Emergence

Keywords: Social construction of markets; category emergence; category labels; proximate social space; geography; market segment

A key element of the social construction of market categories is the existence of a shared symbol in the form of a label (Berger & Luckmann, 1966; Hsu & Hannan, 2005; Scott, 2001). Labels enable communication and activation of expectations and evaluation of products, even without full observation. However, at initial stages of social construction – such as when an organization claims a label that is otherwise novel within a proximate social space – there is limited social understanding associated with the use of the label as well as substantial uncertainty about what social understanding may develop (Kennedy, Lo, & Lounsbury, 2010). We ask which proximate social spaces are most relevant for category emergence, focusing on organizations’ choices to be early claimants of a label relative to incentives and risks that arise within differently defined social spaces.

We frame the logic by viewing market categories as socially constructed institutions that become understood through social interactions in proximate social spaces. Scholars commonly view institutions as providing social understanding (Scott, 1995) and legitimacy (DiMaggio & Powell, 1983) while offering taken-for-granted status in defining customary “ways of doing things” (Nelson, 2005, p. 127). Similarly, categories provide structure and facilitate interactions in the market by helping consumers determine the legitimacy of products based on how well they fit with the expectations and understanding of a category (Zuckerman, 1999). Viewing market categories as socially constructed institutions provides conceptual suggestions for how processes of label emergence may begin and which dimensions of social space will be relevant.

For clarity, we outline several definitions, which Table 1 summarizes. Proximate social space is a limited subset of the environment within which an organization is likely to pay attention and interact; proximate social space can be defined around one or more dimensions, for example, by market segment, geographic region, or market segment within a geographic region. This idea arises in work on the creation of patterns in social structure among organizations (McPherson, 1983) and individuals (Feld & Grofman, 2009; Kossinets & Watts, 2009).

Concept	Definition
Proximate social space	A limited subset of the environment within which an organization (or person) is likely to pay attention and interact over time, enabling processes of institutionalization based on habitualization; it can be defined around one or more dimensions of social space.
Category	A taken-for-granted set of objects (e.g., products) in the market that are perceived as close substitutes because of their common attributes or function; provides actors with expectations and evaluative criteria for assessing products, actors, and interactions in the market.
Emerging category	A category that has not achieved taken-for-grantedness, though some consensus exists among some actors within a proximate social space about the objects or attributes that could be part of a common category. It may never become taken for granted.
Social understanding	The overlap actors within a proximate social space have about the sets of attributes or objects that they associate with an emerging category.
Peers	Organizations that create products that are perceived as sufficiently similar by some actors in the market such that they could be in a common category, even if that category does not yet exist.
Label	A linguistic symbol that provides a name for a category and can activate expectations or beliefs about the category once it is taken for granted.
New label	A name suggested for an emerging category.
Claiming a label	When an actor suggests a label for an emerging or existing category.
Early claimants	The initial organizations within a proximate social space that claim a label.

Categories are taken-for-granted subsets of the market (e.g., sets of products) that are perceived as being close substitutes because of common attributes or function; categories provide expectations and evaluative criteria for assessing products and actors in the market (Durand & Paolella, 2013; Hannan et al., 2007; Zuckerman, 1999). We refer to emerging categories as categories that have not yet – and may never – become taken for granted, yet have achieved some level of social understanding, which is the overlap in the sets of attributes or objects (organizations, services, or physical products) that various actors associate with the emerging category. Peers are organizations that create products that have enough traits in common that some actors could perceive those products as being in a common category, even if that category does not yet exist. The concept of peers is similar to competitors; the more neutral term is relevant for organizations that do similar things and want to be understood within a market and may choose to act collectively (we consider how competitive dynamics among peers may influence strategic choices during the emergence process).

Labels are linguistic symbols that provide names for categories (Hannan et al., 2007; Pontikes & Hannan, 2014). Labels offer actors closely associated with a category – whether as organizations, consumers, or complementary agents – a means to signal their understanding of the category and its norms while also enabling communication about the category (Berger & Luckmann, 1966; Granqvist, Grodal, & Woolley, 2013; Hsu & Hannan, 2005; Kennedy & Fiss, 2013; Scott, 2001). New labels are terms that actors suggest as names for emerging categories; we refer to the process of suggesting names for categories as claiming labels, as attempts to affiliate a label with a particular object or a set of objects. Early claimants are the initial organizations that claim a label in a proximate social space, often with reference to their particular product offering. We use these concepts to develop the idea of categories as institutions that are constituted in social spaces through interactions among social actors.

Because institutionalization of a market category requires establishing social understanding across actors, initiation of a process depends on the existence of a set of peers along with social interactions involving the peers and their products. Social interactions are likely to be concentrated in proximate social spaces because of limited time and attention, which may lead to local search by the focal organizations (Cyert & March, 1963; March & Simon, 1958). Further, proximate social space provides people with shared norms and culture, easing communication and interactions (DiMaggio & Powell, 1983; Scott, 2001). The result is that peer recognition and category emergence processes tend to occur within proximate social spaces.

While the argument for why proximate social space matters is intuitive, it is not obvious which means for delimiting social space is most relevant for category emergence, either theoretically or in particular contexts. Reflecting the nascent nature of this theory, we develop orienting propositions concerning strategic incentives and risks for organizations to be early claimants of a label for an emerging category. The first proposition focuses on the size of the peer group within an organization’s proximate social space. We then consider distinct characteristics of interactions in proximate social spaces defined by market segments or geography to develop a second orienting proposition; we expect the nature of and means for communicating in the face of deviation from existing norms will make early claims to a label more likely in proximate social spaces defined by geography than by market segment. The orienting propositions serve as the frame for empirical analysis of early labeling claims.

We examine the process of claiming labels for an emerging category in the context of a service in the U.S. nonprofit sector that is commonly referred to as fiscal sponsorship. Fiscal sponsorship is a service in which existing 501(c)3 nonprofit organizations serve as financial intermediaries, providing access to and accountability for tax-exempt sources of capital for entities that do not have tax-exempt status. Using archives of organizations’ homepages as well as data from Internal Revenue Service (IRS) information filings, we developed a longitudinal data set that associates organizations with proximate social spaces – defined by market segment (nonprofit mission, such as arts or health) and geography (metropolitan statistical area) – in which they may be likely to interact. We analyze the organizations’ choices to use the label “fiscal sponsor” on their homepage relative to characteristics of the organization and its environment. The analysis offers insights about the choice to be early claimants to a label within different definitions of proximate social space, thereby developing understanding of the dimensions of social space within which organizations act strategically during the early stages of category emergence.

This research contributes to the developing literature on the emergence of categories as part of the social construction of markets. We develop the conceptualization of categories as institutions and then use this logic to consider how category emergence processes are shaped by organizations’ strategic responses within their proximate social spaces. We demonstrate how the social environment creates incentives for an organization to take the risk of claiming a new label in a context of uncertainty about the potential for emergence of the market institution that they are trying to influence. Further, we highlight the importance of considering heterogeneous conceptions of social space, highlighting the role of proximate social spaces defined by market segment or geography in shaping strategic activity in the form of being an early claimant of a label for an emerging category and in responding to peers’ labeling claims.

BACKGROUND

Institutions, Categories, and Labels

We view categories as market institutions, in the sense that fitting a product into a category is a form of meeting and maintaining the socio-normative and cultural-cognitive expectations of an institution as part of the market’s structure. In doing so, we focus on the roles of organizations as strategic actors that seek to influence the social construction of market categories. This view helps develop an argument about the way the social environment – particularly the size of an organization’s peer group within proximate social space – influences organizations to act during early stages of the category and label emergence process.

Scott (2001, p. 56) defines institutions as “social structures that have attained a high degree of resilience” encompassing “regulative, normative, and cultural-cognitive elements that, together with associated activities and resources, provide stability and meaning to social life.” This conception emphasizes institutions as social constructions that change as human actors engage with and enact them. The literature on market categories aligns closely with the notion of institutions that are maintained through cultural-cognitive and socio-normative processes (Berger & Luckmann, 1966; Hannan et al., 2007; Scott, 1995). We focus on processes where taken-for-grantedness arises from socio-cognitive processes, not formal regulation; nonetheless, regulations in either of the social spaces we consider may represent preexisting norms that became formalized and may shape the understanding that actors take for granted.

In the view of categories as institutions, products that do not fit with expectations about a category – often new products but sometimes longer standing goods that fall outside existing categories – are likely to be overlooked or negatively evaluated relative to established categories (Leung & Sharkey, 2014; Zuckerman, 1999). Conversely, a product that is viewed as part of a category (i.e., consistent with existing market institutions) will be assumed to exhibit the default attributes of that category, even if the product actually deviates from these expectations (Hsu & Grodal, 2015). In this way, categories function as heuristics where products that achieve some level of observable sufficiency will be assumed to fit beliefs about a category more generally, even when characteristics of the product deviate from expectations in actuality. Parallels in the institutional and categories literatures extend to the roles of labels as symbols that signal and, in turn, align with norms of a category by using shorthand that has meaning to those who take the institution for granted (Hsu & Hannan, 2005; Meyer & Rowan, 1977; Pontikes & Hannan, 2014).

Hannan et al. (2007) and Scott (2001) each draw explicitly on Berger and Luckmann’s (1966) depiction of institutionalization as a process in which social interactions become habitualized, leading to social understanding and taken-for-grantedness in the form of expectations among those involved. Thereafter, the expectations can pass to actors who were not directly involved in the original social interactions, in part through communication with those who were involved. Emerging categories are in an intermediate stage of the emergence process such that there is enough recognition of peers and commonality across products to create some salience and social understanding, but not yet a label or taken-for-grantedness within the market.

During category emergence, labels may be particularly useful in disseminating understanding of the category. Information flows such as media mentions, word of mouth, and web browsing can vastly expand exposure to descriptions of products and exchange experiences, thus affecting the perceptions of the label and category (Kennedy, 2008; Kennedy et al., 2010; Navis & Glynn, 2010). Eventually these categories and their labels may become taken for granted and transmitted to people as a “way things are done” (reflecting Nelson’s, 2005 terminology) along with legitimating accounts of their use (Berger & Luckmann, 1966; Rosa et al., 1999; Wry, Lounsbury, & Glynn, 2011).

Strategically Claiming Labels to Influence Understanding

Despite their close relationship, a common difference in the theoretical depictions of emergence in the institutional and categories literatures is the extent to which producers may have agency in the process of recognizing and labeling sets of products and organizations as emerging categories. The institutional perspective (Scott, 2001) acknowledges that actors may strategically try to influence the development of social understanding for their own benefit. By contrast, the core theory of categories has typically described labels as being associated with categories by audiences – as opposed to producing organizations – based on observed attributes such that visibility of and matching to attributes is high (Hannan et al., 2007). Nonetheless, an emerging stream within the categories literature acknowledges that gaps between existing attributes of products and those that are observable creates opportunities for strategic use of labels (Granqvist et al., 2013; Hsu & Grodal, 2015).

When an organization claims a label, the label helps consumers identify the set of comparable products and indicates that the product should be evaluated according to the expectations of the category (Berger & Luckmann, 1966; Glynn & Marquis, 2004; Meyer & Rowan, 1977; Navis & Glynn, 2010). This can be particularly influential when product attributes are difficult to observe in full, because it allows consumers to use beliefs about the category in place of the unobserved attributes, thus facilitating evaluation and consumption decisions. For example, when a restaurant labels itself as Tex-Mex, it helps prospective customers decide whether that restaurant is where they want to have dinner, even though the particular attributes of Tex-Mex may vary based on the region in which the restaurant is located.

Firms also have opportunities to make claims to labels even when the product does not reflect a close fit with the expectations that may exist for a category more broadly (Granqvist et al., 2013; Hsu & Grodal, 2015). Thus, claiming labels is an important strategic activity because it facilitates information flows for exchange. This is particularly salient when exchange occurs without full observation of the products, though there may be variation in the extent of matching between the attributes of the product or organization and beliefs about the category.

However, the benefits to strategically claiming labels may be mediated by the extent to which actors continue to believe that use of that label conveys relevant information. As we will discuss later, attempts to use labels strategically during category emergence face added risks because of uncertainty about the meaning of new labels. Organizations considering early claims to labels need to assess the balance of potential benefits and risks for products that are not, but seek to be, part of a taken-for-granted category in the market. Before developing that argument, we discuss the nature of social interactions in proximate social space.

Patterns of Social Interactions in Proximate Social Space

People and, in turn, their organizations are boundedly rational, which may lead their interactions and problem solving to be focused within proximate social spaces (Cyert & March, 1963; March & Simon, 1958). Proximate social spaces are subsets of the environment that align with particular institutional contexts such that people are likely to have developed shared patterns of interaction and beliefs about the continuation of those patterns. These social spaces may persist as areas of focused interactions because of the shared beliefs and ease of interacting based on physical distance (e.g., geographic proximity), symbolic distance (e.g., professional proximity), or the plurality of networks that facilitate search processes.

The literatures on which we draw commonly focus on social spaces defined by market segments or geography, though often without full explanation for why one but not the other may be relevant for the processes being studied. The neoinstitutional literature historically emphasized industries and their associated market segments as key dimensions of social space in which institutions occur: organizations are likely to interact with the same sets of stakeholders, under the same regulations, with similar customers, and drawing from the same professional training programs, thereby leading to patterns of interactions and expectations that are passed on as the way things are done (DiMaggio & Powell, 1983; Greve, 1998; Haveman, 1993). Recent neoinstitutional work has considered geography as a means for defining social spaces in which interactions are likely to occur and lead to persistent patterns and shared expectations (Greve & Rao, 2014; Marquis & Battilana, 2009; Wry et al., 2010).

In much of the categories research, the emphasis has been on market segment distinctions such as coverage by analysts and identifying competitors through mentions in media outlets (Kennedy, 2008; Navis & Glynn, 2010; Pontikes, 2012; Zuckerman, 1999), although geography is sometimes employed as a boundary, such as types of restaurants within San Francisco (Kovacs & Johnson, 2014). While work on category fuzziness and hybrids has begun to consider where market segments may (or may not) break down (Alexy & George, 2013), investigation of market segment and geography as distinct means for defining an organization’s proximate social space has not occurred to date. Other related literatures have also tended to investigate these dimensions separately, such as entry timing studies that tend to emphasize market segments (Fosfuri et al., 2013; Mitchell, 1989) and international business research that emphasizes geography (Chakrabarti, Vidal, & Mitchell, 2011).

The primary arguments about the two dimensions of proximate social space are similar. Market segment and geography each define contexts where ongoing interactions and common histories can create shared expectations that provide social structure (i.e., institutions) for maintaining patterns of social interactions. Interactions may be more likely when people inhabit the same social spaces because of physical proximity or other shared interests, and there may be preexisting social networks that enable information flows. Further, interactions may go into more depth or require less effort because they can draw on shared cultural elements such as language and labels. Hence, social interactions may be more likely and meaningful in social spaces that are relatively proximate – whether defined by market segment or by geography.

Nonetheless, it remains unclear which proximate social spaces will be most important for social interactions needed for early-stage category emergence. Even as proximity in market segment or geography may ease interactions, they are distinct conceptions of social space and may differentially affect the emergence of categories and incentives for claiming new labels. We seek to understand how proximate social spaces defined by these two dimensions – market segment and geography – affect the process of creating social understanding for emerging categories and labels, whether as primary effects or as boundary conditions that shape joint effects. We focus on market segment and geography as dimensions for defining proximity in multidimensional social space because of their frequent use within the literature on which we draw; in the discussion, we address other potential dimensions of proximate social space.

EARLY CLAIMANTS TO LABELS IN PROXIMATE SOCIAL SPACE

Building on the concepts above, we develop two orienting propositions that emphasize the way that organizations assess the benefits and risks of strategically claiming new labels relative to their peers in proximate social space. Although current theory is not sufficiently robust to develop fine-grained hypotheses, the orienting propositions help guide the empirical investigation of how proximate social spaces relate to early labeling claims. The empirical investigation partly tests the base ideas and partly explores extensions of the conceptual base.

While the logic elaborated below focuses on the benefits and risks of being an early claimant of a label and how organizations respond, we do not suggest that this is the only step in the emergence process or that producing organizations are the only relevant actors. We focus on labels because of their importance as the symbols that name categories. However, for a category and label to emerge there must be a set of products to be categorized and peers producing those products. In focusing on label emergence, we assume that the relevant set of products and organizations already exist. Further, we assume that these products share sufficient numbers of attributes that if people (either those associated with peers or other actors) within their proximate social space are able to observe the products or interact with the peer organizations, there is a meaningful probability that the products may be viewed as similar to each other. We also recognize the possibility that early use of a label arises among external actors such as trade associations and government agencies; at some point, though, the organizations would be likely to respond and begin to claim the label as part of the category emergence process.

Benefits and Risks of Early Claiming of Labels: The Impact of the Size of the Peer Group

Products that are not socially understood risk being ignored by consumers or face higher costs because they require greater effort and energy to be understood (Zuckerman, 1999). Organizations that provide such products are likely to have an interest in cultivating understanding for what they do in order to reduce those risks and costs. As we discussed earlier, claiming a label is one action that organizations can take to influence how they and their products are perceived. However, the benefits and risks of claiming a label will differ when the claim is for a new label in the context of an emerging category. While use of a new label can become informative if other peers in the same proximate social space subsequently claim the same label, the understanding of that label is subject to the uncertainty of the social construction process, including the potential for negative valence (Kennedy et al., 2010). The question is what circumstances make the potential benefits of claiming a label and creating social understanding of an emergent category most salient and valuable.

We argue that the salience and potential value of claiming a new label for an emerging category relate positively to the number of peers within a proximate social space. When multiple organizations use the same label to refer to comparable products, repetition facilitates the process through which external actors such as customers come to understand the label as associated with shared attributes of those products and their salience as a possible category (Kennedy et al., 2010). This is similar to how firms’ identification of close competitors can facilitate external understanding and identification of a new category (Kennedy, 2008).

Further, once an organization has claimed a new label and gone through the process of explaining what it means to a variety of actors within a proximate social space, peers may find it less costly to use the same label. Peers may be able to take advantage of the nascent social understanding where their stakeholders overlap with those of the initial claimant by claiming that same label. Indeed, the opportunity for category emergence depends on the existence and recognition of peers, by the peers themselves and other stakeholders. Hence, organizations that want to create social understanding of what they do and recognize the potential for category emergence by recognizing peers within the social spaces they jointly inhabit may choose to participate in category emergence as early claimants of a label. Thus, early labeling claims are likely to be positively related to the size of the peer group within a proximate social space.

Benefits arising from collective or competitive mechanisms may drive organizations to be early claimants of a label relative to the peers in their shared proximate social spaces. From a collective perspective, organizations may recognize the opportunity to create a category and seek to induce social understanding of their product through the initial claims of a label for the category (Aldrich & Fiol, 1994; Navis & Glynn, 2010). From a competitive perspective, organizations may claim a new label with the goal of influencing the understanding that develops; organizations may try to create a strong association of a label with their particular attributes to ensure that as the label diffuses, their own product remains clearly within the emerging understanding of the category. This is a version of preemption, whereby the first organizations attempt to shape the understanding that develops (Lieberman & Montgomery, 1988; Mitchell, 1989). Either the collective or competitive motive may be sufficient to induce organizations to be an early claimant of a label if they believe that there is potential for category emergence within the proximate social space(s) where they interact.

Although organizations may benefit from claiming a new label, they also face risks due to uncertainty about if and what social understanding will develop for the label and category. The label may not diffuse, so that effort to claim a label is wasted. Even if diffusion does occur, there may be persistent ambiguity about the meaning of the label (Pontikes & Barnett, 2015). For example, if the products associated with the label are diverse, other actors might not recognize the similarities of the products and so social understanding of the category and label may not develop. In parallel, numerous labels may be used for comparable products, which may prevent people from recognizing the similarity of the products. Hence, social understanding may not develop, which will tend to have negative effects for products when consumers are market takers (Pontikes, 2012). Thus, variation in the labels used or the attributes of organizations and products associated with the label may prevent social understanding from developing. In turn, the lack of social understanding – whether due to limited diffusion or to ambiguity – may create confusion in the market and so constrain the success of the organization’s products.

Further, labels may become associated with negative valence (Kennedy et al., 2010). While negative valence could be limited to resistance to the label within the market (e.g., reactions from incumbents of related categories), it could be as severe as regulations prohibiting the product associated with the label (e.g., if perceived to be normatively abhorrent by dominant actors). An example is the current challenge facing app-based taxi companies such as Uber. Perception of the emerging category overall could also become negative, making not only the label but also the attributes of the emerging category subject to negative valence. Glynn and Marquis (2004) provide a relevant example in their study of how organizations changed their names after the dot.com bubble burst to disassociate from the negative valence of being an Internet-based business. The risks from negative valence depend on the possibility for extreme reactions, such as rapid change in the social or legal acceptability of the emergent category.

Just as the potential benefits from being an early claimant of a label are likely to correlate with the size of the peer group, potential risks also will tend to increase with the size of the peer group. When there are multiple products and peers, organizations can claim more different labels, which could undermine the opportunity for category emergence. Similarly, the more peers there are, the greater the likelihood that at least one will engage in deviant behavior that casts a negative shadow across the entire emerging category.

Nonetheless, the potential for category emergence does not exist for peers and products that exist as solo entrants. Taken-for-grantedness of a category based on habitualization and generalizations can only arise when there are multiple products across which commonalities can be identified. Further, the opportunity for recognizing peers and similar products increases when there are more of each within a proximate social space. This also extends to the potential for sufficient social understanding and cohesion among peers and their closest stakeholders such that there may be sanctioning within the group of social actors closest to the emerging category according to the developing norms and expectations. In this way, recognition and interaction not only enable development of social understanding but also the norms and sanctioning through which the peers may try to protect it. Thus, we expect that the potential for emergence of social understanding – including norms, sanctions, and expectations – is so strongly dependent on there being a peer group of some size that the benefits will tend to outweigh the risks when the peer group within a proximate social space is larger.

Our summary argument is that incentives to be early claimants of a label increase with the size of the peer group within a proximate social space. Peer density creates more interactions that increase the likelihood that any one peer recognizes the potential benefits of category emergence. Recognition of the potential for category emergence will relate to the choice an organization makes about claiming a label, whether motivated by competitive or cooperative dynamics. This logic leads to our first orienting proposition, which provides a baseline to explore how market segment and geography function as potentially relevant proximate social spaces.

Orienting proposition 1. The larger the peer group within a proximate social space, the greater the likelihood that a focal organization will be an early claimant of a label.

Relevant Proximate Social Space: Market Segment and/or Geography?

The argument so far focuses on the increased benefits of being an early claimant of a label when there are more peers within a proximate social space, which makes it more likely that people are exposed to and will develop understanding of the emerging category. However, differently defined proximate social spaces may differentially enable or inhibit these social processes. We now turn to the types of proximate social spaces in which these processes occur.

These arguments are premised on two assumptions. First, the product being brought into the market is sufficiently general that it has potential to be used across market segments and across geographies – for example, computers, telecommunications, and financial services. Second, we assume that given this level of generality, there is sufficient latent supply and demand within each proximate social space that there is potential for the set of peer organizations to succeed in creating a market category. Given these assumptions, we argue that some of the key distinctions between proximate social spaces that are defined by market segment and geography are factors affecting the ease with which people can interact in terms of the frequency, content, and opportunities for overcoming communication challenges.

Uncategorized products face difficulties because what they do and the value they provide are not taken for granted nor likely to be obvious, in part because there will often be conflict with existing expectations and beliefs about the market by at least some potential stakeholders (Rao, Monin, & Durand, 2003). While we have already discussed how the socio-cognitive perception of the potential category requires a set of products and peer organizations, peer organizations are not the only actors that matter for the emergence process. Organizations need to identify actors throughout the value chain, converting potential stakeholders into realized stakeholders who have sufficient understanding of the product to value it. This process is likely to depend on the frequency and breadth of interactions as well as the means for communicating; these factors may differ with the distinct dimensions for defining proximate social spaces.

Table 2 compares proximate social spaces defined by geography and market segment. The comparisons relate to the ease of interactions that may help either reach potential stakeholders or convert actors identified as potential stakeholders who may in turn become sources of information about the uncategorized product within the market. Identifying and reaching potential stakeholders will relate to the frequency and breadth of interactions, particularly when it is not clear who one might be searching for because of the latent nature of demand and supply of an uncategorized product. Further, the availability of diverse means of communication and the constraint from norms within the shared space will be important for the conversion of latent stakeholders. We initially consider market segment and geography as distinct dimensions for defining proximate social spaces.

Table 2. Factors That Vary between Proximate Social Spaces Defined by Market Segment and Geography and Their Anticipated Sign on the Effects on Social Interactions and Communication Patterns.

Colocation By	Market Segment	Geography
Frequency of interactions		+
Quality of interactions (means for conveying information such as body language, use of objects)		+
Diversity of interactions		+
Lower risks from interactions (less deviation from industry norms; fewer competitive spillovers)		+
Depth of interactions (shared language and norms from institutions in which colocated)	+

Geographic proximate social spaces offer frequency, quality, and diversity of interactions. Studies of regional clusters and agglomeration (Alcácer & Chung, 2007, 2014; Saxenian, 1996) suggest that geographic closeness facilitates identifying and connecting with stakeholders because it is easier to interact frequently via in-person interactions, whether by chance or intention. Personal interactions facilitate high-quality communication. In-person interactions allow people to use physical objects and gestures as part of their communication toolkits, which may facilitate their ability to convey information that is otherwise difficult for the receiver to understand (e.g., making abstract concepts more concrete through demonstration). Because in-person interactions tend to be more frequent among people who are physically colocated, we expect that this should be more relevant within geographic proximate social spaces. In parallel, the diversity of multiple market segments represented within a geographic space may make people more willing to try different narratives or framings as they search for common ground (Kaplan, 2008). These benefits will facilitate overcoming communication challenges that arise when exchange interactions lack preexisting social understanding.

By contrast, interactions among actors who operate in a common market segment may facilitate deep engagements, reflecting shared understandings about customers, suppliers, regulators, and other relevant actors. However, market segment interactions are more likely to be oriented around formal events such as industry conferences and the use of tools such as directories for purposive interactions with specific individuals, especially for actors who are not also colocated geographically. These needs will often limit at least the frequency and diversity of interactions and may inhibit the quality of interactions if there is limited time for engagement.

Moreover, deep interactions with actors within market segments may be most valuable once one knows what one is looking for. By contrast, early in coming to market, organizations are likely to focus on identifying exchange partners, suppliers, and customers. In these circumstances, chance interactions and stakeholder diversity may be particularly valuable. Geographic proximate social spaces will tend to facilitate these types of unplanned interactions, such as run-ins at a local coffee shop or grocery store or while picking up a child from school.

The risks also will vary across market segment and geography. Interactions that deviate from the existing norms of the proximate social space in which the interactions occur may provoke punishment by other actors within that space. Interactions about market-based exchanges, such as during the emergence process for a market category, will have particularly salient norms – that also are more likely to be violated – when the actors are in a proximate social space defined by market segment. Market segments share similar suppliers and consumers, so that problems that arise from a problematic attempt to claim a new label will be highly visible to market segment peers, who fear contamination of their own commercialization efforts. They may be particularly likely to draw on their social networks and other contacts to sanction labeling attempts that they view as improper, such as by fostering regulatory opposition.

Clearly, geographical patterns of interaction also pose risks due to deviation, particularly as the actors often engage in similar social circles. However, to the extent that actors within the same geography operate in different market segments, they may be functioning under different norms of market interaction. As such, there may be greater willingness to allow market-oriented experimentation in norms and language without attempting to impose sanctions. This suggests that the risks from deviating from existing norms of the proximate social space will be lesser in geographic proximate social space than that defined by market segment.

Further, the process of publicly claiming a label raises risks of losing valuable information to a competitor. Such risks may be particularly high early in the development of a market category, while norms and market positions specific to the emerging category are in substantial flux. These risks will be highest in the case of market segments, where peers tend to be rivals. By contrast, the risks of information loss may be lower in geographic proximate social space, owing to lesser direct rivalry and greater diversity in the stakeholder base.

The logic above suggests that early-stage emergence processes will be more likely to occur when proximate social space is defined by geography rather than market segment. Because random interactions are more likely, it may be easier to identify latent stakeholders within geographic proximate social spaces. Further, conversion should be easier within geographic proximate social spaces because of opportunities for overcoming communication barriers and lower risk of violating norms shared by the organization and potential stakeholders.

Orienting proposition 2. The size of the peer group within a geographic proximate social space will have a stronger relationship with an organization’s choice to be an early claimant than the size of the peer group within a market segment proximate social space.

Clearly, organizations exist in proximate social spaces defined along different dimensions concurrently, such as the presence of market segment peers within a geographic region. Hence, the intersection of market segment and geography also might define proximate social spaces. However, the causality within such jointly defined proximate social space is more complex than we are willing to theorize at this point. The joint space has the benefits of geographical proximity in the form of interconnected networks and market segments in the form of shared backgrounds that will provide shared language and practices. Within the intersection, interaction may be more likely and more in-depth, and thus increase the likelihood of opportunity recognition while also reducing communication costs. However, the duality also generates increasing constraints because of the depth of existing norms and expectations. When colocation is entirely based on geography, organizations and stakeholders may have some lenience due to their variation across market segments; by contrast, when proximate social space is defined by the intersection of geography and market segment, there may be much greater constraint because of expectations from both individual dimensions and their union, particularly if patterns of interaction among actors inhabiting that space predate the process of category emergence. This could increase the risk of penalty due to deviation from existing social norms. Hence, we will conduct exploratory analysis with proximate social spaces defined by the duality of market segment and geography.

Moreover, organizations will often prioritize different social spaces at different times or for different goals. For the purpose of identifying how interactions vary in differently defined proximate social spaces, it is useful to think of organizations at extremes: those that interact primarily within their geographic proximate social space and those that interact primarily within their market segment proximate social space. The argument so far applies to patterns of prioritization of social spaces, such that geography may be particularly salient in interactions even as organizations retain connections within their market segment, and vice versa.

EMPIRICAL SETTING: FISCAL SPONSORSHIP AS AN EMERGING CATEGORY

We investigate the orienting propositions in the context of U.S. nonprofit organizations that provide a service that is often referred to as fiscal sponsorship. We take advantage of the fact that nonprofits are associated with both a market segment (mission) and geography (typically metropolitan areas) to identify social spaces in which they may be likely to interact. We follow recent research practices on discourse and the value of written claims to focus on the patterns of adoption of the label “fiscal sponsor” on the homepages of organizational websites.

Fiscal sponsors are a subset of 501(c)3 tax-exempt nonprofits in the United States that participate in a practice wherein they collect and provide oversight for tax-exempt donations and grants to other charitable entities (referred to as sponsees) that are not incorporated as tax-exempt nonprofits with the U.S. IRS. Fiscal sponsors provide a mechanism for foundations and individual donors to make tax-exempt gifts to entities that would not otherwise qualify as charitable. This helps sponsees access capital that would otherwise be unavailable. While the practice is not officially defined or legislated, many practitioners and legal experts argue that the practice falls under the rules for nonprofits that distribute funds to non-tax-exempt entities. These rules require that the 501(c)3 nonprofit (the fiscal sponsor) be held responsible for ensuring that the sponsees use charitable funds consistently with the sponsor’s stated charitable purpose. Thus, fiscal sponsors help sponsees gain access to capital while providing oversight needed by donors and foundations (Colvin & Link, 1993).

Although the practice of fiscal sponsorship has existed since at least the 1960s, it has developed primarily as an activity that organizations practice, rather than an activity with clear regulations and rules. To date, fiscal sponsorship is not tracked by the IRS or state attorneys’ general offices. Even without regulatory institutionalization, though, the practice appears to be increasing in numbers of organizations providing the service (sufficiently visible to be counted) as well as awareness of the service within the nonprofit sector.

While the practice of fiscal sponsorship appears to be gaining understanding, salience in the social spaces where these organizations regularly interact will be critically important if fiscal sponsorship is to become taken for granted even within the nonprofit sector. All nonprofits have stated charitable purposes with at least one dimension of mission (e.g., arts, health, or environment) or geographic orientation (e.g., Boston). Nonprofits are required to engage stakeholders such as board members and donors. While individuals may engage with a nonprofit for numerous reasons, the general assumption is that there is some affinity with the organization’s mission. In the case of fiscal sponsorship, the sponsees must also align with the sponsor’s charitable purpose. The sponsees and sponsors should be attentive to the same social space(s), and the sponsors may have added reasons to focus their attention within that social space. For example, fiscal sponsors that enable independent artists to receive grants may find it valuable to maintain connections with arts funders not only for their own funding but also to help their sponsees access grants for their artwork. Thus, even though the service appears to be gaining broad understanding, distinct social spaces remain important.

These contextual traits provide an opportunity to examine the process in which organizations make early labeling claims relative to social spaces defined in terms of market segments and geography. Nonprofits with a common mission (e.g., education) will tend to engage with common foundations and donors, pay attention to the same professional associations, and recruit board and staff members with particular training and backgrounds. Hence, market segments (based on mission) offer a relevant dimension of social space for attention and interactions, leading institutions to arise and shape future patterns of interactions.

Just as fundamentally, these organizations function in physical space. Nonprofits that provide fiscal sponsorship are large enough that they have staff members. In addition to general interactions where people are located, nonprofits engage numerous stakeholders in their basic activities, including board members, staff, volunteers, funders, and customers. This physical embeddedness makes geography a dimension of social space within which interactions are likely to occur, so that institutions relating to these organizations may also develop within geographically localized contexts. Taken together, this suggests that either – or both – dimensions (geography and market segment) may direct actors’ attention and thus may define proximate social spaces in which interactions occur and institutions arise.

Further, the joint issues of labeling and social understanding are recognized by practitioners and experts of fiscal sponsorship. Among those who are identified as potentially being fiscal sponsors, there are a variety of terms used to reference their activities, including fiscal sponsor, fiscal agent, incubator, and technical assistance. There are also different models defined by distinct structural relationships between fiscal sponsors and sponsees (e.g., in some relationships the sponsee becomes part of the same legal entity as the fiscal sponsor, while others require the sponsees to remain separate legal entities). There are several labeling schemes for these different structural models. In addition to preponderance of labels, there is variation in the attributes (sets of services) that are frequently associated with the practice.

The fiscal sponsor label also faces ambiguity and risk. During a November 2013 meeting of the National Network of Fiscal Sponsors that one of the authors attended, there were extended discussions about the diversity of labels used to refer to fiscal sponsorship as a whole and to distinguish between models of the practice. There were also concerns about the extent to which lack of internal regulation enables organizations to claim any or all of the labels without verifying that they provide sufficient financial oversight to their sponsees. Practitioners expressed concerns that lack of oversight and professionalization may reflect badly on the emerging label, the practice as a whole, and particular organizations that associate with the emerging category. Some were concerned about the risk of IRS actions if self-regulation was insufficient to discourage malfeasance. Practitioners reported that some foundations changed their rules on giving to sponsees due to malfeasance by some practitioners that claimed the label.

Further, qualitative research presented at the conference reported inconsistency in understanding the label and the extent to which it applies to nonprofits affiliated with the practice. In one example, an organization that other practitioners in the room viewed as an ideal-type fiscal sponsor was said to be weakly associated with the label by a funder interviewed, even as that funder raised questions and concerns about working with fiscal sponsors (Prottas, 2013). Hence, fiscal sponsorship is a service about which some providers are concerned about stakeholder understanding and taken-for-grantedness and is an appropriate setting for studying the actions of organizations in the early stages of category and label emergence.

DATA AND METHODS

Data Collection and Construction

In 2005, legal experts and leading practitioners of fiscal sponsorship sought to assess the state of the field by identifying, for the first time, the population of practitioners. The results provide our base data set. The list was generated by employees of existing fiscal sponsors through several stages, including Internet research using key terms, verifying phone calls, and additions from the contemporary members of the steering committee for the National Network of Fiscal Sponsors. The list was designed to be inclusive of organizations engaged in the practice of fiscal sponsorship. Although inclusive, the list may be skewed toward organizations that fit with the notion of fiscal sponsorship as enacted by members of the national network at that time (i.e., that matched the set of attributes that were already associated with the practice); however, to the extent that social construction processes are shaped by perceptions of the actors involved, this remains a highly relevant sample for examining these processes. The set of practitioners identified in the sample are the set of peer organizations in this context.

Although the list was generated in part through web searches that included fiscal sponsor as a search term, the inclusion of additional search terms and the subsequent addition of organizations based on expert knowledge help to offset the risk of bias within the sample. Additionally, the data used to construct the dependent variable are the use of the fiscal sponsor label on the organization’s homepage, as we discuss later. This excludes use of the label on other pages or in the metadata that search engines use to identify websites. Thus, the dependent variable is likely to be more restrictive than what might drive general Internet search results.

Website Archives

We use website archives dating back to 1996 to collect data on the use of the fiscal sponsor label as organizations claimed it, using our retrospective knowledge to identify a term that has gained some traction as an emerging label. Websites are relevant conceptually and empirically as a source when considering processes that are organized around proximate social spaces. The increasing use of Internet technologies for making claims and sharing information that influences consumption decisions increases opportunities for, observability of, and reliance on linguistic symbols such as labels (Phillips, Lawrence, & Hardy, 2004). They also provide a relevant source for identifying the claims organizations make at a point in time. While many of an organization’s interactions occur in person or through other modes of communication and interaction, organizations make claims on their homepages that seek to convey information broadly. In our sample, this includes potential consumers and funders, as well as stakeholders of the sponsees. To the extent that websites are entry points for learning about organizations, labels used on homepages have high potential to influence subsequent impressions and interactions.

Although both targeted and broad communications are likely to play roles in how organizations convey information about what they do, communications that are broadly targeted may be more conservative – and therefore more consistently used in interactions – because of the increased potential to have a negative impact on a broader set of actors. The ease of recording and transmitting broad communications such as websites should increase the importance of such claims to the extent that they are intended to reach numerous actors and are more difficult to refute or deny, in contrast to the more flexible claims made by executives in Granqvist et al. (2013). Discursive work has increasingly argued for research using such types of actions (Phillips et al., 2004). Further, the availability of archived websites means that the process can be assessed as it took place over time without retrospective bias.

Assessing Potential Survivor Bias

Because the list is from 2005 but the data collection goes back to 1996, survivor bias may arise. We assume that closures that would have prevented organizations from inclusion in the 2005 sample but would have been part of the earlier social environment are randomly distributed across environments. The propositions predict a positive relationship between the proportion of peers in a given social space and an organization’s choice to be an early claimant of the label. If any survivor bias is more common in denser regions, then this creates conservative estimates of a positive impact. However, if any bias were concentrated in low-density environments, then they might be low density because of the closures. Because a positive relationship means that low-density areas should be those in which adoption is less likely, if there are fewer peers in the data than existed in the social space, it will make differences in density appear larger in the data than they were in reality, and lead to overestimation of the effect of the density variable.

Supplementary analysis assessed the risk of survivor bias by examining closure rates of nonprofit organizations incorporated prior to the list’s compilation in 2005. By examining correlations of nonprofit closures and the numbers and proportions of peers, we are able to demonstrate that there are no obviously problematic social spaces that would drive the significant results. The appendix reports procedures and tables.

Dependent Variable

The dependent variable is an organization’s first use of the term fiscal sponsor on its homepage. Data were collected using the Internet Archive’s Wayback Machine, which is an archive of web pages produced by web-crawling robots going back to 1996. We manually constructed website histories for each organization in the sample and collected the relevant data for the dependent variable from these websites. Data were collected for 1996–2013. First use fiscal sponsor_{i,j,k,t + 1} is a binary variable that is set to 1 if organization i, located in MSA, j, and part of segment, k, uses the fiscal sponsor label on its homepage in year t + 1.

Independent Variables

Each organization in the sample was matched with its federal employer identification number (EIN). Using this unique identifier, we collected data compiled from federal tax information and incorporation filings (IRS forms 990 and 1023) into the Business Master File (BMF). These files are collected and made available by the National Center for Charitable Statistics (NCCS). All independent variables, as well as data on geographical location and mission-orientation, come from these files. We collected data from 1995 to 2013.

Geography

The IRS data include an address for each of the organizations in the sample as well as metropolitan statistical area (MSA) codes (1999 definitions). MSAs are relevant as the level of aggregation for this study because they correspond to the types of social spaces in which geographic interaction among actors is most likely. One drawback of MSAs is that the classification system does not cover the full landmass of the United States. We fill in missing MSA data with county codes. While counties are less populous than MSAs, they represent a way of conceptualizing geographic space that enables us to include other organizations from the sample.

Peers_−i,j,t is the number of organizations that are identified as fiscal sponsors in MSA, j, and year, t, less the focal organization. Because this variable is likely to correlate with the size of the nonprofit sector in a given geography, we use a control for the number of registered nonprofits that are not peer organizations in the same MSA, j, and year, t: Registered organizations that are not peers: MSA_j,t. Proportion of organizations that are peers: MSA_−i,j,t is the proportion of the nonprofits that were registered with the IRS in MSA, j, and year, t, that were identified in the set of peers other than the focal organization. The proportion variable controls for the number of IRS-registered nonprofits while preserving degrees of freedom; this is the main conceptual variable when geography is the prioritized dimension of social space.

Market Segment

To identify the market segments in the sample, we use six classifications, five drawn from NCCS classes that held throughout the period (arts, education, health, human services, and other) plus community foundations. The six groupings encompass similar missions that may lead the organizations to pay attention to and interact with each other and common stakeholders.

We add community foundations to the five NCCS classifications based on information provided by informants about dynamics of this set of organizations. These organizations often name themselves in ways that identify them as community foundations (e.g., “The Community Foundation for the National Capital Region”) and have a unique identifier in the IRS classes. The foundations occupy a structural position between individual donors and nonprofits such that they reportedly receive requests to act as fiscal sponsors, even though the organizations often view the practice as burdensome and may not wish to draw attention and consumers to the practice. Sixty organizations (32%) in our sample are community foundations, so we want to address the possibility that these organizations have distinct dynamics and social norms.

An alternative classification would be to use more specific groupings that rely on classifications done by IRS and NCCS staff at various points in time. However, we lack knowledge of what information staff had access to or the criteria they used. Moreover, use of the more specific classifications is problematic empirically because the uneven distribution of organizations across the classes yields many organizations without a peer group.

In parallel with the measures used for geographic proximate social space, Peers_−i,k,t is the set of organizations identified as providing fiscal sponsorship within segment, k, and year t, excluding the focal organization. When we use Peers_−i,k,t we also include the corresponding number of registered nonprofits in market segment, k, at time, t, less the number of peers: Registered organizations that are not peers_k,t. Because of concerns about the degrees of freedom, we also consider the density of peers relative to the size of the nonprofit sector within that segment: Proportion of organizations that are peers: Segment_−i,k,t. This variable measures the proportion of registered nonprofits in segment, k, in year, t, identified as peers.

Control Variables

Several controls assess organizational and environmental factors. Small_i,t is a dummy variable corresponding to receipts reported on IRS forms (organizations with no financials = 1; those with reported receipts = 0). The data have an unusual distribution due to IRS reporting rules under which small organizations did not have to report their financials. For this reason, the distribution is bimodal even when we use the natural log of the receipts (after adding 1); hence, the dummy variable accounts for financial size. Having greater receipts may be associated with the number of connections with donors or customers, which may decrease the likelihood of using the label. On the other hand, being large may provide financial security to withstand risks, which may increase the potential that the organization claims the label.

We address two time effects. Age 2005_i is a time-constant variable created by subtracting the year the organization received its incorporation from the IRS from 2005. This controls for age independent of temporal trends. Having more experience may change the tendencies of organizations to be early claimants of a label; being older may be related to understanding emergence processes, the need to be understood by consumers, or attitudes toward risk and change. Year_t accounts for the temporal nature of this process. Although it might be desirable to use year dummies, with 18 years of data the loss of degrees of freedom would be substantial.

Each analysis controls for the proportion of peers in the alternately defined proximate social space that already use the label. Proportion of segment peers that use label_k,t is the proportion of peers in segment, k, that used the label in time, t. The variable captures the influence of prior use of the label within the market segment and also the number of market segment peers (in the denominator); this may have an influence even if organizations prioritize focus and interactions in geographical space. When social space is conceived of as prioritizing market segment, the corresponding control is Proportion geographic peers that use label_j,t.

Variables for Proximate Social Space Defined by Geography-Segment Intersection

The variables for analyzing proximate social spaces jointly defined by market segment and geography parallel those used in the other two sets of analyses. The focal variable is Peers_{−i, j*k,t}, to account for the other organizations identified within the set of peers within the same MSA-segment (e.g., arts organizations in San Francisco or community foundations in New York City). We take into account the number of nonprofits within social spaces defined in this way by using Registered organizations that are not peers_j*k,t. We also consider the effect when we incorporate the number of registered nonprofits within the MSA and segment as the denominator for the variable Proportion of organizations that are peers: MSA-Segment_j*k,t.

Several controls apply here. We control for label use in the same MSA and segment to separate the extent to which organizations pay attention to potential for category emergence among their closest peers (MSA-segment), or to realized label use in their segment or MSA more broadly. We subtract the number of MSA-segment peers from the denominators of the controls for the use of the label within both the geographic and segment spaces to avoid double counting those peers in the regressions. These variables are the Proportion of peers in MSA using label_j,t,−j*k and Proportion of peers in segment using label_k,t,−j*k.

Methods

We explore how early claims to the fiscal sponsor label unfold when proximate social space prioritizes geography, market segment, or the intersection. This shapes our analysis, because different organizations colocate in the same proximate social space depending on the dimension around which social space is defined. In turn, this affects which organizations are identified as early claimants relative to their peers in a proximate social space. Figs. 1(a) and (b) depict how an organization that is an early claimant when one dimension of social space is prioritized is not an early claimant when a different dimension of social space is prioritized.

Fig. 1. (a) Early Claimants Relative to Segment-Based Peers; (b) Early Claimants Relative to Geographical Peers. Implications: Panels (a) and (b) demonstrate how early claimants vary based on which organizations are considered peers. For example, in panel (a), the arts organization in Philadelphia may not be considered an early claimant relative to its proximate peers when proximate social space is conceived as emphasizing market segment, but that same organization may be considered an early claimant when the proximate social space emphasizes geography, because it is the first organization within Philadelphia to make the labeling claim.

Our data is an unbalanced panel due to entries and exits, with organization-year observations that associate each organization with a geography and market segment. Because each analysis uses different subsamples, we have different numbers of claims as well as different sets of organizations and observations in each analysis. For the market segment analysis, there are 11 early claims (125 unique organizations in 6 segments, with 478 organization-segment-year observations; all six segments had at least one early claimant). In the geographic analyses, there are 35 early claimants (186 organizations in 94 MSAs, with 1,550 organization-MSA-years; 22 of the MSAs had at least one early claimant). In the joint MSA-segment analyses, there are 49 early claims (208 organizations in 145 MSA-market segments, with 1,883 organization-MSA-segment-years; 38 of the MSA-segment pairs had at least one early claimant). Online Appendix B reports summary statistics for social proximity based on market segment and geography.

Reflecting the binary dependent variable, the regressions use panel probits. Because we are interested in the early usage of the fiscal sponsor label relative to peers in an organization’s proximate social space, we structure the data so that organizations are included only until they are first observed to use the fiscal sponsor label or until two organizations have used the label within the particular proximate social space. Using the first two claims within each proximate social space fits with the concept of organizations engaging in novel activities relative to proximate peers while providing more statistical power than would be possible with single adoptions.

RESULTS

Social Space Defined by Market Segment

Models 1 through 4 in Table 3 report the results for when proximate social space prioritizes market segment. Model 1 includes controls; the only significant result is the proportion of peers within a focal organization’s MSA that already use the label, which is positive. Model 2 adds the number of registered nonprofits within the same market segment, which is positive and moderately significant. Model 3 separates the number of registered nonprofits within the same segment into those identified as peers and those that are not; only the variable for the registered nonprofits that are not peers is significant. Model 4 combines these two variables to increase the degrees of freedom; this corresponds to the visibility of peers relative to the set of organization that could be peers (as nonprofit organizations).

Table 3. Probit Analysis of First Use of Fiscal Sponsor Label Relative to Peers within Proximate Social Spaces Defined by Market Segment (Mission) or Geography (MSA).

The results do not support P1. Indeed, in Model 4, the results have the opposite sign, although this relates to the positive correlation with the denominator. The implication is that the size of the peer group in the same market segment has little or no relationship to an organization’s choice to be an early claimant of the fiscal sponsor label relative to those peers.

By contrast, the initial positive relationship between use of the label among geographic peers persists. That is, actual adoption by peers within a geographic region correlates to early labeling claims relative to peers within the same segment. We will return to this point later.

Social Space Defined by Geography

Models 5–8 in Table 3 report results for geographic proximate social space, producing stronger support for P1. Model 5 includes controls. Model 6 adds the number of registered nonprofits within the same social space; the number of nonprofit organizations within a geographic area has a moderate positive relationship with whether an organization is an early claimant of the label. Model 7 splits the nonprofits into peers and nonpeers: both are insignificantly positive, although the “number of peers” in geography variable nears significance, as predicted by P1. Model 8 includes the more parsimonious proportional explanatory variable, capturing the effects of both the size of the peer group and the number of nonprofits within a focal organization’s MSA; this has a significant positive value, consistent with P1.

The models that include our two versions of the predictor variable (Models 7 and 8) are not nested and so cannot be directly compared. Nonetheless, we can compare differences of each model relative to Model 5. This comparison suggests a greater increase in the explanatory power for Model 8 than for Model 7, providing support for the use of the proportion as the operationalization for measuring the size of the peer group within a social space.

Among controls, smaller and younger organizations are somewhat more likely to be early claimants of the label, especially in Model 8. Strikingly, the proportion of market segment peers that have already claimed the label is significant in all models.

Social Space Defined Jointly by Geography Segment

Models 9 and 10 of Table 3 use the size of the peer group in the MSA-segment combination. Neither predictor variable is significant. This may reflect small samples. Conceptually, though, the null results reinforce the primacy of geographically defined proximate social space. The proportion of peers already using the label in each dimension of geography and market segment is positive and significantly correlated to the choice of an organization to be an early claimant relative to the jointly defined proximate social space, though the estimate is much more significant for geographically defined peers.

Implications

Taken together, the results provide intriguing implications concerning both propositions. Proposition 1 (early labeling claims are more likely in proximate social space with larger peer groups) is rejected when proximate social space is defined by market segment, while being supported for geography by Model 8. Proposition 2 (geographic social space will have more impact on early claiming than market segment social space) is supported by the fact that we only see results when proximity is defined around geographical spaces.

The results speak to distinctions between incentives to be early claimants and incentives to respond to prior claimants. In all three sets of analyses, it is striking that organizations with larger proportions of peers in the nonprioritized conception of proximate social space that already use the label – whether defined in terms of segment or geography – are more likely to be early claimants. This may relate to greater exposure to the label, essentially providing awareness that then is disseminated by additional organizations in different proximate social spaces. Hence, even if only geographic interactions appear to provide incentives to be early claimants, it appears that organizations tend to react to revealed moves – whether from competitive or cooperative incentives – once other organizations have become early claimants.

Together, the results have strong implications for how geographic and market segment proximity affect label claiming over time, through a combination of potential and realized claimants. The initial impact appears to derive from potential claimants in geographic proximity: increasing density of potential claimants correlates with organizations choosing to be early claimants. In turn, the density of early claimants realized in proximate social space defined by either geography or market segment affects subsequent claimants of the emerging label. Hence, while the trigger for the process may be potential for category emergence within geography, market segment proximity shapes subsequent evolution of claiming labels.

DISCUSSION

This article seeks to shed light on how organizations strategically engage in the social construction of markets. We focus on the social construction of categories as structures in the market that affect exchange, highlighting an action organizations can take to influence category emergence: claiming a label. We draw on the literatures of categories and neoinstitutionalism to develop two orienting propositions that develop logic about how boundedly rational actors interact with their social environments as they strategically choose to claim a label as part of a category emergence process. We develop the concept of proximate social space, which facilitates considering the multidimensionality of social environments and how distinct definitions of social spaces shape social interactions. We assess our propositions and conduct exploratory analysis in the context of U.S. nonprofits that engage in the practice of fiscal sponsorship.

The orienting propositions predict a positive relationship between the density of peers within a proximate social space and the likelihood a focal organization will be an early claimant for a label, while expecting a stronger relationship when proximate social space prioritizes geography over market segment. Having more peers in a proximate social space makes it more likely that actors will be exposed to and recognize the commonalities of peers, and also makes category emergence more beneficial because the spillovers – for instance, in communication – will be greater. We expect the effects to be greater for geographic proximate social spaces because there are more opportunities for random interactions and overcoming communication challenges compared to proximate social space defined by market segment.

Empirically, we find an effect when proximate social space is defined by geography, but not for market segment. We also find a secondary impact: the number of peers that have already made a labeling claim in a proximate social space is associated with increased likelihood that an organization is an early claimant relative to its peers in the alternately defined proximate social space. The patterns suggest a two-stage process in which the trigger for claiming labels is the potential for category emergence within geography, while both geography and market segment proximity then shape the ongoing evolution of claiming labels in an emerging category.

Contributions and Future Research

The research makes three contributions: it provides insights into the role of organizations as strategic actors in the social construction of markets; it develops the notion of proximate social space, highlighting two relevant dimensions; and it advances the conception of categories as institutions that shape the market. Each of these contributions provides opportunities for further research, as does the nature of the empirical context.

Social Construction and Emergence

First, while it is broadly accepted that markets are socially constructed, how the social structures emerge remains unclear (Glynn & Navis, 2013; Kennedy & Fiss, 2013; Koçak, Hannan, & Hsu, 2014; Porac, Thomas, & Baden-Fuller, 2011; Vergne & Wry, 2014). Issues include actions that organizations take to influence market structures and outcomes. This article helps fill gaps in our understanding of the social construction of market emergence by exploring ways that organizations respond to incentives and risks in their social environments, as well as considering different means for conceptualizing social spaces. We focus on one type of category emergence; there are opportunities to continue to parse existing research and to develop research that is more attuned to these key contextual distinctions.

This article focuses on the relatively straightforward case when products of the emerging category are sufficiently distinct from existing categories that efforts to create social understanding are in an essentially empty space. By this we mean that products in the emerging category do not have obvious comparisons or competitors in preexisting taken-for-granted categories. The organizations involved in trying to create understanding for what they do thus should be considering particular benefits and risks when deciding how to communicate and label their product. However, these environmental traits are unlikely to arise in some other instances of category emergence and so require additional theoretical work. It may be useful to classify types of emergence based on the decisions that organizations make relative to their environments that are likely to both shape and influence the benefits and risks that the organizations face.

Consider two key decisions organizations face when deciding to bring a product into the market and make labeling claims. Organizations choose: (1) to produce a product that is either similar to or different from existing categories of products and (2) to position that product either relative to an existing category or as a new type of product. Based on this, our logic applies to what is likely the simplest of these types of emergence: organizations produce something different from existing categories and they use a new label to try to influence the understanding of the product as distinct from other categories.

However, many products develop as recognizable variations on products that already exist. In these cases, there are likely to be distinct benefits and risks to claiming either the existing label or a new label, as well as additional layers of decisions such as the extent of differentiation or dis-identification from existing categories. Claiming an existing label while adding details about how a product differs could enable actors associated with the emerging category to more easily convey information to consumers by activating expectations about taken-for-granted products, but in such a way that it distinguishes them. However, the risk of making reference to the taken-for-granted category, even if in the process of disassociating, is that by activating that category’s expectations, there is potential for the deviation to lead to negative reactions by actors within the market. Conversely, for emerging categories that are closely related to an existing category, it may be more costly or difficult to identify relevant actors and communicate about the new – or different – products without invoking the name of the existing category. Future research could focus on differences of category emergence when products draw clearly on taken-for-granted categories (including hybrids) and those that arise in relatively uncontested (or sufficiently complex) spaces that links to predecessors are less identifiable or plausible. In the case of clear predecessors, there are further questions including when emergence occurs as differentiation (e.g., fully distinct from the taken-for-granted categories), sub-categorization, or replacement or updating of existing category definitions.

Beyond the earliest labeling claims, the way the process unfolds is ripe for research. One of the risks of claiming a label is that because labels and categories involve social processes, their meaning may change in unanticipated and undesirable ways, particularly as a function of actions by other actors within the market. For instance, labels such as “green” and “natural” were meant to distinguish products as different from the status quo in particular ways. However, as those labels began to be valued in the market, additional actors began to claim that those labels for products did not embody the intended attributes of the initial claimants. This use of the label by products and actors that do not embody the traits that are perceived to correspond to the emerging understanding of the label has affected the way those labels are understood. Thus, actors interested in influencing category emergence through claiming labels need to be concerned not only about the initial claims and understanding, but about maintaining the understanding as it diffuses, though how and when they may actually do this remains open.

Proximate Social Space

Second, we develop the concept of proximate social space in a more nuanced way than studies that consider social environments and social construction processes traditionally emphasize (Carroll & Swaminathan, 2000; Chakrabarti & Mitchell, 2016; DiMaggio & Powell, 1983; Fosfuri et al., 2013; Marquis & Battilana, 2009). While many literatures have addressed social environments in various forms, this has typically been done without clearly explaining why particular definitions of social space are focal, why others can be excluded, or when the intersection of the dimensions might be most appropriate. By developing the notion of proximate social space, we draw attention to the multidimensionality of the environments that organizations exist in, and the potential for people to pay attention and interact differently when proximate social spaces are defined in different ways. We focus on two dimensions of social space that are common within the literatures on which we draw: market segment (Carroll & Swaminathan, 2000; Fosfuri et al., 2013; Koçak et al., 2014; Kovacs & Johnson, 2014; Zuckerman, 1999) and geography (Alcácer & Chung, 2014; Chakrabarti & Mitchell, 2013, 2016; Chown & Liu, 2015; Marquis & Battilana, 2009; Marquis et al., 2013; Wry et al., 2010). In turn, we develop insights for why geography may be more important than market segment early in the social construction of market categories.

Market segment and geography are not the only ways to conceptualize proximate social space and the levels we use are not the only ways to operationalize these concepts. While we believe our implementation is intuitively relevant for our context, part of the value of the notion of proximate social space is that it is possible to develop additional insights and nuance about the general patterns of interaction and operationalize the concept appropriately for diverse contexts so that we can better understand general relationships between social environments, human and organizational interaction, and emergence. There are several opportunities to continue to develop our understanding of the multidimensionality of social space and how it relates to category emergence processes. These include – but are not limited to – the granularity at which proximate social spaces are conceptualized and operationalized, potential for intersectionality, and additional dimensions of social space. We discuss each in turn. One note is that underlying many of these opportunities is a need to continue to develop more nuanced understandings about whose perspective or interactions fundamentally shape these processes.

We operationalize our conceptions of proximate social spaces in ways that suit our context and data, though other ways for delimiting proximate social spaces arise even for geography and market segment. Geographies at the level of countries shape cultural understandings and beliefs, particularly as reinforced by national laws and regulations, and so we would expect patterns consistent at the level of countries and different regions within large or diverse countries. Similarly, missions within the nonprofit sector are one way of thinking about market segments, much like SIC codes, but these systems of coding often include nested hierarchies, so that there are choices to be made about which level of the nesting is most appropriate. Further, there are questions about the extent to which formal classifications align with the actual function of the objects or organizations being classified, as has been demonstrated in the case of patents (Kaplan & Vakili, 2015). Similarly, as the work of nonprofits, for-profits, and government organizations overlap in function (e.g., health insurance, microfinance, education), it is worth considering when it is appropriate to look within a formal type of organization (e.g., nonprofit) and when to look between organizational types because of common functions that may be more salient in shaping interactions. In many cases, this may be related to a question of perspective: whose perspective and classification matters in driving the process of interest, and where do gaps in perception affect market functions? Is it the focal organization, a regulatory body, or specific sets of consumers and other stakeholders whose perception and involvement most critically affect these processes? And to the extent that it may be a combination, how do the interactions work and change over time?

Although the intersection of geography and market segment did not influence early claiming in our context, it is an intriguing way of thinking about proximity in social space. In our setting, there are relatively few MSA-segment intersections that had multiple peers available, which means that there many early adopters would have only one or two peers in their proximate social space defined by the intersection. Hence, it would be difficult to observe a positive correlation of number of peers and being an early claimant, even if the dynamics worked that way; that is, the impact of relatively rare cases may be lost in the midst of larger sample analyses. As such, it may not be surprising that there is no effect within the MSA-segment space, because the number of peers in the other dimensions is easier to identify and interact with than those that are peers in the intersection. It could also suggest that the granularity of market segment is not sufficient: the same mission could include diverse groups such as performing arts organizations and visual arts organizations – and perhaps they are not sufficiently close to view each other as close substitutes. This all suggests that although we have done what we are able with the data for this context, it would be valuable to continue to push on these topics using other contexts, possibly some that are more robust in their numbers of early entrants, have clearer data tracking (e.g., recognized comprehensively by somebody earlier in their emergence process), are further along the emergence process such that the emergence patterns are more robust, or involve examining the impact of key organizations and individuals separate from large sample statistics.

While geography and market segment are commonly delimited ways for defining organizations’ proximate social spaces, perception and interaction are carried out by people. One way to identify additional relevant dimensions of proximate social space would be to focus on the traits of individual persons that shape their experiences and patterns of ongoing interactions as key parts of the institutionalization process. In the literatures about individuals, there are many ways of thinking about similarity in multidimensional social space, including foci of activities. This consideration highlights characteristics such as gender and race, as well as arenas of activity such as hobbies, as ways for thinking about where and how people interact with sufficient consistency to give rise to market categories (Feld & Grofman, 2009; McPherson, 1983). For example, people who race sailboats as a hobby may share language and patterns of interaction, even as they represent a variety of professions and geographies. They may develop new ideas in the social space where they interact through sailing to an extent that they would not otherwise achieve because of the constraints or inability to identify necessary but latent resources from professional identity or geographic locality.

Thus, there are multiple opportunities for research to determine what dimensions and level of granularity – both for operationalizing the dimensions and considering their intersections – are most relevant for category emergence processes. However, the potential for extensions comes with challenges. In particular, while the goal of developing the concept of proximate social space was to draw attention to the patterns that arise in social contexts that can be defined in a variety of ways, the application of the concept to infinite definitions of proximate social space may not be productive for elucidating the social contexts in which emergence is most (or least) likely and how organizations generally engage in those spaces. Nonetheless, it does suggest the potential to contribute to both specific and general understanding of characteristics of people and how they conceive of and interact in different social spaces, particularly when trying to shed light on how to identify and understand the most appropriate social spaces (in terms of both dimensions and levels of specificity) for continuing to develop our understanding of emergence processes based on interactions in social contexts. This suggests not only a need to consider the types of spaces that might be exist, but also the benefits of identifying the mechanisms through which social spaces differentially affect social interactions.

Categories as Institutions

Third, to help develop these insights, we articulate an argument about categories as institutions in the market. The literature on categories in markets is increasingly relevant as the broadest sense of the market continues to develop into a complex and cognitively taxing space. Exchange that requires people to make sense of all of information and products to which they are exposed would bring most exchange to a halt. It is for this reason that the categories literature is so powerful – it helps us think about how people develop understanding that spans individuals, enabling communication and the use of heuristics to make decisions and engage in the market (Durand & Paolella, 2013; Granqvist et al., 2013; Hsu & Grodal, 2015; Kennedy & Fiss, 2013; Kovács & Hannan, 2010; Marquis et al., 2013; Rosa et al., 1999).

The categories literature has articulated a process through which social construction occurs: many people and objects interact repeatedly over time until there are social beliefs about the sets of objects, and sets of attributes along with a label to name those sets and activate expectations (Hannan et al., 2007; Koçak et al., 2014). However, the emphasis has been on how audiences develop understanding based on what they perceive. As long as audiences are the primary functionaries in the social construction process, the perspective of organizations as strategic actors that are affected by the understanding that develops remains underemphasized, even as the role of organizations as strategic actors that take actions based on the existing understandings has become increasingly a focus of study (Granqvist et al., 2013; Hsu & Grodal, 2015). Extending the emerging interest in the categories literature on how organizations can strategically shape the understanding of categories as they emerge, ideas from the institutional literature (Scott, 2001) and their shared base in Berger and Luckmann (1966) has helped us to develop propositions about the way that these processes and social structures can be expected to emerge and may similarly provide a useful lens for others interested in the actions of organizations within emergence processes.

Implications of the Fiscal Sponsorship Empirical Context

Empirical research on emergence remains difficult, even as greater quantities and types of data become available, because at the earliest stages of emergence there is little indication of what eventual emerging categories may develop. We chose a context that gives us leverage on our interests concerning differently defined social spaces, though there are data limitations and opportunities to refine and extend the insights of this work through the use of additional contexts.

First, ideal research on emergence would have deep knowledge about peer organizations, labels that might emerge, and the actions organizations might take. We have addressed this by using peer organizations identified by an expert audience and focusing our research on actions around a single label in a context in which there is a category and label which appear to be emergent. There are opportunities for future research by including more comprehensive sets of labeling claims and other social construction actions that organizations might take.

Second, fiscal sponsorship is relevant conceptually and gave us leverage on the distinction between market segment and geography, but data limitations limit the scope of the study. Other contexts could provide opportunities for extended study, such as analyzing proximate social spaces defined by the combination of market segment and geography, as well as different levels of market segment or geography to more clearly determine if there are patterns regarding the levels of granularity for defining proximate social spaces that are most relevant (and when). Relatedly, there may be concerns about the extent to which geography is a dimension of social space that is likely to matter for other types of products and organizations.

This interest in the role of geography may seem oddly timed given the increasing use of communication across distance. However, as with recent research that has emphasized the role of geography in shaping and maintaining patterns of social interaction in meaningful ways, we suggest that the role of geography may persist much longer than we might be inclined to believe at first blush (Chakrabarti & Mitchell, 2016; Marquis, 2003; Marquis & Battilana, 2009; Marquis & Lounsbury, 2007). Further, to the extent that geography is often overlaid with important political and social boundaries, such as countries, but also smaller units like cities, the increased potential for interactions and persistence of patterns of social interactions that are historically informed are likely to be long-lasting.

To some degree, the most likely contexts for moving away from this geographical orientation may be those in which the numbers of people are so sparse that they are most likely to find each other using nongeographically oriented means, such as online communities of people who are affected by relatively unusual diseases. In these cases, the connections to people within a shared geographical space may be less likely, so that other dimensions of social space may be more relevant for shaping the likelihood of finding relevant actors and latent interest.

Nonetheless, to the extent that interpersonal interactions may be both more random and more able to overcome communication challenges through alternative means, it seems likely that geography will continue to affect emergence processes. This is particularly likely to be true when the product or service is consumed locally (e.g., professional services at a physical office, art or cultural experiences at a particular destination, or the purchase and consumption of goods such as food). This is likely to be reinforced by regulations that inhibit or enable new product entry. For instance, specialty food trucks have become popular, but they tend to be geographically limited because they are constrained by municipal regulations that prevent easy transfer and implementation of the new products across geographical areas. Even when geography might not be expected to matter as much, such as in venture capital, interpersonal ties and colocation have effects on raising capital to a degree that might suggest that even as the economy becomes more global and communication easier, geographically defined colocation and interaction is likely to persist in its relevance for the foreseeable future (Sorenson & Stuart, 2001).

Although this suggests that geography continues to be an important means for defining proximate social spaces, open questions remain. What level of geography matters for emergence through social construction and how does that vary for different types of emergence processes? How do formal regulatory institutions provide objective constraints to social construction processes? How do organizations and other actors affect the introduction and diffusion of new regulations that enable or inhibit emergence? What are the contexts or boundary conditions for when geography is likely to matter – or not – for emergence?

As is often the challenge with research on emergence, there are risks of left censoring and survivor bias. In our context, the practice has existed since the 1960s even though the coherence and use of the label “fiscal sponsor” is much more recent, which we determined via discussions with practitioners and searching resources on the subject. Patterns of censoring (e.g., MSAs in which organizations had websites and may have claimed the label before observation) correspond to areas in which claims were observed earliest and contexts where technology and websites may have been adopted early (e.g., around Silicon Valley). We have addressed these empirically and logically but recognize that additional tests in different contexts will be valuable.

Some challenges likely relate to the type of emergence: it may be more difficult to gather comprehensive data on categories emerging in relatively empty spaces, which lack complementary organizations or even source categories from which the entrants differentiate. These challenges also relate to the time frame of the emergence: some ideas take decades to develop and diffuse. When this is the case, even modern access to data and expanded digital archives may not be sufficient to wipe out risks of left censoring and survivor bias. Even so, these contexts may provide valuable insights, particularly when quantitative research is coupled with contextual understanding. While research in contexts that can get closer to causal identification is certainly valuable, there are opportunities to develop our understanding of category emergence in interesting contexts even when causal identification is limited.

REFERENCES

Alcácer, J., & Chung, W. (2007). Location strategies and knowledge spillovers. Management Science, 53(5), 760–776.

Alcácer, J., & Chung, W. (2014). Location strategies for agglomeration economies. Strategic Management Journal, 35(12), 1749–1761.

Aldrich, H. E., & Fiol, C. M. (1994). Fools rush in? The institutional context of industry creation. Academy of Management Review, 19(4), 645–670.

Alexy, O., & George, G. (2013). Category divergence, straddling, and currency: Open innovation and the legitimation of illegitimate categories. Journal of Management Studies, 50(2), 173–203.

Berger, P., & Luckmann, T. (1966). The social construction of reality. Garden City, NY: Doubleday.

Carroll, G. R., & Swaminathan, A. (2000). Why the microbrewery movement? Organizational dynamics of resource partitioning in the U.S. brewing industry. American Journal of Sociology, 106(3), 715–762.

Chakrabarti, A., & Mitchell, W. (2013). The persistent effect of geographic distance in acquisition target selection. Organization Science, 24(6), 1805–1826.

Chakrabarti, A., & Mitchell, W. (2016). The role of geographic distance in completing related acquisitions: Evidence from U.S. chemical manufacturers. Strategic Management Journal, 37(4), 673–694.

Chakrabarti, A., Vidal, E., & Mitchell, W. (2011). Business transformation in heterogeneous environments: The impact of market development and firm strength on retrenchment and growth reconfiguration. Global Strategy Journal, 1(1–2), 6–26.

Chown, J. D., & Liu, C. C. (2015). Geography and power in an organizational forum: Evidence from the U.S. Senate Chamber. Strategic Management Journal, 36(2), 177–196.

Colvin, G. L., & Link, G. (1993). Fiscal sponsorship: 6 ways to do it right. San Francisco, CA: Study Center Press.

Cyert, R. M., & March, J. G. (1963). A behavioral theory of the firm. Upper Saddle River, NJ: Prentice Hall.

DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48(2), 147–160.

Durand, R., & Paolella, L. (2013). Category stretching: Reorienting research on categories in strategy, entrepreneurship, and organization theory. Journal of Management Studies, 50(6), 1100–1123.

Feld, S., & Grofman, B. (2009). Homophily and the focused organization of ties. In P. Hedström & P. Bearman (Eds.), The Oxford handbook of analytical sociology (pp. 521–543). Oxford: Oxford University Press.

Fosfuri, A., Lanzolla, G., & Suarez, F. F. (2013). Entry-timing strategies: The road ahead. Long Range Planning, 46(4–5), 300–311.

Glynn, M. A., & Marquis, C. (2004). When good names go bad: Symbolic illegitimacy in organizations. In C. Johnson (Ed.). Legitimacy processes in organizations (Vol 22, pp. 147–170). Research in the Sociology of Organizations. Bingley, UK: Emerald Group Publishing Limited.

Glynn, M. A., & Navis, C. (2013). Categories, identities, and cultural classification: Moving beyond a model of categorical constraint. Journal of Management Studies, 50(6), 1124–1137.

Granqvist, N., Grodal, S., & Woolley, J. L. (2013). Hedging your bets: Explaining executives’ market labeling strategies in nanotechnology. Organization Science, 24(2), 395–413.

Greve, H. R. (1998). Managerial cognition and the mimetic adoption of market positions: What you see is what you do. Strategic Management Journal, 19(10), 967–988.

Greve, H. R., & Rao, H. (2014). History and the present: Institutional legacies in communities of organizations. Research in Organizational Behavior, 34, 27–41.

Hannan, M. T., Pólos, L., & Carroll, G. R. (2007). Logics of organization theory: Audiences, codes, and ecologies. Princeton, NJ: Princeton University Press.

Haveman, H. A. (1993). Follow the leader: Mimetic isomorphism and entry into new markets. Administrative Science Quarterly, 38(4), 593–627.

Hsu, G., & Grodal, S. (2015). Category taken-for-grantedness as a strategic opportunity: The case of light cigarettes, 1964 to 1993. American Sociological Review, 80(1), 28–62.

Hsu, G., & Hannan, M. T. (2005). Identities, genres, and organizational forms. Organization Science, 16(5), 474–490.

Kaplan, S. (2008). Framing contests: Strategy making under uncertainty. Organization Science, 19(5), 729–752.

Kaplan, S., & Vakili, K. (2015). The double-edged sword of recombination in breakthrough innovation. Strategic Management Journal, 36(10): 1435–1457.

Kennedy, M. T. (2008). Getting counted: Markets, media, and reality. American Sociological Review, 73(2), 270–295.

Kennedy, M. T., & Fiss, P. C. (2013). An ontological turn in categories research: From standards of legitimacy to evidence of actuality. Journal of Management Studies, 50(6), 1138–1154.

Kennedy, M. T., Lo, J. (Yu-Chieh), & Lounsbury, M. (2010). Category currency: The changing value of conformity as a function of ongoing meaning construction. In G. Hsu, G. Negro, & Ö. Koçak (Eds.), Categories in markets: Origins and evolution (Vol. 31, pp. 369–397). Research in the Sociology of Organizations. Bingley, UK: Emerald Group Publishing Limited.

Koçak, Ö., Hannan, M. T., & Hsu, G. (2014). Emergence of market orders: Audience interaction and vanguard influence. Organization Studies, 35(5), 765–790.

Kossinets, G., & Watts, D. J. (2009). Origins of homophily in an evolving social network1. American Journal of Sociology, 115(2), 405–450.

Kovács, B., & Hannan, M. T. (2010). The consequences of category spanning depend on contrast. In G. Hsu, G. Negro, & Ö. Koçak (Eds.), Categories in markets: Origins and evolution (Vol. 31, pp. 175–201). Research in the Sociology of Organizations. Bingley, UK: Emerald Group Publishing Limited.

Kovacs, B., & Johnson, R. (2014). Contrasting alternative explanations for the consequences of category spanning: A study of restaurant reviews and menus in San Francisco. Strategic Organization, 12(1), 7–37.

Leung, M., & Sharkey, A. (2014). Out of sight, out of mind? Evidence of perceptual factors in the multiple-category discount. Organization Science, 25(1), 171–184.

Lieberman, M. B., & Montgomery, D. B. (1988). First-mover advantages. Strategic Management Journal, 9(S1), 41–58.

March, J. G., & Simon, H. A. (1958). Organizations. New York, NY: Wiley.

Marquis, C. (2003). The pressure of the past: Network imprinting in intercorporate communities. Administrative Science Quarterly, 48(4), 655–689.

Marquis, C., & Battilana, J. (2009). Acting globally but thinking locally? The enduring influence of local communities on organizations. Research in Organizational Behavior, 29, 283–302.

Marquis, C., Davis, G. F., & Glynn, M. A. (2013). Golfing alone? Corporations, elites, and nonprofit growth in 100 American communities. Organization Science, 24(1), 39–57.

Marquis, C., & Lounsbury, M. (2007). Vive la résistance: Competing logics and the consolidation of U.S. community banking. Academy of Management Journal, 50(4), 799–820.

McPherson, M. (1983). An ecology of affiliation. American Sociological Review, 48(4), 519–532.

Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83(2), 340–363.

Mitchell, W. (1989). Whether and when? Probability and timing of incumbents’ entry into emerging industrial subfields. Administrative Science Quarterly, 34(2), 208–230.

Navis, C., & Glynn, M. A. (2010). How new market categories emerge: Temporal dynamics of legitimacy, identity, and entrepreneurship in satellite radio, 1990–2005. Administrative Science Quarterly, 55(3), 439–471.

Nelson, R. R. (2005). Technology, institutions, and economic growth. Cambridge, MA: Harvard University Press.

Phillips, N., Lawrence, T. B., & Hardy, C. (2004). Discourse and institutions. Academy of Management Review, 29(4), 635–652.

Pontikes, E. G. (2012). Two sides of the same coin: How ambiguous classification affects multiple audiences’ evaluations. Administrative Science Quarterly, 57(1), 81–118.

Pontikes, E. G., & Barnett, W. P. (2015). The persistence of lenient market labels. Organization Science, 26(5), 1415–1431

Pontikes, E. G., & Hannan, M. (2014). An ecology of social categories. Sociological Science, 1, 311–343.

Porac, J. F., Thomas, H., & Baden-Fuller, C. (2011). Competitive groups as cognitive communities: The case of Scottish knitwear manufacturers revisited. Journal of Management Studies, 48(3), 646–664.

Prottas, J. (2013). How do funders understand fiscal sponsorship? National Network of Fiscal Sponsors Annual Meeting. Boston.

Rao, H., Monin, P., & Durand, R. (2003). Institutional change in Toque Ville: Nouvelle cuisine as an identity movement in French gastronomy. American Journal of Sociology, 108(4), 795–843.

Rosa, J. A., Porac, J. F., Runser-Spanjol, J., & Saxon, M. S. (1999). Sociocognitive dynamics in a product market. Journal of Marketing, 63, 64–77.

Saxenian, A. (1996). Inside-out: Regional networks and industrial adaptation in Silicon Valley and Route 128. Cityscape, 2(2), 41–60.

Scott, W. R. (1995). Institutions and organizations. Thousand Oaks, CA: Sage.

Scott, W. R. (2001). Institutions and organizations. Thousand Oaks, CA: Sage.

Sorenson, O., & Stuart, T. E. (2001). Syndication networks and the spatial distribution of venture capital investments. American Journal of Sociology, 106(6), 1546–1588.

Vergne, J.-P., & Wry, T. (2014). Categorizing categorization research: Review, integration, and future directions. Journal of Management Studies, 51(1), 56–94.

Wry, T., Greenwood, R., Jennings, P. D., & Lounsbury, M. (2010). Institutional sources of technological knowledge: A community perspective on nanotechnology emergence. In N. Phillips, G. Sewell, & D. Griffiths (Eds.), Technology and organization: essays in honour of Joan Woodward. Research in the Sociology of Organizations (Vol. 29, pp. 149–176). Bingley: Emerald Group Publishing.

Wry, T., Lounsbury, M., & Glynn, M. A. (2011). Legitimating nascent collective identities: Coordinating cultural entrepreneurship. Organization Science, 22(2), 449–463.

Zuckerman, E. W. (1999). The categorical imperative: Securities analysts and the illegitimacy discount. American Journal of Sociology, 104(5), 1398–1438.

APPENDIX A: ANALYSIS OF POTENTIAL FOR SURVIVOR BIAS

The earliest known list that attempts to identify the population of fiscal sponsors is from 2005 to 2006. However, because our interest is in the beginning of the label emergence process, when we initially began examining the website archives of these organizations we realized that there were instances of use of the fiscal sponsor label on some of their homepages that predated 2005. Building our data set going back all the way to the earliest observations is in keeping with the theoretical process of interest, but increases concerns about survivor bias. In particular, the concern is that proximate social spaces that have low density of peers identified might be due to factors in those proximate social spaces that make them inhospitable to these types of organizations, and so the low density may be an outcome of a process which leads to closures of organizations that thus should have been included in our data set, but were not because they were no longer operating in 2005.

We performed supplementary data collection and analysis to assess this risk. Because the primary concern is that survivor bias is inappropriately inflating our results, we focus our supplementary analysis on geographically defined social spaces, as this is the set of analyses in which we are most concerned about overestimating coefficients since these are the analyses that resulted in significant results.

For all of the MSAs (and where missing, the county codes in keeping with the data construction for our sample) in our sample at any point in time, we collected data on the number of organizations that were listed in the IRS data files and then calculated the percent change relative to key thresholds. Because we have organizations that entered the sample during the years of observation, we segmented this analysis based on the year that organizations incorporated with the IRS and calculated the percent change within those groups of organizations. While the IRS has historically removed organizations from their data files when they know that they closed or went out of business, organizations that were inactive were not historically removed. This was true until 2011, at which point thousands of organizations had their nonprofit status revoked, and were removed from these data files. Because of this, we calculate the percent change in the number of organizations by geography and incorporation year between both the time of incorporation and 2005 as well as between 2010 and 2011.

Because IRS data are collected and compiled throughout the year, and with some leeway, we use data files from the year after the year of interest for the first set of calculations, which can be thought of as the active changes to incorporated nonprofits. Table A1 explains the years of data collection for t₀, whereas t₁ data was all from 2006. The later data, which can be thought of as the inactive changes to incorporation status all use data from 2010 and 2011 to assess the changes, which are again calculated by geography and incorporation year. The percent change is calculated such that a decrease in the number of organizations in a geography with a particular incorporation date is positive:

math-eqn

In correlating these change variables with the proportion of organizations within a geography that are identified as peers, a positive correlation means that there is a relationship such that decreasing organizations in the MSA are related to an increase in the relative number of peers. A positive correlation between the change variables and the number of registered organizations means that there is a relationship between a decrease in the number of organizations founded at a particular point in time and the total population of nonprofits in that same geography. And a positive correlation between the change variables and the number of peers in an MSA demonstrates that regions that exhibit greater loss of organizations from those particular points in time are not those which have fewer peers in later years. Thus, positive correlations indicate that there should not be concern about the effects of survivor bias based on the distribution of closures across geographies, which is the closest we can get to assessing the possibility of specific dynamics for practitioners of fiscal sponsorship without actually knowing the population of fiscal sponsors as they existed at those points in time.

Table A1 presents the correlations between the change variables and the focal variables from our analysis. It demonstrates that there is little need to be concerned about survivor bias based on this analysis. Even for the correlation that would be of most concern at −0.19, this appears to be attributable to the fact that even though it is positively correlated with both the denominator and numerator, the correlation is larger with the denominator. We replicate these correlations using data from 2006 because of the timing of the list creation. We also replicated these general patterns both including and excluding geographical regions that had problematic percent changes (they were less than −1), which either reflected changes to how the IRS was defining the geographical areas or that there were unusually large movements between MSAs.

Table A1. Correlations between Change Variables and Conceptual Variable and Its Components.

APPENDIX B: DESCRIPTIVE STATISTICS

Table B1. Summary Statistics for Sub-Sample Based on Market Segment (478 Cases).

Table B2. Summary Statistics for Sub-Sample Based on Geography (1,550 cases).

THE SOCIAL CONSTRUCTION OF MARKET CATEGORIES: HOW PROXIMATE SOCIAL SPACE CREATES STRATEGIC INCENTIVES TO BE EARLY CLAIMANTS OF THE FISCAL SPONSOR LABEL

ABSTRACT