Can foreign aid help the development of local industrial production in poor countries? Studies offer a range of reasons why foreign aid is doomed to fail. Anthropologists highlight the exploitative nature of foreign assistance, while economists emphasize the incompetence of international programs. This paper offers a sociological analysis that identifies specific conditions under which foreign aid can lead to the development and upgrading of local manufacturing. Based on a systematic comparison of local pharmaceutical companies in Kenya, Tanzania, and Uganda, I show that foreign aid contributed to the development and upgrading of a local pharmaceutical industry when it provided three resources in particular: markets, monitoring, and mentoring. When donors were willing to procure local drugs, they created markets, which gave local entrepreneurs an incentive to produce the kinds of drugs donors would buy. When donors enforced exacting standards as a condition to access those markets, they gave local producers an incentive to improve the quality of their products. Finally, when donors provided guidance, it enabled local producers to meet the higher quality standards. Foreign aid has structural limits, however, and it is vulnerable to local conditions; state capacity, in particular, is an important constraint on aid's effectiveness.
Locally owned pharmaceutical companies opened in Kenya, Tanzania, and Uganda starting in the 1980s.1 These companies, which produced generic copies of off-patent drugs, did not draw on novel technologies, and in 2000 they still only made basic products such as painkillers, simple antibiotics, simple antimalarials, and vitamins.2 They also did not follow internationally accepted manufacturing standards. By the next decade, though, some drug companies in East Africa began producing antiretrovirals (ARVs) for HIV and artemisinin-based combination therapies (ACTs) for malaria, following quality standards that surpassed those required by national regulations. Important differences across countries nonetheless remained. While top companies in Kenya and Tanzania broadened their product range and improved their quality standards, in Uganda, local companies did not adopt new practices, with the exception of one company with a unique trajectory.
The production of more complex drugs following better quality standards could lead companies to higher value-added activities, contributing to a country's industrial development. In addition, it could give patients who rely on locally produced drugs access to better-quality medicines, contributing to a country's social development. Still, given established practices, why the upgrading? And how to explain the differences between Kenya, Tanzania, and Uganda? Industrial development is often attributed to state policies, including policies that help with the diffusion of technical know-how, or it is attributed to changed market conditions, such as foreign direct investment (FDI). Here, although the state and FDI did play contributing roles in a few instances, they were not the main trigger. In this paper I show that the upgrading of drug companies in East Africa happened as a result of foreign aid.
In general, the effectiveness of foreign aid is highly contested. Anthropologists, for example, tend to highlight the exploitative nature of foreign assistance, while economists often emphasize the incompetence of international programs. Indeed, just as state policies can be more effective or less, so can foreign assistance. I argue that foreign aid led to the upgrading of local pharmaceutical production when it provided three types of resources: markets, monitoring, and mentoring. Markets create demand and therefore new opportunities that shape the decision whether and what to produce; monitoring functions as a disciplinary mechanism that can assure performance, quality, or other standards; and mentoring provides access to technical know-how required for the upgrading. In the pharmaceutical field in East Africa, when donors were willing to procure local drugs, they created a market that gave entrepreneurs an incentive to produce the kinds of drugs donors would buy. When donors imposed exacting standards on potential recipients of funds, they gave local producers an incentive to meet standards that would have otherwise been ignored. And when donors actively guided local producers in how to meet the exacting standards, although they were unable to entirely block the “low road,” they succeeded in paving an alternative “high road.”
These three resources have much in common with resources provided by a “developmental” or “cohesive-capitalist” state—a state with the capacity and coherence to put in place policies conducive to industrial development (Amsden 1989; Wade 1990; Evans 1995; Kohli 2004). Therefore, I refer to foreign aid that creates pockets of industrial development by providing markets, monitoring, and mentoring as “developmental foreign aid.” Although developmental foreign aid can help promote certain industrial sectors that states do not have the resources, enforcement capabilities, or technical skills to promote themselves, I argue that foreign aid cannot serve as an alternative, but only as a complement, to the state. This is partly due to foreign aid's own internal contradictions, and partly due to foreign aid's vulnerability to domestic conditions that only the state can help overcome.
The next section reviews existing explanations for industrial development and upgrading, and shows why they do not adequately explain the trajectories of the pharmaceutical sectors in Kenya, Tanzania, and Uganda. The paper then turns to the current debate on foreign aid, mostly among economists and political scientists, and, building on the case of pharmaceutical production in East Africa and current insights in the literature, offers a novel sociological perspective. Following a methodological discussion, the main part of the paper describes how foreign aid led to the differences in the upgrading of pharmaceutical companies in East Africa. The paper then analyzes the inherent contradictions and domestic conditions that may limit foreign aid's effectiveness, and the implications this has for the role of the state.
CURRENT EXPLANATIONS FOR INDUSTRIAL DEVELOPMENT AND UPGRADING
What could lead manufacturing firms to produce complex products or pursue stringent standards? The literature on the developmental state has convincingly showed that market forces on their own did not lead to the emergence of successful manufacturing industries (Amsden 1989; Wade 1990; Kohli 2004). Instead, the state had an essential role to play, and state characteristics—including state capacity and level of autonomy from economic actors—were closely related to the level of industrial development a country achieved (Evans 1995). Scholars have also identified the types of policies that triggered industrial development, such as infant industry protection, which, through tariffs and other import prohibitions, enabled nascent local industries to develop before having to compete with established international producers. An essential policy for upgrading was that of conditioned subsidies, in which the state provided subsidies only if the firm fulfilled a set of performance measures (Amsden 1989, 2001).
Kenya, Tanzania, and Uganda were not “developmental states,” with the types of policies that helped other countries develop successful industries. In particular, it is difficult to identify policies or regulations aimed at pharmaceutical companies that could lead to the change in the strategies of local drug companies and that would explain the differences between Kenya, Tanzania, and Uganda. In Kenya, interest in the production of ARVs did originally emerge in response to the government's decision to purchase anti-AIDS drugs, and in Uganda, a new joint venture was established thanks to generous government support. But there was no government involvement in Tanzania, and in all three countries, the drugs that local pharmaceutical companies learned to produce did not necessarily match what the government was willing to procure. It is similarly difficult to explain the upgrading of quality standards by pointing at state policies. While in all three countries quality regulations have become more stringent over time and enforcement has generally improved, these policies could not have led to the pursuit of quality standards that were more stringent than what was required by law. Hence, while state policies did play a role in the trajectory of individual companies, they cannot explain the change in the general orientation of the sectors or the differences among the countries.
A different explanation for a strategic reorientation of a manufacturing sector is the involvement of foreign capital. While the development literature has long debated whether FDI helps countries' economic growth or produces stagnation and social inequality (Amin 1974; Evans 1976; Kosack and Tobin 2006; Moran 2006), empirical studies have emphasized the positive contribution of FDI for industrial production (Castley 1997; Kabelwa 2004). More broadly, the recent scholarship on the “developmental network state,” which emphasizes the role of learning in development, suggests that the state could help with upgrading exactly through connecting local firms to a global network (ÓRiain 2004a; Breznitz 2007; Breznitz and Murphree 2011; Negoita and Block 2012; Mehri 2015). In East Africa, drug companies linked to foreign capital had easier access to finance and, especially if the links were with pharmaceutical companies, access to technical know-how. This influenced in important ways the companies' response to available opportunities and incentives. Still, while local companies owned by foreign pharmaceutical companies were more likely to preserve quality standards as required by law, foreign ownership did not necessarily lead to investment in more complex products or the pursuit of quality standards beyond what was locally required.
Others have drawn on insights from economic sociology (Uzzi 1996; McEvily and Zaheer 1999; Owen-Smith and Powell 2004) to suggest that in developing countries, if small firms are in clusters, they are able to overcome some of the major constraints individual firms usually face, including lack of specialized skills and limited access to technology, inputs, market, information, credit, and external services (Giuliani, Pietrobelli, and Rabellotti 2005; Nadvi and Schmitz 1994). According to this literature, government support institutions, such as public research institutes and training centers, may help firms access new knowledge by creating ties to domestic organizations, rather than only to foreign companies (McDermott, Corredoira, and Kruse 2009). In Kenya, close social relations among drug manufacturers contributed to the diffusion of information and technical know-how across the sector. But the knowledge itself was not indigenously produced. It came from abroad.
In short, factors such as state policies, foreign investment, or knowledge networks impacted local pharmaceutical production in East Africa to some extent, yet these factors cannot explain the decision by top pharmaceutical companies in Kenya and Tanzania, but not in Uganda, to invest in the production of more complex drugs and pursue higher quality standards. Rather, it was a combination of the promise of markets, monitoring of quality, and provision of mentoring—all through foreign aid—that led to changed practices in Kenya and Tanzania; and it is the absence of such resources that explains the lack of change (with one exception) in Uganda.
FOREIGN AID EFFECTIVENESS: THE DEBATE
Foreign aid is a contentious terrain, featured in prominent debates in academic journals, popular books, and op-eds. For skeptics, foreign aid reproduces North–South relations of dependence and imposition (Bornschier, Chase-Dunn, and Rubinson 1978; Wood 1986; Ferguson 1990), while supporters consider foreign aid the only sure way to bring an end to poverty (Sachs 2005, 2015). As for effectiveness, skeptics argue that aid not only fails to reduce poverty, improve human development, or promote democracy, but that foreign aid harms—that it breeds corruption, deters democracy and good governance, and ultimately impedes economic growth and increases inequality (Easterly 2006; Radelet 2006; Wright and Winters 2010; but see Prasad and Nickow 2016). Yet others contend that, through the strengthening of resources, capabilities, and skills, aid brings positive results (Kosack and Tobin 2006; Finkel, Pérez-Liñán, and Seligson 2007; Savun and Tirone 2012). Collier (2007) found that official assistance has accelerated gross domestic product growth among the poorest nations by around 1% per year.
Due to the complex relations between foreign aid and economic growth or other outcomes of interest, statistical evidence on aid effectiveness has been inconclusive (Addison, Mavrotas, and McGillivray 2005; Bourguignon and Sundberg 2007; Asongu and Nwachukwu 2017; but see Doucouliagos and Paldam 2011). One challenge is that some types of aid might only affect growth after a long period of time, making the relationship between aid and growth difficult to detect (Radelet, Clemens, and Bhavnani 2004); another challenge is that some types of aid, especially emergency and humanitarian aid, are likely to be negatively associated with growth, since the aid would increase sharply exactly at a time that growth dramatically falls (Radelet 2006). Wright and Winters (2010:62) conclude that “this research agenda has in many ways stalled amid criticism related to poor identification, self-inflicted endogeneity, and the general limitations of cross-country growth regressions.” In economics and political science, some researchers consequently turned to randomized controlled trials (e.g., Duflo and Kremer 2008). Another response to the inconclusive results of macro-level analysis was a turn to sector-specific studies (Gopalan and Rajan 2016). While suffering from their own limitations (Deaton 2010), these new approaches enable a more careful consideration of not only the effects of foreign aid but also the mechanisms through which foreign aid works or fails to work.
Indeed, although the literature on development in sociology and political science has generally ignored the question of foreign aid, a number of studies have offered convincing evidence for positive outcomes, especially when the inquiry shifts from the economy as a whole to specific sectors. So while the earlier literature on dependence viewed foreign aid as negatively as any other form of foreign intervention (Bornschier, Chase-Dunn, and Rubinson 1978; Wood 1986), the literature on the developmental state has been more cautiously positive when discussing the experiences of South Korea and Taiwan in the 1950s, where the magnitude of foreign (American) aid made it impossible to ignore.3 In both cases, scholars concluded that foreign aid had positive, albeit modest and non-determinant, impact on the countries' long-term economic growth (Amsden 1989; Wade 1990; Kohli 2004; but see Amsden 1985).4 In South Korea, aid improved education, and it helped build infrastructure, mines, and factories (Kohli 2004); in Taiwan, “the effects of aid may be said to have been felt in perpetuity … in terms of a multiplication of civil engineering projects and know-how and improvements in the administrative capability of the Taiwan technocracy” (Amsden 1989:91).
This impression of positive but “minor” impact (Amsden 1985:91) greatly improves if individual sectors are considered. One example is the textile industry in South Korea—with the help of U.S.-subsidized loans, textiles eventually became South Korea's major export item (Amsden 1989; Kohli 2004). Another example is the South Korean construction firms that, by serving as civilian subcontractors to the American forces, obtained surplus equipment, upgraded the quality of their construction work to meet Western specifications, and acquired management and quality-control techniques (Amsden 1989:232).
So, while it is difficult to reach conclusive results regarding the effect of foreign aid on the economy as a whole, historical research has been able to show clear effects when looking at individual sectors. Not all foreign aid is equally effective, however. In the following section I identify three resources that make foreign aid instrumental for industrial upgrading.
DEVELOPMENTAL FOREIGN AID: MARKETS, MONITORING, AND MENTORING
In current debates, foreign aid is commonly measured in terms of volume—so that aid effectiveness is inferred if more aid is correlated with, for example, greater economic growth. But no less consequential than “how much” aid is “what kind”—what the aid is used for, what conditions are attached to it, and what type of guidance is provided. Skeptics like Easterly (2008:25), who call for minimal use of foreign assistance, suggest that funds can be effectively utilized only for modest interventions that “make people's lives better.” Examples include vaccination, urban water provision, and fertilizer subsidies. Aid for productive activities is out of the question. Advocates like Sachs do leave room for industrial development, to be promoted through the “provision of export processing zones [and] industrial parks,” and other supportive policies, including “tax concessions and grants, as well as provision of additional infrastructure” (Sachs et al. 2004).
The experience of local pharmaceutical producers in East Africa suggests that aid for productive activities can be effective, but under more specific conditions than the ones offered by Sachs. I argue that tools similar to the ones utilized by a developmental state to promote industrial production as a whole could be provided through foreign aid to advance a specific sector in an otherwise non-developmental context. (I later discuss the limits on the ability of foreign aid to function in lieu of states.) Drawing in particular on arguments developed by the political economist Alice Amsden (1989) in regard to the developmental state, as well as on current insights in the literature on foreign aid, I identify three interventions that together make foreign aid instrumental in the development and upgrading of industrial production: creating markets as incentives to produce specific commodities; imposing monitorable conditions as incentives to improve production practices; and offering technical know-how through mentoring.
According to Amsden (1989:143), if the first industrial revolution, in Britain, was built on laissez-faire, the second, in Germany and the United States, was built on infant industry protection—based on the understanding that new industries cannot develop without state support to create “market reserve” and in other ways limit early competition with established importers and foreign subsidiaries (Evans 1995; Chang 2002:3). Following this logic, many countries in the global South adopted import-substitution industrialization as the basis for their trade and economic policies starting in the 1950s, leading to enhanced industrial production, although not always to sustainable economic growth.
Foreign aid agencies do not normally have the means available to governments to protect infant industries. But donors can create markets. The provision of commodities by aid agencies can come in the form of (imported) in-kind donations, or in the form of funds earmarked for the purchase of those commodities. In the latter case, donations are often tied to or for other reasons rely exclusively on imported goods. But when donations can be used for the procurement of locally produced commodities, the availability of these funds offers incentives for local producers to invest in the production of such goods, even if these commodities are more complex or in other ways extend local firms' existing range of products. Indeed, international aid agencies have been relying on market incentives not only to encourage the production of known drugs but also to encourage the development of new drugs and vaccines, through “advanced market commitment,” in which donors commit to subsidizing the purchase of an as yet unavailable drug or vaccine, once it is developed (Barder, Kremer, and Williams 2006). By creating markets, then, foreign aid can generate conditions that serve a similar function to protectionist measures—namely, providing incentives for local entrepreneurs to produce.
This was the experience of pharmaceutical companies in East Africa. When donors made funds available for the procurement of ARVs and ACTs after 2000, in some cases the arrangements did not a priori exclude the procurement of locally made drugs. In the context of a small private market and a state with limited purchasing power, these new markets were viewed by local producers as exciting new opportunities.5 Even though donor-funded markets were not normally reserved for local producers, the possibility of a market was sufficient to direct the attention of some local pharmaceutical producers to the production of new drugs.
Infant industry protection—and import-substitution industrialization more generally—risks creating an industry incapable of becoming internationally competitive. Amsden's insight was that to prevent such stagnation later industrialization was based on conditioned support—government use of certain requirements, such as meeting export targets as a condition for receiving subsidies, to assure investment in performance standards (Amsden 1989:145–47, 2001). Amsden's insistence on monitorable conditions as necessary for successful industrial development is key to our understanding of the developmental state. Monitorable conditions are similarly key to our understanding of effective foreign aid. Just as effective state support depends on the ability of the state to make recipients follow performance standards, effective foreign assistance depends on its ability to make recipients meet certain standards.
Importantly, the conditions referred to here are not the same as tied aid or the “conditionalities” attached to structural adjustment loans from the World Bank and the International Monetary Fund (Radelet 2006:13), which involve demands unrelated to the effective use of the loans. Rather, here the emphasis is on performance and other standards that relate directly to how the funds are used. The conditions here are therefore closer in their logic to a more recent type of conditionality, in which countries are expected to reach certain political threshold criteria before they can profit from aid and which aim at incentivizing recipients to reform (Molenaers, Dellepiane, and Faust 2015). They are most similar to programs that employ performance-based financing (Grover, Bauhoff, and Friedman 2018) or cash-on-delivery (Birdsall et al. 2012; Kenny and Savedoff 2014), where donors pay only on the basis of measurable, provable results. In such programs, funds are distributed incrementally, and subsequent installments of funds (for example, for vaccination) depend on the achievement of certain benchmarks with earlier installments—for example, a certain number of children vaccinated (Chorev, Andia, and Ciplet 2011). As this example indicates, monitoring is applicable to services, such as distribution of medicine, but also to commodities, such as the production of such medicine.
In the pharmaceutical sector, standards often focus on quality, since the harm from consuming substandard drugs could be significant (UNIDO 2011:51). In countries like Kenya, Tanzania, and Uganda, however, the standards required by the government are lower than the international standards of good manufacturing practices (GMP) recommended by the World Health Organization (WHO), and enforcement is often lax (Losse, Schneider, and Spennemann 2007). In that context, a monitoring mechanism created by foreign aid had a major impact on the quality standards for locally produced drugs. I show that when donor-funded tenders were open to local producers under the condition of meeting international quality standards, local producers in East Africa began to follow standards beyond the minimum that their respective governments enforced.
The literature on knowledge networks rightly emphasizes the role of learning, but assumes that sources of learning exist or could be developed domestically (McDermott, Corredoira, and Kruse 2009). However, some production—including that of complex drugs following high quality standards—requires technical know-how that is not available in industrializing countries (Amsden 1986:253; Lall 1992:112). Indeed, one of Amsden's (1989) arguments regarding later industrializers, including South Korea and Taiwan, is that they industrialized through learning from others; that they developed skills—including market and technical capabilities—by “borrowing” knowledge from elsewhere (Whittaker et al. 2010:445). Reliance on borrowed knowledge explains why entrepreneurs in developing countries have often been immigrants who arrived with education or experience (Kohli 2004:151), or locals who had been working abroad (Saxenian and Hsu 2001; ÓRiain 2004b). Knowledge could also be acquired by hiring foreign consultants (Amsden 1989; Mehri 2015) or, less formally, through commercial channels, such as suppliers of capital goods (Amsden 1989:233–34; Romer 1993).
I argue that learning can occur—and knowledge can be “borrowed”—with the help of foreign aid. Notably, scholars and practitioners have questioned the effect of foreign assistance aimed at capacity building through direct technical assistance provided by “flying-in experts” or the training of locals (Collier 2007:112), and Riddell (2007:203) reports “a growing consensus” that in terms of efficacy and effectiveness, technical assistance “has largely been a failure.” Criticisms of technical assistance include, in addition to high costs, the sense that experts were better in performing their tasks competently than in successfully training nationals to perform those tasks (Riddell 2007). Still, even the otherwise critical Easterly and Williamson (2011:33) accept that “technical assistance can be well-done and productive in some cases,” and others show that technical assistance has been reasonably successful (Collier 2007; Krasner and Weinstein 2014). Technical assistance, or other forms of technology transfer, are likely to be particularly important in schemes that involve performance-based conditions, as described above. In their description of cash-for-delivery, Birdsall et al. (2012:19) call for “a hands-off approach, emphasizing the power of incentives rather than guidance or interference,” but they do not object to technical assistance, as long as it is the recipient's choice. They reason that “such demand-driven technical assistance has a greater chance of being useful to recipients, and because they selected it, they are more likely to apply it” (24). In countries with scarce know-how, it is reasonable to expect that technical assistance will be both useful and welcomed.
Amsden's analysis of the construction sector in South Korea, mentioned above, is one example of the success of mentoring. Pharmaceutical production in East Africa is another. Pharmacy schools opened in Kenya and Tanzania only in 1974, and in Uganda in 1988, and these schools did not provide adequate training in industrial pharmacy and were therefore not able to address the severe shortage of qualified technical personnel (Wangwe et al. 2014). Hence, the know-how required for the production of new drugs or for maintaining high quality standards was not available locally. But it could be provided through training funded by international development agencies, and it was indeed through such training that local pharmaceutical firms learned how to produce new types of drugs and follow high quality standards.6
In short, commercial opportunities, conditions attached to those opportunities, and guidance provided to help meet those conditions hold the key to foreign aid's impact on industrial production. In East Africa, the promise of new markets generated interest in the production of new drugs, monitoring provided an incentive to produce quality-assured drugs, and mentoring provided the know-how.
RESEARCH DESIGN AND DATA COLLECTION
One of the features that makes the pharmaceutical sector analytically attractive for studying industrial development is that the production of medicine can be as simple as mixing ingredients in one's basement or it can be highly technologically complex—so the sector is particularly suitable for looking at the possibility of upgrading. The pharmaceutical sectors in Kenya, Tanzania, and Uganda are especially appropriate for a comparative analysis, given the difference in the trajectories of the three sectors in similar local contexts. The countries have a somewhat common colonial history, and although their economic and political practices diverged in the 1960s and 1970s—with liberalism in Kenya, African socialism in Tanzania, and a period of military dictatorship in Uganda—structural adjustment reforms in the 1980s and 1990s resulted in a drastic convergence. Today, Kenya is still industrially more developed than both Tanzania and Uganda,7 but the challenges facing the pharmaceutical sectors in the three countries—including small domestic markets, weak legal frameworks, lack of management and technical skills, underdeveloped infrastructure, and heavy reliance on imports for machinery and raw materials—are remarkably similar.
My study of the pharmaceutical sector in each country is constructed largely through the experience of individual companies. To establish the emergence and upgrading stories of each firm (including those that have since closed), I draw on 240 semi-structured interviews I conducted, along with materials from the WHO library and archives, and public documents. The interviews in Kenya (95), Tanzania (37), and Uganda (43) were with founders and managing directors of local pharmaceutical firms; industrial pharmacists working in these firms; importers and distributors of drugs; government officials; and representatives of professional and trade associations, international organizations, development agencies, and local and foreign NGOs. I also conducted interviews in India (25), China (15), and South Africa (2) with pharmaceutical companies that export to African countries, representatives of professional and trade associations, and local and foreign NGOs. Additionally, I interviewed officials at the Global Fund to Fight AIDS, Tuberculosis and Malaria, the WHO, the UN Industrial Development Organization (UNIDO), the German Corporation for International Cooperation (GIZ), and the German medical aid organization action medeor.
LOCAL PHARMACEUTICAL PRODUCTION IN EAST AFRICA
In the 1980s and 1990s, locally owned drug facilities opened in Kenya, Tanzania, and Uganda.8 The Kenyan sector has been the largest of the three. By 2010, the estimated value of the pharmaceutical sector in Kenya was about $103 million (US dollars are used throughout the article), compared to $46 million in Tanzania and $27.6 million in Uganda (UNIDO 2010a, 2010b; UNDP 2016). Kenya had 20 firms producing drugs, compared to 9 in Tanzania and 11 in Uganda. The estimated share of the local pharmaceutical sector in Kenya was 28%, compared to 10–15% in both Uganda and Tanzania (UNIDO 2010a; Wangwe et al. 2014). These differences notwithstanding, manufacturers in all three countries generally produced only basic medicines, such as painkillers, simple antibiotics, simple antimalarials, and vitamins (UNIDO 2011:2), and they only loosely followed quality standards.
Given the existing business orientation of companies in all three countries, there was no reason to expect that any would turn around and start producing more complex drugs or pursue more stringent quality standards. And given the differences between the sectors, if any were to do so, they were likely to be Kenyan rather than Tanzanian or Ugandan. However, in the early 2000s, a number of local drug manufacturers in Kenya and Tanzania, and one company in Uganda with a unique trajectory, started to produce ARVs and ACTs, as well as zinc and low-osmolarity oral rehydration therapy (lo-ORS) for diarrhea. Moreover, some of these companies invested in higher quality standards, pursuing not only local requirements but international and European standards.
Because I argue that this investment in new products and higher quality standards was due not to domestic conditions, such as state policies or FDI, but to foreign aid, I next describe how pharmaceuticals were incorporated into donor-funded development projects in the first place.
How Pharmaceuticals Became “Development”: The International Context
The catastrophic spread of HIV/AIDS devastated many poor countries, particularly in the African continent. In 2003, the estimated prevalence of HIV among adults aged 15 to 49 was 8.4% in Kenya, 7.9% in Tanzania, and 8.2% in Uganda.9 Compounding the effect of inadequate health care services, the cost of medicines was initially exorbitant. In early 2000, antiretroviral therapy for one patient for a year cost $10,000–$12,000.10 The rolling out of medicines in many countries was extremely slow as a result. In Kenya, the government started providing antiretroviral therapy only in 2003. The estimated coverage that year was 5% of people in need of treatment (NACC and NASCOP 2012). In Tanzania, by the end of 2004, only 880 people had received treatment through the public sector. Together with 2,000 patients who received treatment from private facilities and research projects, only 0.6% of those in need of treatment were covered (WHO 2007). Uganda was one of the first African countries to respond aggressively to the HIV/AIDS epidemic, and the government provided access to treatment relatively early. By the end of 2002, about 10,000 Ugandans were receiving antiretroviral therapy, an estimated coverage of 11–17% (WHO 2003).
Pharmaceutical companies could charge such high prices at the time in part due to the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), which was signed in 1994 by member-states of the World Trade Organization and obliged governments to grant lengthy patent protection to pharmaceutical products. Following a worldwide public outcry, TRIPS was eventually interpreted so that it allowed, under certain conditions, the manufacturing of generic ARVs and other medicines under patent (Chorev 2012). With the production of generic ARVs, especially in India, by 2002, the price fell dramatically, to $300 per person per year.11
But even at these much lower prices, poor countries could not afford these drugs. As part of a complex political give-and-take in the negotiations leading to TRIPS and in the course of its implementation, rich countries committed funds to help poor countries fight these diseases—one of the eight UN Millennium Development Goals, for example, was “combat HIV/AIDS, malaria and other diseases.”12 In Kenya, Tanzania, and Uganda, the agencies that most significantly shaped access to medicines were the Global Fund, an international financing institution that was established in 2002 specifically to provide support to countries in fighting AIDS, tuberculosis, and malaria, and two U.S. programs: the President's Emergency Plan for AIDS Relief (PEPFAR) and the President's Malaria Initiative (PMI), which were established in 2003 and 2005, respectively. While donations for the procurement of drugs were not new, the scope was unprecedented. By 2006, the funding provided by donors for the procurement of drugs in Kenya was almost six times as much as that provided by the government, and made up half of the total drug market in the country (MMS 2010:2). The Kenyan government provided only about 5–6% of ARV funding (UNIDO 2010a). In Tanzania, thanks to foreign aid, the funds available in 2007 for the public procurement of drugs were five times those of one year earlier (Häfele-Abah 2010:14). In 2014, the Global Fund in Tanzania accounted for about 70% of ARV procurement (Baran 2016). In Uganda, in 2007, the contributions from the Global Fund, PEPFAR, and PMI to programs for malaria, HIV/AIDS, and tuberculosis accounted for about 60% of the total (UNIDO 2010b). And it is hard to overstate the role these funds played in improving access to medicines. By 2016, the treatment provision rate for ARVs was 64% in Kenya, 62% in Tanzania, and 67% in Uganda.13
Although improved access to medicine became central to the global development agenda, that agenda did not initially translate to an interest in local pharmaceutical production in recipient countries. On the contrary, a 2005 report in a World Bank discussion paper series forcefully argued that “in many parts of the world, producing medicines domestically makes little economic sense” (Kaplan and Laing 2005:iii). But developing countries countered the report's economic reasoning by emphasizing the danger of being entirely dependent on imports for essential commodities, especially as the conditions under which TRIPS allowed generic production meant that drug manufacturers in places like India were less likely to produce such drugs (Shadlen 2007; UNAIDS 2013:5). These arguments allowed developing countries to win a few political battles that opened the possibility of local pharmaceutical production.
To begin with, “least developed” countries were able to postpone the time when they would be obliged to implement or enforce patent obligations under TRIPS with respect to pharmaceutical products, most recently to 2033 (UNAIDS 2013; WTO 2015). This meant that generic drugs that would have been legally challenging to produce in countries that were already fully implementing TRIPS could still be produced in least developed countries. In that context, international organizations also became more tolerant of supportive policies of national governments.14 In addition, some donors have designated markets for local drug companies with the explicit goal of supporting local production. The Global Fund, which did not a priori reject the possibility of purchasing locally made drugs, has recently softened its opposition to preferences for local producers.15
In addition, greater emphasis was given to technology transfer in support of local production. This was another unintended consequence of TRIPS. Among the stated objectives of TRIPS was that “the protection and enforcement of intellectual property rights should contribute to the promotion of technological innovation and to the transfer and dissemination of technology.” WHO consequently emphasized technology transfer in its suggested response to TRIPS (WHO 2011a), and both WHO and UNIDO offered extensive trainings for local drug producers (WHO 2009, UNCTAD 2011). UNIDO devised business plans and roadmaps with the goal of strengthening African countries' “ability to produce high quality, affordable pharmaceuticals across all essential medicines” (AUC and UNIDO 2012), while the German government funded efforts “to develop local capacities for the production and marketing of high-quality drug products at affordable prices” (GTZ 2009).
As this recounting of the international political context makes clear, I consider a relatively broad set of actors and practices under the rubric of “foreign aid”: in addition to donor governments, providers of aid include international organizations and public–private partnerships, and aid can be provided not only to a recipient government but also to for-profit enterprises. It is only by including this entire array of resources that we can convincingly capture the effect aid might have. In what follows, I describe these effects in Kenya, Tanzania, and Uganda.
Kenya: Markets, Monitoring, and Mentoring
When the Kenyan government introduced a modest plan to roll out ARVs in 2003, the expectation was that the drugs would be imported. But the emergence of this new market, however small, was sufficient to lead three of the largest local pharmaceutical companies, which normally produced simple drugs requiring little technological sophistication, to learn how to produce ARVs. By 2002, Cosmos Pharmaceutical Limited, Universal Corporation, and Laboratory & Allied (Lab & Allied) had successfully registered ARVs with the Kenyan Pharmacy and Poisons Board. At the forefront of these efforts was Cosmos, one of the most established drug firms in Kenya, which was started in the 1980s by an Indian-born pharmacist, Prakash Patel, and has remained family-owned. In September 2003, Cosmos announced its plan to participate in a government tender for ARVs (Kimani 2003). Cosmos's bid was acclaimed as a “bold move” that “put Kenya in the league of India and Brazil,” two developing countries celebrated for their local pharmaceutical capabilities (Okwembe 2004). A few months later, Cosmos won a 30% component of a $1 million tender to kick-start Kenya's first public HIV/AIDS treatment initiative (Okwembe 2004).
The ability of Cosmos and other top pharmaceutical companies in Kenya to successfully register ARVs suggests relatively high technological competence. (As we will see, other companies in East Africa relied on foreign assistance or other international sources to learn how to produce ARVs and ACTs.) And the interest among Kenyan companies in learning to produce such drugs greatly increased when the Global Fund dramatically enlarged the market by providing funds to the government for the purchase of drugs.16 A local newspaper reported that the Global Fund donations “encourage[d] more companies … to improve their competitiveness in international tenders to supply antiretroviral drugs” (Otieno 2011). Cosmos and Universal registered additional antiretroviral therapies, in 2005 and 2007, respectively, as well as ACTs.
But the Global Fund and other donors did more than just create markets. As the gatekeepers, they were able to set conditions for access to these markets, including quality standards that were higher than the standards enforced by many governments. PEPFAR and PMI, for example, required that drugs purchased be approved by the U.S. Food and Drug Administration (USFDA). European donors required auditing by European-qualified inspectors. The Global Fund, in turn, accepted “products registered by stringently-regulated authorities” (such as the USFDA) or “quality-assured products under the WHO's prequalification scheme” (Global Fund 2012a).
The WHO prequalification (PQ) scheme—a program to evaluate whether the quality standards of drug manufacturers follow international GMP—played a critical role in legitimating generic drugs produced in developing countries.17 And the PQ scheme did not simply assure quality, it arguably oriented generic producers toward quality manufacturing, since many manufacturers invested in quality procedures when they realized that they couldn't access the international drug market otherwise.18 Additionally, although WHO only prequalified specific drugs, the impact of quality control measures spilled over to other drugs produced in the same facility.19
In Kenya, too, the orientation toward quality manufacturing due to conditions set by donors was clear. Because local drug manufacturers' interest in ARV production was initially in response to government tenders, they were “not worried” about quality at first.20 It was only when companies learned that “to enter the big money of donor-funded programs, you need prequalification” that Cosmos and Universal decided to pursue WHO PQ.21 To participate in German-funded tenders, Cosmos, Universal, and four other companies also pursued higher quality standards through European inspection.
These quality standards were not easy to achieve. Of the six companies that pursued European standards, only three were successful. And only Cosmos and Universal pursued PQ, in large part because WHO GMP would have required the other companies to build entirely new facilities. Most manufacturers also did not have the financial resources or technical capabilities necessary to achieve the requirements.22 Cosmos never did achieve the WHO standards; and it took years for Universal to achieve that milestone.
Universal was a relatively new company—it was established in 1996 by two Kenyans of Indian origin, Palu and Rajan Dhanani, who were joined by Pentti Keskitalo, a Finnish citizen—and the company had a number of advantages over Cosmos and other Kenyan companies. Thanks to Keskitalo, in 2005, Universal received funds from the state-owned Finnish Fund for Industrial Cooperation (FinnFund), which provides “long-term risk capital for profitable projects in developing countries that involve a Finnish interest” (FinnFund 2012). FinnFund, which later held 30% of Universal's shares, in 2009 granted the company $10 million of additional financing “to … develop its operations towards quality certification by the WHO” (FinnFund 2009). It took many years, but in 2011, Universal received WHO PQ for its antiretroviral combination of zidovudine and lamivudine (Mutegi 2011). This made Universal one of only three companies in Sub-Saharan Africa to be granted WHO PQ certification at that time.
But Universal's accomplishment was not due only to FinnFund's financial support. The company gave much credit to the mentoring it had received from development agencies. To begin with, Universal received “hands-on” assistance from WHO.23 “We have gone quite a long way to help them,” said Dr. Rägo of WHO, with technical experts spending weeks “working with them in the unit, going through [everything] they do, coaching them.”24 In contrast, Cosmos failed to achieve WHO PQ partly because the company “hit a lot of bottlenecks” and, unlike with Universal, WHO did not give them “the hands-on [training]” they needed.25 Also central to Universal's accomplishment was technical assistance from the German Technical Cooperation Agency (GTZ).26 German humanitarian agencies were interested in the possibility of procuring drugs in East Africa to distribute in crisis-affected areas in the region, but to be eligible for German funds, the medicines had to be of the same quality as required in Europe. Thus a targeted market was created, and as with the Global Fund, quality conditions were attached. Unlike the Global Fund, moreover, GTZ offered technical assistance to help meet the conditions. Specifically, GTZ offered subsidized training (companies paid for the improvements to their facilities) to help local manufacturers reach the quality level of two international instruments used in Europe, the Pharmaceutical Inspection Convention and the Pharmaceutical Inspection Cooperation Scheme (PIC/S) (Losse, Schneider, and Spennemann 2007). Starting in 2005, industrial pharmacists from Europe worked with local manufacturers to bring them up to international standards.27 That training involved not only the identification of concerns, as in most audits, but also how to solve them. “Inspectors [told] you, ‘Now you have done this, but your gap is here.’ Then they [gave] you the way of how to fill that gap.”28 By the end of a three-year period, participating manufacturers were formally audited, and those that passed received a certificate indicating that the factory had been successfully audited by EU-accredited GMP inspectors.29 Universal gave GTZ's training credit for the company's successful quality upgrading. A local consultant agreed: “It is from PIC/S that Universal got WHO prequalification.”30
In addition to Universal and Cosmos, four other large Kenyan drug companies enrolled in the GTZ project. Three of the four—Regal, Lab & Allied, and Elys—were, like Cosmos, locally owned family businesses which had been established in the early 1980s. The fourth, Beta Healthcare, was established in the late 1930s by an Indian trader, later purchased by the large UK retail company Boots, and in the 1980s sold to Kenyan owners. In 2003 it changed hands again, when it was first acquired by Sumaria, a Tanzanian group, and later by a South African pharmaceutical company, Aspen Pharmacare.31 Like Universal and Cosmos, these four companies pursued PIC/S despite the costs, hoping for access to new markets. As one managing director put it, “Of course there was a carrot dangled, that the companies that go through this will get increased tender opportunities.”32 With the help of mentoring, including guided inspections and trainings (Losse, Schneider, and Spennemann 2007:22), Cosmos and Regal, like Universal, passed the audit. Although the other three companies either dropped out or failed the audit, the very process of striving to meet international quality standards led these companies to upgrade their procedures.
In short, with minimal state support, pharmaceutical companies in Kenya became interested in the production of complex drugs and following international quality standards, thanks to three aspects of foreign aid intervention: the creation of new markets, including for ARVs and ACTs; the condition of quality certification to participate in those markets; and the provision of training that taught them how to meet that condition.
Tanzania: More Aid, Greater Role for Domestic Factors
The creation of conditioned drug markets by donors in Tanzania had a similar effect as in Kenya, and the three largest pharmaceutical companies in the country—Tanzania Pharmaceutical Industries (TPI), Shelys Pharmaceuticals, and Zenufa Laboratories—all invested in the production of new, quality-assured drugs. (TPI was later forced to close, as I describe below.)
The three companies were clearly motivated by the creation of new markets. TPI opened in 1980 as a state-owned company and was privatized in 1997, although the state retained 40% of the shares (Wilson 2009). Once privatized, it went through a “modernization plan,” which included “new design, good layout, separate penicillin production unit, new equipment, [and] new heating-ventilation-air-conditioning system” (UNIDO 1997:11). TPI was therefore in a good position to invest in ARVs and ACTs when a market for these drugs was created. By 2009, TPI had registered three ACTs and four ARVs. TPI's first attainable market was the government—in a way that greatly exceeded the experience of Kenyan companies. In 2006, TPI's sales to the government's Medical Stores Department were 40% of its total sales. A year later, when TPI began selling ARVs, it was more than 80% (UNDP 2016). But when TPI planned a new factory for the production of ARVs and ACTs, it is clearly donations that motivated the investment, as I describe below.
Shelys also planned to produce ARVs and ACTs, because “more donations” meant that “government business increased.”33 A German report similarly suggested that the company was “geared towards the donor market” (Losse, Schneider, and Spennemann 2007:21). Shelys originally opened in 1981 and three years later was bought by Sumaria Group—the same diversified Tanzanian group that owned Beta Healthcare in Kenya. Under Sumaria, Shelys grew to become the largest drug company in Tanzania, with almost 60% of value of local production (UNDP 2016). The company changed hands again in 2008, when the South African Aspen Pharmacare acquired 60% of the company; in 2012, Aspen acquired the remaining 40% of Shelys, and Beta Healthcare. Aspen at the time was the largest generics manufacturer in the southern hemisphere and among the top 20 generic manufacturers worldwide; it was also Sub-Saharan Africa's first and largest generic ARV manufacturer. Under Aspen, Shelys's strategic orientation changed, focusing on fast-selling drugs with high profit margins. This strategy meant emphasis on branded generics for the “emerging, growing middle class,” and a move away from government tenders.34 Shelys consequently abandoned its interest in donor-funded tenders for ARVs (GTZ 2009), but continued to broaden its range of products to take advantage of new markets. With the support of the USAID-funded Point-of-Use Water Disinfection and Zinc Treatment Project in collaboration with the Academy for Educational Development (POUZN-AED), Shelys produced PedZinc, a zinc sulfate monohydrate tablet. This was the first African-manufactured zinc treatment for diarrhea (USAID 2011). Unlike the Global Fund, POUZN-AED was mandated to support local manufacturers, including by identifying suitable markets. Interestingly, instead of relying on donor-funded markets, this initiative relied on a domestic market, which had to be created. POUZN-AED was able to convince the Tanzanian government, following WHO and UNICEF recommendations, to revise its national guidelines for treatment of diarrhea to include zinc therapy and a new low-osmolarity formulation of oral rehydration salts (lo-ORS), which Shelys also produced. POUZN-AED also convinced the government to purchase these drugs for the public sector (USAID 2011:7). In addition, POUZN-AED helped broaden the market by convincing the government to approve zinc as an over-the-counter medicine.
Zenufa was a lot smaller than Shelys, but it was newer, and had a “fantastic facility,” so Zenufa, too, was recruited for the POUZN-AED program.35 Zenufa decided to produce zinc not in tablets, as Shelys did, but in liquid form, which promised better profit margins and could be more popular with children. Because syrup was more expensive, the government would not buy it, so Zenufa only targeted the private market. In addition to zinc, Zenufa worked with the non-profit drug research and development organization, Drugs for Neglected Diseases initiative (DNDi), to produce ASAQ, a fixed-dose combination of artesunate (AS) and amodiaquine (AQ) that DNDi had developed as a simplified antimalarial treatment for children.
In short, in Tanzania as in Kenya, donor-funded markets led local drug companies to produce new drugs. Whether destined for donor-funded or domestic markets, these new products were expected to be produced following international GMP.
TPI, the only company in Tanzania that produced ARVs, had an old factory. Rather than upgrading it to achieve the required WHO PQ, TPI was able to receive financial support from the European Union to build a new factory that would be designed to be WHO GMP-compliant.
Shelys opened a new manufacturing plant in 2006, also with the expectation that it would be “built according to International Standards to meet the WHO-cGMP, US FDA and European PIC/S guidelines” (Losse, Schneider, and Spennemann 2007:22), and Shelys successfully passed PIC/S inspection in 2008. Still, the WHO PQ requirement might have contributed to Shelys's decision not to produce ARVs. An Aspen official in South Africa said, “It's pointless having WHO prequalification in [Tanzania], [if we can] get the products from South Africa, where the standard is there already.”36 Yet when in 2008 WHO began to prequalify zinc products, Shelys nevertheless planned to get WHO PQ for its zinc (USAID 2008), and in January 2010, a WHO GMP team that inspected Shelys's facility found only “not critical” deviations (MOHSW 2011). As of early 2018, Shelys has not yet produced a WHO-prequalified drug, however, and interviews suggest that quality standards in the factory fell short of original expectations.
When Zenufa had just opened, it was “looking to become the first pharmaceuticals company in Tanzania to be cGMP certified by the WHO” (BIO 2007:36). The motivation was the donor market, as WHO PQ would “enable [Zenufa] to take part in international calls for tenders” (36). But Zenufa did not produce ARVs, and because it produced zinc for the private market, there was again no need for WHO PQ. Yet Zenufa sought WHO PQ for the antimalarial drug it produced with the support of DNDi. Although “the journey was ridden with obstacles of a technical, administrative and human nature,” in 2016, DNDi could declare that Zenufa was “completely ready” to apply (SDC 2016). Like Shelys, however, as of early 2018, Zenufa has not yet produced a drug that is WHO-prequalified.
Finally, for both the initial production of new drugs in response to potential new markets and the quality upgrading in response to explicit conditions, Tanzanian drug companies heavily relied on donors also for mentoring. Firms even received help with the creation of markets, as I described above. A number of factors made Tanzania an attractive site for dedicated mentoring. As a least developed country, Tanzania was exempted until 2021, and later 2033, from having to implement patent obligations with respect to pharmaceutical products (UNAIDS 2013), which made it a viable place for the production of newly developed drugs. Also, maybe as a former colony, Tanzania has been of particular interest to German development agencies and medical aid groups, which supported local pharmaceutical production in general, but in Tanzania in particular. Additionally, a “strengthened drug regulatory framework” in Tanzania encouraged allocation of aid, since donors saw it “as an opportunity for enforcement of higher production standards on a sustainable basis” (Häfele-Abah 2010:20, emphasis added).
TPI received extensive technical support from the German medical aid organization action medeor.37 For technical assistance, action medeor partnered with Krisana Kraisintu, formerly head of research and development of the state-owned Government Pharmaceutical Organization in Thailand. Dr. Kraisintu helped TPI with the production of ACTs, in 2003, and ARVs, in 2005, still in the old factory (WHO 2009, 2011b; Mhamba and Mbirigenda 2010). TPI's construction of the new factory was also under the guidance of action medeor (Häfele-Abah 2010). TPI committed $963,000 and provided a plot for the construction, while the EU's regional program Aid for Poverty-Related Diseases contributed $6.6 million, and action medeor contributed $660,000 (WHO 2011b:37; IRIN Africa 2012).
Technology transfer to Shelys was provided to help not only with product formulation and quality standards, but also with management and marketing. In addition to helping with the creation of domestic demand, as discussed above, POUZN-AED also helped Shelys with the marketing of PedZinc, which “was affordably priced … and attractively packaged” (USAID 2011:8), and of the company's brand of lo-ORS, SAVE (Goh and Pollak 2016). Relying on the domestic market meant that donors could not impose international quality requirements as they would with donor-purchased drugs. Still, mentoring improved standards. Indeed, “one of the chief incentives POUZN offered Shelys was technical assistance to achieve [WHO] GMP status” (USAID 2011:6), and Shelys invested about $1 million to implement the recommended changes. In addition, Shelys, like the top companies in Kenya, enrolled in the GTZ training project with the objective of reaching PIC/S standards. (When GTZ started the project, Zenufa had not yet opened.) It was thanks to the guidance of POUZN-AED and GTZ that Shelys significantly improved its quality standards (Mhamba and Mbirigenda 2010; MOHSW 2011).
Mentoring was even more foundational for Zenufa. When Zenufa had just opened, it entered into a collaboration with the Belgian Investment Company for Developing Countries (BIO), a government-funded institution that granted local companies in developing countries long term financing at market conditions. BIO provided Zenufa a $3.5 million loan over eight years, which came with “hands-on” technology transfer: “BIO was … involved at the technical level to … advise promoters [Zenufa] on the technologies being used, and on the financial structuring of the project. BIO also persuaded the promoters to adhere to corporate governance standards … by hiring an established auditing firm and appointing an independent director” (BIO 2007). In addition to BIO, in 2008, POUZN-AED provided Zenufa the same support Shelys had received for both production and marketing.38 But in 2011, Zenufa severed its relations with the Belgian company, reportedly because BIO had “too much control,” including requests for “too many reports” and the imposition of “a lot of restrictions.”39 Around that time, Zenufa lost some of its reputation, and the company was said to be “struggling” (various interviews). But in the following years, the firm showed signs of recovery, including the successful collaboration with DNDi for the production of ASAQ. Again, hands-on mentoring was involved, with DNDi helping with both local registration and WHO PQ (DNDi 2013).
There are important differences in the experience of foreign aid by Tanzanian drug companies compared to their Kenyan counterparts. Foreign aid in Tanzania was more interventionist: development agencies offered technical support for production and quality upgrading, but also for marketing and government regulations. And foreign aid in Tanzania depended more heavily on domestic factors. Most significantly, by creating domestic markets, aid agencies could not require international quality standards and had to rely on domestic regulation instead. The regulatory agency responsible for drug registration, inspection of factories, and post-marketing surveillance, the Tanzanian Food and Drugs Authority, therefore played a critical role. (I discuss below the role of the state in closing down TPI.) Lastly, foreign ownership of drug companies in Tanzania was also more consequential than in Kenya. With the shift in ownership from Sumaria to Aspen, Shelys's orientation changed toward less reliance on opportunities created by foreign aid—and therefore potentially less attentiveness to international quality requirements. Still, in both Kenya and Tanzania, foreign assistance—in the form of promise of markets, monitoring of quality, and provision of mentoring—led top pharmaceutical companies to learn how to produce new drugs and pursue higher quality standards.
Uganda: No Aid, No Upgrading (with One Exception)
In contrast to Kenya and Tanzania, in Uganda, only one company was impacted by foreign aid. This company, Quality Chemical Industries Ltd. (QCIL), from its very inception in 2007 was explicitly oriented toward the Global Fund market. The company was a joint venture between a Ugandan drug trading company, Quality Chemicals Ltd. (QCL), and Cipla, an Indian pharmaceutical firm that was the first in the world to offer affordable generic versions of on-patent ARVs (Chorev 2012). The Ugandan owners of QCL saw a unique opportunity in the new donor markets of ARVs and ACTs, especially since Uganda, as a least developed country, was exempted from some TRIPS provisions, which allowed it to still produce on-patent drugs.40 (In 2013, Cipla acquired additional stakes in QCIL, making it the majority shareholder. The company accordingly changed its name to Cipla Quality Chemical Industries Ltd., or CiplaQCIL.)
Cipla was willing to partner with QCL thanks to unprecedented government support, which, others have argued, was granted due to the close relationships between QCL partners and the Ugandan political leadership, including president Yoweri Museveni. In addition to low-interest loans, land, and access to infrastructure, Cipla was guaranteed a domestic market: the Ugandan government agreed to procure medicines from QCIL worth $30 million per year for seven years.41 This commitment tripled the government allocation to essential medicines at the time (UNIDO 2010b). It was later increased to $40 million and extended to 2019.
Still, QCIL's production capacity exceeded these guaranteed government sales, and the company sought to access the donor-funded markets, which required high quality standards. With technical support from Cipla, the QCIL facility was successfully certified by the International Committee of the Red Cross and DNDi (Klissas et al. 2010). The company also sought WHO PQ, which in 2010 proved to be somewhat urgent, as QCIL wanted to participate in Affordable Medicines Facility-malaria, a Global Fund two-year program that subsidized producers of ACTs selling in the private sector and which was to be launched in Uganda the following year. After some negotiations, WHO agreed to fast-track QCIL's application. In June 2010, WHO prequalified QCIL, under license from Cipla, for antiretroviral lamivudine, nevirapine, and zidovudine fixed-dose tablet preparation, and in December that year for antimalarial fixed-dose combination of artemether and lumefantrine. QCIL was the first African manufacturer to receive WHO PQ to manufacture ACTs.
With WHO PQ, QCIL could venture into new markets. By 2016, while 54% of the company's sales were to the government of Uganda, sales of ACTs to Global Fund programs accounted for an additional 22%, and sales to other Sub-Saharan African governments accounted for the remaining 24% (Uganda Business News 2017). QCIL drugs were distributed in Sub-Saharan African countries directly, by the Global Fund, or thanks to government-to-government arrangements. For example, the Ugandan and Rwandan governments signed a trade and investment framework agreement on pharmaceutical products in 2014, in which Rwanda granted QCIL access to its drug market and QCIL committed to work with Rwanda to build a plant to manufacture essential drugs (MOTI 2014). In 2012, the Ugandan and Kenyan governments started negotiating a bilateral framework agreement on health and medicine. Kenya, which depended exclusively on donor funding for its malaria and AIDS programs, was reportedly willing to create a $28 million fund for purchases from QCIL. Four years later, however, “this [had] not yet come to fruition.” Among other obstacles, “Kenya … challenged the price of drugs from [QCIL], which were 25% to 30% higher than the minimum prices” (Khisa 2015). QCIL's pricing practices, as I describe below, offer another illustration of how domestic conditions impact foreign aid effectiveness.
The unparalleled support of the Ugandan government, combined with the resources and technical know-how that Cipla provided, make the trajectory of this company unique. And with the exception of QCIL, markets, monitoring, and mentoring were not available for Ugandan pharmaceutical companies to the same extent that they were in Kenya or Tanzania. In regard to markets, the Global Fund did not rely on Uganda's National Medical Stores (NMS) for the procurement of drugs, and other development agencies did not offer tailored drug markets to individual companies, as POUZN-AED did in Tanzania for example. As a result, donor markets were inaccessible to even the larger pharmaceutical companies in Uganda, including Rene Industries, which opened in 1998 and has remained a family business, and Kampala Pharmaceutical Industries (KPI), which opened in 1992 and was acquired in 1996 by Industrial Promotion Services, a for-profit company under the Aga Khan Fund for Economic Development.42 While KPI launched a new line of treatment for diabetes in 2008 and new antihypertensive products a year later (CEHURD 2013:12–13), and while both KPI and Rene eventually registered ACTs, neither company tried to sell or prequalify them, and neither registered or tried to produce ARVs.43
Indeed, without the promise of potential markets, monitoring of quality standards through international monitoring schemes had no impact on these companies. And while in Tanzania, companies welcomed development agencies' offers of quality training even when the markets created were domestic rather than international, in Uganda, no such mentoring was offered. UNIDO and GTZ did not provide PIC/S training in Uganda, and other development agencies did not offer targeted mentoring for the production of specific drugs.44 (Thanks to Cipla, QCIL did not need mentoring.) Ugandan companies also did not take advantage of the general training provided by GTZ and UNIDO in Arusha, Tanzania: only one manufacturing company from Uganda, KPI, attended the workshops.45 As a result, there was minimal upgrading in Uganda compared to both Kenya and Tanzania. While Rene and KPI improved their quality standards over time and certainly followed local GMP standards, they did not follow quality standards beyond what was required by Ugandan law. We should not discount the technical difficulties that local manufacturers would have faced if such attempts in quality upgrading had taken place—both KPI and Rene would have required new facilities to successfully follow WHO GMP, for example. Yet in Kenya and Tanzania, foreign aid helped improve quality standards even in cases where the companies did not eventually reach international standards, an effort that did not take place in Uganda.
To summarize, by providing markets, monitoring, and mentoring, donors and development agencies triggered an important transformation of pharmaceutical companies in East Africa: firms produced more complex drugs and pursued higher quality standards. Two companies, Universal in Kenya and QCIL in Uganda, successfully achieved the coveted WHO PQ, while other companies passed European audits. And even companies that failed such inspections, by the sheer attempt to achieve them, learned how to produce new drugs following higher standards than were required by law. As a local consultant insisted in regard to PIC/S training in Kenya, “Not only was it a good experience, it made a [big] jump in the quality” of the factories involved.46
INHERENT CONTRADICTIONS, CONTINGENT SCANDALS, AND THE INDISPENSABLE ROLE OF THE STATE
The initiatives supported by foreign aid were not without challenges. WHO PQ opened a number of new markets for Universal, but, unlike QCIL, Universal did not sell drugs to the Global Fund. QCIL, in turn, has been criticized for its pricing, and TPI was forced to stop production following allegations of selling mislabeled drugs. While the experience of Universal points to the inherent contradictions of foreign aid, the experiences of TPI and QCIL point to the vulnerability of foreign aid to domestic political dynamics—and therefore, I will argue, the indispensable role of the state.
One challenge for foreign aid is that funds are often given to those that already have capabilities and resources, leaving behind those that do not (Briggs 2017). The initiatives and programs that benefited pharmaceutical firms with the resources, knowledge or political connections to take advantage of new opportunities did not reach smaller companies. As a WHO official admitted, “We can't coach those who are on the lowest level.”47 But smaller companies were also those with the most worrisome quality standards. So while foreign aid gave top companies tools to take the “high road,” for other companies, the “low road” remained the only option. Because quality upgrading through WHO PQ or similar schemes was voluntary, while foreign aid helped capable companies pursue higher-value-added activities, consumers still relied on the state to protect them from poorly made drugs.
Those companies that did receive foreign assistance still had difficulty competing in international tenders, even if they successfully maintained international GMP. Part of the challenge was that quality requirements—unlike the conditions Amsden (1989) describes—did not force improved cost-competitiveness. Given the high cost of production in Kenya, Universal could not compete with manufacturers in India or China. Lack of competitiveness risked not only the solvency of the pharmaceutical companies but also their willingness and ability to maintain high-quality standards, since lowering quality standards was an effective way to cut costs. In Kenya, this dilemma was so far avoided by the fact that the production of new types of drugs and quality upgrading gave Universal and other high-performing companies access to alternative markets.48 Universal and Cosmos sold small orders of ARVs to the Kenyan government, and their improved reputation, in turn, helped with sales of branded generic medicines in the private market.49 Most consequentially, international certificates “made a big difference” for non-governmental institutional buyers.50 Universal, Cosmos, and Regal increased their sales to the faith-based Mission for Essential Drugs and Supplies, which provided medicines for many health facilities in Kenya, as well as to action medeor, the International Committee of the Red Cross, and Médecins Sans Frontières.51 Even PEPFAR started buying drugs for AIDS-related opportunistic infections from local sources. As the managing director of Universal asserted, “Because of [WHO PQ], we started getting orders from institutions like USAID and UNICEF” (quoted in ACCI 2016, emphasis added). In the case of Universal, its growing reputation may have also been instrumental in its sale, in 2015, to one of the top Indian pharmaceutical firms, Strides Shasun. Similarly, in Tanzania, upgrading allowed institutional buyers to procure from local manufacturers; even USAID purchased local products, if only “for a couple of million [dollars] a year.”52
Industrial policies may help address the challenges of international competitiveness, and some international organizations have come to support procurement preferences (“discounts”) for local pharmaceutical producers in government tenders (Alexander and Fletcher 2012). But the preferential treatment of QCIL by the Ugandan government was harshly criticized by some, as it seemed to be favoring one company over others. According to a formulation that permitted a 15% markup, QCIL charged the Ugandan government more than the government would have paid for imported drugs. In defending the markup, QCIL officials argued that absorbing higher prices was a way for the government to support local industry as well as the country's health needs (BBC 2012). The chairman of Cipla, Yusuf Hamied, forcefully claimed that rather than an arbitrary markup to enrich the company at the expense of the government, the price reflected the company's cost of production, and was therefore a necessary condition if the government wanted to support local manufacturing of drugs. More generally, Hamied claimed,
We believe that a country should be self-reliant and self-sufficient.… But a country has to pay a price for that. If you put up a factory [in Uganda] … you can't compete in a tender on AIDS [drugs] with China and India. So don't even try. But to be self-reliant, you have to pay a price. So if in the tender the price is, say, one dollar, and [a local] factory can supply you at two dollars, for argument's sake, you should accept.53
Critics argued that the funds would be better spent on cheaper drugs.54 And, in addition to principled concerns, accusations were made against officials at the Ugandan drug procurement agency, NMS, of “corruption, abuse of office, misappropriation, illicit enrichment, plunder and wastage of government resources … in complicity with QCIL” (East African Court of Justice 2015, emphasis added). The acting inspector general of government reported that as of 2010, 84% of drugs delivered by QCIL to NMS were imported from India, and that the markup that was supposed to be applied only to drugs manufactured locally was also applied to the imported drugs. This meant that in 2010, NMS allegedly paid $17.8 million more than it should have to QCIL (ACCU 2013). QCIL denied the allegations. And while these accusations were not resolved, a new accusation surfaced: that CiplaQCIL was selling drugs to NMS at a price that was 36% higher than the prevailing global market price and therefore higher than the statutory 15%. CiplaQCIL admitted it charged lower prices on the export market than when it was selling the same drugs to NMS but explained that it was forced to do so to use excess capacity caused, according to the company, by the government's failure to buy all the medicines it had originally agreed to (Nassaka 2016).
As for TPI in Tanzania, before the new factory had a chance to become operational, a scandal involving the old factory forced the closing of both. It began when in August 2012 mislabeled ARV drugs were found in a government hospital, which, according to the Tanzanian Food and Drugs Authority, had been sold to the government by TPI (Saiboko and Tambwe 2012; various interviews). The authority immediately recalled the ARVs, and stopped the distribution of all drugs manufactured by TPI. TPI directors vehemently defended their company, claiming that they did not have the technology to produce the ARVs identified as theirs. According to the deputy director manager, Zarina Madabida, who was also a shareholder in the company, “the product is simply not ours” (Saiboko and Tambwe 2012). She suggested that the case was an act of political sabotage. At the time, she held a Special Seat in Parliament, as a member of Tanzania's ruling party, Chama Cha Mapinduzi (CCM); her husband, Ramadhani Madabida, was the managing director of TPI, and the chairman of the Dar es Salaam Region for CCM (CCMCentral 2013). She claimed that the accusations were meant to tarnish the company's image at a time when CCM was conducting internal elections, in which both she and her husband were vying for various posts. Her claims seemed to be vindicated when in May 2013 the minister for health and social welfare cleared TPI of any wrongdoing (Mugarula 2013). The company was again allowed to produce. But in February 2014, Ramadhani Madabida and five others were charged with supplying counterfeit drugs, obtaining money by false pretenses, and causing a loss for the Medical Stores Department (Kenyunko 2014). In March 2017, he was among 12 CCM members who were expelled from the party for “sabotage” (Nyanje 2017). Already in 2015, Pharmaceutical Investment Ltd., the consortium through which the Madabidas and other investors owned 60% of TPI, had gone bankrupt (Tanzania Daily News 2015). As of December 2017, TPI was still closed.
In the case of both QCIL and TPI, then, informal political dynamics appear to have informed the companies' actions and/or the government's response. QCIL's Ugandan investors were close to the country's political elite, which helps explain the government's financial support of the company, and might have also allowed the marking-up practices. TPI's investors were part of the country's political elite, which colored both the original allegations and the investors' defense. In both Uganda and Tanzania, without a state agency formally investigating the company's behavior, none of it would have come to light—but in both countries, it is possible that political maneuvers, rather than the rule of law, influenced the final outcomes.
The implications for foreign aid are significant, as these scandals suggest that highly interventionist and relatively effective foreign assistance is still dependent on domestic conditions. Importantly, however, the instances described above show not only the vulnerability of foreign assistance to political dynamics, but also the ability of state agencies, in some cases, to minimize the damage. In that way, they reveal the crucial importance of state capacity for aid effectiveness. And in the context of pharmaceutical production, the role of the state is broader than just containing abuse, as the state is needed to monitor the manufacturing practices of companies, many of them small, that do not seek or no longer follow international approval. Foreign aid, therefore, cannot replace or bypass state functions altogether.
In the context of the AIDS pandemic, struggles over a multilateral agreement to tighten intellectual property protection brought about generous drug donation programs. And it was in order to participate in Global Fund or similar donor-funded tenders that pharmaceutical companies in East Africa learned to produce new types of drugs. In turn, it was the requirement for WHO PQ or similar quality conditions that made local companies invest in quality upgrading. The technical know-how needed for the production of new types of drugs and for quality upgrading was also made available by foreign assistance. In Kenya and Tanzania, developmental foreign aid compensated for instances of inadequate local opportunities (by creating markets), poor regulations (by providing monitoring), and lack of technical know-how (by offering mentoring) in support of local industrial production. In Uganda, most companies were not exposed to foreign aid and therefore did not transform in the same way. Although developmental foreign aid could function independently of any active state support, aid effectiveness clearly depends on domestic institutions.
The combination of markets, monitoring, and mentoring in the pharmaceutical sector was often contingent rather than planned. The Global Fund was not designed to support local industrial production, and the WHO PQ monitoring scheme did not have sufficient resources for comprehensive mentoring, so the mentoring ended up being relatively ad hoc, dependent on bilateral funds. Still, there were cases, including action medeor and POUZN-AED, in which the same entity provided all three resources—promise of markets, monitoring, and mentoring. And in some ways, the field of international aid seems to be moving more consciously in the direction identified in this paper as effective, though not necessarily in a coordinated manner. First, interest in industrial development, which has been neglected in recent decades by “traditional” donors in favor of humanitarian assistance and poverty alleviation, seems to be of growing interest again, as manifested, for example, in the greater focus on economic issues of the UN's Sustainable Development Goals, declared in 2015, compared to the original Millennium Development Goals.55 In addition, new donors, especially Japan and South Korea, emphasize economic infrastructure and production facilities in their aid to African countries.56 Second, aid agencies, including the Global Fund, have come to emphasize performance as a condition for their support, as part of a general shift from aid based on need to aid based on merit (Chorev, Andia, and Ciplet 2011; Birdsall et al. 2012). When such funds are needed urgently—for example, for the provision of vaccines or medicines—stopping them due to failed performance raises difficult ethical issues, though not so much when the funds are aimed at industrial goals. But in both cases, this paper suggests, such conditions are more likely to be useful if they come together with technical support. And while technical transfer has long been central to foreign aid, it should be emphasized much more in the discourse on performance-based grants.