Collaborative Propensities Among Transnational NGOs Registered in the United States

Government–NGO Relations: Theory and Evidence

Government–NGO relations have become a topic of much attention and controversy within public administration (Batley, 2011; Perry, 2012) and nonprofit studies (Salamon, 2006; Salamon & Anheier, 1998). Much of this research draws on resource dependence (Pfeffer & Salancik, 2003) and institutional (DiMaggio & Powell, 1983) theories to examine NGO collaborations.

According to resource dependence theory, organizations are subject to external control when they depend on an external actor for a large proportion of a critical resource. Studies of NGO collaborations typically emphasize government–NGO relations that involve significant resource asymmetry in contracting relationships. This narrow emphasis may explain the emergence of more recent scholarship examining the roles of transnational NGOs in public governance suggesting, that resource dependence scholarship may overstate its case. Broadly speaking, resource dependence theory tends to overemphasize the role of external factors, underappreciate the extent of transnational NGO autonomy (Gerstbauer, 2010; Mitchell, 2012), and oversimplify government–NGO collaborations, generally by assuming that NGOs occupy a subordinate role (Najam, 2000) even while an increasing body of evidence fails to support this assumption (Mcloughlin, 2011). ³ Reliance on government funding does not necessarily put transnational NGOs in a subordinate position (Rose, 2011) as most organizations enjoy significant latitude to exercise strategic choice within their environments (Batley & Rose, 2011). If nothing else, the substantial administrative costs associated with contract or grant supervision often afford NGOs substantial freedom of action (Deflin Jr., & Tang, 2008). Moreover, resource dependence approaches typically emphasize the relative amounts of resources provided to organizations, but the manner in which resources are provided may be equally relevant. For example, multiple year grants or contracts may reduce the likelihood of cooptation (Deflin Jr., & Tang, 2008), increase transparency, and improve coordination (Cooley & Ron, 2002). Finally, organizations may be more likely to yield control to donors due simply to donor demand rather than to market characteristics or resource dependence (Barman, 2008).

According to institutionalism, subordinate organizations imitate their dominant counterparts through a process of isomorphism. Evidence of institutional isomorphism among domestic and transnational NGOs can be found within a variety of contexts, especially among organizations that receive government funding. For example, such NGOs are more likely to adopt government sanctioned strategies (Molenaers, Dewachter, & Dellepiane, 2011) and to comply with government accounting and reporting standards, even in the absence of meaningful enforcement (Verbruggen, Christiaens, & Milis, 2011). Some research suggests that government funding increases political advocacy and reduces social capital building (Moulton & Eckerd, 2011), although other research suggests no significant relationship exists (Child & Grønbjerg, 2007). However, as with resource dependence theory, institutionalist scholarship may also overstate its case. For example, differentiation in tactical response, driven by factors including path dependence and variation in resource environments, mute the isomorphic tendency among NGOs (Ramanath, 2009).

Despite research indicating that some NGOs view governments as restrictive, bureaucratic, prone to interfere, and untrustworthy, while some governments view NGOs as lacking capacity, occasionally corrupt, and comparatively less committed (Alam, 2011), NGOs and governments have nevertheless been shown to collaborate productively (Sansom, 2011). Fruitful collaboration is most likely when the participants have complementary mandates, share values, goals, visions, and interests, have achieved domain clarity and consensus, have developed embedded interpersonal ties and mutual interpersonal trust (Tsasis, 2009), and when the NGO has technical expertise, extensive social and political networks, and community embeddedness (Bano, 2011). Over time government–NGO collaborations are characterized by increasing trust (Alexander & Nank, 2009), increasing use of contracts (Sansom, 2011), and an evolution from principal-agency to principal-stewardship that fosters mutual goal alignment (Van Slyke, 2006).

From a transnational governance perspective, government–NGO collaborations constitute only one type of collaboration among many, and one that takes place within a broader sociopolitical context. The persistent emphasis on government–NGO relations within public administration reflects, inter alia, both an ongoing interest in the phenomenon of increased government outsourcing and an increasing appreciation of the crucial roles NGOs play in public governance, whether complementary, supplementary, or adversarial to government policy (Young, 2006).

Compared with public administration scholarship drawing mainly upon resource dependence and institutional theories, international relations scholarship has been much less concerned with government–NGO collaborations and instead has focused on the impact and motives of transnational NGO behavior, particularly whether organizations are driven by “principled” considerations as expressed in their missions, by “self-interested” factors stemming from their need to secure resources, or more recently, by both principles and self-interest (Mitchell & Schmitz, 2012).

Early international relations research animated by social constructivism and its emphasis on shared norms and values identified transnational NGOs as principled actors (Keck & Sikkink, 1998) that promote norm institutionalization at the global level (Glasius, 2006; O’Brien, Goetz, Scholte, & Williams, 2000; Price, 1998) and collaborate or clash with domestic actors in efforts to implement norms in local contexts (Hertel, 2006; Schmitz, 2006; Tarrow, 2005; Wapner, 1995). According to this view, a principled commitment to upholding and enforcing shared norms and values empowers NGOs with ideational resources that are leveraged to influence other actors even though NGOs typically lack the financial and coercive capabilities of corporations and states, respectively (Risse, 2006).

Later research programs depart from these optimistic characterizations, modeling transnational NGOs instead as interest groups (Bloodgood, 2011) or firms (Prakash & Gugerty, 2010) and arguing that NGO behavior is consequently more “self-interested” than “principled” than prior international relations scholarship assumed. Critics argue that the growing number of transnational NGOs coupled with a proliferation of short-term contracts forces organizations to abandon their principled commitments to their missions in the pursuit of self-interested financial security (Cooley & Ron, 2002). Within this perspective, economic factors including resource dependence (Rauh, 2010), government contracting practices (Edwards & Hulme, 1996a, 1996b; Robinson, 1997; Smith & Lipsky, 1998), and competition for resources undermine collaboration and lead to normatively undesirable conditions of organizational dysfunction (Cooley & Ron, 2002). For example, organizations may select issues on the basis of media exposure and donor support rather than genuine beneficiary need (Bob, 2005, 2010) as transnational NGOs are compelled by the imperative of economic survival to become donor-driven, rather than need-driven organizations.

The implications of resource dependence, institutionalist, and international relations theories raise fundamental questions about the autonomy and identity of transnational NGOs. Many transnational NGOs rely upon their identity as principled, independent civil society organizations to establish their legitimacy in lieu of formal democratic accountability mechanisms (Erman & Uhlin, 2010). While modern state agencies derive legitimacy from their political accountability and market institutions derive legitimacy from their efficiency and ability to generate wealth, transnational NGOs traditionally derive legitimacy from their principled commitments to their missions and supposed independence from state and market institutions and forces. Collaborations with corporate and government actors, particularly if they appear to be financially motivated, risk undermining this basis of legitimacy. To better understand the perspectives of NGO practitioners, the remainder of this article examines the collaborative propensities of transnational NGOs and the benefits and obstacles to collaborations as perceived by NGO leaders.

Data and Method

Most empirical studies concerning NGO collaborations use subnational samples of domestic organizations or focus on a single sector of activity. ⁴ This article instead examines data from interviews with top leaders of transnational NGOs registered in the United States and spanning all major sectors of NGO activity. Although transnational NGOs vary widely by orientation, sectoral focus, geography, and organizational structure (Kaldor, 2003; Vakil, 1997; Willetts, 2002), they share a common definition as independent, voluntary, charitable, nonviolent and professional public interest organizations that operate across national boundaries (Kaldor, 2003; Martens, 2002; Willetts, 2002).

Within the United States the Internal Revenue Service (IRS) required organizations with 501(c)(3) tax-exempt status and more than US$25,000 in annual revenues to file IRS Forms 990, which are available for public inspection. Based on these forms, the National Center for Charitable Statistics identified 7,218 transnational NGOs in the United States. According to their data, these organizations spent US$29 billion in 2009 alone (Roeger et al., 2011). The combined average annual expenditures for organizations sampled for the interview study is about US$20 billion, accounting for more than two-thirds of total U.S. transnational NGO expenditures. ⁵

To identify leaders for the interview study, researchers initially sampled 182 organizations from a database of 334 transnational NGOs rated by Charity Navigator. ⁶ Charity Navigator is an online nonprofit rating agency in the United States that evaluated organizations with 501(c)(3) tax-exempt status from the IRS, at least 4 consecutive years of IRS Forms 990 available and public support greater than US$500,000 during their most recent fiscal years. Organizations that reported zero fundraising costs or that were overwhelmingly funded through government grants or fees for services were excluded, along with private foundations, hospitals, hospital foundations, private universities, colleges, community foundations, and public broadcasting stations. The selection criteria effectively exclude very small organizations, new organizations, organizations unreliant on the general public for funding, and more generally, organizations not widely regarded as transnational NGOs, such as hospitals and public broadcasting stations. The size criterion introduces a bias toward oversampling larger, more prominent organizations. For example, median total revenue for sampled organizations is US$13,042,579 compared with US$174,214 for all transnational NGOs recognized by the IRS. ⁷

Researchers employed stratified random sampling to ensure a diverse sample and proportionate representation by size, sector, and financial characteristics. ⁸ One-hundred twenty-three interviews were completed with leaders from the initial sample and 29 replacements were added for a total sample size of 152 organizations. The overall response rate was 68%. ⁹ In the final sample, 81% of respondents were the CEOs, presidents or executive directors of their organizations, 12% were vice presidents, and only 7% were below the level of vice president. Table 1 provides a statistical description of the final sample.

OPEN IN VIEWER

Table 1.

Sample Description.

Sector	Count	Percent
Environment	22	14
Human rights	21	14
Humanitarian relief	32	21
Sustainable development	64	42
Conflict resolution	13	9
Total	152	100
Size
Small	56	37
Medium	64	42
Large	32	21
Total	152	100
Function
Advocacy	34	22
Service	70	46
Both	48	32
Total	152	100

The interview protocol was not designed to test specific models or theories, but to collect baseline information about how leaders understand governance, goals, strategies, activities, effectiveness, accountability, collaboration, transnationalism, communication, and leadership. Nearly all of the interview questions solicited open-ended responses where leaders were free to speak in the own words. Specific protocol questions were developed in consultation with practitioners and pilot tested during workshops in the United States and abroad.

The interviews were conducted at leaders’ preferred locations, usually their offices. Interviewers guaranteed leaders confidentiality to promote candor. The interviews averaged 82.5 min; the shortest was 32 min and the longest was 153 min. Interviewers ultimately collected about 209 hr of digital recording. All the interviews were recorded with the permission of the interviewees and subsequently transcribed.

A team of graduate students coded the complete interview transcripts using computer-aided qualitative data analysis software. The research team created the codebook based on prior expectations, inductive readings of initial transcripts for emergent themes, and input from the coding team. The final codebook contains 413 individual codes. To measure the overall degree of intercoder agreement across all 413 codes, 10 complete in-sample interviews were each coded twice by separate coders. Researchers then calculated scores measuring the percentage of agreement between the two coders of each interview transcript and then averaged the 10 scores. A value of zero indicates complete disagreement, whereas a value of one indicates complete agreement. The overall intercoder agreement score is 0.80, indicating satisfactory intercoder agreement.

Interviewers asked NGO leaders a series of questions about their organizations’ networks and partnerships and whether their organizations specifically collaborated with “NGOs, international organizations, governments or corporations.” Accordingly, researchers created four qualitative codes to indicate the presence of collaborations with each of these four types of actors: NGOs (for example, other NGOs in their field), governments (for example, US or foreign government agencies), corporations (for example, for-profit businesses), and international organizations (for example, intergovernmental organizations such as UN agencies and regional development organizations). These four qualitative codes may also be regarded as four binary response variables indicating the presence of each type of collaborative relationship. Respondents interpreted the meaning of collaboration for themselves, and owing to the large variety of collaborations in which a typical NGO is engaged, interviewers did not probe respondents about the specific form of each collaboration. In practice, leaders’ understandings of collaboration were consistent with the scholarly and professional literature (Fosler, 2002; Sink, 1998; Wood & Gray, 1991), within which

collaboration generally involves a higher degree of mutual planning and management among peers; the conscious alignment of goals, strategies, agendas, resources, and activities; an equitable commitment of investment and capacities; and the sharing of risks, liabilities, and benefits. In this sense, collaboration is akin to, if not the same as, partnerships (including “public-private partnerships”), alliances (including “strategic alliances”), and similar kinds of associations (Fosler, 2002, p. 19).

Overall, and consistent with prior scholarship on domestic NGOs (cf. Suárez, 2010), respondents indicated that their organizations are most likely to collaborate with other NGOs (86%). This is followed by governments (52%), corporations (52%) and international organizations (37%).

Collaborative propensity describes a transnational NGO’s tendencies to collaborate with other NGOs, governments, corporations and international organizations. To identify the collaborative propensities of transnational NGOs, leaders’ response patterns were analyzed using latent class analysis, a model-based analogue to cluster analysis. Latent class analysis posits a latent categorical variable that assigns organizations to clusters to maximize within-group homogeneity and between-group heterogeneity (Ahiquist & Breunig, 2012). However, unlike traditional cluster analysis, latent class analysis enables formal hypothesis testing and provides statistical information for the identification of exemplary cases, which is helpful for selecting illustrative qualitative quotations in mixed-method research.

More technically, latent class analysis generates a latent categorical variable that explains or minimizes the statistical association among a set of response variables. Within each category of the posited variable the response variables are thus mutually or “locally” independent (Goodman, n.d.; Lazarsfeld, 1950; Lazarsfeld & Henry, 1968; Magidson & Vermunt, 2003; McCutcheon, 1987). Local independence is assessed by examining a latent class model’s bivariate residuals, which quantify residual association left unexplained by the model. A bivariate residual between any two binary response variables greater than 3.84 indicates an unacceptable local dependence at the 95% confidence level.

The overall level of model fit for a latent class model is quantified by the L² statistic, which is interpreted similarly to a χ² statistic. However, because the goal of latent class analysis is to explain or minimize statistical association and the L² statistic is a measure of association, a lower, not higher statistic indicates better model fit. An L² p-value greater than 0.05 indicates adequate model fit at the 95% confidence level.

Latent class analysis begins by testing the fit of a model with only one cluster and then incrementing the number of clusters until doing so fails to result in a statistically significant improvement in the L² statistic. The marginal improvement in model fit associated with an increment in the number of clusters is quantified by the marginal L² reduction statistic. A smaller value indicates a smaller improvement and a marginal L² reduction p-value greater than 0.05 indicates that a marginal increment does not improve model fit at the 95% confidence level.

The models summarized in Table 2 contain all four binary response variables listed above (NGOs, governments, corporations, and international organizations) and differ only according to the number of posited categories of the latent variable, or equivalently, the number of clusters. ¹⁰ Model 1 posits one cluster, Model 2 posits two clusters, and Model 3 posits three clusters. The latent class model that fits the data, exhibits local independence, and cannot be significantly improved upon is the preferred latent class model.

OPEN IN VIEWER

Table 2.

Models for Collaborative Propensity.

Model name	Model type	LL	L 2	P-value	Marginal L2 reduction	P-value	Classification error	Local independence
Model 1	One cluster	−345.51	24.15	0.03				No
Model 2	Two clusters	−335.68	4.49	0.77	0.81	0.00	0.14	Yes
Model 3	Three clusters	−335.27	3.68	0.97	0.18	0.97	0.34	Yes

The null hypothesis that there is only one cluster is rejected according to the L² p-value for Model 1, which indicates inadequate model fit. Model 2 provides adequate model fit (according to the L² p-value), explains most of the association among the response variables (according to the marginal L² reduction statistic), represents a statistically significant improvement over Model 1 (according to the marginal L² reduction p-value), and contains no local dependencies (according to the model’s bivariate residuals, which are all less than 3.84). ¹¹ The statistics for Model 3 show that incrementing the number of clusters to three does not result in a statistically significant improvement in model fit over Model 2 (according to the marginal L² reduction p-value). Thus, Model 2 is the preferred latent class model, indicating the existence of two collaborative propensities based on leaders’ response patterns.

The key quantities of latent class analysis are the latent class probabilities, which provide estimates of the size of each cluster, and the conditional probabilities, which describe the profile of each cluster (Magidson & Vermunt, 2003; McCutcheon, 1987). These quantities are presented in Table 3. ¹²

OPEN IN VIEWER

Table 3.

Model 2 Profile.

	Cluster 1	Cluster 2	P-value
Cluster size	0.63	0.37
NGOs	0.81	0.95	0.11
Governments	0.29	0.92	0.00
Corporations	0.38	0.76	0.00
International organizations	0.23	0.60	0.01
			n = 141

As shown by the latent class probabilities in Table 3, about 63% of respondents’ organizations fall into Cluster 1, whereas the remaining 37% fall into Cluster 2. Transnational NGOs in Cluster 1 are much less likely to collaborate with governments, corporations, and international organizations than those in Cluster 2.

To facilitate interpretation, the following two sections provide exemplary qualitative evidence retrieved from the interview transcripts separately for those assigned to each cluster. The exemplarity of a leader’s statements about his or her organization’s collaborations is defined as the probability that the leader belongs to the cluster to which he or she has been assigned. A probability value close to 1 indicates exemplarity, whereas a value close to 0 indicates atypicality. Because the assignment of organizations to clusters is probabilistic or “modal” (Magidson & Vermunt, 2003; McCutcheon, 1987), the underlying posterior membership probabilities are provided parenthetically in the text to indicate the probability that each qualitative quotation exemplifies the cluster it is meant to illustrate. The overall classification errors introduced by modal assignment are displayed for each model in Table 2. Leader’s statements about their organizations’ collaborations have been interpreted within the contexts of the complete interview transcripts, which are organized in the computer-aided qualitative data analysis software for efficient querying.

Cluster 1: The Independent Collaborative Propensity

NGOs are classically defined as operating independently of state and market institutions (Martens, 2002; Vakil, 1997; Willetts, 2002). The profile for Cluster 1 displayed in Table 3 describes such transnational NGOs that prefer to collaborate primarily with other NGOs while generally, although not exclusively, avoiding collaborations with corporations, governments, and international organizations.

Among advocacy and human rights organizations this appears to be due largely to concerns over maintaining organizational independence and legitimacy. For example, one respondent from an environmental NGO (ID = 9, P(Cluster 1) = 0.98) said that his or her organization would only partner with other NGOs

in a extremely minimal way. Never with governments. And corporations, again, extremely limited. Always stuff that can be really isolated. If 3M or GE or whoever said, “We have got a plan,” we would not touch it. We just would not touch it. We could be in an economic crisis and we would not touch it.

“We are totally independent,” the leader explained. “We will praise you when you do the right thing, we will go after you when you do the wrong thing, but we are not part of your team. Ever.” As the leader reveals, intersectoral collaborations can be seen as potentially compromising to an NGO’s ability to credibly hold government and corporate actors accountable for their actions.

A respondent from a small human rights NGO (ID = 24, P(Cluster 1) = 0.98) admitted that although “there are times when we have friendly conversations with . . . government offices, it is not too often.” As a matter of policy, the respondent said, “we are not supposed to take money from any corporation or government,” which could potentially undermine the organization’s ability to sustain an adversarial stance toward government agencies and corporations. The respondent explained that the goal of her organization is to make “positive change in the world as we see it, to move away from the neoliberal economic model toward one of greater equality and justice.” However, “the power of the U.S. government and the big corporations,” which “certainly have dominated the world,” pose major obstacles to mission achievement, according to the respondent’s perspective.

Other leaders are simply cautious in their attitudes toward collaborations with state and market actors. A respondent from a small human rights organization focused on lesbian, gay, bisexual, and transgender (LGBT) issues (ID = 31, P(Cluster 1) = 0.99) collaborates with “labor groups” in Latin America to combat workplace discrimination, “HIV groups” in Africa to decriminalize LGBT sexuality and “feminist and women’s rights organizations because the issues of sexuality have such crossover, and frankly there are a lot of lesbians in each community doing work. Some of us have been around and have known each other for a very long time.” The organization has not yet found significant opportunities to collaborate with governments, which the leader believes are often responsible for LGBT “human rights violations” and “should be [held] accountable” from the outside.

Leaders are also concerned about the operational competencies of non-civil society organizations and are correspondingly selective. For example, a leader from a sustainable development NGO focused on children (ID = 95, P(Cluster 1) = 0.93) admitted that “we work with a lot of corporations but we don’t work with the UN, don’t work with the World Bank, don’t work with USAID.” Although the organization’s corporate partners support its educational and fundraising activities, the leader noted that “we have some very interesting thoughts about the role of government in [providing] social services for the poor. I’m not sure governments are very effective.” Rather than partnering with government agencies, the respondent’s organization focuses its support on grassroots civil society organizations, which it regards as more innovative than government agencies and therefore better able to address the needs of children in the developing world.

Although clearly NGOs exhibiting the independent collaborative propensity differentiate between state and market institutions, they nevertheless tend to place government agencies, for-profit corporations, and international organizations in the same category. Indeed, the profile for the independent propensity (Cluster 1) displayed in Table 3 shows that these NGOs are similarly averse to collaborations with state and market actors.

Cluster 2: The Interdependent Collaborative Propensity

Although NGO scholars widely define NGOs as operating independently of state and market institutions, this definitional characterization belies the empirical reality that many NGOs frequently collaborate with non-civil society actors. These NGOs exhibit interdependence with states and markets, despite the percieved risks for civil society organizations (Salamon & Anheier, 1998). As shown in Table 3, interdependent transnational NGOs are more than three times as likely to collaborate with governments, twice as likely to collaborate with corporations and more than twice as likely to collaborate with international organizations relative to independent transnational NGOs.

One environmental NGO, for example, regularly collaborates with governments and partners with major corporations to certify their products (ID = 14, P(Cluster 2) = 0.92):

We have substantial relationships with companies like IKEA, Kraft Foods and Lavazza Coffee, and some of the larger timber companies. Beyond just being certification clients, at our gala every year we recognize our corporate partners. They are part of the whole thing, the recognition thing. We call attention to some of those companies each year. We have formal [memoranda of understanding] with government ministries in various countries . . . to work toward similar purposes in the environmental sector or the tourism sector, for example.

Such collaborations are valuable for the organization because they provide “funding, support for the policies that we espouse, recognition of our work, and introductions to other relationships with corporate entities,” the respondent explained.

A leader from another environmental NGO (ID = 20, P(Cluster 2) = 0.69) candidly explained that his or her organization works “with governments, not for governments” despite concerns about graft and corruption within foreign government agencies:

Since it’s anonymous, I will say this . . . my strategy is that we work with governments, not for governments. Working for governments is suicide in terms of your money, in terms of your project, in terms of them stealing everything from you . . . but that doesn’t mean that we don’t work with them. Purposely, we create a program where we don’t have to ask for permission. We don’t go and say, “Oh, we’re going to be working in this country,” because then they want to get more involved and get a “little bite.” They want their part.

“We try to avoid that,” the respondent explained, by exercising discretion to identify partners

far enough down the feeding chain that you are not dealing with the politicos; you’re dealing with natural resource people. Again, it’s a small universe. People who are in the forestry aren’t there for politics or for money. Usually you can find people within the ministries [who are] working on similar projects and [solicit] their opinions about everybody who works at those organizations. So you can say, “Oh, this guy is good but he gets all of his money from Colombian drug lords.” So you have their contact information and at least somebody’s opinion concerning what kind of work they are doing and how well they are respected. The U.S. side is a little bit different, but not really different. We go out and similarly look for groups that are doing things that we like and we support them. They have to be a collaboration-based organization. They have to have multiple issues. Once we’ve identified them, then we offer them a menu of services.

Respondents differentiate between U.S. and foreign government contexts and consistently imply that outright graft and corruption are much more prevalent abroad. However, their transnational perspectives view U.S. and foreign government agencies as instantiations of the same phenomenon—state actors as distinct from nonstate actors—and because transnational NGOs typically intervene not when governments succeed, but when they fail, NGO leaders are apt to exhibit caution when forming collaborations with government agencies whether at home or abroad.

For example, a leader from a humanitarian relief organization (ID = 60, P(Cluster 2) = 0.69) said that in addition to their corporate partners, “we have partners that are governments,” but due to graft and corruption concerns, they are chosen very selectively:

You either have to be a nonprofit/NGO group to be a partner with us or a government. Now, it’s not every government. Domestically we work with a lot of different governments, like an aging services department of a county or something like that. We do a lot of that kind of stuff to help them get product and supplies to help their groups. They have very limited budgets, so we feel like that fits. Some governments and some countries we would never work with because the worry about it getting resold or used for something else is too great. So we would rather find an NGO to work with, but we will work with governments.

Incidentally, although the leader’s organization does not directly receive government funding, the respondent explained: “We decided to work with partners that get government funding and not to get it ourselves. The processes are just [too] long and difficult.”

A leader from a large humanitarian relief NGO (ID = 45, P(Cluster 2) = 0.69) explained how his or her organization collaborates intimately with corporations and NGOs, and works with government agencies largely out of operational necessity. Adopting a corporate idiom, the respondent said:

In the global humanitarian assistance program, we basically operate as an intermediary between the wealth and the capacity of pharmaceutical and medical equipment companies and resource-starved communities . . . We seek to establish a strategic connection to their corporate philanthropy and their corporate sustainability and responsibility programs . . . We work with local healthcare providers and intermediaries like the ministries of health and clinic operators to be our in-country partners and so our objective is to supply them and strengthen them . . . We have some places in which the government is an in-country partner—the ministry of health. To the extent that you are working in any of these countries you at least have to have a relationship with the ministry of health. But it would be relatively unusual for us to have a program partnership with the ministry of health.

The leader explained that

the reason for that is that there just are not too many ministries of health in the developing world that run very efficiently. So generally you can find a more effective and efficient partner locally, whether it is a hospital system or a clinic system or a faith-based organization or another international NGO. If you give something to the ministry of health, what value added are they going to be as opposed to what kind of a drain are they going to be? In most cases they are going to be a drain.

There are of course exceptions. A leader from a sustainable development NGO (ID = 133, P(Cluster 2) = 0.92) described collaborations with NGOs, corporations, and a government agency. Typically “we don’t work with governments,” the respondent said, “with only one exception. In Guatemala one of our local partners is the Ministry of Health. It just happened that way” because “they’re able to deliver service at a high level.”

Respondents primarily differentiate between civil society organizations and non-civil society organizations in their conceptualizations of their organizations’ collaborations and tend to interpret their collaborations in loose operational, rather than formal contractual terms (cf. Gazley, 2007). Most respondents’ organizations adopt an independent approach, preferring primarily to collaborate with other NGOs, whether to safeguard the legitimacy derived from their independence or because of concerns over the integrity, efficiency or competency of non-civil society organizations to successfully address operational challenges. A minority is clearly more interdependent with state and market institutions, as these NGO leaders are more willing to blur the line circumscribing civil society, whether out of necessity or to obtain increased funding or recognition. Those NGO practitioners whose organizations exhibit a high propensity for interdependence appear to represent a comparatively smaller subset of the transnational NGO community, although a few notes of caution are warranted. First, some respondents exhibiting the independent collaborative propensity may indeed be engaged in financial relationships with government agencies but simply do not regard their relationships as collaborative. Second, the sampling population excluded organizations overwhelmingly funded through government grants, reducing the proportion of interdependent organizations. Finally, transnational NGOs typically serve foreign beneficiaries and thus may be less likely than domestic NGOs to receive U.S. state and local government support.

Benefits and Obstacles to Collaboration

The benefits of interorganizational collaborations may include increased effectiveness, efficiency, learning, service quality, competitive advantage, expertise, market access, and risk diffusion (Gazley & Brudney, 2007). NGOs seek collaborations for a variety of reasons. NGOs collaborate to improve service delivery, to improve the likelihood of organizational survival, and to enhance institutional legitimacy and competitive advantage (Sowa, 2009). Many collaborations are driven by the need to secure resources, and organizations that perceive environmental uncertainty are more likely to collaborate (Foster & Meinhard, 2002). In government–NGO relationships specifically, the benefits of collaborations may include increased funding, improved service quality and availability, and stronger community building (Gazley & Brudney, 2007). NGOs receiving government funding are more likely to report increased resources as a benefit to collaboration, but less likely to report improved community relations or mission fulfillment (Gazley & Brudney, 2007).

Interviewers asked transnational NGO leaders the following question about the benefits of collaborations in general: “What kinds of benefits, if any, do you see resulting from networks and the formation of partnerships?” Researchers developed ten qualitative codes to capture leaders’ open-ended responses: better results (for example, increased effectiveness and efficiency), broader programs (for example, can help more people or make a larger impact), increased funding (for example, more money), better understanding, (for example, better understanding of substantive issues), local capabilities (for example, collaborating with others with local expertise), learning (for example, exchange of best practices), enhanced visibility (for example, raising organizational profile through collaborations), better access (for example, resolving registration challenges, meeting government requirements, securing accreditation, and obtaining invitations), legitimacy or credibility (for example, improved legitimacy or credibility with donors, policymakers, and peers), and legal (for example, meeting legal requirements and obtaining legal expertise).

As shown in Table 4, better results was the most frequently mentioned benefit to collaborations. Leaders adopting an interdependent collaborative propensity were more likely to have mentioned enhanced legitimacy and credibility as a benefit of collaboration, as these organizations may be more amenable to insider strategies that rely upon acceptance rather than independence as a basis for establishing organizational legitimacy. Interdependent transnational NGOs were also more likely to have mentioned increased funding as a benefit to collaboration, suggesting that financial self-interest may motivate many of these NGOs to engage in collaborative relationships outside of the sphere of civil society.

OPEN IN VIEWER

Table 4.

Collaborative Propensity and Benefits of Collaborations.

	Collaborative propensity
	Independent		Interdependent
	Count	Percent	Count	Percent	χ2(1)	P-value
Better results	49	66	32	62	0.29	0.59
Broader programs	28	38	27	52	2.46	0.12
Increased funding	25	34	28	54	5.04	0.03
Better understanding	27	36	17	33	0.19	0.66
Local capabilities	18	24	20	38	2.90	0.09
Learning	15	20	11	21	0.01	0.90
Enhanced visibility and presence	16	22	8	15	0.77	0.38
Better access	13	18	10	19	0.06	0.81
Legitimacy and credibility	5	7	11	21	5.71	0.02
Legal	1	1	3	6	1.94	0.16
Valid n	74		52

The obstacles to interorganizational collaborations are similarly manifold. They may include mission drift, decreased autonomy and accountability, cooptation, financial instability, evaluation difficulties, inadequate time and resources (Gazley & Brudney, 2007), “problems with governance, roles, or responsibilities guiding relationship[s],” the “complexity of partnership forms and structures,” the “focus on competition versus collaboration,” and “changing mission and objectives” (Babiak & Thibault, 2009, pp. 126-131).

Interviewers asked transnational NGO leaders the following question about obstacles to collaborations in general: “Are there obstacles or challenges that arise in the formation of partnerships and networks?” Researchers developed seven qualitative codes to capture leaders’ open-ended responses: incompatibility of missions (for example, concerns about the consistency or compatibility of missions, agendas, goals or strategies), muddled management (for example, unclear roles and responsibilities, poor accountability, uncertainty about leadership, and a lack of partnership experience), reduction of resources (for example, loss of resources to partners and less funding individually or for the field), time (for example, too time consuming in comparison with going it alone), organizational cultures (for example, differences between organizational cultures and conflict over how to manage the collaboration), loss of control (for example, too many concessions and concerns over intellectual capital or ownership of ideas), and lack of confidence (for example, concerns about reliability and trustworthiness).

As shown in Table 5, incompatibility of missions was the most commonly mentioned obstacle, emphasizing leaders’ overarching concern for safeguarding their missions in collaborative relationships. Leaders adopting an interdependent collaborative propensity were more likely to have mentioned organizational cultures as an obstacle to collaboration, indicating a significant cultural gap between NGOs and non-civil society organizations. ¹³

OPEN IN VIEWER

Table 5.

Collaborative Propensity and Obstacles to Collaborations.

	Collaborative propensity
	Independent		Interdependent
	Count	Percent	Count	Percent	χ2(1)	P-value
Incompatibility of missions	39	55	25	48	0.56	0.45
Reduction of resources	19	27	16	31	0.24	0.63
Muddled management	19	27	16	31	0.24	0.63
Organizational cultures	14	20	20	38	5.27	0.02
Time	19	27	13	25	0.05	0.83
Loss of control	13	18	14	27	1.30	0.25
Lack of confidence	14	20	12	23	0.20	0.65
Valid n	71		52

Discussion

Independent transnational NGOs avoid collaborations with state and market actors as a means of preserving their outsider status and because of concerns over the integrity, efficiency, or competency of non-civil society organizations. These organizations are also less likely than interdependent NGOs to perceive increased funding, legitimacy, and credibility as benefits of collaborations. Finally, independent NGOs are less likely to perceive obstacles to collaborations rooted in organizational cultures.

Interdependent transnational NGOs collaborate with state and market actors out of necessity, to obtain increased funding and to receive recognition, despite concerns over graft and corruption when working with foreign government agencies in the developing world. Compared with independent NGOs, interdependent NGOs are more likely to perceive increased funding, credibility, and legitimacy as benefits of collaboration. Finally, interdependent NGOs are more likely than independent NGOs to perceive organizational cultures as an obstacle to collaboration, indicating that cultural differences exist between civil society and non-civil society organizations.

Furthermore, Table 6 shows that collaborative propensity is statistically associated with organizational sector. ¹⁴ The independent propensity is most common among conflict resolution and human rights organizations, where an outsider status may be regarded as an asset connoting neutrality and legitimacy. The interdependent collaborative propensity is most common among sustainable development, environmental, and humanitarian relief organizations, which are more likely to receive official funding from government agencies (Edwards & Hulme, 1996b). Collaborative propensity is not statistically associated with organizational function (n = 141, χ² = 0.72, p = .70) or size (n = 141, χ²(2) = 3.62, p = .16), although it should be noted that a size effect may be undetectable due to the exclusion of organizations with annual public support less than US$500,000. ¹⁵

OPEN IN VIEWER

Table 6.

Sector and Collaborative Propensity.

	Sector
	Environment	Human rights	Humanitarian relief	Sustainable development	Conflict resolution
Collaborative propensity
Independent	11	17	18	30	10
	52%	81%	64%	51%	83%
Interdependent	10	4	10	29	2
	48%	19%	36%	49%	17%
Total count	21	21	28	59	12
Total percent	100	100	100	100	100
	n = 141, χ2 (4) = 9.37, p = .05

In addition, many transnational NGOs registered in the United States have a religious or faith-based orientation. Despite efforts during the administration of President George W. Bush to cultivate deeper government relationships with faith-based organizations, such NGOs are actually more likely to adopt an independent collaborative propensity. The relationship between faith-based and collaborative propensity is shown in Table 7.

OPEN IN VIEWER

Table 7.

Faith-Based and Collaborative Propensity.

	Faith-based
	No	Yes
Collaborative propensity
Independent	53	31
	53%	79%
Interdependent	47	8
	47%	21%
Total count	100	39
Total percent	100	100
	n = 139, χ2(1) = 8.44, p = .00

Compared with secular NGOs, respondents from faith-based organizations are 19% less likely to have mentioned better results as a benefit of collaborations (n = 128, χ² (1) = 3.98, p = .05). In addition, these leaders are 24% less likely to have mentioned time (n = 128, χ² (1) = 6.93, p = 0.01) and 18% less likely to have mentioned organizational cultures as obstacles to collaboration (n = 128, χ² (1) = 4.33, p = .04). These results suggest that although faith-based NGOs are less likely to view inadequate time and cultural differences as barriers, they also see fewer pragmatic benefits of collaboration than do secular organizations. ¹⁶

A final consideration when analyzing coded interview data is whether the results truly reflect respondents’ views and not error introduced through researcher bias or baiting. Coder bias would be evidenced by statistical association between collaborative propensity and coder identification, while interviewer baiting would be evidenced by statistical association between collaborative propensity and interviewer identification. Appropriate independence tests do not present evidence of coder bias (n = 141, χ² (4) = 8.18, p = 0.09) or interviewer baiting (n = 141, χ² (6) = 5.65, p = 0.46) at the conventional 95% confidence level, suggesting that neither source of error is present.

Heterogeneity in collaborative propensity reflects deep underlying philosophical differences within the NGO community concerning the appropriateness of maintaining relationships that extend beyond the sphere of civil society. These differences encompass not only government–NGO relations, but collaborations between NGOs and corporations and international organizations as well. From the perspective of many transnational NGO practitioners, government agencies are seen as non-civil society organizations, invoking concern about organizational independence, identity, and legitimacy. Indeed, the most valuable roles that transnational NGOs play in society may be as independent advocates and critics, roles predicted on their autonomy from state and market institutions (O’Connell, 1996). Collaborations with non-civil society organizations, although often necessary to improve the scale and effectiveness of programs, undermine this identity and may erode the legitimacy of transnational NGOs as authentic representatives of an independent global civil society.

Conclusions and Implications

Practitioners have identified a “need for a radical rethinking, reinvention, and renegotiation of governance paradigms” (Brazee et al., 2012, p. 18) while scholars have called for a more complex ontology of governance in which governing authority is understood to be shared among state, market, and civil society institutions (Rosenau, 1999). A transnational governance perspective responds to these calls to provide additional nuance and context to collaborative governance scholarship. Although resource dependence theory and institutionalism tend to focus attention more narrowly on government–NGO relations, and international relations scholarship tends to focus on normative characteristics, a transnational governance perspective informed by NGO practitioners reveals a deep philosophical division in leaders’ attitudes toward intersectoral collaborations rooted in concerns over the impact of such collaborations on organizational identity and legitimacy. Although findings are limited to larger transnational NGOs registered in the United States, the relatively large sample size, cross-sectoral perspective, and mixed-method approach offer a combination of scope and detail uncommon in the field. Nevertheless, future scholarship may improve upon generalizability by relaxing selection criteria and sampling from more and larger populations.

The preponderance of public administration scholarship concerning NGO collaborations adopts the perspective of the public manager, animated by an assumption that NGOs are often more capable and efficient service providers than government agencies. Missing from these accounts, however, are the perspectives of transnational NGO practitioners, many of whom exhibit strong convictions about the appropriate role of NGOs vis-à-vis state and market institutions. Indeed, analysis reveals that organizations that do choose to pursue collaborations outside of the sphere of civil society are more likely to be motivated by financial incentives, raising crucial questions about the impact of government funding on the principled character and grassroots authenticity of transnational NGOs as the organizational embodiments of global civil society.

Research suggests several implications for public managers. Interdependent transnational NGOs are most common in the sustainable development, environmental, and humanitarian relief sectors. Public managers could build trust and promote more effective collaborations with these organizations by providing funding, recognition, access, and acceptance, and by emphasizing mutual goals and the pragmatic benefits of cooperation. Independent transnational NGOs are most common among conflict resolution and human rights organizations. Managers could achieve better results with these organizations by demonstrating respect for the identity of the NGO as an independent civil society actor. Moreover, scholars and managers alike may do well to avoid conceptualizing the transnational NGO as a delivery vehicle for official aid or as a subordinate contractor to government agencies. Although research in public administration examining domestic human service organizations often portrays NGOs as subordinates in government–NGO collaborations, from a transnational governance perspective an independent transnational NGO may perceive such a characterization negatively as threatening to its legitimacy and dismissive of its identity.

More research in public administration is needed to study transnational collaborative governance involving governmental and nongovernmental actors, including government agencies, NGOs, corporations, international organizations, and faith-based organizations, and the reasons why organizations choose to collaborate or not to collaborate within and across sectors. This research could benefit from a greater awareness of, and sensitivity to, the divergent collaborative propensities among transnational NGOs and the philosophical as well as practical issues that government–NGO collaborations raise within the transnational NGO community. By calling attention to issues of organizational authenticity, identity, and legitimacy, often underemphasized in collaborative governance scholarship, a transnational governance perspective appropriately broadens the views of scholars and public managers seeking to address social and political challenges in an ever more collaborative and globalized governance environment.