Doing ‘Us-Them’ differently: the identity work of frontline aid bureaucrats in translating aid effectiveness policy rhetoric into practice

ABSTRACT Bilateral aid agencies often face implementation challenges in internal efforts to address long-standing aid fragmentation and effectiveness issues. This article introduces the organizational identity concept to understand better these challenges by examining how Swedish International Development Cooperation Agency (Sida) frontline staff understand their role and organizational goals in light of shifting demands to coordinate and align Swedish government agencies’ (SGA) aid engagements. SGAs implement 10-15% of Swedish bilateral aid annually. A recent government strategy prioritizes strengthening partner countries’ public institutions and partnerships in line with the Agenda 2030 for sustainable development. Analysis of interviews and focus group data reveals a general shift among bureaucrats beyond the traditional us-them funder identity, to embrace a range of other identity orientations in the Sida-SGA relationship. The various orientations reflected Sida frontline bureaucrats’ diverse interpretations of their individual authority and socialized sense-making of ambivalent organizational changes, as they grapple with questions of ‘Who we should be?’ and ‘What we should do?’ on the frontline. The study provides a fine-grained view of the essential attitudes, skills and behavior on the frontline that influence aid relationships and the implementation of aid effectiveness principles, adding nuance to the existing aid effectiveness literature.


Introduction
Although the volume of aid continues to increase to support global poverty reduction and sustainable development efforts, aid fragmentation and effectiveness remain significant concerns for many bilateral aid agencies.Following the 2005 Paris Declaration on Aid Effectiveness and subsequent commitments (OECD 2022), aid agencies have made internal changes to improve the quality of aid and its development impact.These changes include, among others, letting frontline staff make their judgment to work contingently and flexibly, and explicitly linking these changes to more significant aid impact on partner countries (Honig and Gulrajani 2018).However, there is growing recognition that the implementation of aid effectiveness's coordination, alignment, and harmonization principles is insufficient to overcome the negative side effects of uncoordinated and fragmented aid (Leiderer 2015;Lundsgaarde and Engberg-Pedersen 2019).Translating rhetoric into practice requires changing the roles, power relations, attitudes and behavior of individual professionals and agencies involved (Groves and Hinton 2004;Eyben 2014).
Research suggests that bilateral aid agencies face more internal challenges than expected in their efforts to deliver on their aid effectiveness commitments.Bilateral aid agencies, typically federal ministries, government offices, departments, and agencies, provide direct development assistance to a low-income partner countries through grants, loans, in-kind services or expertise, to other governments, civil society, and multilaterals (viz., United Nations and international organizations).While bilateral aid offers donor countries advantages such as more control over aid funds, promoting their strategic interests, and developing longterm relationships with specific partner countries, these advantages also make bilateral aid more susceptible to the political influence of the donor nation, more politicized (e.g.tied to specific conditions), and even less cost-efficient than multilateral aid (Gulrajani 2015(Gulrajani , 2017)).It is, therefore, of no surprise that the aid effectiveness literature presents opposing views of bilateral aid agencies' performance.Recent studies suggest that most bilateral aid agencies still need to fulfill their pledged standards of good practices and have performed poorly on several aid effectiveness metrics (Palagashvili andWilliamson 2020, 2021).This has led some to suggest that there is growing aid effectiveness 'fatigue' (Leiderer 2015), or 'lost momentum' of aid effectiveness commitments (Lundsgaarde and Engberg-Pedersen 2019).These arguments highlight the complexity of implementing aid effectiveness and the importance of attending to the diverse relational settings in which aid effectiveness principles are implemented (Lundsgaarde and Engberg-Pedersen 2019).Given that bilateral aid constitutes 70% of today's total official development assistance (179 billion USD) for the majority of OECD Development Assistance Committee (DAC) donor countries, understanding bilateral aid agencies' inner workings behind the gaps between policy commitments and practical realities of implementing aid effectiveness principles, and how they balance prevailing tensions, is critically needed.
The aid effectiveness literature has primarily focussed on policy design and structural factors that influence aid agencies' implementation of aid effectiveness principles.However, previous research has emphasized the discretionary power of frontline aid bureaucrats in shaping aid relationship and aid effectiveness (e.g.Brinkerhoff 2002;Eyben 2006).In this article, frontline aid bureaucrats are public employees who are the primary implementers and interpreters of bilateral aid policies at the frontline and have day-to-day contact with development brokers or aid agents (i.e.organizational actors associated with implementing bilateral aid) and people affected by their decision (Falanga 2019).However, research has yet to provide a fine-grained understanding of how frontline aid bureaucrats translate aid effectiveness policy commitments into new roles and new aid relationship while balancing prevailing tensions or organizational conflicts.
To address this knowledge gap, this article responds to repeated calls (e.g.Hupe and Hill 2007;Honig 2014;Gulrajani 2015) to incorporate organizational theory by utilizing he concept of organizational identity (OI).OI has rarely been used in the aid relationship or aid effectiveness literature.OI offers a valuable perspective for examining how organizational members collectively comprehend their roles and interpret organization goals and objectives (Fiol 1991;Ellis and Ybema 2010).By analysing how aid bureaucrats construct their identity when implementing policy commitments in a given aid relationship, we can gain a better understanding of the organizational tensions that arise, how aid agencies and their members balance these tensions, and the values that are important in the aid relationship and aid effectiveness principles.
To answer the research question, the study conducted a single case study that examines the identity work of the Swedish International Development Cooperation Agency (Sida) frontline staff.The study aims to understand how they position themselves to meet new aid effectiveness policy demands to more effectively coordinate and align the aid engagements of Swedish government agencies (SGA).Sida works with embassies, non-profits, and government agencies to execute Sweden's bilateral aid strategies.Despite its highly decentralized aid management model granting high-level program and financial authority to its frontline staff in headquarters and embassies, Sida still had areas needing improvement, such as engagement in fragile states, alignment between Swedish development actors, and Sida staff's capacity in adaptive programming, collaborative performance/ results, and knowledge-management (OECD 2019).The new organizational demands were a result of a recent government strategy to strengthen partner countries' public institutions and partnerships, as implied in the Agenda 2030 for sustainable development.A pool of 20-25 SGAs (out of 340; Statskontoret 2023) act as development intermediaries or aid agents, implementing about 10-15% (around 600 million SEK or 57 million USD) of Sweden's total aid each year.The new strategy explicitly specified Sida's formal responsibility for implementing the strategy and creating conditions for more effective aid engagement and effectiveness of SGAs.SGAs' aid engagements could be traced back to three decades ago yet, along with the Sida-SGA aid relationship, has been underexplored in academic research.
This study analysed the OI narratives collected from key informant interviews and focus group discussions (FGD) using reflexive interpretation techniques guided by Why, What if, So What questions (Gabriel 2018).The analytical approach enabled a better understanding of the link between OI, aid relationship, and aid effectiveness by attending to the significant details of the identity narratives in the situated organizational environments.The study contributes to the knowledge on aid effectiveness by demonstrating the diversity of interpretation and construction of identity among frontline aid bureaucrats in translating ambivalent organizational demands to implement aid effectiveness principles.The next section outlines the theoretical approach, followed by the study's materials and methods, results, and finally, the article's contributions and implications are discussed.
Theoretical background Albert and Whetten (1985) defined organizational identity (OI) as the central, distinctive and enduring features that organization members collectively share.Since then, the concept has evolved in many directions across the fields of organization, public administration and international organizations, but has yet to bridge with aid relationship or effectiveness literature.OI research has suggested that identity work will intensify when there are conflicts or misunderstandings around roles, and identity narratives are neither static nor completed as multiple versions are constantly being reconstructed, negotiated, and re-imagined by different actors (Caza, Vough, and Puranik 2018, 898).This article adopts a pragmatist philosophical perspective for a richer and more realistic view of such multi-faceted human behavior across multiple levels of analysis.It emphasizes a more comprehensive account of the structural, historical and relational contexts in which contemporary public organizations operate (Farjoun, Ansell, and Boin 2015).
This pragmatist perspective assumes that public organizations have to balance continuity and adaptability, acknowledging that stability and change of OI coexist (Albert and Whetten 1985;Gioia, Schultz, and Corley 2000;Golant et al. 2015).Therefore, a dynamic view of OI with an adaptive potential is used in the study to entail 'who we should be as an organisation' (Gioia, Schultz, and Corley 2000) and 'what members perceive, feel and think about their organisation' (Hatch and Schultz 1997, 357).In short, OI is the outcome of ongoing collective sense-making (Fiol 1991) by which members share, discuss and negotiate their perceptions of what their organization is or should be.In the study context, this means that the identity work of aid bureaucrats is highly dependent on what has been socialized into them internally and externally.This also presents frontline bureaucrats as social constructors, and OIs as products of such social constructions, produced in interactions (Czarniawska-Joerges 1994, 196) within and outside the organization.
The literature has indicated that sense-making is often 'grounded on constant struggles to construct identities' (Gabriel 2018, 263).In this sense, OI (re)construction can be understood as triggered by events that disrupt members' or stakeholders' expectations (Weick 2001), such as changes in the internal and external environment (Ravasi and Schultz 2006), external shocks or critiques (Breit 2014).This phenomenon is similar to what Ramalingam (2013) has described as 'aid on the edge of chaos', referring to the current aid context of rising complexities needing urgent reorganization, and resilient partnerships with local actors to adapt to changing needs.In such challenging times, members of aid organizations may reconstruct OI to address prevailing ambiguity, dilemmas and conflicting demands.Scholars have suggested that members may generate multiple OI accounts, drawing on different resources (e.g.knowledge, skills, expertise and influence) as conscious reciprocal interactions from within or outside the organization (Weick, Sutcliffe, and Obstfeld 2005;Cornelissen 2012).In decentralized contexts with high organizational autonomy and frontline authority, the nature of resource mobilization would look quite different.Therefore, according to Scott and Lane (2000, 43), it is crucial to recognize OI as negotiated cognitive references or by-products emerging from complex, dynamic, and reciprocal social relations that are embedded within different sensemaking (sub)systems.
Research suggests that resources are likely to be internally influenced by references to the organization's history and values to connect past, present and future (Ravasi and Schultz 2006).For instance, OI may be reconstructed from evoking memories of the past (Schultz and Hernes 2013), or enforced through leaders' messaging with historical referencing (Golant et al. 2015).However, unmanaged, fragmented sensemaking (Maitlis 2005) and competing OIs (Pratt and Foreman 2000) could result in organizations being pulled apart (Kreiner et al. 2015).As mitigating fragmentation is one pressing concern in aid effectiveness reform agenda, understanding the intra-organizational sense-making dynamics and the interactions embedded in different (sub)systems of the organizational membership is vital.
In partnership-based inter-organizational contexts, resource mobilization is likely influenced by members' use of different cognitive references (mental frameworks to understand and organize information, to justify their sense-making of self and identity) (Thornton, Ocasio, and Lounsbury 2012, 3).For example, Rondeaux (2006) has suggested that bureaucrats combine a public service reference and a public managerialism reference to describe the identity of their partnershipbased organization.Others have further suggested that cognitive references support members' OI generation and negotiation with stakeholders in expanding available resources within socially acceptable boundaries (Kraatz and Block 2008;Kraatz et al. 2016;Reissner 2019).Reissner (2019) even highlights that there may be tensions and ambiguity in which references are considered essential or socially accepted.Therefore, understanding how these tensions and ambiguity affect the way aid bureaucrats construct their OI in frontline interactions with aid agents could provide a strong foundation to explore the key elements of OI that support or undermine the implementation of aid effectiveness policy commitments.
In inter-organizational literature, the us-them divide is central to the conceptualization of OI as negotiated cognitive images or by-products of asymmetric power relations.Ellis and Ybema (2010) argue that this aspect creates uncharted territory for scholarly investigation of what the organization and its members do and value, not just in policy design but also in frontline reality.This aspect is highly relevant to the study of frontline bureaucrats' identity work in implementing aid effectiveness commitments within a specific aid relationship.For instance, Ybema, Vroemisse, and van Marrewijk (2012) present a nuanced us-them OI account of an aid-funded Dutch human rights organization.They show that staff members deconstructed differences between themselves and their partners for egalitarian partnership-building purposes by actively seeking to 'smooth out, trivialise or upend differences' through presenting themselves as 'strange' and others as 'normal'.The intention was to level out power differentials typical in aid relationships and constructing an inclusive 'we' in relationship-building (Ybema, Vroemisse, and van Marrewijk 2012, 48).This example resonates with Elias and Scotson's (1994) notion that social interactions involve inclusion and exclusion dynamics.It is crucial to understand who is 'in' and 'out', who 'we' are and who 'they' are in the us-them divide.This aspect also highlights the importance of sensitizing research to the fundamental heterogeneity and situatedness of OI in the interplay of power, bureaucratic procedures, and relationships.In the intra-and interorganizational context of the case study, unveiling this us-them aspect requires multi-level analysis of how individuals-and organizationally shared OI shape aid relationships as part of ongoing organizational changes.

Materials and methods
A single case study was conducted between December 2018 and December 2019 to answer the research question and to reveal distinct, previously inaccessible theoretical and empirical insights (Yin 2009).Sweden was selected for its theoretical usefulness as a bilateral donor country that has persistently affirmed its aid effectiveness commitments and entrusted high-level authority to frontline aid bureaucrats of its executing agency, Sida.About five months prior to this study, the Swedish government launched a new 2018-2022 aid strategy, 'Strategy for capacity development, partnership and methods that support the 2030 Agenda for sustainable development' (hereafter referred to as the 'Strategy') (GOS 2018), which explicitly specified Sida's formal responsibility to implement the Strategy and to better support other Swedish government agencies' (SGA) aid engagements and effectiveness.Approximately 10%-15% of Sweden's official development assistance budget is disbursed annually to a network of 20-25 SGAs to strengthen public institutions in partner countries for achieving poverty reduction and sustainable development.At the time of the study, Sweden had bilateral development cooperation with some 35 countries in Africa, Asia, Latin America and Europe.
Access to the case was gained through a grant funded by the Strategy in a research project aimed at exploring how SGAs experience the changing bilateral aid contexts.To address potential bias resulting from the research funding source and the author's prior professional aid experience, a pilot study workshop was conducted in Stockholm on 5 December 2018 with a core group consisting of SGAs' aid project managers and Sida representatives.The group members regularly met to discuss common issues in the bilateral aid contexts.Twelve representatives from the group attended the workshop, which presented the research purpose and its independent nature.The workshop also provided a contextualization of the current Sida-SGA collaborative relationship and challenges, and identified three SGAs for an in-depth examination of Sida's frontline identity work in relation to them.
Semi-structured interviews were conducted as the primary means of data collection to elicit in-depth qualitative OI accounts from Sida frontline bureaucrats.During the study, there were 29 focal points in Sida headquarters who each managed one or more SGA relationships.To answer the research question, the researcher purposively selected 12 samples based on their frontline interactions with three focal SGAs on strategic or operational levels.The three focal SGAs (given pseudo names of Alpha, Beta, Gamma in randomized order) represented the public administration of environment, land, and statistics.Sida's frontline interactions with them were selected because the three SGAs had the most extensive bilateral aid history, programmatic and geographical spread, offering a situated perspective over time and change contexts.In 2018, the disbursement from Sida to these three SGAs together constituted 24% of the total disbursed to all aid-involved SGAs.The 12 informants were interviewed between February and May 2019, nine from three Sida departments at the headquarters and three from embassies based in three different continents with the most intensive SGA interactions (see Appendix A for informant details).Pseudonyms were used to protect their identity, and only in the article where they would not be identifiable.This qualitative case research aims to provide an in-depth and contextualized understanding of the phenomenon (i.e.organizational identity) (Alvesson and Sandberg, 2011).Information saturation was considered reached following the 12 interviews, as no new relevant themes emerged to answer the research question (Guest, Bunce, and Johnson 2006;Small 2009).All informants provided written informed consent to participate in the study, and all interviews were recorded.A similar set of broadly framed questions guided the interviews, covering the topics such as the informant's background, their role and SGA interactions, tensions experienced in organizational changes, and opportunities and concerns to promote SGAs' aid engagements and effectiveness.
In addition to the interviews, observations from two FGDs in Stockholm provided complementary data.The first FGD on 30 September 2019 with 17 participants representing six SGAs, including Sida, validated the preliminary findings around challenges and concerns in the changing context.The FGD also identified priority areas of improvement, coordination and foreseeable challenges moving forward.The second FGD on 16 December 2019 with also 17 participants representing seven SGAs, including Sida, mapped out in more detail the policy, procedures, and practices in Sida's and SGAs' coordination mechanisms that influence their interactions and aid effectiveness.The FGDs generated further insights into the us-them divide and how collective identity in the ongoing organizational changes was referenced, negotiated and reconciled in a relatively power-neutral setting.
This study conducted multi-level analyses of the collected data in three main reflexive steps.This reflexive approach to the data analysis went beyond pattern recognition to gain a deeper understanding grounded in the data and account for the social, cultural, and historical particularities, complexities and contradictions of the case (Alvesson 2011;Miles, Huberman, and Saldaña 2014).First, 'in vivo' coding approach (Strauss and Corbin 1998) was used to stay as close as possible to informants' own words (e.g.'not just a funder') and to identify interesting details of their cognitive references of OI.Second, a reflexive gaze, guided by questions of Why, What if, So What? (Gabriel 2018, 152) directed the interpretation coding techniques.This step complemented the theoretical perspective to attend to the significant details of the identity narratives dynamically interrelated to the organizational environments in which the informants were embedded.For instance, asking Why led to the code of organizational situatedness (e.g.'operationalisation vagueness'), What if led to code of cognitive referencing (e.g.'top-down control' and 'negotiated bottom-up feedback' mechanism), and So What led to the code of socially accepted sense-making choices and the us-them positioning in the change (e.g.'aid effectiveness value-creation aspirations' and 'concerns' of change).This approach enabled theoretical imaginations of the studied phenomena and made the plausible link between identity work at the frontline, the intended relationship and desired outcome of organizational demands.Finally, the interpretations from Step 2 were reconciled with extant literature on frontline bureaucracy and aid effectiveness, and aggregate themes on OI narratives were drawn up.

Results
The study revealed the complexity of managing evolving organizational demands with limited Sida frontline resources in both the headquarters and field offices in the embassies, where the needs of a large number of smaller projects across various implementing SGAs and other organizations had to be met.In the case context, the study identified a general shift from a historically dominant funder identity, which prioritized non-profits, to a multi-dimensional identity, which prioritized SGAs as aid agents.Although the technical details that set the two contexts apart were beyond the scope of this study, the interviews indicated that a series of internal and external events triggered the new organizational demands into the second context.These events included the United Nations' adoption of the Agenda 2030 in 2015 for sustainable development with an explicit dedication of goals to strengthen public institutions (SDG#16) and revitalize global partnerships (SDG#17), the Sida Director-General's emphasis on prioritizing SGAs as key aid agents, and the launch of the Strategy in July 2018, which called for collaborative partnership approaches, engagements in fragile states, and sustainable and synergistic aid effectiveness.

Pre-SGA prioritization context: who we are and what we do
The most prevalent theme in this context was Sida's traditional funder identity which typically featured Sida as a 'financier … We finance other development actors'.Despite SGAs and Sida both being highly autonomous from the 'ministerial rule' (Ahlbäck Öberg and Wockelberg 2021) in the domestic context, regardless of SGAs' aid engagement forms (viz., government mandate, strategic framework or service agreement), the frontline interactions would always 'put Sida or the embassy as the funder'.This meant that the same logic of frontline interactions with non-profits, consulting firms and multilaterals also applied to SGAs, creating bureaucratic competition.
Informants with fund-management responsibilities notably demonstrated this identity.Typical frontline bureaucracy in line with this funder identity would entail 'quality and financial accountability control exercises every year', 'appraising every project that Sida supports, following up on project finances, identified and new risks, practical challenges, reviewing the annual report they submitted, and following up with further discussions with them'.This identity signaled top-down control and a strong us-them divide in frontline interactions, and position Sida above SGAs in the aid's value chain network.As one informant seasoned in various Sida positions justified the importance of keeping the us-them divide, Seven-eight years ago, when there was much focus on internal control within Sida, the approach was that we should be at arm's length from SGAs to control their implementation … This is basically still the model we get.We assess their application to see if they have the capacity to do this and manage funds.(Kristina) Some informants argued the prevalence of this funder identity as permeating some frontline attitude and behavior to-date, reflecting different us-them cognitive referencing sub-systems within Sida.As one headquarters informant with a multilateral background and broad collaborative program portfolios (involving SGAs, universities, non-profits and consulting firms) articulated, The dialogue and attitude here at Sida, not always but often, is rude and arrogant.There is an elite culture here that we Sida in certain ways are better.I've been in meetings where Sida staff are being condescending to staff from other SGAs.It's completely unacceptable … People work in silos and can't think outside the box when the political dynamics are changing.(Erik) The dominance of the funder identity had created some contentious moments in the past, especially where Sida's bureaucratic routines entangled in between local (Sweden) and global (aid) relational contexts were not be well understood or accepted by SGAs.One embassy informant recalled a dramatic event in the past and how it was resolved, We wanted to do an audit review … The general director of involved SGA went berserk, saying 'Sida cannot come and do an audit on us!We are the same state agency bodies'.The people at their international department got upset as well.We managed to convince them to do it and it was quite useful because it helped them their project management procedures in place because that's not always something a state agency has.(Anna) We should note that there had already been customized rather than standardized frontline practice within Sida in the pre-SGA prioritization context.However, one other prevalent feature associated with the funder identity in this context was a shared cognitive image within Sida that working with publicsector (and SGAs, by association) was 'not glamourous work … with people fighting over it', or 'sexy subject like civil-society cooperation', and ended up as 'an add-on' for bureaucrats who might or might not have relevant public administration background or motivation to engage SGAs meaningfully.While informants generally recognized that some SGAs were doing a 'fantastic work' that Sida 'could never do', others acknowledged the 'political pressures to engage SGAs' even though SGAs might not be 'the most cost-effective' aid agents compared to multilaterals or consulting firms, or 'have the capacities', 'linguistic skills', or 'context-relevant expertise' to scale up to Sweden's 35 partner countries, fragile states with weak public institutions in particular.
These features before the SGA prioritization context constituted the basis of potential tensions and us-them orientations with varying control-trust interactional dynamics in the SGA prioritization context.

SGA prioritization context: who we should be and what we should do
Sida had introduced numerous internal organizational changes into the SGA prioritization context.We should note that some of these changes were universal changes in Sida's frontline interactions with all aid agents, not necessarily specific to SGAs' aid engagements.One notable change was a strengthened focalpoint structure comprising one overall coordinator and 29 SGA-specific focal points across Sida's departments.Focal points served to 'answer any questions from any SGAs who wanted to engage with Sida', as 'entry points to Sida' and find 'a good fit' from the myriad of 49 thematic, global, regional and country Swedish aid strategies.Frontline interactions were also strengthened on the senior management level (through the annual SGAs' forum myndighetsforum) and, initiated on the director-general level, to draw on Sida's global experience to support SGA's national implementation of the SDGs.Although not exclusively for the SGAs, crossdepartmental reference groups were also strengthened to appraise and review all bilateral aid projects to break Sida's organizational 'silos'.Additionally, Sida's training and meeting center in Härnösand, some 427 km north of Stockholm, had started adapting their training support to the changing needs of SGAs.
These initiatives were widely recognized by informants as essential to SGAs' aid engagements and effectiveness.While an explicit call for a strategic identity change was not evident in this study, there was a clear general shift beyond the mere funder identity in terms of what informants felt about who Sida should be and what Sida should do in the prioritization context.
The most prevalent theme was the dialogue-partner identity.One informant articulated what this meant for the frontline, 'when SGAs come to us sometimes to ask for help, to get a point across, if we agree or not.If we do, we help them push that point on every level we have access to' (Jan).For others, this would be like a critical friend giving feedback or shedding light on some blind spots, for example, on a new project idea in a partner country with some known high risks or difficulties, by asking 'why, who would be the contact points [in partner country], what would be the purpose' (Maria).As some informants described, this was to add a sense of realism as SGAs were generally not 'development cooperation experts' or that they might be 'too invested in their baby and may need to know when to pull the plug'.It was clear that within Sida there was 'no set principle or model for how that dialogue works, because you can't box development or a reality into a simple format'.This reference seemingly signaled the importance of a reciprocal, flexible or even bottom-up feedback mechanism, and suggested that bureaucrats had a high-level authority to reflect their cognitive images of frontline interactions.
In line with the dialogue-partner identity, several identity features were prevalent, but not dismissing the funder identity.These included problem-fixer, knowledge-broker, true-partner and team-player orientations.Problem-fixer tended to feature frontline interactions relatively proximate to the cognitive image of the funder identity, aimed at, for example, as one embassy informant described, 'trying to solve problems for existing projects, keep them running smoothly and accomplish what's intended, and that the [tax] money is being well used, free from corruption scandals' (Johan).Knowledge-broker¸true-partner and team-player orientations, on the other hand, would feature interactions relatively distant from the cognitive image of a funder.As one embassy informant seasoned in various Sida positions in the headquarters and multiple embassies articulated, We have hired SGAs to do that job and need to think of them as true partners and we are players in the same team.By playing our different roles, we are more likely to reach the results we all want.Development cooperation is not SGAs core competency.But we can pool our knowledge and sharing our development cooperation competence.When we adopt an 'us and you' position or hold up information and knowledge from them, we are in trouble.(Birgitta) These varied us-them identity orientations reflected bureaucrats' heterogeneous interpretations and enactment of their frontline authority.There would always be 'a fine line to cross' given the 'administrative and legal barriers' embedded in the funder identity in the existing aid administration system, which were not always 'compatible' (Lena) with these new organizational demands and SGAs' expectations in frontline interactions.
Although the notion, as articulated in this quote from a headquarters informant, 'in bureaucracies, we are all individuals' (Anna) showed up repeatedly in all the interviews and the group discussions, the study's reflexive interpretative analysis suggested that there were more dimensions than to attributing the bureaucrats' interpretation and enactment of a specific identity orientation to frontline authority alone.The most prevalent theme was the aspirations and values being prized in the internal and external environments in which the bureaucrats made and exchanged senses about their role and SGAs' aid engagements in the organizational change contexts.
Tracing the source of the us-them divide revealed two sense-exchanging categories underpinning the negotiated, collective and reciprocal nature of the sensemaking process in the case's organizational cultural context.The first category primarily sought to exchange cognitive references to improve coordination and complementarity of SGAs' aid engagements with other aid actors in Sweden and partner countries.As this embassy informant explained why one should reject the notion of Sida just being a funder, The agencies are very good, specifically with the EU integration projects.But they have varying degrees of capacity to implement poverty reduction and organisational change projects … Pushing them to fragile states is insane.It can't just be solely up to them to implement whatever they see fit.We have, therefore, different roles and different responsibilities.(Kristina) The second category primarily sought to exchange cognitive references to balance the sometimes conflicting demands and trust-control dynamics between the funder and dialogue-partner identities in frontline practice.As this headquarters informant with decades of private and public-sector experience and global aid portfolios described how the identity dynamics should be managed and guided by aspirations for SGAs' aid engagements and effectiveness potentials for SDGs, In practice, we need to build a bulk of trust with them to look at new ways of doing things, and to take advantage of Sida field offices to build partnership of mutual benefit … This is a balance, of course, since we have a rigid kind of control and follow-up of taxpayers' money.It demands getting to know the agency you work with, their project details and progress … contribute with the flexibility more as a team member … share context-specific knowledge to help them stand on their own and build local partnerships in the long run independent of aid to build real sustainability.We have so much to gain as a public sector in the North from contexts different from ours.It's ideological but I think this is a very good way to go.(Anders) In the same vein, trust-building emerged as a crucial feature in the dialogue-partner identity for making and exchanging senses of cognitive references within Sida and at the frontline with SGAs.As airing dirty laundry implied in this quote from a seasoned embassy informant, 'everybody is reluctant to talk about intimate problems off the bat, or, admit something less than successful in print.You have to win the confidence' (Johan).
Moreover, coordination of sense-making and exchanges also emerged as another essential feature in the dialogue-partner identity.Despite the focalpoint coordination structure, many informants would still use 'ambiguous', 'tricky', and 'ad hoc' to describe the lack of 'institutionalized' information exchange or coordination within Sida to strengthen synergies between different thematic areas of SGAs' aid engagements.Several informants referenced recent examples where non-coordination of a dialogue event by the headquarters ended up undermining a similar effort by an embassy.Specifically, the lack of institutionalized coordination within tended to undermine the opportunity for junior bureaucrats to mobilize authority or trust to deviate too much from the funder identity and associated bureaucratic features.

Discussion and conclusion
This article has responded to the calls to attend to the discretionary power of frontline bureaucrats and engage with organizational theory to better understand the implementation challenges of aid effectiveness within bilateral aid agencies.Drawing on the OI theoretical perspective, this study explored how bilateral aid agencies' frontline bureaucrats construct identity in response to shifting yet ambiguous organizational demands to implement aid effectiveness principles.The central revelation of the empirical case study was the multi-identity narratives and the complexity of implementing aid effectiveness principles on the frontline.Bureaucrats made heterogeneous interpretations of their frontline authority to express who they should be and what they should do when faced with organizational conflicts.On the one hand, the organizational conflicts were derived from bureaucratic competition (Kilby 2011) for a greater share of frontline resources to help coordinate and align a large number of smallscale aid projects, or initiate new ones in unchartered contexts.On the other hand, the conflicts were a result of tensions between demands for flexible relationships and rigid bureaucratic procedures.In light of the organizational conflicts and tensions, frontline bureaucrats might draw cognitive references through social exchanges with internal and external organizational stakeholders to make sense of the value-creation aspirations and concerns of the shifting organizational demands.These findings align with the public administration literature that suggests frontline bureaucrats in contemporary democratic governance systems are not mere executors of public policies, but they can navigate back-office and frontline functions by creating an extended social network of references for decisionmaking (Falanga 2019).This study provides a situated understanding of how frontline bureaucrats balance control-trust dynamics in their identity work to engage in stakeholder relations, draw cognitive references to legitimate their interpretations of their authority, roles, and forms of frontline interactions.This study demonstrates that a trust-control balancing act on the frontline can create opportunities for exploring how shifting between multi-identity narratives may enable flexibility, ownership and collaboration with aid agents.In contrast, solely on controlling risks associated with new organizational demands, given high agency fragmentation, may enforce historically dominant identity, perpetuating asymmetric power relations and bureaucratic competition.

Theoretical contribution
The Paris Declaration on Aid Effectiveness and subsequent agenda emphasize inclusive partnership, ownership and collaboration as organizational principles to avert power inequality and hegemony in aid relationships.By focusing on the identity work of bilateral aid agencies' frontline bureaucrats and their relationship with development intermediaries, the study contributes to the literature in several main ways.
First, it enriches our understanding of the link between organizational identity, aid relationship, and aid effectiveness by providing fine-grained insights into the complexity of implementing aid effectiveness principles in a decentralized yet non-coordinating context within a bilateral aid agency.The study revealed that multi-identity narratives co-exist in the social and cognitive sub-systems of a given aid relationship.The socialization of individual frontline bureaucrats amplifies the legitimacy of specific aid effectiveness values and aspirations to make sense of the shifting organizational demands (i.e.prioritization the support to SGAs as bilateral aid agents).This influenced how they reinterpreted shifting their role, discretionary power, organizational routines, and aid relationship.
Frontline bureaucrats could reference a mere policyimplementer but interpreters or even policy-entrepreneur (someone who uses their influence, knowledge, and skills to advance the policy) (Mintrom and Norman 2009), based on their sense-exchanges.Frontline bureaucrats' bottom-up referencing opens up opportunities to gather local knowledge (e.g. from aid agents and their local partners) and customize routines or procedures to lessen the effect of bureaucratic impediments (Bovens and Zouridis 2002), enhance reciprocity and aid effectiveness policy implementation success.
Second, the study extends the literature on frontline bureaucracy by providing rich insights into the discretionary power of frontline bureaucrats in a bilateral aid agency when faced with organizational conflicts and bureaucratic competition.The study showed that bureaucrats' situated and perceived authority could influence cognitive references and the enactment of specific sense-exchanging references in the identity work.In the case context, for example, embassy bureaucrats had distinctively different authority from those at the headquarters, whereas bureaucrats with specific sectoral or professional proximity to the aid agents would gravitate toward a specify type of identity work in their sense-exchanging process.
Third, the study enriches knowledge of organizational identity in development studies literature by showing how frontline aid bureaucrats combine different identity orientations while preserving their funder identity to navigate conflicting organizational demands between stability and change.The study identified an emerging and co-existing dialogue-partner identity, featuring problem-fixer, knowledge-broker, true-partner and teamplayer characteristics in frontline interactions.The study also found that each identity orientation could allow us-them dynamic to play out differently in relation to the aid agents (de)prioritized, in/excluded in the shifting organizational demands.The study highlights the usefulness of the OI theoretical perspective in examining the actual power distribution implicated in closing the aid effectiveness policy rhetoric and practice gaps.
Finally, one essential element in the current development cooperation context is Agenda 2030 for sustainable development which necessitates development actors orient their roles and actions.This study provides early lessons about the potential contradictions between the different dimensions of the Agenda to carry forward.

Practical implications
This OI study suggests that implementing aid effectiveness principles require aid agencies to focus on the development of essential social skills among frontline bureaucrats.These skills include the ability to shift or combine different mind-sets, the capacity to mobilize relevant information and knowledge, and exert influence to guide decision-making, and the ability to balance trustcontrol dynamics in various relationship-building and dialogue forms with aid agents (and local stakeholders) affected by the changes.Furthermore, such skills would require not just extensive knowledge of the inner workings of the aid organizing systems, procedures and routines, but more importantly, the willingness to go outside the box and across organizational and professional boundaries to adjust roles and aid relationships to the changing institutional demands, stakeholder expectations, and needs.To improve frontline bureaucrats' social skills, aid agencies must institutionalize internal and external knowledge-exchange as a fundamental frontline function and skill.This institutionalization of OI work would support the acquisition of the essential skills necessary to make good, rather than make do with politically driven aid effectiveness policy commitments in various aid relationship contexts.

Future research directions
Future research can expand on the study's findings and address the limitations of a single case.Quantitative research with larger samples in other OECD donor countries, comparting the aid agencies' OI narratives in relation to other types of development intermediaries (e.g.non-profits, universities, consulting firms) would help to better understand the generalizability of the findings.Additionally, future research could explore the trust and control dynamics and trade-offs, and the resulting aid effectiveness outcomes implied in different identity orientations or combinations.As Lotta et al. (2022) point out, the nature of frontline bureaucracy may differ in the developing world.Therefore, future research could investigate how aid agents construct their identity in increasingly collaborative relationships at various levels of the aid value chain, including upstream, downstream, in the new context of the Agenda 2030.

Funding
The research received funding from Swedish International Development Cooperation Agency (Sida) under contributio no.12848 and D.nr.2017-10917.