Legitimacy and Legitimation Practices: An Analysis of TSMO Networks 1 Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 University of Bamberg Press Legitimacy and Legitimation Practices: An Analysis of TSMO Networks Author: Takuma Shibaike University of Calgary Email: takumi.s.com@gmail.com Private transnational organizations have grown in number and in influence. However, sociologists and political scientists often study them separately, either as transnational social movement organizations (TSMOs) or the larger category of international non-governmental organizations (INGOs). In this paper, I examine the determinants of TSMO legitimacy by drawing on the literature on INGOs. In so doing, I call for bridging the disciplinary gap between sociology and political science. Empirically, I find that legitimation benefits already prominent organizations more than those that are not. Networking thus helps reproduce the hierarchy among the TSMOs, challenging the earlier notion that TSMOs are horizontally networked. However, I also find that Southern TSMOs are more likely to gain legitimacy than Northern TSMOs once they are visible to their peers. The analysis of TSMOs thus cautions our bias to study Northern INGOs and generalize the findings to INGO population. Overall, my findings reveal that the incentives and strategies that INGO research has documented exist among TSMOs despite their counter-hegemonic ambitions. Keywords: international nongovernmental organizations; organizational legitimacy; organizational networks Introduction Private, non-profit transnational organizations are regularly engaged in the pursuit of social change, working on human rights, eco-systems, gender equity, and more. The work of such transnational organizations has been noted in seminal studies of transnational social movements (Boli & Thomas, 1999; Keck & Sikkink, 1998; Tarrow, 2005). The typical picture of these groups is of principled actors seeking to hold the powerful to account (Smith & Wiest, 2012). Existing research documents that some private organizations consistently wield vast authority, commanding the attention of global media outlets, policymakers, corporations, and their peers (Bob, 2005; Carpenter, 2014; Dauvergne & LeBaron, 2014; Stroup & Wong, 2017). Both sociologists and political scientists have extensively analyzed these organizations but used different terms to describe them, with the former preferring the term “transnational social movement organization” (TSMO) and the latter employing “international nongovernmental organization” (INGO). Conceptually speaking, TSMOs are a special subset of INGOs, defined as INGOs “explicitly involved in work to change the dominant political and social order” (Smith & Wiest, 2012: 46). As INGOs include pro- system organizations, TSMOs and INGOs may seem to fall into different conceptual categories. However, little empirical investigation has been done to demonstrate the differences. The question is not just about the terminology of transnational private organizations. The difference in the terminology also signifies different ways in which research is conducted among sociologists and political scientists. While sociologists have studied the processes in which TSMOs organize progressive movements (Caniglia, 2001; Smith, 2001, 2002; Smith, Chatfield, & Pagnucco, 1997; Smith, Gemici, Plummer, & Hughes, 2018; Smith, Plummer, & Hughes, 2017), research on the differences in organizational legitimacy has been relatively scarce. By contrast, political scientists have focused on the differences in organizational legitimacy, demonstrating the effects of INGO legitimacy on international norms and regulations (Bob, 2005; Carpenter, 2014; Stroup & Wong, 2017). The analysis of organizational legitimacy is important because neither TSMOs nor INGOs have coercive power to singlehandedly affect social changes. Instead, they must rely on legitimacy—a right to act—acknowledged by their audiences. Although the analysis of social movement processes offers a useful mailto:takumi.s.com@gmail.com http://dx.doi.org/10.20377/cgn-114 2 University of Bamberg Press insight into TSMO behaviors, the analysis of organizational legitimacy helps us understand the effect of TSMOs on the broader political and social phenomena. Taken together, I ask why some TSMOs gain legitimacy while others do not. In so doing, I draw on the literature on INGO research. This paper has two main contributions. First, it reveals the extent to which the choices of TSMOs influence their own legitimacy. In recent years, the legitimacy of civil society as a whole has been challenged (Deloffre & Schmitz, 2019), but we know little about how individual organizations build or defend their legitimacy. This paper evaluates whether TSMOs are capable of shaping their own legitimacy through strategic choices, or their legitimacy is shaped by structural pressure exerted by shared expectations about organizations (DiMaggio & Powell, 1983). Second, this paper emphasizes the need to bring the literatures on INGOs and TSMOs closer together. To do this, I leverage a newly updated TSMO dataset1 and evaluate whether explanations for INGO legitimacy are able to account for the variation in TSMO legitimacy. In particular, I examine whether the anti-systemic goals of TSMOs challenge the current understandings of INGOs. The results of my analysis below show that legitimation through networking (i.e. claiming a connection to another organization) is relatively rare but effective among TSMOs. My empirical evidence suggests that networking is a deliberate attempt to leverage connectivity for organizational legitimacy rather than driven by functional needs. However, I also find that the effect of networking on organizational legitimacy is larger for prominent TSMOs than for low-profile ones, suggesting that the existing hierarchy among TSMOs is persistent. While some scholars characterize TSMO networks as horizontal (Smith, 2002; Smith et al., 2018), my findings support the view that the division and inequality are salient characteristics of TSMO networks (Hughes, Paxton, Quinsaat, & Reith, 2018). These findings are consistent with the established canon explaining INGO behaviors (Lake & Wong, 2009; Murdie, 2013), underscoring the benefit of studying INGOs and TSMOs together The following sections are organized as follows. First, I define TSMOs and discuss how I conceptualize and measure TSMO legitimacy. Second, I briefly review the network-based studies of INGOs and TSMOs to motivate my analysis. I then empirically evaluate arguments about organizational legitimacy through quantitative analysis. I demonstrate that legitimacy is primarily associated with legitimation attempts. However, legitimation advantages already prominent TSMOs rather than those that are not, helping the reproduction of hierarchy among the TSMO community. The final section discusses the implications of my findings. Understanding the Legitimacy of Transnational Social Movement Organizations The concept of TSMOs was introduced more than two decades ago by Jackie Smith and her colleagues (Smith et al., 1997). Smith and Wiest (2012: 46) define TSMOs in terms of their goals and objectives. A transnational private group is a TSMO if it is “explicitly involved in work to change the dominant political and social order.” The definition of TSMOs by itself does not imply ideological orientations, but Smith and Wiest (2012) recognize that they are mostly progressive and even anti- systemic, since right-wing, nationalist groups tend to avoid publicizing their work (exceptionally, see Bob (2012) for global right-wing mobilizations). Organizations whose primary focus is research, religion, service provision, or education are also excluded (Smith & Wiest, 2012: 70). TSMOs are thus a subset of the national and INGO population with an anti-systemic focus. Currently, there is no database that catalogs the entire population of INGOs. Scholars often use the Yearbook of International Organizations as a relatively comprehensive dataset on INGOs (for its 1 Smith, Jackie, Dawn Wiest, Melanie M. Hughes, Samantha Plummer, Brittany Duncan. 2017. Transnational Social Movement Organizations Dataset (TSMOD), 1953-2013. [Computer file]. Pittsburgh, PA: World Historical Dataverse [distributor]. And Hughes, Melanie M., Samantha Plummer, and Jackie Smith. 2017. Transnational Social Movement Organization Networks (TSMOnet), Version I, 1993, 2003, & 2013. [Computer File]. Pittsburgh, PA: World Historical Dataverse [distributor]. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 3 University of Bamberg Press sampling bias, see Bush and Hadden’s (2019) comparison with US charity database). The Yearbook updates every year and covers both intergovernmental organizations (IGOs) and private organizations. For private organizations to be included, they must hold three or more offices internationally or conduct internationally-oriented activities. However, the Yearbook does not publish its data longitudinally or in a manner that is amenable to cross-sectional studies. In building the initial TSMO dataset, Smith and her colleagues reviewed the historical records of the Yearbook and sampled the organizations that fit the above definition. In the context of transnational organizations, legitimacy is defined as the belief that an actor is right to act in the eyes of audiences. It does not mean that legitimate actors always induce behavioral compliance by their audiences (Lake, 2009), but their claims will be heard and taken seriously. If the Brookings Institution offers a view on American politics, for example, other think tanks may contrast their positions with Brookings’ to legitimate them as challengers. As Hudson (2001: 348) points out, legitimacy is “a socially-constructed quality that can be ascribed to an NGO by stakeholders coming from different perspectives.” As such, legitimacy is inherently relational, must be conferred by others, and cannot be claimed singlehandedly. Although the study of TSMOs tends to assume that any TSMO is legitimate, substantial differences exist between individual TSMOs. To measure the legitimacy of TSMOs, I use a dataset on networks of TSMOs (Smith et al., 2018). This network data also comes from the Yearbook, but importantly, network ties are self-reported collaboration between TSMOs. The nature of self-reporting could harm measurement validity if we analyze actual cooperation and collaboration among TSMOs. However, self- reporting works well for the purpose of measuring organizational legitimacy. For example, if A reports a tie to B, B should also have collaboration with A, but this does not necessarily mean that B reports a tie with A. TSMOs do not have to validate their ties when reporting to the Yearbook, so they may aspirationally name other organizations as a way to legitimate themselves before their peers. Conversely, even if a TSMO works with other organizations, it may not mention some unimportant ones. From this vantage point, the number of incoming ties can be interpreted as an indicator of organizational legitimacy. The more legitimate a TSMO, the more likely it is to have incoming ties from other organizations seeking to benefit from association with a legitimate actor. This legitimation practice is widely adopted by private organizations, including for-profit firms, which often use association with high-status firms as a quality signaling in market settings (Kilduff & Krackhardt, 1994; Podolny, 1993). To use a more concrete example of Amnesty International, TSMOs may be interested in Amnesty International’s global reach and stated commitment to representation of diverse views on human rights. Other might appreciate Amnesty’s effectiveness in securing attention from global media outlets (Powers, 2018). Whatever the reason may be, TSMOs like Amnesty are likely to end up with having more incoming ties than outgoing ties because peers want to leverage connections with legitimate organizations. Conversely, I use outgoing ties to approximate the intensity of legitimation efforts. My use of peer networks builds upon existing efforts to use incoming ties as an indicator of organizational legitimacy. Mitchell and Stroup (2017) measure organizational reputation based on the self-reporting of INGOs during interviews with INGO staff. Similarly, Murdie and Davis (2012) create a matrix of INGO network ties based on the Yearbook data, and Stroup and Wong (2017) provide a ranking of INGOs in terms of authority based on the number of incoming ties. To be sure, legitimacy may be observable in other places. For example, differences in INGO legitimacy can be observed from the frequency by which media outlets quote INGOs: legitimate INGOs are more often cited to boost the credibility of news stories (Thrall, Stecula, & Sweet, 2014). However, a focus on peer networks has two major advantages. First, data on nonpeer organizations would introduce selection bias. We cannot observe TSMOs if they do not receive any attention, even if they are engaged in legitimation practices. The TSMO dataset allows us to observe negative cases since all TSMOs are observable regardless of the attention they receive. Second, because of counter-hegemonic ambitions, Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 4 University of Bamberg Press TSMOs may be less interested in the response of pro-system actors. Pro-system actors, such as states, corporations, and INGOs that frequently partner with states and corporations, typically undermine rather than help transnational social movements (Smith et al., 2017). This context in effect creates a hard test for my hypotheses based on INGO research. Because INGOs include pro-system actors, if the insights of INGO research are able to account for TSMO legitimacy among peers, they should also work when we analyze TSMO legitimacy among other groups. Understanding Legitimacy in Transnational Private Organizations Existing research has adopted network perspectives to understand the legitimacy of INGOs. I highlight four findings from network theories to motivate my analysis of TSMO networks. First, the choice of strategies within networks can shape structural properties, such as international norms, which in turn affect the legitimacy of agents. For example, Goddard (2006) argues that entrepreneurial agents in international networks are able to effect changes in normative structures. Others have explored how organizational strategies, in the context of networks, affect advocacy output (Luxon & Wong, 2017) or their ability to affect international norms (Wong, 2012). Although INGOs and TSMOs are constrained by international structures, network theory allows us to examine if and how organizations can effect changes in broader social structures. Second, networks between INGOs and other types of actors can be seen as co-constitutive of legitimacy. For instance, given substantial attention to the “democratic deficit” in global governance, IGOs are increasingly “opening up” to INGO participation (Bernauer, Böhmelt, & Koubi, 2013; Tallberg, Sommerer, & Squatrito, 2013). INGOs generate accountability for IGOs, and in turn, participating INGOs increase their voice in global governance. Research has documented that some TSMOs also leverage their networks to increase their legitimacy. Smith (2001) argues that appeals to international law provided legitimacy to the TSMOs protesting the WTO in Seattle. Similarly, the work of environmental TSMOs was legitimated by national elites and marginalized states (Caniglia, 2001; Smith, 2014). Third, networks among INGOs are frequently analyzed as attempts to build legitimacy. In the analysis of what she terms free-riding behavior, Murdie (2014b) argues that outgoing ties by human rights INGOs to other INGOs are attempts to increase their legitimacy. Carpenter (2011) also argues that agenda setting by INGOs at the center of advocacy networks create “bandwagoning effect” where smaller INGOs join the coalition and bolster their legitimacy by joining the advocacy winners. Finally, the effect of networks on legitimacy is well-established in sociology, although the focus tended to be domestic organizations. Kamens (1977) shows how universities and colleges create certain discourse about education that legitimates particular organizational structures. Organizations also leverage their connections to influential companies and local communities (Galaskiewicz, 1979). Similar to IGO-INGO relations, corporations and charitable organizations have mutually co-constitutive legitimacy (Galaskiewicz, Wasserman, Rauschenbach, Bielefeld, & Mullaney, 1985; Miles & Cameron, 1982). Others show how the structure of networks affected the legitimacy of small- to medium-sized corporations in the United States (Human & Provan, 2000). In short, the existing network literature on INGOs and organizational sociology has demonstrated how ties among actors, while also useful for other purposes, are mechanisms for building and defending legitimacy. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 5 University of Bamberg Press Who Is Legitimate? In this section, I describe different levels of legitimacy among TSMOs, using the latest TSMO dataset.2 The dataset by Smith et al. (2018) has two files. One file reports a variety of organizational attributes, including the location of headquarters, the issue areas of operation, and the years of founding, every odd number year from 1981 to 2013. The other file reports TSMO networks in 1993, 2003, and 2013.3 These networks capture all sorts of “horizontal,” collaborative self-reported connections between TSMOs. In other words, they do not indicate formal hierarchies, such as founder-founded relationship or official membership to an umbrella organization.4 As I discussed above, however, the nature of self- reported ties creates a de facto hierarchy among TSMOs in terms of organizational legitimacy. Both files are based on the entries of the Yearbook. I merged these two data files as cross-sectional data to overcome observation bias. Because data on TSMO networks only include organizations that have ties, they miss a large number of organizations that are not involved in any networking efforts. To avoid the selection bias of observing only positive cases, I needed to compare TSMOs that engaged in networking and ones that did not. Thus, my analysis treats incoming and outgoing ties as organizational attributes. Since network data are available only in 1993, 2003, and 2013, I also took the subset of organizational data in these years. The number of organizations in my dataset is 4,946 (1,045, 1,879, and 2,022 in 1993, 2003, and 2013, respectively). Figure 1: Histogram for legitimacy of TSMOs Descriptively, the TSMOs that are legitimate or involved in legitimation are relatively uncommon. We can see this from both incoming ties and outgoing ties (see Table A1 for summary statistics). As Figure 1 shows, the majority of the TSMOs receive no tie (73.1%). This pattern by itself challenges the widespread portrayal of TSMOs as a highly networked community (Carroll, 2007; Juris, 2008; Wolfson & Funke, 2017). More importantly, it suggests that organizational legitimacy is rare, or very few TSMOs are legitimate enough to induce others to claiming association. Yet, the finding is consistent with existing 2 The same as above. 3 These years were chosen by the coders, Smith et al. (2018). I focus on the ties that are described as any collaborative relationships, coded as “1” in the TSMOnet dataset. 4 For more, see the codebook of the TSMOnet. Hughes, Melanie M., Samantha Plummer, and Jackie Smith. 2017. Transnational Social Movement Organization Networks (TSMOnet), Version I, 1993, 2003, & 2013. [Computer File]. Pittsburgh, PA: World-Historical Dataverse [distributor] Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 6 University of Bamberg Press INGO research. For example, Stroup and Wong (2017) find that only a handful of INGOs have authority before multiple audiences and the vast majority of them have none. In terms of legitimation, the majority of the TSMOs also do not send ties (67.1%), as shown in Figure 2. Some TSMOs have extreme values due to the self-reporting nature of the Yearbook, suggesting that the strategy of networking is not uniformly distributed among TSMOs. The low number of outgoing ties suggests that TSMOs do not just “name drop.” It represents costly networking, in which TSMOs strategically send ties to other INGOs rather than sending them automatically by working in the same issue area. Figure 2: Histogram for networking effort of TSMOs The scarcity of observed networking behavior among TSMOs does not necessarily mean that they lack an interest in developing and defending their legitimacy. As a point of comparison to INGOs, I use Stroup and Wong’s (2017) list of top fourteen INGOs in terms of the amount of attention paid by a variety of audiences, such as states, corporations, and peer INGOs. Among the fourteen INGOs, eleven are identified as TSMOs.5 As Table 1 reports, not all leading INGOs are highly legitimate in the TSMO community, which suggests that TSMOs might be unique in terms of their ideas about what a legitimate organization should look like (see also Table A2 for 1993 and 2003 data and Figure A1 for network diagram). As discussed above, Smith and Wiest (2012) define TSMOs in terms of their counter-hegemonic goals and objectives, a position that is not always welcomed by other actors, especially states and corporations. Below, I take a closer look to investigate which factors explain the legitimacy of TSMOs. 5 Ten TMOS were leading INGOs in 1993, and they were 11 in 2003 and 2013. International Campaign to Ban Land Mines were not included in1993, presumably because it was too young (established in 1992). Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 7 University of Bamberg Press Table 1: Legitimacy ranking of TSMOs from 2013 data What Explains Legitimacy? If some TSMOs stake their legitimacy on being counter-hegemonic actors, what are the organizational attributes that satisfy such expectations? Here, I derive several hypotheses from the literature on INGOs. First, networking behavior, such as making a claim to work with other INGOs, can be used as an attempt to increase legitimacy. For example, Bob (2005) argues that insurgent groups seek to obtain the endorsement of highly legitimate INGOs. I expect that the more they send ties, the more others will respond in kind, thereby creating legitimacy (through incoming ties) for themselves. H1: The more ties a TSMO will send, the more legitimacy it will gain among its peers. Another plausible explanation is that affiliation with IGOs increases TSMO legitimacy by sending credible signals. Murdie (2014a) finds that INGOs with IGO consultative status are more likely to be effective in international development, as consultative status works as a costly signal to local and international donors. It is costly because the requirements for such consultative status are heavy, and an INGO must have the capacity to go through the long process of obtaining consultative status. Moreover, consultative status gives INGOs standing before states in IGOs and helps them link to other INGOs. Murdie (2013) finds that human rights NGOs with IGO consultative status are more likely to be connected with one another. However, as social movements frequently target IGOs as neoliberal institutions (Ayres, 2004), consultative status may tarnish TSMO legitimacy by signaling pro-system positions. INGO signaling behavior is measured by the number of IGO affiliations. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 8 University of Bamberg Press H2: A TSMO will gain more legitimacy as the number of IGO affiliations increase. The divide between the global North and South is a critical element that counterhegemonic movements seek to address (Evans, 2012; Smith et al., 2018; Smith & Wiest, 2012). And yet, the existing literature suggests that the connection with Northern INGOs is essential for Southern INGOs to achieve their goals, mainly because Northern INGOs enjoy larger social and economic resources necessary to influence states and IGOs (Bob, 2005; Carpenter, 2007; Keck & Sikkink, 1998). Northern TSMOs may therefore induce networking behavior of others. Here, TSMOs are treated as Northern INGOs if they have primary or secondary headquarters in the global North. H3: TSMOs located in the global North are more legitimate than those located in the global South. Finally, organizational age may be associated with TSMO legitimacy, as TSMOs may be able to establish their brand associated with their issue areas of operation over time. Household names, such as Amnesty International and Oxfam, were indeed established more than half a century ago. Operating for a long time also signals an ability to survive in a changing environment. Organizational age is computed as the data year (1993, 2003, or 2013) minus the year of TSMO founding.6 H4: On average, older TSMOs are more legitimate than younger TSMOs. Estimation methods To test the hypothesized relationships, I control for other factors that might confound TSMO legitimacy. First, the size of a TSMO can shape prominence and thus organizational legitimacy. Typically, size is seen as budget, but the dataset does not report the budget of TSMOs. Instead, I use the number of countries where a TSMO operates as the measure for organizational scale, with the assumption that operating in more countries indicates a greater fiscal capacity. Second, TSMOs working in multiple issue areas may have greater legitimacy because they can bridge different clusters of organizations (Granovetter, 1977; Murdie & Davis, 2012). I therefore control for whether or not a TSMO operates in multiple issue areas. Even though the dataset is panel-structured, the variation among the indicators within each TSMO is over time quite low. For example, once a TSMO is established in the global South, it is uncommon to move its headquarters to the global North. Therefore, rather than including time fixed effects or lagged dependent variables, I resorted to a pooled regression with organizational fixed effects. Organizational fixed effects account for the unobservable, idiosyncratic effects on organizational legitimacy that are specific to each TSMO (Vaisey & Miles, 2017). It is important to note that the dependent variable is not tie activation itself, but the number of incoming ties. In network terms, it is called indegree centrality. While tie activation is estimated with exponential random graph models (ERGMs) (Cranmer & Desmarais, 2011), indegree centrality is often estimated by regression analysis (Faris & Felmlee, 2011; Lyle & Smith, 2014). In my case, the dependent variable is a count variable with many zero observations (73.1%). To account for overdispersion and zero- inflation, I adopted a zero-inflated negative binomial (ZINB) model. ZINB regression assumes that observed zeros are generated by two different, unobservable processes: the zero-inflation stage estimates the latent group membership of the sample that could have never received a tie (structural zeros), while the conditional stage estimates the number of ties for the rest of the sample, some of which happened to receive no tie (incidental zeros) (Hendrix & Wong, 2013; Long, 1997). In my interpretation, structural zeros 6 A few TSMOs indicated negative ages, which were dropped from the analysis. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 9 University of Bamberg Press may be generated by visibility. In other words, if a TSMO is not visible to its peers, it could never be mentioned. By contrast, incidental zeros may result from the lack of organizational attributes necessary to TSMO legitimacy, even if the organization is known by others. As a robustness check, I also conducted ERGMs to explicitly account for network dependency (see Tables A3 and A4), but the results are consistent with my findings below. Determinants of legitimacy Table 2 reports the results of regression analysis. Observations with missing data were removed, which gives us a total of 3,447 observations. For robustness checks, Model 2 (ZINB without random intercepts), Model 3 (negative binomial model: NB), and Model 4 (zero-inflated Poisson model: ZI Poisson) are also reported. Note that the zero-inflation stage estimates the membership for the structural- zero group, so negative coefficients mean that such variables are, in short, positively associated with visibility of a TSMO among peers. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 10 10 University of Bamberg Press Table 2: Results of regression analyses on the legitimacy of TSMO Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 11 University of Bamberg Press Across all models, I find robust support for H1 that networking behavior is positively associated with TSMO legitimacy. In the zero-inflation stage, outgoing ties are negatively associated with the structural-zero membership. This means that sending a tie to other TSMOs can increase organizational legitimacy among peers and it also increases their visibility among the broader TSMO community. In short, networking as a legitimation practice works to gain both visibility and legitimacy. However, the marginal effect of networking might be extremely limited for smaller TSMOs. Figure 3 illustrates the marginal effect of networking behavior for two TSMOs— Amnesty International and International Union of Catholic Esperantists (IUCE)—while everything else is held at the mean for each TSMO. The marginal effects differ between these organizations because they are dependent on the values of other independent variables, but they are illustrative of the fact that already legitimate TSMOs can easily legitimate themselves with networking, while low-profile ones like IUCE have a much more difficult time. Figure 3: Predicted legitimacy of Amnesty International and Catholic Esperantists Collectively, this effect would help reproduce the hierarchy of legitimacy among TSMOs despite the repeated claims that TSMOs are often horizontally organized (Smith, 2002; Smith et al., 2018). Given this finding, it also makes sense that many TSMOs do not engage with networking (see Figure 2) because low-profile TSMOs, which are the vast majority of TSMOs in the dataset, do not gain much legitimacy from it. Low-profile TSMOs may be better off focusing on local constituencies rather than seeking collaboration with other TSMOs, which incurs transaction and administrative costs. I also find support for H2 that IGO consultative status increases TSMO legitimacy. The effect of each additional affiliation with an IGO is substantially larger than that of networking in the conditional stage. In the zero-inflation stage, the effect is much smaller than networking and not statistically significant; that is, no evidence supports that affiliation with IGOs by itself can increase the visibility of a TSMO. It is only when the TSMO is known among other TSMOs that IGO affiliation can increase organizational legitimacy. This may be because, while IGOs serve as a venue in which TSMOs can claim credit over outcome (Betsill & Corell, 2008), civil society participation in IGO meetings may not be accessible to outsider TSMOs and thus does not increase organizational visibility. The results also suggest that, despite the critiques that IGOs are coopted by neoliberalism (Bond, 2012; Charnovitz, 1996; Mooney, 2012; Willetts, 2006), affiliation with IGOs does not damage TSMO legitimacy. Perhaps, pragmatism is Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 12 University of Bamberg Press important for the legitimacy of counter-hegemonic movements, and anti-systemic positions can be more about market positioning. For example, Sea Shepherd Conservation Society, which rejects pro-system advocacy methods (Eilstrup-Sangiovanni & Bondaroff, 2014), accepted funds from the Dutch National Postcode Lottery, which is run by a marketing agency.7 Contrary to my expectation in H3, having headquarters in the global North is negatively associated with TSMO legitimacy. In the zero-inflation stage, the substantive effect is positive but statistically insignificant. That is, I do not find that Northern TSMOs are more visible in the TSMO community, perhaps because the majority of TSMOs are headquartered in the global North anyway (74.8%). It is still puzzling, however, that they receive fewer ties, ceteris paribus, than Southern TSMOs once they are known. One could argue, given that the current structure of global governance favors Northern INGOs (Barnett & Walker, 2015), Southern TSMOs that brought themselves to visibility might exhibit exceptional qualities. Alternatively, organizations with counter-hegemonic ambitions, such as TSMOs, might favor actors from the global South. The result also points out the problem of observation bias. We know that the density of ties among TSMOs is higher among Northern countries (Hughes, Peterson, Harrison, & Paxton, 2009), but it does not follow that Northern TSMOs are on average more legitimate than Southern TSMOs. I find limited support for H4, and organizational age has two different effects in the zero-inflation and conditional stages. In the zero-inflation stage, organizational age is negatively associated with the structural zero group, meaning that TSMOs are more likely to gain visibility as long as they stay active. Surprisingly, however, in the conditional stage, organizational age is negatively associated with legitimacy. There are two possible reasons for this. First, older TSMOs have likely amassed long track records that other TSMOs can exploit to challenge their legitimacy. Second, organizations that have survived for a long time may have developed a narrow, path-dependent vision. If TSMOs have been successful in effecting change in the past, they could be trapped in the past success model (Yanacopulos, 2015). More broadly, this finding cautions us against extrapolating our knowledge from leading INGOs. For example, Amnesty International and the Friends of the Earth were established more than 50 years ago, but these TSMOs with high levels of legitimacy should be seen as anomalies rather than typical longstanding organizations. Organizational size, measured by the number of countries in which TSMOs operate, had a statistically significant effect on organizational legitimacy. This is unsurprising in the sense that large organizations should have relatively large operational budgets that allows them to collaborate with many other TSMOs. However, the effect of organizational size on visibility was statistically inconsistent. Overall, the marginal effect of organizational size was smaller than networking or IGO affiliation. The result thus strongly suggests that being a large organization is not the only determinant of organizational legitimacy. It does seem to help, but strategic actions, such as networking and IGO affiliation, have greater impacts on the legitimacy of TSMOs. Finally, I did not find the evidence of brokerage by TSMOs working in multiple issue areas, such as human rights and development. While Murdie and Davis (2012) find that such "hybrid" organizations play an important role in bridging different network clusters, multi-issue TSMOs are not necessarily more legitimate than the rest of TSMOs. Perhaps, brokerage is not a widely adopted tactic among TSMOs, as Hadden (2015) finds in the networks of climate change organizations. In Appendix 3, I conducted an analysis based on ERGMs to estimate the chance of legitimation between two TSMOs, which generally supports the findings presented here. 7 http://www.theideatree.ca/sea-shepherds-8-3-million-euro-award-and-the-idea-tree/ (Accessed June 11, 2019) Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 http://www.theideatree.ca/sea-shepherds-8-3-million-euro-award-and-the-idea-tree/ 13 University of Bamberg Press Conclusion In this paper, I evaluated self-reported networks as proxies for the legitimacy of TSMOs. The self- reported nature of TSMO data allowed me to exploit the incentives and disincentives of associating with other organizations in a publicly accessible forum. A TSMO that reports a network tie is, whether genuine or not, engaging in a strategy of legitimation. I demonstrated the insights of INGO research can also explain how different strategies and organizational attributes affect TSMO legitimacy. The main empirical finding is that, while TSMOs can improve their legitimacy by leveraging connectivity to other TSMOs, such networking helps those that are already prominent rather than lesser known TSMOs. There are also a few counter-intuitive findings. Among TSMOs, Southern organizations are more likely to gain legitimacy than Northern TSMOs and organizational age is negatively associated with legitimacy (although it does help organizational visibility). The analysis begs further research as to why the hierarchy of TSMOs is so persistent despite their counter-hegemonic ambitions. I suspect it has to do with the credibility of the claims (Gourevitch, Lake, & Stein, 2012). Perhaps, there is a bias in how TSMOs interpret claims to have network ties. A legitimate TSMO’s claims to work with other organizations should be true, 1) as their reputations negate their need to make spurious claims about partners; and 2) their prominence makes it more likely that misrepresentations will be reported. Such a display of collaborative behavior might further increase the legitimacy of the TSMO, potentially creating “free-riders” (Murdie, 2014b). Other TSMOs, on the other hand, may aspire towards having connections with better-known peers, but paradoxically, the lack of erstwhile interpretation makes it hard for others to interpret whether a reported tie is genuine or not. Moreover, even if such TSMOs genuinely reported ties, the networked TSMO may not see them as important and legitimate enough to report on a public forum. In short, peer organizations may not want to react to some TSMOs based on their claims, but they trust other TSMOs because of different levels of legitimacy. This insight of the winners keep on winning seems to be a logic that is true in other social relations, such as corporations and individuals (Sauder, Lynn, & Podolny, 2012). More broadly, structural pressure to maintain the hierarchy among TSMOs is persistent, and strategic choices have positive but limited effects on most TSMOs. The results also suggest that explanations for INGO behaviors generally work for TSMOs despite the latter’s inclination against the status quo, with the possible exception that Northern TSMOs are less likely to be legitimated than Southern TSMOs. This finding illuminates the challenges of being an anti- hegemonic organization. In other words, it is no surprise that TSMOs prefer to work with Southern organizations in rejection of the current world order. It is more surprising that other determinants of INGO legitimacy, such as IGO affiliations, translate to positive effects on TSMO legitimacy. Perhaps, TSMOs are strategic about how they can achieve counter-hegemonic goals, even if doing so requires some collaboration with pro-system actors. My analysis also contributes to efforts to understand the networking behavior of transnational private organizations. I have shown that sending an outgoing tie is an attempt at legitimation that actually works despite its limited effect. This is a different conclusion from a prominent interpretation in the INGO literature. Murdie (2014b) argues that INGOs with more outgoing ties than incoming ties are “free-riders.” She finds that INGOs that have been less active in advocacy events in the past are more likely to report outgoing ties, which she argues as evidence of these INGOs not actively contributing to the network. While Murdie may be partially correct that free-riding could happen around extremely prominent TSMOs, my finding that outgoing ties increase the expected number of incoming ties offers a different insight. Thus, this paper contributes to the call for more research on the strategic concerns of TSMOs and INGOs. While the structural pressure exerted by shared expectations about TSMOs is strong, it does not Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 14 University of Bamberg Press mean that TSMOs are impossible to construct their own legitimacy. Recent research on INGOs has emphasized the structural properties of issue areas, such as the density of organizational populations, in legitimating certain forms of organizations (Abbott, Green, & Keohane, 2016; Bush & Hadden, 2019), but the agency of INGOs should not be left aside. Most importantly, if TSMOs collectively act, as often stated in their goals and objectives, their definition of legitimacy may shift in the way that would be more conducive to the emergence of entrepreneurial transnational private organizations. Acknowledgement I thank Jackie Smith, Samantha Plummer, and Melanie Hughes for making the latest TSMO dataset available. I also thank Wendy Wong, Sarah Stroup, and the participants of conferences at the University of Pittsburgh, ISA, and ARNOVA for constructive comments. References Abbott, Kenneth W., Green, Jessica F., & Keohane, Robert O. (2016). Organizational ecology and institutional change in global governance. International Organization, 70(2), 247277. Ayres, Jeffrey M. (2004). Framing collective action against neoliberalism: The case of the antiglobalization movement. Journal of World-Systems Research, 10(1), 11-34. Barnett, Michael, & Walker, Peter. (2015). Regime change for humanitarian aid. Foreign Affairs, 94, 130-141. Bernauer, Thomas, Böhmelt, Tobias, & Koubi, Vally. (2013). Is there a democracy–civil society paradox in global environmental governance? Global Environmental Politics, 13(1), 88107. Betsill, Michele Merrill, & Corell, Elisabeth. (2008). NGO diplomacy: The influence of nongovernmental organizations in international environmental negotiations. Cambridge, MA: MIT Press. Bob, Clifford. (2005). The marketing of rebellion: Insurgents, media, and international activism. New York, NY: Cambridge University Press. ———. (2012). The global right wing and the clash of world politics. New York, NY: Cambridge University Press. Boli, John, & Thomas, George M. (1999). Constructing world culture: International nongovernmental organizations since 1875: Stanford University Press. Bond, Patrick. (2012). Politics of climate justice: Paralysis above, movement below. Scottsville, South Africa: University of Kwazulu- Natal Press. Bush, Sarah Sunn, & Hadden, Jennifer. (2019). Density and decline in the founding of international NGOs in the united states. International Studies Quarterly, 63(4), 11331146. Caniglia, Beth. (2001). Informal alliances vs. Institutional ties: The effects of elite alliances on environmental tsmo networks. Mobilization: An International Quarterly, 6(1), 37-54. Carpenter, R. Charli. (2007). Setting the advocacy agenda: Theorizing issue emergence and nonemergence in transnational advocacy networks. International Studies Quarterly, 51(1), 99-120. ———. (2011). Vetting the advocacy agenda: Network centrality and the paradox of weapons norms. International Organization, 65(1), 69-102. ———. (2014). "Lost" causes: Agenda setting and agenda-vetting in the global issues networks. Ithaca, NY: Cornell University Press. Carroll, William. (2007). Hegemony and counter-hegemony in a global field. Studies in Social Justice, 1(1), 36. Charnovitz, Steve. (1996). Two centuries of participation: NGOs and international governance. Michigan Journal of International Law, 18, 183-286. Cranmer, Skyler J., & Desmarais, Bruce A. (2011). Inferential network analysis with exponential random graph models. Political Analysis, 19(1), 66-86. Dauvergne, Peter, & LeBaron, Genevieve. (2014). Protest inc.: The corporatization of activism. Cambridge, UK: Polity Press. Deloffre, Maryam Zarnegar, & Schmitz, Hans Peter. (2019). INGO legitimacy: Challenges and responses. In T. Davies (Ed.), Routledge handbook of NGOs and international relations: Routledge. DiMaggio, Paul J., & Powell, Walter W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organization fields. American Sociological Review, 48(2), 147-160. Evans, Peter. (2012). Counter-hegemonic globalization. In G. Ritzer (Ed.), The wiley-blackwell encyclopedia of globalization. Hoboken, NJ: Wiley-Blackwell. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 15 University of Bamberg Press Faris, Robert, & Felmlee, Diane. (2011). Status struggles: Network centrality and gender segregation in same-and cross-gender aggression. American Sociological Review, 76(1), 48-73. Galaskiewicz, Joseph. (1979). The structure of community organizational networks. Social Forces, 57(4), 1346-1364. Galaskiewicz, Joseph, Wasserman, Stanley, Rauschenbach, Barbara, Bielefeld, Wolfgang, & Mullaney, Patti. (1985). The influence of corporate power, social status, and market position on corporate interlocks in a regional network. Social Forces, 64(2), 403-431. Goddard, Stacie E. (2006). Uncommon ground: Indivisible territory and the politics of legitimacy. International Organization, 60(1), 35-68. Gourevitch, Peter A., Lake, David A., & Stein, Janice Gross (Eds.). (2012). The credibility of transnational NGOs: When virtue is not enough. New York, NY: Cambridge University Press. Granovetter, Mark S. (1977). The strength of weak ties American journal of sociology (Vol. 78, pp. 1360-1380): Elsevier. Hadden, Jennifer. (2015). Networks in contention: The divisive politics of global climate change. New York, NY: Cambridge University Press. Hendrix, Cullen S., & Wong, Wendy H. (2013). When is the pen truly mighty? Regime type and the efficacy of naming and shaming in curbing human rights abuses. British Journal of Political Science, 43(3), 651-672. Hudson, Alan. (2001). NGOs’ transnational advocacy networks: From ‘legitimacy’to ‘political responsibility’? Global Networks, 1(4), 331-352. Hughes, Melanie M., Paxton, Pamela, Quinsaat, Sharon, & Reith, Nicholas. (2018). Does the global north still dominate the international women’s movement? A network analysis of women’s international nongovernmental organizations, 1978–2008. Mobilization: An International Quarterly, 23(1), 1-21. Hughes, Melanie M., Peterson, Lindsey, Harrison, Jill Ann, & Paxton, Pamela. (2009). Power and relation in the world polity: The INGO network country score, 1978–1998. Social Forces, 87(4), 1711-1742. Human, Sherrie E., & Provan, Keith G. (2000). Legitimacy building in the evolution of smallfirm multilateral networks: A comparative study of success and demise. Administrative Science Quarterly, 45(2), 327-365. Juris, Jeffrey S. (2008). Networking futures: The movements against corporate globalization: Duke University Press. Kamens, David H. (1977). Legitimating myths and educational organization: The relationship between organizational ideology and formal structure. American Sociological Review, 208-219. Keck, Margaret, & Sikkink, Kathryn. (1998). Activists beyond borders: Advocacy networks in international politics. Ithaca, NY: Cornell University Press. Kilduff, Martin, & Krackhardt, David. (1994). Bringing the individual back in: A structural analysis of the internal market for reputation in organizations. Academy of management journal, 37(1), 87-108. Lake, David A. (2009). Hierarchy in international relations. Ithaca, NY: Cornell University Press. Lake, David A., & Wong, Wendy H. (2009). The politics of networks: Interests, power, and human rights norms. In M. Kahler (Ed.), Networked politics: Agency, power, and governance. Ithaca, NY: Cornell University Press. Long, J. Scott. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, CA: Sage Publications. Luxon, Emily M., & Wong, Wendy H. (2017). Agenda-setting in greenpeace and amnesty: The limits of centralisation in international NGOs. Global Society, 1-31. Lyle, Henry F., & Smith, Eric A. (2014). The reputational and social network benefits of prosociality in an andean community. Proceedings of the National Academy of Sciences, 111(13), 4820-4825. Miles, Robert H, & Cameron, Kim S. (1982). Coffin nails and corporate strategies: Prentice Hall. Mitchell, George E., & Stroup, Sarah S. (2017). The reputations of NGOs: Peer evaluations of effectiveness. The Review of International Organizations, 12(3), 397-419. Mooney, Pat. (2012). Civil society strategies and the stockholm syndrome. Development Dialogue, 61(September), 327-331. Murdie, Amanda. (2013). The ties that bind: A network analysis of human rights international nongovernmental organizations. British Journal of Political Science, 44(1), 1-27. ———. (2014a). Help or harm: The human security effects of international NGOs. Stanford, CA: Stanford University Press. ———. (2014b). The ties that bind: A network analysis of human rights international nongovernmental organizations. British Journal of Political Science, 44(01), 1-27. Murdie, Amanda, & Davis, David. (2012). Shaming and blaming: Using events data to assess the impact of human rights INGOs. International Studies Quarterly, 56(1), 1-16. Podolny, Joel M. (1993). A status-based model of market competition. American Journal of Sociology, 98(4), 829-872. Powers, Matthew. (2018). NGOs as newsmakers: The changing landscape of international news. New York, NY: Columbia University Press. Sauder, Michael, Lynn, Freda, & Podolny, Joel M. (2012). Status: Insights from organizational sociology. Annual Review of Sociology, 38, 267-283. Smith, Jackie. (2001). Globalizing resistance: The battle of seattle and the future of social movements. Mobilization: An International Quarterly, 6(1), 1-19. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 16 University of Bamberg Press ———. (2002). Bridging global divides? Strategic framing and solidarity in transnational social movement organizations. International Sociology, 17(4), 505-528. ———. (2014). Counter-hegemonic networks and the transformation of global climate politics: Rethinking movement-state relations. Global Discourse, 4(2-3), 120-138. Smith, Jackie, Chatfield, Charles, & Pagnucco, Ron. (1997). Transnational social movements and global politics: Solidarity beyond the state. Syracuse, NY: Syracuse University Press. Smith, Jackie, Gemici, Basak, Plummer, Samantha, & Hughes, Melanie M. (2018). Transnational social movement organizations and counter-hegemonic struggles today. Journal of World-Systems Research, 24(2), 372-403. Smith, Jackie, Plummer, Samantha, & Hughes, Melanie M. (2017). Transnational social movements and changing organizational fields in the late twentieth and early twentyfirst centuries. Global Networks, 17(1), 3-22. Smith, Jackie, & Wiest, Dawn. (2012). Social movements in the world-system: The politics of crisis and transformation. New York, NY: Russell Sage Foundation. Stroup, Sarah S., & Wong, Wendy H. (2017). The authority trap: Strategic choices of international NGOs. Ithaca, NY: Cornell University Press. Tallberg, Jonas, Sommerer, Thomas, & Squatrito, Theresa. (2013). The opening up of international organizations. Cambridge, UK: Cambridge University Press. Tarrow, Sidney. (2005). The new transnational activism. New York, NY: Cambridge University Press. Thrall, A. Trevor, Stecula, Dominik, & Sweet, Diana. (2014). May we have your attention please? Human-rights NGOs and the problem of global communication. The International Journal of Press/Politics, 19(2), 135-159. Willetts, Peter. (2006). The cardoso report on the un and civil society: Functionalism, global corporatism, or global democracy? Global Governance: A Review of Multilateralism and International Organizations, 12(3), 305-324. Wolfson, Todd, & Funke, Peter N. (2017). Contemporary social movements and media: The emergent nomadic political logic and its nervous system. Wong, Wendy H. (2012). Internal affairs: How the structure of NGOs transforms human rights. Ithaca, NY: Cornell University Press. Yanacopulos, Helen. (2015). International NGO engagement, advocacy, activism: The faces and spaces of change: Springer. Complexity, Governance & Networks – Vol. 8, No. 1 (2022), p. 1-24 DOI: http://dx.doi.org/10.20377/cgn-114 Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 17 DOI: 10.20377/cgn-114 University of Bamberg Press Appendix 1: Statistic N Mean SD Min Max # of Incoming Ties 4,938 0.607 1.711 0 30 # of Outgoing Ties 4,938 1.448 5.439 0 181 IGO Membership 3,651 2.053 3.159 0 44 HQ North 4,756 0.744 0.436 0 1 Organizational Age 4,756 24.672 21.687 0 174 Multiple Issue Areas 4,938 0.212 0.409 0 1 Country Count 4,919 21.837 28.605 0 188 Table A1: Summary statistics 1993 Table A2: Legitimacy ranking of TSMOs Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 18 DOI: 10.20377/cgn-114 University of Bamberg Press 2003 Table A2: Legitimacy ranking of TSMOs Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 19 DOI: 10.20377/cgn-114 University of Bamberg Press 1993 Figure A1: Network diagrams Green nodes are leading INGOs identified by Stroup and Wong (2017). Node size is proportionate to legitimacy (# of incoming ties). Isolates are removed. Graphs were generated by the Force Atlas algorithm in Gephi. Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 20 DOI: 10.20377/cgn-114 University of Bamberg Press 2003 Figure A1: Network diagrams Green nodes are leading INGOs identified by Stroup and Wong (2017). Node size is proportionate to legitimacy (# of incoming ties). Isolates are removed. Graphs were generated by the Force Atlas algorithm in Gephi. Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 21 DOI: 10.20377/cgn-114 University of Bamberg Press 2013 Figure A1: Network diagrams Green nodes are leading INGOs identified by Stroup and Wong (2017). Node size is proportionate to legitimacy (# of incoming ties). Isolates are removed. Graphs were generated by the Force Atlas algorithm in Gephi. Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 22 DOI: 10.20377/cgn-114 University of Bamberg Press Appendix 2: Exponential Random Graph Models I conduct the analysis of exponential random graph models (ERGMs) to estimate the chance of having a tie between two TSMOs. Although the dependent variable is different from the ZINB model discussed in the main text, it measures a similar outcome: what drives TSMOs to confer legitimacy on another organization. The advantage of ERGMs is that they can estimate the effects of different logics of tie activation, while explicitly taking account for network structures through Markov chain Monte Carlo simulations (Cranmer & Desmarais, 2011). ERGMs assume that the observed network is just one realization of many possible networks and simulate a network many times to evaluate if a tie between two nodes occurs randomly or systematically. First of all, I consider reciprocity. This network statistic estimates the likelihood of receiving a tie back when a TSMO sends a tie to another TSMO: when j sends a tie to i, what is the chance of i sending a tie back to j? In my interpretation, this represents the mechanism of direct legitimation. I also consider three types of network clustering. The first clustering is a mechanism by which TMSOs sends a tie because the tie recipient successfully signals its legitimacy through its association with a legitimate TSMO. In network terms, this relationship is called transitive: i sends a tie to j when both i and j send a tie to k. In my interpretation, this represents the case where k is locally seen as a legitimate TSMO, and thus being associated with k increases the chance of legitimation by peers. The second clustering is called cyclical: i sends a tie to j when j sends a tie to k and k sends a tie to i. I doubt this is a common phenomenon because deferential relationship should not be circular. I included this network statistic to evaluate how much more likely transitive relations are to occur relative to other kinds of triadic relations. Finally, I also consider network centralization because the network data, as suggested in Figures 1 and 2, have highly uneven distribution of ties among TSMOs.8 The primary purpose of this network statistic to help Markov chain Monte Carlo estimation converge by constraining the degree distribution of a network. To explore homophily effects, I included headquarter locations (global North) and United Nations Economic and Social Council (ECOSOC) affiliation, given that these are salient explanations in the regression analysis above. I chose ECOSOC since it is an IGO that is wellknown and used widely in the study of INGOs (Murdie, 2014a). I also account for organizational age, which estimates the chance of a tie when age difference between TSMOs increase by one year.9 Finally, I consider the issue areas of operation to evaluate functionalism. This is statistically equivalent to the inclusion of other organizational attributes, just like the case of homophily. I sperate this mechanism from others because in real-world practices the functional logic that motivates a tie among TSMOs should not be organizational similarities per se, but the functional needs to achieve something in the issue area in question. Table A3 reports the results of estimations. Because the size of a network varies by year, the results are also reported by year. Coefficients are changes in the log-odds of a tie, so they can be interpreted in the same way logistic regression coefficients are interpreted. The number of edges serve as a baseline likelihood; when simulation includes this network statistic alone, the likelihood of a tie equals the density of a network, which can be understood as a random assignment of realized ties in the network.10 8 I use indegree distribution since the network is a directed network. 9 More accurately, it estimates a change in log-odds of a tie. 10 Network density = # of the observed ties / # of all possible ties. In a way, this is like the “intercept” of regression analysis. Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 23 DOI: 10.20377/cgn-114 University of Bamberg Press Table A3: Results of ERGM for 1993, 2003, and 2013 Across all years, I find that a tie is much more likely to occur as a result of reciprocity than the other mechanisms I examined. Cyclical relations are statistically significant, but the coefficient is negative. This means that a tie is unlikely in a cyclical relationship, which is in line with my argument about how legitimation works. In fact, transitive relations have the second largest effect among independent variables. That is, association with a legitimate organization does help increase peer legitimacy. To look at the effects more carefully, I derive the chance of tie activation from the cumulative probability distribution of the latent variables of interest. In 2013, the chance of a tie in any given dyad (i.e., baseline likelihood) is 0.1%. The very low baseline is expected given that the network is both large and directed. The chance of a tie, however, increases to 16.7% when I consider reciprocity. That is, when A sends a tie to B, B sends a tie back to A with a 16.7% chance.11 This is significantly higher than that of transitivity (0.7%), in which association with a legitimate TSMO increases the chance of tie activation. Thus, networking increases the legitimacy of a TSMO directly rather than indirectly. Signaling through a third party seems a reasonable strategy at first glance, but perhaps because peers are more knowledgeable about other organizations around them, it does not have a substantively important effect. The chance of a tie via homophily is also quite low. The ECOSOC affiliation is negatively associated with tie activation, and the statistical significance is inconsistent (not significant in 2013). A tie is negligibly more likely among Northern TSMOs, and so is the effect of organizational age. However, the 11 The chance of reciprocity is very high in 1993. I suspect this is due to a smaller size of the network (441 vs. 509, 727). Note that network matrices expand quickly with an additional node. Complexity, Governance & Networks – Vol.8, No. 1 (2022), p.1-24 24 DOI: 10.20377/cgn-114 University of Bamberg Press chance of a tie via functionalism is slightly higher, 0.4%. Overall, while I confirm that different mechanisms are at work in tie activation, direct networking is by far the most effect way of increasing legitimacy. Table A3 summarizes the estimated chance of tie activation. Year Baseline Reciprocity Northern TSMO Issue area 1993 0.1% 67.4% 0.1% 0.3% 2003 0.1% 54.2% 0.2% 0.7% 2013 0.1% 15.1% 0.1% 0.4% Table A4: Probability of a tie based on reciprocity and shared attributes