local preferences of stable residents, entrepreneurs are expected to enter communities marked by limited local attachment and by high residential stability. We therefore advance:

On Franconia

Franconia has 37 counties, or Landkreise, spread across three subregions that differ in terms of number of breweries, founding events, consumer expectations, and residential stability. Figure 2 shows a map of Franconia and the number of entries by county. Twenty-two foundings took place in Upper Franconia, 25 in Middle Franconia, and 28 new entrants opted for Lower Franconia.

OPEN IN VIEWER

Figure 2.

Entries by county in Franconia, 1989–2012.

Upper Franconia is home to 13 counties and to the industry cluster Bamberg. Around this city, there exists an agglomeration of mainly traditional beer producers: over 85% of the breweries in Upper Franconia are more than 100 years old (see Figure 3a). This cluster possesses the world’s highest brewery density and stands more than any other region in Germany for local and traditional production. The breweries in this region are small (yearly output is, on average, 21,600 hectoliters) and distribute their product mostly in hometowns and neighboring counties (within a radius of 65 kilometers). While non-traditional beer types are prevalent in all regions in Franconia, Upper Franconia offers highly typical brews that are rare in other parts of Franconia, such as Rauchbier (smoked beer). Consequently, traditional beers are an important part of the composition of the beer portfolio of breweries in Upper Franconia (see Figure 3b). Newcomers are typically small and family owned in Upper Franconia, and most of them offer food as well.

OPEN IN VIEWER

Figure 3.

Breweries’ characteristics: age (a) and beer portfolio (b).

The region of Lower Franconia offers a different context in terms of incumbent breweries and of their degree of local attachment. Beer is primarily brewed in the eastern part of the subregion, whereas wine production is common in the west. Although most of the breweries are 100+ years old, 41% of the producers in this region were founded in the last 50 years (Figure 3a). Of the foundings in Franconia 37% were established in the 12 counties of Lower Franconia. Beers like pilsner and wheat beer are produced here in quantity, while some producers engage in experimenting with ingredients such as honey, pumpkin, and rye. In fact, 76% of the beers offered in Lower Franconia are considered not typical of the Franconian beer industry (Figure 3b). Breweries tend to distribute their product over a radius of 80 kilometers and produce almost 24,000 hectoliters yearly. Newcomers are small and aim to serve mostly their neighbors (radius of distribution of 7 kilometers, on average).

The region of Middle Franconia has a strong influence from both Lower and Upper Franconia. In this region, approximately 60% of the incumbent breweries are 100+ years old; however, 40% of the producers were born less than 100 years ago (see Figure 3a), and approximately 33% of the newcomers in the last 24 years chose Middle Franconia to found a microbrewery. The 12 counties in Middle Franconia also offer diversity in terms of beer types. The region primarily produces pale beer, which is typical in Bavaria but less common in Franconia. One of the world’s strongest beers is produced in this subregion. ³ In total, 80% of the output produced by breweries in Middle Franconia is in non-typical Franconian beer types (i.e., bock, wheat, and pale beer; see Figure 3b). Breweries are larger in Middle Franconia (average yearly output is over 31,000 hectoliters), and newcomers tend to be slightly larger than in other parts of Franconia.

To better understand the variance in consumer preferences across the various regions of Franconia, we conducted a set of structured interviews using a questionnaire involving 1300 residents in different counties of Franconia. The interviews took place in beverage stores and supermarkets, where between 10 and 20 customers were approached in each store. We mainly focused on consumers’ preferences towards breweries of different size and community attachment to the local beer tradition.

Figure 4a suggests that 87% of the respondents in Upper Franconia felt favorably towards small breweries while also showing strong resistance to large and contract breweries. ⁴ In contrast, most of the consumers interviewed in Lower Franconia showed tolerance towards contract breweries (75%) and big breweries (71%). Consumers in Middle Franconia instead show generally strong preferences towards small and local breweries, while they are also somewhat open to large and contract breweries. Concerning community attachment to beer tradition, the survey revealed substantial differences in consumer preferences with respect to traditional beer bottles and to the sponsorship of local events. Figure 4b suggests that more than 50% of the consumers in Upper Franconia (56%) prefer breweries that play a role in their local communities by, for example, sponsoring school bazaars. Beer is consumed on a regular basis, and consumers prefer beers that come in traditional Franconian bottles (flip-topped, thick large bottles made of brown glass). As far as Lower Franconia is concerned, consumers exhibit limited interest in beers that are filled in traditionally Franconian bottles (30%), yet they view locally engaged producers favorably. In turn, nearly 50% of the consumers in Middle Franconia prefer beers that come in a traditional Franconian type of bottle, while only half of the respondents expect breweries to be involved in their communities.

OPEN IN VIEWER

Figure 4.

Consumer preferences for small breweries (a) and community attachment to beer tradition (b).

The mechanics and consequences of local attachment in Franconia

A first step in our empirical work was to grasp the organizational dimensions that affect the perceptions of local attachment among community members in our context. Qualitative data such as interviews with experts suggest that Franconian breweries are perceived as locally attached to their community when they are small, i.e., produce in small quantities; when they sell their products regionally, i.e., in geographic proximity to their location; and when they exist in the community for a long time. As one of our experts, ⁵ who works in a public office for fostering Franconian culture, said:

Franconian breweries are typically small businesses with a strong local character, particularly in Upper Franconia, where there are many microbreweries with their own inn and food […] Franconian beers are often not to be found in supermarkets or beverage stores but rather at their own production points […] Franconian breweries also play an important role in our regional identity, who we are in our towns, and we like to drink our local beer.

Two more experts reported that local attachment is often obtained through participation in community activities such as sports contests or local festivals. Participation in local activities and, therefore, rooting in the community is facilitated when breweries are small and exhibit an enduring experience in the community. As another expert reported in the media:

It is not without reason that we talk about ‘Bier braucht Heimat (beer ought to reflect its local roots)’. The oftentimes regionally limited distribution [of products] generates something like a certain [local] pride. (Jäschke Operational Media, 2014)

The view of these as well as other experts and brewers resonates with the view of the Association for the Promotion of the Franconian beer culture (Verein zur Förderung der Fränkischen Braukultur, 2016),

Franconia stands for a beer tradition that is ancient yet vivid and many-sided. Small and micro-breweries produce unique beer types here, like smoked beer, zwickel, and unfiltered cellar beer […] Multiple small businesses brew for hundreds of years only to serve customers in their own inn or at their own beer garden and beer cellars […] Without a doubt, beer has a more important meaning particularly in Franconia than in the rest of the [German] Republic […] Local attachment and rootedness is very important here.

According to this evidence, small-scale production is associated with satisfying local tastes, e.g., by offering specialty beers as mentioned above. This response to local tastes is something that mass production usually neglects, since it serves standard tastes that exist across different regions (Beck, Swaminathan, Wade, & Wezel, 2016; Carroll & Swaminathan 2000; Verhaal et al., 2015). The geographic reach of the products also matters because products that are available everywhere lose regional appeal. Note that the geographic reach matters independently of the size of the production, as even small producers that serve different regional markets can lose local appeal. Also, an organization’s age emerges as a relevant factor. The older an organization is, the more opportunities it has had to bond with the local population and to establish solid roots in the local community.

Having collected converging qualitative evidence about the relevant dimensions of local attachment, we screened local newspapers to develop a qualitative understanding of the community reactions towards newcomers. Some portions of the informal interviews with customers and master brewers revolved around this topic as well. The evidence we retrieved from these sources suggests that there is significant variation across regions; Upper Franconia, in particular, seems to impose more constraints to new entries. For example, Klosterbräu—a brewery founded in 2006 in Marienweiher (Upper Franconia)—closed down after six years of operation. Marienweiher exhibited high attachment from incumbent organizations and stable residential patterns. When interviewed by the local newspaper about the reasons for exiting the market, its owner said:

We did not have any support from the population. […] We could simply not satisfy many of the consumers in Marienweiher. There was only negative feedback the whole time. To live and let live is our motto. We did not get much of this [from the population], though. (Frankenpost, 2012)

Brewers in Upper Franconia appear reluctant to embrace new breweries. In alignment with its slogan, ‘stay local to your roots’, the master brewer of Seelmann distrusts new breweries that lack tradition. He says, ‘typical beer drinkers enjoy their regular brands [= breweries] and would be skeptical about new ones’. Likewise, when interviewing another brewer concerning new entrants in Upper Franconia, he said, ‘well, the new ones are not so enticing; nothing has evolved [with tradition]. They just open their brewery, but…’ When asked about new entrants in other parts of Franconia, he acknowledged that larger cities like Nuremberg in Middle Franconia—and the largest city in the cluster—could be more-welcoming arenas for new breweries:

probably the [new breweries] are an enrichment [there around Nuremberg]. However, this is not good in the surroundings of Bamberg.

A similar skepticism towards newcomers was also detected among local consumers in Upper Franconia. As a group of consumers in Bamberg told us about newly established breweries:

Community members in Lower Franconia exhibit a different view. Armed with the motto ‘a town without its own brewery is simply soulless’ (Nachrichten am Ort, 2012), founders can seize new opportunities in the local market. That is the case of Martins Bräu (Schonungen), whose entry into the market was described by a local newspaper thus: ‘When Ulrich Martin opened his brewery, he ultimately revived the tradition of our place’ (Schweinfurter Oberland-Kurier, 2009). In a similar vein, an entrepreneur who founded Pfarrbräu in Karlstadt explained, ‘people have enough of the big ones around [Lower Franconia]’ (Mainpost, 2010). He also claims that this gave him the chance to enter the market.

Both the qualitative evidence and the survey point to newcomers being confronted with particular skepticism in Upper Franconia. Since incumbent organizations in this region are usually quite small and possess an impressive tradition, we read this evidence as supporting our theoretical reasoning. Because each region exhibits substantial variation across counties, and because counties emerged as the key location of attachment for incumbent organizations, we resort to multivariate analysis at the county level to test our hypotheses.

The dataset

To capture new entrants, we consulted various archival sources, including (i) several editions of the ‘Fränkische Brauereikarte’ (Franconian brewery map), by Mack (1989, 1990, 1991, 1992, 1993, 1996, 1997), (ii) the first two editions of the Brauerei Atlas (brewery atlas), by Boris Braun (2004, 2010), and (iii) Verzeichnis der Hausbrauereien, by die Förderergemeinschaft von Brauerei-Werbemittel-Sammlern (2006, 2011). Seventy-five new entries were detected over the years 1989–2012. Demographic data on inhabitants were collected from the Bayrisches Landesamt für Statistik und Datenverarbeitung (Bavarian Agency of Statistics and Data Processing).

Methods and measures

Dependent variable and statistical approach

We constructed a panel dataset spanning 24 years (1989–2012). Our unit of analysis is the Franconian Landkreis or county. Our choice is motivated by the fact that communities (within each region) are the level at which local attachment becomes more relevant. Franconian counties are administrative subdivisions and autonomous bodies where decisions about local infrastructure and laws are made. By focusing on counties, we are able to account for intra-regional differences, particularly in terms of demographic, institutional, and economic environments.

We are interested in the likelihood of founding a new brewery in a given county in a given year. The final dataset consists of 864 yearly observations containing 75 entry events. ⁶ The dependent variable is coded ‘zero’ if there were no entries in a given county/year and ‘one’ if a county experienced one entry in a given year according to standard event history procedures. Since it is possible that a single county will witness several consecutive foundings over the same observation period, we estimated multiple episode models in which the clock for the first episode started to ‘tick’ with the beginning of the observation period (1989). When the founding of a brewery took place in a county, the clock was reset to zero and started to ‘tick’ again until the next founding happened. (In very few cases, only six, we observed more than one founding in a given county/year. In these cases, we set up an (artificial) episode of half a year between entries.) The dependent variable is modeled as a continuous-time hazard rate r(t), where t denotes the duration of an episode.

r ( t ) = lim Δ t → 0 Pr ( t ≤ T ≤ t + Δ t | T ≥ t ) Δ t

(1)

The hazard rate denotes the limit of the probability that a county witnesses an entry in the interval t, t+1. Since there are no clear expectations about the pattern of the hazard rate, we opted for piecewise constant exponential models. The duration of an episode was divided into eight different periods, in which the hazard rate was held constant even though it freely varies across periods. The length of such periods was defined as 3 years each, which sums up to 24 years of observation.

Independent variables

Local attachment

The extent of local attachment of incumbent breweries is the key independent variable of our study. According to our qualitative evidence, locally attached producers are those that are small, sell their products regionally, and have existed in the community for a long time.

The first dimension, scale production, is captured with yearly output (measured in hectoliters). We operationalize the second dimension, geographic reach of the products, with the radius of distribution of a brewery’s products (measured in kilometers). The third dimension is the organization’s age, measured in years since its establishment. It is worth noting that only one case of relocation across communities is observed in our sample.

We took the actual values of these variables and transformed them into standardized measures ranging between zero and one, so that values close to one represent small output and a low radius of distribution. ⁷ Age was instead log-transformed to indicate that its influence on expectations increases at a decreasing rate. That is, ancient breweries (300+ years) convey only a slightly stronger degree of tradition compared to 200-year-old breweries. We took the average score of these dimensions for breweries within the same county/year and then multiplied these values to capture a compound measure of the extent of local attachment to incumbent breweries. ⁸ The higher the score obtained from this index, the higher the degree of local attachment of breweries to a given county in a given year. In the results section, we present also the estimates obtained from each individual dimension as a robustness check.

Residential stability

The second independent variable of our study is the degree of residential stability that accounts for demographic patterns that may challenge the existing and lasting system of values of a community. As stated elsewhere, residential stability serves as a proxy for how stable the expectations of members of a community are (Carroll & Torfason, 2011). We operationalize residential stability by looking at population changes by county from one year to another. Thus, we first calculated the difference in number of inhabitants between year t and t+1, then took the absolute value, and inverted it to account for stability in residential patterns. ⁹

Control variables

Because our aim is to model the effects of community characteristics net of regional variations, we captured unobserved and time-invariant differences among Franconian regions by including two dummy variables: Upper Franconia and Middle Franconia. Lower Franconia was used as the reference category. We also controlled for calendar time to account for possible changes in market opportunities that might occur over time. Several control variables were then computed to account for county characteristics that may affect new entries. To ensure that our arguments on the effects of local attachment on market entry do not just correspond to density dependence effects, we controlled for the number of incumbents in a given year in a given county. Therefore, its linear and squared terms are taken into consideration (Density and Density squared). To further capture the competitiveness in the region as well as the potential availability of resources, we also controlled for the level of entrepreneurial activity in Bavaria by accounting for the lagged number of new businesses that were registered each year (Entrepreneurial activity Bayern). In a similar vein, the number of breweries that closed down in the previous year (Failures) helped control for availability of resources and competitive environment.

A few more measures are related to the potential demand for beer in the focal county. Qualitative interviews with master brewers revealed that Population is one of the most influential factors for the attractiveness of a market location in the Franconian beer industry. We controlled for the number of inhabitants who are at least 18 years old by county/year. This variable also helped control for county size and resource effects on market entry. We also controlled for the inflow of tourists in each county by looking at the number of booked rooms in each county per year. Touristic counties experience higher demand for beer and are therefore likely to attract more resources that support the entry of new breweries. However, a permanent flow of tourists also brings heterogeneity in the preferences producers should satisfy, and thus, they are likely to alter local community norms. We lagged these variables.

The extent of tradition of the focal community affects entries as well. We account for the degree of ideological conservatism by controlling for the proportion of people who voted for the conservative party, CSU (Christian Social Union of Bavaria) in the second vote, or Zweitstimme. Despite its name, the second vote is important in Germany because it determines how extensive the representation of a political party in the Bundestag will be. Hence, the percentage of second votes for the CSU indicates the extent of conservatism of a given county. We used a moving average of three years to account for consistency in people’s political behavior. As a second control variable for the extent of communities’ engagement, we use people’s engagement in community activities and volunteering work. We relied on the percentage of time individuals spend working for their community in each county/year (Volunteer time per person). We obtained these data from the Bayrisches Landesamt für Statistik.

Table 1 provides descriptive statistics of the variables included in our models. Table 2 presents descriptive information about new and incumbent breweries in terms of small output, degree of local distribution, and tradition. On average, newly founded breweries are small, with an output less than 700 hectoliters of beer a year, ¹⁰ and they serve their products within a 14-kilometer radius.

OPEN IN VIEWER

Table 1.

Descriptive statistics.

	Mean	S.D.	Min	Max	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Local attachment	4.22	1.23	0	6.17	1
2. Residential stability	−0.46	0.69	−14.22	0	0.15	1
3. Density	9.27	12.64	1	85	0.24	−0.05	1
4. Density sq.	245.54	927.49	1	7225	0.17	−0.09	0.94	1
5. Entrepreneurial activity Bayern (lag)	132.42	11.96	120.54	158.84	−0.12	−0.04	−0.03	−0.03	1
6. Failures (lag)	0.17	0.45	0	4	0.08	0.02	0.32	0.28	0.02	1
7. Log of population 18+ (lag)	4.4	0.45	3.39	6.07	−0.33	−0.4	0.12	0.11	0.08	0.01	1
8. Tourism: # of booked rooms (lag)	435.78	447.89	18.27	2558.3	−0.22	−0.23	0.004	−0.04	0.05	0.002	0.51	1
9. Prop. second votes for CSU (ma)	0.49	0.07	0.32	0.69	0.36	0.16	0.38	0.3	−0.26	0.1	−0.06	0.15	1
10. Volunteer time per person (%)	0.18	0.01	0.17	0.19	0.09	0.04	−0.01	−0.002	−0.22	0.04	−0.01	−0.04	0.14	1
11. Upper Franconia	0.35	0.48	0	1	0.07	0.12	0.38	0.28	−0.01	0.14	−0.29	−0.14	0.09	−0.002	1
12. Middle Franconia	0.32	0.47	0	1	−0.02	−0.17	−0.15	−0.13	−0.01	−0.03	0.24	0.04	−0.36	0.002	−0.5	1
13. Calendar Time	11.24	6.96	0	23	−0.12	−0.05	−0.04	−0.04	0.8	0.01	0.08	0.06	−0.3	−0.03	−0.002	−0.02	1

OPEN IN VIEWER

Table 2.

Small output, local distribution, and tradition for newcomers and incumbents.

	Newcomers				Incumbents
	Mean	S.D.	Min.	Max.	Mean	S.D.	Min.	Max.
1. Smallness (hectoliters)	647	684	3	2500	28’513	123’543	3	2’100’000
2. Local distribution (km)	14	30	4.5	245	65	200	3	5000
3. Tradition (age)	–	–	–	–	224	158	1	960

Additional analyses

We ran multiple sets of additional analysis to test the robustness of our findings. As far as local attachment is concerned, we ran additional models that estimate the effects of each single dimension of our composite measure. Tables 4 and 5 illustrate the effects produced by disaggregating the components of the local attachment index into its constitutive pairs (Table 4) and individual components (Table 5). We find consistency between our main set of analyses and most of our additional analyses. Particularly, the effects induced by smallness and local distribution (in Tables 4 and 5) when coupled to residential stability align with our expectations. The influence of age of incumbent breweries is the weakest, however. This unexpected ¹¹ result requires further investigation.

OPEN IN VIEWER

Table 4.

Robustness checks: pairwise combinations of local attachment dimensions.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Entries	Entries	Entries	Entries	Entries	Entries
1989–1991	−9.720*	−10.171*	−12.576**	−13.405**	−11.835*	−11.932*
	(4.685)	(4.778)	(4.727)	(4.797)	(4.767)	(4.753)
1992–1994	−9.897*	−10.291*	−12.929**	−13.680**	−12.243**	−12.334**
	(4.650)	(4.729)	(4.692)	(4.757)	(4.718)	(4.705)
1995–1997	−10.066*	−10.458*	−13.152**	−13.897**	−12.539**	−12.639**
	(4.746)	(4.823)	(4.745)	(4.810)	(4.819)	(4.805)
1998–2000	−9.466*	−9.859*	−12.620**	−13.354**	−11.966*	−12.063*
	(4.686)	(4.762)	(4.716)	(4.772)	(4.766)	(4.750)
2001–2003	−9.481*	−9.867*	−12.734**	−13.478**	−12.129*	−12.232*
	(4.708)	(4.775)	(4.713)	(4.765)	(4.819)	(4.806)
2004–2006	−8.948†	−9.293†	−12.434*	−13.146**	−11.555*	−11.669*
	(4.823)	(4.883)	(4.849)	(4.890)	(4.915)	(4.896)
2007–2009	−9.038†	−9.433†	−12.438**	−13.231**	−11.858*	−11.969*
	(4.771)	(4.833)	(4.789)	(4.842)	(4.952)	(4.936)
2010–2012	−8.090†	−8.484†	−11.809*	−12.531**	−11.801*	−11.929*
	(4.660)	(4.732)	(4.648)	(4.719)	(4.878)	(4.855)
Local attachment: size, age	−0.858**	−1.040**
	(0.161)	(0.176)
Local attachment: size, local distribution			−4.760**	−5.803**
			(0.794)	(0.792)
Local attachment: local distribution, age					−0.004	0.051
					(0.193)	(0.222)
Residential stability	0.969*	2.192*	0.840†	2.431**	0.842**	0.385
	(0.378)	(0.984)	(0.448)	(0.827)	(0.297)	(0.624)
Interaction: size, age * residential stability		−0.399†
		(0.228)
Interaction: size, local distr. * residential stability				−2.313**
				(0.838)
Interaction: local distr., age * residential stability						0.121
						(0.196)
Density	0.166**	0.172**	0.184**	0.189**	0.066*	0.065*
	(0.033)	(0.034)	(0.031)	(0.031)	(0.029)	(0.029)
Density sq.	−0.002**	−0.002**	−0.002**	−0.002**	−0.001*	−0.001*
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Entrepreneurial activity Bayern (lag)	0.031**	0.036**	0.033**	0.040**	0.039**	0.038**
	(0.010)	(0.010)	(0.010)	(0.010)	(0.011)	(0.011)
Failures (lag)	−0.469	−0.462	−0.457†	−0.458†	−0.441	−0.441
	(0.286)	(0.284)	(0.274)	(0.272)	(0.281)	(0.280)
Log of Population (18+) (lag)	−0.083	−0.158	−0.048	−0.153	0.491	0.502
	(0.365)	(0.364)	(0.355)	(0.365)	(0.402)	(0.402)
Tourism (# of booked rooms, lag)	0.0004*	0.0005*	0.001**	0.001**	0.001*	0.001*
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Prop of second votes CSU (ma)	4.622*	4.801*	2.840	3.151	−0.751	−0.714
	(2.282)	(2.280)	(2.177)	(2.199)	(2.350)	(2.349)
Volunteer time/person (%)	23.123	26.452	43.860†	49.047*	16.974	16.592
	(24.254)	(24.833)	(24.099)	(23.859)	(22.262)	(22.430)
Upper Franconia	−0.501	−0.469	−0.857**	−0.862**	−0.389	−0.400
	(0.354)	(0.354)	(0.310)	(0.305)	(0.349)	(0.349)
Middle Franconia	0.245	0.284	−0.061	0.028	−0.058	−0.060
	(0.368)	(0.365)	(0.352)	(0.345)	(0.308)	(0.306)
Calendar time	−0.061*	−0.062*	−0.036	−0.037	−0.057*	−0.056*
	(0.024)	(0.025)	(0.023)	(0.023)	(0.024)	(0.024)
Observations	864	864	864	864	864	864
N_fail	74	74	74	74	74	74
N_sub	110	110	110	110	110	110
Log pseudo-likelihood	−103.3	−102.0	−96.15	−94.33	−125.0	−124.9

Robust standard errors in parentheses.

**

p <0.01, * p <0.05, † p <0.1.

OPEN IN VIEWER

Table 5.

Robustness checks: individual dimensions of local attachment.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Entries	Entries	Entries	Entries	Entries	Entries
1989–1991	−12.777**	−13.564**	−9.267*	−8.846†	−14.440**	−14.678**
	(4.674)	(4.787)	(4.636)	(4.620)	(4.799)	(4.769)
1992–1994	−13.045**	−13.729**	−9.716*	−9.300*	−14.891**	−15.121**
	(4.637)	(4.746)	(4.571)	(4.553)	(4.749)	(4.721)
1995–1997	−13.281**	−13.972**	−10.010*	−9.591*	−15.224**	−15.467**
	(4.693)	(4.804)	(4.673)	(4.658)	(4.841)	(4.810)
1998–2000	−12.788**	−13.466**	−9.423*	−9.001†	−14.685**	−14.925**
	(4.650)	(4.747)	(4.611)	(4.599)	(4.783)	(4.752)
2001–2003	−12.952**	−13.649**	−9.542*	−9.117*	−14.937**	−15.184**
	(4.648)	(4.743)	(4.646)	(4.633)	(4.847)	(4.820)
2004–2006	−12.749**	−13.409**	−8.974†	−8.548†	−14.465**	−14.728**
	(4.782)	(4.867)	(4.727)	(4.711)	(4.953)	(4.918)
2007–2009	−12.817**	−13.563**	−9.201†	−8.779†	−14.789**	−15.046**
	(4.716)	(4.812)	(4.762)	(4.744)	(4.989)	(4.959)
2010–2012	−12.210**	−12.919**	−9.047†	−8.604†	−14.913**	−15.193**
	(4.612)	(4.720)	(4.635)	(4.623)	(4.886)	(4.852)
Size	−4.635**	−5.564**
	(0.791)	(0.800)
Local distribution			−2.974*	−3.606*
			(1.349)	(1.558)
Age (log)					0.299	0.376
					(0.209)	(0.234)
Residential stability	0.892*	2.473**	0.746*	1.945	0.812*	0.163
	(0.436)	(0.922)	(0.316)	(2.192)	(0.324)	(0.681)
Interaction: size * residential stability		−2.218*
		(0.965)
Interaction: local distr. * residential stability				−1.273
				(2.443)
Interaction: age (log) * residential stability						0.161(0.189)
Density	0.166**	0.169**	0.090**	0.091**	0.059*	0.059*
	(0.030)	(0.031)	(0.028)	(0.028)	(0.028)	(0.028)
Density sq.	−0.002**	−0.002**	−0.001**	−0.001**	−0.001†	−0.001†
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Entrepreneurial activity Bayern (lag)	0.031**	0.038**	0.040**	0.041**	0.039**	0.038**
	(0.010)	(0.010)	(0.011)	(0.011)	(0.011)	(0.011)
Failures (lag)	−0.453†	−0.454†	−0.436	−0.437	−0.449	−0.449
	(0.271)	(0.269)	(0.284)	(0.285)	(0.277)	(0.277)
Log of Population (18+) (lag)	0.289	0.168	0.188	0.170	0.687†	0.697†
	(0.345)	(0.363)	(0.359)	(0.367)	(0.383)	(0.381)
Tourism (# of booked rooms, lag)	0.0005*	0.001*	0.001**	0.001**	0.001**	0.001*
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Prop of second votes CSU (ma)	2.322	2.557	−0.452	−0.430	−2.023	−1.968
	(2.132)	(2.161)	(2.213)	(2.216)	(2.402)	(2.400)
Volunteer time/person (%)	40.581†	45.729†	23.169	23.982	22.406	22.188
	(23.560)	(23.553)	(23.147)	(23.130)	(21.427)	(21.447)
Upper Franconia	−0.617†	−0.594†	−0.640†	−0.657*	−0.427	−0.448
	(0.316)	(0.314)	(0.335)	(0.333)	(0.349)	(0.349)
Middle Franconia	−0.180	−0.080	−0.085	−0.085	−0.220	−0.223
	(0.335)	(0.325)	(0.280)	(0.280)	(0.318)	(0.314)
Calendar time	−0.029	−0.031	−0.060**	−0.061**	−0.046†	−0.044†
	(0.023)	(0.024)	(0.023)	(0.023)	(0.024)	(0.024)
Observations	864	864	864	864	864	864
N_fail	74	74	74	74	74	74
N_sub	110	110	110	110	110	110
Log pseudo-likelihood	−97.21	−95.54	−122.8	−122.8	−124.1	−124.0

Robust standard errors in parentheses.

**

p <0.01, * p <0.05, † p <0.1.

We deepen our analysis by recognizing that age alone may fail to capture ‘tradition’ unless it is coupled with the production of traditional beer types. Building on this premise, we coded the 100+ beer types offered by producers in Franconia and classified them into categories of typicality, i.e., how truly Franconian they are, with the help of seven experts in different regions of Franconia. We decided to maintain classification differences across regions to allow experts from Upper Franconia to classify beers in a different way than do experts in Middle and Lower Franconia. For example, experts in Upper Franconia see smoked beer as highly typical, while experts in Lower Franconia see it as almost typical (see Table 6 for further details on beer types within each category and values). We thus developed a scale of three categories: typical Franconian beer, almost typical Franconian beer, and non-typical Franconian beer. We assigned values between 0.1 and one to these categories, where a value of one stands for the highest degree of tradition and 0.1 for the lowest. We then asked experts to classify each beer type into one of these categories. Because some beer types did not match any of the above categories (e.g., organic beer) or conveyed tradition through their names, though their content was rather unclear (e.g., Fränkischer Urstoff, Old Rye), we created one more category, other beers, for these cases. In cases in which experts from the same region categorized beers differently, we opted to read online reviews from beer enthusiasts to resolve discrepancies. Later, we summed scores across beer types for the same brewery and divided this value by the number of beer types produced. We then took the average of the value obtained for a focal county in a given year. Higher values indicate a greater degree of typicality for the county (see Table 6).

OPEN IN VIEWER

Table 6.

Scale of typicality. ^a

Categories	Examples of beer types	Score
Typical Franconian beer	Dark, cellar, country, zwickl, unfiltered	1
Almost typical Franconian beer	Export, pale, pilsner, full	0.8
Other beers	Specialty grain, organic beer	0.5
Non-typical Franconian beer	Old, steam, diet	0.1

a

We maintained classification differences across regions to allow experts from Upper Franconia to classify beers in a different way than experts in Lower and Middle Franconia. Lager, smoked, beer mix, bock, and steamed are among the beer types that are seen in different categories across regions.

We then re-ran our main models upon splitting the data into two subsamples, one marked by low levels (typicality <0.45) and another by high levels (typicality >0.70) of typicality. As expected, the estimates of these analyses show that the amplifying effect of residential stability in deterring entries in communities with high local attachment of incumbents (Hypothesis 2a) is also traceable for the age variable when considering communities with high typicality. The whole set of results discussed in the present paper appears stronger in this subset of observations (results available upon request).