Academic debate on Rwanda has significant thematic gaps, and does not usually make use of a theoretically informed comparative framework. This article addresses one thematic gap – the distinctive approach of the RPF-led regime to political involvement in the private sector of the economy. It does so using the framework of a cross-national study which aims to distinguish between more and less developmental forms of neo-patrimonial politics. The article analyses the RPF's private business operations centred on the holding company known successively as Tri-Star Investments and Crystal Ventures Ltd. These operations are shown to involve the kind of centralized generation and management of economic rents that has distinguished the more developmental regimes of Asia and Africa. The operations of the military investment company Horizon and of the public–private consortium Rwanda Investment Group may be seen in a similar light. With some qualifications, we conclude that Rwanda should be seen as a developmental patrimonial state.
Different African countries attract different kinds of research. Especially since the 1994 genocide, Rwanda has attracted a large volume of research relative to its size, with special attention to a particular set of topics. Subjects particularly well covered include justice and reconciliation,1 rural poverty and land,2 power and ethnicity,3 the management of knowledge,4 political space,5 and Rwanda's role in the Congo wars.6 This pattern of interest, understandable given the circumstances, has nonetheless resulted in gaps in research coverage, including some topics that usually figure prominently in country bibliographies: the macro-economy, state–business relations, public administration, public services, and political parties.
For related reasons, Rwanda is often also treated as sui generis. It has only occasionally figured in multi-country academic studies focused on generic issues of socio-economic or political development. The focus tends to be more normative than theoretical, with much emphasis on governmental shortcomings in relation to standards of social equity, liberal democracy, or human rights. Lately, one concept from the mainstream of political and social science – ‘high modernism’ – has become prominent in analyses of country policies. This concept is well adapted to supporting a critical view but is not especially helpful in thinking about what distinguishes the Rwandan regime from relevant comparators.7
This article addresses one of the more significant substantive gaps in the Rwanda literature, and does so in the context of a cross-national comparative study with theory-building ambitions. The focus is on a well-known but under-documented aspect of the Rwandan Patriotic Front (RPF) as a ruling party: its approach to political involvement in the private sector of the economy. Central to this is the relationship between the RPF and the private-sector operations of the holding company formerly known as Tri-Star Investments and now trading as Crystal Ventures Ltd (CVL), in which the RPF owns 100 per cent of the equity.
Our analytical framework derives from research led by Tim Kelsall in the Business and Politics stream of the Africa Power and Politics Programme (APPP).8 It draws on a systematic comparison of business–politics linkages across a large set of countries and regimes of South-East Asia and sub-Saharan Africa. The working hypothesis is that while business–politics interactions have neo-patrimonial qualities under most regimes – that is, there is a blending of impersonal/bureaucratic and personal/clientelistic features – it is important to distinguish between more and less developmental forms of neo-patrimonialism.
The article is structured as follows. First we set out the elements of a typological theory which distinguishes what we call ‘developmental patrimonialism’ from the modal type of business–politics interaction in sub-Saharan Africa. We then situate Rwanda's historical trajectory as a country in relation to the modal pattern. The next two sections advance an argument about how present-day Rwanda seems to differ from the modal pattern and to approach the developmental patrimonial type, referring particularly to the relationship between the RPF and Tri-Star/CVL. We then expand the discussion to include the roles of the army's Horizon group of companies and a public–private consortium, the Rwanda Investment Group (RIG). The article concludes that, with some qualifications, the concept of developmental patrimonialism captures and helps to illuminate some distinctive features of Rwanda's current political economy.
Developmental patrimonialism in Africa
In the APPP framework, a country is said to display developmental patrimonialism when the ruling elite acquires an interest in, and a capability for, managing economic rents in a centralized way with a view to enhancing their own and others' incomes in the long run rather than maximizing them in the short run. We hypothesize that developmental patrimonialism represents a more likely scenario for achieving economic transformation9 and social development than the one suggested by orthodox advocates of ‘good governance’.10
We also draw on the arguments of Mushtaq Khan about the necessary role of central management of rents and other ‘primitive accumulation’ in the early stages of capitalist development.11 Khan's political economy is certainly heterodox, but increasingly he is in good company in this respect.12 If he is right, the elimination of rents – and the superseding of patrimonialism – is neither feasible nor desirable until capitalist economies are far more established than at present in most of sub-Saharan Africa. The generation and sharing out of major rents is not only the key to social order and political stability, as argued by Douglass North and his colleagues;13 it is also an essential means of financing the learning costs involved in the acquisition of technological capabilities by an initial generation of domestically owned firms.14
Considerable interest therefore attaches to the question of whether and under what conditions long-horizon rent centralization becomes both attractive and feasible for a country's politically dominant groups and individuals – its political elite. Feasibility, it is assumed, centres upon the ability of leaders to impose the necessary disciplines, first within the political elite itself and then within key elements of the techno-bureaucracy. An important question is what kinds of leaders acquire this sort of ability and under what circumstances?
Most of the instances of developmental patrimonialism identified in a survey of regime types in seven resource-poor tropical African countries were historical examples.15 They belonged to the immediate post-independence era in countries like Côte d'Ivoire (1960–75), Kenya (1965–75), and Malawi (1964–78). The mechanisms used by these regimes to centralize rent management were varied, but they usually entailed a concentration of power around the person of the President or other ‘big man’. There was systematic clientelism and informal use of state resources, and hence a blurring of the distinction between public (state) wealth and the private wealth of the rulers, corresponding to the standard definition of neo-patrimonialism.16 However, as in comparable regimes in Asia (such as Indonesia and Malaysia), rents were deployed in ways that assisted national development; they did not just serve to enrich the big man and his cronies.
The question obviously arises whether regimes such as those of present-day Ethiopia and Rwanda are also to be considered examples of the type. For Rwanda, the answer to this question is not completely straightforward, as blurring of the boundaries between public wealth and the personal wealth of members of the ruling group is not a notable feature of the model. However, developmental patrimonialism covers a cluster of subtypes in which the personalistic element in the centralization of rents is a variable. On this understanding, we argue that there are reasons for including today's Rwanda in the category.17
The article is based on research carried out in Rwanda between 2007 and 2011. Fieldwork included 82 semi-structured confidential interviews or other personal communications, several repeat interviews, site visits to CVL companies, and direct observation of key events, including the November 2010 CVL group management meeting. We benefited from exceptionally good access to nearly all the key players in the recent history of the RPF's business operations, an advantage that researchers in this kind of field have not often enjoyed. We also interviewed disinterested but well-positioned observers, and some with reasons to be resentful about RPF business operations or related government decisions.18 This put us in a position to triangulate information from different and potentially conflicting formal and informal sources, and make judgements on the balance of evidence. In this way, we were able to investigate and validate or invalidate many of the claims – from the serious to the completely scurrilous – which circulate as Kigali gossip or on the internet.
Our analysis is focused on two questions. First, what features of the RPF-led regime correspond to the model of developmental patrimonialism, and with what qualifications; and second what does the Rwandan experience tell us about the general relevance of the model for understanding the options facing sub-Saharan Africa today?
We are aware that any attempt to categorize the regime in Kigali is going to be controversial. Interpretations of almost every aspect of Rwanda and its history are exceptionally polarized. Even scholarly and analytical writings tend to be pigeonholed by their critics as either apologias for or attacks upon Paul Kagame and the RPF. Compounding the problem, Rwandan public policy has been characterized by unnecessary secrecy and its inevitable counterpart, unrestrained rumour-spreading. One of the effects of this is a remarkably low level of knowledge and understanding about some of the topics that concern us, even at quite senior levels in government and in the resident international community.
We do not expect to avoid being caught up in controversy. However, we should stress that the theory of developmental patrimonialism is not primarily an evaluative exercise but an effort to disentangle the empirical dynamics at the heart of different types of African regimes, past and present. All of the regimes that appear to us to exemplify the more developmental form of neo-patrimonialism pose ethically difficult questions about trade-offs between liberal freedoms and human rights, on the one hand, and development outcomes (and thus other human rights) on the other. These issues merit discussion but we firmly believe that such discussions are only fruitful when they are grounded in a good understanding not just of all the relevant facts but also of systems and linkages.
Rwanda and the African modal pattern
As a developing country, Rwanda suffers from a number of quite severe disadvantages in addition to its violent political history. Landlocked, under-endowed with natural resources other than land and climate, and with an exceptionally unfavourable person–land ratio, it continues to be extremely poor in income and human-development terms. The formal sector of the economy is tiny. As a destination for private investment, it is geographically ill-placed and lacks compelling attractions. Nonetheless, the country's economic growth in recent years has been impressive, averaging over 8 percent per year in real terms since 2005.19 While this is growth from a low base, it compares favourably with what is being achieved by its better-endowed and more favourably located East African neighbours. In terms of poverty reduction and human development, a huge amount remains to be done, but the latest survey data show income poverty falling from 57 to 45 percent between 2005/6 and 2010/11, and substantial progress on maternal and child mortality rates.20
Some see Rwanda's recent progress as a good example of the gains from applying internationally recognized best practices in economic management and governance reform. There is something to this. Rwanda under Kagame is in some respects a star pupil of the Washington Consensus. The RPF-led government has been strongly committed to the private sector as the engine of development. It has recently adopted a good deal of orthodox best-practice thinking in its approach to investment promotion. That is, it has defined its role as facilitating and enabling private investment, with official bodies concentrating on the provision of pure public goods, including policy guidelines, information, standards, and regulation. Rwanda was rewarded for its strenuous efforts to meet the criteria of the World Bank's Doing Business survey when, in 2010, it was categorized as second-top global improver (after Kazakhstan), reaching a global ranking of 58th out of 183 countries and a regional ranking of 4th in sub-Saharan Africa.21 However, to emphasize these aspects of the situation would be to miss what is most distinctive, interesting, and widely relevant about the Rwandan experience. It would capture only one of the features that distinguish the Rwandan case from what may be called the African modal pattern.
In our understanding, the modal pattern in sub-Saharan Africa today is one in which ‘rent seeking’ is widespread and uncontrolled, and associated with both political and administrative corruption. The pattern may be seen as having five interrelated elements.22
First, rent extraction is a major source of personal enrichment for the political class as a whole as well as for private business. Second, the political leadership is either unwilling or unable to deny access to rent-taking opportunities by its major supporters, because it is largely by distributing such opportunities that it remains in power. Third, policy making is driven in part by the exigency of creating rents for allocation to supporters and for replenishing the campaign funds of the party in power. Fourth, corruption reaches down to the lowest levels of the public service, partly because clientelism has become systemic and partly because the political class lacks the moral authority to maintain a clean administration. Some public servants perform their duties with dedication for idiosyncratic reasons but the dominant pattern has become the exploitation of public office for private gain. As a consequence, finally, appointments and promotions within the civil service have lost their connection to considerations of merit and effectiveness.
We would maintain that, historically, Rwanda has displayed all the elements that, other things being equal, tend to reproduce the African modal pattern. If, as we are going to argue, Rwanda today is somewhat exceptional it would therefore be a mistake to treat this as an historical inevitability. Two sorts of evidence underpin this proposition.
First, the pre-1994 regimes in Rwanda were characterized by all or most of the features of the African modal pattern. Rwanda's long history as a state means that, in Pierre Englebert's terms, there is a high degree of ‘state legitimacy’,23 which contributed to aspects of governmental effectiveness then as it does today. The introduction of multi-partyism came only at the end of the 1973–94 Habyarimana regime and helped to destabilize it. Before that there was a one-party state under which clientelistic rent seeking was generalized, competitive, and disorganized. Businesses needed ‘godfathers’ within the administration or the military, and through this mechanism the small business sector financed politics.24 The Akazu, the apex of the clientelist system, was a powerful network close to but not effectively controlled by the President.25
Second, there have been repeated indications since 1994 that many actors in the RPF-led political system, not to speak of opposition elements abroad, would have found it both easy and desirable to resume the practices established under the former regime. Kagame's rather austere sense of public duty and strait-laced line on corruption are not to everyone's taste. Some, at least, of the early defections from the RPF government may be best explained in these terms. Of course, the question why former allies of Kagame eventually fell out with him and went into opposition or exile is controversial. The murkiness surrounding what was the eventual fate of some of these people, Hutu moderates and RPF originals alike,26 does not help in reaching a clear position on the matter. No doubt, each case was specific, and in most of them more than one issue was at stake. Nonetheless, a plausible suggestion about some of the prominent early defectors is that they were eager for a more ‘flexible’ approach, permitting greater scope for using public funds to buy the political support of key players, ‘the working methods of the old regime’.27
Even some of the more recent high-level defections are attributed by some observers to an underlying discontent with the rigours of leadership discipline under Kagame. The arrangements described below, in which politically generated opportunities for profit are comprehensively institutionalized and centralized, are seen as a little too austere. The arrests and court judgements that follow even minor infractions seem to many to be unduly severe, even if it is also true that offences are often pardoned and reputations restored quite swiftly.28 All of this tends to confirm that Rwanda could well have gone, and could still go, the way of the African modal pattern.
From the moment when Kagame secured his leadership of the RPF and the government in 2000, the system in Rwanda has distanced itself from the African modal pattern on each of the five points listed above. We argue that the key to this is a set of arrangements for managing economic rents in a centralized way and deploying them with a view to the long term. This has several dimensions, involving the relationships between private businesses and the government on the one hand, and the army and the ruling RPF on the other. In the next two sections of the article, we focus on the most directly political of these relationships: the RPF's own business operations.
Management of economic rents: the Tri-Star dimension
A key element in the centralization of rent management is the role in the economy of the private holding company Tri-Star Investments/CVL, which is fully owned by the RPF.29 Tri-Star grew out of the Production Unit maintained by the rebel army during the war of 1990–4. Its initial funding came from political contributions to the war effort by supporters, as part of a wider effort to attract people with business experience and capital from the Rwandan diaspora.30 Within a few years of the RPF's accession to power, companies owned wholly or mainly by Tri-Star were in metals trading, road construction, housing estates, building materials, fruit processing, mobile telephony, and printing, as well as furniture imports and security services.
In each of these fields, the Tri-Star subsidiary was initially the largest domestic firm, if not the largest investor in what was an extremely limited and depleted private sector.31 Today, CVL has a 50 percent stake or more in 11 companies operating in Rwanda.32 Most of these are the leading national company in their sector. With one exception, the competition they face comes primarily from either regional (usually Kenyan) or international (including notably Chinese) firms. In addition to this core portfolio, CVL has a minority stake in several joint ventures in Rwanda.33 Outside the country, it owns a chain of coffee shops in the USA and UK,34 and an air charter company registered in South Africa.35
The scale of the operation is substantial in relation to the national economy. In 2009, the group's turnover, referring to majority shareholdings only, was about US$35 million. In this restricted sense profits were around US$7 million after the payment of US$0.8 million in taxes. With the contributions from minority shareholdings, including in the Rwanda branch of the mobile phone company MTN, turnover was US$167 million, post-tax profits US$47 million, and taxes paid US$24 million. While the majority-owned CVL firms represented well under 1 percent of Rwanda's US$5.3 billion GDP, the whole group represented over 3 percent, a significant proportion given the limited size of the formal sector as a whole.36 The taxes paid by the larger group were equivalent to around 9 percent of all direct taxes paid in the fiscal year 2009/10.37
As discussed later, the management systems of Tri-Star/CVL have changed and improved over time, but there have been two constants: RPF ownership and a private-sector legal status. The holding company has been governed throughout by a board of directors nominated by the RPF leadership and managed by an executive chair or chief executive officer. Profits from the operations of the subsidiaries have been taxed and then either reinvested or returned as dividends to the RPF.38 The central relationship is between two formal organizations with their own corporate identities and rules.
We would argue that this arrangement is accurately described as a system for centralizing economic rents. We leave for later the question of what forms of ‘rent’ are included in Tri-Star/CVL earnings and the significance of this for the development of the Rwandan economy. In this section, we assume the element of rent is significant and focus on how the rent-centralization mechanism helps to shape the wider political economy, setting the Rwandan situation apart from all the five features that define the African modal pattern.
In the modal pattern, rent extraction is a major source of personal enrichment for the political class as a whole. In the current Rwandan pattern, this is not the case: Tri-Star/CVL profits accrue to the RPF corporately.
At the beginning, Tri-Star companies were concerned exclusively with responding to the acute material shortages that characterized the immediate post-war situation, using RPF financial reserves to import goods and even pay civil servants' salaries.39 In this phase, the party was ‘sponsoring the government’;40 financial returns were hardly an issue. As the Rwandan economy recovered from its deep post-genocide depression, however, the Tri-Star/CVL firms became profitable. Benefits then flowed back to the RPF in the form of dividends, at which point they became subject to the decisions of the RPF's national executive committee.41 What the party did with these funds has not been researched by us and is probably not researchable. However, the important point is that there is no evidence of direct profit taking by individual politicians or military leaders.
For example, the most controversial form of rent extraction in which the new Rwandan elite was involved for a period, namely wartime minerals trading out of Congo, was in part in the hands of a Tri-Star subsidiary, Rwanda Metals. The holding company records indicate that Rwanda Metals did contribute to the corporate balance sheet of Tri-Star, although profits were not large, especially by comparison with the MTN investments discussed below.42 Because of the international outcry about exploitation of Congolese natural resources, Rwanda Metals was sold off to a Botswana-based firm soon after the effective end of the war in 2002. We do not have any data with which to assess whether Congo minerals trading made any individuals rich in Rwanda. However, if it did, Tri-Star was not the channel through which it happened.
In the African modal pattern, the political leadership maintains itself in power by distributing rent opportunities to its major supporters. At least since 2000, this has not been the pattern in Rwanda. In the terms used by Bruce Bueno de Mesquita and his co-authors, the political system of the country has a moderately large ‘selectorate’ (enfranchised citizens) and a smaller winning coalition (the subset of selectors whose support is decisive to the retention of power by the incumbents).43 More crudely, it is not at present the votes of the general population that keep the incumbents in power. Under these conditions, the theory – which applies equally to regimes that are formally democratic or authoritarian – predicts a leadership preference for providing private goods selectively to members of the coalition, as the cheapest and most reliable means of political survival.44 Contrary to the theory, the RPF and its allies are gambling on the ‘expensive’ option of building support on a broad base by demonstrating an ability to provide more and better public goods.45
In Rwanda since 2000, policy has been driven rather exclusively by the view that economic and social development – underpinned by adequate provision of essential public goods by the state – is the only feasible route to overcoming the ethnic divisions and violent conflicts of the past. This is formally articulated in a document called ‘Rwanda Vision 2020’. Moreover, contrary to what happens with equivalent documents in most countries of the region, this is a real point of reference for ministers and civil servants.46 The assumption underlying the vision is that, if economic and social progress occurs fast enough, a new generation will emerge who are capable of fully assuming their national identity as Rwandans rather than privileging what divided them in the past. Many critics of the regime see this as naïve and argue that reconciliation needs to be attended to in a more direct fashion.47 But in so doing they also confirm that this is indeed the vision that drives policy.
In the African modal pattern, policy making is driven away from a public-goods focus by the need to create rents to allocate to supporters and more generally to finance politics. One of the things that enable the Rwandan regime to deviate from this pattern is the role played by Tri-Star/CVL. The RPF's running costs and campaign funding needs are met in the first instance by party member contributions, but these are able to be supplemented by distributed profits from the holding company. This has the side-effect of removing the pressure on the ruling party to raise political funds by methods that involve fraud, kickbacks, or other corruption.48
Readers who approach Rwandan reality from the angle of democratic ideals may well find it unfair that an already dominant political party can finance its activities in this way.49 However, in the analytical and comparative perspective we adopt, the contrast with the way incumbents normally exploit their position to sustain themselves in power is the more significant feature.
In the modal pattern, corruption has become systemic. In contrast, according to most surveys corruption is quite uncommon in Rwanda's public service at any level, and corruption with impunity is largely absent.50 ‘The least corrupt country I have worked in’ was the characterization offered by one experienced ‘Africa hand’ now in business in Kigali.51 The explanations are to be found at two levels. First, alleged corruption at high levels has been vigorously suppressed at regular intervals and this has shaped expectations most of the way down the chain of the public administration. Second, the RPF/Tri-Star link has made high-level political corruption unnecessary, which has given Kagame and his immediate circle the necessary moral authority to enforce a zero tolerance policy.
Finally, the combination of factors noted above has enabled the civil service to follow a pattern that differs from the one observed in recent years in most other countries of the region. It was necessary and possible after 1994 to reconstruct the Rwandan civil administration from scratch. Former civil servants were rehired, but ministries received a good deal of fresh staffing, much of it drawn from returnees from the diaspora and, increasingly, recent products of the National University or universities in Uganda and other countries. The resulting civil service is exceptionally youthful and there has been much learning on the job. In this context, recruitment and promotion decisions are able to be based to a large extent on merit and effectiveness. In addition, administrators are motivated and disciplined by an unusually effective form of performance-based contracting linked to the neo-traditional practice called imihigo.52
In at least five ways, then, Rwanda deviates from the African modal pattern. Our contention has been that this is partly explained by the way rent management is effectively centralized though the RPF/Tri-Star link. To complete this part of the argument, consider what evidence would compel us to reject our claim about centralization of rent management in Rwanda. It would have to be rejected, obviously, if we were to find one instance or more of personal rent capture by a leading or middle-ranking politician which was not effectively sanctioned. For example, if a minister or other prominent public figure were to go unpunished after undertaking a large-scale scam at the expense of the state, we could not maintain the hypothesis that rent management was centralized.
For Kenya, Uganda, or Tanzania, it would be possible to cite many such instances.53 We would maintain that there are no such cases in Rwanda. In Kigali, new appointees to permanent secretary or parastatal management positions are teased with the question ‘Have you got your pink [i.e. convict's] uniform ready?’54 Numerous examples may be cited where the mere suspicion of financial wrongdoing has been sufficient to prompt immediate and effective sanctions, legal and otherwise. Among the prominent individuals affected are several who were not known to be in any other respect in dispute with the leadership and, in that sense, candidates for victimization.55
The ‘Rwanda Briefing’ document which has circulated widely on the Internet under the names of four prominent exiles makes the argument that Kagame sets very high standards of integrity for those who serve under him, but does not live by the same rules.56 The document alleges that the President is ‘responsible for financial impropriety and theft of public resources on a grand scale’. However, the route by which it reaches this conclusion does not inspire confidence. It rests partly on non-credible claims about how Tri-Star/CVL is managed57 and partly on concerns about Tri-Star/CVL's relationship with the rest of the private sector that are widely shared, as discussed next, but have nothing to do with Kagame's personal wealth.
Tri-Star as example of long-horizon rent deployment
It appears, then, that one of the features distinguishing the Rwandan business–politics regime is an effective centralization of control of rents. Can we also distinguish the post-2000 Rwanda pattern from the equally centralized but short-termist kleptocracy that characterized a number of ill-reputed post-colonial African regimes? To what extent does it differ, for example, from the pattern in Zimbabwe, where it is also the case that the ruling party, ZANU-PF, owns significant private businesses?58
This is not a question of whether the intentions of the leadership are more benevolent or otherwise praiseworthy. The issue may be restricted to whether rent management is directed towards the short-term enrichment of members of the political class and its allies, or alternatively towards ‘growing the pie’ of the national economy, maximizing opportunities for long-term accumulation.
We believe the evidence places the RPF-led regime clearly in the second category. To some extent, this puts us at odds with a widely held view in Kigali, which emphasizes the various ways Tri-Star/CVL companies have benefited from government decisions and take advantage of their dominant position to beat down competitors. This view has been influential among the country's international donors,59 and it is sometimes claimed that prospective foreign investors have been put off from coming to Rwanda by the knowledge that their potential rivals in the domestic market are part of a conglomerate owned by the ruling party.60 The government itself has become sensitive to the suggestion that the role played by the party-owned enterprises is anti-competitive and thus a hindrance to the emergence of a healthy private-sector-led economy. However, we think this concern is in several respects exaggerated.
No doubt, the allegations about restrictions on competition have something in them. The members of the conglomerate certainly gain from decisions to buy goods and services from each other.61 Tri-Star's financial power (a combination of its own resources and its financial credibility as a borrower from national banks) does give it a major advantage over would-be competitors from the domestic private sector, and this is no doubt resented.62 At various stages, the subsidiaries have been in a position to earn monopoly profits, as the CVL website itself concedes.63 Unsuccessful private business people in Kigali do sometimes explain their troubles by saying that they have been unwilling to bid for contracts against Tri-Star affiliates because they cannot believe that the playing field will be level.64 These are all relevant considerations.
On the other hand, the Tri-Star/CVL firms do not seem to have either the overweening market power – after foreign competitors are factored in – or the ability to mobilize political leverage that is often attributed to them. Now that business is booming across East Africa, most of the CVL subsidiaries are facing quite intense competition from regionally or internationally owned firms; if they are able to grow, it is because markets are expanding.65 Their managers naturally reject suggestions that government purchasing offices or the Rwanda Revenue Authority do them any favours. Specific examples and testimony from other sources add credibility to these claims.66
What distinguishes the conventional view from ours, however, is not so much a different view of the facts, but a different interpretative framework. The conventional approach takes a textbook view of competition, one that assumes the existence of a fully fledged capitalist economy. Following Khan, we believe that the challenge which applies to Rwanda's situation, even today after a decade or more of economic recovery, is not that of defending a mature private-enterprise sector against monopoly power, but that of getting capitalism started. It is a misreading of economic history to treat these challenges as the same.
For foreign firms, there are many more significant disincentives to investment in Rwanda than the possibility of unfair competition from the biggest of the local firms. They include the implications of small domestic market size for any investments in import substitution; the barriers to export production created by the country's landlocked status; land, power, and transport issues; and the comparatively limited opportunities for quick wins in natural-resource extraction, particularly in comparison with Uganda and Tanzania.67 Crucially, they also involve the risks associated with investing in an immature business environment, in which professional skills and support services are in short supply. The risks are especially large in sectors characterized by major backward and forward linkages and externalities, such as commercial horticulture.68 As for domestic investors, many have only emerged as significant players thanks to the initial learning undertaken, and the example provided, by the politically driven operations of the Tri-Star/CVL companies.69 As different interviewees put it, Tri-Star/CVL has been an ‘ice breaker’ for other firms in Rwanda's formal economy; it has shown the way, ‘preaching by example’. As a result, there has been more ‘crowding in’ than ‘crowding out’ of other local firms.70
Our contention is that Tri-Star/CVL is playing a critical role in getting capitalism started in Rwanda. It did so initially with investments that met urgent social or political needs and subsequently earned the monopoly profits that would have accrued to any first-comer. In several cases, it is reasonable to treat the profits made as a return on risk taking and investments in learning that eventually helped to crowd in other domestic players.
The early ventures included providing housing for returnees and private security services for official buildings and installations. Investments in basic import substitution followed, such as bottled water and basic dairy products. Subsequent investments were in furniture imports, mobile telephony, road construction, housing estates, building materials, fruit processing, and printing. In most of these fields – most strikingly in the cellphone example detailed further on – the Tri-Star subsidiary was at first a pioneer in activities where there was little interest from the domestic or (more relevantly, perhaps) the diaspora private sector. For some years, the majority of firms were not highly or even moderately profitable and were supported by the RPF primarily for their social benefits. Although they operated like private companies, they were run by party cadres, some of them with little or no business experience, and were probably not very efficient. The introduction of the accounting and reporting systems that would allow us to judge efficiency only came later.71
The management of the Tri-Star/CVL group seems to have gone through three distinct phases. In the earliest phase, management styles within the group resembled those of the parastatal sector, but over two subsequent periods of reform (roughly 2005–8 and 2008 to the present) the companies have come to be managed more in accord with the norms of private corporate business.72 Today, according to the CVL CEO, the model is that of ‘early-stage venture capitalism’. The orientation is towards creating firms that are attractive partners for international direct investors, not just large players in domestic terms. The company's board – now consisting largely of RPF members with banking or financial-management backgrounds – is determined to raise efficiency and management reporting to the necessary standards.73
A continuous feature of Tri-Star/CVL operations throughout the three periods is that they have been to a greater or lesser extent risky, and in several cases involved heavy initial learning costs.74 Thus, what emerges as most significant is the willingness of Tri-Star/CVL to use its financial clout to fund investments with high expected social benefits and/or positive economic externalities, including those associated with venture capitalism.
This interpretation applies clearly to the most important early Tri-Star investment, in MTN Rwanda. Tri-Star largely funded the initial establishment of the MTN cellphone network in Rwanda at a time when MTN and other global operators found the country's potential subscriber base uninteresting and the government of Rwanda was in no position to take the lead financially. The results of this venture were spectacularly successful, leading the MTN parent company to expand its equity share. Not long after setting up in Rwanda (July 1998), it went on to establish a network in Uganda (October 1998), with Tri-Star as one of the shareholders. Initially, Tri-Star held approximately 65 percent of the equity in the Rwandan operation and MTN South Africa 26 percent, with the government of Rwanda contributing the balance through the then parastatal Rwandatel. In the following years, Tri-Star progressively transferred holdings to the South African parent company, reducing its share to 40 percent by 2007. That year, MTN International assumed majority control (55 percent) when Tri-Star sold it a 15 percent stake, anticipating the entry into the market of two new providers – Tigo and the now privatized Rwandatel. Tri-Star realized 5–10 times its original investments as a result of these sales.75 In October 2011, the remaining 10 percent government stake and a further 15 percent Tri-Star (now CVL) holding in MTN Rwanda was sold to the parent company, leaving CVL with a 20 percent stake.76
In other words, Tri-Star helped to generate a demonstration effect and a learning experience in which one of the beneficiaries was an international firm. It thereby ensured not only that Rwanda entered the world of mobile telephony earlier than it would otherwise have done, but also that the network that was established was at least partly owned by domestic capital. There have been spill-over benefits for the wider information and communication technology field, with new IT firms being established by Rwandan entrepreneurs who cut their teeth negotiating with MTN on behalf of the government.77
In other sectors, too, the emphasis has been on using financial clout to enable local players to undertake the risks and learning associated with getting established in competition with international suppliers. This is particularly applicable to building and road construction, where some international firms, including increasingly Chinese companies, have not only experience but a financial capacity that allows them to be free of risk-averse Africa-based banks.78 As experience in Uganda confirms, it is hard for local firms to achieve operational competitiveness with international and particularly Chinese firms in these sectors unless there is some mechanism for financing start-up costs and learning-by-doing.79
In all of these operations, there is awareness that competitiveness does not depend only on having a supportive and patient financial backer. Tri-Star firms have had extremely open recruitment policies for managers, engineers, and other technical specialists. In a number of cases, diaspora professionals have been head-hunted but increasingly the firms recruit by means of open advertising within the East African region and beyond. They can and do hire globally to meet needs in some technical areas.80 A willingness to hire internationally for the sake of creating competitive national firms has been noted as a distinguishing feature of regimes that we have elsewhere characterized as ‘developmental patrimonialisms’.81 This would appear to be one of the features that distinguish the policies of such regimes from those of the African modal pattern, in which ‘jobs for the boys’ and jobs for locals take precedence over a given firm's efficiency and competitiveness.
Horizon and RIG
What the Tri-Star/CVL experience illustrates is that, in very underdeveloped capitalist economies, an incipient private sector may benefit from being led from the front and not just facilitated with the provision of a business-friendly environment. Other dimensions of the business–politics relationship that has emerged in Rwanda are also consistent with that principle. They complete the picture on central management and long-horizon deployment of economic rents.
To begin with, the privatization process was more actively supervised than has typically been the case within the African modal pattern. At the outset (1997–2004) the policy was what one interviewee called ‘all-out privatization’ and followed the main principles of the Washington Consensus quite closely. But subsequently a number of recently privatized firms, including Rwandatel, were taken over and then privatized for a second time when the first buyers proved incapable of providing the promised injections of capital and know-how. More recently, Rwandatel went into receivership after Libya's Lap-Green failed to meet some of the conditions of its purchase, and is to be sold for a third time. Policy makers today are more tough-minded about the likely benefits of privatization and there is stronger interest in the option of bringing private-sector discipline into remaining and newly created state-owned companies.82
The government has adopted a relatively activist stance in at least two other areas. First, it encouraged the army to create an investment arm with which to undertake socio-economic projects and create productive enterprises. The result was another holding company run on private-sector lines, Horizon Group. Second, it brokered the creation of a private investment consortium that brought together a group of the richest domestic and diaspora entrepreneurs. The consortium is known as the Rwanda Investment Group (RIG).
Horizon Group's first venture was a construction company established with an initial gift of equipment from the government and a team of military engineers seconded from the army. It undertook a series of projects for the government, including building irrigation dykes and constructing coffee-washing stations, ‘to avoid the Chinese doing everything’.83 At an early stage, it established a cassava-growing operation and a dairy (Laiterie Nyabisindu). Subsequently, it moved into comprehensive urban site development, first on land bought from the Housing Bank in Kigali and later in collaboration with CSS-Zigama, the military's micro-finance initiative.
Horizon is now also in pyrethrum processing,84 as the owner of the Sopyrwa plant in Musanze, in the former Ruhengeri Province, which is linked to 24 large producer cooperatives in the area. Sopyrwa, created as a parastatal in the early 1970s but recently privatized, was close to bankruptcy following the global credit crunch. Horizon was approached with a view to a takeover by the former owners, a group of private entrepreneurs, in view of the company's track record in agri-business undertakings.85
The operations of Horizon subsidiaries are increasingly diverse. Thus, Horizon Logistics is moving into logistical support to Rwandan peacekeeping forces in a number of locations, taking over from international firms. On the other hand, Galaxy Management Systems has recently been established to assist the Kigali City Council with street naming and house numbering, considered a precondition for efficient taxation.
Horizon Group is run as a private firm. Its board does not include serving military officers, although its CEO was seconded from the army following a previous posting with the military bank. However, as with Tri-Star, its social and political purposes are important, and according to the CEO profitability is no more than a co-equal objective.86 The group's interest in rural construction arose from the perceived need to restore export agriculture to something approaching its previous condition. Urban housing was signalled as a vital matter when competition between returnees and displaced people for access to the limited housing stock became acute in the later 1990s. The intervention in pyrethrum was necessary to avert the collapse of a privatized parastatal, which would have harmed employment and smallholder incomes in the still politically fragile north-west. Like Tri-Star, Horizon has a robustly internationalist approach to filling skills gaps in its firms, with business efficiency and the meeting of strategic social objectives taking precedence over commitments to local hiring and capacity development.87
The Rwanda Investment Group SA (RIG) is also a holding company but of a different character, a public–private partnership. It was created in May 2006, at the instigation of President Kagame and in response to the difficulty of raising funds for large projects in the absence of a local capital market. Currently, it has 41 shareholders, including 31 individuals, four medium-sized companies, and six institutional investors including the Rwanda Development Bank, major insurers, and Crystal Ventures Ltd. The initial start-up capital totalled US$25 million. In effect, RIG brings together ‘nearly all’ of the richest and best-known businesspeople of Rwanda and the diaspora, along with the major public financial institutions.88 At present it operates with a fairly restrictive minimum subscription of US$3.6 million, but the intention is to seek international partners and in due course float shares publicly.89
RIG's mandate is to raise capital for investments of particular national interest without relying on international capital markets or the local branches of foreign banks. Social objectives are less prominent than in the cases of CVL and Horizon, but the group's mission is described as a form of ‘economic patriotism’.90 A local public–private partnership appeared necessary to the founding members of the consortium at the time when the country's largest cement factory CIMERWA, a Chinese–Rwanda government joint venture under the Habyarimana regime, was being privatized and needed a substantial capital injection. As a result, RIG acquired a 90 percent stake in CIMERWA, with the government of Rwanda holding the balance.91
RIG subsequently invested heavily in peat mining and methane gas extraction from Lake Kivu (both potential solutions to Rwanda's acute electric power shortage). It is in a public–private partnership with the government for the establishment of the Kigali Industrial Park and several other schemes.92 These are all initiatives that funding sources with no ‘patriotic mandate’, or willingness to underwrite risks, might well have considered unsuitable.93
A developmental neo-patrimonialism?
It seems, therefore, that the business–politics relationship in Rwanda rests upon a centralized management of rents and their utilization in a long-horizon perspective. In different ways, the RPF's holding in Tri-Star/CVL, the army's involvement in the Horizon Group, and the creation of RIG all exemplify and contribute to this pattern. In other words, the Rwanda case complies with the main features of the developmental patrimonialism model. It agrees, moreover, with Khan's thesis about the constructive role that utilization of rents can play in incipiently capitalist economies.
Rwanda's developmental patrimonialism differs from the model provided by earlier regimes of this type, such as Houphouët-Boigny's Côte d'Ivoire or Banda's Malawi. The allegations of the ‘Rwanda Briefing’ notwithstanding,94 the distinctive features of the regime do not seem to include a blurring of the distinction between the resources of the state and the private income or wealth of the ruler or ruling group. Generally, the boundaries between government operations and the private-sector operations of the RPF and the army are clear and quite formalized. We have argued, moreover, that one of the effects of the Tri-Star/CVL arrangement has been to enable the political elite to enforce an unusually strong anti-corruption line. For this and other reasons, the country scores well on conventional business-climate criteria.
The implication would seem to be that Rwanda since 2000 has displayed some distinctive features compared with other exemplars of developmental patrimonialism. Nonetheless, the concept of developmental patrimonialism centred on long-horizon rent management helps to illuminate what is most distinctive and important about politics–business interactions in Rwanda. Other resource-poor, landlocked African states surely have something to learn from the politically inspired economic activism represented by Rwanda's Tri-Star/CVL, Horizon, and RIG.