Friend or Patron? Social Relations Across the National NGO–Donor Divide in Ghana

The article examines the institutions governing relations between grant using national NGOs and grant giving international donors in three regions of Ghana (Upper West, Northern and Greater Accra Region). Formal procedural rules and professional norms can be viewed as necessary to minimise opportunities for informal patronage, rent-seeking and corruption made possible by the unequal access to resources. However, semi-structured interviews, life histories and observation highlight the positive role informal networks, connections, personal contacts and friendship play in enhancing collaboration between donors and national NGOs. Friendships originating in kinship and ethnicity, school links and past collaboration offer opportunities for influencing and resource mobilisation, but can also weaken NGO sustainability. Informal contacts and face-to-face interactions also build trust and strengthen lines of accountability, with non-adherence to shared norms resulting in sanctions and reputation loss. These findings affirm the positive role of informal relations, and highlight how they can complement formal rules and professional norms governing NGO–donor relations rather than undermining them. It throws a very different light on the role of informal institutions than that fostered by a discourse of clientelism and provides a more nuanced conceptual foundation for assessing ‘formalisation’ as a normative strategy. Cet article analyse les relations entre les ONGs à niveau national réceptrices de subventions, et les donneurs internationales qui octroient des subventions, dans trois régions au Ghana (le Haut Ghana Occidental, le Nord du Ghana, et la région du Grand Accra). Les procédures formelles basés sur des règles, et les normatives professionnelles, peuvent être perçues comme nécessaires afin de minimiser les opportunités de clientélisme informel, d’avantages personnels et de corruption, dû à l’accès inégal aux ressources. Cependant, des entretiens semi-structures, le récit d’histoires et l’observation soulignent le rôle positif que les réseaux informels, les connections et contacts personnels, et les amitiés jouent dans l’amélioration de la collaboration entre donneurs et ONGs à niveau nationale. Les amitiés basés sur des liens ethniques et de parenté, et les liens fondés sur une expérience scolaire partagé ou des collaborations antérieures, offrent l’opportunité d’influencer et mobiliser des ressources, mais peuvent aussi affaiblir la durabilité de l’ONG. Les contacts informels et les interactions face à face construisent ainsi de la confiance et renforcent les chaînes de responsabilité ; la non-adhérence aux normes partagés aboutit à des sanctions et à la perte de réputation. Ces résultats confirment le rôle positif des relations informelles, et soulignent comme elles peuvent complémenter les règles formelles et les normes professionnelles qui gouvernent les relations entre ONGs et donneurs, plutôt que de les miner. On éclaire le rôle des institutions informelles d’une façon très différente au discours du clientélisme, et on fournit une fondation conceptuelle pour évaluer la « formalisation » comme une stratégie normative.


Introduction
This article examines the institutions governing relations between donors and national NGOs in Ghana. The existing literature on NGOs have mostly focused on formalisation of rules as part of the design of donors' aid chain framed normatively around patron-client relations (Wallace et al. 2006). This literature has paid much attention to formal procedural rules and professional norms on governance and accountability requirements (see Ebrahim 2003;Agyemang et al. 2017) to the neglect of the informal aspects of such relationships. The existing literature is silent on how the culture (i.e. the social relations, informal networking, values and ideas) that governs donor-NGO relations influences their daily operations and the effects of attempts to formalise such relationships (Eyben 2006;Yarrow 2011;Fechter 2012;Sundberg 2019). In particular, there are relatively little empirical research on how informal networks and personal connections enhance collaborations between donors and national NGOs from the perspective of sub-Saharan Africa. This article addresses this knowledge gap.
In doing so, it seeks to answer the following research questions: How can the culture governing donor-NGO relations in Ghana best be described and explained? Who gains and who loses from attempts to formalise or strengthen the explicit rules governing donor-NGO relations in Ghana? In answering these questions, this article draws on qualitative in-depth interviews, life histories and personal experiences of Ghanaian donor representatives and NGO staff to illustrate the culture governing their relations 1 . This article highlights that the Ghanaian development sector is characterised by power structures between NGOs and donors. This arises because of a limited number of donors who control critical resources crucial for the survival of NGOs. Importantly, given the increasing number of NGOs albeit fewer number of donors, this presents a challenge to donors about how to allocate their scarce resources among relatively potentially abundant NGOs. Understanding the culture governing donor-NGO relations provides useful insights into how donors allocate their scare resources.
We show that although donor agencies seek to formalise their relations by enforcing formal procedural rules and professional norms, Ghanaian donor representatives and NGO workers draw on their social relations, shared values and ideas and informal networks in helping them allocate resources and manage their development work. By doing so, this article shows that personalised relationships, friendships rooted in kinship and ethnic ties, past professional experiences and old-school associations are significant elements of the culture governing donor-NGO relations. It further demonstrates how personalised relationships within the Ghanaian development sector operate alongside explicit bureaucratic and professional rules that govern the allocation of donor funding for NGOs. Based on our empirical evidence, we argue that formalising donor-NGO relations does not necessarily improve the quality and cost-effectiveness of development interventions. Rather rule formalisation creates opportunities for the co-existence of personalised relationships and formal bureaucratic orders which becomes a strategy for managing and achieving the desired development goals by donors and NGOs. Our empirical findings challenge arguments about the culture of Friend or Patron? 60/2020 Kumi and Copestake unequal donor-NGO relations. In particular, we contest the belief that donors are selfinterested, corrupt institutions and rent-seeking patrons who disburse funds to maximise their own benefits while NGOs are complicit in such clientelism to the extent that weak formal rules and professional norms allow. We show the limits of such arguments by highlighting that personalised relations, contacts and friendships between donor and NGO representatives serves as a catalyst for 'getting things done'. To this end, efforts to promote development effectiveness should pay particular attention to the less formalised or personalised nature of relations because 'relationships matter' significantly for the success of development interventions. We argue that it is the personalised nature of relations or what Eyben (2010) calls 'relationalism' that sustains the existence and effectiveness of the aid industry in Ghana. Therefore, investing in mechanisms that promote the practice of 'relationalism' provides enormous benefits for the development sector. Building relationships is central to the creation of partnerships that produce meaningful development outcomes.
This article contributes to the empirical literature on donor-NGO relations by showing how personalised relationships enhance collaborations between donors and NGOs. In doing so, it highlights how informal practices within donor and NGO circles manifest itself and its importance for the sustenance of the development sector. Building informal social ties and relational resources is a critical strategy for resource mobilisation and NGOs' operations. We show that informal networks and personal connections are strategic responses that help in negotiating and manipulating strict donor requirements and conditions which create opportunities for influencing donor agendas and policies. Understanding how NGOs use their networks and personal connections to navigate the donor landscape is crucial for policy deliberation because it provides a strong foundation for appraising the potential of NGO leaders and donor representatives to organise in ways that can enhance their agency and influence over development policy and practice.
The remainder of this article proceeds in five parts. Following this introduction, Section 2 discusses formalisation as conceptual framework underpinning this study. This is followed by a review of the literature on informal networks among NGOs in Section 3. The research methodology is presented in Section 4. The research findings are presented and discussed in Section 5. The last section presents some concluding remarks.

Conceptual framework of formalization
This article draws on the concept of formalisation in examining the culture that governs how development work is managed by NGOs and donor representatives and the effects of efforts to formalise their relationship in Ghana. Formalisation provides a useful framework for understanding the values, norms and beliefs and how NGO staff use shared norms and trusts including social networks and connections in forming collaborations with donors. The concept is also used in shedding light on how attempts to impose more explicit order into donor-NGO relations produce unintended consequences for development interventions.
While formalisation is not a new phenomenon, it has received an increased attention Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 6 especially in the public administration literature (Pandey and Scott 2002;DeHart-Davis et al. 2013;Kaufmann et al. 2018;Borry et al. 2018). Within institutions, formalisation is considered as impersonal and relies largely on fixated rules and procedures that structure the behaviour of actors. In this article, formalisation is understood as the active process by which a specified agency seeks to impose more explicit order onto a system or process. It is defined as the ''degree to which rules define managerial and employee roles, authority, relations, communications, norms and sanctions, and procedures in organisational activities'' (Lakshman 2015: 170). Formalisation therefore is about the extent of explicit formulation of norms. It also denotes the intensity of written rules, procedures, instructions and communications within an organisation and focuses on their strict application with the aim of ensuring predictability and stability (Pandey and Scott 2002;Kaufmann et al. 2018). Formal organisations and rules help in structuring collective behaviours and individual actions within society (Mica et al. 2015).
Two forms of formalisation exist: job codification and rule observation. Job codification focuses on the governance of individual behaviour through the provision of detailed norms and rules about how specific tasks are to be performed. Codification takes the form of written or formalised rules that enables a common language to be spoken across organisations which demonstrates the collective nature of organisational processes (Pandey and Scott 2002). Formalised rules also convey organisational legitimacy as it increases the likelihood of compliance (DeHart-Davis, 2009;DeHart-Davis et al. 2013). On the other hand, rule observation is also about the extent of supervision towards conformance to the standards set by job codification (Hage and Aiken 1967). To this end, organisational rule following becomes an important aspect of formalisation as it shapes the effectiveness of bureaucratic structures, demonstrates the strength of public sector accountability and how transparency and equity are delivered (Borry et al. 2018). Within many institutions, written rules a play a critical function by virtue of their functional nature and can have organisational and societal benefits when pursued through rational means. At the same time, they have the potential of producing unintended effects for organisations if they are illogical or burdensome. In this regard, formalisation can be enabling and constraining depending on its content and the context within which it is applied (DeHart-Davis et al. 2013;Borry et al. 2018).
Within the aid industry, formalisation of organisational management has received attention in the literature (Roberts et al. 2005;Eyben 2010). The emphasis on formalisation is on managing for development results and forms an important part of what Eyben (2010) calls 'substantialism'. A substantialist world view focuses primarily on 'pre-formed entities' and attach secondary importance to social relations. This has therefore led to an increasing formalisation of relationships between donors and organisations including NGOs embedded in the aid chain (Eyben 2010, p. 385). For instance, donor-NGO relations are mostly enforced through contractual agreements characterised by several accountability requirements. Formalised accountability requirements create opportunity for donors to hold NGOs to account for their actions due to the principal-agent problems that confronts donors (Ebrahim 2003). As Copestake et al. (2016) demonstrate, donors as principals lack information about NGOs as agents in terms of their commitment and ability to achieve intended project Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 7 outcomes.
Faced with principal-agent problems, donors have resorted to 'development managerialism' characterised by the application of rational tools for planning and measurement within private aid channels (Roberts et al. 2005;Wallace et al. 2006). This is premised on the idea that it promotes efficiency, effectiveness and transparency. As part of managerialist thinking, formalised relations between donors and NGOs are assumed to help in the prevention of financial malfeasance, ensure compliance and produce value for money. The belief is that the absence of formalised structures is associated corruption, clientelism, nepotism and a means of control (Blundo and Olivier de Sardan 2006). In this regard, formalisation of relations has become a mechanism for increasing organisational efficiency and effectiveness. As DeHart-Davis (2009) highlight, formalisation increases the likelihood for achieving effective organisational rules.
However, formalisation of relations between donors and NGOs increases donors' power. For example, Agg (2006) argues that in order to receive a formalised partnership agreement from the Department for International Development (DFID), NGOs had to prove how they could contribute to the UK government's priorities and targets. Similarly, Elbers and Schulpen (2013) document power asymmetries in donor-NGO relations where decision-making and partnership governance structures were unilaterally set by donors. In addition, due to resource dependence, donors unilaterally define accountability requirements and funding conditions for NGOs (Ebrahim 2003). This includes the obligatory use of for example, logframe as part of funding reporting and accountability requirements. Power relations associated with formalised relations undermine transformative learning as it prejudices mutual accountability associated with trust-based relations (Eyben 2006). However, the adoption of trust-based relations by donors is limited because of their emphasis on substantial approach that focuses on cause-and-effect or rational thinking. From a rational choice perspective, developing and maintaining relationships between donors and NGOs is considered a transaction cost. Importantly, maintaining 'informal' relationships goes against the ethos of formalisation and rational choice thinking (Eyben 2010).
Notwithstanding, some authors have argued that informal relations constitute an important aspect of donor-NGO relations (Alikhan et al. 2007;Yarrow 2011;Aliyev 2015). Informal relation is understood as ''a form of interaction among partners engaging in dialogue, the rules of which are not pre-designed, and enjoying relative freedom in the interpretation of their roles' requirements'' (Misztal 2000, p. 46). The understanding of informal relations as used in this article does not only portray its 'perceived negative connotations' including neopatrimonialism but also shows the extent to which they can facilitate the processes of 'getting things done' because they are socially acceptable. In what follows, we review the literature on informal networks in the NGO sector.

Informal Networks among NGOs
NGOs operate within an institutional environment which consists of the task and general environment. The task environment focuses on elements within which NGOs have direct contact with while the general is concerned about factors that affect them indirectly. As part of the task environment, NGOs have direct relationship with stakeholders which occur at the level of formal and informal institutions. To this end, informal networks have become important resources employed by NGOs in ensuring their continual survival (Yarrow 2011;Aliyev 2015). Informal networks are built on shared norms and values including trust.
Trust is very important for NGOs in their relationship with stakeholders because it determines the extent of their resource mobilisation and legitimacy (Keating and Thrandardottir 2017). Trust is also critical in helping NGOs build shared sense of norms and values. For this reason, social trust between NGOs and stakeholders including donor agencies and local politicians is not only maintained through formal contractual agreements, but also informal ties. Trust strengthens cooperation among partners because it enhances legitimacy, credibility and reputation. It also ensures mutual accountability in networks in achieving collective goals (Lyon 2000;Romzek et al. 2012).
Trust-based personal relationships are important among development workers because it helps them in negotiating their complex lifeworlds (Hilhorst 2003). For instance, Eyben (2011, p. 246) found that trust-based personal relationships were required by development professionals in Bolivia ''to take them through the rocks and rapids of negotiating a multidonor budget support facility''. Similarly, Bachmann (2016) illustrates how personal relations and interactions between EU diplomats in Nairobi created opportunities for collective diplomacy. Recent studies (see Fechter 2012;Heuser 2012) have also shown that social relations including friendships and the primacy of the personal among development workers facilitates development work. The social and the professional are intertwined in development work. For example, Hueser (2012) demonstrates that among foreign and local NGO workers in Indonesia, social relations through friendship create a sense of belonginess, helps in creating identities and also bridging gaps in knowledge between foreign and local development workers.
Informal networks are also crucial for especially international assignees or expatriates working in NGOs as they provide informational, emotional, instrumental and financial resources in their host country (Claus et al. 2015). Similarly, Yarrow (2011) demonstrates the importance of social relations among development workers in Ghana. He maintains that informal relations help workers to create formal networks with stakeholders including politicians, media and development organisations. Everyday life of development workers involves developing informal relations including trusted friendship, family, ethnic and religious ties. These are built on trust and self-regulation which results in social interactions that are honest in nature (Hilhorst 2003;Romzek et al. 2014).
The examples mentioned illustrate that not all informal relations result in corrupt practices Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 9 because it is directly linked to the way development organisations including NGOs and donors operate. The activities of development organisations are sustained largely by informal practices that take place on a regular basis but remain largely unnoticed. This is what Hilhorst (2003) calls 'the real world of NGOs' which is crucial in determining how NGOs survive and achieve their missions in an ever-changing operating environment.
This understanding of informal relations is in sharp contrast to the argument by Chabal and Daloz (1999) that individuals engage in the work of civil society to further their own interests. Similar accounts have been given by Smith (2010) about how personal elites in the Nigerian NGO sector engage in development work to further their parochial interests and corrupt practices. For these authors, social relations stifle development and by doing so portray a negative impression framed mostly as neo-patrimonialism. This Western understanding of informal relations fails to appreciate contextual differences between Western and nonwestern societies (Ledeneva 2018). However, as some scholars have argued, informal networks including friendships are a normal way of organising in many parts of the world (Yarrow 2011;Eyben 2011;Harrison 2017;Sundberg 2019). For example, in their study of informal contacts and networks among NGOs in Post-Communist Europe, Grødeland and Aasland (2011) found in Slovenia, Bulgaria and Romania that NGO representatives perceived informal practices as part of their culture. They also highlight the importance of informal practices and relations in helping people get quick solutions to their problems compared to using institutionalised structures. In this regard, informal networks among NGO workers become a social norm because it revolves around cultural products such as traditions.
Similarly, Aliyev (2015) documents the significance of informal networks by NGOs in Georgia and Azerbaijan and argues that it helps them collaborate with stakeholders including government officials and peer NGOs as well as receiving funding from donors and philanthropic institutions.
It is worth noting that reciprocity, moral obligation and affection are important factors for building stronger informal networks and relationships. For example, guanxi (i.e. business connections and networks) in China has been used by NGO leaders in building their organisations. Ru and Ortolano (2009) maintain that environmental NGOs in China use their guanxi in overcoming administrative bureaucracies and challenges associated with registering an NGO. In the case of Ghana, Yarrow (2011) highlights how reciprocal assimilation among development workers leads to the formation of stronger informal relationships. Informal relationships are built around friendship which originates from educational and professional backgrounds (Heuser 2012). This leads to the building of personalised trust developed over a long period of time. In fact, informal networks among development workers are also developed through mechanisms including engagement in sporting, cultural events, during implementation of donor-funded projects and participation in political activism (Aliyev 2015;Eyben 2011;Yarrow 2011).

Bath Papers in International Development and Well-Being
Friend or Patron? 60/2020 Kumi and Copestake 10

Methodology
This article is informed by a qualitative research methodology in exploring the culture that governs donor-NGOs relations in Ghana. The use of qualitative research was based on the need for a deeper understanding and critical insights into the dynamics of the social relations existing in the Ghanaian NGO and donor space. This helped in exploring the perception of NGO and donor representatives on the importance of their personal connections. This article is a by-product of a broader doctoral research which examined the resource mobilisation and survival strategies of national NGOs in Ghana (see Kumi 2017). Data collection for this research took place between July 2015 and July 2016. The research involved fifty-nine (59) national NGOs operating in health, education and agriculture sectors in the Northern, Upper West and Greater Accra regions of Ghana.
The selection of the three regions was informed by the following reasons. First the Northern and Upper West regions (hereafter, The North) are the most marginalised part of Ghana which makes poverty very severe. For example, according to the Ghana Living Standards Survey (GLSS) 6, poverty incidence for Upper West and Northern a region stood at 70.7% and 50.4% respectively (GSS 2014, p. 13). The concentration of NGOs and donors is therefore highest in the North because of the high poverty incidence and the peculiarities of development challenges. Tamale, the capital of the Northern Region is considered the NGO capital of Ghana (Kumi, 2017). Second, the selection of the Greater Accra region was informed by its importance as the headquarters of many donor and government agencies. Moreover, in terms of geographical coverage, Greater Accra is located in a different agro-ecological and climatic zone (i.e. dry-south east coastal plain) and has different degree of urbanisation and proximity to donor agencies compared to the North.
Aside from the 59 national NGOs, nine bilateral and multilateral donor agencies, philanthropic institutions and international NGOs participated in the research. A criteria sampling was used in selecting NGOs and donor agencies 2 . In total forty-two NGO staff involving executive directors, heads of programmes and project officers were interviewed. For donor agencies, eleven representatives ranging from Country Directors to Programme Managers and Officers were interviewed. In addition, nine key informants including academics, development consultants and government officials at the Ministry of Finance, National Development Planning Commission and Department of Social Welfare were also interviewed.
Data collection involved the use of life histories, participant observation and semi-structured interviews which provided opportunity for exploring and understanding how the relationships between NGOs and donor representatives is shaped by their personal and life courses. In fact, exploring the life course of NGO leaders and donor representatives provided much insight into their social affiliations developed over the years through their engagement in development work. As Lewis (2008) highlights, using life histories in third sector research helps in providing historical depth and ethnographic details especially as individuals reflects on their life and provides a better sense of 'being there'. More importantly, life histories provide valuable insights through its actor-oriented perspective that seeks to reinforce Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 11 element of agency including how NGO leaders influence donor policies and practices. The discussion of structure and agency is very important especially when researching issues of social relations and informal networks (Hilhorst 2003). Similarly, life history interviews and personal narratives have been used in understanding the motivations for Ghanaian development workers (Yarrow 2011).
The semi-structured interviews were made up of open-ended questions and this gave interviewees opportunity for expressing their opinions in detail. Gillham (2005) highlights that elite interviews be loosely structured because of the sensitivity with the information they provide. Interviewing NGO leaders and donors' representatives on their informal networks also raises methodological challenges because they are often associated with clientelism and corrupt practices. For this reason, being tactful in approaching interviewees and also building rapport was crucial. Although it was explained to interviewees that this research was for academic purpose, some were skeptical about our intent of asking questions relating to their personal relationships and networks. This is because our research focused on resource mobilisation, donor funding mechanisms and relationships with NGOs and donors. Personal connections among development workers sometimes create cooperation and tension and therefore understanding these issues was critical. In addition, combining semi-structured interviews and personal life histories was a useful exercise for the interviewees to open up by reflecting on their work trajectory. Given the assurance of anonymity offered to interviewees, some were able to voice dissent and became critical of each other which sometimes they were reluctant to express through official feedback reporting systems 3 . This provided the opportunity for them to share their negative and positive experiences of informal relationships in their daily operations. Aside from semi-structured interviews and life histories, one of the authors participated in several meetings and programmes organised by NGOs and donors. During such periods, field notes were used in recording events and the feelings of participants in their natural contexts. This provided opportunity for self-reflexivity during data analysis.
The interview data were recorded and transcribed with the consent of interviewees. These were coded in NVivo 10 and analysed using thematic and discourse analysis. The inductive, iterative approach to qualitative analysis informed our analysis. Thematic analysis was used in used in identifying emergent themes from the data while discourse analysis was employed in understanding the nature and how interviewees framed their relationships. In complying with ethical requirements, approval for this research was given by the University of Bath Ethical Committee. To validate the research findings, the transcripts were sent to interviewees for their feedback which was incorporated into the final analysis. This ensured data accuracy. In the next section, we present and discuss our key findings. The above statement sheds light on the negative aspect of focusing on ethnicity in developing collaborations. Ethnicity and tribal affiliations were perceived by interviewees as divisive and resulted in allegations of favouritism and nepotism among NGO leaders and donor representatives. Nonetheless, the use of kinship ties and social identities was critical for the survival of NGOs due to their resource acquisition potentials (Aliyev 2015;Harrison 2017). The findings suggest that kinship ties provided the conduit and served as bridges in accessing resources within their operating environment. The issue of ''who you know'' was therefore a recurrent theme emphasised in relation to funding applications. However, it was explained that this resulted in perceived discrimination in the funding process. A section of interviewees argued that some donor representatives had interests in some NGOs because they were established by their close relatives and friends which made them use their connections in securing funding. This was explained by an NGO leader as follows: ''you establish relationships with them and once they have your contact, periodically, they will be sending you funding windows and tell you what they expect from the proposal'' (Interviewed on 1 st December 2015, Tamale).
However, interviewees suggested that NGOs established by 'cronies' in anticipation of upcoming funding opportunities could not survive for a long time and often collapsed as a Some interviewees acknowledged that discrimination exists in the Ghanaian NGO sector mainly because of limited funding opportunities and an increasing number of organisations. For this reason, discrimination was expected because ''at any point in time, there is discrimination everywhere'' as stated by one NGO staff (Interviewed on 14 th March 2016, Accra). Discrimination arising from personal connections was raised by NGOs that were unsuccessful in their funding applications. For this reason, they suggested that their competitors won donor grants based on their social connections and not only the quality of their proposal. One NGO leader lamented: Some donor representatives have established their own local NGOs, so they are always looking for support for their NGOs. So, they will advertise a call and when you apply, they don't mind you. But some particular NGOs are always getting the funding [….]. So, unless you have strong connections, otherwise whatever proposal you write does not matter. I must admit it has not been very easy at all (Interviewed on 7 th January 2015, Wa).
The above statement indicates that despite the explicit formalised, bureaucratic and professional rules and norms governing how donors should allocate their funds to NGOs based on organisational competencies, track record and compliance with formal application rules, interpersonal relationships between donor representatives and NGOs facilitated the selection of implementing partners. The findings therefore suggest that relationalism operated alongside the bureaucratic or formalised structures put in place by donor agencies. However, when asked about personal involvement in funding applications, donor representatives dismissed such claims by arguing that ''all NGOs are judged on the same principles rather than whom you know'' (Interviewed on 15 th December 2015, Accra). This was unsurprising as donors always want to portray themselves as neutral and impartial in their relationship with NGOs and intended beneficiaries. Given that the legitimacy of development interventions hangs on the perception of their neutrality, some donor staff argued that they do not want to be accused of using their personal connections in favouring some NGOs because of its potential to affect their credibility and legitimacy. The use of personal connections often  (Smith 2010;Harrison 2017).
This also reflects the argument by Unsworth (2009) about potential barriers including a technical managerial approach to development that hinder donors from recognising that politics is central to their development interventions. Donors often consider themselves as politically neutral partners and have an apolitical imaginary of their development interventions. For instance, an interviewee argued that ''our development partners, they are also politicians and they wouldn't want to come and mingle in our issues'' (Interviewed on 23 rd March 2016, Accra). This was reflected in the explanations given by donor representatives who argued that funding proposals were judged using standard criteria where funds were awarded on merits. In some instances, independent consultants were hired to evaluate these proposals as articulated by the Head of Programmes of a multilateral donor agency: We have standard criteria which could be referred to as a marking scheme and that's what we use. To ensure integrity and robustness of the grant making process, grants are not assessed by our staff alone.
We have a team of independent assessors alongside the staff. For this call, we had six independent assessors and they have a copy of the assessment criteria. They are entering their scores of the proposals and later they will send them back to us (Interviewed on 1 st February 2016, Accra).

Friendship through old-school links
Another avenue through which friendships between donor representatives and NGOs leaders were was through educational affiliations including school-based networks described by an interviewee: ''the fact is that you have school mates who are working in the development sector, you develop relationship with them'' (Interviewed on 25 th April 2016, Wa). Friendships developed from old school associations were mainly through meetings and reunions. Such avenues provided opportunities for building stronger bonds as well undertaking programmes that kept or bind interviewees together. Some interviewees explained that school-based network meetings served as ''family gatherings'' where they reunited with old friends to ''ignite memories of our youthful days'' (Interviewed on 25 th May 2016, Accra). It was also the place where vital information and contacts were exchanged. Strong social ties developed from school networks were also crucial in staff recruitment within the Ghanaian development sector. The research data suggests that social events including old-school unions played an important role in the development of informal networks among development workers. Consistent with the findings of Eyben (2011), social events led to the integration of social life into the working practices of donor representatives and NGOs. For example, interviewees mentioned several instances where they attended annual meetings and end of year dinners which created opportunities for social bonding. Being members of such networks served as spatial cleavages within which information about potential funding opportunities and getting to know the big men and people in high places were shared. Speaking about the importance ] some funding opportunities you cannot find them in the newspapers but through your contacts'' (Interviewed on 20 th April 2016, Tamale). This was informed by the idea of helping a brother which highlights the significance of positional embeddedness in which network members enjoy informational and social benefits. The network structure provides access to valuable information and appropriate resources which are useful for increasing cooperation (Granovetter 2017).
Friendship developed with school mates were perceived differently from those established on ethnic and tribal lines because they were national in scope and not confined to specific geographical areas. Thus, old school friendships traverse different regions and extend beyond ethnic groups. For example, interviewees described about how their school mates came from the ten regions of Ghana and most of them were development workers employed in different sectors including education, health and agriculture. Emphasis was often given to the social diversity within these school-based social networks which provided opportunities for members. Some respondents stressed that their school mates were 'big men' occupying highstatus positions in donor and government institutions which gave them access to critical resources. A respondent described a common narrative by stating that ''the boss was my classmate. I went there and told him we need money and immediately he said, come and take the forms and submit the application. So, at that level, because I knew him, it was easier'' (Interviewed on 23 rd April 2016, Tamale). Another interviewee added: ''for these funding opportunities, we were introduced by a friend [school mate] who is a big man now and we put in our proposals'' (Interviewed on 1 st April 2016, Accra). These excerpts demonstrate the significance of NGO leaders benefiting from their affinity with school mates who occupy higher positions in the funding hierarchy. Big men used their social and personal connections to operate behind the scene in helping NGOs through their performance of generosity.
It is important to clarify that within school-based networks, the so-called big men used their positions of power to help their acquaintances in the NGO sector in securing resources for project implementation which benefits the public. This indicates that for NGO-donor relationships, big men used their positions not for financial gains but rather to help some NGO leaders achieve their goals of promoting development. Thus, 'big men' within the Ghanaian development sector use their educational capital in facilitating development rather than for personal gains. This contradicts earlier studies suggesting that big men use NGOs to further their interests (see Chabal and Daloz 1999;Mohan 2002). The findings indicate that Ghanaians working as representatives of bilateral and multilateral donor agencies were not motivated by greed and self-maximising attitudes that allows only a few elites to benefits from donor funds. Rather they used their positions of power in helping their colleague NGO leaders secure funding for project implementation which serves the good of the society. As one NGO leader explained: ''I told you that getting funding for projects is always about who you know, people that you have in high places. If you know someone, you've already qualified 50% (Interviewed The rise of school-based networks and its associated 'big-menship' also signifies the emergence of new forms of ties built around former schoolmates which is a reflection of the multiple realities within which development workers are situated. Similarly, Alikhan et al. (2007) and Yarrow (2011) highlight how development workers in Ghana drew on their friendship developed during school and social activism days in accessing resources for their operations. This is an indication that donor relationships are based on informal and interpersonal relationships and not only limited to institutionalised business relationships. The findings are supportive of current research on the importance of friendship in the creation of social identities (Eyben 2011;Fechter 2012;Heuser 2012). The findings on of the use of interpersonal relations in resource acquisition also raise questions about the extent to which acquiring the latest development jargon helps NGOs to secure funds. In fact, the literature highlights that becoming a professionalised organisation and identifying with the latest buzzwords and fuzzwords helps NGOs to secure funding (Cornwall and Brock 2005;Roberts et al. 2005). However, as our findings show, in the case of the Ghanaian NGOs interviewed, developing and sustaining personal relationships with donor representative is a key determinant of their ability to secure funding. Our argument is not to downplay the importance of 'formalised application procedures', however, we emphasise the centrality of personal connections in resource acquisition.

Friendship through past professional experiences of implementing projects
The empirical evidence indicates that friendships developed through historical engagements and past professional provided opportunities for collaborating professionally. According to one NGO director who used to work for an international NGO (INGO) in Tamale, he always drew on his work-related ties with other INGO and donor staff for professional engagements. Similar sentiments were shared by other interviewees who argued about the importance of maintaining previous work-related ties and experiences. Another NGO leader who works as a consultant expressed this same argument by putting it this way: Who you know matters so much. If you don't have strong connections or you have not worked in the sector before, you won't get funding or donors asking you for collaboration because they want track record. I worked for an INGO and had colleagues who are now in big-big places. And before that I worked with a consultancy firm that gave me exposure which made many people to know me and my experience. So, my track record and existing contacts in the field has helped us (Interviewed on 12 th September 2015, Tamale).
The above statement demonstrates the dynamism associated with the relationships between donor representatives and NGOs. This demonstrates that the relations between donor representatives and NGO leaders are always in fluidity because they are being reworked through their job mobility. When donor and NGO staff change jobs, they carry with them their professional networks and connections. In this regard, career friendships are sustained and Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 17 persist over a long time which often informs current relationship between some donor and NGO staff. This finding mirrors Lewis (2008) about how people who worked for Oxfam in the past ('ex-fams') served as boundary spanners by facilitating relationship building between the UK government and the third sector.
In the case of the Ghanaian development sector, such social connections and personal contacts are important mediums for sending information about suitable and dangerous organisations to collaborate with. These connections although are outside of formal procedures, they are critical in helping donors in the choice of partners to engage with. For instance, donor representatives often emphasised the importance of communicating among friends who shared information and experiences of working with national NGOs. One representative recalled: Sometimes from informal conversations with fellow donors, we have a very good idea which NGOs are working here. I may talk to a colleague at DANIDA, DFID or Christian Aid that we want to do agroforestry, then he will say, talk to these NGOs, they have very good record. They are reliable and credible, or they will warn you, be careful of these NGOs because if you work with them, they will give you problems (Interviewed on 28 th April 2016, Accra).
Given that development workers tend to know each other and are sometimes close friends, exchanging information was a common phenomenon. This echoes Hilhorst's (2003) argument that development workers often bring their personal contacts and connections to their work and blurs the distinction between official and unofficial discourses. As the findings highlight, official selection for NGOs involves calls for grant proposals by donors. NGOs could submit a ''catchy proposal because that is what donors always want'' (Interviewed on 15 th December 2015, Tamale). However, they might not be able to deliver on the actual project. In this regard, using social connections and norms of trust with former donor and NGO colleagues helps in verifying and deciding on the right NGO to collaborate with. This produces a complex reality and shows how formal and informal processes are interwoven together with regards to donors' selection procedures for partners.
Interviewees also described how formal engagements with some donor agencies evolved out of informal discussion and such relations involve elements of reciprocity. Accordingly, interviewees suggested that friendship within the Ghanaian development sector revolves around reciprocity which helps in deepening relationships. Many interviewees claimed they resorted to tactics such as invitation to lunch breaks, dinners, investing time and resources to visit each other and also sending souvenirs and Christmas and birthday cards. Other means of showing solidarity and reciprocity was through attendance at social programmes such as marriage, naming ceremonies and funerals. Some interviewees explained that attending such programmes helped in keeping the informal relationships between donor representatives and NGO leaders because ''it is those elements that breaks the ice because exchanging one good favour is part of the informal relations by which most of the time, we get things done'' as  Putnam's (1993) argument that individuals are embedded within their social environments influenced by norms of trusts and reciprocity. Direct reciprocity ensures the existence of cooperation between NGO leaders and donor representatives over longer timespan because they are part of the moral norms of society.
Having presented and discussed the findings on the institutions governing the relations between donors and NGOs, we turn to our second research question.

Who gains and who loses from attempts to formalise or strengthen the explicit rules governing donor-NGO relations in Ghana?
Interview data suggests that informal networks and personal connections among development workers are important for the achievement of project outcomes. As part of mechanisms for maintaining their relationships, interviewees explained that they had contact persons both within NGOs and donor agencies who oversaw project implementation t as one project officer puts it: Being able to have a focal person at the donor office who I can talk to and tell that these requirements we're supposed to adhere to, these are the challenges we're facing. How do you we address it together? (Interviewed on 26 th February 2016, Accra).
This officer acknowledges the importance of maintaining direct contact and constant informal communications with donors because they served as mechanisms for getting access to key stakeholders. Many NGO leaders suggested that contact persons at donor offices were well connected to key stakeholders such as government officials and politicians. Using diplomacy in accessing government officials was a widely shared perspective among donor representatives. One donor staff at an Embassy in Accra explained that NGOs and donors have different capacities in influencing policies because ''we occupy different seats at different tables. I sit here as an officer looking at the technical work, but I also do more of political diplomacy. I engage in diplomacy, lobbying and all that just to make things okay'' (Interviewed on 17 th May 2016, Accra). This suggests that donor representatives are able to lobby governments through diplomatic means but this remains largely limited for NGOs' leaders. For this reason, building informal relations and rapport with contact persons who serve as gate keepers provides opportunities for getting access to high government officials. Doing so contributes immensely to how NGOs are able to achieve their project outcomes. Speaking about how donors influence government through diplomacy in relation to concerns raised by NGOs, a donor staff noted: We handle issues in very different ways. We engage in diplomatic efforts and informal discussions with government officials and discuss the issues. That's our role and we have tried a number of times which really saves NGOs much more time because they don't have such diplomatic channels (Interviewed on 13 th December 2015).
This statement suggests that informal engagements with donor representatives help them to lobby governments on behalf of NGOs. This is consistent with Yarrow's (2011) observation that interpersonal relations are used in linking NGO staff to public actors including politicians and the media. However, as the findings highlight, donor representatives act on behalf of NGOs in promoting the good of society rather than their own self-interests especially when engaging government on issues such as NGOs' advocacy and civic space.
The argument being made here is that personal connections rather than formalised rules and procedural norms grant NGO and donor representatives the opportunity for influencing key actors like politicians to undertake specific initiatives that benefits entire society rather than specific individuals. Interviewees echoed the sentiment that informal relations are critical ingredients for getting some donor representatives to act quicker on some decisions or initiatives which otherwise they would not if such requests were made only through formal procedures because of the bureaucracies involved (Interviewed on 28 th January 2016, Tamale). Notwithstanding, some interviewees widely claimed that relying too much on personal relations often led to allegations of co-optation or NGOs being tagged as donor puppets which affected their credibility and legitimacy with intended beneficiaries. This in turn negatively affected their survival prospects and potential roles as watchdogs. Some interviewees explained that many NGOs due to their close connections with donors are less reluctant to critique their relationship even when things are not working rightly. This raises questions about the legitimacy of such NGOs because of their inability to represent the views, needs and interests of their intended beneficiaries. Such concerns have led to NGOs facing crisis of legitimacy which undermines their autonomy (Walton et al. 2016).
It is worth mentioning that although the relationship between NGO staff and their contact persons at donor offices is expected to be formal, they extend beyond the formalised lines of communication for information exchange. As noted by donor representatives, their understanding of the Ghanaian context creates opportunities for engaging in informal discussions on project outcomes that were perceived as difficult. This gave them the room for manoeuvring and also adapting their requirements to meet NGOs' needs. In doing so, they admitted these informal channels of communications were meant for getting feedback on their engagements. Donor representatives and NGO leaders explained that they wanted to develop shared visions and a better understanding of their performance on projects outcomes together because they had a common interest of improving the lives of intended beneficiaries. For this reason, using informal discussions as a feedback mechanism was very helpful. The ability of NGOs to use informal discussions through what one NGO director calls ''head-tohead discussion'' (Interviewed on 26 th May 2016, Tamale) was due to their resource interdependency. This helps in negotiating unfavourable funding requirements because donors rely on NGOs for achieving their project goals and reach out to their intended beneficiaries. Again, NGOs use their local knowledge and embeddedness to influence the conditions set by donors. Negotiating donor conditions requires informal conversations and relationship building as shown by the following extracts: Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 20 They have a system in place where you can renegotiate the work plans you submitted in your proposal and say that I think the plan we presented might be problematic because of factors prevailing in the environment. And therefore, we propose we do it this way and they allow for that. Sometimes even the deadlines for report submission, we also negotiate (Interviewed on 14 th January 2016, Tamale).
Having that focal and contact person responsible for the project who you're constantly having interactions with, that's very useful because you develop relationships just beyond the life of project. When you have informal relations and you send in a report, you can get feedback quickly (Interviewed on 3 rd November 2015, Accra).
According to interviewees, having regular informal conversations made it easier for timely feedback because as one NGO staff explained ''not all concerns need to be addressed formally because ''donors are humans and not buildings who understands realities with NGO work'' (Interviewed on 4 th March 2016, Wa). This sentiment was shared by some donor representatives who argued that not having informal conversations makes their relationships too mechanical where interactions only happened during reporting periods. In fact, they explained that face-to-face interactions was faster and better than formal reporting especially when concerns need to be raised urgently with NGO leaders and vice versa. This happens as donor representatives often visited NGOs during project implementation, which creates opportunities for informal engagements that help in influencing donor agenda. This is captured in the following extract: As part of their [donors] monitoring and evaluation process, they would come and visit the organisation and project sites to get a sense of what is going on. When they come, we interact informally, and they give us their feedback. Sometimes we also negotiate on their reporting'' (Interviewed on 16 th May 2016, Accra).
Interview data suggests that interviewees preferred addressing their concerns through informal means rather than relying on conventional formal channels. According to some interviewees, the aim here was to ensure that they ''do not put each other in a tight corner because formal reporting is seen as you're complaining too much'' (Interviewed on 12 th February 2016, Wa). Similar account about how informal feedback mechanisms were used by in sharing contextual information with donors has been reported in the literature (Agyemang et al. 2017). This indicates that informal feedback compliments formal feedback mechanisms. The existence of formal and informal compliant procedures is akin to the facilitative behaviours reported by Romzek et al. (2012) in their study of network organisational actors.
The empirical evidence suggests that sometimes informal relations make donor representatives blush over some requirements by becoming more lenient in how they dealt Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 21 with NGOs not meeting standard procedures. It was also due to their understanding of the local context within which NGOs operate. This clearly shows how the development of informal relations helps in influencing and shaping the practices of donors. However, the ability of NGOs to influence donor requirements through informal means depended on organisational characteristics including leadership style and personality of project officers.
In what follows, we discuss how informal relations are used as accountability mechanisms by donor and NGO staff.

Informal relations and accountability
Evidence from this research shows that outside formal and official reporting requirements, NGO leaders and donor representatives dwell on informal relations as accountability mechanisms especially for non-performing organisations. There was a shared understanding among NGO leaders that they benefited from structured informal interactions and norms from donor representatives in supporting their weak governance structures. This occurred through regular informal interactions and dialogues which provided opportunities for resolving conflicting accountability requirements (Romzek et al. 2012). For instance, an NGO leader stressed that some donors prefer more face-to-face discussions rather than communicating through official channels such as emails. They explained that there were higher chances of misunderstanding messages sent over emails. In this regard, face-to-face interactions resulted in better cooperation because such discussions were opened and gave a better understanding of what was expected from each partner so that ''you can hold them [donor representatives] by their words'' as one NGO staff stated (Interviewed on 18 th April, Tamale).
Verbal assurances became informal accountability mechanisms because it helps NGO and donor representatives to engage in follow-up communications to ensure that commitments made were attained. Face-to-face interactions with contact persons at donors' office according to NGO staff helped in assessing their performance and also ensured that they communicated their inability in meeting reporting requirements. This assertion was supported by this quotation from a donor staff: If you're not able to submit the report, you need to communicate. It is a matter of regular communication such as phone calls. Partners need to communicate with us if they realise that it will not be possible to deliver the outcomes within the project timelines. They either ask for a budget extension or for a no cost extension (Interviewed on 14 th March 2016, Accra).
Informal relations were useful in facilitating committed listening with donors which allowed for negotiated accountability. This is in sharp contrast to the findings of Agyemang et al. (2017) who found that NGO workers in Ghana were unable to negotiate accountability requirements by donors. As the findings in this research demonstrate, interpersonal relationships helped NGOs to engage in strategies for influencing donor requirements. In addition, maintaining regular communication through face-to-face interactions was useful in in trust and confidence Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 22 building. The development of trust was identified by interviewees as one of the outcomes of their informal networks because it created a sense of shared norms and values. Shared norms are informal code of conduct for assessing the behaviour of actors which in turn determines their credibility and reputation. It was emphasised by interviewees that informal networks were the building blocks for interpersonal trust which also led to mutual accountability between donors and NGOs. Trust building was recognised as very important because it determined the ability of NGO leaders to offer suggestions for change to donors' accountability requirements. This is akin to the idea of relational embeddedness (Granovetter 2017). From a relational embedded perspective, trust has multidimensions comprising of personal good will trust, personal competency trust and social trust (Eng et al. 2012).
Interviewees explained trust was developed through frequent direct interactions and friendships. Such friendships created opportunities for relying on what one staff calls the ''goodwill of friends'' to deliver on project outcomes. In doing so, they believed that their friends had the competency to achieve any agreed goals because people can be trusted in following rules and requirements. Consistent existing literature (see Lyon 2000;Eyben 2010;Romzek et al. 2014), our analysis showed that trust was understood as confidence and knowledge of the ability of their partners. Interviewees saw themselves as trusted partners which helped in sustaining their relations. Trust therefore served as checks and balances or accountability mechanism in sustaining their relationships. This mirrors Lyon's (2000) findings on the importance of trust in sustaining networks among traders in Ghana where he notes that traders trusted each other mainly because of the establishment of long-lasting trading relationships. A similar finding about the importance of trust-based relationships has been reported in the literature (see Eyben 2010).
Directly related to the discussion of trust is reciprocity and favours in networks. Although partnership between NGOs and donors is formal, some interviewees stressed that they often exchanged in informal favours. For example, NGO leaders explained that they reciprocated their donors by recommending fellow NGOs with credibility for future partnerships. They also provided background information on potential communities that donors would want to undertake future development projects. On the other hand, donor representatives reciprocated NGOs through the writing of recommendation letters for grant applications and also shared information on potential funding opportunities by other donors as illustrated in the following quotation: Because we have worked for some donors like SNV, UNICEF and WaterAid before, at times when a new project is coming, they can just recommend you. At times, they can even pre-inform you that this grant opportunity is coming, so you are aware and the moment it comes, you put in your proposal. Having these organisations as our referees has helped us a lot (Interviewed on 22 nd April 2016, Tamale).
NGO leaders explained that donor recommendations enhanced their credibility and reputation because they were perceived positively by other stakeholders. One respondent stated that ''now donors are even recommending us to potential donors, so it has increased Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 23 our visibility'' (Interviewed on 4 th May 2016, Accra). For this reason, it could be argued that adherence to informal accountability mechanisms brought rewards. Reciprocity was also expressed through collaborations in consortia. NGO leaders explained that they reciprocated good actions through future funding partnership with organisations with perceived limited capacity which enhanced their credibility and survival prospects. As in this research, exchanging favours helps in strengthening relationships between donors and NGOs which results in getting things done.
Despite the positive aspects of reciprocity, there was a widely held perception among NGO leaders that reciprocity by donors was much limited. This happened when NGOs freely shared information about difficulties faced in the implementation of projects. However, such information was sometimes interpreted by donors as failure by NGOs which affected their future funding and partnership potentials. Accordingly, donor representatives stressed that some NGOs abused the trust they had developed in them and failed to deliver on project outcomes. This resulted in sanctions including blacklisting from future funding opportunities, termination of existing contracts and relationships, damaged reputation and complete avoidance of relationships. The following extracts from an interview with a blacklisted NGO explains how sanctions imposed by donors affect resource mobilisation potentials: I have not really applied for funding from them [donors] because we are a bit cautious. We discussed with them and paid the money they said we had embezzled, and the contract was terminated. So, I particularly became a bit uncomfortable with their funding (Interviewed on 29 th April 2016).
The above statement indicates that abusing informal relations and accountability mechanisms was detrimental to the survival prospects of NGOs. For instance, interviewees emphasised that organisations that had failed to demonstrate their trustworthiness found it difficult to collaborate with others. Trust serves as the lubricant and the glue that keeps relationships together and has become an important attribute for collaboration (Bryson et al. 2006). During interview, a donor representative shared his experience of working with a nontrustworthy NGO which has informed their future partnership decision: Because we don't want to hang out our dirty clothes in public, we kept quiet. But we know ourselves, so next time when there is an opportunity; I don't go near such NGOs because I don't trust them based on our past experience (Interviewed on 12 th January 2016, Accra).
The empirical evidence suggests that the level of trust-based relationships changes over time with experience. The findings clearly demonstrate how shared norms including trust enable and constrain informal accountability in donor-NGO relations. Adherence and non-adherence to informal accountability mechanisms result in rewards and sanctions including enhanced resource mobilisation potentials and damaged reputation respectively.

Conclusion
This article has examined NGO-donor relations in Ghana, revealing the importance of informal institutions -including friendship, school-based networks and shared norms of trust and reciprocity. In so doing, it qualifies the often assumed antipathy between formal and informal relations, revealing instead how they work hand-in-hand. This observation challenges a normative assumption of the primacy of a formalised relationship between NGOs and donors, arguing that this understates the role of informal elements such as personal connections, contacts and friendships in 'getting things to be done'. Again, it shows that while NGO and donor representatives work within bureaucratic structures, they do not embrace these structures blindly and uncritically; rather they exercise their agency through the use of informal and personal connections to influence and work around these structures. Influencing donor agenda through personal connections seem problematic because this is often perceived to be clientelistic and corrupt. However, that there is no iron law linking informal relationships do not necessarily lead to corruption and patronage that undermines development. Viewing informal relations as synonymous with corruption presents a partial and unduly negative view of the culture governing donor-NGO relations.
We show that Ghanaian development workers, including NGO and donor representatives bring their personal connections to their work, including those arising from kinship, ethnic, and educational ties. Furthermore, in the complex environments within which these development professionals operate, it is often the personal elements (including friendship and trust-based networks) that ensure project outcomes are achieved. This supports the argument put forward by Eyben (2010) that strong and healthy social relationships need to be viewed more explicitly as an integral part of healthy and effective NGO-donor relations. This is not to deny the importance of formal rules and professional norms, but to emphasise how it is their interaction with informal rules and norms that leads to positive as well as negative development outcomes.
This finding can be linked to longstanding discussion of development partnership. Informal accountability based on shared norms has the potential to support fairer, more open and honest inter-agency relationships underpinned by ideals of peer and downward accountability alongside upward accountability. Informal accountability mechanisms are complements and not substitutes to formal accountability rules because of their ability to reward (through enhanced reputation, information sharing on future funding opportunities and collaboration in consortia) and to sanction (through loss of opportunities, termination of contracts, severing of relationships and damaged reputation) actors within NGO-donor networks.
Our findings have implications for development policy and practice. First, the role of informal institutions needs to be made explicit when donors seek ways to improve 'capacity', foster 'organisational development' and promote better leadership. Second, a more nuanced understanding of the role of informal relationships is also relevant to how donors seek to strengthen their own programme management capabilities: in the distribution of centralised Bath Papers in International Development and Well-Being Friend or Patron? 60/2020 Kumi and Copestake 25 and within-country, and in balancing local and international recruitment of staff, for example. Third, informal practices affect inclusion and exclusion of both NGOs and individuals from networks, consortia and funding.

Notes
1. While we acknowledge that the complex nature of the role of NGO leaders and their ability to inhabit different lifeworlds by maintaining a wide range of connections with actors such as national or local level politicians, donor agencies and community leaders (Hilhorst, 2003), our analysis focuses on their relationships with donor representatives rather than politicians or community leaders. 2. The following criteria was used in selecting NGOs: a) The NGOs must have been in existence for 5 years and must have had experience with donor funding; b) must have an annual budget of not less than GH₵ 20,000 (US$5, 263.15) and at least 4 paid employees; and c) the NGO leaders are willing to grant access. The selection criteria ensured that large, medium and small-sized NGOs operating at the national, regional and district levels were included in this research. For donor agencies, the selection criteria are as follows: a) the organisation must be a bilateral, multilateral, international NGO and philanthropic development agency providing funding for national NGOs; b) the donor representative must be a key personnel (e.g. Country Director, Chief of Party, Programme Director, Programme Officer, Grants Manager) with formal responsibility for and/or practical knowledge of NGO funding landscape; c) Sampled NGOs must have indicated they have received funding from the donor; and d) the donor agency must be active at the time of data collection and willing to be interviewed. 3. It is important to clarify that our understanding of personal relations and how it is applied by donors and NGOs focus on how respondents explictly talked about them rather than affording it a self-evident and empirical status.