Return to Index

"Failed States and International Security II:
Sources of Prevention, Modes of Response, and Conditions of State Success and Renewal"

Purdue University, West Lafayette
April 8-11, 1999

Development in Fragile/Failed States

by Georg Sørensen

Department of Political Science
University of Aarhus
DK-8000 Aarhus C
Denmark
Tel: +45 8942 1258
Fax: +45 8613 9839
E-mail: georgs@ps.au.dk

We are living in a time and age where we expect 'development' (socio-economic, political, etc.) to be the rule and the lack of development to be the exception. According to this view, the issue for discussion in this paper concerns the identification of measures needed to get development underway in fragile/failed states. The implicit assumption is that development is always possible provided that the proper steps needed to promote it are taken. It is relevant to emphasise that this view is not a statement of fact; it is a piece of (development) ideology. We assume, and firmly believe, that development is possible in any country, no matter how adverse the circumstances may be. It is a matter of finding the proper measures and then get on with it. This ideology is fairly new; it was created in the context of post-World War II decolonisation. We only have go to back to the 1930s to find an entirely different ideology, dominating especially in the colonial motherlands. On that view, only some, maybe even rather few, colonies would ever be able to stand on their own feet and thus achieve development. A leading British authority on the colonies, Margery Perham, emphasised in 1941: "Africans must have foreign rulers and for a long time to come" (quoted in Betts 1998:16). The League of Nations explicitly recognised Mandates, areas to remain under 'tutelage' of 'advanced nations'. Most peoples in Sub-Saharan Africa were seen to require an "indefinite period of European tutelage" and the 'primitive' peoples of South West Africa and the Pacific were likely to "remain wards of the states-system for centuries, if not forever" (quoted in Jackson 1990:73).

The change between the traditional pre-war view and the contemporary post-war, developmentalist view, is not hard to understand. The traditional view was sustained by Western belief in its own superiority; "the African", Lord Lugard remarked, "holds the position of a late-born child in the family of nations, and must as yet be schooled in the discipline of the nursery" (quoted in Jackson 1990:71). When that belief in superiority was replaced by a more egalitarian outlook, it seemed natural to take the next step and consider 'development' to be just as possible and naturural there as it was and is here. That conclusion, however, is not warranted. Removing a paternalist outlook has nothing to do with empirical conditions for development of course. The possibility of development in the previous colonies cannot be taken for granted. Successful development requires certain preconditions. Those preconditions may be present or not present in any particular country, South or North. My argument in this paper is that some basic preconditions for development are largely absent in fragile/failed states. Unless they can be created, development on a significant scale will not be forthcoming. The underlying assumption is a pessimistic one: it might not always be possible to create those necessary preconditions for development (some of what follows draws on my previous writing on the subject, including Sørensen 1993; 1996; 1998).

Fragile states are deficient in several respects. The first problem is economic; there is a lack of a coherent national economy, capable of sustaining a basic level of welfare for the population. The poorest, least developed countries are in Africa south of Sahara, but there are also considerable pockets of poverty in Central America and Asia (Burma, Nepal, Bhutan). (It should be added that a relatively large part of the world's poor are in India, a country of enormous internal economic variation. Other large countries in the Third World (China, Brazil, Indonesia) also have many poor). The second problem in fragile states is political, concerning the institutions of the states and their legitimacy in the population. States that function well sustain a number of activities which are more or less taken for granted by their citizens: security against external and internal threat; order and justice in the sense of a functioning rule of law; and personal freedom including basic civil and political rights. Fragile states sustain such function only to a limited extent or not at all.

On the one hand, the institutions of the state are weak, lacking capacity, competence, and resources. On the other hand, power is frequently concentrated in the hands of state elites who exploit their positions for personal gain. The system is known in Sub-Saharan Africa as 'personal rule' or 'the strongman' (Jackson and Rosberg 1982; Sandbrook 1985). The most important positions in the state apparatus, whether it be in the bureaucracy, military, police or in the polity, are filled with the loyal supporters of the strongman. Loyalty is strengthened through the (unequal) sharing of the spoils of office. The strongman thus controls a complex network of patron-client relationships. The functions of the state only in a very limited sense has to do with producing public or collective goods. The state apparatus is a source of income for those fortunate or clever enough to control it. Such a state is by no means a source of security, order, and justice for its citizens; it is more of a threat, an apparatus against which the population must seek protection.

Against that background there has to be a lack of legitimacy. Vertical legitimacy is low because large parts of the population have no reason to support the government; and the government has no authority in the sense that people support or follow its rules and regulations. Horizontal legitimacy-peoples' sense of belonging together in a nation-is also low because the state is captured by specific groups; it is not a state for the whole people. Christopher Clapham has emphasised that such systems comprehensively lack "the capacity to create any sense of moral community amongst those who participate in them, let alone among those who are excluded" (Clapham 1996:59).

There is a partial overlap between lack of economic substance and lack of political-institutional substance. But not all economically wanting states are weak in political-institutional terms; Uruguay, Chile or Costa Rica are Less Developed Countries (LDCs) economically, but they are not weak states (although they may feature some elements of weak statehood). The reverse also holds; not all states that are weak in political-institutional terms are LDCs. Yugoslavia earlier and Bosnia today, as well as Russia are in many respects weak without being LDCs. Yet in most cases there will be an overlap, so that states which are politically-institutionally weak are also LDCs. It is in this category that we find those states that are in danger of becoming 'failed states':

The concept of fragile state is an ideal type. Real-world states approximate that type in varying degrees. Most of the fragile states are in Sub-Saharan Africa. But the least developed Central American states and the Central Asian states coming out of the former Soviet Union and even some states in Europe (e.g. Albania) share many of the characteristics of fragile states outlined here.

State failure means that a state breaks down in decisive respects. Compared to fragile states, breakdown-again in ideal typical terms-is a matter of degree: the problems related to fragile statehood magnify. That means the distinction between fragiled state and failed state is blurred: failure is when fragility intensifies. William Zartman describes failure as follows: "As the decisionmaking center of government, the state is paralyzed and inoperative; laws are not made, order is not preserved, and societal cohesion is not enhanced...As a territory, it is no longer assured security by a central sovereign organization. As the authoritative political institution, it has lost its legitimacy...As a system of socioeconomic organization, its functional balance of inputs and outputs is destroyed..." (Zartman 1995:5).

Describing fragile/failed state indicates the absence of those basic preconditions that must be established for development to take place: first and foremost a state machinery and a government that can help push and sustain such a process. The crucial importance of the state for development is the starting point for the deliberations which follow. This view of the state should be differentiated from two alternative views that I reject. One is the radical neo-liberal view which holds that bureaucrats in particular and states in general contribute nothing to development; sometimes they are not even considered 'zeroes' but 'minuses' in the development equation. The neo-liberal view is wrong; it is sufficient here to recognise that no successful process of development has taken place in this century without a critical role played by the state. Markets have not created development on their own anywhere.

The other view that I reject is diametrically opposite of neo-liberalism. It is the statist view. Statists assume, without prior questioning, that states are by definition good for development. The implication is 'the more state the better', i.e. maximising the role of the state mean maximising the positive developmental influence of states. When we reject both statism and neo-liberalism, the pertinent research questions come into view: Why are the preconditions for a positive role of the state in development not present in failed/fragile states? And: Is it possible to bring about a situation where those preconditions are present?

A complete discussion of the role of the state in development ought to include, as a minimum, three topics: (a) state autonomy, i.e. the proper degree of freedom of manoeuvre for the state; (b) state capacity, i.e. an efficient bureaucratic machinery guided and shielded by by a political elite that gives priority to development; and (c) statecraft, i.e. the ability to formulate proper policy responses to given development challenges. I cannot cover all three topics in this paper; focus will be on state autonomy with some excursions to the two other issues.

State Autonomy and Development

As indicated, state autonomy is freedom of manoeuvre for state elites and their bureaucratic machineries. (This really requires a discussion of relations between state elites and bureaucrats but I cannot go into such an 'opening up' of the state here). As noted by Theda Skocpol, a minimum of such autonomy means that states "may formulate and pursue goals that are not simply reflective of the demands or interests of social groups, classes, or society" (Skocpol 1985:9). We might add: goals that are not simply reflective of the demands of external actors, be it economic or political interest groups. Maximum autonomy is a situation where states are completely independent of internal and external interests. Such a possibility is purely theoretical, but there have been situations approximating it, for example the rein of Mao Zedong in the People's Republic of China in the 1960s; he could do almost anything he wanted with nearly a billion people completely at his mercy.

Given there are different degrees of state autonomy what is the appropriate amount of autonomy in terms of maximising a positive contribution to development? Both extremes, too little and too much autonomy, appear to be counterproductive. Too little autonomy produces a captured state responding to the narrow needs of its interest group masters but not to a larger development agenda; to much autonomy produces a freewheling, unrestrained state, where things can go very wrong, even when there are good intentions on part of the state elite, as demonstrated during the so-called Great Leap Forward and the Cultural Revolution in China (Sørensen 1991).

My contention is that fragile/failed states are plagued by a special kind of 'captured autonomy'; such states are autonomous in the sense that they are not significantly constrained by forces outside of the state apparatus; yet the state is captured in the sense that the elite controlling the state is exploiting that control for benefit of its own narrow interests. The long rule of Mobutu in Zaire is a frequently used example of this situation. After all, he was one of the richest men in the world when he finally died. What gave Mobuto an extreme measure of autonomy? Domestically, there were no social groups outside of the state sufficiently strong to challenge his position. Yet Mobutu's neo-patrimonial state was not about development; it was about enriching himself and a small group of followers. His state services were up for sale to the highest bidder in context of patron-client relationships. It is sometimes said that the state autonomy required for development is especially autonomy from classes and groups involved in zero-sum activities, that is, speculation, corruption, usury and the like. That is, "classes which derive wealth from unproductive activities or which are otherwise hostile to industrial development" (Hamilton 1987:1243). But in Zaire those groups were exactly Mobutu himself and his clique, the state elite was thus part of the development problem and not at all of the solution. When the political elite is itself the strongest zero-sum group in society we can hardly expect it to act as a developmental state in any meaningful way. That kind of autonomy is a real problem for development.

What turned Mobutu into a predator? I have noted the lack of domestic constraint on him, but what about the international context? Many people will think about Mobutu as highly constrained by external forces. He was head of a desperately poor country, economically and politically dependent. This is all true. There is even a book by Sean Kelly called 'America's Tyrant' (1993), which records how Mobutu was dependent on the CIA which saved him from coup attempts by rival military factions on more than one occasion. One explanation for Mobutu then, is that he could conduct his dirty deals under CIA protection. He was a puppet of imperialism and that's the end of that. His autonomy was all domestic and not in the least external.

There is a lot of truth in that story but it is misleading in a basic sense. That is because it stipulates that Mobutu was completely in the pocket of the CIA. He was not, for the simple reason that he was the leader of a sovereign state. Formal independence is often downplayed by political economists, but that is a mistake. Formal independence, that is, juridical sovereignty, is of the utmost importance. At the moment of independence a new political, economic, social, and cultural sphere is created which has some substantial amount of autonomy. This new 'inside' of domestic sphere can still be influenced by external forces of course, but the conditions of operation are very different from before. On the one hand, there is a new need on the part of outsider for finding domestic allies; that implies some sort of bargaining situation between insiders and outsiders. Mobutu was not merely in the pocket of the CIA, he bargained with them: you do this for me, I do this for you; he had bargaining autonomy. On the other hand, interventions in sovereign states cannot be conducted in complete ignorance of the rules of international society. After all, the basic norm of juridical sovereignty is non-intervention which means that acts of intervention have to be justified. So both in the domestic and in the international sphere the rules of the game change in ways which provide increased autonomy to domestic actors.

Development of international norms after the Second World War has increased the value of juridical sovereignty. The reason is simple: in the old days states were constantly at each others throats. War between states was an important aspect of state building as emphasised by Charles Tilly's well-known phrase: 'states made war and war made states'. An important ingredient in war was the conquest of enemy territory: the stronger swallowed the weaker. We have countless examples of this in Europe. According to international norms after WW2 however, borders are sacrosanct. They can only be changed with the consent of the affected parties. That new norm has provided additional autonomy for weak states. No matter how weak, they will not be swallowed by stronger states. Yet this can also be a free pass for predatory elites to run states into the ground: no matter what the extent of misery and dissolution, we still pretend that there is a Somalian state. The weak are no longer swallowed, they are merely allowed to disintegrate. In other words, the realist logic of states having to successfully compete for survival in an anarchic international system where the strong consume the weak is replaced by a new logic according to which predatory state elites can go to any extreme in terms of creating chaos and violent conflict and lack of development without paying the ultimate price: termination of the state.

Compare Mobutu to Jiang Kai Shek on Taiwan in 1950. Jiang had perhaps even more domestic autonomy than Mobutu. He had just arrived from the Mainland after being badly beaten by Mao, and thus had no constituency on Taiwan. At the same time, he was an even bigger rascal than Mobutu, responsible for even more corruption and killing of innocent people during his time on the Mainland. In short, a disgusting fellow with as much domestic autonomy as Mobutu. Why was he not a predator? What made him into one of the developmental state heroes, heading the perhaps most singularly successful development model in the whole postwar era? The simple answer is external pressure, two kinds of it. The first was from the Mainland: Jiang was afraid Mao would come and eat him for breakfast. The second was from the Americans. Dismayed with his performance on the Mainland, the Americans gave Jiang an ultimatium: behave decently or lose our support. They backed it up with advisors, pushing for agrarian reform and a host of other development measures.

That brings us to an important conclusion: External pressure or constraint which reduces full autonomy of the state is not in itself negative or counterproductive. It need not lead to underdevelopment. It can indeed lead to development. External pressure in itself is not the problem. It all depends on the concrete content of that pressure. In the absence of domestic constraint, of which I will have more to say below, external pressure or constraint is the only remaining disciplining force. If domestic forces are unable to discipline the state, external forces need to take on the job. The alternative is predation and maldevelopment.

External pressure on the state: the road to development?

External pressure, then, is frequently necessary. What is the content of the optimal external pressure? The abstract answer is straightforward: the optimum external pressure is that which provides maximum push and pull in the right direction. That means pursuing policies which help solve development problems in general and captured autonomy problems in particular. One item is stopping all support for predatory state elites for reasons of narrow national economic or political interest, and providing additional support for state elites moving in the right direction. Such policies are at least partially forthcoming, under the umbrella of 'good governance'. Increasing use of political conditionalities in the context of giving aid are an improvement; at the same time, old friendships die hard. In a world of sovereign states which must look after themselves in basic ways, we cannot expect that national economic and political interests will cease to play a role. The logic of sovereign statehood is to look after your own state's best interests first. This can still (fortunately not always) lead to some support for predatory state elites as demonstrated, for example, by the winter 1995 difficulties of deciding an oil boycott against the Nigerian military. Yet the overall tendency is at least for a more consistent support than during the Cold War to state elites conducting decent attempts to promote development.

Meanwhile, we cannot envisage a situation where external support in the form of economic aid comes to more than a tiny fragment of the resources of strong, substantial states. Denmark takes pride in spending more than 1 percent of its GNP on development aid; it is much more than most countries, which are around the 0.3 percent average for OECD. Yet Danish external aid is nowhere near the resources spent on internal redistribution from the Danish rich to the Danish poor. Such redistribution amounts to more than 30 percent of GNP, or more than a hundred times the average spent on external aid. Nobody would expect any radical revision of these percentages. That is another instance of the sovereign state first and foremost taking care of itself.

Establishing the maximum external pressure could mean the creation or re-re-creation of external threat in more fundamental ways. In operational terms it means that international society accepts secession and the creation of new states in a much more radical way than has been the case so far. The idea appears attractive; risks of secession would push state elites towards getting their acts together and better accommodating ethnic and other groups. Robert Jackson makes the point in a recent analysis:

the international support of existing African jurisdictions sustains two important effects: (1) the perpetuation of domestic and personal insecurity by blocking the formation of jurisdictions that might be less arbitrary, more cohesive, more legitimate, and therefore better protectorates of their populations; and (2) the perpetuation of underdevelopment (Jackson 1992:91).

Yet given the current situation in many weak states, secession could also lead to even greater problems. Both in the former Soviet Union and especially in ex-Yugoslavia, secession and the drawing up of new borders has been accompanied by massive acts of so-called ethnic cleansing (Bennett 1995). In Sub-Saharan Africa, the creation of many new states could lead to even higher levels of human suffering. The question is also where to stop; which groups should be allowed the create new states following which specific criteria? A recent study on state collapse in Africa is very cautious on this point, emphasising that "the logic of secession works against seceding states, threatening an infinite regress of self-determination" (Zartman 1995:268).

It is a genuine dilemma: the old borders are holding together groups that fight against each other because they cannot agree on forming a political community. The result is insecurity and underdevelopment. Splitting them up appears to create more problems than it solves. The cannot stay together and they cannot split up; damned if you do and damned it you don't. If there is a way out, it would seem to entail that the international community clarified its policies as regards secession, setting forth unambiguous conditions under which groups have the right to secede. Such standards must be combined with a more consistent enforcement of international law protecting the rights of ethno-political groups as well as an improved machinery for preventive diplomacy in the case of ethno-political conflict (see the proposals in Gurr 1994).

Exercising the proper amount and type of external pressure will often be difficult because, as already indicated, in a world of sovereign states other considerations are bound to play a role. This is also clear in the case of state failure. On the one hand, there must be limitations to the extent of human suffering that international society can tolerate without trying to do something. If basic human rights are more than articles on a piece of paper, they require some sort of action when governments or guerrilla groups conduct massive attacks on civilians. That was at least the idea behind Boutros Boutros-Ghali's 1991 statement about the duty of states to meet human rights or face intervention: "the time of absolute and exclusive sovereignty ... has passed. Its theory was never matched by reality" (quoted from Helman and Ratner 1992-3:10). Yet sovereignty is not easily set aside. Article 2 of the UN charter emphasises that "Nothing contained in the present charter shall authorize the United Nations to intervene in matters which are essentially within the domestic jurisdiction of any state". Put differently, there have to be compelling reasons for setting aside the principle of non-intervention.

The third element in the intervention equation is the national interest. The countries that conduct such intervention are compelled to look after the security of their own soldiers. No sovereign state would probably accept large casualties in the name of humanitarian intervention. There are also more parochial elements in the national interest, such as economic, political and related interests. All in all, humanitarian intervention that is worthy of the name in the sense that the action undertaken is humanitarian both in its motivation, it implementation, and its outcome is really hard to find, as demonstrated in a recent analysis by Adam Roberts (1996). Roberts concludes that the "twin principles of sovereignty and non-intervention remain fundamentally important" and "only in very extreme cases...can they be directly overridden by overt cases of humanitarian intervention" (Roberts 1996:30).

In sum, external pressure, sometimes including intervention, is needed in order to avoid predation and underdevelopment. But external pressure in a society of sovereign states is subject to the limitations following from the fact that considerations concerning national interest and international law will always be involved. So even if external pressure may be direly needed it may often not be forthcoming in the optimal ways as seen from the point of view expressed here, namely the need to confront the problem of captured autonomy in fragile/failed states. In a world of sovereign states, then, the pressure needed to discipline the state will most often have to come from domestic sources.

Domestic pressure on the state: the road to development?

From previous scholarship including Peter Evans (1995), Richard Sandbrook (1985), Gordon White (1984) and many others, as well as from the examples given above, we know that a very high degree of domestic autonomy may very well be counterproductive in development terms; it can easily lead to predation. To be developmental, states not only need a measure of autonomy, they also need to be embedded in society. Embedded in what way? The simple answer is: the developmental state is connected to social classes and groups pushing for economic development and not connected to social classes and groups impeding economic development. Who is in the latter category? That has already been indicated: the classes and groups involved in zero-sum activities, such as traditional agrarian powerholders, foreign and domestic speculators. We saw that part of the state's problem is that there may be zero-sum groups in the state itself, such as sections of the military or the bureaucracy or some state enterprises.

Who is in the former, development-promoting group? According to Peter Evans, they are first and foremost domestic industrialists: "Connections that privilege industrialists allow the developmental state to focus on a project of industrial transformation, to keep its involvement selective, and to avoid having its bureaucratic capacities overwhelmed" (1995:234). In a broader perspective, state ties to popular groups such as associations of workers and peasants, are also important. While industrial ties tend to produce growth, broader popular ties tend to produce welfare, and we can count on both achievements as a part of a broader development process.

How is the general situation in terms of domestic autonomy in the Third World? In gross oversimplification, it appears that East Asian state have been connected to the industrialists, a few states here and there have been connected to popular forces (Costa Rica, Kerala in India, maybe some of the Latin American countries today), a great many states in Asia and in Latin America are at least partially (frequently more than that) connected to zero-sum groups, and the Sub-Saharan African states are generally only embedded with themselves and that has most often led to predation. There is a caveat here which should be drawn out in the open. The notion of 'embedded' or 'connected' as a description of the relationship between the state and groups in society is potentially unclear, to say the least. What does such embeddedness actually mean? Peter Evans emphasises that it does not entail a dependency relation where the state is merely expressing the interests of groups in society. 'Embeddedness' means that on the one hand, state elites and their bureaucracies are able to impose decisions on groups in society, for example the introduction of competition to improve efficiency and quality. On the other hand, 'embeddedness' means connections which provide channels of information and which pave the way for effective implementation: "Efficacious states combine well-developed, bureaucratic internal organization with dense public-private ties. The recipe works only if both elements are present" (Evans 1995:72). In other words, the connections implied in 'embeddedness' contain a special kind of state power which has nothing to do with coercive power. It is sooner what Michael Mann has called "infrastructural power", i.e. "the institutional capacity of a central state...to penetrate its territories and logistically implement decisions. This is ... 'power through' society, coordinating social life through state infrastructures" (Mann 1993:59).

Defining embeddedness in this way is clarifying, but it also points to the difficulties which fragile states confront in terms of meeting those conditions of 'proper' embeddedness. The primary problem, of course, is that the groups of social actors required for optimum embeddedness are simply not present in fragile states. On the state side, there is not a bureaucracy which features "well-developed bureaucratic internal organization". In contrast, there is more often a disorganised, incompetent and corrupt bureaucracy. On the civil society side, it is clear that the presence of domestic industrialists which is so important according to Peter Evan's analysis (which faithfully reflects the conditions for successful development in East Asia) is a condition which is not met in most fragile states for the simple reason that in order for groups of industrialists to exist a certain basic level of industrial development is a necessary pre-condition and that pre-condition is not met by most fragile states.

But then there are at least groups of workers and peasants with whom the state can connect in productive, developmental ways, even in fragile states? Such groups are indeed present, but very often they are small and weak. Because industrial development has not progressed very far or even at all, there is no real industrial working class. Most 'workers' in the city are in services, or in petty trade. In the countryside, most people are not peasants in any classical sense; they are villagers which scrape off a minimum subsistence level from the land. In short, the notion of proper embeddedness might not get off the ground at all in fragile states because the very groups in state and civil society required to forge those links of embedded autonomy are either not present, or, more frequently, so weak that the proper domestic pressure on the state, as expressed in the notion of embedded autonomy, can not emerge on any significant scale.

Confronting these problems forces us to think of problems of socio-economic development and problems of political development as part of a larger whole which must be thought about in a more coherent manner than is sometimes the case when using the concept of 'embedded autonomy'. We need to combine the ideas about a strong, developmentalist state, better capable of promoting economic and social development, with the idea of a more democratic state which is responsive to society, legitimate, and under the rule of law. Many observers used to see strong, developmentalist states as basically non-democratic, using the experience of the authoritarian developmentalist states in East Asia (especially South Korea and Taiwan) as the empirical reference. This view was inspired by Chalmers Johnson's emphasis on the need for 'soft authoritarianism' which could provide political stability and order, particularly in the sense of avoiding political pluralism which might challenge the goals of the developmentalist elites. According to Johnson, Japan showed the way in this respect, in displaying "an extremely strong and unsupervised state administration, single-party rule for more than three decades, and a set of economic priorities that seems unattainable under true political pluralism during such a long period" (Johnson 1987:137).

Yet it was never quite clear that authoritarianism was a necessary element in a developmentalist state (Hamilton 1987; Baeg Im 1987). And comparative analysis can quickly reveal that the special conditions surrounding a 'soft authoritarian' developmentalist state are not present elsewhere, least of all in Africa. In other words, any attempt to make a general claim on the basis of the East Asian experience that authoritarianism will invariably help create a strong, developmentalist state must be rejected. The different variants of 'soft' and 'hard' authoritarianism found in Africa, Latin America, and South Asia have all failed to help generate a developmentalist state (Sørensen 1993).

Against this background, several scholars have started thinking about the combination of the merits of a strong, developmentalist state and a more democratic, responsive state. One example is the recent work by Merilee Grindle (1996). According to her analysis, a strong state in the developmentalist sense has the following capacities: institutional capacity; technical capacity; administrative capacity; and political capacity where the latter includes legitimate authority and responsive and representative government. A similar view, in most respects, emerges in recent World Bank report on the state in a changing world (1997): the bottom line of the analysis is that the state needs to be effective as well as responsive; i.e. developmentalist and democratic. The underlying idea of this combination is that democracy will help discipline the state, making it more responsive to society, less inefficient, corrupt and incompetent. A more effective state, in turn, can take the lead in promoting development processes which will help change society and thereby help 'produce' groups more conducive to rapid socio-economic development..

This may sound very good in theory, but practice has been somewhat less successful. Early elections in fragile states have not produced the expected results. Rather, they have superimposed a thin layer of democratic coating upon a system of personal rule without major changes in the basic features of the old structure. According to one astute observer, "Elections appear to be the wrong place whence to start a process of democratization in a collapsing, conflict-ridden state. In recent years, African elections have typically been organized in a hurry, in some cases before parties had time to consolidate or armed movements had agreed to disarm. As a result, losers have found it easy to reject election results, and voters had little choice but to vote on the basis of ethnic or religious identity...Elections held under wrong conditions can be a real setback for democratization (Ottaway 1995:245; for a similar view see Dahl 1992 and Elklit 1994).

The other major drawback of a democratization which focuses on elections is that there is little change for the better in economic politics of the new regimes. A recent analysis found that "elections may actually increase the use of patronage...Traditional patron-client relations have often been critical in winning recent elections, indicating that the nature of African politics has not changed despite the new liberalization. Ghana, Nigeria, and Kenya have all reported massive overspending as governments sought to reward traditional supporters, notably members of particular ethnic groups and civil servants, to smooth the transition process or to gain votes...The particular circumstance of political liberalization in Africa cause leaders' horizons to be relatively short and therefore to induce particular strategies such as clientelism which may be unnecessary where democratic structures are more institutionalized" (Bienen and Herbst 1996:38-9).

In sum, democratisation appears to be a sound long-term answer to the question of how to create the optimum domestic pressure which can substitute a situation of captured autonomy for one of developmentally conducive embedded autonomy, but democracy takes a considerable amount of time to develop, especially where the societal conditions are adverse, as is mostly the case in fragile states. Therefore, early results from scattered electoral processes should not be expected. Democracy cannot be installed overnight. There might even be a need for developing new models of democracy different from the standard liberal models of the West, and better suited to fragile state conditions, but this is by no means an easy task.

Conclusion: the long road to development and the risk of failure

This paper has identified captured autonomy as the major impediment to development in fragile states and has speculated about the possibilities for external and internal pressures to change the situation and thus create better conditions for development. An overall conclusion will have to be basically pessimistic; while there are both external and internal pressures at work, they appear, for a variety of reasons, to be insufficient to improve substantially on the situation in fragile states. In other words, captured autonomy may continue to characterise fragile states and therefore development will not be forthcoming.

Given the domestic situation in most fragile states, where state elites are most often either too autonomous or, alternatively, closely connected to the wrong groups in society, there is, in the short and medium run, no alternative to external pressure and thus to external constraint. If domestic forces are unable to discipline the state, external forces need to take on the job. The alternative is predation and maldevelopment. What does this mean in concrete terms? It means, for example, that the IMF should not continue to bail out the ruling oligarchy in Russia because it only postpones the day of reckoning for a sick economy and a predatory elite. Instead, it should apply maximum pressure for real economic and social reform. In Sub-Saharan Africa, it means that international institutions and individual donors should sustain the pressure for achieving better governance. What we should do is not to argue for an easing of the pressure. Instead, we should point out all the instances where policies are short-sighted, counterproductive, and in pursuit of narrow donor-interests. The struggle then, is for changing the content of the pressure for good governance in the direction of the best possible measures. It is not for removing the pressure, because there is no viable alternative to it in the current situation.

In the longer term, however, the domestic situation must change. I have pointed to democratisation as the long-term solution to the problem of captured autonomy, but the short and medium-term obstacles to democratisation have also been identified. Some basis for optimism can be found in countries that have been successful in spite of the adverse conditions they have faced; it is relevant in this context to briefly introduce the cases of Botswana and Mauritius. On independence in 1968, Mauritius was a poor sugar-economy with deep ethnic cleavages in the population; Botswana was a cattle-economy with a population divided into eleven different tribes. The country had the good fortune of discovering diamonds on its soil, but several other African countries have had rich mineral deposits and have still been unable to convert that potential into broader development. How could Mauritius and Botswana succeed economically and politically (in establishing functioning political democracies)? A recent analysis identifies the following factors behind that economic and political success:

the fact that talented political leaders were personally committed to democratic government, and to economic development; the creation of a competent, politically independent state bureaucracy with personnel policies based largely on merit, but with a composition that is reasonably representative of their societies; the development of a public realm that is capable of imposing at least modest checks on the actions of the state, and that is characterized by a balance between universalistic and particularistic norms, and by a pragmatic recognition of the important representative role of tribal/ethnic organizations and institutions (Carroll and Carroll 1997:470).

These are interesting and convincing answers to the question posed above; yet it is also clear that such answers beg new questions. Where do talented leaders committed to democratic government and economic development come from? As indicated by Carroll and Carroll, it is reasonable to assume that success cultivates success: once a competent leadership has been established and has demonstrated a decent track record chances are good that capable leadership will continue. But what about that crucial turn-around phase, where success is by no means secure and leadership might as well turn out to be narrow-minded, egoistic, and self-serving? What is it that brings forward the Mandelas instead of the Mobutus? If it is not pure coincidence then the question merits further research. The hopeful answer that such leadership is more or less automatically created by holding elections has not been confirmed by events. It should also be emphasised that good and honest leaders are not enough, especially if they are committed to bad policies. Julius Nyerere of Tanzania is an honest man who has done much good for the country, but his policies of a state-led economy and a basically non-democratic polity also led to disastrous results.

As for the second item, a good bureaucracy, it is clear that early decisions were made in both Botswana and Mauritius not to sacrifice competence for 'nativization' of the public service and to base recruitment on merit. At the same time, however, Carroll and Carroll stress that the Weberian ideal of an impartial public service probably should be abandoned when it comes to developing countries with many different ethnic groups in the population. In the two countries examined, care has been taken to make the bureaucracy representative of subgroups in society without sacrificing merit. Such a representative bureaucracy "is more likely to consider a wide range of views and interests in making decisions. Indeed, the simple existence of a representative bureaucracy if often taken by the public as evidence that the state is responsive and legitimate" (Carroll and Carroll 1997:473).

The third item above concerns the need for a civil society to constrain the state. In many poor Third World countries there is no civil society in the traditional Western sense of the word. At very low levels of development, there is no business class, no middle class, not even a well defined class of peasants. Partly as a consequence of this, there are few autonomous, strong secondary organizations based on universal membership criteria. According to a recent analysis by Goran Hyden, "the prime contemporary challenge is how to restore a civic public realm. The trend of postindependence politics in most African countries has been to disintegrate the civic public realm inherited from the colonial powers and replace it with rivaling communal or primordial realms, all following their own informal rules" (Hyden 1992:23). Yet the good news from Botswana and Mauritius is that these communal and primordial realms can act as at least "a modest check on the power of the state" (Carroll and Carroll 1997: 479). In other words, under favorable circumstances traditional social forces can perform some of those functions that we would normally expect demand the presence of a more fully developed civil society. That is, ethnically divided societies can sustain democracy even though their civil societies have been weak, and ethnic divisions continue to exist. Perhaps the main message from the experience of these countries is the importance of competent leadership and some measure of institutional innovation. None of the countries may live up to the highest demands of liberal democratic procedures and institutions, but they should be credited with substantial democratic success anyway.

In overall conclusion, I would like to bring in the micro-perspective. My discussion so far has been excessively 'macro', focusing on the larger structures and processes of development. But 'macro' is basically a compilation of many 'micro' processes. Zooming in on them leads to a more nuanced view of what's going on in fragile states; even the weakest states have strong spots; sub-departments, institutions, offices that are staffed with fairly well-educated, competent, and honest civil servants. In every society, there are groups and organisations, however small, that do what they can to promote development. Bringing such elements of 'strong state' and 'strong society' together can create 'state-society synergies' (cf. Evans 1997) which help promote local processes of development. Examples of this can be found in any contry, including any fragile or failed state. And so there is always a place to begin; what emerges as maldevelopment in the overall macro picture may contain several constructive micro-elements of successful development efforts. Strengthening and supporting those is the big challenge; the only way ahead is the long haul. There are no quick fixes when it comes to development in fragile and failed states.

References

Baeg Im, H. (1987). "The Rise of Bureaucratic Authoritarianism in South Korea", World Politics, vol. XXXIX, no. 2.

Bennett, Christopher (1995). Yugoslavia's Bloody Collapse-Causes Course and Consequences, London: Hurst and Company.

Betts, Raymond F. (1998). Decolonization, London: Routledge.

Bienen, Henry and Jeffrey Herbst (1996). "The Relationship between Political and Economic Reform in Africa", Comparative Politics, October, pp. 23-42.

Carroll, Barbara Wake and Terrance Carroll (1997). "State and Ethnicity in Botswana and Mauritius: A Democratic Route to Development?", The Journal of Development Studies, vol. 33, no. 4, April, pp. 464-86.

Clapham, Christopher (1996). Africa and the International System. The Politics of State Survival, Cambridge: Cambridge University Press.

Dahl, Robert A. (1992). "Democracy and Human Rights under Different Conditions of Development", pp. 235-52 in Asbjørn Eide and Bernt Hagtvet (eds.), Human Rights in Perspective, Oxford: Blackwell.

Elklit, Jørgen (1994). "Is the Degree of Electoral Democracy Measurable?", pp. 89-11 in David Beetham (ed.), Defining and Measuring Democracy, London: Sage.

Evans, Peter (1995). Embedded Autonomy. States and Industrial Transformation. Princeton: Princeton University Press.

Evans, Peter (1997). "Sustainability, Degradatoin and Livelihood in Third World Cities: Possibilities for State-Society Synergy", paper for UN21 Project, New York: UN.

Grindle, Merilee (1996). Challenging the State: Crisis and Innovation in Latin America and Africa, Cambridge: Cambridge University Press.

Gurr, Ted Robert (1994). "Peoples against states: ethnopolitical conflict and the changing world system", International Studies Quarterly 38:3.

Hamilton, Clive (1987). "Can the Rest of Asia Emulate the NICs?", Third World Quarterly 9:4.

Hyden, Goran (ed.) (1992). Governance and Politics in Africa, Boulder: Lynne Rienner.

Jackson, Robert H. (1990). Quasi-states: sovereignty, international relations and the Third World, Cambridge: Cambridge University Press.

Jackson, Robert (1992). "The Security Dilemma in Africa", in Brian Job (ed.), The Insecurity Dilemma. National Security of Third World States, Boulder: Lynne Rienner.

Jackson, Robert and Carl G. Rosberg (1982). Personal Rule in Black Africa: Prince, Autocrat, Prophet, Tyrant, Berkeley: University of California Press.

Johnson, Chalmers (1987). "Political Institutions and Economic Performance: The Government Business Relationship in Japan, South Korea, and Taiwan", pp. 136-65 in Fred C. Deyo (ed.), The Political Economy of New Asian Industrialism, Ithaca: Cornell University Press.

Kelly, Sean (1993). America's Tyrant. The CIA and Mobutu of Zaire, Washington: The American University Press.

Mann, Michael (1993). The Sources of Social Power, vol II. The rise of classes and nation-states, 1760-1914. Cambridge: Cambridge University Press.

Ottaway, Marina (1995). "Democratization in Collapsed States", pp. 235-51 in I. William Zartman (ed.), Collapsed States. The Disintegration and Restoration of Legitimate Authority, Boulder: Lynne Rienner.

Roberts, Adam (1996). Humanitarian Action in War, ADELPHI Paper 305, IISS, Oxford: Oxford Univertisty Press.

Sandbrook, Richard (1985). The Politics of Africa's Economic Stagnation. Cambridge: Cambridge University Press.

Skocpol, Theda (1985). "Bringing the State Back In: Strategies of Analysis in Current Research", in P.B. Evans et al. (eds.), Bringing the State Back In, London: Cambridge University Press.

Sørensen, Georg (1991). Democracy, Dictatorship and Development: Economic Development in Selected Regimes of the Third World, London: Macmillan.

Sørensen, Georg (1993). "Democracy, Authoritarianism and State Strength", pp. 6-35 in Georg Sørensen (ed.), Political Conditionality. London: Frank Cass.

Sørensen, Georg (1996). "Development as a Hobbesian Dilemma", Third World Quarterly, vol. 17, no. 5, pp. 903-16.

Sørensen, Georg (1998). Democracy and Democratization. Processes and Prospects in a Changing World, 2nd. ed. Boulder: Westview Press.

White, Gordon (1984). "Developmental States and Socialist Industrialisation in the Third World", Journal of Development Studies, 21:1, pp. 97-120.

World Bank (1997). World Development Report 1997. The State in a Changing World, New York: Oxford University Press.

Zartman, I. William (1995). "Introduction: Posing the Problem of State Collapse", pp. 1-15 in I. William Zartman (ed.), Collapsed States. The Disintegration and Restoration of Legitimate Authority, Boulder: Lynne Rienner.