In January 2013 President Mwai Kibaki of Kenya launched a $14.5 billion project to create Africa's first world-class technology city: Konza City. Amidst media fanfare, the president announced that the project would establish Kenya as a major hub in the global software services industry. Drawing on recent research into the development of the Indian software industry, this paper critically assesses each of the four assumptions that underpin the project. Rather than facilitating the creation of a dynamic software services industry, the paper suggests the project may, in fact, retard wider technical progress in the country.
Background
In 2006 the Kenyan government announced that it intended to capitalise on the telecommunications revolution to transform the country into ‘one of the major IT outsourcing destinations in the world’ (State House 2006). By 2008, that intention had crystallised into a National ICT Master Plan, with objectives, time frames and supporting policy. Revised in 2012, the plan now states that by 2018 Kenya will be home to a large and rapidly growing software services industry in which firms, both local and foreign, will provide low-end, IT-enabled services (ITES) such as call centres and data transcription alongside higher-end customised software applications to an array of international clients (Kenya ICT Board 2012, iii). It is an ambitious undertaking but is given succour by the example of India – a country with close links to Kenya – which has established itself over the past 20 years as the developing world's pre-eminent software services industry (Kenya ICT Board 2013).
The $14.5 billion Konza City Project is the starting point and centrepiece of the state's strategy to establish a major software services industry in Kenya (Kenya ICT Board 2011). Located approximately 60 kilometres southeast of the Kenyan capital, Nairobi, Konza City is currently a 5000-acre tract of barren savannah.1 However, by 2017 it is expected that the first phase of a state-of-the-art city, with high-specification business and residential complexes and served by the most advanced telecommunication and transport infrastructure possible, will be built and ready for use.2 It is assumed that such facilities, when combined with Kenya's educated, low-cost, English-speaking labour force and commercially supportive environment, will convince transnational corporations (TNCs) to relocate their call centres to the city (Konza City 2011).
Once ensconced in Konza City, the presence of these call-centre subsidiaries of foreign companies is believed to lead to several further developments which will result in both a broadening and a deepening of the software services industry in Kenya. First, it is assumed that the call-centre subsidiaries will catalyse the emergence and growth of local Kenyan software service firms (Kenya ICT Board 2012, 19–20). Second, it is expected that the call centres will evolve into suppliers of higher-end customised software services (Kenya ICT Board 2010, 1). Third, it is presumed that the clustering of firms, foreign and local, within Konza City will lead to positive spillovers, enhancing the local capabilities and international competitiveness of Kenyan software service firms (Summit Strategies 2011, 13–14).
Surprisingly, despite the high profile of the project and the scale of government resources devoted to it, a critical examination of it and the wider software strategy has thus far been absent.3 The Kenyan media has overwhelmingly supported the state's claims about the desirability of the Konza City Project and endorsed the aforementioned assumptions underpinning the project (East African 2013). Public dissent has been limited to a few ‘maverick’ blog posts (see Akunga 2012; Makau 2012) and some disparaging comments below online articles praising the project (see Chao 2014). However, with the government in Nairobi recently sanctioning additional state investment in the project (Mark 2014), a possible portent of a rapid escalation in its financial commitment, a broader and better-informed debate about the Konza City Project and the positives that supposedly stem from it is urgently required.
Taking this as point of departure, the paper considers the validity of the aforementioned assumptions underpinning the project. It does so by systematically evaluating each assumption in light of original recent research into the mechanisms and processes behind the successful development of the Indian software services industry, the model for, and inspiration behind, the Kenyan state's software strategy. As such, the next section will consider the likelihood of Konza City attracting foreign direct investment (FDI) in the form of call centres. The following section will assess the prospects of such FDI catalysing the growth of local software service firms. The fourth section then evaluates the notion that developing a call-centre industry is a first step in cultivating a more advanced software services industry. The fifth section then examines the notion that clustering will lead to positive spillovers from foreign subsidiaries to local firms. The paper concludes by summarising the key findings and identifying the most salient issues for further investigation. Information on the Indian software services industry is derived primarily from the author's own research (see Saraswati 2008, 2009, 2012, 2013) while the Kenyan context is provided by secondary sources, primarily government reports and media articles.4
Assumption I: The Konza City Project will attract FDI
In 2009 the Konza City Project was initiated with the Kenyan government procuring 5000 acres of land on the main trunk road between Nairobi and Mombasa.5 The location is considered commercially ideal, lying between the country's two major cities and just 50 kilometres from its international airport. Moreover, while construction of Konza is still in its preliminary stage, the plans for the site are nothing short of spectacular. Konza is expected to be at the international cutting-edge not only in term of telecommunications links and commercial buildings, but also in the wider sphere of urban design.6
The assumption is that these investments, when combined with Kenya's well-educated, low-cost, English-speaking workforce and business-friendly environment, will attract FDI in the form of TNC call centres to Konza City. The Kenyan government has therefore been busy showcasing the city to potential investors. The key element of this marketing campaign has been to label Konza City ‘Africa's First Technopolis' and ‘Silicon Savannah’ in a bid to gain the attention of the global business media and, by extension, the international corporate elite (Bloomberg Businessweek 2013; Manson 2013). This gimmick appears to have worked well, with CNN, CNBC and Bloomberg all running documentaries on the project and the Financial Times and Wall Street Journal endlessly churning out Silicon Savannah-themed articles. With the first phase of Konza City due to be completed in 2017, the expectation within the Kenyan government is that TNCs will soon be signing up for slots in the city's business complexes (East African 2013).
The Indian experience, however, suggests that impressive infrastructure, even when combined with a suitable workforce and a commercially supportive environment, will not be sufficient to attract such FDI.
First, the example of India shows that the building of ICT parks did not immediately result in a surge in FDI in the form of call centres. India began to establish ICT parks (in India they are referred to as software technology parks) in the early 1990s, alongside implementing various incentives that constitute a ‘business-friendly environment’ such as tax breaks and import waivers on inputs for exporters of software. However, despite state-funded marketing events around the world to showcase India as a desirable offshore platform for software programming, by 1998 only five TNCs had established subsidiaries in the country (Giarratana, Pagano, and Torrisi 2005, 225).7
Second, the influx of software-related FDI into India from the late 1990s onwards was a result of growing awareness of the export success of Indian software firms (Desai 2005; Giarratana, Pagano, and Torrisi 200 5; ICMR 2004). By harnessing advances in telecommunications technology throughout the 1990s, India's leading software service firms had grown at double-digit rates and had become the most profitable firms in the sector (Singh 2002). This success demonstrated to TNCs that India was a viable export platform for software services and convinced them to establish subsidiaries in the country (ICMR 2004; Saraswati 2008).
As such, the assumption that Konza City will attract FDI is open to question.
Assumption II: FDI will catalyse the emergence and growth of local software firms
If Konza City had been built in the 1990s, it would not have attracted FDI. However, as a result of the success of India as an offshore base for the export of software services, investors in the twenty-first century may now be less wary in establishing offshore subsidiaries than they were at the tail end of the twentieth. As such, the proven success of domestic firms may no longer be a precondition in attracting FDI. TNCs may eventually establish call centres in Konza City. Therefore it is worth examining the second key assumption underpinning the Kenyan state's strategy – that software services-related FDI will catalyse the emergence and growth of local software service firms.
Evidence from India again casts doubt on the validity of this assumption.
First, and as already pointed out, the arrival of FDI into India was the result of the success of India's software firms. By the time the first software-related TNC subsidiary was established in the late 1980s, hundreds of export-oriented Indian software firms were already active (Dossani 2005, 18). TCS had first started exporting software services to the US in 1974, aided by the Indian state's 1972 Software Export Scheme (Grieco 1984; Heeks 1996). And by 1980 there were 21 Indian firms exporting over $4 million a year in software (Dossani 2005, 18).
Second, the real catalyst for the growth of the Indian software service firms was the domestic market (Pingle 1999). This market was peculiarly large as – through an idiosyncratic policy agenda – the majority of computers manufactured in India at the time did not have pre-installed software (in technical terms, they were unbundled).8 It was this large domestic market that allowed Indian software firms to not only expand rapidly through the 1980s but develop the technical skills and commercial acumen required to secure software contracts in the more competitive export markets (Athreye 2004).
Assumption III: Call centres will form the first step in a transition to higher-end software services
While FDI will probably not catalyse the growth of indigenous companies, the Indian experience suggests that it is possible to develop a large export-oriented software services industry through local firms. Encouragingly, some 50 Kenyan software service firms are already in operation (Were 2010). Most are engaged in call centres or some other form of ITES such as data transcription and almost all are small, localised operations (Dihel et al. 2011). Nevertheless, there is a strong likelihood that some of these firms can rapidly develop, both in revenues and in global orientation. This is best evidenced by KenCall, a Kenyan call-centre firm which was serving dozens of international clients (African Millionaire 2011) and registering revenues of $4.7 million (Isenberg 2009) within just four years of starting up.9 It is, therefore, worth considering the third assumption underpinning the Kenyan state's strategy – that call centres will form the first step in a transition to higher-end software services such as customised software applications.10
This assumption is not supported by the Indian experience.
First, while the late 1990s saw a proliferation of Indian ITES firms such as EXL and Daksh (both established in 1999), none of these firms ever migrated into higher-end software service provision. Most went out of business. Those that did not have either remained ITES firms (e.g. EXL) or been acquired by software service firms (e.g. Daksh, which was acquired by IBM in 2004). Crucially, the leading Indian software service firms such as TCS and Infosys did not begin with ITES and then migrate up into customised software services. Rather, they expanded into ITES after they had established themselves in customised software applications.
Similarly, the TNC subsidiaries initially engaged in ITES activity in India did not evolve into providing more advanced software services. The only substantive expansions undertaken by such subsidiaries in India have been horizontal. For example, there are many cases of subsidiaries at first providing only front-desk ITES such as call-centre work but then slowly expanding into back-office ITES such as data transcription. There have also been instances where subsidiaries engaged in back-office ITES have expanded into front-desk ITES. However, the vertical expansion of a subsidiary initially specialising in the delivery of ITES into one also providing customised software applications has been rare. This is because the development of higher-end services typically requires close contact with the firm's headquarters and its major markets, and both of these tend to remain in the West.
Therefore, while Kenya may eventually become an important base for ITES, it is unlikely – ceteris paribus – that this will form the first step in a transition into higher-end software services.
Assumption IV: Clustering will benefit local firms in terms of capabilities and competitiveness
While ITES firms and subsidiaries are unlikely to transition up into providing higher-end software service firms, a large export-oriented ITES industry in the country is a possibility. However, key to this scenario unfolding is the development of local firms into major players in the ITES sector, thereby insulating Kenya from Western TNCs' predilection for continuously shifting subsidiaries from one offshore base to another. Part of the rationale behind centring the Kenyan software services industry in Konza City is the assumption that such clustering will enhance the competitiveness of local firms by facilitating knowledge spillovers with TNC subsidiaries.11 It is therefore worthwhile considering whether this assumption is valid.
Unfortunately, the Indian experience suggests clustering will not only fail to generate knowledge spillovers from foreign to local firms but could actually impede the development of local firms.
First, clusters are unlikely to produce spillovers. Studies have shown that in the Indian context the useful technical or managerial learning between firms (particularly from foreign to domestic firms) did not arise from physical proximity with each other but through local firms taking on contract work with foreign clients (Desai 2005). Moreover, if clustering did induce positive spillovers, one could expect the leading Indian software service firms such as TCS and Infosys to relocate from their existing premises to ICT parks when such parks were established. However, instead of relocating, they had their own premises listed as ‘individual software parks', paving the way for the state to bring the advanced telecommunications infrastructure to them (Saraswati 2008).
Second, the Indian experience suggests that clustering may in fact be detrimental to the growth of domestic software service firms. In the software services industry, retention of employees is critical for firms hoping to increase their competitiveness (Wiggins et al. 2003). High attrition rates create considerable impediments. In so far as proximity amongst firms makes it far easier for employees to job skip, concentrations of firms in the same area may in fact impede domestic firms from moving up the value-chain. IBM, Accenture and other Western software service firms have used their presence in such clusters to help lure away the most capable software employees from the more established Indian software service firms (Saraswati 2012). It is more than likely that should TNCs establish ITES subsidiaries in Konza, they too will poach the best employees of Kenyan firms.
Further research
This paper has not categorically stated that the Kenyan state will fail to realise its software services goal. It has, instead, argued that the assumptions underpinning the Konza City Project, as well as those informing the wider strategy, are not supported by the Indian experience. This suggests that a more open and frank debate is needed about the Konza City Project than has hitherto been the case. While the contours of such a debate go beyond the purview of this paper, four areas for further research and investigation stand out.
First, who is driving the Konza City Project and why? Is it a genuine attempt to promote Kenyan development? Or have vested interests hijacked the process and subverted it for their own economic or political benefit? A favourable starting point would be to examine how the Konza City Project emerged as the centrepiece of the software services strategy which had hitherto relied on already-existing, multi-purpose ICT parks. Furthermore, an investigation into the connections between the construction companies charged with building the city and the government agencies responsible for overseeing it may also yield fresh insights into the political economy of the project.
Second, should developing a global software services hub in Kenya be placed on a par with agriculture and manufacturing in Kenya's 2030 vision? Is it likely to generate developmental benefits and social returns comparable with transformations in the agricultural and manufacturing sectors? The case of India is telling. Notwithstanding the industry's contribution to government finances and an improvement in India's global image, the chief beneficiaries of the software services industry in India have been the well-educated urban classes.
Third, and related, what effect might a dynamic software services industry have on other sectors of the economy? The Indian example suggests that a software services industry might in fact come at the expense of greater IT diffusion throughout the country and, by extension, the wider benefits such diffusion brings.12 This is because the industry has successfully lobbied the Indian government to enforce intellectual property rights (IPR) in software, deeming it an important element in sustaining its global competitiveness (John 2011; Tessler and Barr 1997, 6). Given that pirated software is the chief means by which IT is taken up by firms and individuals in developing countries (Porwal 2009), the relatively strict enforcement of IPR in India has impacted negatively on the uptake of IT in the country. India's international ranking in IT diffusion has been falling since 1998 (UNCTAD 2006) while at a sub-national level the states with software hubs (and therefore the most zealous opponents of pirated software) have seen their IT diffusion rankings fall vis-à-vis other states (Das 2010).
Fourth, what are the genuine policy lessons that might be gleaned from the successful development of software service industries in other countries? Again, a closer reading of the Indian experience is useful. Not only does it undermine many of the assumptions driving Kenya's current software strategy, as well as raise warnings about the potential adverse effects a software services industry could have on IT diffusion, but it also points to several policy initiatives not currently on the government radar. In particular, the Indian example highlights the important role played by domestic firms in driving the growth of a software services industry, the need for government to work closely with such firms in devising effective policy, and the importance of the domestic market in building up firm capabilities which can be subsequently utilised in export markets.13
Summary and conclusion
This paper assessed the Kenyan government's software strategy, focusing on the Konza City Project and the claims that it will help establish the country as one of the world's major software service hubs. Such an examination is important for several reasons. First, it was pertinent given the state's increasing financial commitment to the Konza City Project. In a developing country such as Kenya, where government resources are limited, ensuring beneficial returns on such massive state outlays is vital. A nascent white elephant – an idiom for an expensive yet useless project – should be terminated as quickly as possible. Second, despite the growing cost of project, scrutiny of the government's claims has been almost absent from the discourse. Critical voices have been confined to the periphery and drowned out by a chorus of enthusiastic rhetoric from the government and media. And third, the potential role of IT in fostering economic growth in sub-Saharan Africa is an increasingly prominent theme in development policy discourse. Kenya's strategy is therefore of considerable interest to the wider development community.14
The paper chose to evaluate the Kenyan state's claims regarding the importance of the Konza City Project by comparing and contrasting them with the actual evolution of the Indian software services industry. This comparative analysis was deemed apt for several reasons. First, the software services industry in India is the largest and most sophisticated in the developing world and its experience is therefore particularly telling. Second, India's software success has encouraged the Kenyan government to see software services as an industry with much developmental potential and inspired its policy agenda.15 And third, recent original research has shown that the Indian software services industry developed in a manner very different from that espoused by the World Bank and other Western organisations. Whereas the dominant Western narrative has emphasised the integral role of FDI and the importance of tax breaks for said investors (Miller 2001), evidence indicates that it was domestic firms backed by government support that propelled the growth of the industry.16
The findings of the comparative analysis showed that each of the four assumptions driving the government's software strategy was severely flawed. First, the experience of India suggests that the technological sophistication of Konza City is unlikely to trigger inflows of FDI, even when combined with an attractive pool of labour and a supportive commercial environment. Second, even if such FDI was to arrive, the Indian experience implies it is unlikely to catalyse the emergence and growth of local software firms. Indian software firms developed for other reasons elaborated on earlier. Third, call-centre firms, whether local or foreign, are unlikely to form merely a first step in a transition to higher-end services as such a transition has been absent in India. And fourth, the Indian experience reveals that clustering of foreign and local firms together has, at best, a neutral effect on the capabilities of local firms.
The article also proposed several areas for further research. First, it noted that the political economy of the strategy needed to be investigated, in particular the question of whether or not vested interests were unduly influencing policy. Second, it recommended a study into the direct developmental returns a dynamic software services industry would generate, noting that for all of India's success in the software services industry, the benefits to wider society and economy have been limited. Third, it stressed the need to situate the growth of the software services industry within the wider developmental context, identifying the connection between a thriving software services industry in India and that country's poor rate of IT diffusion. Fourth, it suggested that the Kenyan software strategy could be improved by a closer examination into the experiences of India and other developing countries that have successfully developed software service industries.
To conclude, it is hoped that the findings and suggestions put forward in this paper will encourage greater academic and journalistic scrutiny of the Konza City Project and, by extension, help to refine Kenya's development plan in general and its software strategy in particular.