From mainstream to niche: how value regimes shift in emerging economy upgrading

Efforts of emerging economies to upgrade into global market capabilities are often conceptualized as either discrete choices or ongoing experiments. Mediating between these perspectives, this study uses the concepts of value regime and economic imaginary to examine micro-dynamics in upgrading. Based on the case of global business services in Kenya and South Africa, this study shows that upgrading efforts center around mainstream or niche value regimes which align local resources with specific demands and which are guided by economic imaginaries. These are shared theories of value creation and upgrading success which are linked to global discourses and which legitimize certain local investments while hindering others. Findings show how both Kenya and South Africa initially followed mainstream imaginaries according to which meeting global standards attracts clients and helps achieve scalable success. Misalignments with material realities later promoted niche imagi- naries which focus on the value of local specificity and exclusiveness and which promoted the rise of impact sourcing and specialty services. Findings indicate how the interplay of material and discursive processes affects upgrading dynamics.


Introduction
Upgrading in global value chains has long been discussed as a potential driver of industrial development in emerging economies (Humphrey and Schmitz, 2002;Lee et al., 2014;Blazek, 2016;Kleibert and Mann, 2020;Gereffi, 2019). Upgrading is often defined as a joint effort of local governments and businesses to "make better products, make them more efficiently, or move into more skilled activities" (Schmitz, 2004, p.1), in order to "catch up" with global markets (Lee and Malerba, 2017). Upgrading not only involves the development of higher-level capabilities across firms within a particular location or region (Whitfield et al., 2020), but also the development of individual skills, technology and education infrastructures (Radosevic and Yoruk, 2016). However, upgrading efforts often fail (Blazek, 2016;Horner, 2014), mostly due to high cost and quality pressures from competitive global markets (Brandt and Thun, 2016), and limited opportunities to enter high-value adding market segments (Pipkin and Fuentes, 2017;Thun, 2018). Sub-Saharan Africa is a good example of a region that has been facing severe challenges in upgrading into globally competitive capabilities Friederici et al., 2017;Foster et al., 2018;Whitfield et al., 2020;Kleibert and Mann, 2020).
Thus, one key challenge for local governments and businesses in emerging economies is to make choices about which capabilities to invest in. One often cited distinction refers to the choice between mainstream industry capabilities that emphasize global standards and scalability vs. more niche market capabilities that emphasize location specificity and exclusiveness (e.g. Van Grunsven and Hutchenson 2016, Brandt andThun 2016, Keijser et al. 2021). Yet, a parallel stream of research suggests that the idea of "choice" can be misleading as upgrading often turns out to be a longer-term process of trying to "reach a better deal" (Ponte and Ewert, 2009;Mann and Graham, 2016). This dynamic may involve processes of upgrading as well as downgrading (Blazek, 2016;Blazek et al., 2019) and de-coupling from global production networks (Horner, 2014), as a response to global cost and competitive pressures (Werner, 2016). However, we still lack an understanding of when and how government and business actors shift from one upgrading strategy to another. To better capture these micro-dynamics, I adopt the concepts of value regimes and economic imaginaries from the political economy literature (Jessop, 2010;Levy and Spicer, 2013;Levy et al., 2016). Value regimes denote stabilizations of configurations of economic processes and normative values, supported by imaginaries in terms of dominant discourses and theories of value creation and success that make present and future investments into certain practices and orders more legitimate than others (Jessop, 2010). In this perspective, mainstream and niche approaches to upgrading can be understood as alternative value regimes along with competing imaginaries of upgrading. A value regime perspective thereby highlights the interplay of material processes and discursive dynamics, hence providing a more embedded upgrading perspective (see also Hughes et al. 2015). Based on this understanding I ask: What are key drivers and components of mainstream value regimes vs. niche value regimes in upgrading, and what are key dynamics of shift from one to the other?
This study examines this question for the empirical case of global business services in sub-Saharan Africa, specifically Kenya and South Africa. Global business services are a highly fragmented and competitive global industry that lacks global lead firms and that forces local decisionmakers to make upgrading choices under conditions of high uncertainty and complexity. Extending prior related studies of upgrading efforts in this region (Graham, 2015;Graham et al., 2015;Mann and Graham, 2016;Foster et al., 2018;Keijser et al., 2021) and based on original empirical data, this study finds that both Kenya and South Africa initially invested into a mainstream value regime of global business services focusing on global standards, scalability and competitiveness, driven by dominant global discourses and the example of lead economies in this sector. Later, however, both countries increasingly shifted towards niche value regimes, focusing on impact sourcing, regional sourcing and specialty services. This shift was promoted by a growing misalignment between local investments and global client expectations, which contradicted the assumption that investing into global standards leads to scalability and upgrading success. This led to the formation of niche value regimes and corresponding imaginaries with a stronger focus on local specificity and local/regional demands. This shift was assisted by a reframing of "upgrading success" and the mobilization of locally specific capabilities and business models that emerged previously but were neglected under the mainstream value regime.
Based on my findings I formulate theoretical propositions to guide future research. More broadly, my findings contribute to recent research on shifts in upgrading (Horner, 2014;Blazek, 2016;Blazek et al., 2019) by specifying micro-dynamics underlying these shifts, especially under conditions of high uncertainty and lacking the guidance of global lead firms. By taking a value regime perspective, I propose a middle position between the assumption of distinct "choices" of upgrading paths on the one hand (Van Grunsven & Hutchenson, 2016), and the observation of serendipitous learning on the other hand (Ponte and Ewert, 2009). I argue that imaginaries can play a key role in uniting key governance actors and in guiding and stabilizing investments based on shared theories of value creation and upgrading success. Findings promote a more embedded, socio-political understanding of upgrading opportunities and constraints (Hughes et al., 2015). In addition, they have important policy implications.

Upgrading dynamics in emerging economies: towards a value regime perspective
Development and economic geography scholars have long debated the potential and limitations of "upgrading" as a way to stimulate economic growth in developing countries (Humphrey and Schmitz, 2002;Schmitz, 2004;Lee et al., 2014;Blazek, 2016). Upgrading is typically understood as a complex undertaking that involves the mobilization of territorialized resources and capabilities, and their effective utilization in global markets and production systems (Coe et al., 2004;Liu, 2017;Henn and Bathelt, 2017). These production systems are typically specified as "global value chains" (GVCs) (Gereffi et al., 2005;Kano et al., 2019) or "global production networks" (GPNs) (Coe et al., 2004;Coe and Yeung, 2019). The GVC literature in particular has discussed different forms of upgradingprocess, product, functional and inter-sectoralas ways to increase and benefit from global market participation (see Humphrey and Schmitz 2002). This study focuses in particular on functional upgrading, i.e. moving into higher-skilled business functions, which is considered one of the most rewarding but also challenging forms, as it requires investment into marketable, often higher-level capabilities (Blazek, 2016).
Upgrading processes are driven by both local efforts and conditions, and global or regional demand and expectations. Locally, upgrading efforts typically involve decisions and investments from both firms and government (Lee et al., 2014;Liu 2017;Coe et al., 2004). The potential role of firms and entrepreneurs is often seen in driving capability and market development (Su and Hung, 2009), and in creating new business models promoting upgrading (Sturgeon and Kawakami, 2011). Governments are expected to promote upgrading by financing and stimulating human capital development (Su and Hung, 2009), by promoting job creation (Vernay et al., 2018), and by attracting and connecting foreign business partners with local firms Lee et al., 2014). In addition, governments may become important buyers of products and services (Horner, 2017). Government policies (top-down) and firm strategies (bottom-up) are seen as interrelated dimensions and drivers of local capability development (Behuria, 2019;Jungwirth and Mueller, 2014;Vernay et al., 2018). Pipkin and Fuentes (2017) call this dynamic "induced search" since government policies and initiatives and firm strategies jointly determine patterns of search for investment opportunities, clients and suppliers, and new business models.
At the same time, upgrading is affected by global and regional demand conditions and expectations. Many industries are dominated by global lead firms whose location and supplier selection strategies significantly affect how and to what extent latecomer economies can participate in existing global value chains and production networks (Ivarsson and Alvstam, 2010;Amendolagine et al., 2019;Pietrobelli and Rabellotti, 2011). In the presence of lead firms, local governments and businesses may engage in so-called "strategic coupling", i.e. various initiatives to set up and coordinate relationships with lead firms to stimulate upgrading (Coe et al., 2004;Yeung, 2009). Local industries may benefit from strategic coupling through knowledge and technology transfer (Amendolagine et al., 2019;Bathelt et al., 2018) or, indirectly, by being pressured to adopt higher standards and/or innovate to be able to meet global demands (Pietrobelli and Rabellotti, 2011). However, both opportunities and constraints of strategic coupling are strongly influenced by strategic choices of lead firms. For example, lead firms may choose to stimulate incremental process innovation among local suppliers, while limiting opportunities for functional or product upgrading (Blazek, 2016;Pipkin and Fuentes, 2017). Yet, in some industries, strategic coupling may not be an option, because global markets are rather fragmented, lacking dominant buyers. In this study I focus on such an industryglobal services sourcing.
Especially in industry contexts where global markets are highly fragmented, local decision-makers face high uncertainty and complexity in terms of where to invest in. Two main perspectives have been dominant in assessing upgrading approaches under these conditions. One stream of research emphasizes the importance of making deliberate choices as to which market segment to target. Typically, studies differentiate between global mainstream and niche markets (e.g. Van Grunsven and Hutchenson 2016, Brandt and Thun 2016). Upgrading into mainstream capabilities within a particular industry requires meeting global standards, cost efficiency expectations and scalability. By comparison, upgrading into more niche markets, such as regional market segments, requires a focus on more location-specific resources and capabilities and specific knowledge of niche customer demands, whereas scalability and global competitiveness become less important (see e.g. Van Grunsven and Hutchenson 2016, Brandt andThun 2016, Keijser et al. 2021). Opting for a particular segment may entail path-dependent processes (Martin and Sunley, 2006), e.g. in terms of investment in training and government support. However, joint efforts of governments and businesses may also promote the creation of new development paths , focusing on new industries or industry segments.
Another stream of research is much more skeptical about the notion of "deliberate choice" and instead conceptualizes upgrading as a longerterm process of trying to "reach a better deal" (Ponte and Ewert, 2009;Mann and Graham, 2016). This dynamic may involve experimental processes of upgrading and downgrading (Blazek, 2016;Blazek et al., 2019) as well as de-coupling from global production networks (Horner, 2014), as a response to global cost and competitive pressures (Werner, 2016). Decoupling and downgrading, for example, may imply abandoning highly competitive markets and moving either into lower-cost production (Ponte and Ewert, 2009;Blazek et al., 2019) or smaller market segments (Blazek, 2016). This perspective on local industry development as a continuous learning process has been triggered by the observation that initial attempts of upgrading often fail, for example because regions lack the skill base to attract global clients and lead firms (Cammett, 2007); because lead firms switch to lower-cost locations (Altenburg and Meyer-Stamer, 1999;Altenburg et al., 2008;Moreira, 2006;Jenkins et al., 2007;Foster and Azmeh, 2019), or because upgrading efforts lead to the crowding out of local firms, and hinder rather than stimulate economic development (McGrath, 2018). However, our understanding of when and how governments and businesses shift from one upgrading approach to another is still rather limited. Whereas the "deliberate choice" perspective underconceptualizes processes of learning, many studies that follow a more experimental view on upgrading say little about the conditions under which a collective shift from one approach to another is likely to occur. To better conceptualize these important dynamics, this study employs the concepts of value regimes and economic imaginaries from the political economy literature (Jessop, 2010;Levy and Spicer, 2013;Levy et al., 2016). Value regimes are defined by Levy and Spicer (2013, p. 673) as "the broader political-economic settlement linking an imaginary with specific set(s) of technologies, production methods and market structures". Value regimes thus link economic imaginaries with materialized structures and relations. Economic imaginaries refer to shared cognitive frameworks or dominant discourses that legitimize certain practices, economic structures and institutional orders (Jessop, 2010). Part of these frameworks are shared "theories" of success and value creation. Importantly, imaginaries denote both present frameworks and future orientations, and therefore not only reflect present structures but also guide future investments (Levy and Spicer, 2013). In a later conceptualization, Levy et al. (2016) further specify value regimes as consisting of economic models of value creation and distribution, normative and cultural values, and governance mechanisms. These configurations turn into value regimes as they stabilize and mutually support each other.
In the context of upgrading, both global mainstream and niche approaches can be thought of as alternative value regimes, as they form around distinct investments into particular capabilities, economic relations and principles of value creation. Importantly, however, in order for them to come into being and stabilize as specific configurations, value regime studies suggest that supportive imaginaries are required that are shared by critical governance actors. Levy and Spicer (2013) demonstrate this with the case of "climate imaginaries" enabling and constraining investments into certain technologies and economic practices. These dynamics, however, may be disrupted when substantial contradictions arise between dominant imaginaries and the material realities of economic relations on the ground. Such misalignments open windows of opportunity for competing imaginaries to come forward and attract new coalitions of actors (Levy and Spicer, 2013). Imaginaries and material dimensions of value regimes are therefore recursively inter-related, driving stability at certain times, and driving change in others. Based on this general understanding, the micro-dynamics of shift in upgrading will be examined in more detail below. First, the specific context of upgradingglobal business serviceswill be introduced.

The case of upgrading in global business services
Global business services are an industry that emerged from the trend of global services sourcing, starting in the mid-1990s. Initially, especially large firms from the U.S. and Western Europe with administrative overhead experimented with sourcing business processes from abroad to cut costs, including IT infrastructure, finance and accounting, and customer services (Doh, 2005;Lewin and Peeters, 2006). This trend was facilitated by decreasing costs of long-distance communication and coordination thanks to advanced information technology (Metters and Verma, 2008) along with an increasing digitization, modularization and standardization of business processes (Sinha and Van de Ven, 2005;Blinder, 2006). These factors in combination created market opportunities for specialized service providers, initially mainly from the U.S. and India, who took on various tasks from clients by hiring qualified, yet low-cost labor to perform these tasks from globally dispersed locations (Athreye, 2005;Ethiraj et al., 2005;Manning et al., 2008). Importantly, unlike suppliers in manufacturing, which tend to specialize in intermediate products within particular product value chains, business service providers typically specialize in corporate functions, such as Call Centers or HR, that are in demand across industry boundaries (Sako, 2006). This is one reason why their client base is often rather diverse and not industry-specific . However, despite a growing global provider industry, many especially larger clients also operate their own service centers abroad, creating economies of scale and scope across product lines. Both governance models continue to co-exist .
From the very beginning, this industry provided employment and growth opportunities primarily in developing countries, due to the availability of a growing pool of qualified labor at lower costs in these countries. In the 1990s and early 2000s, service capabilities developed in particular in India (Arora et al., 2001;Dossani and Kenney, 2007;Patibandla and Petersen, 2002). In later years, however, a growing number of countries across different regions, such as Eastern Europe, Southeast Asia, Latin America and Africa, entered this industry (Kearney, 2004;GlobalServices, 2008;Manning, 2013). While some countries did so quite successfully, such as the Philippines and Ukraine (see e.g. DaxxTeam 2020), others have failed to grow a significant market share or declined in importance (Kearney, 2019). At the same time, this industry has differentiated into various co-existing market segmentsfrom mainstream business services that are dominated by Indian and US-based providers and that target especially large global clients, to a range of more niche players across the world, including specialized software services, call centers focusing on specific languages or special customer needs, and local subcontractors of mainstream providers (see e.g. Couto et al. 2008, Kearney 2019, Manning 2013. The level of global spread and differentiation into different market segments makes global business services a fascinating case for studying upgrading strategies and dynamics. One important dynamic that affects the potential of any one country or region to grow a substantial business services industryand thus is likely to affect the success of upgrading strategiesis what Manning (2013) calls the ambivalent effect of commoditization. Most generally, studies have shown that business services, from IT services to call centers and software, have the tendency of becoming more commoditized, i. e. more standardized and thus less firm-, user-or even industry-specific (Langlois, 1992(Langlois, , 2003Davenport, 2005). As a consequence, a potentially increasing number of clients across industries and regions take an interest in sourcing these services from qualified suppliers. This has made upgrading into so-called "mainstream" services, including tech support and call centers, an attractive option for job creation in emerging economies. However, high degree of commoditization also lowers switching costs for clients, thus leading to high levels of global competition for client projects.
At the same time, various more "niche" service offerings keep emerging that are more client and location-specific, which has expanded the range of upgrading options for emerging economies. One example I detail further below is the practice of "impact sourcing", which focuses on hiring and training of disadvantaged staff for business services jobs, thus combining the busines case of outsourcing with corporate social responsibility (CSR) (Kannothra et al., 2018). This practice in part targets CSR-conscious clients, and builds on alliances with local community organizations that tap into underutilized labor pools . Another niche segment are various forms of local and regional sourcing that specialize in targeting regional clients, based on exclusive client access (Kleibert, 2017;Keijser et al., 2021). A related segment is the growing trend of subcontracting between mostly urban mainstream service providers and rural or suburban vendors specializing in serving a limited number of regional clients. In comparison to mainstream services, those niche segments are less vulnerable to global competitive pressures, since they make more use of location-specific resources .
Next, this study examines in more detail the upgrading strategies in global business services of two countries in sub-Saharan Africa -Kenya and South Africa. Both initially invested into more mainstream business services before gradually shifting towards more niche capabilities (see also Graham 2015, Mann and Graham 2016, Foster et al. 2018, Anwar and Graham 2019. This study examines the micro-dynamics of the initial strategic shift from mainstream to niche, from a value regime perspective. Notably, in more recent years, the industry as a whole has experienced another shift towards platform-based contract work (Lehdonvirta et al., 2019), the so-called "gig economy" Graham, 2020, 2021). However, for the purpose of this study, the focus will be on the initial shift, yet, with interesting implications for future research.

Data and methods
I adopt a qualitative case approach to examine initial drivers of the formation of mainstream value regimes and micro-dynamics of shift towards niche value regimes in the context of global business services. Qualitative methods can be used to explore complex phenomena about which little is known and/or about which a novel understanding is needed (Strauss and Corbin, 1998). Dynamics of upgrading in emerging economies is a phenomenon of this kind.
Specifically, I use a multi-case design (Yin, 2003) in combination with "mechanism-based theorizing" (Davis and Marquis, 2005) to inform future research. This approach focuses on "the wheelwork or agency by which an effect is produced" (Hernes, 1998, p. 74). The notions of "value regime" and "imaginary" are thereby used as sensitizing devices. More specifically, I look at mechanisms leading to the formation of value regimes and promoting shifts between value regimes. I focus on the context of sub-Saharan Africa, which has been described as a promising destination for upgrading, while facing persistent economic, political and social challenges (George et al., 2016;Barnard et al., 2017). Specifically, in the context of global business services, several studies in the early 2000s and 2010s suggested that Africa is growing into an important global hub for such service capabilities (Abbott, 2013). Indeed, in particular Egypt, Morocco, South Africa, Kenya and Nigeria have often been mentioned as emerging business outsourcing hubs (Abbott, 2013;Manning, 2013;Daxx Team, 2020). At the same time, several observers cautioned that the outsourcing industry in Africa might be more of a hype than a real upgrading opportunity, noting in particular the example of Kenya (Kamau, 2016;Graham, 2015;Mann and Graham, 2016). Yet, more recently, studies also suggest that aside from mixed experiences with mainstream outsourcing services, such as call centers, tech support and IT infrastructure (see e.g. Anwar and Graham 2019), sub-Saharan Africa is becoming an important hub for niche markets such as "impact sourcing" (Wright, 2019;Manning et al., 2017; IS, 2019) and regional sourcing (Keijser et al., 2021). This makes sub-Saharan Africa a fascinating case for studying upgrading dynamics in the context of global business services.
Within this context, I selected Kenya, a lower-middle income economy, and South Africa, an upper-middle income economy, as particularly interesting cases of latecomers in global business services (see also Graham 2015; Mann and Graham 2016; Anwar and Graham 2019). Both countries are often listed as the two most prominent examples of sub-Saharan African countries entering that industry (Daxx Team, 2020), also reflecting the growing importance of IT and services industries in both countries (see in detail Table 1) At the time most data was collected, mainly between 2012 and 2014, South Africa had already achieved some considerable success in mainstream outsourcing services, in particular call center services catering to UK clients (see also Anwar and Graham 2019). However, South Africa was struggling to diversify beyond that. By comparison, Kenya was largely failing to get into mainstream services, while showing more success in so-called impact sourcing and regional sourcing (see also Mann and Graham 2016). At different stages of development, both economies were therefore interesting examples of emerging economies facing both opportunities and challenges of entering an increasingly established market.
I used multiple sources of evidence to increase the robustness of findings, including 26 semi-structured interviews of about one hour each with industry experts, government actors, CEOs and managers of service providers, along with industry reports, government programs and prior

Industries
Agriculture accounts for 34% of GDP (Statista, 2020a). Kenya is a lead exporter of black tea and cut-flowers, coffee and vegetables. Kenya's manufacturing sector is rather underdeveloped, accounting for only 16% of GDP (Statista, 2020a). Services remain the largest income-generating sector, contributing 43% to the country's GDP (Statista, 2018). While tourism remains the largest service domain, so-called "ICT-enabled services", including mobile banking and business services, have been on the rise, with an estimated 8% contribution to GDP and 250, 000 jobs by 2020 (Mwangi, 2019). This is also why in particular the capital Nairobi has become known as a thriving startup ecosystem especially around mobile applications and tech services (Gakweli, 2020).
Manufacturing is relatively small, providing 13% of jobs and 15% of GDP. Services, including tourism, financial services, real estate, and business services are by far the largest employer (72%) and income-generating sector (67%) (Statista, 2020b). Within global business services, South Africa has been specializing mainly in call center work for UK clients ( Harris, 2012) studies. I carried out data collection in two rounds. The first round was conducted in Kenya in 2012. The main objective at that time was to better understand Kenya's potential to grow into a global outsourcing hub. I conducted 13 semi-structured interviews with managers of service providers (6), policy-makers (5) and industry experts (2) from local universities. In selecting interviewees, I sought to capture both firm strategies and government approaches to entering and competing in this industry. Firm selection was facilitated by the Kenya Information and Communication Technology (ICT) Board. During this trip I also came across the emerging practice of impact sourcing. Four out of the six providers I interviewed had adopted an impact sourcing model at that time. The second round of data collection in South Africa (in 2014) followed the principle of 'theoretical' replication' of the first round (Yin, 2003). That is, both case contexts are comparable by showing important similaritiesboth are sub-Saharan African countries known for business services, while also showing some theoretically interesting differencesin particular the fact that South Africa at that time succeeded more in mainstream business services than Kenya did. Like in the first round, I again conducted 13 interviewswith service providers (7), representatives of government and a major industry association (3), and other industry experts (3). All interviews were transcribed verbatim. For data analysis, I first did a cross-tabulation of responses across cases, focusing on the interplay of government and firm choices in coselecting upgrading strategies. I paid special attention to how key players in business and government framed "value" and upgrading success, both initially and in later years. Second, I used timelines and interviewees' specific narratives to better understand the formation of mainstream vs. niche imaginaries and how these imaginaries legitimized selective investments in certain local resources and capabilities, while opposing the use of resources that do not fit the dominate imaginary. Third, I identified three important phases in the upgrading processes: (1) Initial alignments of local investments with global demands in line with a mainstream imaginary; (2) Misalignments that called the mainstream value regime into question; and (3) Realignments around niche imaginaries. First, however, I give a brief outline of the history of global business services in both Kenya and South Africa.

The history of global business services in Kenya and South Africa
Fig. 1 displays a timeline of major events affecting the development of the global business services industry in Kenya and South Africa. The timeline presents both government and private sector initiatives, focusing on events between 2000 and 2017. In the timeline, events in regular font are associated with the mainstream value regime, whereas events in italic font are associated with the emerging niche value regime.
Since regaining its independence in 1963, Kenya has gone through a volatile economic history. After an initial growth period until the mid-70 s, a combination of rising oil prices, ineffective government interventions in the private sector, and lack of export incentives contributed to an economic decline in the 1980s and 1990s (Kimenyi et al., 2016;Meilink, 1982;Thugge et al., 2009). With a new political leadership starting in 2003 (under Mwai Kibabki), the Kenyan government formulated an "Economic Recovery Strategy for Wealth and Employment Creation" (see timeline), focusing on stimulating economic growth, private sector infrastructure initiatives, improved education and poverty reduction (Thugge et al., 2009).
Following a rather healthy recovery trend, the government then developed a long-term strategy in 2007 called "Vision 2030 ′′ . The vision was built on three pillarseconomic, social and political. The objective of the economic pillar was to "maintain a sustained economic growth of 10 percent per annum for 25 years", emulating the success of the "Asian tigers" (Thugge et al., 2009). As part of this strategy, six priority growth sectors were selected: tourism, agriculture, wholesale and retail trade, manufacturing, financial services, and business process outsourcing (BPO). The government's ambitious goal was to make Kenya the "Top BPO Destination in Africa" (see timeline; Thugge et al. 2009, Mann andGraham 2016). Concrete targets included: creating 7500 direct BPO jobs through a BPO park and other initiatives by 2012, attracting at least ten large multinational captive operations and five leading IT suppliers. This plan was further fueled by the installment of undersea fiber optic cables in 2009 (Graham, 2015;Mann and Graham, 2016). Alongside this national agenda, the Information and Communication Technology (ICT) Board was set up to implement the Vision 2030. A leading member of the ICT Board remembered:  In the early and mid-2000s, a lot of enthusiasm was shared by both government actors and businesses around the prospects of mainstream BPO services in Kenya. Private sector initiatives in this domain began around 2002 with small local firms offering remote service jobs to foreign firms (Mann and Graham, 2016). Later, several key outsourcing service providers started operating in Kenya. Most prominently, celebrated entrepreneur Nik Nesbitt launched Kencall in 2005, winning multiple awards for his call center in the coming years (see timeline; Mann and Graham 2016). Other important entrants included Techno-Brain (a multi-service provider) and CraftSilicon (a provider catering to financial service firms). One CEO remembers: However, in the early 2010s, several niche players, such as Digital Data Divide (DDD), also set up businesses in Kenya, especially around the business model of "impact sourcing" (IS), supported by NGOs and community partners. IS in Kenya included the hiring and training of youth from urban slums or people with impaired hearing for tech support and data entry jobs as a way to generate opportunities in disadvantaged communities. One representative of an impact sourcing service provider (ISSP) explains their approach: "So we hire through […] NGOs that work in the slums. And then the other set are, hearing impaired, the deaf. And again there's an NGO you know for the deaf that we hire through." However, IS models did not initially fit the mainstream agenda of the Kenyan government. Interestingly, despite the government's initial ignorance of this sub-sector, the practice of IS actually began with business model experiments in particular in Kenya and South Africa. The US-based Rockefeller Foundation was a leading initial supporter of these early experiments (Browning, 2019). Over time, however, IS grew into an important market segment, with a global market share of around 12% in 2014 (Everest Group 2014;Impact Enterprises, 2015). Following this trend, several local outsourcing firms, such as TechnoBrain, started adopting IS models. After initial reluctance to recognize this trend, the Kenyan government later supported IS initiatives more actively in their business promotion policies.
The early 2010s also experienced a shift in the local BPO industry from serving global clients to focusing more on local and regional clients, e.g. governments (Mann and Graham, 2016). This shift was in part stimulated by the government around 2012, which adopted a more niche-oriented view on BPO upgrading, after the feasibility of developing a mainstream BPO hub was called into question (see also Mann and Graham 2016). Struggles of mainstream BPO businesses, including the formerly celebrated entrepreneur Nik Nesbitt, stimulated this shift in thinking. Below, the specific mechanisms of shift from a mainstream value regime to a niche value regime in the case of Kenya will be discussed in greater detail.
By comparison, South Africa's history of outsourcing services began in the mid-1990s. From that time into the mid-2000s, mainstream, highly standardized call center work, through providers such as Aegis, Capita and Teleperformance, established itself as a relatively successful sector, especially around Cape Town (Harris, 2012;Anwar and Graham, 2019). A combination of supply and demand contributed to this. Availability of qualified, low-medium cost English-speaking staff with time zone proximity and cultural affinity with the UK proved to be particularly appealing to clients in the UK (Harris, 2012): "So traditionally, we focused on the UK and the decision was made several years ago, also more at the national level to focus in terms of linage on the value proposition of the region in terms of attracting the UK market. So things like time zone, compatibility, […] cultural affinity with the UK compared to say the United States, so that was the target market and traditionally we have always focused very strongly on that." (Government official) The South African government supported this industry with the setup of the industry association BPeSA (Business Process Enabling South Africa) in 2002 and the launch of a BPO park in 2007 (see timeline). Furthermore, the government launched a tax incentive program in 2011 that subsidized each additional call center job created. The focus on call center work was further stimulated by the perception that South Africa would not be competitive enough in IT and software work, as well as other "non-voice" business services, which were dominated by Indian providers: "India is out of the picture. I think India does so well on back office non-voice, and I do not think anybody can compete with them on that.
[…] But I think on voice culture affinity South Africa is got to be a good option for the UK." (CEO of BPeSA) Interestingly, however, while the upgrading focus of the government and lead businesses was on mainstream call center work, various niche initiatives emerged. Notably, in 2008, the Monyetla work readiness program was launched (supported by the government) to provide entrylevel BPO skillsvery much in line with the idea of impact sourcing. Similar initiatives followed, such as Harambee, which is supported by Rockefeller Foundation (Anwar and Graham, 2019). In the following years, these initiatives would form training and recruitment alliances with multiple BPO firms (see below). At the same time, several South Africa-based BPO firms began shifting from voice work to non-voice work around 2012-2014, including a growing interest in legal process outsourcing. Yet only in the mid-2010s, the government would increasingly acknowledge such niche market opportunities. Indicative of this shift is the active participation of BPeSA in the newly formed Global Impact Sourcing Coalition in 2016. Next, the dynamics of shift from mainstream to niche value regimes in Kenya and South Africa are examined in more detail.

Initial alignments: mainstream imaginary and value regime
To understand the dynamics of upgrading in global business services in both Kenya and South Africa, it is critical to capture the interplay between value regimes and imaginaries. In order for both mainstream and niche value regimes to be formed and sustained, I find that two types of alignment are critical: (1) the alignment between local resources/ capabilities and global/regional demand and expectations, and (2) the alignment between local-global economic exchanges and a coherent imaginary of upgrading. The latter helps channel investments into local conditions that are in line with the imaginary, while hindering investments that are not. The importance of imaginaries stems from the fact that upgrading is a continuous process, linking past to potential future investments that are guided by dominant imaginaries.
I find that decision-makers in both Kenya and South Africa initially favored upgrading into mainstream business services, guided by what I call a "mainstream imaginary" (see Fig. 2). This imaginary focused on upgrading into business services in line with global standards in order to provide scalable work and services at globally competitive prices and quality. One major driver for both the Kenyan and South African government to initially focus on mainstream business services was their promise of scalability and significant job creation to combat high levels of unemployment. Local resources, such as labor, and capabilities, such as local business models, were seen as potentially compatible with global demand for mainstream business services. At the same time, local resources and business models that deviated from this imaginary were seen as distractions and obstacles in achieving mainstream success.
The promise of the mainstream. In the case of Kenya, the mainstream imaginary was embodied by the BPO vision 2030 that was formulated by the government in 2007 (see above; Fig. 2). The notion that African economies like Kenya can successfully participate in mainstream global business services was widely shared among various international organizations (Friederici et al., 2017) and consulting groups (e.g. Kearney 2004, GlobalServices 2008. Investing into mainstream service capabilities was recommended as long as a qualified English-speaking workforce was available at relatively low costs, and as long as governments has the appropriate legal and technological infrastructure in place, i.e. they could "provide […] tax incentives and telecom providers can deploy connectivity" (Vashistha and Khan, 2008). Fueled by this optimism, the Kenyan government was confident in emulating India's success in this industry. The secretary of Kenya's "Ministry of Information and Communication" put it this way: "I joined this ministry in 2005, and at that time we were thinking well, we have so many youths who are highly qualified, we could do exactly what India is doing." Many business executives initially bought into this vision, sharing the optimism around adopting widely accepted business models for Kenya. One executive put it this way: " [Kenya] got the people, it's got the infrastructure, it's got the business environment, it's got the technology, it's got the head start, it's got companies like [us]. There are not companies like this really in Uganda, Tanzania and I do not know, Zambia, Zimbabwe, but we are the only ones really. So with the three, four companies like us, I think [we are the] head and shoulders […] compared to anybody else in Africa." In line with the mainstream paradigm, Kenya's government envisioned the BPO industry as a "mass employer" that could hire large amounts of (otherwise unemployed) graduates. This quote by a representative of the Kenya ICT boards illustrates this logic of thinking: "Everything here is designed to address […] the problem of unemployment. That is why BPO is important because it's a mass employer.
[…] We also have a strength in the BPO space […] because we have a lot of graduates every year who are unemployed with the degrees, and so if they get [special training] [….] they will be fit to work in any BPO center." Similar to Kenya, the South African outsourcing industry also initially favored a mainstream approach to building this industry, focusing on scalable services and attracting big brands. More specifically, South Africa focused on establishing voice call for UK clients, and on scaling up employment for these types of services (Harris, 2012;Anwar and Graham, 2019).
Part of this business model has been a focus on attracting established outsourcing firms from India and Europe to set up call center operations in South Africa to serve UK clients. One appealing aspect for the government in particular has been to demonstrate to the media and public South Africa's ability to attract "big brands" and to thereby establish the country as a "global player" in this business. This quote from the CEO of BPeSA illustrates this: For the South African government, being able to demonstrate success in stimulating economic growth through direct job creation was a major incentive to focus on standardized and labor-intensive mainstream services (such as voice call). The following quote from the BPeSA CEO illustrates this way of thinking: "We created 3000 jobs last year, 3000 jobs the year before, and the potential is there for, yeah, amazing growth from a South African point of view.
[…] Almost 100% contact center. So the recent growth is being predominantly through telco and retail companies from the UK." To further stimulate job creation according to this logic, in 2011, the South African Department of Trade and Industry created an incentive program by financially rewarding every additional job that is created and maintained for at least three years. The objective of this program was also to make the South African BPO sector more cost-competitive at the global level (Harris, 2012). This program was well-matched with the needs of call center operators, many of whom would further scale up operations with the help of this subsidy. The neglect of local specificity. The mainstream imaginary that guided government programs and investments of local businesses led to a neglect and partial ignorance of resources and capabilities that did not match mainstream expectations. This was true in both Kenya and South Africa in particular for the emerging niche market of impact sourcing. In Kenya, for example, IS was already growing as a business model when the mainstream imaginary was still dominant. Even though the mainstream model was visibly failing (see also below), many perceived IS as a threat rather than as a strategic opportunity for Kenya. This is because it conflicted with the idea of making Kenyan providers compatible with mainstream global demands. This quote by the CEO of a mainstream provider illustrates that: "So what you are appealing to now in the corporation is more or less sentimental part of why they should move to Africa than it is perhaps to the pure economics.
[…] It definitely hurts us, because then it promotes Africa as a low-end outsourcing thing. And all the people pick up on it because they really care about helping people when it comes to Africa at the moment." Similarly, in South Africa, the focus on attracting and serving big, established firms has long been a barrier for more entrepreneurial niche players. For example, even though IS firms had been on the radar of the government, the idea of supporting them would contradict the established focus on "big brands". The following quote from a South African government official illustrates this: "I mean you can have impact sourcing, […] but, it has not built up any significant amount of support politically or in terms of our project focus. We are still looking to attract the big brands to more established companies rather than following the [impact sourcing] model." Likewise, from the viewpoint of businesses, the focus on offering voice call services to major UK clients as the most acceptable market focus has prevented mainstream providers to experiment with IS practices. Especially, the perceived inability of IS staff to meet global client expectations, in terms of language and communication skills, has led providers to ignore IS opportunities: "We could certainly use them for non-voice a lot because obviously they are not then talking over the phone and there is no culturalization issue. [..] There is all sorts of foundational skills. There is cultural training. There is listening skills. There is questioning techniques. There is understanding. It goes around all the different aspects of the soft skills for call centers." (Provider manager) However, the mainstream imaginary was not only an obstacle for IS to take off. Other service providers, whose business model was less focused on scale, found it difficult to survive. For example, one successful firm in South Africa offers specialized advisory services to individual clients with health problems. By utilizing local expertise on various health issues, including obesity, this company has been able to recruit highly qualified staff whose services are in demand globally. However, for a long time, this company stayed "under the radar" as it neither met the profile of mainstream UK voice-call services nor did it specialize in high-volume contracts.: The situation only changed when both governments and businesses perceived a growing misalignment between the core goals and assumptions of the mainstream imaginary and the perceived (in)ability of local businesses to serve global mainstream demand, as I detail next.

Growing misalignments: Mainstream value regime called into question
The historical outline of Kenya's and South Africa's business services industry (see Fig. 1) suggests that both economies experienced a shift over time from more mainstream-oriented services to a stronger focus on niche market segments (see also Graham 2015, Mann and Graham 2016, Anwar and Graham 2019. Previous studies indicate that this shift was driven mainly by global competitive pressures, e.g. by early movers India and Philippines (Mann and Graham, 2016;Kleibert and Mann, 2020;Anwar and Graham, 2019). These studies further argue that the Kenyan government in particular was overly optimistic about the promises of global connectivity (Kleibert and Mann 2020;Friederici et al., 2017), which promoted a growing disconnect between hope and reality (Graham, 2015). At a closer look, however, my findings suggest that the shift in orientation from mainstream to niche in both Kenya and South Africa happened through a combination of particular mechanisms, which are highlighted next.
Local conditions vs. global expectations. Findings indeed suggest that there was a growing misalignment between global/regional demand and expectations, and local resources and capabilities, in particular in Kenya. For example, local businesses underestimated that Kenya as a sourcing destination was lacking acceptance among many global clients. These quotes by business executives in Kenya's outsourcing industry illustrate this dilemma.
"For the demand side, there were definitely challenges.
[They] came from unfamiliarity and ignorance in many ways, and unawareness. So the businesses in the States and the UK, they had no idea that Kenya was a destination." (CEO of Kenyan provider) "So I think the big challenge to the Kenyan industry is that it is just not sophisticated enough, and it does not have established players.
[…] because the barriers to entry appear to be pretty doable.
[yet] the barriers to entry to develop a very good business are actually high, but people can't see that. That's kind of a challenge." (CEO of Kenyan provider) Importantly, the inability of local businesses to attract global clients, despite having a suitable technical infrastructure, qualified personnel and a competitive cost structure, conflicted with the assumption of the mainstream imaginary according to which meeting global standards can drive upgrading success and global competitiveness.
The problem of scalability. Likewise, the assumption that building mainstream capabilities while lead to scalable job opportunities conflicted with realities on the ground, especially in the case of Kenya. While both government and businesses perceived scalability as a benefit of mainstream upgrading, they failed to acknowledge scale as a condition for services to attract global clients. In other words, certain "scale thresholds" will need to be met in order for global customers to consider contracting with Kenya-based providers. In practice, especially at the beginning, none of the service providers in Kenya had the capacity to compete with counterparts in India and the Philippines based on scale. Because of this, and because of the highly standardized nature of typical mainstream BPO services, Kenyan providers would largely fail to secure any significant client contracts in this market segment. The following quotes illustrate that: "I know we have had some inquiries from large Indian BPOs who saw that Kenya might be cheaper, but they wanted to have a larger talent pool available.
[…] They want let's say 2000 software engineers, it's not possible. They will never find them, and then they go away disappointed." (Manager of provider operating in Kenya) "[You need] a certain amount of scale. I mean we are small in the BPO sector. [….] They are not giving us work for 1000 people, but if they give us work for 50 people then we can have 50 people." (Manager of provider operating in Kenya) By comparison, when it comes to providing scalable services, South Africa had more success in offering large-scale contracts, especially with voice-call services. However, the commitment to scale came at a significant cost, as mentioned above. In particular, those businesses providing more knowledge-intensive services were disadvantaged by not benefiting as much from scale-based tax incentives by the government.
Reputation and professionalism. Finally, both Kenya and South Africa underestimated global reputation effects and the differentiating perception of "professionalism" in mainstream business services. This can be illustrated in particular by the case of the Philippines as growing global competitor with a significant reputation advantage over sub-Saharan Africa. In the case Kenya, for example, one business executive emphasizes the (underestimated) role of reputation: "The reputation of the country […] is important for some clients. And there's a whole set who we never even meet because they are not even willing to consider. And even amongst those who we do meet there are some who say […] I do not care that the Philippines has separatist movements and whatever, because they know how to structure the outsourcing sector there and I know that I'll be safe." (Manager at Kenyan provider) Such insights would lead to a fundamental shift in thinking about competitiveness in Kenyan government circles. Whereas previously the government would focus on making the Kenyan BPO industry "competitive" on a global scale, later the emphasis would become to protect Kenya from global competition by carving exclusive market niches, for example by creating local and regional demand. I discuss this niche upgrading strategy further below.
In South Africa, there was greater awareness in government and industry circles right away about the competitive "potential" of South Africa vis-à-vis major industry players such as India. Experts realized that South Africa would not be able to capture market shares in nonvoice IT and software services. Instead, the mainstream outsourcing discourse in South Africa focused on voice call services targeting UK clients. However, especially the emergence of the Philippines as a major global call center destination would be seen as an increasing threat to South Africa's competitive positioning and identity as a UK-serving call center hub. In fact, despite South Africa's time zone proximity, more and more UK clients (like their US counterparts) would choose to contract with professional, low-cost and large-scale providers in the Philippines. This quote from the CEO of South Africa's industry association BPeSA elaborates on this growing concern: "Traditionally South Africa is about voice. It's about voice quality at value, so it's not saying, 'We are cheap,' because we are not. If you want cheap go to India, because it's commercially far more understood, a far more commodity-based environment. Philippines just in the time that I have been here, they have obviously upped their game, they have created a voice proposition that's in the same ballpark as us I would suggest, whereas they always used to service the American market.
[But] the UK market is becoming more accepting of that kind of slightly American Oriental twang that comes out of the Philippines." This perception would increasingly call into question the notion that South Africa's ability to compete in this industry solely relies on building scalable mainstream voice call capabilities. Instead, the outsourcing discourse in this country opened up to a wider range of options.
Findings therefore suggest that the mainstream value regime was called into question only after the major building blocks of the mainstream imaginarymeeting global standards to provide scalable and globally competitive servicesconflicted with perceptions of what it takes to win over mainstream global clients in practice. This would lead to a growing demand for a new imaginary and fitting value regime. In the course of this realignment processes, all of a sudden, previously neglected resources and capabilities, e.g. impact sourcing and specialty services, would come to the foreground and provide templates for a niche re-orientation.

Emerging realignments: towards niche value regimes
Following the growing misalignment between the initial mainstream imaginary and the realities of limited competitiveness of local resources and capabilities in the mainstream global market, both Kenya and South Africa increasingly shifted in their upgrading efforts towards niche value regimes that are guided by a "niche imaginary" whose building blocks are quite different from the prior mainstream focus (see Fig. 2). Whereas the latter emphasizes global standards, scalability and global competitiveness, the "niche imaginary" focuses on location specificity and protecting local businesses from global mainstream players. As detailed next, the findings suggest that local resources and capabilities that were previously neglected under the mainstream value regime now came to the forefront in the framing of niche value regimes as potential assets. For example, the perception of impact sourcing shifted from being incompatible with global demand to being a location-specific niche value proposition. At the same time, new marketsboth in niche areas and local/regionally marketswere identified as equally valuable as global mainstream markets. These realignments would eventually promote new governance and business models, including local partnerships to further exploit local specificity, while relying less on scale.
The shift to niche value regimes was arguably further supported by changes in the global discourse around upgrading in global business services. For example, as mentioned earlier, impact sourcing (IS) gained traction in popularity as a viable alternative business model, as reflected by various consulting reports in the mid-2010s (Everest Group, 2014; Impact Enterprises, 2015), also leading to the foundation of the Global Impact Sourcing Coalition. Also, upgrading into local and regional markets as alternative to global markets has become an increasingly legitimate option (Mann and Graham, 2016;Keijser et al., 2021). However, as detailed below, the formation of niche imaginaries and aligning value regimes in both Kenya and South Africa was also to a great extent a bottom-up entrepreneurial process of "re-imagining", in response to prior challenges with global mainstream upgrading and guided by previously neglected business models. The reframing of upgrading success. Most fundamentally, the shift towards more niche-oriented imaginaries involvedand depended upon -the reframing of "upgrading success". Whereas the mainstream imaginary was dominated by criteria such as "scalability", "standards: and "global competitiveness", the emerging niche imaginaries embraced notions of "specialization", "differentiation", and "high-value adding services". This new vocabulary helped decision-makers develop a new individual and collective identity, which would, in turn, help justify more niche-friendly investment decisions. In Kenya, for example, this involved moving away from seeing India as a role model to appreciating the potential value of locally specific resources and capabilities. The Permanent Secretary of the "Ministry of Information and Communication" in Kenya put it this way: "So the idea was that BPO is what will help us reach our goal of becoming a middle income country, but our thinking about that has really changed and become a bit more sophisticated.
[…] We have done some outsourcing here. Then we realized we could do a strategy [that is totally different] from India. If you want to build a capacity, then build it with your people with your work.
[…] So we changed it from our focus on BPO, on outsourced jobs to creating internal efficiencies, automate our services and outsource locally." At the same time, businesses in Kenya would reframe their own business model by taking on more of a "niche identity" as an alternative to high-volume global services. The following quote by a manager of a Kenyan service provider illustrates that: "We have to be more niche based. You know we are not going to be able to put 15,000 people to work on some data entry project at a minimum cost it's just, it's never going to be possible. So we have to be in things that are a little more specialized, where the market is smaller, where the teams that the client's needs are smaller." The idea of focusing on "smaller", "more specialized" projects was adopted primarily in the community of ISSPs. Instead of focusing on winning large contracts, most ISSPs in Kenya would build their capacity around flexibly employing staff across smaller scale projects, thus creating sufficient scale through diverse, yet specialized offerings. This diversified approach also aligns with the mission of IS to provide rich training opportunities for disadvantaged staff.
In South Africa, there has been a similar trend towards deemphasizing large-scale operations as the primary criterion of success. This can be nicely illustrated with the emergence of legal process outsourcing (LPO) as a niche model. LPO includes legal support services such as document analyses for law firms. One appeal of South Africa as a destination for LPO at that time was access to a large pool of highly qualified young law professionals in need of jobs. Also, the alignment of the legal systems of UK and South Africa provided an important differentiating advantage. Yet, LPO players in South Africa were initially suffering from the pressure to "scale up" in line with the dominant call center regime. Only gradually, the notion of "higher value" services became an acceptable alternative to "high volume".
"It's interesting because it's not the kind of numbers that you would be speaking with other industries. […] If there is a call center, they will be talking about thousands, right. […] But it's a different value proposition. We are moving up the skill set in terms of what BPO can do and as you do that the numbers are going to naturally decrease because you are doing a service that is more highly skilled and so a few people do it. So, creating something at that level of skill with 50 jobs in Cape Town in a country with 28% unemployment is a very exciting thing to be doing as well." (CEO of LPO firm) Over time, the idea of providing higher value services as a way to diversify business service capabilities became more accepted, also at the government level (see also Lago 2019, Anwar and. This government representative explains: "So historically our focus has been on voice, […] like low end voice.
[…] More and more the discussion is now moving into that space around where what are the higher value services that we can start offering from the region and how do we sort of progressively help the industry to mature from just the simple sort of frontline voice to more sophisticated services and I mean we are focusing on legal processes outsourcing.
[…]" (Government representative) The value of local specificity. The increasing importance of specialization and differentiation in niche imaginaries promoted a shift towards valuing (rather than neglecting) locally specific resources. This can be best illustrated with the case of impact sourcing (IS). For example, long before the government in Kenya officially acknowledged IS as an important business model, several impact sourcing service providers (ISSPs), with the help of international funding, set up operations in Kenya. Yet, only with the demise of the mainstream imaginary, IS was increasingly acknowledged as an alternative to mainstream outsourcing. One CEO explains: "Kenya is to some people […] the NGO capital of the world, which means that it is very much of an NGO way of thinking.
[…] Many of the foreigners, expatriates and so on are in that space. And the way they think is different from the way people think on Wall Street […] So they have created this thing called impact sourcing. It's kind of like fair trade, but impact sourcing, where they can make an impact on the communities by outsourcing." ISSPs from the start targeted clients that link their sourcing strategies to criteria of corporate social responsibility (CSR). In doing so, Kenya as a sourcing destination and "development hub" would become a credible choice for particular clients, such as NGOs and social mission-driven firms. IS thus conveys a sense of care that utilizes specific aspects of Kenya's image in the world. This CEO of a Kenyan ISSP describes how clients opt for ISSPs: "So, they care a little. They care to the extent that […] they do not like the idea that they will be sending work to some digital sweat shop. You know they like that our workers see the work as an opportunity to improve their lives and are keen to come to work." (Manager at ISSP) Over time, the Kenyan government began to increasingly incorporate the development aspect of IS to appeal to global clients. For example, whereas initially my interview partners conveyed to us in side conversations that urban slums were mostly ignored as a target of economic impact related to outsourcing, later on, promotion material would incorporate success stories of youth in slums seeking job and training opportunities through IS providers (Rockefeller 2013).
I observed a similar dynamic in South Africa. Because of the initially strong focus on globally competitive voice call services targeting UK clients, many providers in South Africa showed little interest in employing so-called "impact sourcing" staff, many of whom have little formal education and very limited experience in directly communicating with international clients. This quote by a manager illustrates these initial concerns: "So, I think the restriction is have they finished school and can they speak English, […] can they be understood, do they articulate clearly. I do not want to use the word accents because we have all got accents. You have got an accent, I have got an accent, but we understand each other. But if my accent prevented you from understanding me I wouldn't be useful in this conversation." However, with the emergence of alternative niche imaginaries, the one-sided perception of IS staff as a liability shifted towards a more balanced view. Several service providers would now reframe IS staff as a "conditional asset". Specifically, they would acknowledge that IS training and employment can become part of outsourcing models out of South Africa with more differentiated target markets -"regular staff" working for "international clients" vs. IS staff working for domestic and regional clients: "So we can put [impact sourcing staff] into a domestic programs where they would fit in and can actually work within the call center.
[…] Very few of them would be able to make it into the international programs. So it's more the domestic clients where we will put them and we will do the full training." One important dynamic in this reframing process, especially in the context of South Africa, has been the acknowledgement of cost savings of IS hiring practices due to greater loyalty of staff and lower risk of poachingtwo advantages that are often associated with IS, but that were largely ignored when the mainstream imaginary was still dominant. When new niche imaginaries were forming, both government officials and business executives would increasingly frame IS as a potential solution to the immanent risk of poaching and wage inflation in mainstream business services. Especially due to the high concentration of mainstream voice call centers in urban areas along with the initial emphasis on creating many disposable low-paid jobs, the South African outsourcing industry faced a significant employee turnover problem (see also Anwar and Graham 2019). One government official explains: "We have almost been a victim of our own success. [We] have done well like you said to bring Amazon and certainly a lot major special UK clients coming in. And now, it seems to be that the challenge is there is lot of poaching between the operators and the sort of negative impacts the initial investments as had." (Government official) In more concrete terms, outsourcing firms in South Africa would face attrition rates of up to 80% per year, driving up costs of retraining and rehiring significantly. This manager elaborates: "We have an attrition rate of 10% and it's actually the bottom end. In some areas we have got 80%. I am looking at like best scenario, about 10%, which means that industry requires 14,000 almost 15,000 people annually just address the attrition, and that's not taking into account new jobs that are coming in. So we have got a huge skills deficit. So what's in it for a company is that companies need people. There is a cost attached to attrition because when somebody leaves, you to have to recruit a new person, train a new person; and for the first three months, they are not operate at 100% productivity, they are way lower. So it costs the company as we choose people." By comparison, attrition rates of IS are typically much lower. Studies suggest that the annual percentage of IS staff leaving their employer is 20-40% lower than the attrition rate of regular outsourcing staff (Browning, 2019). One major reason is the greater loyalty of staff and stronger embeddedness of IS models in local communities. Based on this observation and following the gradual move away from the sole emphasis on mainstream business services, CEOs would now increasingly see IS as a valuable alternative: "I prefer [IS staff] to be honest because their tenure tends to be longer. Their dedication of work ethic is a lot better." (CEO of service provider) Focus on local and regional markets. The new emphasis on specialization and protecting local players from global competition promoted the active exploration of new local or regional markets as alternatives to global mainstream markets (see also Keijser et al. 2021). For example, Kenya's government and businesses increasingly recognized local and regional governments as potential clients (see also Mann and Graham 2016). Similar to large domestic firms, government agencies were looking to cut operational costs by outsourcing IT infrastructure services and by making their operations more efficient. This was seen as an opportunity to help the struggling outsourcing sector land attractive contracts and generate employment. One government official explains: "We are not going to kill ourselves [to] get this work from the US you know, because there is inefficiencies here [inside the government] which you can take advantage of, and grow, build your capacity, and out of this capacity you can partner, the partnerships with these organizations. It's a totally different strategy." (Government official) "So the idea that if government is really using ICTs, it will create a stronger ecosystem. So it's just about government creates that demand […]. It will help grow the economy.
[…] And then we have government shared services and government applications." Responding to these new market opportunities, Kenyan providers started building capacities to serve domestic clients, particularly governments and public institutions. Government clients had a strong political interest in favoring domestic providers to generate employment and growth. In other words, both government organizations and Kenyan providers jointly discovered outsourcing of public services as a potentially lucrative and protected niche market.
"[The government outsources these services] because they want to improve employment, improve the business environment so that they can support local industries in the area. If an international company wants to be in the local tender, especially a Kenyan government tender, they have to have a tie with a local company." (CEO of Kenyan provider) New business and partnership models. The increasing emphasis on locally specific resources and capabilities as well as the acknowledgement of new niche markets promoted new business and partnership models in support of niche imaginaries and value regimes. Interestingly, these models also enabled a greater scalability of niches. Whereas in the past, niche models were ignored in part because of a perceived lack of scalability, the overall increasing emphasis on niche markets later on stimulated efforts towards developing new business and partnerships models that would combine a niche orientation with scalable training and employment.
This can be illustrated again with the case of IS in South Africa, in which various new alliances emerged in support of IS. Some of these alliances were not specific to South Africa, yet South African government-affiliated organizations, such BPeSA, made sure to become part of them. When IS was introduced as a concept by Rockefeller Foundation in 2011, it was mainly treated as a development initiative that co-evolved with the regular outsourcing industry. However, in 2016 the Global Impact Sourcing Coalition (GISC) was founded by the Rockefeller Foundation and the BSR Sustainable Futures Lab, in collaboration with major ISSPs, Google, Facebook and established industry associations such as BPeSA. In 2019, more than 60 companies were members of GISC representing 1.2 m BPO workers around the world (Browning, 2019). The GISC created a business standard for IS that allows companies to evaluate and assess their hiring models. GISC also connects buyers and suppliers through a comprehensive ISSP directory.
GISC has promoted the exchange of knowledge and best practices in IS, and has helped give further legitimacy to IS as a business practice. It was also instrumental in helping IS get recognized by the International Association of Outsourcing Professionals (IAOP) (Frinton, 2019), which in 2018 launched the Global Impact Sourcing Awards to recognize BPO buyers, providers, and influencers who have successfully promoted this business model. The active engagement of BPeSA as an associate member of GISC as well as the membership of major industry players such as Microsoft and the South Africa-based call center operator Teleperformance have helped promote IS to mainstream service providers and clients.
In addition, BPeSA along with major outsourcing firms has taken a proactive approach in scaling up IS business models through dedicated partnerships with NGOs and community organizations. One central operational challenge of IS, as indicated above, has been the potential gap between the skill set IS staff has and the skill set expected from mainstream providers facing global clients. Alliances with community organizations can mitigate that challenge by using specialization advantages of these organizations in recruiting and training IS staff and in offering "readiness" trainings for various outsourcing jobs.
One example of such an alliance is the partnership between the South Africa-based outsourcing firm Webhelp and the non-profit organization Harambee (Gibson, 2019). Harambee's mission is to partner with governments and businesses in Rwanda and South Africa to tackle youth unemployment through work readiness trainings on behalf of business partners. Webhelp uses Harambee's expertise to get access to a large pool of unemployed youth in townships in South Africa. Combining social impact with benefits of hiring employable and potentially loyal staff at low to moderate labor costs has become a very viable alternative to mainstream practices of hiring and training. Similar alliances include partnerships between the Maharishi Institute and Digital Solutions Group, Vodacom and Harambee, CCI and CareerBox (BPeSA, 2019). Partnerships like this also help develop a skilled IS workforce at scale, which has led to "blended" business models combining mainstream services with a strong social impact dimension.

Discussion: towards a dynamic value regime perspective on upgrading
This study has focused on micro-dynamics of upgrading in a highly fragmented and competitive industry contextglobal business services. Inspired by the concepts of value regime and economic imaginary from political economy (Jessop, 2010;Levy and Spicer, 2013;Levy et al., 2016), this study has examined drivers and components of mainstream value regimes vs. niche value regimes in upgrading, as well as dynamics of shift from one to the other. Thereby, this study has taken a middle road between the idea that local governance actors make deliberate upgrading choices (Van Grunsven & Hutchenson, 2016;Brandt and Thun, 2016), and the idea that upgrading resembles more of an ongoing process of learning and experimentation (Ponte and Ewert, 2009; see also Mann and Graham 2016, Blazek 2016, Blazek et al. 2019, Horner 2014. Based on my findings in the context of global business services in sub-Saharan Africa, I promote a dynamic value regime perspective that takes strategic choices seriously while also giving insights into shifts and dynamics. Next, four propositions are developed to guide future research on upgrading. First, my findings demonstrate how value regimes can emerge in the context of upgrading and how economic imaginaries guide upgrading choices in support of these regimes. Extending prior research on value regimes (Levy et al., 2016;Jessop, 2010;Levy and Spicer, 2013), I argue that in the context of upgrading in highly fragmented global industries, such as global business services, imaginaries become particularly important as mechanisms guiding upgrading choices, because business and government actors cannot rely on specific expectations or demands from lead firms that might co-determine a location's upgrading path (see for the latter, Ivarsson and Alvstam 2010, Amendolagine et al. 2019, Pietrobelli and Rabellotti 2011. Instead, in the case of global business services, local economies often face rather anonymous and diverse global markets composed of clients of different size from across geographies and industries (see e.g. Manning et al. 2017). Imaginaries can reduce this complexity by focusing attention on particular ways of creating value to upgrade into such an industry. My findings suggest that imaginaries in upgrading in general and in global business services in particular specifically concern (a) the perception of the potential value of local resources and capabilities, (b) the perception of global demands and expectations, and the (c) perception of potential alignments between (a) and (b). As such, these imaginaries may guide and legitimize concerted efforts of businesses, government actors and local institutions to develop and invest into local resources and capabilities in line with particular global or regional demands and expectations. This, in turn, may promote the emergence of value regimes in which such investments and alignments stabilize over time.
In particular, I distinguished two types of value regimes along with corresponding imaginaries: the mainstream value regime and niche value regimes. The mainstream value regime focuses on investing in potentially scalable resources and capabilities that meet global standards and that can attract global demand based on meeting these standards. In my empirical context, both Kenya and South Africa initially took this approach specifically by investing into call center capabilities targeting English-speaking clients around the world (mostly UK in case of South Africa) (see also Graham 2015, Anwar and. I find in my interviews that these choices were to a large extent guided by a certain imaginary of upgrading success and underlying value propositions. In Kenya, this imaginary was formulated in the BPO vision 2030 (see also Mann and Graham 2016). In South Africa, this imaginary was the foundation for tax incentives favoring large-scale operations. As a result, investing into immediately scalable jobs, e.g. call center seats, was seen as more valuable than investing into less scalable, but more knowledge-intensive services. Likewise, making call centers and call center staff fit global demand was given priority over exploiting locally specific resources, e.g. impact sourcing staff.
By contrast, later on, both Kenya and South Africa, to a different degree, shifted towards niche value regimes, which are based on a different theory (or imaginary) of value creation and upgrading success. Niche value regimes emphasize the value of local specificity and prioritize protection from global competition. In both Kenya and South Africa, impact sourcing, i.e. the hiring and training of disadvantaged staff for business services, was identified as a potential niche segment which can exploit local specificity while limiting global competition. In addition, in Kenya, both government and businesses identified regional markets as viable niche alternative to global markets (see also Mann and Graham 2016). In South Africa, more knowledge-intensive services, such as legal process outsourcing, were identified to further differentiate from highly competitive mainstream domains, such as voice-based call centers (see also Anwar and Graham 2019). Importantly, I find that the adoption of a new understanding of upgrading success along with the appreciation of local specificity as an asset rather than liability were key in legitimizing the investment into such niche capabilities. In sum, I propose: P1: Upgrading choices in the context of highly fragmented global industries are likely guided by imaginaries, i.e. shared "theories" of value creation in line with perceived global demands and expectations. Such imaginaries may channel investments into certain local resources and capabilities to meet these expectations. These investments help create value regimes that stabilize over time in line with the imaginaries that guide them. This, in turn, leads to a neglect of resources and capabilities that do not fit the dominant imaginary.
My findings further help us understand the drivers of developing certain value regimes and corresponding imaginaries of upgrading, especially in the context of upgrading into highly fragmented global industries, such as global business services. Most importantly I find that in both cases -Kenya and South Africalocal government and business actors initially bought into a mainstream imaginary before embracing more niche-focused approaches. While I am not suggesting to generalize based on just two cases, several other studies on upgrading dynamics do imply a similar pattern (e.g. Blazek 2016, Blazek et al. 2019, Horner 2014. My findings help further explain why and how these shifts happen. As mentioned above, the initial choice to develop a mainstream imaginary of upgrading in both Kenya and South Africa was not the result of exposure to expectations of any particular global lead firm. Instead, I find that two main drivers were important here. On the one hand, I find that the initial optimism around mainstream global business services as an upgrading opportunity was driven by the widely shared view in the dominant global discourse that Internet connectivity, English language capability and a qualified labor force are key ingredients to entering this industry. This optimism was shared by consulting groups (e.g. Kearney 2004, GlobalServices 2008 and international organizations (see for an overview, Friederici et al., 2017). It shows that imaginaries, especially in their initial formation, are often linked to dominant broader discourses from which they derive their legitimacy (see also Jessop 2010, Levy andSpicer 2013). On the other hand, I find that especially Kenyan government actors referred to the success of India in global business services when legitimizing their own approach. Similarly, South African governance actors frequently referred to the Philippines as a core global competitor and peer. This indicates that especially in embracing mainstream imaginaries, mimicking industry leaders and isomorphism play a major role (DiMaggio and Powell, 1983).
In turn, the tendency to follow the dominant discourse and/or leading economies in the industry also explains why decision-makers in latecomer economies may often initially opt for mainstream approaches rather than niche approaches to reduce uncertainty and to gain legitimacy (see also Friederici et al., 2017, Graham 2015, Mann and Graham 2016. In contrast, niche imaginaries, which focus on the value of local specificity, initially lack legitimacy and globally successful templates. "My findings suggested for example that government officials and business leaders in both Kenya and South Africa initially framed staff from disadvantaged backgrounds, and any business models promoting them, as inferior to what global clients ask for. This observation parallels studies in the tradition of postcolonial theory according to which business practices and conditions especially in former colonies are seenboth by local decision-makers and foreign observersas "lacking" and "inadequate" when compared to standards imposed by the Global North (Chakrabarty, 2000;Hughes et al., 2015). I proposeWe propose: P2: Mainstream upgrading choices, which focus on scalability, meeting global standards and achieving global competitiveness, are likely driven by the need to reduce uncertainty and gain legitimacy, especially in the context of highly fragmented global industries. Mainstream choices are initially preferred over niche choices because local governance actors are likely to follow dominant global upgrading discourses and the example of lead economies, both of which feed the mainstream imaginary that guides local investments.
However, my findings also suggest how micro-dynamics may lead to a shift from mainstream to niche value regimes in upgrading, especially in the context of a global industry that is highly fragmented. From a structural perspective, upgrading into highly fragmented industries, such as global business services, on the one hand, induces complexity and uncertainty leading local governance actors to favor mainstream approaches, while, on the other hand, providing a range of alternative niche market entry opportunities, due to their high level of fragmentation (see also Manning et al. 2018). Yet my findings suggest that niche upgrading approaches in this industry are likely to be taken only after initial mainstream upgrading efforts have been made. Similarly, previous studies have shown that economies consider "decoupling" or "downgrading" as an option in reaction to "failed" upgrading attempts (e.g. Blazek 2016, Blazek et al. 2019, Horner 2014) rather than as an initial choice. In this regard, one could argue that the overall legitimacy of niche upgrading has been growing beyond particular countries which may have promoted their adoption (see e.g. Keijser et al. 2021 on local and regional upgrading), yet in the cases of Kenya and South Africa the formation of niche imaginaries and value regimes was very much anchored in collective experiences of the challenges of mainstream upgrading.
More specifically, taking a value regime perspective, my findings suggest that one key driver for niche value regimes to emerge is the growing contradiction between assumptions made about value creation and upgrading success in mainstream imaginaries and material realities on the ground. As a result, the legitimacy of the mainstream value regime is called into question and a window of opportunity opens for niche imaginaries to form and related value regimes to emerge. In the case of Kenya, for example, the perceived mismatch between scale expectations of global clients and the scale limitations of many Kenyan providers was an important trigger for this shift. While the mainstream imaginary conceptualized scalability as an opportunity of upgrading into highly standardized services, it did not anticipate that being unable to scale up operations can be a constraint in winning contracts. Addition, in both Kenya and South Africa, there was a mismatch between the assumption that local skills that meet global standards can attract global clients and the perceived reality that global clients prefer global rivals, such as the Philippines, due to their edge in professionalism. This called the legitimacy of building and maintaining mainstream value regimes into question and increased efforts towards developing alternative imaginaries with a new theory of value creation and upgrading success.
Especially in the case of Kenya, the new focus on local and regional markets in line with locally specific capabilities was a reaction to challenges of reaching global clients (see also Mann and Graham 2016). However, I also identified an interesting dialectic between mainstream and niche value regimes. One important, yet under-researched, property of value regimes is their discriminating effect on resources and capabilities that do not fit the theory of value creation in the dominant imaginary. In case of both Kenya and South Africa, in particular impact sourcing (IS) is an example of a locally specific business model that initially contradicted efforts towards matching global mainstream expectations. Similarly, in South Africa, knowledge-intensive services, such as legal process outsourcing, were neglected in the mainstream value regime because of their limited immediate scalability. Interestingly, it is exactly those services that moved to the core of future niche value regimes. The subsequent success of previously neglected capabilities can be partially explained by the innovation effect of resource constraints. For example, Porter (1990) argued that the success of Japanese businesses in developing innovative solutions (technologies, furniture) to deal with spatial constraints was mainly a result of these very constraints. In other words, certain niche service providers in Kenya and South Africa managed to survive and grow despite opposing dominant norms, thanks to ongoing "experimentation" (see also Grabher, 2007). As a result, they combined resilience with innovativeness which made them highly competitive candidates for alternative niche imaginaries and value regimes emerging around them.
Another example of this dialectic is the growth of new IS partnership models in South Africa once niche approaches became more legitimate.
In fact, such partnership models already existed prior to the emergence of the niche imaginary, yet they evolved largely "under the radar", lacking government recognition and support (see also Manning et al. 2017). The principle of exploiting linkages with communities has a long tradition in economic practice in many countries in sub-Saharan Africa (Barnard et al., 2017;Manning et al., 2017). Yet only when niche imaginaries started to form following the erosion of mainstream models, these partnership models were "re-discovered" as innovative approaches to elevating locally specific resources. In fact, today, such partnership models are even framed as approaches to achieve scalability outside the global mainstream upgrading framework (BPeSA, 2019). I propose: P3a: Niche upgrading choices, which focus on local specificity and protecting local businesses from global competition, are likely taken only after mainstream upgrading is considered as an initial choice. Niche upgrading is likely to be a response to the decreasing legitimacy of mainstream value regimes, following a misalignment between mainstream imaginary and the limited ability of local businesses to meet global mainstream expectations. P3b: Niche upgrading choices that follow mainstream upgrading choices in the context of highly fragmented global industries often center around previously existing local business models that proved resilient and innovative despite prior neglect by mainstream imaginaries, which makes them highly competitive contenders for niche imaginaries.

Further implications
My findings have important implications for future research. Specifically, I discuss how I can inform research on (1) the dynamics of upgrading, and (2) the interplay between government and businesses in upgrading. I also discuss some important policy implications.
First, my study connects the idea that upgrading strategies focus on particular choices (Van Grunsven & Hutchenson, 2016;Brandt and Thun, 2016), while also unfolding as continuous journeys of experimentation (Ponte and Ewert, 2009;Blazek et al., 2019;MacKinnon, 2012;Horner, 2014;Blazek, 2016;Blazek et al., 2019). I add"? to this debate by specifying micro-dynamics that can explain the stabilization of certain choices, but also the shift from one upgrading strategy to another. Specifically, I have argued that a value regime perspective can help better capture these dynamics. Value regimes, in terms of stabilizing configurations of economic models of value creation and distribution, along with normative and cultural values, and governance mechanisms (Levy et al., 2016), manifest in the context of upgrading in terms of alignments of investments into certain local resources and capabilities with global or regional demands for such capabilities. These alignments are based on shared principles and theories of value creation that are anchored in so-called imaginaries (Jessop, 2010;Levy and Spicer, 2013). Importantly, I find that imaginaries are not formed arbitrarily but are linked to dominant global discourses of value creation and upgrading. This has profound implications since it implies that upgrading strategies are both enabled and constrained by shared understandings of what upgrading success means, what upgrading strategies are viable to achieve success and to what extent local resources need to be selected or adjusted to be "useful" in support of these strategies.
More broadly, this implies that socio-political contexts, including global discourses, need to be taken (more) seriously in research on upgrading and global value chains (see e.g. Hughes et al. 2015, Gereffi, 2019Gereffi et al. 2021). For example, my findings suggest that consulting reports and international development policy agendas played an important role in legitimizing initial upgrading approaches (which turned out very flawed) in particular in Kenya (see also Mann and Graham 2016, Friederici et al., 2017, Gereffi, 2019. At the same time, postcolonial theory can be a useful critical lens to better understand why governance actors in many emerging economies make certain upgrading choices to adjust to global demands (Hughes et al., 2015), thereby often ignoringat least initiallythe potential of very viable locally specific resources, capabilities and business models which could be exploited for alternative markets.
However, a value regime perspective can also help illuminate under what conditions shifts to a different upgrading approach are likely. Prior research suggests that value regimes may be questioned when substantial contradictions arise between dominant imaginaries and the material realities of economic relations (Levy and Spicer, 2013). It is thus important for future research to better understand when such windows of opportunity for change open. My findings suggest that these dynamics are neither driven by material relations or discursive processes alone, but by the interplay between the two (see in general also Hughes et al. 2015). This also implies that rather than focusing entirely on negative market feedback as a driver of change (Pipkin and Fuentes 2017;Pietrobelli and Rabellotti, 2011), it is important to understand how (and when) market feedback affects the legitimacy of dominant value regimes. Likewise, this suggests that path dependencies in upgrading and economic development are not driven by specialized investments alone (see e.g. Martin and Sunley 2006, Blazek et al. 2019), but they are also affected by shared cognitive frameworks guiding these investments. This may require future studies to pay more attention to narrative processes and identity work (e.g. Kemp et al. 2012, Pike 2009; see in general Snow andAnderson 1987, Sveningsson andAlvesson, 2003) as antecedents of upgrading shifts in conjunction with changing material conditions. Another important implication of the value regime perspective is the ambivalent effect of dominant value regimes and supporting imaginaries on the neglect of locally embedded resources and capabilities that do not "fit" the dominant theory of value creation. On the one hand, this perspective suggests that once value regimes emerge and stabilize, they do so by selectively investing into certain resources and capabilities to match perceived demand conditions. Thus, the dominance of a value regime always implies the neglect of potential alternative uses of resources in a particular location. This does not imply that these resources are not valuable, but that their value may be underutilized under an existing regime. On the other hand, this very neglect may paradoxically promote innovation through ongoing experimentation "under the radar" (see also Grabher 1997), which can become an important driver of shift towards alternative value regimes once the legitimacy of the established one is called into question. More studies are needed to better understand these important "hidden" dynamics in upgrading processes.
Second, my findings contribute to a better understanding of the interplay of government and business in upgrading processes. Prior research has often viewed this interplay as a "functional" division of labor whereby firms and entrepreneurs drive the development of capabilities, business models and markets (Su and Hung, 2009;Sturgeon and Kawakami, 2011), whereas governments stimulate human capital development and job creation, primarily through financial incentives (Vernay et al., 2018;Horner, 2017;Pipkin and Fuentes, 2017). In doing so, they "induce searches" for business opportunities (Pipkin and Fuentes, 2017).
My findingsyy certainly confirm some of these interactions. Yet, I also make important additional observations. First of all, a value regime perspective suggests and my findings show accordingly that lead actors in both government and business jointly engage in the formation of value regimes and supporting imaginaries. The stabilizing force of value regimes and imaginaries not least derives from how they unite key government and business actors, along with their policies and investments, towards a shared understanding of upgrading success and value creation (see also Jessop 2010, Levy andSpicer, 2013). In other words, rather than viewing business strategies merely as responses to policies, or policies as reactions to business choices, future studies need to pay equal attention to discursive and institutional processes connecting government and business actors. In this regard, intermediary organizations, such as industry associations, become critical as meeting places for government representatives and business leaders. In the case of South Africa, for example, the industry association BPeSA served as a meeting place for discussing upgrading strategies, including the role of impact sourcing. At the same time, future studies need to better understand how businesses and policy-makers outside of these circles potentially develop alternative imaginaries that may become templates for future value regimes.
In this regard, more research is needed to better understand which stakeholders are driving the formation of particular value regimes. Even though this goes a little beyond the scope of this study, my findings indicate that the formation of mainstream value regimes is more of a "top-down process" whereas niche value regimes emerge more "bottomup". As for the former, especially the Kenya case suggests that the initial mainstream imaginary was introduced mainly by governance actors when promoting their 2030 national economic policy agenda. The importance of global discourses in informing mainstream imaginaries further suggests a strong "top-down" tendency in how these imaginaries get adopted. By comparison, in both Kenya and South Africa, nichebased imaginaries formed rather "bottom-up" with a strong impulse of ongoing experimentation of locally specific business models on the ground. The lack of global templates for niche approaches further suggests the relative importance of "bottom-up" circulation of niche ideas and the potential role of entrepreneurial business leaders carrying these ideas. However, further studies are needed to better understand these important dynamics.
This study also has some important limitations. First, this study focuses on the contexts of Kenya and South Africa, which may limit the generalizability of findings. While both contexts were deliberately chosen to represent sub-Saharan Africa's attempt of upgrading into global business services, there is a need for future studies to compare the emergence of value regimes and imaginaries in upgrading across different emerging economies and regions. Second, due to data constraints, the findings presented here rely primarily on interviews and industry reports, whereas "hard data" on the performance and success especially of African companies in global industries such as business services is very hard to access. Even very comprehensive surveys, such as the Offshoring Research Network survey, lacked statistically valuable information on service providers from Africa (see e.g. Manning, 2013;Manning et al., 2017). Third, this study focuses primarily on business models related to global business services until the mid-2010s. Since then this industry has experienced a major shift towards freelance "micro-work" using digital platforms (Lehdonvirta et al., 2019) which is often related to the "gig economy". This trend has also affected Africa's participation in this industry Graham 2020, 2021;Kleibert and Mann, 2020). Future studies need to examine how the gig economy is affected the emergence of value regimes and imaginaries in the context of global business services and beyond.
Finally, my study has important policy implications. First of all, the example of global business services in sub-Saharan Africa showcases how the typical overemphasis of governmentsas well as consulting groupson generic location factors, such as IT infrastructure, low labor costs, English language capabilities etc., can be futile in directing attention to highly mature business segments which are difficult to compete in and which underutilize more locally embedded resources. Governments thus need to be more reflective of the bias towards mainstream business models. Second, learning about what is often called "key success factors" in different business segments, especially at the global level, is crucial before making investments. For example, having a good understanding of requirements in scale, professionalism and branding, as well as being aware of constant cost pressures, is critical especially for upgrading into established markets. Third, governments need to escape the straight-jacket of scalability. While generating employment is certainly a worthwhile policy goal, generating a high number of low-paid, standardized jobs is not necessarily the most effective way. Diversifying into higher-skilled and more locally specific niches may also generate cumulative scaling effects over time, while often creating more attractive jobs that are better protected from both local and global competition.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.