The trajectory of much of sub-Saharan Africa over the past thirty years has been a standing rebuke to advocates of the free-market Washington Consensus. Successive waves of structural adjustment programmes, international conflict mediation, good-governance monitoring and the best efforts of numerous western-funded ngos appear to have done nothing to halt rural crisis, ethnic conflicts and spreading shanty towns. In the eyes of many critics, they have merely exacerbated the situation. Along with the impoverished narco-economies of Central America, the blight-struck ex-Soviet republics and the sprawling slums of Cairo, Kolkata or Jakarta, this is where globalization has most visibly piled up misery and destitution, just as it has accumulated undreamt-of wealth in Manhattan or Mayfair.

The popularizing works of Paul Collier—first The Bottom Billion and now Wars, Guns and Votes—are a robust rejoinder to such views . Collier is a former Director of Development Research at the World Bank and currently runs the Centre for the Study of African Economies at Oxford. His research teams make use of the most advanced econometric techniques to identify the factors causing states to ‘fail’ and the policies that could make them succeed. Their diagnoses are based on sophisticated statistical analyses of a wide array of datasets, and Collier’s policy propositions are presented as logical deductions from them. He has some trenchant criticisms of World Bank and imf practice, above all their apparent fetishization of elections and too-rapid departure from post-conflict situations. He is unrelenting in his blasts against the greed and corruption of the local ‘villains’, who have blighted the development of the poorest countries in Africa and elsewhere; and even more so of Western critics of free trade—Christian Aid, apparently ‘infiltrated by Marxists’, comes in for particular lambasting in The Bottom Billion. On the other hand, there is genuine warmth in his praise for the few ‘heroes’—generally us-educated finance ministers and central bankers, such as Nigeria’s Ngozi Okonjo Iweala and Charles Soludo, or Ghana’s Kwesi Botchwey—who have followed responsible paths of ‘economic reform’. His books aim to garner popular support for econometric techniques, as applied to such intractable problems as world poverty and endemic conflict, as well as his preferred policy solutions. The Bottom Billion explains that statistical evidence will be used to ‘smash’ stereotyped images of ‘noble rebels, starving children, heartless businesses and crooked politicians’. In the process, Collier hopes the reader will get ‘a flavour of how modern research is done, and the thrill that comes from cracking intractable questions’.

The dramatic expansion of econometric and quantitative-modelling techniques has been one of the most significant trends throughout the social sciences since the 1990s. Originally elaborated within the rational-choice framework of American neo-classical economics, mathematical models of risk-analysis or game theory have now spread into the ‘political’ domains of military conflict, state forms and ethno-linguistic identities. The resulting discipline—part economics, part statistics, part quantitative political science—now plays a central role not only in scholarship, but in formulating policy options for global institutions. The dense thicket of algebraic equations with which econometric studies are hedged normally ensures them a narrowly specialized readership. In the past, Collier’s work on civil war has produced such results as Uw = {p(D).T; M; C}, where the utility of choosing rebellion (Uw) is a function of the probability of victory (p), the gains to rebels upon victory (T), the potential for government defence spending (D), the expected duration of the conflict (M) and the co-ordination costs of mobilizing for rebellion (C). By contrast, Bottom Billion and Wars, Guns and Votes are breezy reads, sprinkled with anecdotes, which keep the datasets and algebra out of sight, though fully referenced.

Among critics of neo-classical economics, econometric approaches to issues such as civil war and social breakdown tend to provoke a set of almost visceral objections. First, the expression of complex, historically produced structures and motivations through the binaries of individual rational-choice calculation—will I lose or gain by rebelling against the government?—is condemned for its extreme economic reductionism. Second, econometrics is often accused of simply corroborating what the data it deploys has already shown. This was certainly a charge levelled against Collier’s early work, a 1986 study of Tanzania’s ujamaa villages, which crunched figures from a dataset of village surveys to show that intra-village income differences were greater than inter-village ones. The conclusion had already been substantiated by a sophisticated body of empirical work, which had also pointed to the existence of pre-ujamaa rural class differentiations. Collier and his collaborators simply overwrote this research in the name of ‘rigorous’ statistical methods. A third charge is that of opaque causality: statistical risk-analysis typically shows up correlations between, say, poor economic performance and the likelihood of civil war, but in itself cannot tell whether the lagging economy is exacerbating social tensions, or whether the tensions are causing the economy to lag—delaying investment, for example.

There are good grounds for all these objections. Nevertheless, in rejoinder it might be said that any model requires some form of reduction, while the transformation of social and economic data into quantifiable, comparable series may reveal surprising patterns, which can generate new questions and insights into unforeseen outcomes or unintended consequences—correlations between, say, girls’ education and falling birth-rates, or rural unemployment and civil war. As to the problems of causality, sophisticated models have increasingly been able to introduce controls for this. This work deserves to be given serious critical examination, not just because of its overwhelming predominance in development research and policy-making, but because the problems it addresses are so stark. In sub-Saharan Africa, for example, gdp per capita, health, mortality and adult literacy rates have all deteriorated significantly over the past decades; over a third of the population is classified as undernourished; and in some regions, civil wars, with their accompanying mass rapes and population displacements, appear almost endemic. It would be foolish to pretend that the solutions to these problems are obvious. Well-informed proposals for diagnoses and cures should be considered on their merits.

The fundamental premise of The Bottom Billion is that the shape of world capitalist development has undergone a radical change since the 1980s. For forty years, Collier argues, the ‘development challenge’ had consisted of ‘a rich world of one billion people facing a poor world of five billion people’. This is no longer the case: ‘Most of the five billion, about 80 per cent, live in countries that are indeed developing, often at amazing speed. The real challenge of development is that there is a group of countries at the bottom that are falling behind, and often falling apart.’ As a result, ‘we must learn to turn the familiar numbers upside down: a total of five billion who are already prosperous, or at least on track to be so, and one billion who are stuck at the bottom.’ Why is this unfortunate sixth of humanity not benefiting from the economic conditions that are delivering prosperity elsewhere? According to Collier, the 58 states in which the ‘bottom billion’ live have become ‘stuck’ in what Jeffrey Sachs has termed ‘development traps’, which hinder economic growth. Collier proceeds to define four such traps—endemic conflict, natural-resource dependence, lack of coastal access, ‘bad governance’—and draws on the results of his team’s statistical analyses to calculate how the inter-relations of these and further factors (population size, education levels, ethnic heterogeneity, political systems, military budgets, gdp growth rates) might mitigate or exacerbate a country’s prospects.

Collier sets out to ‘explain the conflict trap’—the propensity of poor countries to have civil wars or military coups—‘statistically’, drawing on the University of Michigan’s ‘Correlates of War’ database. His team found three variables associated with civil war: gdp per capita of less than $2,700, a slow growth rate and dependence on primary commodities, such as oil or diamonds. By contrast, neither ethnic oppression nor inequality correlate statistically with civil war, although ‘ethnic dominance’ may do so. Thus, for the growth rate, Collier estimates that the ‘typical’ risk of civil war, 14 per cent, falls to 11 per cent if the economy grows by 3 per cent. By contrast, he analyses the ‘natural resource trap’ in terms of political factors. Collier’s team finds that a newly established democracy with natural resources will grow more slowly, since politicians will use resource windfalls to buy votes rather than for investment. However, if there are sufficient checks and balances to penalize patronage and corruption, the growth rate is higher. Botswana is, apparently, the shining example of this combination of factors.

The third trap is to be landlocked and surrounded by ‘bad neighbours’—poorly run states with collapsing transport infrastructure—which prevent the trapped country from getting its exports out onto world markets. Bad neighbours also prove to be poor markets for the landlocked country’s goods, limiting growth ‘spillover’. Collier argues that the effects can be precisely quantified: for every 1 per cent of a neighbour’s growth, the average country will grow 0.4 per cent, but the beneficial effect rises to 0.7 per cent if it is landlocked, and to 0.9 per cent if it is an African landlocked country. Finally, there is the ‘bad governance’ trap, again addressed in terms of economic outcomes. Here Collier’s team may justifiably be accused of discovering the obvious: that an economic turnaround is more likely to be sustained in a country with higher income and a larger, better-educated population; less apparent, perhaps, are the findings that growth is less likely to be sustained if the country’s president has been in office for a long time, if the terms of trade have become favourable or if the country has recently emerged from a civil war. The last variable suggests to Collier that the post-conflict situation is typically ‘fluid’ and presents specific opportunities for targeting aid and interventions.

Having determined by these means a set of instances in which outside intervention can tilt the risks in favour of peace and growth, Collier correlates these with the international policy instruments on offer to produce an ‘agenda for action’. His principal policy recommendation is for a sustained military presence by G8 or ‘international community’ forces to ensure that ‘fragile post-conflict situations’ do not revert to war. ‘We might think that peace should be a struggle waged by citizens of the country itself’, he writes, but ‘the evidence’ (not cited here) ‘is against such internal solutions’—‘breaking the conflict trap and the coup trap are not tasks that these societies can readily accomplish by themselves.’ A long-term military commitment—at least a decade—is required. He cites the 2002 British intervention in Sierra Leone as the model of a sustained military presence assisting the reconstruction of economic and social infrastructure. In The Bottom Billion, Collier’s team price a typical civil war at $64bn, and so manage to demonstrate the benefits of military intervention in Sierra Leone were 30 times the expenditure it involved, since they include the $64bn that reversion to civil war would have cost. Aid should be hedged with conditionalities: contingent on good governance, for example; technical services supplied by outside specialists will generally deliver better results than local development attempts. Economic policy should be geared to growth—‘that means sezs’—which will in turn attract investment flows, underwritten by the stable security situation. Collier advocates an international set of laws and charters, which would reduce corruption by upholding agreed norms in commercial and political life, helped by pressure from civil society. Finally, he suggests that the rich countries should drop all trade barriers to ‘bottom billion’ exports, while maintaining existing tariffs for Asia, so that African countries can gain a foothold in world markets.

In sum, Collier’s agenda follows from his chosen instruments: an external military presence, international standards, better communications for the landlocked, aid for exporting, support for brave ‘reformers’, better policy coordination between ‘donor countries’ and a clear focus on the ‘bottom billion’, rather than the bottom five-sixths. With its rousing combination of do-gooding and can-do, The Bottom Billion has been taken up—as Collier informs us in a postscript to the paperback edition—by the ‘great and good’ across the globe: Ban Ki-Moon, Robert Zoellick, Dominique Strauss-Kahn, Peter Mandelson, Bono, Bob Geldof. ngos are enthusiastic about his recommendation for voluntary international standards on resource extraction; the un Secretary-General is interested in his post-conflict guidelines. Collier’s proposals are reiterated and extended in his latest book Wars, Guns and Votes, which focuses directly on the question of democracy: is the bottom billion ready for it? In principle, Collier welcomes the spread of electoral representation in Africa and elsewhere since the end of the Cold War. But he warns that all too often, in post-conflict situations, ‘elections are the signal for peacekeepers to depart’, resulting in a higher risk of reversion to ethnic division, rebellion and civil war. His team’s statistical models draw on Polity iv and World Bank data to show that, in poor countries, democracy leads to increased political violence—loosely defined here as ‘assassinations, riots, political strikes, guerrilla activity, civil war’. As empirical evidence, Collier points to the Kenyan and Nigerian elections of 2007, and the Zimbabwe referendum of the same year.

The principal problem is ethnic division. Collier here qualifies his earlier rational-choice approach to democracy, in which the electorate was seen as so many million homines economici: in ethnically heterogeneous societies, this is likely to be trumped by ‘expressive voting’ based on ethnic identity, resulting in ‘frozen’ voting blocks. If this analysis is correct, it has ‘huge implications’—for ‘the whole modern approach towards failing states had been based on the premise that they would be rescued by democratic elections’. If ‘bottom billion’ societies lack the security and accountability necessary for democracy to flourish, then ‘we’—the G8 has now been sidelined in favour of the presumably more reliable us, UK and France—must provide them. Aware that the rich world may be squeamish about further military occupations after Afghanistan and Iraq, Collier proposes that a credible threat of armed might—as offered by British guarantees to Sierra Leone, or French ‘protection’ for its ex-colonies—may suffice. Africa only needs to supply ample military bases for the new intervention force. In another adjustment, Wars, Guns and Votes attributes a cost of $20bn to civil wars—a substantial, and unexplained, drop from the $64bn of Bottom Billion, but still ensuring a positive cost–benefit outcome for international intervention. The new book also takes a far more benign view of domestic military coups: now seen not as instances of political violence, but as potentially useful, so long as the generals can be given sufficient ‘guidance’ on economic policy. Collier suggests Zimbabwe may be a suitable case for treatment.

What should we make of Collier’s claim that his proposals are based on scholarly and statistical ‘rigour’? The problems begin with the notion of the ‘bottom billion’. There is a constant slippage in these books between the nation-state, the primary unit of analysis, and individuals as a fraction of the global population. Within the nation-states themselves there is no coherent account of social differentiations, beyond the ubiquitous appeal to ‘ethnicity’ and the occasional passing mention of ‘rent-seeking elites’. Nor is any evidence offered for Collier’s founding premise that the other billions are ‘on track’ to prosperity. It is far from clear that this is the case for the half-billion Indians who live on less than the purchasing power equivalent of $1.25 a day, for example. Again, if we take Collier’s figure of $2,700 nominal gdp per capita as the threshold for a country to enter a virtuous development cycle, then nearly 3 billion fall beneath it. Both locally and on a world scale, the accumulation of capital systematically creates large numbers of people ‘at the bottom’: in the UK 14 per cent of the population and in the us 12 per cent live below their country’s poverty line. In The Bottom Billion Collier is coy about naming the 58 states he deems ‘trapped’, on the grounds that ‘stigmatizing a country tends to create a self-fulfilling prophecy’: he refers to them instead as ‘Africa+’, perhaps because African countries are used to being stigmatized. Wars, Guns and Votes dispenses with these scruples and provides an alphabetical list which includes most of sub-Saharan Africa; the ex-Soviet republics in Central Asia; Laos, Cambodia, Myanmar; a sprinkling of Himalayan kingdoms and impoverished islands; and the oddities of Bolivia, Yemen and North Korea. The vast discrepancies in historical experience and political-economic formation among these states speak for themselves.

With respect to the statistical analysis, this is as good as the data used and the models being tested. Collier and his colleagues take gdp growth as a standard measure, but African output and growth figures are known to be frail, especially for countries in conflict and for transnational comparisons, where differential purchasing power has to be calculated. Gini co-efficients are even more approximate. Collier’s measures of ‘good governance’ rely on the scores produced by the Polity iv project—ranging from –10 for a hereditary monarchy to +10 for a democracy—or the World Bank’s Country Policy and Institutional Assessment index, whose rankings on ‘economic management’, ‘structural policies’, etc., are decided annually by the Bank’s own staff. Both suffer from the shortcomings of ‘expert opinion’, and may be subject to bias in the selection of factors or in their weighting when combined to produce a single country score. Thus, for example, one-party states in Africa have varied in their levels of democracy, freedom of association, and freedom of the press; it would be patently wrong to equate them all with dictatorships. The standard index of press freedom is compiled by the staff at Freedom House, who rank each country on a 3-point scale: ‘free’, ‘pretty free’ or ‘not free’. Other social factors may be judged by the 4-point scales of Afrobarometer opinion polls: ‘Are you very satisfied, fairly satisfied, not very satisfied, not at all satisfied . . .’ Similarly, in representing degrees of ethno-linguistic diversity as a number for each country, the Ethnolinguistic Fragmentation indices that Collier’s team relies on inevitably discount the complexities of social and historical context. How should Tanzania’s ethnic and first-language diversity be balanced by its Swahili lingua franca, for example? There are further doubts as to whether the ethnic groups enumerated in these indices (created from 1980s Soviet ethnographic studies) would reflect how people identify their own loyalties today. The role of ethnicity remains untheorized in its application here; we are offered no explanatory analysis of why it should be important.

If ethnicity is constantly invoked, class appears nowhere as a variable in Collier’s work. There is no recognition here of the particular forms of capitalism that have developed in Africa, nor the class formations they have produced; nor, importantly, of how ethnicity and class may interact. Just as striking, all Collier’s variables are internal: gdp, natural resources, ‘governance’, for example; external factors—foreign intervention, global commodity prices, dollar exchange-rate fluctuations—get no mention. Yet it is precisely these differences in the so-called ‘explanatory variables’ that need to be explained. The point is ironically illustrated in Wars, Guns and Votes when Collier provides a thumbnail sketch of Côte d’Ivoire’s development, in which the French colonial state, the imf, the World Bank, the French Treasury and French armed forces play determining roles at nearly every turning point. Another example is his comparison of Botswana and Nigeria as two resource-rich, ethnically diverse societies: Collier ‘finds’ that Botswana experienced stronger growth because its electoral system was subject to ‘checks and balances’—so much so that the same party has won every election—whereas Nigeria has relatively weak political restraints. In fact, Botswana is dominated by one ethnic group, the Tswana, who comprise 79 per cent of the population (their dominance suggesting, according to another of Collier’s ‘findings’, that the country ought to be prone to conflict). Botswana’s peculiar status as a High Commission Territory surrounded by South Africa, which controls its diamond trade, is excluded from the analysis.

As Christopher Cramer has pointed out in a World Development article, ‘Homo Economicus Goes to War’, the weaknesses of the datasets are compounded when they are used, as so often by Collier, as proxies for some other variable. Thus Gini co-efficients stand in as quantifications of rebels’ grievance levels, or represent the self-preservation motives of elites. Ethno-linguistic fragmentation figures are used as a measure of rebels’ co-ordination costs; the preponderance of primary commodities in a country’s gdp counts as a proxy for lootable goods, and thus for ‘greed’ as a motivation in civil war. Yet clearly, coca in Colombia, opium in Afghanistan, oil in Nigeria or diamonds in Angola have played very different functions in the civil conflicts there. As Cramer notes, mainstream econometrics practitioners have long cautioned against the ‘institutionalized abuse’ of certain techniques: the arbitrary prioritization of particular variables, empirical weakness of data and loose application of proxies; where this is the case, peer review will offer no safeguard.

Methodological problems are perhaps at their most acute in econometric analyses of rebellion and civil war. The models employed tend to involve two players, and to treat ‘civil war’ as a single phenomenon, rather than one of overlapping conflicts. Historically, however, these conflicts have been immensely complex: in Angola, for example, conflict began as an anti-colonial struggle against the Portuguese, became over-determined by Cold War rivalries between the us and ussr, with South Africa and Cuba playing central roles, and was continued into the 1990s by dint of foreign backing for Savimbi. Collier’s work in this area has notably ignored the question of external intervention, state aggression or coercive social relations, concentrating instead on rebels’ motivations, understood on rational-choice lines. While earlier researchers had analysed these in terms of ‘grievances’—deduced from number-crunching a country’s gdp growth rates, Gini figure, political-repression score and ethno-linguistic fragmentation tally—Collier and his fellow researcher Anke Hoeffler argued that ‘greed’ was a greater mobilizing force. They proceeded to ascertain ‘greed’ figures by combining the share of primary commodities in overall exports, the proportion of young males in the population and national education levels.

In Wars, Guns and Votes, Collier has abandoned greed, and ‘motivational’ explanations in general, in favour of a ‘feasibility hypothesis’ for rebellion: poor-country populations will rebel if it is feasible. Clearly, the presence in the country of the international military force that Collier is calling for is statistically likely to lower feasibility levels, even if historically foreign forces have generally provoked resistance. But again, Collier’s chosen variables arbitrarily exclude key external factors in Africa’s ‘civil’ wars. us funding had propped up the kleptocratic dictatorship of Mobutu for three decades: its abrupt withdrawal at the end of the Cold War was a crucial precipitating factor in the Zaire/drc conflicts that followed. Regional spillovers, or sharp drops in living standard as a result of imf-imposed policies, have been frequent catalysts of conflict; none of these factors feature in Collier’s explanations.

Perhaps the most fundamental objection to current econometric practice is that it rests on prior assumptions and post-hoc hypotheses which remain systematically unexamined. State-of-the-art development theory now centres, firstly, on mobilizing an army of researchers to reduce complex social phenomena to quantifiable, comparable series of data—the process of reduction itself usually involving value judgements which are scarcely questioned; and secondly, on the models themselves. Far less critical attention is given to the theoretical assumptions underpinning the ‘hypothetical leap’ between the statistical result and the researcher’s ultimate explanation of it. Social, historical and political determinants have been reduced to a set of numbers at the beginning of the process. At its end-point they return, in disembedded form; or embedded only in the commonsense—which is to say, ideological—assumptions of the researcher. If it is to do useful work, econometrics must recognize its place, as a lower-order set of tools which may generate correlations or discrepancies whose elucidation requires more richly theorized—more conceptually and empirically developed—forms of enquiry.

Collier’s work is not informed by any explicit, overarching theory of development or any historical perspective that might inform one; nor does he offer any social analysis. There is an implicit theory of human behaviour, which is radically reductionist—individual economic self-interest rules. In this view, history appears to be a continuum of ‘14th-century reality: civil war, plague and ignorance’. But these countries had their own 14th centuries and now find themselves in the 21st, playing a highly subordinate role in global capitalism. No understanding of how they got there can ignore the impact of colonization. In Collier’s models, colonial history is reduced to two numbers, one representing the colonial power—Britain, France, Portugal, Belgium, Germany—the other, the length of time that the country was colonized. The identity of the colonizing power does have a bearing upon the ex-colonial country’s legal system, educational set-up, lingua franca and financial institutions; but it tells us nothing about the pre-colonial system, the different processes by which the European power made its peace with local rulers; nor about the ending of colonial rule, and the extent to which ruptures or continuities determined the nature of the ex-colonial state. Such considerations help to inform a richer explanation of how a country has developed, and provide a deeper explanatory framework for civil wars, social conflicts or institutional forms—social and political questions, not purely statistical ones.

It is misleading to paint a picture of endemically low growth rates in sub-Saharan Africa, or in the other ‘bottom billion’ countries. In Africa, growth rates in the 1960s and early 1970s were comparable to those of Southeast Asia or Latin America. This was a period of African industrialization, based on import substitution; with improvements in economic management, this might have enabled several countries to take advantage of export markets. The droughts in the early 70s were a real blow, turning food self-sufficiency into food imports; most African economies were severely affected by the oil-price hikes, and still more so by interest-rate rises after 1979. But a possible industrial rehabilitation was stalled in the early 80s by World Bank and imf opposition to isi, and promotion instead of primary commodities and ‘getting prices right’. Collier fails to point out that African countries were forced to pursue primary-commodity exports as a consequence of World Bank conditionalities; as noted, these are not included in his list of explanatory variables. Collier takes the familiar line on trade, criticizing rich-country barriers to the exports of the ‘bottom billion’ but also, and far more fervently, tariffs set by developing countries. In his account, Africa simply ‘missed the boat’ in the 1980s and ceded global markets to Asia; only Mauritius—hardly a typical African country, if an African country at all—managed to ‘climb on board’. It was not simply low wages that attracted investment to the Asian Tigers, however, but their well-educated populations and skilled labour from the 60s on. Here again, a historically informed account of colonial and post-colonial social structures must be a factor in any satisfactory explanation.

Today, the international institutions have changed their minds again: the fact that resource-rich African governments misallocated windfall gains from primary-commodity price booms, and then suffered from slumps in prices, has led to the belief that diversification out of primary commodities is the appropriate strategy. Of course it always was, which is why countries pursued import substitution. That they did this badly in contrast to East Asia, whose ability to engage with world markets was precisely a consequence of prior isi, does not mean the policy was wrong then; but equally it means that an export-oriented strategy on the back of a failed isi is not likely to be successful now. Yet that is the strategy Collier presents. The problem is that primary commodities remain African countries’ chief export earners. Without these earnings stabilizing—or better, increasing—in the short run, they cannot accumulate resources for diversification. Collier’s argument also depends on a substantial ‘spillover’ factor, but this neglects historical evidence that poles of growth have not ‘spilled’ into neighbouring countries, or even into poorer areas of the same country.

The inter-relation of political and economical factors in African development is immensely complex—more so, perhaps, than in the long-stabilized societies of the rich world. Any explanation of why these governments have ‘failed’, if that is the appropriate term, requires some theorization of the post-colonial state and the forces that sustain it. How far is the state dominated by local as against foreign capital and in whose interests does it operate? How important are workers’ and peasants’ organizations, and to what extent have they been co-opted into ruling party and government structures? How should we understand contemporary manifestations of what John Saul once called the ‘dialectic of class and tribe’? In Collier’s model, the roadblocks and demonstrations of the mas in Bolivia would be dubbed ‘political violence’ and hence ‘bad for growth’. But mas has succeeded in renegotiating the prices paid by multinationals for Bolivian natural gas, and there has been a sustained increase in growth.

Fundamental to Collier’s agenda is the notion that globalization can work for Africa, as long as the ‘international community’ is willing to use its military and economic leverage to impose its preferred policies upon an unwilling set of native politicians, with the help of Western-educated local ‘heroes’ and a fresh set of carrots and sticks. Some may see Collier’s proposals as less a new start for Africa than a rehashed version of an all-too-familiar imperialism, albeit now of the ‘voluntary’ kind advocated by Robert Cooper, as more acceptable to a world of human rights and cosmopolitan values. On the evidence of The Bottom Billion and Wars, Guns and Votes, however, the intellectual case for the new developmentalist imperialism has yet to be made.