Donors dealing with ‘aid effectiveness’ inconsistencies: national staff in foreign aid agencies in Tanzania

ABSTRACT In the name of ‘aid effectiveness,’ public foreign aid is meant to be an equal partnership between donors and recipients of aid, while at the same time proving its efficiency to taxpayers in donor countries. Moreover, as state institutions, public aid agencies are required to follow their own bureaucratic regulations, and increasingly so also those of their partner institutions, while simultaneously managing aid in the most cost-efficient way. This article turns the spotlight on a category of aid workers who help foreign aid agencies manoeuvre through these conflicting objectives: the desk officers employed locally by donor agencies in aid-recipient countries. The article centres on Tanzania, a country at the forefront of the aid effectiveness agenda, illustrating well the tensions it embodies. Tanzanian desk officers advance donor conditionality and circumvent heavy bureaucratic regulation by tapping into their resources as locals. Such resources involve their identity as citizens with a right to hold the Tanzanian government accountable for how it spends development money. They also involve desk officers’ personal networks in the Tanzanian development industry, which help agencies expedite aid interventions – a resource important enough to be assessed by some foreign managers in the recruitment of national staff.

The past two decades have seen major changes within the institutional practice of development cooperation, largely converging under the banner of 'aid effectiveness'. 1 A key component of the aid effectiveness agenda concerns promoting national ownership and development partnership. It involves efforts to transfer the power over aid-supported development onto the state leaderships of aid-recipient countries, and aligning aid interventions with the priorities, regulations and systems of recipient governments. At the same time, foreign aid agencies are increasingly called upon to measure and show the results of aid vis-à-vis home audiences in donor countries, in order to legitimise the spending of tax revenue on overseas development assistance. 2 Aid priorities are to a growing extent influenced by public opinion in donor countries, and by rising demands on the efficiency and transparency of foreign assistance in terms of ensuring that aid reaches its beneficiaries at minimum transaction cost in time and money. Such donor expectations have been criticised as unrealistic, both in terms of the time it should take to effect anticipated development, and with regards to proving causality between money spent and results obtained. 3 How do foreign agencies work towards these seemingly contradictory objectives? This article proposes one way in which they do so, which centres on the everyday work of locally recruited desk officers in public foreign agencies in aid-recipient countries. It takes Tanzania as a case example, a country which has been at the forefront of the aid effectiveness agenda and which illustrates well the tensions inherent in the parallel tenets of recipient countries' ownership of development, on one hand, and demands for certain kinds of measurable development results, on the other. This tension, I will show, relates to what Richard Rottenburg highlights as two sets of inconsistent objectives towards which public aid agencies are mandated to work: progress as well as emancipation, and effectiveness as well as legitimacy. 4 I illustrate how Tanzanian desk officers help their agencies pursue these objectives by tapping into two kinds of resources as 'locals.' One pertains to their identity as citizens and taxpayers in the aid-recipient country. The other concerns their personal networks in Tanzania's development industry, a resource important enough to figure in the assessment of candidates during recruitment of national staff. Foreign aid agencies rationalise the use of informal, personal networks that circumvent their partner country's formal public procedures, by appealing to the dysfunction and inefficiency of the Tanzanian state apparatus. Such appeals echo a common criticism of African bureaucracies as disorganised, slow, and corrupt. Meanwhile, they merit closer examination in view of existing research highlighting how civil servants in aid agencies and state bureaucracies in donor countries use personal networks to bypass their own administrative rules and systems.
The paper begins with a brief account of the aid effectiveness agenda and its role in Tanzania, followed by a discussion of Rottenburg's theory of foreign aid's two inconsistent objectives. After a short account of the study's research methods, three sections of empirical analysis follow, ending with a tentative reflection on national desk officers as potentially re-politicising foreign aid.

Aid effectiveness and its contradictions
Recent decades' structural and political changes in foreign aid have implied a move towards what some scholars label the 'partnership era,' 5 largely consisting of two trends. 6 The first involves donor endeavours to harmonise and coordinate their activities amongst each other to ease the administrative burden placed on aid-recipient countries. Tanzania, for example, is the host of hundreds of domestic and foreign development actors with numerous and often opposing (if not rivalling) objectives. Aid fragmentation has taken a serious toll on recipient governments' capacity to administer donor interventions and respond to donor requests for information, response, and influence. 7 Solutions have been identified in, for example, the formation of development partner groups and joint funding arrangements. The first trend also involves moving away from donorimposed conditionality towards recipient country ownership and long-term commitments by donors to align with the policies, programmes and administrative systems of their partner countries. This sprung out of a shared recognition among donor states of the failure of structural adjustment programmes imposed on aid-recipient countries during 1980s and 1990s. 8 Instead of effecting inclusive growth and poverty reduction, privatisation reforms and public sector cut downs had caused mass unemployment and rising costs for basic services. Aid conditionality was (and still is) criticised not only for violating the sovereignty of recipient states and imposing neo-colonial rule in the global South, but also for undermining recipient governments' accountability towards their citizens. 9 Integral to efforts of enhancing national ownership has been to 'localise' aid, implying the transference of power and influence from foreign to partner country actors and investing in the contribution to development policy and programming of local expertise and knowledge. 10 One example of aid localisation has been to increase the share of national staff in international organisations, such as national desk officers employed by public aid agencies.
The second trend involves an enhanced focus on measuring, identifying, and conditioning the disbursement of aid on 'development results.' In foreign aid discourse, this is called results-based management, and implies that aid allocation is determined based on performance as estimated by the recipient government's policy strength and the measurable outcomes of those polices. 11 This trend, too, has its origin in the perceived failure of foreign aid during the 1970s and 1980s, though in this case 'aid fatigue' and its proposed solutions were largely centred in donor countries. As the legitimacy of aid eroded in the eyes of taxpayers and politicians in donor countries, higher demands were placed on aid agenciesand thus on their recipient partnersto demonstrate that public funds were properly managed. This paved the way for a new form of conditionality focused on 'governance,' led by the World Bank with its emphasis on and systems to enforce 'good governance.' 12 Today's focus on results-based management is the most recent form of aid conditionality, Fraser and Whitfield write, which has come to supplement rather than replace the 1980s' economic conditionality (structural adjustment schemes) and the 1990s' governance conditionality. 13 It basically concerns ensuring that investments made obtain the desired objectives in terms of development 'on the ground.' In recent years, this has implied high expectations on the traceability of development assistance, where aid is implemented, monitored, and assessed quantitatively according to presumably logical and predictable chains of causality (outlining 'input→activities→output→outcome→impact'). 14 In March 2005, the principals of results-based management and national ownership were formalised when over 100 governments (and development institutions and NGOs) signed the Paris Declaration on Aid Effectiveness. The Declaration stipulates five partnership commitments: (i) ownership -'Partner countries exercise effective leadership over their development policies and strategies, and coordinate development actions'; (ii) alignment -'donors base their overall support on partner countries' national development strategies, institutions, and procedures'; (iii) harmonisation -'donors' actions are more harmonised, transparent and collectively effective'; (iv) managing for results -'managing resources and improving decision-making for results'; and (v) mutual accountability -'donors and partners are accountable for development results.' 15

Two challenges facing public donor agencies
There is an inherent tension between the basic tenets of aid effectiveness: supporting recipient country priorities and systems, on one hand, and holding recipient governments accountable and demanding they demonstrate development results, on the other. 16 In the words of Fraser and Whitfield, the partnership era is marked by 'an ambitious, externally defined project for the transformation of African countries, but it requires local actors to drive the project and has little confidence that they are willing or able to do so.' 17 This tension relates to Richard Rottenburg's discussion of two types of inconsistencies facing foreign aid actors. 18 One concerns how development agencies are mandated to work towards two different objectives: progress and emancipation. Progress is defined and assessed in terms of technical modernisation and material wealth, and is pursued by identifying and trying to eliminate deficits in these regards in recipient countries. The progress objective upholds industrialised democracies as the most developed and having the most knowledge of how to eliminate development deficits. Thereby, it asserts the universality and objectivity of the expertise of the global North and thus an unequal relationship between donors and recipients of aid. Emancipation involves a recognition of the contextual nature of development and the need for local expertise in effecting and assessing development. Moreover, it speaks of a duty to respect recipient states' sovereignty and right to participate in decision-making. The objective of emancipation, in contrast to that of progress, affirms an equal relationship between donor and recipient and the relative nature of development knowledge. 19 The aid effectiveness agenda also speaks to a second inconsistency discussed by Rottenburg, which involves the fact that state development actors should treat what he calls effectiveness and legitimacy as equally important objectives. 20 For public aid agencies, it implies that they are, on one hand, meant to use taxpayers' money in the most efficient way they can in order to maximise development results. On the other hand, it means they need to follow bureaucratic procedure, which is not always the most efficient. In fact, development banks and other public aid actors are what Rottenburg calls institutionalised organisations, drawing on Meyer and Rowan. 21 They are organisations with an official mandate and clear rules and regulations, whose existence is justified by their capacity to follow these rules and regulations. The same is true for state actors in the countries receiving aid, one could add. In the partnership era, when donors are meant to use recipient governments' financial and administrative systems, donors face the contradictions embodied in institutional organisations also in their partnering institutions. The latter are often perceivedboth by donor and recipient partnersas less efficient than the bureaucracies of donor states (hence, partly, the many 'governance' programmes and civil service reforms initiated by donors 22 ). The increasing emphasis placed on showing development results and aligning with partner systems thus means that effectiveness goals are challenged by legitimacy demands about following due procedure in two bureaucratic realmsthose of the donor and the recipient state.
Researchers have pointed to different ways in which aid agencies deal with these two inconsistencies. Rottenburg highlights practices of 'loose coupling' and 'ritual evaluations,' where foreign aid actors tie official objectives to actual practices in only vague ways, avoid evaluating their own work in manners that may reveal contradictions, assess results only according to the progress narrative, and place blame on 'cultural factors' among recipient partners when evaluations point to operational failure. With regards to dealing with the conflicting objectives of progress and emancipation, the type of donor strategy attracting most attention by critical development scholars has concerned the depoliticisation of aid, that is, donors' couching aid conditionality in purely technical terms to make it seem as if it does not infringe on partners' political decision-making. 23 In this article I bring to light another way in which foreign aid agencies deal with these two contradictions, grounded in their everyday collaborations and relationships with recipient partners. More precisely, it centres on the practices of a particular group of donor employees, notably the desk officers recruited locally in aid-recipient countries, in this case Tanzania.

Aid effectiveness in Tanzania
Tanzania has been at the forefront of the Paris Declaration. Its own problematic experiences of foreign aid significantly influenced the formalisation of the principals of national ownership and results-oriented aid. Since its independence in 1961, 24 Tanzania has been one of the biggest recipients of official donor assistance in Sub-Saharan Africa. 25 President Julius K. Nyerere's policies of African socialism or Ujamaa (literally meaning 'familyhood') attracted substantial donor support in the 1960s-1970s. Formalised in the Arusha Declaration of 1967, they revolved around nationalisation, selfreliance, and communal agriculture. However, untenable economic policies, combined with the global economic crisis and the rise to power of conservative, neoliberal governments in Western donor states, brought on wide-ranging changes in donor approaches in the late 1970s and beginning of 1980s. 26 Substantial cut-backs in aid were made, and, despite much resistance from the Tanzanian leadership, Tanzania became one of the subjects of structural adjustment in the mid-1980s. While the reforms proposed by the IMF and the World Bank diminished budget deficits and freed large volumes of foreign aid, they also implied comprehensive reductions in public spending and social services. Moreover, they allowed for considerable foreign interference in Tanzanian policy-making. Relations between donors and the Tanzanian government became increasingly tense and climaxed in a major corruption scandal and the consequent freezing of aid flows. The group of Tanzanian and foreign experts tasked with evaluating the roots of the breakdown of relations attributed responsibility both to the weakness of the Tanzanian government and to the invasiveness and micromanagement of donors. 27 A set of basic principles were drawn up to direct future cooperation. Centring on national ownership, development partnership, transparency, mutual accountability, and efficiency in aid delivery, these principles came to guide the Tanzanian government's approach to aid management, and encompassed the establishment of joint policy dialogue forums and the preference for un-earmarked support to the state budget. 28 Tanzania thus became a pioneer and inspiration in the new 'partnership era', and has been described in international development circles as a 'model country of ownership.' 29 When Dr. John Magufuli was elected President in 2015, the Tanzanian government embarked on a forceful programme of cutting public service costs and fighting corruption in the name of development, state sovereignty and ending aid dependence. While this reform agenda was initially praised by the international community, donors' readiness to support it, like previous Tanzanian reforms, continues to depend on their ability to influence policy-making through participation in government planning committees, dialogue forums and working groups The increasingly authoritarian tendencies of Magufuli's administration in recent years, manifested, for example, in its crackdown on LBTQ persons, birth control, and teenage pregnancy, have sparked political condemnation and aid suspension among major donors. 30 While the full effect of these reactions remains to be seen, they indeed indicate donors' continued involvement in Tanzanian political affairs.

Research methods and empirical focus
Empirical findings are based on open-ended interviews with forty development professionals and on observations of the eighteen aid agencies in Dar es Salaam for which they worked. Thirty-four of these professionals were Tanzanian citizens employed on local contracts, meaning they worked for the embassy or field office of a certain ministry or agency, as opposed to its headquarters in the donor country. In this article, I call them national staff. The organisational structure of public foreign aid agencies differ significantly. In some donor countries, foreign aid is managed by one or several state agencies, having their own offices separate from the donor country's embassy in recipient countries. In other donor countries, foreign aid is managed by a ministry rather than a separate agency, and is an integrated part of the embassy in the recipient country. For reasons of simplicity, I generally refer to central structures of public aid management located in the donor country as headquarters, decentralised structures located in Tanzania as field offices, and the larger institutional structures managing foreign aid on behalf of a donor state as agencies.
All but one 31 of the national staff I interviewed held technical positions, rather than administrative or diplomatic ones, meaning they were in charge of the everyday management of aid interventions. Their tasks included assessing funding proposals; designing and following up the implementation of aid interventions; commissioning audits, evaluations, and various kinds of consultancy work; and pursuing continual dialogue with agency partners. They were generally referred to as 'desk officers,' though their specific titles could be, for example, programme officer, project manager, or analyst. The study also included interviews with six foreign staff members in management positions. They shared the nationality of their employer and had been sent out from donor headquarters to lead their country's development cooperation with Tanzania in its field office in Dar es Salaam. I call them posted staff.
All interviews were conducted in English and centred on the study's thematic focus, notably national desk officers' professional role, position, and contribution in relation to their posted colleagues. The recruitment of interlocutors was initially made through email to field office managers (requesting permission to interview them and any of their national employees), but also, eventually, through snowballing, where one interlocutor introduced me to his or her colleague at another agency. Data collected from interviews were processed using NVivo, a computer software programme for categorising and analysing qualitative data.
The agencies drawn upon in this article are public agencies, meaning that the aid provided comes from state revenue in donor countries. The agencies include certain United Nations bodies and one international development bank. However, the majority represent individual countries belonging to the Organization for Economic Co-operation and Development/Development Assistance Committee (OECD/DAC), the OECD's forum for discussing questions concerning aid, poverty reduction, and development, and the main actor behind the aid effectiveness agenda. These government agencies have particular functions and mandates in the current aid architecture. Efforts to harmonise and align foreign assistance with the priorities and systems of aid-recipient governments have implied a turn away from project-based support to programme-and sector-wide approaches in the delivery of aid. At the same time, foreign aid has become increasingly fragmented and complex as a growing share of aid work is outsourced to commercial firms, NGOs, or academic institutions. 32 In the face of demands to harmonise aid interventions, public aid agencies play important roles as contractors, coordinators and steering agents of aid, and as representatives in negotiations and advocacy with recipient partners and other aid actors. 33 Moreover, bilateral agencies especially, interweave foreign development assistance with their governments' national and international security agendas and economic interests. In Tanzania, bilateral aid constitutes on average half of all assistance to the country, but has at times exceeded 80 per cent. 34 My initial interest in national development experts arose during my employment at the Swedish International Development Cooperation Agency (Sida). It was at Sida's office in Kigali during 2009-2010 that I first became familiar with the complex roles that my Rwandan colleagues were tasked with performing on behalf of the Swedish government. My experience of having been employed in a bilateral aid agency in East Africa also came to influence my relationship with interlocutors during my field research in Tanzania. It happens that development practitioners, especially from donor countries, feel unjustly criticised or misrepresented by researchers seen to portray them as unprofessional or unethical (see e.g. Mosse 2005). I believe my personal background in aid work contributed to a sense of shared professional identity and mutual understanding with my interlocutors. This probably made it easier for me to inquire into some of the more sensitive questions addressed in my research, such as civil servants' use of personal networks to influence formal work procedure. Meanwhile, my identity as a posted, rather than national, aid worker, hailing from a European donor country, may have affected my relationship with Tanzanian staff differently from how it affected my relationship with posted staff, though much, I believe, depended on the topic of inquiry. While it sometimes influenced my explorations of national-posted inequalities, 35 it seemed less pertinent in relation to questions about aid workers' strategies to advance aid effectiveness, as will be described below.

National staff as citizens: combining the objectives of progress and emancipation
The identity of my Tanzanian interlocutors as citizens of their agency's host country emerged as an important asset in negotiations with their foreign employer's Tanzanian partners. This seemed especially to be the case when it came to exerting pressure on Tanzanian government partners to align with donor regulations or expectations regarding the speed of work processes. By taking on the role of the Tanzanian citizen holding her state accountable, national staff recast the interests of their foreign agency as the wishes and expectations of the Tanzanian citizenry. That way, national staff helped their agencies to unite the potentially conflicting objectives of pursuing donor agendas and promoting national ownershipor what Rottenburg would call the inconsistent goals of progress and emancipation. Anna, a Tanzanian programme officer managing support for basic education on behalf of a European bilateral agency, explained how she usually dealt with situations where her agency suspected its Tanzanian partner of corruption and violating agreements made.
As far as I am a Tanzanian, if I know that corruption is taking place within my government, I call those people. Because I'm part of Tanzania! I am a Tanzanian! So I will let them know, and I will let the embassy know. … Because it's my country! You know? We are getting support! So we have to be very wise, we have to be very careful to make sure that this support is reaching … you know, and that it's really bringing the expected results.
In other words, Anna would always confront suspicions of corruption in the Tanzanian government, though not primarily as a foreign government representative with financial investments in the Tanzanian state, but as a fellow Tanzanian citizen who expected the government to lawfully redistribute aid money to the general public. A similar approach was voiced by Patrick, a programme officer working with rural development on behalf of another bilateral European agency. He explained that as a Tanzanian, he had the privilege of being able to 'talk straight' to his agency's partners. He could reach out to his government and urge them to 'pull up their socks,' as he phrased it. 'I tell them "You know what! If this money is not going to the project, that means we are denying the Tanzaniansour fellow Tanzaniansthe support they were expecting".' When his agency's partners showed unwillingness to 'cooperate,' his job as a national was to remind them of their obligation towards the Tanzanian citizens like himself. Failing to honour the agreement made with the donor agency, in Patrick's view, ultimately meant that the Tanzanian government violated the social contract between the state and its citizens. 'Sometimes I work as a mediator to encourage my fellow Tanzanians who are working in the government that we have a duty to make sure that the funds that have been given generously by other people reach our … the Tanzanians.' Those were things that were easier for him to say than for his colleagues from Europe, he inferred. One reason for this was that he was able to speak in Swahili with his agency's local partners, which allowed for a higher degree of honesty. The opportunities for more frank dialogue when using Swahili was brought up by several of my interlocutors. It was seen to allow them to be more 'black and white' with their agency's partners, because they were then speaking as one Tanzanian to another. In fact, 'there are many things that are easier for me to say than for the [foreigners],' Patrick continued.
A foreigner would never say such a thing. And if he is bold enough to say that, it will cause a lot of problems. I mean if you are a poor man, and then a foreigner says 'that's why you are poor!' it's an insult! But if I tell my colleague that 'You know we are poor! You see! We need to come out of poverty! These people have come to help us out of poverty!' then … Although development cooperation was much about diplomacy, Patrick acknowledged, as a Tanzanian he was less constrained by diplomatic language. That was why he could, when he needed, 'be blunt' with his agency's partners. By appealing to his identity as a fellow Tanzanian, Patrick was able to put pressure on agency partners in ways that those partners would most likely interpret as paternalistic and neo-colonial if coming from an agency representative from Europe or North America. Certainly, neither Patrick nor his agency's partners consider national donor employees like himself representative of 'the poor' in Tanzania, but rather members of their country's small urban clique of academic development professionals. Many of these professionals depend on future job opportunities among their agency's local partners, which limits the extent to which they are actually able or willing to put pressure on these partners. Their insistence on their negotiating strength as citizens should be considered in light of their need to continually prove their professional value vis-à-vis their foreign employers who often only offer performancebased, short-term contracts. Still, by virtue of being born and raised in Tanzania, national staff are able to recast foreign pressures in a local mould, making them seem not like the exigencies of a Western government, but as the demands of Tanzania's own peopleperhaps not its primary development beneficiaries, but certainly its champions and experts.
National staff as colleagues: combining the objectives of effectiveness and legitimacy Another identity available for national staff when dealing with Tanzanian partners on behalf of their foreign employer is that of friend, colleague, or acquaintance. National staff often tap into their personal networks as a way to access information and persons, and thereby speed up work processes while officially respecting Tanzanian government systems and procedures. In doing so, they allow donors to circumvent the problem of having to align with slow-moving partner bureaucracies in the name of national ownership and transparency, and are able to deliver (more) timely results as expected by agency headquarters and public opinion in the donor country. Contrary to the case where national staff use their legal status as Tanzanian citizens, it is not necessary to be Tanzanian, as many of my interlocutors pointed out, to enjoy personal contacts in the Tanzanian civil service. A foreigner with a long record of reputable work in Tanzania may be able to draw on a similar network of friends in high places. Yet, due to the high turnover of staff sent out from headquarters, such persons are few in number. Many Tanzanian desk officers, meanwhile, enjoy friendship networks stretching across both public and private realms of the domestic development scene. They are built over time through for example growing up in the same region, attending the same secondary schools and universities, working together at the same offices, or sharing extended family. In Tanzania, it is not uncommon that most university-educated professionals involved in aid-funded work on either the donor and recipient side within a particular sector, such as health or education, know each other or at least know of each other. This is grounded in recent changes in the practice of aid work and the recruitment of aid workers in foreign donor agencies.
First of all, positions designed for nationals in donor agencies have grown in number and some have been assigned more responsibility and influence. This is partly due to donor commitments to aid localisation and efforts to support and use partner country expertise. But it also has financial motives. Following the global financial crisis beginning in 2008, many donor governments have cut aid budgets and sought ways to reduce the administrative costs of aid. One measure involves replacing posted staff in field offices with national ones who cost much less. 36 Qualifying for a job as national desk officer often requires a Master's degree, as well as prior work experience from project management. 37 The large majority of my Tanzanian interlocutors had previously worked for multiple public or private organisations engaged in development work in Tanzania. Ninety per cent had a Master's degree, and an even larger share had studied at either one of three universities in Tanzania. 38 Meanwhile, the size and number of higher education institutions in Tanzania is limited and less than four per cent of Tanzanian citizens reach tertiary education. 39 Moreover, national desk officers are often thematic experts, recruited to manage their agency's operations in a particular development sector. Hence they often have academic degrees and past work experience in one particular field.
Finally, because public foreign aid today operates at more centralised levels compared to earlier decades' project-based support, the closest collaborating partners of desk officers in foreign agencies are usually persons sharing similarly advanced academic and professional backgroundssenior officials in the Tanzanian ministries and state agencies, directors of NGOs, or managers of consulting firms. Hence at both donor and recipient sides, Tanzanians working in international development form part of a rather small professional elite. 40 One Tanzanian programme officer, specialised in renewable energy, recounted how the first task he had been assigned at his current job in a multilateral aid agency was to assess the evaluations made of two projects he had previously helped develop while employed in the private sector. Ties to former employers were especially common among foreign agencies' contracted partners. Tanzanian consultants hired for audits, evaluations, baseline studies, or technical support could often be my interlocutors' former colleagues, employers, or university teachers.
Virtually all Tanzanian desk officers believed their personal networks were critical for their agency's success in collaborating with the Tanzanian government. These networks created, above all, fast lanes through the heavy and often confusing Tanzanian bureaucracy. They were meant to speed up and facilitate donor agencies' access to information and documentation, help them obtain formal approvals, get agreements signed, and make Tanzanian officials attend donor-organised meetings and events. In practice, the use of fast lanes often implied replacing or complementing formally sanctioned means of communication with less formal ones. Instead of writing, signing, stamping, and delivering in person official letters and then waiting for a formal response from a Tanzanian official or office, national staff used their personal phones to contact the concerned official (or a friend close to that official), or simply walked over unannounced to his/her office for a face-to-face meeting. This would be seen as highly disrespectful if attempted by someone unknown to the official, I was told, not least a foreign expert newly arrived in the country. Here again is hence an example of how locally employed staff in overseas agencies are able to domesticize foreign exigencies, advancing donor interests in ways that could be interpreted as neo-imperialist if attempted by a foreigner. In this case, national staff were able to violate aid recipients' formal regulations and systems by couching such violations as personal favours exchanged between colleagues, friends, or kin.

Rationalising the use of fast lanes: 'the problem of African bureaucracy'
On one hand, both national and posted aid workers defended the bureaucracy of the Tanzanian administration, at least to the extent it did not illicit bribes or other forms of corruption. Respecting bureaucracy meant respecting the importance of following formal procedures in the name of public accountability and transparency, making sure government actions were quality-checked, documented properly, and prioritised in a democratic way. On the other hand, many Tanzanian employees acknowledged that following formal procedure took 'too long' for the effective management of aid, at least according to the donor agencies and their regulations. My Tanzanian interlocutors, quite a few with a professional background in the Tanzanian civil service, often brought up the difficulty of relying on formal procedures when cooperating with the Tanzanian government. If one wanted to 'get things done', working through informal channels and systems was the only option.
National and posted aid workers alike raised a number of problems with the Tanzanian bureaucracy that explained its inefficiency. The most frequently cited described it as overly centralised and burdened with too many levels of authority, meaning that even minor inquiries made by donors had to pass through multiple instances of approval before a response could be expected from the government.
Another challenge concerned the complexity of the public administration, where parallel and sometimes contradictory regulations, procedures, and systems made the bureaucracy impenetrable for an outsider. To understand how things worked and what was needed to follow due procedureso as not to step on toes or breach regulation which could block or delay processesone needed inside information only obtainable through personal contacts. Omar, a Tanzanian development analyst working for a North American agency, explained it in this way: 'In other cultures, once you have procedures, regulations, it doesn't matter if you know somebody or you don't. But in our case it matters. Sometimes you need to push things through to make them happen. And you can do that if you know somebody.' Omar gave the example of trying to get a document signed by a government agency. 'For me I think this is straightforward. It just needs a signature. It doesn't need even ten minutes [of work]. But they'll say "We have our own way of working. So you wait and we will get back to you." If you know somebody, then you can go and talk to him and say "Release this one, we really need this document for us to process the disbursement".' With a personal connection in the ministry, it was easier to ask the government to 'go the extra mile,' as Omar put it.
A third challenge in collaborating with the Tanzanian state was that civil servants usually ignored persons they did not know personally. Prosper, a young programme officer administering support to natural resource management on behalf of a European bilateral agency, put it as follows: "If somebody doesn't know you then … at least in African governments … ' Prosper chuckled while he searched for words. ' … then they don't pay attention to you. But if you know them and they know you, you can easily push things forward.'" Prosper gave as example the director of a Tanzanian ministry division with which his employer collaborated. The director of that division had been his former housemate when they had both pursued postgraduate studies in Europe.
We were living in the same flat, sharing food and all … so if I need anything from the ministry, I just have to call him. … But if I didn't know him, maybe he'd be like "who is this guy? Why doesn't he [turn to my staff?]" But now I can just go there and talk to him, I don't even have to make an appointment.' My interlocutors' descriptions of the challenges facing the Tanzanian state administration bear resemblance to anthropological analyses of state bureaucracies in West Africa. Thomas Bierschenk describes the latter as marked by parallel and conflicting regulations, much as a result of decades of externally imposed administrative reforms. 41 On top of plural official rules operate the de facto rules, dictated by 'local professional cultures, bureaucratic customs, traditions and routines, accepted rules and practical knowledge.' 42 Consequently, any actor attempting to navigate through these bureaucracies need to relate to multiple, both formal and informal, rules and routines. Blundo and Olivier de Sardan describe West African administrative systems as punctuated by favouritism, where stringpulling and the general exchange of favours are routine practice. 43 Social networks transcend the sphere of the public office, meaning that civil servants take on the role of brokers and patrons. 44 The favouritism marking professional life is an extension of African social life, Blundo and Lemeur write, where the relational capital of individuals is particularly extensive. Eliminating favouritism from professional life is close to impossible 'without insisting on a complete rupture between normal social life and professional life.' 45 This means that it is difficult for many civil servants to say no to requests for favours made by relatives, friends or acquaintances, even if it implies having to bypass official rules. 46 The importance of personal relations has been similarly highlighted in African non-governmental sectors, though in other respects. Thomas Yarrow shows how individual friendships are key to the survival of NGOs in Ghana, in the face of a controlling state and international donors with diverging political interests. 47 'Being connected' enables NGOs to protect their ideological and organisational autonomy, helping them to mobilise and access funds and resources, gain media attention, and engage in formal collaborations amongst each other. However, because these friendships are not formally accountable or transparent, their use among NGO workers contradicts donor conceptions of 'good governance,' where informal relations are generally associated with personal interests and therefore corruption. 48 My research suggests, however, that foreign donors not only accept but benefit from employees' personal friendships. This benefit concerns neither help in respecting social codes and coping in a dysfunctioning work place (as in the case of West African bureaucrats) nor support to maintain political and economic autonomy (as in the case of Ghanaian NGOs). Rather, it concerns improving the efficiency of aid delivery. Meanwhile, the rational is the same in all three cases: non-transparent and nonaccountable relationships are considered a necessary evil serving the greater good of societythey are a (personal) means justifying a (public) end.
All three cases also point to practitioners' ambivalent approach to the use of friendships in their work, recognising the risk that such relationships may be abused for personal interest. 49 National desk officers' willingness to respect the Tanzanian government's formal procedures competed with pressures to speed up collaborative process to a pace acceptable to their foreign employer. This tension is well illustrated in their approach to Tanzania's new government administration. President John Magufuli has become known for 'cleaning up' among corrupted officials and formalising administrative procedures. 50 Nicknamed the 'Bulldozer,' he has vowed to ensure that no person in Tanzania will enjoy preferential treatment in dealing with the state. 51 This political agenda has generally been applauded by the international community, and it was also commended by my Tanzanian interlocutors. On the other hand, many national staff admitted, if not complained, that fast lanes they had previously enjoyed thanks to personal contacts had been blocked by Magufuli's administration. Consequently, it had gotten harder for them to do their job, at least at the same pace as before.

Personal networks as professional merit
The two different rationales used to justify favouritism and clientelism in African public bureaucracies (efficiency and social pressure) converge in the context of international development cooperation in Tanzania. They form part of the same pressure chain that has its origin in demands for efficiency by donor governments (and their citizens). Without the pressure on Tanzanian civil servants to 'do a favour' for friends in foreign aid agencies, national staff's personal networks would serve little purpose to the management of such agencies.
Foreign officials argued that the ability of national staff to draw on their connections in the Tanzanian government was important enough to influence recruitment processes. Personal networks were described as a criterion assessed in the employment of Tanzanian desk officers. Gabriel, a posted manager in a bilateral European agency, recounted the selection process of candidates for two national posts. One had concerned a position as political analyst. The agency had chosen a person with personal connections to several parliamentarians and political party officials. For the position as programme officer in a particular sector in which the agency was involved, it had been important to recruit somebody who knew people in the four government institutions with which the agency collaborated. Tom, another foreign official managing aid on behalf of different bilateral agency, pointed out how happy he was that several of his Tanzanian employees were friends with a number of permanent secretaries in the Tanzania government. Those connections, he argued, were 'excellent [in order] to get access.' He was also thankful that one of his Tanzanian employees had a friend working in Tanzania's Immigration Office. His agency had recently begun to offer services to fellow nationals setting up private businesses in Tanzania. Tom's agency exemplifies a reorientation of many donors towards the private sector and international trade, fuelled by the recession of the global economy and consequent aid cuts, as well as shifts towards political conservatism in many donor states, favouring aid that meets domestic commercial interests. 52 Donor governments, like Tom's, that encourage compatriot companies to invest in markets in Tanzania often assign their embassies or aid agencies in Dar es Salaam a responsibility to assist these companies in establishing themselves locally. Assistance implies helping companies not only to navigate through local bureaucracy and attain the necessary work permits and visas, but also to understand and exploit local market dynamics and the actors they involveand this is understood as greatly facilitated by personal contacts and networks. Hence, as donors seek new development opportunities through market-based competition, personal connections and networks may become even more important.
Inquiries into Tanzanian staff's networks were primarily done by looking at the latter's previous work places during recruitment. No foreign manager was said to actively investigate national staff's networks by asking about non-work related friendship ties. Such ties were described as welcome though unanticipated 'perks,' not requirements. Moreover, no foreign manager said they used their Tanzanian employees to advance donor conditionality or speed up work processes in the manners described by my Tanzanian interlocutors. Neither did my Tanzanian interlocutors ever describe their strategizing as something explicitly requested by their foreign managers. Meanwhile, the line between explicit demand and implicit expectation can be thin in contexts marked by structural inequality and labour precariousness. Half of the agencies in the study only hired national desk officers on temporary, performance-based contracts and many Tanzanian desk officers said they enjoyed less professional authority than their posted colleagues with equivalent experience and qualifications. 53 Hence, many Tanzanians felt a continual pressure to prove their professional value and contribution to their agency. Inequalities between national and posted staff are entrenched in donors' human resource management, in which national and posted staff are treated as exclusive and hierarchically stratified employee categories. National staff earn on average four times less than their posted colleagues and are rarely able to attain management or representative positions in their agencies. 54 This dual system has its roots in colonial government. Though the 'native clerks' were indispensable for the enforcement of imperial rule in Africa and the submission of colonial subjects, they were treated as second class employees and were excluded from senior positions in the colonial administration. 55 The dual employment system was officially motivated by the shortage of skilled labour in the colonies and the subsequent need to attract and reward colonial personnel. 56 Certainly, it was also tied to racialised norms about expertise and its bearers, norms which have been observed to linger inside the walls of today's foreign aid organisations. 57 While my research does not disclose if or to what extent agency managements actively use national desk officers in the manners described above, it does suggest that foreign managers are sensitive to the risk of being associated with or openly accused of clientelism and disrespecting formal procedure. When talking about the importance of friendshipbased entries into the Tanzanian government and private sector, several managers added that it primarily had the function of offering the agency information on how the 'system worked'. 'You know networks have a bad ring to it,' as one manager explained. When saying that networks mattered a lot, she argued, she did not mean her agency used its Tanzanian employees to bypass formal systems. Rather, national staff were connected to people who could inform the agency on whom they should turn to, what documents to submit, and how to formulate themselves in communication with Tanzanian partnersinformation, she added, that would otherwise be very hard to access. Meanwhile, other managers, like Gabriel, described the personal networks of national desk officers as a critical compensation for the malfunctioning of Tanzania's formal state systems. Regrettably, he argued, it was impossible to get anywhere in Tanzania without knowing 'the right people.' Building and using personal networks was not a professional practice restricted to national staff. Several Tanzanian desk officers recounted examples of posted colleagues successfully forging friendships with Tanzanian officials in ways that had offered the agency inroads to policy influence in strategic areas of development cooperation. Apart from their individual personality, foreigners' networking success had largely owed to the duration of their stay in Tanzania (longer than the usual 2-3 years of overseas postings). Several studies, not least in Tanzania, have similarly pointed to foreign donor staff's use of informal channels to influence cooperation with host country partners, from arranging 'back-door side-meetings' to steer health sector policy 58 , to organising cocktail parties and informal house receptions to improve agency leverage in 'policy dialogue' with government officials. 59 Personal networks in donor bureaucracies: questioning African exceptionalism Foreign managers' argument that informal, personal channels are necessary in dealing with the Tanzanian state given the latter's malfunctioning administration deserves scrutiny in light of observations made of the use of personal networks within donor bureaucracies. As Blundo and Olivier de Sardan write, while the gap between official and unofficial is more pertinent in African state administrations, it exists everywhere. 60 Beyond the numerous studies made on corruption and informal systems in Western bureaucracies, research has revealed how personal networks, favouritism and clientelism are found at the centre of some of the major donors operating in Tanzania The European Commission, Shore writes, is marked by a surface of bureaucratic rationality under which resides a parallel administrative system. 63 This system involves alternative rules and routines for promotion, recruitment, and influence-making, which has gained general acceptance among Commission staff. Similar to how my interlocutors in Tanzania pointed to the dysfunction of the formal administration in justifying their use of personal networks when dealing with the Tanzanian state, Shore's interlocutors rationalised the EU Commission's parallel regime as necessary on the grounds that the official system did not workthat 'you needed to "bend the rules" just to get things done.' 64 McKeown observes a similar attitude among high officials in the World Bank and the governments of the United States and India. Friendships between these officials greatly influenced political decision-making during the process leading up to the Indian government's policy path towards the Green Revolution in the 1960s. Policy change was the product of negotiations among a network of people who had known each other for a long time and in a variety of organisational settings, McKeown writes. Echoing my observations of the upper strata of the development industry in Tanzania, McKeown writes that high-level policy-making is a 'small world,' where friends and acquaintances made in one professional setting may be used to attain new goals in later working relationships. 65 Thanks to personal friendships, informal arrangements enabled subsequent formal agreements, where official decisions merely confirmed what had already been negotiated informally. The analyses of Shore and McKeown are both examples of how government actors involved on the donor side of development cooperation seek to 'get things done' by using informal channels that bypass their own formal systems and regulations. Hence, personal friendships are employed by foreign agencies to circumvent not only the administrations of the aid recipients but their own, donor bureaucracies. Moreover, such circumvention is enabled by the pre-existence of informal, parallel systems within these regimes.

Conclusion
While foreign aid is meant to be a partnership between donors and recipients of aid, it also entails proving the efficiency of aid to the taxpayers in donor countries. This is a contradiction that aid actors have faced for decades, but it has become more tangible with the aid effectiveness agenda and its explicit calls for national ownership and results-based management. Research has pointed to different ways in which donor agencies deal with these conflicting objectives, the most frequently addressed being the depoliticisation of aid. It involves how donors conceptualise aid conditionality as purely technical to downplay ways in which it challenges the sovereignty of aid-recipient states by interfering in domestic political affairs. In this article I have highlighted another way in which aid agencies deal with this contradiction, as well as the contradiction posed by pressures to be both efficient and follow due procedure. It concerns the day-to-day management of aid in the field offices of donor governments in Tanzania. More precisely, it spotlights Tanzanian desk officers working in these agencies and how they further donor interests and circumvent formal regulations and procedures by tapping into their resources as locals. Such resources involve national employees' role as Tanzanian citizens with a right to hold the Tanzanian government accountable for how it spends donor money invested in national development. They also involve national staff's ability to draw on their personal networks in Tanzania's development industry to facilitate collaborative processes in which their foreign agency is involved. The objective is usually to speed up the pace of development interventions, but it may also be to put pressure on Tanzanian partners to abide by donor conditionality. Like the 'technicalisation' of aid, the work of national foreign aid staff has a depoliticising function in that it may reconceptualise donor exigencies as favours between friends. But it also has a repolitisising function in so far as national foreign aid workers may couch donor demands as the interests of the general public in the partner country. The share of national staff is growing in many donor agencies. Ironically, this growth partly reflects donor efforts to 'localise' aid, which are meant to transfer power from foreign to host country actors. In so far as national desk officers, elite as they may be, serve as representatives of the ultimate stakeholders of development effortsthe citizens of aid-recipient statesthey may open up opportunities for foreign donors to impose their terms for support in a more openly political way without being accused of Western imperialism and patrimonialism. They can do so by creating an appearance of transferring the origin of demands for partner government accountability from the foreign donor to the domestic population, thereby seemingly attaining the essence of national ownership.