Introduction

Nongovernmental organizations (NGOs) have become much more central to the delivery of foreign aid over the last few decades, reflecting global trends in government that prioritize contracting out for services and other forms of public–private partnerships (Ebrahim 2005a; Edwards 1999; Kettl 1997; Salamon 2012; Smith and Lipsky 1993). Just as the New Public Management (NPM) movement promoted the market and private sector initiatives as alternatives to public agencies in many Northern countries, foreign aid began to shift away from direct funding of governments in favor of NGOs and decentralized development programs (Brinkerhoff 2008; Brinkerhoff and Brinkerhoff 2002; Edwards and Hulme 1996; World Bank 2000). Concerns about corrupt or weak institutions in recipient countries added impetus for decentralization and devolution in foreign aid, as did the emergence of rights-based approaches to development and donor optimism about NGOs as agents of democratization (Nelson and Dorsey 2003; Schmitz and Mitchell 2016).

Besides appealing to donor countries because of their potential to improve the efficiency and effectiveness of foreign aid, NGOs are attractive to donors because of their potential to foster a democratic civil society and spur civic engagement (Brass 2012; Brown 2008; Gulrajani 2011; Schofer and Longhofer 2011). Even though the twin concerns of efficient service provision and civil society promotion frequently are framed as reinforcing policy goals in Northern donor countries, the relationship between foreign aid and NGOs is not clear (Fafchamps and Owens 2009; Watkins et al. 2012). In practice, a tension may exist between funding organizations perceived as having management capacity and funding grassroots organizations that play a critical role in efforts to build community and civil society (Brinkerhoff 2008; Krause 2014).

The need for reporting to donors and other forms of upward accountability, for instance, often privileges NGOs that have highly trained staff and utilize diverse evaluation tools (AbouAssi 2012; Ebrahim 2005a; Kilby 2006). NGOs without management structures and processes that facilitate interactions with donors consequently could struggle to procure aid, even if they have significant local connections and community ties. Moreover, instead of fostering local civic activity, accountability demands can pull NGOs from the very grassroots work that makes them attractive to some donors in the first place (Hilhorst 2003; Chahim and Prakash 2014; Gugerty 2008, 2010; Gugerty and Kremer 2008; Girei 2016). As a recent study on the role of NGOs in foreign aid explains, “NGOs must prioritize their functional accountability to donors (in terms of targets and outputs) over their broader goals of empowerment for poor or marginalized groups” (Banks et al. 2014, p. 713).

Using an original dataset constructed from 135 interviews with leaders of NGOs operating in Cambodia, we contribute to research at the intersection of NGO management and foreign aid by exploring the relative importance of managerialism (professionalization and rationalization) and local embeddedness for direct bilateral funding (Maier and Meyer 2011; Roberts et al. 2005). We find that both of our measures of managerialism (professionalization and rationalization) are associated with receiving foreign aid. Local embeddedness, however, decreases the odds of having foreign aid, and interactions professionalization and rationalization do not alter this basic finding. Taken together, our results suggest that managerialism provides legitimacy for NGOs as a signal of capacity to donor agencies, yet the results also raise significant questions about the relationship between foreign aid and the strength of local civil society.

Bilateral Foreign Aid: Effectiveness, Motivations, and Practices

The literature on foreign aid largely concentrates on three issues: the effectiveness of aid (on outcomes such as growth, democracy, good governance); the motivations for foreign aid (altruism versus self-interest); and aid fragmentation (variability in donor practices). Research on aid effectiveness has fueled a growing awareness of recipient country characteristics and governance as important conditioning factors for effective aid. Several cross-national, longitudinal studies suggest that foreign aid contributes to economic growth most in countries with strong governance practices and reliable institutions (Bourguignon and Sundberg 2007; Burnside and Dollar 2000). Since countries with the weakest institutions tend to be the poorest, an implication of this work is that the effectiveness of aid tends to decrease as human suffering increases. Some evidence indicates that donors are responding to these findings, prioritizing effectiveness by directing a lower percentage of aid to the most impoverished nations (Berthelemy 2006; Berthelemy and Tichit 2005; Dollar and Levin 2006).

Diversity in the institutional characteristics of recipient countries provides an important demand-side explanation for the performance of aid. Attention to the supply-side of aid, or the motivations of donors, is at least as important to consider since the overall performance of aid is regarded as positive but quite weak. A long literature examining aid motives among bilateral donors suggests that strategic and political motivations trump altruistic, humanitarian motivations. Donor countries often favor countries with which they have colonial ties (Alesina and Dollar 2000), commercial interests (Berthelemy 2006; Younas 2008), or geo-political interests (Alesina and Dollar 2000). Most studies nevertheless find that bilateral donors exhibit mixed motives (Hoeffler and Outram 2011; Nunnenkamp and Ohler 2011). While self-interest may predominate, recipient need and recipient merit also explain aid flows, and considerable variability exists even within these general patterns (Berthelemy 2006).

Besides differing along a general spectrum of motivations for providing foreign aid, instructive examples reveal a range of reporting requirements and strategies for disbursing aid—often characterized collectively as aid fragmentation. The current USAID Forward initiative seeks to increase aid directed towards domestic organizations to 30 percent by 2015, but in 2010 less than 10 percent of funding went to recipient country institutions (USAID 2013). In Norway, by contrast, approximately 30 percent of bilateral aid is channeled through NGOs (Kindornay et al. 2012). With respect to reporting, Ebrahim (2003) found that the Norwegian development agency required only brief annual reports from recipient countries, while some agencies demanded evaluations for every funded program. Birdsall and Kharas (2014) similarly document a wide range of donor performance with respect to reporting requirements and associated burdens placed on recipients, the large majority of which are not standardized or coordinated across donors.

The weak results on aid effectiveness, coupled with the array of donor motivations and practices, have contributed to an aid reform movement to reduce the burden of aid fragmentation. A series of aid declarations, beginning with the Paris Declaration in 2005, the Accra Declaration of 2008 and culminating most recently in the Busan Partnership for Effective Development Cooperation in 2011, called for better aid management on the part of donors and increased participation in aid processes for aid recipients. The Paris and Accra Declarations in particular asked donor countries to commit to five principles: country ownership, aid alignment, harmonization, managing for results, and mutual accountability (OECD 2008). In this current environment, understanding both the breadth of bilateral funding for NGOs and the explanations for that funding has considerable relevance for development research.

Bilateral Foreign Aid for NGOs: Managerialism and Local Embeddedness

Despite the diverse literature on the motivations and practices of donors and the available evidence on aid effectiveness, which NGOs receive bilateral foreign aid has received little research attention (Fafchamps and Owens 2009). We suggest that donors may face a tradeoff between supporting NGOs perceived as having conventional management capacity and investing in NGOs that have valued “relational” qualities relevant to the institutionalization of a robust civil society (Brinkerhoff 2008; Schmitz and Mitchell 2016). Blending the foreign aid literature with research on NGOs and development management, we elaborate two lines of argument to explore the issue: managerialism and local embeddedness. Managerialism refers to “the bundles of knowledges and practices associated with formalized organizational management” (Roberts et al. 2005, p. 1846). Responding to critiques about the lack of specification for managerialism, in the following section, we separate the concept into organizational rationalization and professionalization (Hvenmark 2016). We then discuss foreign aid for NGOs in relation to local embeddedness and civil society.

Managerialism (As Rationalization and Professionalization)

As the global NGO sector continues to expand and evolve, these organizations face increasingly diverse coercive, mimetic, and normative pressures in the external environment (DiMaggio and Powell 1983; Meyer and Rowan 1977; Meyer et al. 2013). One such pressure is rationalization, which orients organizations towards explicit purposes and goals and formalizes organizational structures, processes, and procedures (Bromley and Meyer 2015; Hwang and Powell 2009). Since many NGOs have missions that are difficult to measure or assess, institutional sociology indicates that they will embrace a variety of tools and approaches to help them appear legitimate and competent as service providers (Drori et al. 2006; Meyer et al. 2013). Although the effectiveness of any individual tool is unclear, and the adoption of a greater number of tools does not necessarily equate with greater capacity, NGOs increasingly embrace rationalized practices such as strategic planning, hiring consultants, and restructuring (Hwang and Powell 2009; Krause 2014; Suárez 2011).

Another important component or dimension of rationalization is monitoring and evaluation. Many nonprofits conduct audits and regularly update their budgets in order to keep track of their financial performance. Similarly, a growing number of NGOs engage in program evaluation, whether formative or summative, in order to get a sense of the impact their work is having (Thomson 2011; Carman 2011; Barman and MacIndoe 2012). Despite these trends, many studies demonstrate that effective monitoring and evaluation still presents significant challenges for the public sector and the nonprofit sector; the difficulties of linking outcomes to the efforts of organizational programs are pronounced in development contexts (Carman 2011; Ohemeng 2011). Under these uncertain or ambiguous conditions, internal efforts to monitor and evaluate could confer external legitimacy. Even though the collection of performance information alone does not offer conclusive evidence of capacity, we suggest that donor countries will treat the effort as a positive signal.

H1a

NGO rationalization will have a positive association with bilateral foreign aid.

The growing complexity of organizational structures constitutes one aspect of management changes in NGOs, and a related but distinctive dimension is the professionalization of NGO work itself.Footnote 1 The sociological literature on professionalization initially focused on the elaboration of work boundaries in sovereign fields like medicine and law, stressing clear barriers to entry and unique claims to specialized knowledge (Hwang and Powell 2009; Bromley and Orchard 2016). The number of professional fields has expanded over time, including management and development, creating far more ambiguity in who can claim professional status. Aligning with this evolution, organizational professionalization now encompasses two related trends: the growing reliance on full-time, paid staff; and the shift towards a labor force with specialized academic credentials (Hwang and Powell 2009; Suárez and Esparza 2016).

Many studies document and sometimes are critical of, the ongoing shift away from a volunteer, “amateur” workforce in NGOs to a staff-centered workforce (Karl 1998; Smith and Lipsky 1993). Professionalization of this nature occurs when staff positions replace volunteer positions in mature fields, and it also occurs as activities that once took place in family contexts—like elder care and child care—come under organizational control. The emergence of NGO management programs in educational institutions is just one example of the second aspect of professionalization, the growth of academic training among NGO staff (Lewis 2015). Whereas employees once demonstrated credibility and utility through their commitment to a social mission, specialized degrees are becoming a common complement (Suárez 2010). Although volunteers may be just as effective as staff, and un-credentialed employees could be just as competent as employees with formal degrees, we suggest that professionalization confers legitimacy. As a result, donor agencies will view professionalization as a signal of capacity.

H1b

NGO professionalization will have a positive association with bilateral foreign aid.

Local Embeddedness

Embeddedness directs attention to the enabling and constraining effects of relational ties, describing purposive action as influenced but not determined by social context (Granovetter 1985). Extensive research finds that organizational ties matter for economic performance, for learning or knowledge acquisition, and for many other outcomes (Podolny and Page 1998). The concept of embeddedness is useful for characterizing diverse linkages between NGOs and the development environment, some of which could be salient for explaining which NGOs receive foreign aid (Swiss and Longhofer 2016). For instance, NGOs sometimes become members of intermediary organizations because they seek to develop new ties and expand their organizational networks (Brown and Kalegaonkar 2002). NGOs likely learn new practices and improve their services through memberships in intermediary organizations, yet the memberships themselves also enmesh NGOs in a network of committed development organizations.

Besides offering training programs to develop capacity, many of these intermediaries operate as hubs for NGO industries (i.e., health education, arts) or the nonprofit sector as a whole. Donor agencies may be attuned to NGOs with memberships in intermediary organizations because participation suggests dedication to a common enterprise, a commitment to communities of practice, and the broader social sector (Appe 2016). The local embeddedness of NGOs that belong to multiple intermediary organizations contrasts with domestic NGOs that do not belong to any affinity groups or infrastructural organizations. Local embeddedness also contrasts with “briefcase NGOs,” international development organizations that produce projects but make little effort to cultivate relationships (Cooley and Ron 2002; Krause 2014). If donors prioritize the institutionalization of civil society, NGO memberships in intermediary organization may be associated with foreign aid:

H2a

NGOs with more ties to intermediary organizations will have a positive association with bilateral foreign aid.

Local embeddedness also could matter to donor agencies that seek to promote civic engagement and strengthen democracy. Attempts at “funding virtue” began with the wave of democratization that started with the fall of the Berlin Wall and gathered steam in many parts of Africa during the 1990s and 2000s (Ottaway and Carothers 2000). A critical literature suggests this wave of funding has a tendency to undermine the very ‘grassroots’ nature of civil society organizations (Brinkerhoff and Brinkerhoff 2002; Brown 2008; Chahim and Prakash 2014; Ebrahim 2005a, b; Edwards and Hulme 1996; Gugerty 2008; Henderson 2002). Nonetheless, NGOs continue to receive large amounts of assistance, in part driven by donor belief that domestic or “indigenous” organizations can better address local needs; the focus on participatory and “community-driven development” by organizations such as the World Bank demonstrates the ongoing desire to reach community organizations (Mansuri and Rao 2013).

The emergence of rights-based approaches to development also has contributed to changes in donor behavior. NGO activities such as public education, advocacy, and coalition building are quite prominent in rights-based approaches to development, and strengthening the voice of communities is as salient for NGOs as direct service provision (Nelson and Dorsey 2003; Schmitz and Mitchell 2016). As a purposive effort to foster citizen rights, civic engagement and other activities associated with democratic participation, donor agencies may prioritize NGOs that are established in the host country or prefer NGOs that employ local citizens (Bebbington and Riddell 1995; Brinkerhoff 2008). Thus, we expect donors to perceive embeddedness in the local context as a valuable organizational characteristic, a form of relational or cultural capacity that provides legitimacy for NGOs. Consequently, we argue that

H2b

NGOs with more staff ties and historical roots in Cambodian society will have a positive association with bilateral foreign aid.

The argument about managerialism draws attention to NGO processes and staff backgrounds that we expect donors to consider in their funding decisions. We suggest that managerialism confers legitimacy as a signal of functional service capacity—donors will perceive NGOs that adopt management tools and NGOs that utilize credentialed, paid staff as appropriate for delivering programs and implementing programs. By contrast, our argument about local embeddedness draws attention to NGO relationships and social bonds that we expect donors to consider in their funding decisions. We suggest that local embeddedness confers legitimacy as a signal of relational capacity—donors will perceive NGOs that join intermediary organizations and NGOs that have roots in the local context as appropriate for delivering programs and services.

Although the general expectation is for managerialism and local embeddedness to have an independent positive association with bilateral foreign aid, local embeddedness could be especially valuable when considered in conjunction with professionalization and rationalization. Rather than investing in local NGOs with a wide range of managerial practices, it may be the case that donor countries prioritize local NGOs with greater levels of rationalization and professionalization. Instead of capacity building, donors would be “picking winners”—providing aid to a segment of local NGOs that hire staff with extensive credentials and adopt extensive accountability practices. These NGOs might have little in common besides their approach to management and their local ties, but if donors want to demonstrate support for NGOs in the host country but are risk averse, managerialism and local embeddedness might not be independent. To account for this potential interaction involving our two core lines of inquiry, we propose:

H3a

Local embeddedness will have a positive association with bilateral foreign aid in organizations that are rationalized.

H3b

Local embeddedness will have a positive association with bilateral foreign aid in organizations that are more professionalized.

In sum, if donor countries mainly stress relationship building and civil society in a country, the local embeddedness of organizations could be far more relevant than organizational management practices. By contrast, if the promotion of civil society is less relevant than direct service provision, donors may privilege NGO management practices and give little consideration to community ties, in effect treating NGOs as development contractors (Cooley and Ron 2002; Fafchamps and Owens 2009). If no tension or tradeoff in exists at all, a third possibility is that management practices and local ties will matter equally for funding. A plausible final pattern would be for donors to give special consideration to local organizations with management capacity.

Country Context, Data, and Methods

Cambodia is an impoverished country with a long history of governance problems and social conflict. Although Cambodia became a sovereign nation-state in 1953, in 1975, a communist group led by Pol Pot seized control of the country. The Khmer Rouge, as the insurgents were known, perpetrated genocide and promoted agrarian policies that led to famine. Some studies indicate that as many as 2 million people died during their short reign (Gottesman 2004; Kiernan 2008). In 1979, Vietnam invaded Cambodia to overthrow the Khmer Rouge, and the United Nations later intervened to restore stability. The Paris Peace Accords were signed in 1991, and in 1993, the United Nations oversaw democratic elections in the country. The development community has been very involved in Cambodia since that time, particularly through the provision of foreign aid. Official Development Assistance (ODA) comprised 9 % of Gross Domestic Product in 2009, and the proportion of public expenditures that come from aid has remained steady at nearly 90 % since 2005 (CDC 2010a, b; Sato et al. 2011).

Cambodia is one of the most aid-dependent countries in Asia, and besides its reliance on aid, Cambodia confronts many of the challenges of aid fragmentation (Ear 2007, 2013). The country receives aid from a wide variety of countries, all of which have their own priorities, projects, and goals. According to a recent study, aid fragmentation in Cambodia is much higher than the average for all aid recipient countries, comparable to levels in the most dependent countries like Mozambique and Ethiopia (Sato et al. 2011). Table 4 in “Appendix 1” presents an overview of bilateral foreign aid to Cambodia from 1998 to 2007, demonstrating the breadth of support. Japan is the largest donor, providing almost 35 % of bilateral aid, but the other 65 % is distributed among more than fifteen bilateral donors. While ODA provided by bilateral donors remains the predominant form of foreign assistance, NGOs are central to the delivery of this aid in Cambodia.

A Cambodian government analysis estimates that NGOs ‘‘provide or manage approximately 20 percent of all aid to Cambodia’’ (CDC 2010b, p. 14). The NGO sector in Cambodia has grown dramatically over the last two decades, especially after bilateral and multilateral organizations re-entered the country in 1993 (CCC 2010, 2012; Khieng 2014). In 2002, approximately 607 NGOs were operating in Cambodia, of which 407 were local organizations, and these numbers have continued to expand (CCC 2010, 2012; Ear 2012). The largest programming areas for NGOs are public health, rural development, social and community development, and education (CCC 2010, 2012). By 2006, NGOs employed just under 25,000 people in Cambodia, approximately 1200 of whom were international staff (CCC 2010, 2012).

Data

Our sampling framework was designed to provide a nationally representative sample of local and international NGOs in a country that has received relatively little research attention. We first defined an NGO population based on the 1240 organizations that were listed on the Cambodian government’s Council for the Development of Cambodia (CDC) online database in 2010 (www.cdc.khmer.biz). We then chose 230 organizations at random from this database, divided evenly between local and international NGOs. Of this original group we confirmed that 16 had closed and another 34 had no current contact information. This left 180 organizations, from which we were then able to interview 135 organizations, or 75 percent. To address the issue of nonresponse (and closings), we adjusted the sample using survey weights obtained from a multivariate response function. This weight also was adjusted to correct for the original sampling imbalance between local and international NGOs. While we assume that the government database undercounts the actual population of NGOs, likely missing some of the smaller, informal or “grassroots” associations, the weight is based on the available data.

The data collection process began with an initial proposal and survey instrument that was shared with a small group of Khmers (Cambodians) and foreign consultants who work in the NGO sector in Cambodia. Together with a local consulting firm we finalized the survey instrument, translated it into the Khmer language, and then piloted it in four NGOs in Phnom Penh. For the final data collection, the interviews took place in NGO offices or, in a few cases, in local cafes. The interviewee was allowed to choose the language of the interview (English or Khmer). To help with the consistency in the interview across languages, a small group of interviewers (4) carried out approximately 15 interviews together at the beginning of the data collection period together with at least one of the two main researchers. The remaining interviews were completed between October 2010 and June 2011, and on average, the interview took about 75 minutes to complete (Suárez and Marshall 2014).

The survey itself included questions on seven topics: NGO characteristics (i.e., staff size, age, main areas/fields of work, governing board), finances and strategic planning, network memberships, government relations, monitoring and evaluation, training and organizational learning, and leadership. We chose to focus on these areas for two reasons. First, we had substantive interests in topics such as performance measurement and public–nonprofit relationships. Second, we wanted to be sure to collect as many measures of organizational characteristics as possible in order to create the best models for the outcomes that interested us. As we describe in more detail below, since foreign aid is an important source of revenue for development NGOs, we collected sufficient data to explore the research question in this paper.

Dependent Variable

The dependent variable is a dichotomous indicator for whether or not an NGO has direct funding from a donor in the form of bilateral donor assistance. We asked organizational leaders to discuss their sources of revenue, and with bilateral donor funding, we followed up by asking about the specific agencies that provided funding. Based on our sample, a total of 36 NGOs in the sample had funding from bilateral donors, or 22 percent. Table 1 demonstrates that a few NGOs had resources from multiple governments, with 43 government grants and contracts mentioned in the interviews. Australia and the United States provided the greatest number of contracts, 9 and 8 respectively, but 16 different countries are represented. The table also reveals that the majority of funds go to international organizations, not local NGOs. Interestingly, not a single NGO in our sample, local or international, had funding from the Cambodian government.

Table 1 Governments providing foreign aid to NGOs operating in Cambodia
Full size table

Independent Variables

Table 2 summarizes the independent variables used in the analysis, divided into the various categories introduced in our conceptual discussion—professionalization, rationalization, and local embeddedness. We expect professionalization to be positively associated with government funding, and our indicator is based on three measures that cover different aspects of the concept within these organizations: leader education, staff education, and staff training. Based on our interviews, 81 percent of NGOs send staff members to professional training, 83 percent of NGO leaders have at least an undergraduate degree, and 57 percent of NGOs have a majority of staff with an undergraduate degree. To create a single measure, we incorporated factor analysis based on tetrachoric correlation coefficients (due to the dichotomous nature of the indicators), and the three variables load onto one factor with an Eigenvalue of 1.17. Table 5 in “Appendix 2” section demonstrates the association between each component variable and relevant controls on bilateral aid.

Table 2 Variables, means, and standard deviations
Full size table

Our second group of measures focuses on rationalization as measured by the collection of performance information, or data used for monitoring and evaluation in NGOs. More than 90 percent of organizations in the sample report having an annual budget, and most update their budget on a quarterly basis. Moreover, about 60 percent of the sample reports having had an external financial audit. We also found that 37 percent of the organizations reported using a census or survey database, and many incorporated diverse needs assessment strategies. We asked about program evaluation as well, finding that 13 percent employed formal data collection processes by using questionnaires, interview protocols and impact indicators. A relatively large group of NGOs (32 percent) also incorporated a formalized evaluation model, often outsourcing the work to external evaluators that come from a home office or are hired by the NGO. We included all of the variables in Table 6 in “Appendix 3” section in a factor analysis using polychoric correlation coefficients. All of the indicators load onto one factor that has an Eigenvalue of 2.58 (for more information see Marshall and Suárez 2014).

Our third group of indicators addresses the local embeddedness of NGOs. The first measure is an indicator of network participation, or the total number of NGO networks to which the NGO belongs (average is 2.2 networks). The second is a factor based on tetrachoric correlation coefficients for local embeddedness. The indicator is based on three dichotomous measures (Cambodian executive director, all staff are Cambodian, the NGO has no international office), all of which load onto one factor with an Eigenvalue of 2.62. Table 5 in “Appendix 2” section demonstrates the association between each component variable and relevant controls on bilateral aid.Footnote 2

It is important to clarify that we used factor analysis to create measures for professionalization, rationalization, and local embeddedness in order to maximize the data we had available to us. Including variables that are correlated with one another in the same model can produce inflated coefficient standard errors, or multicollinearity. Factor analysis is a powerful data reduction technique that can reduce the likelihood of multicollinearity by combining variables that measure the same underlying concept or trait. Because our sample is fairly small and we are therefore limited in the number of independent variables that we can include, factor analysis also enables us to utilize the available degrees of freedom more efficiently.

In order to test these various arguments and reduce the possibility of omitted variable bias, we include a number of general controls. This begins with three indicators for primary field of expertise, designed to capture potential variation related to programmatic focus (health, education, and human rights). We also include NGO age, an important control because some management practices could develop with time and experience. The average age in the sample is almost 13 years, with a standard deviation of nearly 10 years (the indicator is logged to reduce skewness). We include an indicator for NGO size based on the number of staff employed by the organization as well. The average staff size is 52, but the standard deviation is large because of a number of very big NGOs in the sample. This indicator is logged to reduce skewness.

The final control deals with resource diversity. Nonprofits often have multiple sources of revenue for implementing their missions, ranging from market-based income from program services to grants from foundations and subcontracts from other nonprofit organizations. Some sources of income can provide legitimacy for a nonprofit, demonstrating to other donors that they are capable at procuring funds (Hwang and Powell 2009; Khieng 2014). Governments may consider multiple forms of support for NGOs to be a sign of competence, increasing the likelihood of bilateral foreign aid (Brown 2008; Lewis 2006).Footnote 3

The nature of our data means we are unable to utilize maximum likelihood measures to assess goodness of fit (or for other purposes) in our models. A core assumption of maximum likelihood is that cases are independent of each other. Because we incorporate weights to address oversampling of international NGOs and nonparticipation bias, however, maximum likelihood and the statistics based on it become inappropriate. As an alternative, we rely on F-tests in the models we present below (for more information see Stata Corporation (2011)). Also, since our data are cross sectional, we do not draw strong causal inferences about the relationships among our variables, seeking instead to explore the factors associated with bilateral aid.

Results

Our empirical models begin with bivariate regressions presented in the first main column of Table 3. The regressions in that column serve to establish a baseline for understanding how the controls and the other substantive variables are related to bilateral donor funding. As the results show, all of the independent variables except for network memberships produce statistically significant effects. In addition, the age and size of an NGO are associated with bilateral funding. These results are quite straightforward, particularly if the size and the age of an NGO tend to serve as a proxy for other more substantively meaningful indicators of capacity. The last set of controls, the primary fields of work for NGOs, does not have strong relationships with bilateral funding. Here again the results are reasonable since foreign aid flows to many sectors of Cambodian society. From there, we turn to multivariate regression models, presenting the results for each conceptual frame and the controls.

Table 3 Summary of predictors of bilateral government funding
Full size table

Model 1 considers the influence of professionalization on bilateral donor funding. The results demonstrate that professionalization has a statistically significant, positive effect on the likelihood of having foreign aid, as hypothesized. Organizations that send their staff to trainings and are led and staffed with more university graduates are more likely to have government funding than organizations that score lower on these measures of professionalization. Model 2, which also focuses on the management of NGOs, shows the relationship between the production of performance information and the presence of bilateral aid. NGOs that engage in greater collection of performance information are significantly more likely to have foreign aid than NGOs that adopt fewer tools for assessing performance. Taken together, models 1 and 2 provide strong support for the association between NGO management practices and bilateral donor funding.

Because the indicator for network memberships was not statistically significant in the bivariate analysis, we do not present a separate model for that independent variable. Model 3 clarifies the association between our two indicators for local embeddedness and bilateral funding. Interestingly, the effect for local ties is statistically significant but negative, demonstrating little support for the idea that donor countries privilege domestic nonprofits as a way to build capacity in civil society. Whether international NGOs and NGOs that are staffed by foreigners are actually more capable than their local counterparts is unclear, but such organizations are much more likely to report having bilateral foreign aid. As before, the coefficient for the number of network memberships is not statistically significant, suggesting that participation in local intermediary organizations is not especially salient for explaining foreign aid to NGOs. Even though networks are important in a variety of different development contexts, they may play a much larger role in providing training than in conferring legitimacy (Brown 2008; Brown and Kalegaonkar 2002).

Model 4 presents the results for the analysis when all three lines of argument are considered together, along with the controls. The results for professionalization and rationalization are robust, revealing that both indicators have a positive association with government funding. Since both indicators retain their explanatory power, these results suggest that forms of training (formal and informal) and the collection of information on performance are related but empirically distinctive aspects of NGO managerialism. The coefficient for local embeddedness remains negative and statistically significant, indicating that donor agencies do not privilege or prioritize NGOs with more linkages to Cambodia over organizations staffed by foreigners or with bases in other countries. The results in fact suggest quite the opposite. Even when taking into account a variety of controls and several indicators for management practices, the most locally embedded NGOs are far less likely to receive foreign aid than their international counterparts. Models 5 and 6 consider potential interactions between professionalization and local embeddedness and between rationalization and local embeddedness as a final test. The models offer little support for an interaction effect. Even when combined with high levels of organizational rationalization, local embeddedness seems to offer no advantage for organizations in terms of receiving foreign aid.

Discussion

Since the first forum on aid effectiveness in Rome in 2002, foreign aid agencies have faced numerous potentially conflicting mandates. Calls for more efficient, effective, and transparent aid often have been met with demands on partners for increased reporting and evaluation. At the same time, donor agencies have faced increasing pressure to create genuine partnerships with recipient countries and to deliver aid in ways that support local organizations without unduly burdening recipients. In this paper, we have used data from one heavily aid-dependent country, Cambodia, to examine whether one of these effects tends to predominate. We find evidence to suggest that donor countries privilege managerial NGOs and NGOs with weaker ties to the local context in Cambodia; rationalized, professionalized, and NGOs with fewer local ties are more likely to receive bilateral donor funding.

Our findings are subject to a number of empirical caveats. First, we measure only direct funding of NGOs by bilateral agencies. Some NGOs, particularly local organizations, receive bilateral funding that has ‘passed-through’ another organization, typically an international NGOs or private consulting firm. Although we know which organizations in our sample have funding from NGOs, we could not discern if the original source was foreign aid. We also are hampered by the cross-sectional nature of our data. We cannot assess the extent to which donor funding responds to professionalization and rationalization or causes it. Finally, while our data shed light on a very under-studied context, our results should be generalized with caution. Some have suggested that aid should be channeled around governments in fragile states, towards NGOs and private agencies that might be more reliable (Bourguignon and Sundberg 2007). Since Cambodia is a repressive state with high levels of corruption, additional research is needed to determine if our results hold in countries with more robust, democratic institutions.

The study nevertheless makes several contributions to the extant literature on NGOs and donor funding and sets and agenda for future work. We contribute to NGO studies by specifying two distinctive aspects of managerialism (Hvenmark 2016; Meyer et al. 2013). Professionalized NGOs, those with more trained and educated staff, and rationalized NGOs, those with more extensive monitoring and evaluation regimes, are more likely to have foreign aid. These NGOs actually may not be better at implementing projects than other organizations, but we suggest that the adoption of these management practices provides legitimacy (Hwang and Powell 2009; Karl 1998). Interestingly, while the effects for managerialism are not entirely surprising, decisions about foreign aid often have political motivations, yet we find that management practices still matter a great deal (Riddell 2007).

We do not discount the importance of politics and other motivations that differ across contexts, but our results suggest either that professionalization and rationalization matter for funding despite these other issues or that most funding is for technical projects. Some research on foreign aid in Cambodia lends support to the latter interpretation, concluding that “Cambodia’s experience since 1993 suggests that most projects have been donor-driven in their identification, design and implementation, to the detriment of capacity development” (Godfrey et al. 2002, p. 369). This approach to funding also helps to explain our results for local embeddedness. If donors prioritized capacity building and the construction of civil society, we might expect to see support for local organizations to “outweigh” professionalization and rationalization, but we do not find evidence that donors fund even the most managerial local NGOs.

Our findings add to ongoing discussions about the role of NGOs versus governments in the development process as well. None of the NGOs in our study receive funding from the Cambodian government, raising questions about the role of NGOs in public service provision in the country. If international donors stop supporting these NGOs, there is little indication that the Cambodian government will fill in the gaps in program services or continue with ongoing projects (Brinkley 2011). This issue has been a concern for development scholars for several decades, particularly in relation to capacity development (Edwards and Hulme 1996). Given that we find little evidence that foreign aid serves the direct purpose of capacity building, or even that funding prioritizes local organizations in a direct manner, the long-term implications for development remain unclear.

Finally, we add to discussions about how development funding shapes the priorities and goals of NGOs, particularly in local community contexts (Bebbington 2005; Brown 2008; Edwards and Hulme 1996). Because rationalization can produce an instrumental focus that devalues social relationships in an effort to achieve technical efficiencies, performance could come at the expense of local ties and a community focus (Girei 2016; O’Dwyer and Unerman 2008; Townley et al. 2002). If professionalization and rationalization provide legitimacy for NGOs, as we argue, local civil society may not benefit or grow in a sustainable manner through foreign aid. Future research can build from our study by clarifying the longitudinal connections between professionalization, rationalization, and local embeddedness in development contexts.

As NGOs are likely to remain important recipients of foreign aid, more research is needed to understand which types of organizations receive support and why. NGOs often are described as pivotal to the production of social capital and democracy in international discourse, and these organizations also deliver programs and services to communities in need (Longhofer and Schofer 2010; Schofer and Longhofer 2011). We shed light on one aspect of the aid chain with our investigation of direct aid disbursement, demonstrating clear prioritization of managerial and international NGOs. These results set an agenda for new studies on subcontracts to local organizations and on how those interactions matter for the growth of local civil society.