Infrastructure and the Politics of African State Agency: Shaping the Belt and Road Initiative in East Africa

Infrastructure development has experienced a political renaissance in Africa and is again at the centre of national, regional, and continental development agendas. At the same time, China has been identified by African policy-makers as a particularly suitable strategic partner. As infrastructure has become a main pillar of Sino-African cooperation, there has been growing analytical interest in the role of African actors in shaping the terms and conditions and, by extension, the implementation of infrastructure projects with Chinese participation. This follows a more general African “agency turn” in China–Africa studies, which has shifted the research focus onto the myriad ways in which African state and non-state actors shape the continent's engagements with China. This article is situated within this growing body of literature and explores different forms of African state agency in the context of Tanzania's planned Bagamoyo port, Ethiopia's Adama wind farms, and Kenya’s Lamu port. We posit a non-reductionist and social-relational ontology of the (African) state which sees the state as a multifaceted and multi-scalar institutional ensemble. We show that the extent and forms of state agency exerted are inherently interrelated with and, thus, highly contingent upon concrete institutional, economic, political, and bureaucratic contexts in which African state actors are firmly embedded. In doing so, we make the case for a context-sensitive analysis of various spheres of state agency in particular conjunctures of Sino-African engagement.


Introduction
Infrastructure development has experienced a political renaissance in Africa and is again at the centre of national, regional, and continental development agendas, with the African Union's Agenda 2063 striving to " [c]onnect Africa with world-class infrastructure" (African Union 2015). Much of Africa's infrastructure was inherited from the colonial era, during which it served the dual purpose of economic exploitation and territorial control. After independence and in the light of 'modernisation' discourses and practices, infrastructure development in sub-Saharan Africa surged in the 1960s and 1970s, largely driven and financed by the Bretton Woods institutions (Mold 2012). Yet, many of these infrastructure projects were "regarded as too expensive and unsuited to African requirements" (Nugent 2018, p. 22). For instance, before the turn of the twenty-first century, around 30% of the roads built in the 1970s were not in use anymore (Zawdie and Langford 2002). In light of structural adjustment and austerity policies, the continent witnessed a 'prescribed' retrenchment of African governments from largescale infrastructure development throughout the 1980s and 1990s .
Currently, Africa's "re-enchantment with big infrastructure" (Nugent 2018) is part and parcel of a global "emergent regime of infrastructure-led development whose ultimate objective is to produce functional transnational territories that can be 'plugged in' to global networks of production and trade" (Schindler and Kanai 2021, p.40). Infrastructure has again become central to global development discourse and practice. The narrative that infrastructure needs to be built, upgraded, and rendered more efficient to increase African states' ability to deliver economic and social development is now hegemonic (see Schindler and Kanai 2021;Cissokho 2022). Over the past two decades, African governments, regional organisations, the private sector, as well as multilateral and bilateral development financiers have mainstreamed infrastructure in their development planning and mobilised resources to close the so-called 'infrastructure gap', which the African Development Bank (AfDB) estimates at $68-108 billion per year (AfDB 2018, p. 63).
Against this background, China has been identified by African policy-makers as a particularly suitable strategic partner, becoming Africa's largest bilateral source of infrastructure funding. According to the Infrastructure Consortium for Africa (ICA), infrastructure finance totalled at $100.8 billion in 2018, with China contributing about a fourth thereof ($25.7 billion) (ICA 2018, p.7). As infrastructure has become a main pillar of Sino-African cooperation, there has been growing analytical interest in the role of African actors in shaping the terms and conditions and, by extension, the distribution of costs and benefits of infrastructure projects with Chinese participation (Taylor 2020;Wissenbach 2019;Zajontz 2020a, forthcoming;Mohan and Tan-Mullins 2019;Chiyemura 2019a, b;Gambino 2021). This follows a more general African "agency turn" in China-Africa studies which has shifted the research focus to the myriad ways in which African state and non-state actors shape the continent's engagements with China (Alden and Large 2019, p.13; see for instance, Gadzala 2015; Carmody and Kragelund 2016).
This article speaks to this growing body of literature by exploring different 'spheres' of African state agency, namely the agency of political elites in negotiating Tanzania's planned Bagamoyo port, bureaucratic agency in the context of Ethiopia's Adama wind farms and agency of local governance actors in the case of Kenya's Lamu port. As Mohan and Lampert (2013, p.93) underline, "the ability of African actors to exercise […] agency is highly uneven and can have as much to do with African politics as it does with the politics of China-Africa relations". In this paper, we thus respond to recent calls for a "contextual approach […] that considers the contours and specificities of African agency" (Links 2021, p.124). Our context-sensitive analytical approach towards various spheres of state agency focuses on ways in which African state actors are embedded in and interact with specific political, legal, institutional, and economic contexts.
As such, we seek to transcend reductionist and generalising assumptions about a lack of African agency in light of power asymmetries between Africa and Chinacommonly understood to be tilted in the latter's favour. Certainly, China's economic and political clout in global affairs gives Chinese actors leverage in their engagements with their African counterparts. However, as this paper shows, the distribution of material capabilities and soft power at the international level is by far the only determinant for the degree of African agency in concrete Sino-African encounters. We show that, notwithstanding existing power asymmetries, African state actors shape the development of infrastructure projects with Chinese involvement in myriad ways and at different scales (see Nantulya 2021). Our analysis demonstrates that political, economic, and legal-bureaucratic contexts within African states do matter and crucially condition infrastructure projects with Chinese involvement.
The article draws on primary data collected through extensive field research on Sino-African cooperation in the infrastructure sector in China, Ethiopia, Kenya, and Tanzania. This data was triangulated with relevant secondary literature and primary textual sources, notably media reports and official documents, such as government and company reports and press statements. Semi-structured elite interviews were conducted in Tanzania throughout 2019 with over thirty Tanzanian government officials (including senior civil servants and top officials from relevant ministries and state agencies), civil society representatives, and academics. Passive observation was undertaken in both Dar es Salaam and Bagamoyo. The Ethiopian case study is based on over fifty elite and ethnographic interviews with Ethiopian federal and sub-national state officials, as well as representatives of the Chinese government and Chinese enterprises. Interviews in Ethiopia were conducted between April 2017 and January 2018. The Kenyan case study draws on interview data and ethnographic observations with Chinese government officials, representatives from Kenyan national and sub-national state agencies, and Chinese state-owned enterprises, conducted between June 2018 and July 2019 with over forty participants.
The article proceeds in five sections. First, we outline Africa's current Chinesebacked infrastructure boom and Africa's gradual integration into the Belt and Road Initiative (BRI) and situate this work within the recent literature on Sino-African relations. Here, we propose an explicitly non-reductionist understanding of agency which does justice to the variety of African actors involved in negotiating and mediating Chinese infrastructure projects in East Africa. Section two analyses Tanzanian developmentalist state agency in negotiating the Bagamoyo port and Special Economic Zone. Section three discusses Ethiopian bureaucratic agency in the planning processes of Adama I and Adama II wind farms. Section four deals with Kenya's Lamu port, highlighting the agency of local governance actors vis-à-vis Chinese contractors in the context of pre-existing centre-periphery relations. Lastly, the paper concludes by underlining the key findings and suggesting future avenues for research.

The Belt and Road Initiative, Infrastructure, and the Role of 'the' State
Infrastructure is central to the BRI. Initially aimed at closing infrastructure gaps in China's immediate regional neighbourhood, the initiative has now developed into a trans-regional-if not global-project which promotes multi-sectoral connectivity across Asia, Europe, Africa, South America, and even the Arctic. Although African countries were formally included in the BRI only in 2017, many already existing infrastructure projects have now been incorporated in this connectivity initiative (see Gambino 2022). For instance, the Lamu Port-South Sudan-Ethiopia Transport (LAPSSET) Corridor, a regional project initiated by East African governments, is now branded as a BRI 'project catalysing connectivity' (Belt and Road Forum 2019). Even though Chinese financing is marketed as having 'no strings attached', it is often accompanied by a stringent set of requirements, particularly with regards to the choice of contracting firms (Alden and Jiang 2019; Mohan and Tan-Mullins 2019). China Export Import Bank (China Exim Bank), for example, requires 70% of contract procurements (such as machineries and materials) to originate from China, and it was estimated that 89% of Chinese-funded projects are implemented by Chinese contractors (Hillman 2018). A growing number of Chinese firms are also constructing non-Chinese funded infrastructure projects, as in the case of the port of Lamu in Northern Kenya discussed below. A 2017 McKinsey report estimates that Chinese firms won about half of all international Engineering-Procurement-Construction (EPC) contracts and, by value, 42% of World Bank tenders in sub-Saharan Africa (Sun et al. 2017, p.39). In East Africa, Chinese firms have constructed 41.9% of all infrastructure projects (Deloitte 2016, p. 21). Between 2009 and 2015, the revenues of Chinese construction companies from projects in Africa almost doubled-from $28 billion to $54.7 billion (CARI 2021).
While Chinese involvement in Africa's infrastructure sector has been considered as potentially transformative by decision-makers and scholars alike (see Bräutigam 2019; Soulé-Kohndou 2020), it has prompted lively debates around the social, economic, and financial impacts of certain Chinese-funded projects (Taylor 2020;Carmody et al. 2021). Increasingly unsustainable sovereign debt levels-partly related to Chinese loans for large-scale infrastructure-in countries like Djibouti, Ethiopia, Kenya, and Zambia have caused controversies and fiscal pressures for African governments (Zajontz 2021a). Moreover, poor employment and working conditions in Chinese-built infrastructure projects have also caused contestation from governmental agencies, civil society actors, and workers (see Oya 2019;Gambino 2020a;Chiyemura 2021). In light of public controversies about some infrastructure projects with Chinese involvement, it is paramount to scrutinise the role played by African state actors in their negotiation, planning, and implementation.
In the context of Africa's recent infrastructure boom, "the state as the driver and promoter of development is now back on the agenda" (Wethal 2019, p. 492; see also Péclard et al. 2020). As Schindler and Kanai (2021, p. 40) underline, "infrastructureled development" has brought back state-led spatial and development planning with the ultimate aim of "getting the territory right" for smooth and seamless integration into global trade and value chains. As the state plays a pivotal role in the provision, operation, and regulation of a country's infrastructure, a myriad of state agencies and institutions (at various levels of governance) are involved in the development of said projects. These range from central government institutions, such as presidencies, treasuries, line ministries, specialised agencies, and state-owned corporations, to authorities at the sub-national and local scales, for example provincial administrations, townships, and community assemblies. In other words, state agency in the infrastructure sector unfolds in various spheres.

Spheres of African State Agency: State Actors and Their Contexts
Our 'working definition' of African agency is the ability of Africans (as individuals or as collectives) to shape their engagements with external actors in ways subjectively seen as safeguarding and advancing actors' interests and objectives. We root our analysis in a dialectical understanding of the agency-structure relationship. As Wight (2006, p. 281) argues, "[t]he idea of an agent acting in a structural vacuum, or structures acting without agents, is logically impossible". This also applies to state actors. To transcend reifying and voluntarist conceptions of the state, our analysis of state agency-or rather state agencies-is rooted in a social-relational ontology of the state according to which state actors and institutions are firmly embedded in and interrelated with structural contexts (Jessop 2016, pp. 53-59; in China-Africa studies see Lampert and Mohan 2015;Ziso 2018, pp. 37-39). As Zajontz (forthcoming) cautions, "if not appropriately conceptualised, there is a risk to reduce African state agency to elite agency and to misconstrue actions of political elites and officials as detached from state-society relations, historically specific state forms and wider political-economic structures".
A more nuanced understanding of (African) state agency requires us to first transcend the "'state-as-agent' thesis" (Wight 2006, pp. 177-178), which mistakes a highly complex entity for an anthropomorphic and unitary subject. Jessop (1990, p. 367) rightly reminds us that "it is not the state which acts, it is always specific sets of politicians and state officials located in specific parts of the state system". The state is thus to be understood as a dynamic and multilayered "institutional ensemble" which provides "a set of institutional capacities and liabilities" necessary to exert state power (Jessop 1990, pp. 269-270, footnote 13).
A social-relational ontology of the state implies that the ways in which state officials and agencies act are conditioned by historically specific state-society relations (see Migdal 2001). We concur with Hagmann and Péclard's (2010, p.550) suggestion that statehood (in Africa) is constantly (de)constructed and 'negotiated' by a variety of actors in a host of "negotiation arenas and tables." In Migdal's (1994, p.8) words, "[w]e need to break down the undifferentiated concepts of the state-and also of the society-to understand how different elements in each pull in different directions, leading to unanticipated patterns of domination and transformation." A nuanced analysis of state agency in China-Africa relations must acknowledge the multiscalarity and multifacetedness of the state. For one, it must account for "the manifold moves and efforts made at local, national and international levels to arrive at new arrangements towards the organisation of public authority" (Doornbos 2011, p. 201). Secondly, it must account for the complex spatio-temporality of 'the' state, that entails myriad modalities of governance across time and space (see for instance Boone 2003;Olivier de Sardan 2011;Jessop 2016).
This paper sheds closer analytical attention on three (out of many) 'spheres' in which state agency is exerted by different state actors, namely the agency of political elites, bureaucratic agency, and agency of local governance actors. The first sphere in which African state actors shape the terms and conditions of Chinese projects and investments are the top echelons of the state. The agency of political elites in Sino-African relations has received increasing scholarly attention (see, for instance, Hodzi 2018; Mohan and Tan-Mullins 2019). Indeed, state leaders play a pivotal role in "defining and mediating the external expression of state preferences" (Brown 2012(Brown , p. 1892. Especially at an early stage, negotiations with Chinese financial institutions, investors or contractors are usually headed by cabinet ministers and other top officials. As Soulé-Kohndou (2020) argues, delivering infrastructure is a key means of gaining electoral support and, hence, consolidating political power for leading politicians, both in democratic and less democratic contexts. Moreover, it is usually the presidency, the prime ministerial office, and relevant top officials which-in consultation with various social groups and organised interests-set the (political) agenda in the infrastructure sector. It is also these top-level politicians who crucially determine the political context for the bureaucracies and state agencies to see infrastructure projects through.
Hence, Mohan and Tan-Mullins underline that "the agency of Southern political elites shapes how infrastructure is financed, funded and utilised, which are ultimately questions of 'who benefits?'" (2019, p. 1370). For instance, throughout the 2010s, Zambia's ruling party has strategically used Chinese loans to advance its ambitious 'development-through-infrastructure' agenda (Zajontz 2020a). The keen interest of Chinese banks and firms to fund and construct infrastructure was strategically employed by Zambian elites to pursue infrastructure-led development, whilst simultaneously furthering certain vested interests with the help of 'not so public' procurement processes (Zajontz 2020b, pp.123-130). The case study of the negotiations over a Chinese investment in a mega-port and Special Economic Zone in Bagamoyo, Tanzania, reveals how political elites-by means of the powers vested in their offices, or more abstractly put "positioned-practice-places" (Wight 1999)-can set the tone for negotiations of infrastructure projects. As will be shown, the agency of state actors in the sphere of political elites is thereby not autonomous but very much constrained or enabled by the balance of political powers within the state, hence the wider political context in Tanzania (see Jessop 2016, p. 54).
The second sphere where African state actors exert agency when dealing with Chinese stakeholders is that of African bureaucrats, interchangeably understood as government officials, civil servants, as well as technical and administrative experts. Such actors, by means of their access to various forms of institutional and regulatory power, are able to exercise "control over the process and outcome of negotiations with China in accordance to their country's national regulations" (Soulé-Kohndou 2019, p.191). They do so by interpreting and applying the 'rules of the game', including (in)formal laws, standards, norms, procedures, routines, and conventions. Therefore, political decision-making in the infrastructure sector is highly dependent on bureaucratic expertise and practices. Recent studies are recognising the role of bureaucrats in enacting African agency in the context of Africa-China relations. In her study of Benin-China engagements in infrastructure development, Soulé-Kohndou (2019, p. 202) concludes that "bureaucrats located in ministerial departments in charge of reviewing calls for tenders, monitoring the execution of public works and closing projects" are able to exercise agency by applying (or potentially bending) procedural standards and legal-bureaucratic norms.
In the case of Ethiopia, as shown by , bureaucrats, at various governance levels, are accorded with responsibilities to 'officially' plan, broker, coordinate, negotiate, agree, implement, and manage infrastructure projects in line with and conforming to the values, beliefs, and interests of political forces with access to the state. In this paper, we will show how Ethiopian government officials were able to exercise agency through planning of the arrangements around the finance and development of the wind farms in ways subjectively seen to advance the political and economic interests of their country. The bureaucratic sphere of state agency is thereby firmly embedded in the politics of broader political and governance structures of the Ethiopian authoritarian developmental state.
While political elites and state officials, by means of their (often unhinged) access to state institutions and resources, are certainly key actors in shaping their countries' relations with China, they are evidently not the only actors involved. The third sphere through which African actors exert agency and (re)shape their encounters with Chinese actors is local governance. As an analytical category, governance can be defined as "any organised method of delivering public or collective services and goods according to specific logics and norms, and to specific forms of authority" (Olivier de Sardan 2011, p. 22). In the African context, political decentralisation and devolution-driven by endogenous and exogenous forces, such as structural adjustment programmes (see for instance D'Arcy and Cornell 2016; Kanyinga 2016)caught pace in the 1980s and 1990s when centralisation was seen as a leading cause for inequality, marginalisation, and conflict. Although sub-national authorities in devolved systems play a key role in the provision of goods and services (including infrastructure), the decision-making processes for the development of large-scale infrastructure remain highly centralised (Péclard et al. 2020).
In Kenya, since the 2010 constitutional reform, the newly-formed county government, as well as social and political actors at the local level of governance, have constantly renegotiated their relations with the national government. This dynamic has also affected infrastructure projects with Chinese participation. At the same time and in light of the growing participation of Chinese actors in Kenya's development agenda, local governance actors are also engaging and negotiating with Chinese stakeholders directly (Wang and Wissenbach 2019;Gambino 2020b). Both these dynamics unfold in the Chinese-built port of Lamu, where the local administration and communities of fishers and mangrove farmers are shaping the decision-making and implementation processes for this port. The sphere of local governance thus refers to the critical juncture of 'layers' of power across time and space, which add to and intersect with one another when a 'local' political authority is formed (Bierschenk and Olivier de Sardan 1998). As will be shown below, the devolved system in Kenya has "created multiple opportunities for a diverse range of actors" (Chome 2020, p. 4), who reshape the development of infrastructure projects so as to include their 'local' interests.

Transforming the State from the Top: The 'Bulldozer Effect'
Magufuli's election in 2015 followed a period of waning popular support for the ruling party Chama cha Mapinduzi (CCM) and growing levels of dissatisfaction within the electorate regarding widespread corruption and sluggish economic development (Tsubura 2018;Paget 2021). The 2015 presidential race was marked by fierce political debates about overdue developmental benefits from foreign investment and by demands for an indigenisation of Tanzania's economy (Schlimmer 2018). Having campaigned on a populist platform that promised higher public revenues from foreign investment and a resolute fight against corruption and tax avoidance, Magufuli, once in office, pursued a "[c]onfrontational and authoritarian approach towards the private sector" (Andreoni 2017, p. 37). This applied not least to foreign investment, including investments from the 'all-weather friend' China. A Tanzanian scholar suggested that Magufuli tried to capitalise on growing negative public sentiment towards Chinese investments, which had become manifest in a popular narrative that "this country is going to be colonised again" (Interview, M. Shangwe, 15th November 2019; see also Kinyondo 2019). In a foreword to the book Tanzania's Industrialisation Journey, 2016-2056, Magufuli wrote: "Most importantly, we must seize control of our economy and destiny. This will require courageous leadership, self-confidence, ingenuity, hard work and economic patriotism" (Magufuli 2017, p. viii).
The "resource nationalism" (see Jacob and Pedersen 2018) and state-steered developmentalism pursued by the Magufuli administration took on increasingly autocratic forms, as national development was stylised as an overarching goal that should not be compromised by political debate and contestation (Paget 2021). Magufuli earned himself the nickname 'Bulldozer' due to his uncompromising anti-corruption campaign, radical government interventions in the economy, and repressive measures vis-à-vis the political opposition, the media, and civil society. Institutionally, he worked towards transforming the Tanzanian state apparatus into an autocratic developmental state. His tenure was characterised by centralisation of decision-making power in the presidential office, a tightening of control and oversight of the treasury, line ministries, and state agencies, as well as a rigid scrutiny of public spending, borrowing, and procurements. Magufuli aligned his cabinet and newly installed top bureaucrats with his 'economic patriotism', with frequent dismissals of top officials creating an atmosphere of the constant threat of the president's zero-tolerance approach towards negligence and malfeasance (Polus and Tycholiz 2019;Andreoni 2017). This was passed on along the 'line of command', with senior officials and bureaucrats in relevant ministries and state agencies swiftly subscribing to Magufuli's economic-nationalist rhetoric and practices.
A former Tanzanian senior official, for instance, acknowledged that the government had previously failed to maximise developmental benefits of infrastructure projects in negotiations with Chinese investors: With the Chinese, sometimes you have to open your eyes and look beyond the curtains because they might be speaking, but at the back of their mind, The political and institutional changes leading up to and following Magufuli's election in 2015 caused the Tanzanian government to reassess the Bagamoyo project and to pursue a confrontational strategy vis-à-vis the main investor CMPort.

'Win-Win' Contested: Autocratic Developmental State Agency
The Bagamoyo project is an instructive case that documents how (changing) political contexts condition the agency of African state actors vis-à-vis Chinese financiers, investors or contractors. Once lauded by both sides as a transformative BRI project, the Magufuli administration pursued a confrontational strategy by fiercely contesting the terms and conditions put forward by the investors. In June 2019, Magufuli himself outspokenly criticised that This project has very difficult conditions. They are exploitative and awkward. We can't allow it. […] In fact, the investors wanted to tie our hands in developing Tanga port, which is very crucial for the oil pipeline from Uganda and others in Mtwara and Kilwa. These are our oldest ports. (quoted in The Citizen 2019) Here, Magufuli referred to so-called 'stabilisation' or 'adverse action' clauses, i.e. contractual provisions that protect the investor from revenue losses as a result of changing legislation/regulation or competing projects. Such clauses have become standard state-backed guarantees in investment agreements and public-private partnerships and intend to minimise risks and secure returns for investors (Hildyard, 2016, 37-38). A former senior government official confirmed that CMPort expected the government not to develop ports within a radius of 300 miles, which would affect all of Tanzania's current major seaports in Dar es Salaam, Kilwa, Mtwara, and Tanga (Interview, 27th November 2019).
The Tanzanian government was also apprehensive of a loss of sovereignty rights with regard to the authority over ports operations as well as over customs and taxation, since, according to the Director of Tanzania Ports Authority (TPA), "taxes, calculations and audits were set to be undertaken in China" (Kakoko, quoted in Musa 2019). A former senior official confirmed that "they [CMPort] wanted to collect revenue themselves" (Interview, 27th November 2019). A Tanzanian academic suspected that the question of who would be in control of the port could have been decisive for the government's hitherto suspension of the project: "one of the things China is interested in, worldwide, is ports and control of major ports. So, it is not just a question of assisting to construct a port but eventually who owns it and who controls it". The interviewee conjectured that relinquishing the operation of Bagamoyo port to a foreign corporation was incompatible with the "kind of nationalism" pursued by the Magufuli administration (Interview, N. Kamata, 5th February 2019).
Another point of contention has been the terms of the envisaged Build-Operate-Transfer (BOT) contract. Whilst CMPort sought a duration of 99 years, the Tanzanian government wanted to limit the contract period to 33 years (Interview, former Tanzanian senior official, 27th November 2019). The investors furthermore expected state-guaranteed compensations of any losses incurred during the implementation of the project-according to Kakoko a "demand [that] can render the country bankrupt" (quoted in Musa 2019). In addition, CMPort sought substantial tax exemptions, based on the argument that projected cargo volumes were still too low for the port to run profitably. Former TPA Director Kakoko revealed that the investors called for exemptions from land tax, workers compensation tax, skills development levy, customs duty, and value-added tax (The Citizen 2019). He justified the government's stance as follows: "Even [if] it was agreed that some taxes be waived, we would have to examine the percentage of exemptions. However, it seems they wanted to invest for free, which would be akin to selling our freedom" (Kakoko, quoted in Musa 2019). Demands for regulatory exemptions caused further controversy. An interviewee confirmed that "they [CMPort] wanted not to be touched by the labour laws, not to be touched by immigration laws.
Tanzanian officials emphasised that the government's initial intention was the establishment of a SEZ to boost manufacturing and related service industries, with the port providing supplementary infrastructure (Interview, Tanzanian top official, 15th November 2019; Interview, former Tanzanian senior official, 27th November 2019; Interview, senior official Tanzanian Ministry of Works, Transport and Communication, 15th November 2019). Yet, at the instigation of CMPort, "the port became the major-and the Special Economic Zone the minor" in the course of the negotiations (Interview, former Tanzanian senior official, 27th November 2019). The Tanzanian government remained adamant and requested detailed projections as to how the investment sum would be used to establish manufacturing and production sites and create jobs within the SEZ (Ibid). As a top official explained, Bagamoyo becomes only attractive if we have that industrial city there. If that matures, they [the SEZ and the port] will go together and it doesn't matter who will get the project for doing the same [the construction of the port].
[…] So now we are looking at things objectively, we want to discuss -it is not as desperate so that we are not taken for a ride.
[…] The issue is here: This set of conditions we are still not agreeing [with]. If you [the investors] have another idea on them, you can come and we'll continue discussing. We are not in a hurry. (Interview, 15th November 2019) While overall the 'grand power asymmetry' between China and Africa (with which much research and commentary remains preoccupied) has evidently not changed, the shifting balance of forces in Tanzanian politics has caused the government to strategically reassess and clearly demarcate the conditions under which a mega-port and SEZ in Bagamoyo are deemed to serve Tanzanian national interests. Magufuli's autocratic developmentalism and related adjustments of governance procedures prompted a contestation of the omnipresent "notion of 'win-win' which is 1 3 argued to have predominantly benefited associated Chinese contractors and suppliers in the past" (Makundi et al. 2017, p. 346).
The Bagamoyo project has shown that the 'sphere' of the elite agency is deeply embedded in dynamic state-society relations. The changing balance of political forces that had swept Magufuli into the presidential office led to the gradual transformation of Tanzania into an autocratic developmental state under his aegis. Following the systematic centralisation of political power and control in the presidency and informed by populist-nationalist calculus, rhetoric and action, Magufuli and his inner circle pursued a confrontational approach vis-à-vis the Chinese investor. The Magufuli administration thus effectively shelved the project over terms that were deemed detrimental to Tanzania's national interests. While negotiations have resumed under Magufuli's successor, President Samia Suluhu Hassan, it seems likely that certain terms that were initially proposed by the investor will remain taboo for the Tanzanian government. The territory for Bagamoyo port has been marked.

Planning Adama Wind Farms: Bureaucratic State Agency in Practice
This section uses Adama I (53 MW) and Adama II (151 MW) wind farms as case studies to expand the work of Soulé-Kohndou (2019) and that of others (see Chiyemura 2019a; Phillips 2019) to characterise and contextualise how Ethiopian state and government officials have exercised agency throughout the planning processes of these Chinese-backed wind farms. We believe a focus on planning is essential in the sense that the possibility of exerting agency is gained, maintained, or lost at this stage. As noted by an Ethiopian government official involved in the planning processes of the two wind farms, "if you fail to plan, you are planning to fail, it's high time we take care of our infrastructure development processes'' (Interview, former Ethiopian ambassador to China 1999China -2004, former member of the EPRDF-Executive Council, 12th May 2017). This section begins by accounting for the authoritarian developmental state model adopted by the Ethiopian government aimed at bringing about 'positive' change to the lives of ordinary Ethiopians through infrastructure-led development. Doing so allows us to understand structural dynamics related to the centralisation of state power and how governance modalities enable or constrain the exercise of agency. The second part of this section will then discuss how the Ethiopian bureaucracies exercised agency seen through the lens of planning and brokering processes around the Adama wind farms.

Electricity Infrastructure Development in Ethiopia
Ethiopia, like the rest of the African continent, has a huge population without access to electricity. According to the Ethiopian Ministry of Water, Irrigation and Electricity (MOWIE) (2019), only 44% of the 110 million Ethiopians have access to electricity, of which 33% are covered by the grid network and 11% by off-grid sources. By the end of March 2017, Ethiopia Electric Utility (EEU) had 2.3 million customers, and only 6,000 towns had access to electricity (Federal Democratic Republic of Ethiopia 2017). Access to electricity is particularly limited in rural areas. Although more than 85% of the Ethiopian population resides in rural areas, about 90% are without electricity supply (ibid). This is mainly caused by the lack and shortage of electricity infrastructure. Developing electricity infrastructure was therefore identified by the then ruling coalition, Ethiopian People's Revolutionary Democratic Front 2 (EPRDF), as a priority sector (Lavers et al. 2021).
At least until 2018, the EPRDF-led government pursued an authoritarian approach to development by centralising decision-making powers in the hands of top party leadership, which tightly controlled the state and government at different governance levels (Bayu 2019). In this context, the ruling party was subjectively seen to have hegemony of value creation, and relative embedded autonomy to independently formulate and implement policies against contending social and market forces (Zenawi 2012). This period was marked by a top-down approach to the formulation of development policies and plans. The Council of Ministers-in consultation with the Prime Minister and members of the EPRDF coalition-set the development strategy and direction, while the design and implementation of development projects were decentralised to relevant bureaucracies which were constitutionally and legally mandated to do so.
Several policy instruments, such as the Growth and Transformation Plan (GTP) I (2010/11-2014/15) and II (2015/16-2019/20), were formulated and implemented to improve public infrastructure in the energy, transport, and communication sectors, as well as to establish export-oriented industrial zones. In the GTPs, electricity infrastructure was identified as a key driver and enabler of set development targets. The government planned to increase electricity generation capacity from 2000 Fig. 1 Energy projects implementation framework. Source: Authors' compilation from field data to 8000 MW (MOFEC 2010). However, only 4180 MW were installed by the end of GTP I. In GTP II, the goal was to increase the energy generation capacity to 17,208 MW (MOFEC 2016, p.179). To achieve these targets, the Ethiopian government outlined an implementation framework, as shown in Fig. 1.
Between 2004 and 2018, the Ethiopian economy grew at an average Gross Domestic Product (GDP) growth of 10.44% and reached over $80 billion in 2018 (Wolrd Bank, 2018). The magnitude of Ethiopia's economic performance considerably improved its economic clout in East Africa. Despite poor performance in the 'good governance' indicators, Ethiopia has continued to attract investment from both the West and the East regardless of the government's disapproval of "economic policies ostensibly favoured by the principal Western donors" (Clapham 2018(Clapham , p. 1157. This all comes from the government's ability to proficiently play one donor against the other to its advantage and also the ability to show value for money through high and "visible returns on their investments" (Clapham 2018(Clapham , p. 1157. As Furtado and Smith (2007, p. 24) remark, "this gives the [Ethiopian] government substantial bargaining power, which it exercises at times by refusing to compromise on its policy agenda." As explained by a MOWIE bureaucrat, "although we are collaborating with the Chinese in many sectors, the selection and prioritisation of the wind farms was […] and will remain an Ethiopian initiative" (Interview, Ethiopian government official from MOWIE, 11th November 2017).
The authoritarian developmental approach by the EPRDF-led government translated into a rigid implementation ecosystem, which provided the context for Ethiopian bureaucrats to exercise agency in the planning processes for infrastructure development.

Planning as Agency
Planning for the development of wind farms in Ethiopia can be traced back to 2001. At the time, the Ethiopian government-faced with erratic climatic conditions that were adversely impacting its hydropower-dominated generation capacity-demonstrated interest in deploying alternative energy sources, such as solar and wind. Yet, the government repeatedly cited a lack of adequate information and data to translate this interest into concrete policy frameworks. Consequently, upon the initiative of then Prime Minister Meles Zenawi, MOWIE lobbied for the country to be selected to participate in the Solar and Wind Energy Resource Assessment (SWERA) programme funded by the United Nations Environment Programme (UNEP) and Global Environment Facility (GEF). In 2004, the SWERA programme concluded that Ethiopia was endowed with wind resources to the tune of 100GW (Ethiopian Rural Energy Development and Promotion Centre 2007). The proactive engagement with international development partners by relevant state agencies is a concrete example of the ability of the Zenawi government to exercise agency resulting in the country being selected as one of the beneficiaries.
The participation in the UNEP-GEF allowed the Ethiopian government to undertake further planning and thus alter the bureaucratic context in which the country's planned energy transition was to take place. MOWIE subsequently invited the German Society for International Cooperation (GTZ, now GIZ)-Technical Expertise for Renewable (TERNA) to conduct feasibility studies for the deployment of wind energy in Ethiopia. Ashegoda, Harena, and Adama were highly recommended amongst 11 suitable sites (GTZ-TERNA. 2005), resulting in the development of the Ashegoda wind farm. Following complications related to logistics as well as the capacity of turbines in the context of Ashegoda-implemented by the French firm Vergnet (Economic Consulting Associates 2018)-bureaucrats swiftly re-assessed the situation and pressed for contractual changes to reach agreed targets.
At the same time, EEP and MOWIE, strategically reoriented and diversified their engagements with international partners to avoid further delays in implementing the government energy transition agenda. As part of these efforts, In the energy sector, the two sides agreed on a new Wind and Solar Master Plan to be fully financed by Chinese counterparts. The Master Plan found that Ethiopia had a wind energy potential of 3.03 TW; 51 wind farm sites were identified and proposed for development with a potential installed capacity of 6720 MW (Hydrochina Corporation 2012). Out of these 51, MOWIE-in collaboration with EEP-selected eight wind warms to be incorporated in the GTP I (Interview, Ethiopian government in the Ethiopia-China Development Cooperation Directorate at MOFEC, 30th June 2017). Subsequently, the Ethiopia-China Development Cooperation Directorate in MOFEC-under the auspices of the Prime Minister's office-concluded an EPC + Financing agreement for Adama wind farms with China Exim Bank (Interview, Ethiopian government official from MOWIE, 11th November 2017). In line with China Exim Bank's common practice, the project was tendered in a closed bidding process that involved only Chinese firms and the technology needed to be sourced from China. HydroChina Corporation 3 (HydroChina) and China Geo-Engineering Corporation Overseas Construction Group (CGCOC) won the contracts (see Table 1 for more details). Nevertheless, by adopting an EPC + Financing scheme, the Ethiopian government-informed by their previous (negative) experiences in the development of Ashegoda-shifted the project delivery risks from MOWIE, a public authority, to the contractor (Interview, Ethiopian government official from MOWIE, 11th November 2017; Interview, Ethiopian government in the Ethiopia-China Development Cooperation Directorate at MOFEC, 30th June 2017).
From the government's initial proactive outreach to UNEP-GEF to the awarding of the Adama contract, the planning process was marked by strategic learning and regular adjustments by the relevant Ethiopian bureaucracies. Planning is therefore crucial because "structures of thought and action may be creatively reconfigured in relation to actors' […] desires for the future" (Emirbayer and Mische, 1998, p.971). Such creative reconfigurations also took place in the context of Ethiopia's energy transition. Reflective of their country's structural limitations, bureaucrats in EEP, MOWIE, and MOFEC successfully negotiated an EPC + Financing agreement with "a favourable 2% interest rate and a loan repayment plan of 13 years and a grace period of 7 years" (Chiyemura 2019b, p. 1). Furthermore, as an official in MOWIE highlighted, the SWERA and the Chinese-sponsored Master Plan significantly increased the government's planning capacity and, by implication, their ability to make informed decisions (Interview, 11th November 2017).
In an interview, the former Ethiopian ambassador to China underscored the importance of "proper project preparations" for the success of development projects (Interview, 17th May 2017). In the case of the energy sector, the Ethiopian government's political and development priorities were crucially co-determined by bureaucrats within relevant state institutions. The case study of the Adama wind farm has shown that Ethiopian bureaucracies crucially influence project implementation by planning, setting, and strategically reformulating development targets, leveraging necessary data and technical capacity, as well as initiating the development cooperation with Chinese actors. Thus, despite existing power asymmetries between Ethiopia and China, the specific political-institutional contexts in which Sino-African engagement occurs enable and condition the exercise of (bureaucratic) agency. Table 1 Adama wind farms project terms and conditions. Source: Authors' compilation from field data a By the time of conducting fieldwork, the project capitalization report was not yet completed so the interest incurred during construction is not factored in. Total investment cost may be more if interest incurred during construction is added b Official data does not state whether it was Preferential Export Credits, Export Sellers Credits or Mixed Credits. Simply it is recorded as a government concessional loan. For more details about the financing models, see (Massa 2011;OECD 2015) and also Export and Import Bank of China Preferential Facilities on http:// engli sh. eximb ank. gov. cn/ tm/ en-TCN/ index_ 640. html

Reshaping Lamu Port: the Centrality of Local Governance
Since 2016, the Chinese state-owned contractor China Road Bridge Corporation (CRBC) has been building Phase I-the first three berths-of Lamu port in Northern Kenya. This infrastructure project, valued at $480 million, is envisioned to become one of the largest ports in East Africa (LCDA 2020). Phase I is currently being financed by the Kenyan government, which, despite seeking funding from international organisations, prioritised the construction of the port to kickstart the development of the transport corridor initiative planned for the Northern regions of Kenya, and beyond (see Gambino 2020b). In fact, Lamu port is part of a broader project, the LAPSSET Corridor, which aims to connect Kenya to South Sudan and Ethiopia through cross-border infrastructure, such as brand-new highways and a railway. This regional corridor is expected to not only connect Northern Kenya to neighbouring countries, but also to global capital through Lamu port.
Lamu port is one of the flagship projects of Kenya's development agenda, Kenya Vision 2030, which heavily relies on the construction and refurbishing of infrastructure to lead the country to "a rapidly industrialising middle-income nation" (Government of Kenya 2007, p. i). This is in line with the development agendas of many other African nations, as well as those of regional and continental organisations (Nugent 2018;Péclard et al. 2020;Zajontz 2021b). Beyond its overarching goal of 'transforming' Kenya's economy, the country's development agenda also envisions that "[b]y 2030, it will become impossible to refer to any region of [the] country as 'remote'" (Government of Kenya 2007, p. viii). The development of LAPSSET Corridor in Northern Kenya fits within this objective, as this region has historically been a marginalised space since resources have often been redirected towards the development of regions considered to be 'productive'.
In the colonial era, connectivity initiatives were concentrated in the regions inhabited by or related to the activities of white settlers (Swanison 1980; see also Kanyinga 2016), particularly in the Central Region and Rift Valley of Kenya. In the racialised visions of the British colonial administration, "there were no outside economic reasons to break the underdevelopment loneliness of Northern Kenya" (The National Christian Council of Kenya 1972, p. 29). This trend of disregard towards the 'pastoral' North persisted after independence, as these regions are home to agricultural and pastoral communities regarded by the national government as 'backwards', inhabiting 'remote' regions, and often "an inconvenience" (Mosley and Watson 2016, p. 453; see also Cormack and Kurewa 2018). In the blueprint of Kenya Vision 2030, the implementation of infrastructure and other development initiatives in these spaces (see Odhiambo 2013), is expected to "turn history on its head" (Government of Kenya 2011, p. 12). For instance, during the inauguration speech of Lamu port in May 2021, President Kenyatta made clear: "marginalisation… you don't want that word ever again" (Kenyatta 2021). This exemplifies the (at least rhetorical) shift from the abovementioned patterns of marginalisation.
The Lamu port project is being implemented against the background of said pre-existing contextual dynamics. As will be shown below, mirroring existing imbalances of power in centre-periphery relations, the institutional and legal authority for project implementation and agenda-setting are concentrated in the hands of national-level state actors. This impacts the ways in which and the extent to which 'local' actors engage with Chinese contractors participating in infrastructure projects in the periphery. The following subsections will highlight how, despite contextual power asymmetry that characterise centre-periphery relations and broader China-Africa relations, local governance shapes infrastructure projects with Chinese participation.

Local Governance in the Context of Centre-Periphery Relations
Centre-periphery relations are a crucial factor determining the room for manoeuvre and degree of influence of local governance actors in reshaping Sino-African infrastructure projects. Concerning centre-periphery relations, Boone (2003, p. 9) suggests that "rulers operate within different structural or strategic contexts", the dynamics of which are to be understood through the analysis of socio-political (and economic) trajectories in the periphery (see also Nugent 2019). This points to the need to bring politics of the 'periphery' to the centre of analysis in the study of infrastructure projects with Chinese participation.
The 2013 elections saw the formation of the first county governments in Kenya, following the promulgation of a new constitution in 2010, which represented an "ambitious and rigorous experiment in democratic decentralisation" (Subera 2013, p. 32). The extent to which these reforms have addressed the unequal allocation of state resources and legal competencies remains questionable. This is particularly true for (mega)infrastructure projects 'at the margins', as the mandate for their agenda-setting, planning, and implementation remains in the hands of the central government. Against this background, the rhetorical focus on 'opening up' of the Northern part of Kenya and 'connecting' these territories to the national market and the global economy not only speaks to a trend of respacing (Engel and Nugent 2010) through the construction of physical infrastructure (see for instance Lesutis 2019), but also to the reshaping of people's perceptions of their futures (Enns and Bersaglio 2020) and their identities (Kochore 2016).
It is important here to remember that, as much as infrastructure projects are envisioned to promote 'connectivity' and 'integration', they also imply disconnection and dispossession. Indeed, some might "end up being marginalised by the 'modernity' to come" (Lesutis 2019, p. 605; see also Appel et al. 2018;Harvey 2018). For instance, some members of the county government in Lamu argue that the LAPSSET Corridor still leaves Lamu residents as "mere spectators", as opposed to participants of development initiatives. For them, the parameters and conditions of 'development' continue to be negotiated in Nairobi (Interview, Senior Official, County Government, Lamu, 13th July 2019). Different 'economies of anticipation' (Elliott 2016;Greiner 2016;Chome 2020;Chome et al. 2020;Aalders et al. 2021;Müller-Mahn et al. 2021) are thus emerging vis-à-vis the development of infrastructure projects in the Northern regions of Kenya, which, in turn, has affected projects with Chinese involvement. On the one hand, in anticipation of the construction of the Lamu port project, political and business elites at the 'centre' moved to acquire land in Lamu, speaking to the not uncommon practice of lucrative land grabbing in the (previously 'untapped') peripheries. On the other hand, Lamu communities, the county government, and other local governance actors are attempting to reshape these infrastructure projects so as to comprise their own agendas and interests.

Centring 'Peripheral' Agency
In the broader Kenyan context, local governance actors have often engaged with Chinese contractors. For instance, popular demands and protests in the context of the construction of the Nairobi-Mombasa Standard Gauge Railway resulted in negotiations between CRBC and different county governments along the railway route (Wang and Wissenbach 2019). Similarly, in Lamu, negotiations take place between the current county government (2017-) and CRBC with regard to job openings and other opportunities in the construction site as they become available (Gambino 2020b). This followed calls from community leaders to include labourers from the Swahili groups inhabiting the Lamu archipelago in the soon-to-be port. Even though the centralised LAPSSET Corridor Development Authority has the (presidential) mandate to oversee and implement the Lamu port project, these examples show that local governance actors have carved room for manoeuvre. Such 'peripheral' agency has prompted Chinese companies to formulate and then redeploy risk mitigation strategies with regard to controversies around local content requirements.
This implies that project implementation of large-scale infrastructure with Chinese participation is highly contingent upon the local political context. In 2012, Lamu elders and LAPSSET officials formed the LAPSSET Steering Committeechaired by the County Commissioner-with the aim of increasing support for the project (Interview, Secretary, County Government, Lamu, 19th March 2019). This committee was a response to community protests (SaveLamu 2011) and a court case submitted against the national government by Lamu communities (High Court of Kenya, 2012; see below). While it did not initially prove conducive to increasing the agency of local governance actors, it provided a platform for engagement amongst different stakeholders. The LAPSSET Steering Committee, however, was short-lived, being dismantled soon after the election of the first county governor of Lamu Issa Timamy (2013)(2014)(2015)(2016)(2017). This decision was likely motivated by party politics (Chome 2020), as the sitting members of this committee had been associated with the benefactor-or patron-of Timamy's political opponent Fahim Twaha, who would eventually win the 2017 elections.
Timamy's time in office was characterised by struggles with the national government over land allocation for LAPSSET Corridor components, which exposed Twaha as a benefactor of land deeds re-allocation (Nema 2017). This further embittered in-county resentment, particularly amongst the Timamy-Twaha factions, the latter of which, in the meantime had been "edged out of the instruments of local governance" (Chome 2020, p. 15). In 2017, however, Twaha emerged as the winner of the county electoral race, becoming the second governor of Lamu. Twaha's government represented a shift from that of Timamy's.

3
On the one hand, Twaha pursued a less confrontational approach toward the central government due to his affiliation with the ruling party in Nairobi. On the other hand, in his role as Lamu county governor, Twaha also faced the necessity to address the concerns and demands of his constituency.
In 2018, following the Petition n22 of 2012 discussed above, the High Court of Kenya ruled that, in the Lamu port project inception, there had been a violation of the new 2010 Constitution due to lack of inclusion of the county government. The court also ordered to compensate fishers for the disruption caused by the port construction and its future operations. Overall, the ruling recognised the exclusion of the county government and called for more inclusive forms of community participation in the planning and implementation of the Lamu port. Consequently, calls for participation in the national development agenda regained prominence in Lamu politics. In 2019, Twaha's government initiated a committee aimed at promoting the inclusion of specific (and at times diverging) 'local' interests in the development of the Lamu port project (Interview, state official, 17th March 2019). In contrast to the previous LAPSSET Steering Committee the new Lamu port committee also includes the Chinese contractor CRBC. The decision to include CRBC in the new Lamu port committee signals the establishment of direct and more formal channels of engagement for local governance actors to raise their concerns and interests.
In the context of the Lamu port projects, by raising, voicing, and manifesting concerns and grievances, local governance actors have renegotiated the insertion of community interests in the national development agenda. Local governance actors have gained room for manoeuvre to promote their specific (and at times diverging) interests in development projects that are otherwise largely characterised by centralised and top-down decision-making processes. The Lamu case has shown that the agency of supposedly 'peripheral' actors is not so peripheral after all. Indeed, local governance actors crucially influence the implementation of large-scale infrastructure projects with Chinese participation.
In the context of the Lamu port projects, by raising, voicing, and manifesting concerns and grievances, local governance actors have successfully renegotiated the insertion of community interests in the national development agenda. This has happened with the support of civil society organisations, vocal individuals, and by means of resistance in the form of public protests or litigation. Local governance actors have gained some room for manoeuvre to promote their specific (and at times diverging) interests in development projects that are otherwise largely characterised by centralised and top-down decision-making processes. Indeed, most of the negotiations in the agenda-setting, planning, and implementation phases of infrastructure projects with Chinese participation are undertaken by national-level actors. The case of Lamu has shown that the agency of supposedly 'peripheral' actors is not so peripheral after all. As discussed with regards to labour relations, local governance actors can, under certain conditions, influence the implementation of large-scale infrastructure projects with Chinese participation.

Conclusion
Inspired by Links' recent call for a "contextual approach […] that considers the contours and specificities of African agency" (2021, p.124), this article explored different spheres of African state agency in the context of three African infrastructure projects with Chinese participation. Rooted in a social-relational ontology of the state which sees the state as a multifaceted and multi-scalar 'institutional ensemble' (see Jessop 1990), our analysis has shown that infrastructure projects with Chinese participation are shaped by various actors in three (out of many) highly dynamic 'spheres' of the state. We have shown that the extent and forms of state agency exerted are inherently interrelated with and, thus, highly contingent upon concrete institutional, economic, political, and bureaucratic contexts in which African state actors are firmly embedded.
First, in the case of Tanzania's envisaged Bagamoyo mega-port and special economic zone, Tanzania's transformation towards an autocratic developmental state under late President Magufuli has resulted in a thorough reappraisal of the project and a confrontational negotiation strategy towards the investor. Second, the Ethiopian case study revealed that the EPRDF-led authoritarian developmental state model created a political and institutional context which allowed the state bureaucracies to crucially influence project implementation by planning, setting, and strategically reformulating development targets, leveraging necessary data and technical capacity, as well as initiating the development cooperation with Chinese actors. Third, the case of the Kenyan port of Lamu showed that the agency of local governance actors reshapes the implementation of infrastructure projects with Chinese participation. Indeed, local governance actors negotiate the inclusion of their (at times diverging) interests vis-à-vis the implementation of the national development agenda.
There are broader conclusions for the study of African agency-or rather agencies-in Africa-China relations (and beyond). The extent to and the ways in which African state actors exert agency is conditioned but not determined by undeniable power asymmetries between China and Africa that are all too often invoked. For an assessment of African agency (or the lack thereof) the concrete contexts in which actors find themselves matter profoundly. This finding calls for further theoryinformed empirical investigations into the complex politics of African agency vis-àvis China and other external actors.

Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.