Aid Unchained: Examining Development Project Management Practices at Aid Chain Interfaces

By analysing a comparative case study investigating a development project implemented in Uganda and Vietnam, the article aims to understand how donor directives travel and translate into actual practices in aid chains. Making use of Norman Long’s concept of the interface, we focus on the interfaces between organizations to examine the negotiation of everyday project practices. Based on practice theory, our analysis unpacks how directives are filtered through the power relationships that shape practices at the various interfaces. We find that organizational relations between southern organizations are just as power-laden as north–south relations. Our analysis also shows that neither the management directives nor the freedoms that were granted to the participating organizations resulted in uniform practices and that practices did not have the same implications for organizations. Hence, the aid chain concept tends to simplify the complexities inherent in project systems comprising a multiplicity of vertical and horizontal organizational relations.


I. Introduction
There is no straight connection between policy and outcomes.
- Long, 2001: 25 A recent evaluation of the Strategic Partnership Programme (SP) financed by the Ministry of Foreign Affairs of the Netherlands found that 'The long-term commitment and flexibility that MFA [Ministry of Foreign Affairs] provides to N-CSOs 1 [Northern-CSOs] is not always transferred to S-CSOs [Southern CSOs], many of which are still bound to annual contracts, activity-based budgets and strict reporting requirements' (Ministry of Foreign Affairs of the Netherlands, 2019: 13).While the policy of the Ministry provided leeway in certain aspects, these freedoms had apparently failed to reach all of the participating organizations.The policy framework underpinning the SP acknowledges the risk that the design of aid chains can constrain the organizational autonomy CSOs require to assume political roles.To this end, the Ministry even commissioned research to study the effects of aid chains on programme design (Ministry of Foreign Affairs of the Netherlands, 2017: 39).Yet, the programme was not able to mitigate these risks.So why did the directives 2 in the SP not reach all organizations across the aid chain equally?
In this article, we take a closer look at organizational relationships across the aid chain by examining a development project implemented in Uganda and Vietnam, which is part of the aforementioned SP .The project was implemented through Oxfam Novib, the Dutch civil society organization that received the biggest share of the funding from the Ministry within the SP.Uganda and Vietnam were chosen as research sites as they both implemented the same thematic components and relied on the same administrative structures as Oxfam Novib acted as executive affiliate for both country offices.The article aims at unpacking how directives are filtered through the power relationships deriving from the complexity of the aid chain both vertically and horizontally.
We make use of Long's interface approach in order to understand how directives are renegotiated at the different interfaces of the aid chain.Existing studies often analyse north-south aid relations, depicting how northern directives curtail the operational space for southern organizations (e.g., Bawa, 2013;Nabacwa, 2005;Suárez and Gugerty, 2016).This uni-directional image is equally reflected in the concept of the aid chain.While we do not deny that directives can be constraining, we add to this literature by examining the relations across the entire aid chain, including horizontal relationships between southern organizations.In order to contribute to a more nuanced understanding of how donor directives translate into actual practices, we analyse our data by means of a practice theory approach.Thereby we show that directives do not necessarily travel across the aid chain: their meaning is mediated through organizational relationships at the various interfaces, which also results in non-conforming practices.Due to an overt focus on power relations between international donors and implementing partners in the development literature, ironically, many of the existing inequalities within the project implementation system are concealed.
Our article suggests rethinking and retheorizing how policy models are transformed in practice by those working with them.
The remainder of the article is organized as follows: first, we embed our article in the available literature that deals with aid chains and anthropological accounts of project relations.We then present practice theory as the theoretical backbone of the article and illustrate our use of Long's interface approach.In the empirical section, we introduce our research design, methods and case study.We successively analyse our data examining how project management practices within and across organizations relate to the project's directives.The article closes with a discussion of the findings and a conclusion.

II. Literature Review: Aid Chains and Their Social Lives
The concept of the 'aid chain' is largely taken for granted in publications and seldom scrutinized. 3The majority of authors use it without defining it and refrain from commenting its connotations (e.g., Bawole and Langnel, 2016;Bebbington, 2005;McGee, 2013;Tembo, 2004).The few publications that deal more thoroughly with the concept of the aid chain use it to refer to the catenation/arrangement of development organizations in a vertical hierarchy.Funding and requirements are said to flow from the donor via international, national and subnational NGOs, acting as intermediary organizations, to the beneficiaries (the exact number of tiers depends on the project), while results flow up the chain (Cotts Watkins et al. 2012: 287;Elbers, 2012: 23;Koch, 2009: 109;Wallace et al., 2007: 13).
According to this model, policy derives from the top level of the aid chain from where it is handed down (Williamson, 2010: 57).Such a view is often substantiated by principal-agent rhetoric, as donors (principals) rely on NGOs (agents) in order to implement their programmes, underlining the hierarchical and contractual relationships between organizations (Burger et al., 2018;Cotts Watkins et al., 2012;Michaelowa and Borrmann, Progress in Development Studies 24, 2 (2024) pp.97-112 2006;Winters, 2010).Some authors explicitly highlight the north-south dimension in CSO funding relationships (Bawa, 2013;Nabacwa, 2005;Suárez and Gugerty, 2016).In line with a transaction-cost approach, organizations in aid-receiving countries are sometimes depicted as mere sub-contractors in the literature (Cotts Watkins et al., 2012: 287; see also Bradley, 2009: 107).
While the degree of vertical control that is exercised through funding relationships certainly differs from programme to programme, the literature cited above falls short in that it takes policies at face value.A rich body of literature, with which this article is aligned, conceptualizes project practices as relational and embedded in everyday life instead of derived from top-down mandated policies (e.g., Eyben, 2010;Peters, 2020;Watanabe, 2019).For instance, Hilhorst aptly demonstrates in her ethnography of a Philippine CSO how development interventions can be conceived of as 'interlocking projects of different actors with diverging life worlds, aims and ideas'.Additionally, Bierschenk (1988) describes the struggle about decision-making power and resources as an 'arena of negotiation'.Also, the discrepancy between policy and practice has been the subject of many anthropological accounts of development projects (e.g., Beck, 2017;Krause, 2014; see also Fejerskov, 2018 for an overview of the different strands of literature that engage with how ideas and practices in development change when they move across contexts).In his seminal work, Mosse conceptualizes policy as an institutional practice and analyses the social life of development projects, oscillating between the diversity of interests and the production of common narratives that serve these interests.He sheds light on the agency of seemingly subordinate actors in development projects who, despite their room for manoeuvre, work towards sustaining the dominant policy models, as it is in their interest to keep the representation of the model alive (Mosse, 2005: 10).However, in the academic literature on civil society relations, there is an overt focus on analysing north-south power dynamics between donors and CSOs (Elbers and Arts, 2011;Kamstra and Schulpen, 2015;Kumi and Copestake, 2022).Our article contributes to filling this gap by including (funding) relations between southern organizations both horizontally and vertically.
Another research strand takes concrete concepts or buzzwords (such as aid effectiveness, adaptive management or ownership) that dominate international fora as a point of departure and dissects their workings (e.g., Brown, 2020;Cornwall and Eade, 2010;Gutheil, 2023;Keijzer and Black, 2020).Hasselskog (2022) traces, for instance, how discussions around ownership are increasingly about technical aspects such as procurement, public financial management and results management instead of domestic agenda setting, thus depoliticizing the ownership principle.In a similar manner, this article examines the actual workings of the aid chain.

III. Theoretical Framework: Practices at the Interface
In line with previously cited authors, our aim is to advance beyond the analysis of written policy models, instead relying on a practice theory approach to understand how these models are implemented and how actors engage with them in the aid sector.Practice theories can be described as a family of related theories building on the philosophical work of Wittgenstein and Heidegger (Hui et al., 2017: 1).Countering structuralist and methodological individualist approaches to theorizing social life, practice theory grapples with the dialectical relationship between structure and agency (Reckwitz, 2016: 245;Ortner, 2006: 2).Practice theory can be described as a 'general theory of the production of social subjects through practice in the world, and of the production of the world itself through practice' (Ortner, 2006: 16).Practice theory, therefore, serves to balance between the conceptualization of project management practices as solely driven by individual choices or structural constraints.Practices are 'patterned actions that are embedded in particular organized contexts' (Adler and Pouliot, 2011: 5).They can best be understood from within their communities of practices, where their meanings are determined (Adler and Pouliot, 2011: 16).Hence, practices are contextualized and relational; they rely on shared meanings that are constantly re-produced and re-negotiated (Schäfer, 2016: 13).Consequently, practice theory facilitates the countering of Weberiantype depictions of project management practices as rational and impersonal, and that can be applied independently of the context.Practice theory reveals that different logics, meanings and discourses are tied to project management practices (even within a single project).This shows that practices (re)-produce social relationships and therefore have a stratifying effect.As our case study especially deals with project management practices across the 'aid chain', we find Norman Long's concept of the social interface useful for grappling with encounters between different actors and institutional realities.When referring to social interfaces, Long means 'social situations or arenas in which interactions become oriented around problems of bridging, accommodating, segregating or contesting social, evaluative and cognitive standpoints' (Long, 2001: 65).Interfaces thus provide us with sites at which we can analyse how practices are negotiated between different actors.Interface situations do not necessarily imply that there are two opposing sides (even though these situations are potential sites for conflict), but rather point to the complex encounters between different interests, modes of rationality and legitimacy (Long, 2001).Interface analysis grapples with the linkages, interactions and relationships between parties and analyses how the interface itself becomes an organized entity, subject to specific rules and practices (Long, 2001: 69).Instead of taking policy models at face value, development projects are analysed as processes that are shaped by the specific dynamics they encounter and create (Long, 2001: 69).

IV. Methodology The Case Study
The article relies on a comparative case study of two projects, which are part of the SP.The SP is a five-year programme (2016-20) funded by the Dutch Ministry of Foreign Affairs and comprises a budget of around 1 billion € (Ministry of Foreign Affairs of the Netherlands, 2017: 11).Its aim is to support CSOs' lobbying and advocacy capacities in different sectors, for which 25 CSO consortia were chosen through a tendering process.Of these 25 consortia, all but one are headed by a lead Dutch CSO that receives the funds.In total, there were 366 country partnerships implemented in 72 different countries.In most consortia, the funds are channelled further to partner organizations in the Global South.The SP is underpinned by the Dialogue and Dissent Theory of Change, which emphasizes the social transformative approach of the Ministry representing 'a shift in focus from aid aimed directly at combating poverty through service delivery to aid aimed at tackling the root causes of poverty and inequality through lobby and advocacy' (Ministry of Foreign Affairs of the Netherlands, 2017: 2).CSOs are seen as vital actors in their own right, who are able to empower the marginalized and challenge existing power asymmetries and structural inequalities (Ministry of Foreign Affairs of the Netherlands, 2017: 5).In line with recent calls to make development management more adaptive, 4 the SP encourages the use of flexible theories of change (ToCs) and advocacy strategies, as well as the use of customized planning, monitoring, evaluation and learning (MEL) systems (Ministry of Foreign Affairs of the Netherlands, 2017: 9).
While a number of lead CSOs in the Netherlands were contacted and interviewed, Oxfam Novib in consortium with Stichting Onderzoek Multinationale Ondernemingen (SOMO) was the only one that agreed to grant access to its project partners for a larger case study.With a budget of 77 million € for the fiveyear period, Oxfam Novib received the biggest and participant observation.We visited both countries twice in order to conduct participant observation during project activities, such as meetings, workshops and trainings (Uganda: December 2018/September 2019; Vietnam May 2019/February 2020).In addition, project documents, such as guidelines, project reports, monitoring and evaluation frameworks and outcome harvesting forms were reviewed to compare written statements with practice.The interviews inquired into project management practices during design and implementation.Findings were shared with the interview partners and interviewees were confronted with other stakeholders' statements in order to cross-check their views and detect inconsistencies.In total, we spoke to 52 individuals and held 10 group discussions.Document analysis was used to prepare the interviews and field visits and triangulate our findings.All interviews were subsequently transcribed and coded with the help of MaxQDA software.
There are several limitations to the methods used in this article.Although it would have been ideal to spend more time observing practices instead of talking about them, this was not possible due to the geographical spread and variety of actors as well as the limited resources of the research project.While practice-based research is often equated with participant observation (see, e.g., Schmidt, 2012: 49), we agree with those researchers who critically question the assumption that the visual is authentic in itself (Nassehi, 2006: 231;Schäfer and Daniel, 2015: 45).While the meanings of practices are to a certain extent public and collective, it is the researcher's task to understand the social space in which they are enacted, in order to decipher their specific meanings.Therefore, we found that a combination of interviews, document analysis and participant observation was most amenable to the goal of understanding that social space and helped us to neither take observations nor talk at face value.
In addition, due to limited resources in the research project, it was not possible to include more projects funded by other donors share of funding from the Ministry within the SP.Among 17 of Oxfam Novib's partner countries, Vietnam and Uganda were chosen.Multiple selection criteria were used: first, two different continents were supposed to be represented in order to allow for cultural variety.Second, Oxfam Novib had to be the executing affiliate for both countries (put simply, this means that the respective country office makes use of Oxfam Novib's administrative structures) to allow for comparability.Third, both countries needed to implement the same thematic components in order to be comparable (the SP as implemented by Oxfam Novib has three thematic components but not every country implements every component).Vietnam and Uganda fulfilled all the criteria.In both countries, the 'Financing for Development' (F4D) and 'Right to Food' (R2F) thematic components were implemented, which are separate programmes under the SP .The countries' differences in terms of their CSO landscape and their history of relating to Oxfam Novib, as well as the thematic differences, allowed for the variety needed to contextualize practices.

Data Collection and Analysis
The research is multi-sited and the research approach can best be described by Reinhold's concept of 'studying through' (Reinhold, 1994).Studying through describes a strategy which follows a policy across different locations through to those affected by them (Wedel et al., 2005: 10;Wright and Reinhold, 2011: 87).Instead of only relying on one specific organization or set of actors, the focus of the analysis lies on the practices of project management within the project's boundaries.We started interviewing different lead CSO partners in the Netherlands and subsequently focused on our Oxfam case study by covering the partner countries' offices and partner CSOs on the national and subnational level in Uganda and Vietnam.In order to identify practitioners' day-to-day practices, we relied on semistructured interviews, informal conversations as control groups.In order to reduce this bias, interviewees were asked to compare this practice to the practices of other donor-funded projects they were working on.This helped to put our findings into perspective.While our data are limited to CSOs, we are convinced that our findings are indicative of relationships found in other forms of development, such as in bilateral aid interventions as Haley Swedlund's research suggest (Swedlund, 2017).We include additional literature describing project practices in the analysis section, in order to support our findings.More research should be conducted to validate this claim and to include projects that work, for example, through contractors, which adds another layer of complexity.

V. Analysis
In this section, we will analyse actors' and organizations' project management practices regarding project design and adaptation, templates and directives, cross-organizational collaboration and reporting.We show how policies are transformed in practice by those who work with them on an everyday basis.By making use of Long's concept of the interface, we further explore how collaboration between and within organizations takes place, turning interfaces into organized entities despite the variety of practices and their varying meanings.

Design and Adaptation
We find that the project design is not solely shaped by the Ministry's framework but is mediated just as much by the common history that organizations and individuals share.While the global ToC was developed by Oxfam Novib, country offices were requested to come up with their own country-specific ToC to further specify their project goals.Instead of publishing a call for proposals, the country offices invited CSO partners they had worked with previously to jointly develop the project.The project was framed by the project managers in the country offices rather as a continuation of past interventions: 'because to Oxfam this is a programme, different projects come in and build upon where previous projects stopped' 5 (Interview, Oxfam Uganda Country Office, December 2018).The CSO partners were also happy about the opportunity to continue their ongoing work without a heavy proposal writing stage.Thus, the interface between country offices and CSOs was in this instance characterized by routine interactions building on a joint history of collaboration.
The country offices welcomed the thematic focus of the programme, too, as it fell in line with strategic objectives and previous programmes.However, Oxfam senior programme managers at the strategic level in Uganda raised the issue that it was generally hard for them to follow an institutional or strategic vision, as they were dependent on the funding that was coming in, which was tied to certain outcomes: The country offices thus face difficulties, when it comes to the financing of their strategic plans.For the CSO partners, this is less problematic, as they are highly specialized and simply looked for a continuation of their activities.The practices involved in the design phase were not monocephalous: they did neither all derive from the donor, nor did they have uniform meanings for all actors.The Ministry's vision that CSOs are actors in their own right, who can set the agenda, materialized at the interface with Oxfam Novib, but lesser for the country offices and CSO partners.They could develop their own programmes within the ToC, but they were not involved in choosing the overall thematic focus.While this was evaluated positively by the CSO partners, the country office in Uganda was more critical.
As has been found by other researchers, interventions do not operate in an institutional void but are always part of a flow of events (Hilhorst, 2003: 204f;Long, 2001: 32).While for the Ministry, the meaning tied to the practice of designing a new policy framework equals a paradigm shift, for Oxfam and their partners, the design practices emerged out of a previous collaboration.In both Uganda and Vietnam, the F4D's approach of working with community groups built upon previous projects during which these groups had been founded.In a similar manner, Peters found in her study of a development intervention in Angola that the programme's decision 'to form community groups and hold public forums as central activities were settled on more for pragmatic reasons than for their theoretical promise to achieve the project's stated goals.They were quite simply activities to which different professionals and organizations could agree and jointly deploy' (Peters, 2020: 120f).Beck raised a related point, describing how the Guatemalan women targeted by new micro-finance providers were already familiar with the proceedings of getting a loan.Their expectations significantly shaped how the new projects were implemented (Beck, 2017: 96).
A monocephalous 6 view of development projects obscures not only how different actors are involved in decision-making and approvalseeking but also suggests that adaptation is not possible.On the contrary, we found in our research that project managers were often concerned with adapting activities, in the sense of changing timelines and budget lines.In a workshop we attended, one organization complained that they could not make use of the budget devoted for rent, as the rent was already covered by a different donor.In the end, they could use the budget line for a contribution to the salary of administrative staff.Smaller changes thus took place during implementation, while bigger changes were mostly included as part of the annual planning.Even though changes up until 10% per budget line (which is quite significant, as there were only seven budget lines) were permitted without seeking approval at the level of Oxfam Novib or the donor, common practice was that they were subject to the country office's approval.However, some raised the issue that approval was too slow in cases in which they filed for changes: You must seek approval, and then the bureaucracy becomes a problem.I have to write to my supervisor, who should relate my concern to the head of the institution, then the head of the institution writes to the country office and that takes time.Just following up on a few things would result in so much time wasted' (Interview, National CSO 4 Uganda, December 2018) Thus, it is not only the monocephalous donor bureaucracy but also internal communication lines in CSOs that lengthen the process of seeking approval.The more interfaces are involved, the lengthier is the process.Another CSO interlocutor explained that the timespan needed for approval processes 'really depends on how far away you are from the decision maker' (Interview, National CSO 3 Uganda, December 2018).Approval seeking is a relational exercise, which is facilitated by personal contacts.This is easier in smaller and less hierarchical organizations.In one instance, a project manager wanted to reassign the budget that was designated for a workshop to a campaign opportunity that had arisen.As her organization is very small and the campaign was considered a critical opportunity, the head of the organization directly related to the country office.In bigger organizations in which superiors might not be very accessible, staff need to pitch their proposals and might need to follow up, which became problematic in case the issue was time-sensitive.In Uganda, the heads of the participating CSOs all had a cordial relationship with the country office programme manager, which was based on a long-standing trusting relationship.On top of that, the CSOs were in a comparatively good negotiation position as they were major players in implementing Oxfam's programmes, which facilitated approval seeking at this interface.One head of the organization complained that 'in other instances, especially for funding agency x, this is not the case.We do not even have a focal person, there is no telephone number, nothing.Just an online system where you can fill in boxes with your requests.You might get an answer or not, but mostly we do not even try to change things' (Interview, National CSO 2, Uganda, December 2018).

Templates and Directives
While there are requirements defined in the ToC, outcome areas and monitoring and reporting templates, rules are not fixed and standardization in a project is never total.We find that templates and to a certain extent tools are subject to negotiation.While the framework sets the scene, especially in the design phase, project managers, in particular, at Oxfam Novib and its country offices make their own decisions, interpreting the directives or recommendations received from the Ministry.Apart from a country-specific ToC, country offices also had to come up with a MEL plan.While the ToC was a requirement by the Ministry, the use of certain tools for planning, monitoring and evaluation was up to the CSO consortia.Oxfam Novib used its own MEL methodology and preselected seven corporate outcome areas with matching outcome and output indicators, on the basis of which country offices and partners had to formulate their own outcomes and indicate who was going to be responsible for achieving and measuring the outcome.MEL requirements did thus not derive from the donor, but in this case from the international CSO.
While country offices and partners had to link their MEL methodology with Oxfam Novib, the interpretation of the template differed.Some presented detailed outputs and relied mostly on the given indicators, others structured the outcomes according to years, while again others presented intermediate outcomes.That the content of MEL methodologies is subject to negotiation can be further exemplified by the F4D team at Oxfam Vietnam, who suggested to Oxfam Novib that they would rather use their own capacity assessment tool instead of the one stipulated in the guidelines, to which Oxfam Novib agreed.
Although the interfaces between the country offices (in Vietnam and Uganda) with Oxfam Novib are technically the same, the research reveals that these interfaces are independently organized entities and that there is variation even within country offices.In Vietnam, we were surprised that the F4D and R2F components were implemented quite differently, even though they were part of the same funding scheme at the same country office and thus located at the same interface.The astute project coordinator of the F4D team did not start a new project, but integrated the funding in her ongoing programme.This enabled the implementation of a larger intervention to which the F4D contributed.In order to get approval for this approach, she needed to negotiate with the programme manager at the interface with Oxfam Novib.Her good track record and the trust she had built with Oxfam Novib over time enabled her to carve out space for her approach.Her colleague in the same country office who led the R2F programme did not engage in such negotiations.While he agreed that it was good to link different projects working on similar topics, he perceived the mixing of different donor grants in one project as too risky.Thus, he did not challenge business as usual at the interface with Oxfam Novib.
These different approaches at the same interface gave rise to different project practices.For F4D, the joint approach meant that the project coordinator had to make sure that even though the partners submit only one report, she extracts the outcomes and shares them with the different donors who contributed to that outcome area.It also meant that she had to look for new funding when projects ended to continue activities.For R2F, the allocation of partner contributions for each donor was kept neatly apart.
The negotiations of requirements were also dependent on the relationships of actors at the different interfaces.In Uganda, partners particularly emphasized how accessible the country office was.One interviewee uttered, this kind of partnership I think for Oxfam is quite unique.By the time you reach a decision everybody is already aware, if you don't implement they will know why, as they have been part of the struggle with you.(Interview, National CSO 5 Uganda, December 2018) This particular pilot intervention in the R2F aimed at increasing milk production by feeding cows with a special type of pasture.However, as the rainy season started late, the pasture had not grown as expected, which delayed project results.Partners also involved the country office in their national level advocacy campaigns and their suggestions, for example, as to simplify templates or receive training with regard to certain topics, were taken up.However, not everyone shared this experience.In Vietnam, one partner raised the question of whether the Oxfam country office perceived itself rather as a partner or donor.On the one hand, they wanted to be kept up to date, but on the other, they were not an official member of the alliance and thus did not take part in any alliance activities.When we talked to the country office about this, we were told that this was a conscious choice to foster organizations' autonomy and ownership.However, the partner perceived this as ambiguous: if they were not part of the alliance, they should not follow up on the partner's activities as closely as they did.
Thus, the data reveal that the meaning of templates and rules is not fixed, but interpretations diverge and emerge in implementation at the different interfaces.While a close relationship between CSOs and country offices facilitates the adaptation of plans and requirements if it is considered trustworthy, a close relationship that is considered to assume a monitoring function hinders joint advocacy.A distant relationship that is only involved in monitoring but does not meddle in implementation can in turn also be considered a good working relationship (even though it has a tendency to obstruct adaptation).Swidler and Cotts Watkins (2017: 8) point to the relational dynamic of project management practices in a similar manner, describing how 'these rituals of monitoring and evaluation bind the brokers to the donors, and vice versa'.As country offices assume both the functions of supporting and monitoring partners' work, the interface is the site at which these modes of rationality compete and struggle for domination.CSO partners are active negotiators in the process of defining shared meaning of their relationship with the country offices, thereby (re)producing social order.They strategically decide to what extent they involve the country offices in their projects, which gives rise to an independently organized interface that is subject to its own emerging rules.The level and kind of involvement practised by actors can shift the meaning of the relationship and transform practices at the interface.

Cross-organizational Collaboration
While the metaphor of the chain suggests that organizations are chained to each other, we rather find that constellations of project actors are not fixed, but change over time.For instance, in Vietnam, there was considerable movement in the F4D alliances.While one of the alliances added two new members in early 2017 to expand their reach, four organizations that had previously come together to form a new alliance on tax issues in a different programme, joined the F4D in 2018.Later on, one of the organizations decided to leave this alliance because it considered the work as too political.Similarly, Peters demonstrates how partnerships that were initiated in the inception period of the programme she studied could not be realized because of disagreement over methods and personal animosities (Peters, 2020: 140).
In Uganda, one of the national organizations in the F4D had to leave the project because irregularities with regard to its financial management had been detected.This organization was in charge of funding partners at the subnational level and coordinating their work.When the organization left, the country office had to step in to fill the void.After consultations with all partners, they agreed that one of the subnational organizations should assume the coordinating role at the subnational level.The organization was thus promoted to a sub-granting organization and received a new contract for the last two years of the project period.This significantly altered practices at the interface with the Oxfam country office.Entering into a fiduciary relationship required much more frequent exchange and guidance.As the subnational organization was new to sub-granting and its structures were rather informal, they had less autonomy to negotiate at the interface with the country office if compared to the national CSO partners.The way they spent their budget was closely monitored and when they were hiring a new coordinator the country office participated in the job interviews to make sure that the new person would be well-versed in reporting.
While overall the national CSOs were in a better position to negotiate at the interfaces than the subnational ones, this also developed over time.One of the national CSOs worked for the first time with the Oxfam country office and explained that they felt that their work was not well reflected in the activities and indicators in the project design.Over time, they managed to build trust and carve out more space for the topics they deemed important: One of the approaches we used is that every time we had an activity or an engagement, we always invited Oxfam, we invited all participants and we kept telling them about the importance of [the topic we work on].We told them about the usefulness again and again, and we even had a training in which all partners were involved.(Interview, National CSO 7 Uganda, December 2018) When we returned to the project site the second time, we observed that this organization had become an integral part of the project and had started involving the local authorities in their activities at the subnational level, which considerably shaped the further direction of the project.Instead of being chained to the superior organizations, organizations and their relative positioning move during projects.In addition, they are entangled with other donors and in many instances in our research also with the same donor in several projects.Also, staff rotation was a serious challenge for many organizations, which meant that interface relations had to be built from scratch.

Reporting
In our case study, neither reports nor funds flow up or down the chain through all organizations.In case of the funds, CSOs receive the money from Oxfam Novib directly and not from the country offices.While some of the subnational organizations are funded through national CSOs, others receive their funds from the country office, thus skipping a level.Looking at the practice of reporting more closely, our findings clearly contradict the ideal depiction of the aid chain, stating that requirements are handed down the aid chain and reports flow it up in a linear fashion.Oxfam Novib provided the country offices with reporting templates, demanding a narrative as well as quantitative information to report in the International Aid Transparency Initiative as required by the Ministry.The country offices were responsible for collecting results from their partners to write the consolidated report.While in Vietnam, the country office gave their partners a more general, simplified reporting template, in Uganda project managers handed down (almost) the same version of the template to the partners, even though Oxfam Novib did not recommend that.Some of these partners handed down the templates even further to subnational partners.The partners (especially in Uganda) complained to the country office about the reporting workload and often about the complexity of the templates, which they found hard to understand.
Their suggestions were followed up by training and Oxfam Novib agreed to simplify certain elements in the templates and give more guidance and instructions on technical aspects.
Our data thus show that in development projects with many actors, the reporting practices are not equally distributed and are subject to negotiation among actors.Although in Vietnam, the 'translation' of results into official categories is mostly done at the country office; in Uganda CSO partners carry a significant part of the burden.Most interestingly, the majority of subnational organizations, who are the most direct witnesses of the results, are not part of the formal reporting chain.The official communication by Oxfam Novib in this regard was that they did not expect the templates to be handed down to the partners, because on the one hand the terminology and logic of the templates is hard to understand for those who do not work within the Oxfam system, and on the other because they wanted to foster partners' autonomy.The reason for the lack of template customization could be out of convenience (let the partners do the work) or out of a lack of awareness.
Hence, reports are not simply passed from one organization to the next, but reporting can be characterized better as a relational exercise.The final report is a product of multiple iterations of commented versions that do not flow in a uni-linear manner but go back and forth between stakeholders vertically and horizontally.Organizations require information from other actors (i.e., CSOs request information from the subnational level or other national CSOs), in order to compile project results and the higher levels get back to the lower levels with feedback and explanations as to how templates are supposed to be filled out.Reporting is thus a joint production that links stakeholders by creating a common version of project realities.Biruk (2019: 204) specifically points to this performative dimension of paperwork, stating that 'it is less important what documents stand for than how they arrange people around themselves'.A comment in one of the reports reveals that this is not only with regard to the lived events, but also with regard to other project documents: 'Explain how these norms and attitudes are shifting.It is important to make reference to the baseline reports and link your explanations' (Report 5: 14, comment 2).Creating such joint project representations is not only an exercise of nurturing collegial relations, but also an act of sustaining funding by satisfying 'multiple accountabilities' (Sullivan, 2017).
This does not mean, in turn, that the contractual conditions did not have an effect on reporting practices.The partners worked in alliances contributing to the same goals, which meant that they were dependent on receiving information from their colleagues and from the organizations at the subnational level.As they did not have a fiduciary relationship, and there was no budget line for information sharing, this was perceived as challenging at times.To overcome this challenge, joint review and reflection meetings were organized, which helped the partners to relate their contribution to the larger intervention.Before that, partners had to liaise informally, which was not remunerated.As Peters has aptly shown in her work, many of the social and relational tasks in development work can be classified as shadow or invisible work, as they are neither recognized as work nor paid (Peters, 2020: 148).
In summary, the practice of reporting exemplifies the non-linear relations in development projects.While Oxfam Novib did not recommend using the same templates for partners, the Ugandan country office still adhered to that practice.Similarly, a number of actors in both country offices were convinced that it was Oxfam Novib's wish to conclude oneyear contracts with partners, instead of longer term contracts (which was not the case).These practices were based on experience rather than on what had been communicated.In large development projects, the discourses and practices of multiple actors compete and not all stakeholders interpret them in the same way.This can lead to actors personally curtailing flexibility and not making use of the room provided for as the examples above suggest.As practices are routinized, new practices need repetition to stabilize.

VI. Revisiting Directives in the Aid Chain
We find that management directivesregardless of whether they are constraining or providing more flexibility and autonomy-do not necessarily travel across the aid chain.In cases in which they reach the different actors and organizations, they neither result in uniform practices nor have uniform meanings.Being the donor, the Ministry selected thematic areas and determined liability and accountability requirements.While this constrained operations for CSOs in different ways, it is noteworthy that in terms of operational requirements, such as MEL and reporting, Oxfam Novib and the country offices had significant control over procedures and practices.This also meant that some of the flexibilities provided for in the policy were not harnessed.Overall, requirements were not equally distributed and their distribution was subject to change over time while organizanegotiated at the different interfaces.It is not our intention to claim that management directives do not matter; we are convinced that they matter just as much as organizations' relationships, histories and institutional exigencies.In line with Beck (2017: 6), we argue that development projects are not 'pre-packaged products' but can rather be characterized as an 'ongoing series of interactions'.While policies suggest a certain level of control and direction, this turns out to be a mirage in practice.Consequently, policies should be read as documents of practice and active producers of reality, rather than a reproduction of reality (Höhn, 2013: 112;Reckwitz, 2016: 256).
In light of our case study, we contend that the concept of the aid chain might be analytically unhelpful and normatively undesirable as it perpetuates the image of a stable system that risks victimizing organizations receiving donor funding.In particular, its rational choice underpinnings grossly simplify project relations by reducing the analytical focus to dyadic transactions, contracts and information asymmetries.As Peters has aptly shown in her analysis of a development project in Angola, principal-agent thinking obscures 'the nature of the social and relational work conducted by development agents at all levels' (Peters, 2020: 153).The concept of the aid chain takes policy at face value by assuming that actors' behaviour is solely shaped by directives that trickle down.Focusing on actual practices, like we have done in this article, is one way of highlighting the complexity of project relations, showing that development interventions consist of a multiplicity of interfaces and not only between a dominant donor and submissive implementing partners.
Aid chain thinking overestimates the practical effects of single projects, whereas realities for organizations are much more diverse.Practices are determined by organizations' engagements in various horizontal and vertical relations due to the multiple funding sources they rely on: Development projects take place in organizational systems that almost resemble forms of multilevel governance, in which many organizations and organizational levels are interwoven and mutually contingent.Within these, different forms of power flow in many different directions, gaining strength and legitimacy from diverse sources, with meaning continuously negotiated and translated among and between the different levels.(Fejerskov, 2018: 140) In addition to highlighting the contingency of organizational relations and the multiple avenues through which power flows, the article also illustrates how management frameworks are transformed by those working with them.In addition, we showcase that interfaces are crucial sites at which power is negotiated.Due to an overt focus on power relations between international donors and implementing partners in the literature, ironically, many of the existing inequalities within the project implementation system are concealed. 7As southern CSOs act as sub-granters and are made fully liable for losses, they carefully evaluate how much autonomy they can allow their sub-grantees.Thus, funding relations between southern partners are also characterized by efforts to striking the delicate balance between partnership and donor roles.In addition, organizational procedures within southern organizations have in several instances prevented the uptake of more flexible practices that the donor allowed for.This had repercussions for other organizations, for instance, with regard to their contract duration and reporting duties.Southern organizations are just as much as northern organizations caught up in their own organizational bureaucracies and need to streamline the institutional exigencies of different donor organizations to work efficiently.

VII. Conclusion
In this article, we have analysed everyday project management practices in two development interventions, identifying how donor directives are implemented in practice throughout the aid chain.By making use of practice theory, we examined how practices emerge and are negotiated in interactions between organizations and actors at the different interfaces in a development project.We discovered that different discourses and rationales compete at the interface, giving rise to a variety of practices both between and within organizations.Project management directives neither resulted in uniform practices nor had the same meanings for different actors.This led us to the conclusion that the metaphor of the aid chain is analytically unhelpful in examining aid interventions.Neither requirements, results nor funding move in a linear fashion across organizations as the concept of the aid chain suggests.
We thereby position ourselves against the body of literature that depicts southern organizations as passive receivers in aid interventions and mere recipients of funding and requirements.Instead of conceptualizing southern CSOs as being outside the aid system, we show that they are an integral part of the system: the management of aid happens at multiple interfaces, by no means excluding southern organizations.Southern CSOs actively shape project implementation and navigate the constraints they are faced with.By rejecting the metaphor of the aid chain, we by no means intend to negate power relationships in aid interventions.On the contrary, our analysis has shown that CSOs are faced and have to deal with a number of conditions (such as a fixed budget, a thematic focus and monitoring and evaluation (M&E) systems that are not their own).However, we also found that these conditions are not constraining for all actors in the same way, just as the freedoms provided for are not beneficial for all actors alike.We, therefore, advocate for a more nuanced analysis of relationships in development projects and especially for an analysis of everyday practices that glimpses behind written policy models.The approach of 'studying through' can help to dissect the workings of policy processes, illuminating the messiness, contradictions and unintended consequences of policies in practice.
As has been found elsewhere (Gutheil, 2020), management frameworks do not work as universally effective policy levers-their practical implementation can only be understood from studying the context, relationships and organization of the project's interfaces.This has crucial implications not only for development research but also for development programming.The recent initiatives introducing more adaptive ways of project implementation were introduced to counter some of the unintended effects of donors' desire for control, which is limiting operational space for implementing organizations (McCulloch and Piron, 2019).While in theory, adaptive management might address these concerns by shifting power to the implementing organizations, it is a fallacy to believe that this can be achieved by introducing such models by simply funnelling directives into the lead organization and the respective practices are expected to follow across all organizations.Practices do not link at will with other