Targeting and strategic intervention

Paul Spicker

Department of Political Science and Social Policy University of Dundee Dundee DD1 4HN United Kingdom E-mail: p.spicker@dundee.ac.uk
Paper presented to the 2nd International Research Conference on Social Security, Jerusalem, 25-28th January 1998.
This paper is a theoretical examination of the relationship between targeting, selectivity, residual welfare, incentives and strategic intervention. The relationship between the issues is often taken for granted: residualism implies selectivity, selectivity is a form of targeting, and both targeting and incentives have played an important role in the development of the New Right's arguments for the restriction of benefit. But the principles on which policy is made do not determine the focus of policy, and the focus of policy does not determine the pattern of response. Selectivity is not necessarily residualist, targeting is not the same thing as selectivity, and arguments about incentives are independent of arguments about targeting. An understanding of strategic intervention helps to clarify both the links between the concepts and the distinctions which need to be made in their application.
Criticisms of targeting and incentives are often based in a general concern about the type of welfare system they seem to imply. Targeting is associated with selectivity and residual welfare. Selectivity, Townsend argues, fosters hierarchical relationships of superiority and inferiority in society, diminishes rather than enhances the status of the poor, and has the effect of widening rather than reducing social inequalities. Far from sensitively discriminating different kinds of need it lumps the unemployed, sick, widowed, aged and others into one undifferentiated and inevitably stigmatised category.1 Selectivity is, by this account, a right-wing approach, adopted by those who wish to differentiate the poor from the rest of society. Arguments about incentives are used as a means of restricting benefits to those who are most vulnerable. They are part of a discourse which represents poverty and dependency in terms of the individual choices of the dependent person. "Within such discourse", Hartley Dean writes, 'incentives' underwrite the fitness of the individual to survive; compulsion is the proper guarantee of productivity and independence; punishment is the proper response to idleness and dependency.2 The ideological link between these concepts is so well established that there is a danger of assuming that, in some way, the concepts are necessarily connected. In this paper, I wish to argue not simply that they are not necessarily connected, but that there is no direct route from these concepts to the kind of right-wing ideology with which they are most frequently associated. The idea that targeting, selectivity, incentives, or even residual welfare are necessarily punitive or stigmatising is based in a muddle - an understandable muddle, given the way these concepts have been associated historically, but a muddle nevertheless. When the concepts are unravelled, it becomes possible to think of the methods and approaches involved in different ways.
P Townsend, 1976, Sociology and social policy, Harmondsworth: Penguin, p.126. H Dean, 1991, Social security and social control, London: Routledge, p.183. 2
Distinguishing concepts Residualism requires selectivity, and selectivity is a form of targeting. They are part of a family of closely related, interlinked concepts, and as such they can be difficult to distinguish in practice. The basic method which I propose to use to distinguish between them is an application of set theory. For any two sets A and B, ! if there is no case where an item belongs to A and not B, and there is no case where an item belongs to B and not A, then the two sets are identical. ! if there is at least one item which belongs both to A and B, and there is a case where an item belongs to A and not B, and there is a case where an item belongs to B and not A, then the two sets overlap and are not identical. ! if there is at least one item which belongs to both A and B, and there is no case where an item belongs to A and not B, and there is at least one case where an item belongs to B and not A, then A is a subset of B. ! if there is no case in which at least one item belonging to set A also belongs to set B, the two sets are distinct. The figure below shows two overlapping sets (A and B), two sets which are distinct (A and C) and one set which is a subset of another (C is a subset of B).

Set Theory

(A + ~ B) (A + B)

(B + ~ A)

(B + C)
To show, for example, that residualism and selectivity are discrete, it is necessary to show that - there is at least one case in which a residual measure is not selective, and - there is at least one case in which a selective measure is not residual. Residual welfare The term 'residual welfare' was initially coined by Wilensky and Lebeaux, in contrast with the idea of institutional welfare.3 It refers in the main to welfare which is offered as a safety net, for those who have no other kind of provision available; it is 'residual' because it deals with those who are left over. The contrast between residual and institutional welfare reflects the transition from the Poor Law of the nineteenth century to the Welfare State which came after it. This has meant that the idea of residual welfare is overlaid with a 'public burden' model - the phrase is Titmuss's4 - and the idea that residual welfare is particularly stigmatised and divisive. This has been the general experience of residual systems. Mishra describes the residual model, by contrast with institutional and socialist models, in the following terms: Three models of welfare5 Residual Attitudes to: State intervention Need as basis of provision Range of services Population covered Level of benefits % of national income spent on welfare Means testing Clients Status of clients Minimal Marginal Limited Minority Low Low Primary Paupers Low Optimal Secondary Extensive Majority Medium Medium Secondary Citizens Medium Total Primary Comprehensive All High High Marginal Members of High Institutional Socialist
Wilensky H and Lebeaux C (1958) Industrial Society and Social Welfare, Free Press, New York, 1965. 4 R M Titmuss, 1974, Social Policy: an introduction, London: Allen and Unwin. 5 Mishra R., 1981, Society and social policy, 2nd ed., Macmillan, London, pp.101-134) 4
Orientation Role of non-state services

Coercive Primary

Utilitarian Secondary

Solidaristic Marginal

The picture of residual welfare is clearly recognizable, though the assertion that need is only a marginal criterion is very questionable; a focus on need is central to the selection of the residuum to be served. It should also be said that elements of the ideal type are not required by others. There is no evident reason why residual welfare should treat clients as paupers, or why it should be coercive. Conceptually, it is possible to represent residual welfare in positive terms. Residual welfare is based in the argument that people should not fall below a basic minimum standard. This identifies residual welfare with the idea of a safety net, or a guarantee of protection, and such guarantees are usually applied generally and comprehensively. If people have a general right to welfare, there must either be comprehensive service, or a guarantee of service when necessary. If there is a guarantee, but there is not general coverage, there is a residual system. Education and health services are increasingly like to be comprehensively provided, but there are many other spheres of activity - like housing and social work - in which the residual model is alive and well, and necessary for people's welfare. In the terms used in continental Europe, the provision of residual services could be seen as a means of extending solidarity to those who are excluded. There is a distinction to be drawn between a residual system, in which the role of social services is confined to a limited sphere of activity, and a residual benefit, which has the role of supplementing or complementing other forms of provision. Even if there are objections to residual welfare as a general principle, it is difficult to see how these can be extended to specific residual benefits. The French RMI, a means-tested form of social assistance designed to offer support to the excluded, is certainly a 'residual' benefit, intended to mop up those not covered by the patchwork quilt of mutualistic and state benefits. But the RMI is framed in the following terms: Every person who, by reason of his age, his physical or mental condition, or of the economic or employment situation, finds himself unable to work, has the right to obtain from the collectivity appropriate means of existence. The social and professional insertion of people in difficulty is a national priority. To this end, a minimum income for insertion (RMI) has been instituted, implemented in the circumstances specified by the present law. This RMI constitutes one of the elements of a general provision in the struggle against poverty, directed toward the suppression of every kind of exclusion, in particular in the fields of education, employment, training, health and housing.6 This seems to me to be an example of a residual benefit within an institutional framework; it is not the same thing as a residual system of welfare. Other interpretations are possible, of course. One may be to cast doubt on the validity of the RMI as a model; if it is residual, it must have the smell of sulphur and the sign of the cloven foot. The other possibility is to

Commission Nationale d'Evaluation du Revenu Minimum d'Insertion, 1992, RMI: Le pari de l'insertion, Paris: La documentation Franaise, vol 2 p 789. 5
accept that it is a residual system, but that residualism can be founded in concern for the situation of the most disadvantaged, and the intention to offer a comprehensive system of social protection. If that is right, there is nothing intrinsically wrong with residualism. Selectivity 'Selectivity' is used ambiguously; it has been used to refer specifically to selection on the basis of low income (and so, to means-tested benefits), but also to refer to selection on the basis of need.7 These propositions can be seen as equivalent if needs can be met through financial provision; a person who can employ a housekeeper does not have the same need for domestic help as someone who cannot, even if their functional capacities are the same. But where needs cannot simply be described in financial terms - which may be the case in benefits for old people and physically disabled people - selective benefits might not be means-tested. The arguments against selectivity have been well-rehearsed. The first problem is that the process of identifying and testing entitlement is likely to be intrusive, complicated and stigmatising. Second, there are boundary problems, in determining who should be selected, and who should be excluded; the best-known example is the 'poverty trap', which arises because of the withdrawal of benefits from people whose incomes rise. Third, there is the problem of low utilisation, because of the barriers which selectivity creates for access. Fourth, selectivity can produce perverse incentives; means-testing penalises people who have made private or independent provision, and tests of physical needs creates disincentives to rehabilitation. Fifth, selectivity is administratively costly. There are also inefficiencies; the 'paradox of targeting', described by Keen, is that the greater the need, the more expensive it becomes to deal with those needs.8 This means that lower unit costs can be achieved by excluding those in greatest need. Selectivity is widely associated with the residual model of welfare, because concentration on people in need seems to require some mechanism by which people in need can be identified. The association with means-testing is particularly strong, because the effect of means testing is to confine benefits to those who are poorest. However, it is true neither that residual welfare must be selective, nor that selective welfare must be residual. Residualism can operate without any selective process. Residual welfare deals with a residuum, and the residuum may reflect not a conscious process of selection and rejection of people who are wealthier, but the relative opportunities of the poor and the rich. The residualisation of public housing in the UK has happened, not because public housing was confined to those in need, but because those who can afford to buy have superior options.9 Soup kitchens are not in general selective - there is no need to impose a test of entry - but they are residual.

Contrast, e.g., Reddin M., 1970, 'Universality versus selectivity', in W. A. Robson, B. Crick (eds.), The future of the social services, Penguin, Harmondsworth; A Seldon, H Gray, 1967, Universal or selective social benefits, London: Institute for Economic Affairs. 8 M Keen, 1991, Needs and targeting, London: Institute for Fiscal Studies. 9 R Forrest, A Murie, 1983, Residualisation and council housing, Journal of Social Policy, vol. 12. 6
Conversely, selective benefits are not necessarily residual. The means-tested pensions in Australia deal, not with a residuum, but the bulk of the elderly population. Medical care is generally selective, in the sense that the care offered is strictly geared to a professional assessment of need, but the principle of a universal health service has for many years been seen as the cornerstone of institutional welfare in the UK. There are always exceptional circumstances, and Titmuss suggested that any welfare system would require a combination of universal and selective social services. The real challenge resides in the question: what particular infrastructure of universalist services is needed in order to provide a framework of values and opportunity bases within and around which can be developed acceptable selective services provided, as social rights, on criteria of the needs of specific categories, groups and territorial areas. ?10 This is a mix, not of principles, but of methods. Titmuss realised early on that there was no necessary link between the method and the model it might represent, and his opposition to selectivity was confined to the situation in which it might be seen as stigmatising and divisive. To show that selectivity and residualism are discrete, I had to establish that - there is at least one case in which a residual measure is not selective, and - there is at least one case in which a selective measure is not residual. Both conditions are satisfied. Targeting Targeting means only that policies have to be directed at someone or something. The occasional use of 'targeting' as a euphemism for means-tested, residual services has disguised the importance of the concept for social policy. Social policies have to affect someone, and any attempt to identify a client group specifically can be referred to as 'targeting'. In Social Policy, I considered the implications of targeting a range of different groups: individuals, households, families, communities, and blocs.11 There is no intrinsic reason why the target should be the needs of particular individuals, let alone poor individuals; it may be possible, for example, to attribute needs to broader categories of people (like single parents, or residents of particular neighbourhoods)12. In the targeting of services, four main problems reduce efficiency. The first is deadweight; people receive the service or benefit, but their circumstances are not materially affected by the measure. Second, there are spillovers; people are helped who it was not intended or necessary to help. Third, there are inefficiencies arising through administrative costs; there may be a trade-off between gains through greater precision, and the costs of selection. Fourth, there is failure to reach those at whom the policy was targeted. This is usually described as low takeup, but it is not quite the same thing: as Atkinson points out, it does

Titmuss R.M., 1968, Commitment to welfare, George Allen and Unwin, London, p.122. P Spicker, 1995, Social policy: themes and approaches, Hemel Hempstead: Prentice Hall. 12 N Gilbert, H Specht, P Terrell, 1993, Dimensions of social welfare policy, (3rd ed) Englewood Cliffs NJ: Prentice-Hall, pp 84-85. 7
not follow, because people are entitled, that they were within the target group, and their failure to claim is incidental.13) A concern with effectiveness, rather than efficiency, focuses attention mainly on the last of these issues; deadweight and spillovers can be tolerated, but low takeup tends to imply that a measure will fail to meet its goals. The problem of containing administrative costs also argues for tolerance of these externalities. In the context of developing countries, the World Bank has argued for broad-based 'indicator targeting', picking on regions, age groups, gender or other kinds of common characteristic.14
If selectivity focuses on people in need, targeting does not always do so; it makes perfectly good sense to talk about measures which target women, old people or children. These are not necessarily 'selective' in the usual sense of the term, because there is no requirement for the groups to be in need. Targeting poor women, frail elderly people or excluded ethnic minorities would be selective, because the groups are in need. I do not think, though, that a measure can specifically target people in need without being selective, because that is what selectivity means. If, for example, targeting means something vaguer than selection, like 'aiming to reach' a group, it might be done in an indeterminate way. Setting up an advice agency in a poor area, for example, might be represented as a means of 'targeting' those in need without selection. But the difference here seems to imply that selectivity is individualistic, and targeting is not; there is no reason to accept that interpretation. If there is no distinction between targeting those in need and selectivity, there is no case in which selectivity is not targeted. To show that selectivity and targeting are discrete, I would have had to establish that - there is at least one case in which targeting is not selective, and - there is at least one case in which a selective measure is not targeted. The first condition is met, but not the second. This means that selectivity is a sub-set of targeting.
A B Atkinson, 1989, The takeup of social security benefits, in Poverty and social security, Brighton: Wheatsheaf. 14 World Bank, 1990, World Development Report, Oxford: Oxford University Press. 8
The relationship between residualism, selectivity, and targeting can be represented graphically as follows:
There are two main alternatives to targeting. These are to distribute benefits to everyone, and to distribute benefits indiscriminately. Distribution to everyone is exceptional - most 'universal' benefits are, in fact, categorical, and aimed at a broad class of recipient (like children or old people). The best truly universal example I can think of in the UK was the distribution of water supplies to all homes between 1948 and 1974 - coupled with the 9

sanction that homes without water supplies were condemned as unfit for habitation, and demolished. Non-targeted services are often used by the voluntary sector, and by local government, sometimes experimentally (to see what will happen) and sometimes in the belief that they are generally a good thing. I am not referring to allocation by lot (though there are examples of its use in welfare provision15), but rather to provision which is offered to all comers. Adult education, for example, is not usually targeted; education is a good thing in itself, and it is made available to whoever happens to want it. Marriage guidance counselling is often self-selected; if people want it, they can have it. (Some services which at first may appear not be targeted in fact are intended to benefit the whole community, and it is the community which is the target group. Community education, for example, works on the principle that the introduction of skills of negotiation and political development will benefit a community as a whole, whoever receives them. Parks, reclamation of waste areas or conservation work are intended to improve the ambiance of a community overall, without concern about the specific beneficiaries.) The alternatives have limited application to social security, largely because measures which are based in provision or aimed at communities tend to have different names, but in any case there is little scope for indiscriminate financial provision within an area, If policies are going to produce intended effects, they have to be directed towards a target population, and for practical purposes targeting is pretty much unavoidable. The problems of targeting are, however, similar to the problems of selectivity. There are problems of identifying the target group, boundary problems, and potentially problems of creating barriers to access. This is not specifically an objection to targeting on the basis of need; these issues apply whether or not targeting is done on the basis of need; targeting groups, areas, communities faces the same kinds of problem. For target groups which are not identified in terms of need, there are also problems in equity; the difficulty of targeting women is that poor men may receive less than rich women, while targeting old people may mean that a minority ethnic group is disadvantaged further relative to the majority. Although there may be problems, it is difficult to think of good arguments against ever directing resources to particular people or groups; the issues which have to be considered are how to overcome the problems, not whether targeting should happen in the first place. Incentives Incentives and disincentives are influences on behaviour. Social security policy produces a range of potential incentives and disincentives: there are rewards for behaving in particular ways, and punishments for behaving in others. Although the concept of incentives is widely used in the discussion of social security, and can also be found in discussions of taxation and management, it is surprisingly rarely found in discussions of labour economics. Atkinson describes the discussion of work incentives as 'partial partial' analysis, looking at only one side of the labour market, which is itself

J Elster, 1992, Local justice, Cambridge: Cambridge University Press. 10
intimately connected with many other factors.16 But this hardly reflects its influence in politics, where it is assumed that "both common sense and economic theory suggested that people might work less if they were given money."17 What economic theory is this? It is difficult not to see this as a case of 'do-it-yourself' economics, assuming that political nostrums have, in some way, the backing of social science. The political arguments on incentives are one of the best illustrations of a principle expounded by Keynes: that practical men are often the slaves of some defunct economist. Charles Murray writes: 1. People respond to incentives and disincentives. Sticks and carrots work. 2. People are not inherently hard working or moral. In the absence of countervailing influences, people will avoid work and be amoral. 3. People must be held responsible for their actions. 18 These arguments are drawn directly from the Benthamites, who argued that poor relief had to be made unpleasant, because otherwise there was no reason why people should not claim. Bentham argued that man was moved by pursuit of pleasure, or fear of pain; things which people should avoid had to be made painful.19 Discussions of incentives are generally based in a crudely utilitarian theory of human motivation. Murray believes that 'sticks and carrots work'. If there is a lesson to draw from economic analysis, it is that sticks and carrots do not work, or at least that they do not work straightforwardly. Most problems of this kind are multi-factorial, and the responsiveness to stimuli varies according to circumstances. This is generally expressed in terms of 'elasticity'. Costs and benefits have to be balanced against each other, and against other considerations. Even within its own constraints, the Benthamite position on welfare ignores a simple fact of life, which is that many of the conditions in which people claim benefit - like unemployment, sickness, disability or bereavement - are already unpleasant, and there are personal, financial and social costs involved. It does not immediately follow, because unemployment benefits increase, that more people will become unemployed, any more than it follows that an increase in death benefit leads to more people being dead. Economics is generally concerned, not with the reactions of individuals, but with aggregate populations. In economic theory, the behaviour of groups of people is taken to reflect such influences because individual variations in behaviour are cancelled out in aggregate behaviour. Incentives are determined by considering the effect of changing conditions on aggregate behaviour. For example, the reduction of tax on overtime work increases the rewards for such work. If the supply curve is positively elastic to changes in price - that is,

A Atkinson, G Mogensen, 1993, Welfare and work incentives, Oxford: Clarendon Press, p.22. 17 Institute for Research on Poverty, University of Wisconsin, 1973, The New Jersey Graduated Work Incentive Experiment, United States Department of Health, Education and Welfare, p.1. 18 C Murray, 1984. Losing ground, New York: Basic Books. 19 Bentham J., 'An introduction to the principles of morals and legislation' (1789) in M Warnock (ed.) Utilitarianism, Collins, Glasgow 1962. 11
if the quantity supplied is proportionately related to the price is - the reduction of tax might be considered an incentive to work. (There is some empirical evidence to suggest that the supply of labour is fairly inelastic in practice, because differences in the population cancel each other out.20) But an individual may perceive the change differently; for example, the extra rewards may reduce the need to work. Although it is possible to speak of incentives in individual terms, in the same way as it is possible to speak of supply or demand in individual terms, the existence of an incentive is not sufficient to determine an individual's choices. There is sometimes a tendency to refer to any inducement or penalty as an 'incentive', which presupposes that the inducement has a direct relationship to motivation. We read, for example, that because there is a financial gap, "incentives to work are much lower if families have children or couples have an unwaged spouse."21 This is not about incentives; it is about financial gains for low paid work, measured in terms of replacement ratios and tax rates. The relationship of these ratios to incentives has been assumed. The existence of an incentive depends on the context in which financial inducements are applied. Whether a social security benefit affects incentives to work depends, amongst other things, on the level of the benefit, the structure of wages, the labour market, and the propensity to work. As it turns out, the empirical evidence which exists about the impact of benefits on the labour market is equivocal. The strongest example appears to be that of women whose partners are not working.22 Atkinson and Mogenson, reviewing the evidence, do report some individual effects, but aggregate figures do not show any association between benefits and participation in the labour market.23 Because unemployment is primarily related to economic activity rather than social security benefits, there is no relationship apparent between the relative generosity of benefit systems and the levels of unemployment prevailing over time. The evidence of other incentive effects - such as the impact of means-testing on personal saving, or the impact of natalist policies - is no less unsatisfactory; there are certainly individuals whose conduct may be altered, but the influence is likely to be countered by other individuals who are influenced by other factors entirely. Eithne McLaughlin's work on the motivation of unemployed people shows that it is not true that people will only work if they have an increase in income; that the source of income matters as well as its amount; and that women's behaviour in any case reflects a

C V Brown, D A Dawson, 1969, Personal taxation, incentives and tax reform, London: Poltical and Economic Planning; P Johnson, 1994, Taxes and benefits, equality and efficiency, in A Glyn, D Miliband (eds) Paying for inequality, London: IPPR. 21 C Giles, P Johnson, J McCrae, J Taylor, 1996, Living with the state, London: Institute for Fiscal Studies, p.95. 22 A Dilnot, M Kell, 1989, Male unemployment and women's work, in A Dilnot, I Walker (eds) The economics of social security, Oxford: Oxford University Press; R Davies, P Elias, R Penn, 1994, The relationship between a husband's unemployment and his wife's participation in the labour force, in D Gallie, C Marsh, C Vogler (eds) Social change and the experience of unemployment, Oxford: Oxford University Press. 23 A Atkinson, G Mogensen, 1993, Welfare and work incentives, Oxford: Clarendon Press. 12
different pattern of constraints, requiring choices not between work and leisure, but between paid and unpaid work.24 The strongest arguments relating to individual 'incentives' are not usually about incentives at all, but about rewards. The concern that people should not be better off on the dole than they are working than workers is that it is not fair for workers to be worse off. The main concern about the poverty trap is not that people will not work because of high rates of withdrawal, but that it is not fair that they should be so little rewarded. The effect of means-testing on the propensity to save is unproven; as far as I know, no-one has actually shown that means-tested benefits discourage thrift. But they penalise conduct which governments wish to encourage, and that is morally inconsistent. Arguments about equity are clearly important, but the tendency to confuse arguments about morality with arguments about economic effects is unfortunate, and it only serves to muddy the water. Strategic intervention In Social Policy, I used the term 'welfare strategy' to refer to a group of inter-related policies, with a common approach or purpose.25 A strategic intervention is not a strategy, but a method used in pursuit of the strategic aim; I use the term to refer to intervention designed to produce specific types of effect. Nearly all forms of intervention have some effect on social relationships, but this does not mean that the effect is deliberate. Redistribution which is intended to be socially just, for example, is undertaken because it is right in itself, not because it will produce beneficial effects. Some social policies, like intervention in child care and criminal justice, are undertaken not because they have been proven to be effective, but because there is a moral obligation to try to do something. Not every intervention is intended to produce effects; we might distinguish teleological interventions, which are intended to produce effects, from deontological ones, which do not. Strategic interventions, in their very nature, are teleological; they are interventions from which further consequences are designed to follow. The promotion of welfare is, of course, a teleological objective, and strategic intervention to promote welfare is characteristic of social policy. The pattern of strategic intervention depends, obviously enough, on how the promotion of welfare is operationalised; the most common strategies include economic growth, social protection, redistribution, and the development of solidarity. Interventions may be conceived as a part of a general strategy, but it is possible that the method may be the only method employed. This may be because the intervention is believed to address the issues comprehensively. It seems naive to rely on a single method of intervention for a comprehensive strategy - it is more typical for governments in developed countries to introduce a package of measures - though there are examples, of which the Beveridge plan is the most striking. More commonly, where a particular method is selected, it is believed to be a key to other issues. In economics, this is widely practised, through the use of single

E McLaughlin, 1994, Employment, unemployment and social security, in A Glyn, D Miliband, Paying for inequality, London: Institute for Public Policy Research. 25 Spicker, 1995. 13
instruments like the interest rate or the exchange rate. The approach has been influential in social work through systems theory, where change can be brought about by key intervention within a specific sub-system.26 Strategic intervention is multi-dimensional. The values which inform policy constitute one dimension; the focus of policy (including targeting and universality) is another dimension; the means by which policy is pursued (such as regulation, provision and subsidy) is another. Policy formation, finance, service delivery and the role of users all have claims to be considered as dimensions in their own right. The range of permutations is large, and the outcomes are complex. Interventions, which can themselves be very diverse, are not certain to be internally consistent, let alone consistent with each other. In this light, it is difficult to identify any clear, simple role for targeting, selectivity or incentives. The most useful conceptual tool available for distinguishing the elements in this kind of problem is systems theory. Describing complex issues as 'systems' begins from the assumption that different elements being considered are inter-related, and that they stand part of a whole. This does not mean that they work in harmony - there may be contradictions and tensions between the elements - but that the elements have a structural relationship to each other. It is not necessary, for interventions to be strategic, to show that the same methods or purpose are applied repeatedly. What matters, rather, is complementarity - the extent to which the methods can be understood as contributing to strategic objectives. The relationship between targeting and selectivity on one hand, and incentives on the other, is contingent; they do not overlap directly, though a measure may involve both simultaneously. If these approaches have any systematic or structural relationship to each other, it is because they are part of a strategy - and so, that they are in some sense complementary. Historically, the approach they are associated with has often been individualistic, market-oriented and punitive. From the preceding analysis, however, it should be clear that there is no reason to suppose that these associations are necessary. A residual benefit can complement a universalist strategy; a universal service (like maternity visiting) can be used to identify residual problems. The relationship between methods and approach is not self-evident. Targeting is generally necessary for strategic intervention, because any measure which is intended to be effective has to be aimed at someone. (It does not, of course, have to be aimed at the intended recipient. The disadvantage of women can be addressed by targeting men, as has happened recently in the extension of paternity leave in the European Union. Equally, targeting does not necessarily call for targeting of benefits; it may imply targeting of costs or disincentives.) From a strategic perspective, the principal issue is that of effectiveness - not efficiency, which generally implies some compromise of goals to cost, but of maximum achievement of goals. Because effectiveness is most directly compromised through failure to reach the target group, this tends to argue for a more broadly based (and so less efficient) approach than precise targeting would imply.

B Compton, B Galaway, 1973, Social work processes, Homewood Ill: Dorsey; M Payne, 1991, Modern social work theory, London: Macmillan. 14
Selectivity may have a role in strategic intervention, and it may not. The importance of selectivity in an institutional system is primarily as a means of extending the scope and coverage of services, particularly in circumstances where a degree of flexibility is required. It may have an important residual function, in providing protection where no other form of protection is available. It is also, for what it is worth, directly redistributive. Incentives (and disincentives) are also not necessary for strategic intervention, but they have a special place in policy formation, for two reasons. The first is that they are the principal measure which is available to governments to bring about intended effects without the use of coercion. There are areas in which governments are reluctant to turn to coercion possibly for moral reasons (as in the case of controlling family size), but more usually because coercion does not always work (as in the case of prohibition). Social security may be used as a means of rewarding socially valued behaviour (like raising children: natalism, or encouraging births, is an important basis for benefit provision in France) or deterring undesirable behaviour (like failing to support one's family in the UK or US). The alternatives - propaganda, exhortation, and education - are limited in their effectiveness. The second is that policy measures may have unintended incentive effects; social security policies are accused of discouraging people from taking up employment, weakening families, and encouraging dependency.27 There is little empirical evidence to demonstrate such effects to any major extent, but there is a literature which is concerned with the marginal effect of altering aspects of policies - changing entitlement dates, for example, may change the period at which job-searching begins.28 Once it is accepted that a key intervention can have an impact on behaviour outside the immediate purpose for which the intervention is designed, it becomes necessary, for any teleological intervention, to examine that impact. (The same is not true of deontological action; if something is right, it is perfectly possible to argue that it should be done even if it has negative consequences. Ruat coelum, fiat justitia: 'let right be done, though the heavens fall'.) Because many impacts are measurable only as a marginal, aggregate effect, they are understood in terms of incentives. Incentives, selectivity and targeting can, of course, be combined, because the incentive can be targeted on a particular group - like manufacturing industry, or part-time employees - or on a particular group in need. But incentives do not have to be targeted - the supply and demand curves they influence commonly reflect the preferences of broadly defined populations - and they certainly do not have to be targeted on individuals. On the contrary, if the economic theory incentives are linked with is meaningful, there is some reason to suggest that they cannot sensibly be targeted on individuals. Incentives work because they affect aggregate behaviour, not individual behaviour. That means that the target group needs to be considered in aggregate. Untying the knot

H Parker, 1995, Taxes, benefits and family life: the seven deadly traps, London: Institute of Economic Affairs. 28 A B Atkinson, J S Flemming, "Unemployment, social security and incentives", in Midland Bank Review, Aug 1978. 15
The arguments on targeting and incentives have been colonised by particular partisan approaches, giving the political right a claim to a monopoly of economic insights and closing the minds of the left to strategic alternatives. But incentives and targeting can be used for many different kinds of strategic intervention, and strategic interventions can be done with many purposes. In their nature, they tend not to be liberal, because liberalism argues for the minimisation of intervention by government, and the implementation of a strategy requires purposive intervention. (The position of the New Right, which has argued for strong state action to bring about a liberal rgime, is paradoxical if not downright inconsistent.) It follows that the kind of ideal rgime with which these ideas have most often been popularly associated is also a rgime in which the scope for their application is very limited. Although links between residual welfare, selectivity and targeting can be found, they are not overwhelmingly strong. Residual welfare can be achieved without selectivity or targeting; targeting can be done without selectivity. Selectivity is only a particular form of targeting. There should, then, be no problem about arguing for each of these issues on their own merit. Arguing for residual welfare does not commit us to harsh or stigmatising policies; arguing for selectivity does not commit us to residual welfare; arguing for targeting does not commit us to selectivity. There are, of course, arguments to say that certain methods tend to lead to certain results. But it is misguided and naive to suppose that because, for example, groups dealt with selectively are stigmatised, they will not be stigmatised when they are not dealt with selectively29; that because residual welfare is often accompanied by complaints about the burden of welfare expenditure, such complaints disappear when welfare is not residual30; or that because arguments about targeting and incentives are used by the political right to oppose benefits, there are not arguments that can be made about targeting and incentives from the left in order to support them. If we want to protect the vulnerable, to redistribute income or to eliminate forms of social disadvantage, we may want to target particular groups. If we want to promote collective action or solidaristic arrangements, or to deter adverse selection, it is entirely appropriate to use incentive or disincentive arrangements to achieve these aims. These are methods - tools, if we will - and what we do with them is up to us.

II. Current state of the decentralization process
A. History of the decentralization process in Argentina
Despite its federal constitution, Argentina is a highly centralised country because of its centralising political culture inherited from colonial times and the concentration of human, financial and natural resources in the province of Buenos Aires. The 1853 Constitution included no national municipal legislative system and made no reference to municipal autonomy. It gave each province the authority to approve its own constitution and lay down rules for the municipalities within its jurisdiction. Municipalities were thus considered to be part of each provincial government system and for more than a century were perceived and mere administrative dependencies of the provinces. During the 1950s, the new constitutions of Chubut, Ro Negro, Neuqun, Misiones, Formosa, and Santiago del Estero adopted a wider definition of municipal autonomy.
25) Subsecretaria de Gestin Municipal. 26) Instituto Federal de Asuntos Municipales. 27) In Argentina, tax revenue redistribution schemes are known as regmenes de coparticipacin. 28) Jefatura De Gabinete De Ministros.
During the military governments of 1966-73 and 1976-83, local elections were abolished and municipalities reduced to mere administrative units of the central government. The return to democracy saw a movement in favour of greater municipal autonomy, reflected in the reform of the provincial constitutions in Crdoba, La Rioja and Salta. Since the 1994 reform, municipal autonomy has been mentioned in the federal constitution, although its scope and content are left to the discretion of each province. Local governments degree of autonomy varies, consequently, from one province to another. The reform also introduced a major change for the provinces, which received ownership over their own mineral resources. Taxes on oil, gas and minerals have greatly bolstered provincial budgets. This transfer strengthened some outlying regional centres, with an effect on the national distribution of population. For example, the oil-producing province of Neuqun has attracted greater internal migration than Buenos Aires.

Bolivia features a number of indigenous models of direct democracy. In rural areas these are ayllus, farmers unions and communities, and in urban areas, neighbourhood groups.120 These community structures were long excluded from the local government system in various ways. Since the jurisdiction of municipalities was restricted to urban areas, rural areas went unrepresented. Nor were indigenous community organisation legally recognised or allowed to stand for municipal election, which were, and still are, reserved for political parties.
112) The Enlared website listed 11 intercommunalities at the end of 2005. 113) Federacin de Asociaciones Municipales de Bolivia. The Bolivian Association of Women Councillors of Municipalities (ACOBOL) also belongs to FAM. 114) Agencia Suiza para el Desarrollo y la Cooperacin. 115) United States Agency for International Development. 116) Enlared municipal: http://www.enlared.org.bo/2005/agencia/cgdefault.asp 117) Programa de Apoyo a la Democracia Municipal. 118) This project is funded by the Federation of Canadian Municipalities (FCM) and implemented by FAM in the Lake Titicaca intercommunality and in Santa Cruz department by the departmental association of municipalities. 119) This project is carried out with support from the Canadian Atlantic CED Institute (ACEDI) and International Development Research Centre (IDRC). 120) Juntas de vecinos.
However, citizens organised at departmental level in the 1970s and 1980s to form pressure groups called civic committees121 or committees to defend regional interests.122 The latter played an important part in the regional development cooperation that supplied most public services. This movement led in 1992 to a decentralization bill. As part of its decentralization policy, the central government began to reorganise and institutionalise the channels for the participation of civil society, mainly through the 1994 popular participation law. These organisations were then legally recognised with a territorial basis and defined as basic territorial organisations (OTBs). OTBs may be farmer communities, indigenous areas within rural municipalities, or neighbourhood groups in urban municipalities. This definition therefore excludes professional organisations such as producer organisations, chambers of trade or professional associations. Its effect is to divide social movements, whose participation is achieved territorially. The law grants OTBs a number of rights with respect to municipal authorities. They may propose, monitor and supervise the execution of construction work and the supply of public services, take part in the management of the environment and sustainable development, approve education and health authorities and access information on the resources available for popular participation. OTBs do not exercise these rights directly, but via watch committees. These committees are responsible for monitoring the popular coparticipation funds transferred to municipalities, at both scheduling and execution and assessment phases. A committee may freeze the funds if it is not satisfied with the use the municipality is making of them. This system was enhanced in 1994 by the creation of a social control fund, an independent public unit of the Ministry of Human Development designed to support the watch committees. Since this fund did not manage to receive any funding, it was replaced by the Ministry of Popular Participation.

D. Results of the decentralization process in Bolivia
The centralising colonial legacy in Bolivia has been strengthened by the countrys experience of territorial dismemberment and military governments. The power-sharing process aimed at first, in the late 1970s, at delegating power to the intermediate level, namely the departments, via the departmental development corporations (CDDs). These corporations, supported by their civic committees, played a key role in supplying public services, including at local level, until the mid-1990s. They imposed a de facto decentralization process on the central government and fought to have it recognised institutionally. In the mid-1990s, the central government shifted the process towards participatory municipalisation. The popular participation law passed in 1994 dissolved the CDDs and transferred their resources to municipalities. It recognised traditional community structures, defined as basic territorial organisations (OTBs), and involved them in municipal management, mainly via the watch committees. The administrative decentralization law passed in 1995, concerning departments, did not strengthen their autonomy. It did grant them an assembly, but it is elected indirectly. Lacking representativeness, it had difficulty in standing up to an executive appointed by the President of the Republic. The election of prefects, adopted in 2005, may help change this.
121) Comits civicos. 122) Comits de defensa de los intereses regionales
Ten years after these laws were passed, their initial results may be reviewed. Civil society has indeed taken over the areas of participation opened up at municipal level. However, this participation does not appear to have raised the efficiency of public action and has helped strengthen corruption and patronage. Those, at least, are the conclusions of the director of the La Paz multidisciplinary studies centre (CEBEM)123. He considers that the OTBs have focused on obtaining immediate advantages, neglecting the construction of a strategic vision of development. Financing, he claims, has been allocated by discretionary methods and the OTB leaders have been co-opted by politicians. Nor are the OTBs representative of civil society as a whole, since their strict definition excludes professional organisations such as producer organisations, chambers of trade or professional associations, who have not been involved in the participatory process. The popular participation law has contributed to dividing social movements. Departments, which manage nearly half of national public investment124 and such important sectors as education and health, have also been frequently criticised. They do not appear to have met social demand, either for health, education and basic sanitation or for the protection of the environment, employment or poverty reduction. The lack of transparency in their management has also been pointed out, and their poor qualifications and political management of staff. In 2003, the Bolivian political system entered a severe crisis with popular riots directed at public buildings. This crisis highlighted the economic, social and cultural divisions between the western and eastern regions and the inadequate channels for social demands, and led to the overthrow of President Gonzalo Snchez de Lozada. His successor, Carlos D. Mesa, promised a revision of the hydrocarbons law that grants international oil companies ownership of the gas extracted and is the bone of contention between the two regions.125 He also called a constituent assembly to redefine the Bolivian social compact. Decentralization is one of the central topics for this assembly, which will need to find a formula that is appropriate for the weakness of central government and the opposing interests of the western and eastern departments. The departments, led by Santa Cruz and Tarija, are preparing different proposals for departmental autonomy. The results of the elections on 4 December 2005 to elect a new President of the Republic, a legislative assembly and department prefects are likely to modify the architecture of Bolivian institutions, through a new constitution that is to be debated and a transfer of powers to the regions.

A. Human resources

1. General
In 1994, the number of municipal employees in Brazil was estimated to be two million, 26% of all public sector employment. This figure increased by 77% from 1985 to 1991 as municipal resources rapidly grew. Salaries as a proportion of municipal expenditure rose from 15% to 23% during that period. Note that the federal constitution restricts the salary proportion to a maximum of 65% of municipal resources. Unlike the rest of Latin America, local agents in municipalities in Brazil enjoy security of
137) Rolnik, Raquel: Descentralizacin y Federalismo en el Brasil in: on-line journal of the University of Alcal, Descentralizar el Estado, December 2003. 138) Applying to real estate sales and donations. 139) Consolidaao das Leis do Trabalho, CLT.
employment and full-time status. The 1967 constitution required municipal staff to be hired in an open manner by a statutory recruitment system. In practice, this requirement has been distorted by patronage practices, with the establishment of a special category of staff covered by the private labour code139. Employees hired via the CLT during the military regime enjoyed the same stability as statutory employees. With the return to democracy, municipalities inherited staff who were supernumerary, unqualified, unmotivated and undisciplined. The 1988 constitution attempted to solve these problems by requiring municipalities to choose a single employment system for all their staff, except for those employed in a confidential capacity, imposing a system of hiring by merit and providing for career plans. The Constitution allows each municipality to decide its own statute for local civil servants, define the organisation of its human resources and establish pay scales and working conditions. These measures must comply with a number of principles established for the public sector as a whole: hiring by merit, except for people employed in a confidential capacity, job protection after two years service, pensions indexed on inflation after 35 years service for men, 30 years for women, trade union rights, including the right to strike, dismissal procedures, with the right of appeal. The freedom of each municipality to decide its own statute for local civil servants reduces agents ability to move sideways between municipalities and hampers the development of a municipal career system at national level. Despite these constitutional measures, because of the high degree of staff stability, attempts to introduce a merit-based human resources system, begun in the 1940s, have often failed in the face of patronage140. In 1991, promotion by merit was still rare. As in the past, there is no guarantee that legal measures will suffice to eliminate a deep-seated tradition. Posts for people employed in a confidential capacity are a third form of municipal employment that is not subject to the constitutional provisions governing permanent employees. These posts are many and various, and are distributed by the mayor within or outside the statutory system, although the 1988 constitution encourages internal promotion.

195) Fronterizo. 196) Since 1992. 197) Alcalde. 198) Concejo municipal. 199) Mayors have been elected by popular vote since 1988.
A. History and division of powers
Following its independence from Spain in 1810, Colombia first adopted a federal system, and then in 1886, in reaction to the unlimited autonomy that system gave municipalities, a unitary constitution. For a century, the country had extreme political centralisation. The first municipal legislation of any importance was not passed until 1913. However, in the 1960s, a process of demunicipalisation began with the introduction of a system of shared powers reflecting the extensive involvement of central government in the supply of public services at the local level. Responsibility for education and health was transferred to central government, land registry to a central government agency, urban transport and social housing to autonomous administrative agencies. Although some large municipalities set up semi-autonomous authorities to manage the electricity and water supply, by 1981, most municipalities only managed the following services: road maintenance, abattoirs, markets and solid waste disposal. Between 1970 and 1986, the legitimacy of the political elites was in crisis. The reasons were inequalities in the government distribution of resources,200 the inadequacy of public services, uncontrolled urban development, the lack of any channels for participation, and the presence of a patronage system. This crisis of government showed itself in lower poll turnout, a culture of political violence, and a wide movement of popular contestation, which took the form during that period of more than 200 civic strikes201 called by local groups.
Background to decentralization 1968-82
Type Political Budgetary Advances made Political crisis (1977-82): increasing numbers of civic strikes demanding efficient public services - Cration of a redistribution of sales tax - 13 % of ordinary revenues goes on health and education in municipalities and departments et de lducation dans les municipalits et dpartements - 1968 government reform: delegation of functions

Administrative

Source: Evaluacin de la Descentralizacin Municipal en Colombia, Departamento Nacional de Planeacin, 6-7/05/2002.
These movements gained the support of local governments and modernising elements in the conservative and liberal parties. The politicians launched a reform of Colombian administration. Decentralization was, therefore, a response to the crisis of legitimacy.
200) In 1979, the three largest municipalities accounted for 72% of local public expenditure, and their per capita expenditure was six times that in small municipalities. 201) Paros cvicos.

Jaime Castro, a former government minister, proposed five key points, which were incorporated in the new municipal code, Cdigo de rgimen Municipal No1333 of 25 April 1986: - direct election of mayors - local referenda - increased budgets for municipalities, begun in 1983 - administrative decentralization - citizens participation in municipal affairs From 1986, therefore, as part of the municipal reform programme, municipalities were granted major areas of responsibility: water supply and sanitation, environmental health, construction and maintenance of schools, clinics and roads, social housing, agricultural expansion, urban transport and land registry. The municipal departments and autonomous central government agencies, such as the municipal development institute (Instituto de Fomento Municipal), which concerned themselves with school buildings, water supply and environmental health, were closed. Responsibilities for education and basic health services were gradually transferred to provinces and municipalities, together with financial transfers earmarked either for salaries or investment. However, the transfer of powers was not followed by the allocation of the technical and human resources required by the new tasks and new technical and financial responsibilities. For example, the hasty closure of the national water agency INSFOPAL caused problems in supply.
Background to the municipalisation 1983-90
Domaine Political Budgetary Avances Constitutional reform: directly elected mayors - change in system for redistributing sales tax share: reform of transfer system (Law 12 of 1986) - reform of local taxes (Law 14 of 1983) - creation of a joint financing model - distribution of responsibilities (Law 12 of 1986) - distribution of responsibilities for health and education (Law 10 of 1990 and Law 29 of 1989)
The public campaign by students forced the adoption of a new constitution in 1991. The purpose was first to restore the governments legitimacy by opening up politically (incorporation of former guerrilla groups in the system) and introducing participatory democracy (decentralization of political power). The new constitution defined Colombia as a decentralized nation and a number of articles granted greater political and financial autonomy to local governments. Municipal terms of office were extended, and the number of councillors was required to be odd. The discretionary system used for central government transfers was replaced by a system based on constitutional guarantees, and the amount of this financing was increased. The Constitution also introduced the election of governors, and created provinces, regions and indigenous territorial entities. It enshrined citizens participation in public management.

205) One-off taxes to finance specific infrastructure operations.
In 1992, there were 75,738 municipal employees in Colombia, 8.8% of all public employees. Since 1938, many attempts have been made to introduce a merit-based career system at both central and local levels, with no real success. Following the period of political violence, known as la Violencia (1948-53), the Frente nacional agreement of 1958 provided not only that the two main political parties would alternate in government every four years for sixteen years, i.e. until 1974, but also that they would share out public posts at all levels. This parity system had the effect of institutionalising political patronage throughout the country down to municipal level. Although each municipal council could establish its own scales of pay and responsibilities, the parity system in practice prevented the development of a career system. After 1938, Colombia set up an administrative career structure. The 1991 constitution provided for a career civil service throughout the public sector, including municipalities, except for posts for people employed in a confidential capacity, and this was implemented by legislation in 1992. Colombia has an excessive number of special regimes, a variety of standards and a disciplinary regime so hard to implement that it is ineffective for applying sanctions.206 But gradually the civil service is becoming structured, particularly with more formal access mechanisms: all posts must be filled by competitive examinations held by the national civil service commission (and for which public and accredited private universities may prepare candidates). Article 2 of Law 909 specifies the following criteria: merit, in-service training for civil servants, and civil servants' responsibility for the work done. In order to increase professionalisation and transparency in the civil service, responsibilities were defined at national level. This has not been done at local and regional levels. With respect to training, an important role is played by the Higher School of Public Administration (Escuela Superior de Administracin PblicaESAP; Law 443 of 1998, Article 58). It was created in 1968 as the national-level public university-status establishment leading to employment in the administrative department of the civil service, which is part of the administrative and civil service career system. It is responsible for scientific and technical research and development, training, improvement and extension for the national and territorial public administration. However, faced with an increase in demand and despite the size of its territorial network, ESAP does not have enough staff to fulfil its role. For that reason, the legislation authorises the certification of higher education establishments to train regional and local administrative staff and hold competitive examinations for careers in the administration (Law 715, Article 36). At local level, each municipality is responsible for arranging training for its staff, from either ESAP or accredited higher education establishments. The choice of training institution must be open to tender.

B. Citizens participation
Until the 1980s, Colombia saw little citizens participation at local level. The many attempts to introduce legislation to enable the creation of district councils were rejected by the political elites, who feared the development of local democracy. The existing community action committees, juntas de accin comunal, were co-opted by the elites and used to maintain the traditional patronage system. Following the civic strikes of the 1970s, the national congress of civil movements and popular organisations, held in 1983, called for measures to enable genuine citizens participation in local government.
206) The latest standards are laid out in Law 909 of 2004 on public employment, administrative careers and public management.
The political component of the 1986 decentralization reform comprised three objectives, namely the consolidation of participatory democracy, the development of participatory, direct democracy and the improvement of governability. After 1986, a large number of instruments for citizens participation were produced, achieving formal advances in the creation of forums and channels for citizens participation. At present, Colombia is one of the countries best endowed with instruments for citizens participation. Some of them are intended to reinforce representative, such as the election of mayors and governors, local administrative assemblies (JALs),207 and the mayors platform vote (voto programatico), whereby a mayor who does not apply his or her electoral promises may be dismissed. Other instruments strengthen participatory direct democracy: - popular consultation, referendum, peoples law initiative; - creation of general participation forums such as planning councils; - creation of sectoral forums to link the community to the management of public affairs, such as the attendance of consumer groups at board meetings of companies supplying utilities such as electricity, water, telecoms or solid waste treatment,208 creation of development and social control committees for public services, participatory committees for health, social security, territorial organisation committees, etc. Various factors have reduced the effectiveness of these measures: - Some of these mechanisms, such as the JALs, are not mandatory for territorial entities. The creation of a participatory forum depends entirely on the willingness of mayors and municipal councils to set up bodies that in some cases are in competition for their own responsibilities; - Direct democracy instruments, such as referenda and consultations, depend for their implementation on the cooperation on the executive. In practice, left to the initiative of municipal councils, they have been little used; - The regulations applying to these participation mechanisms makes them difficult to use and discourages the organisations and citizens concerned, particularly in the case of consultations, dismissal of elected officials, or the participation of users in public service board meetings. Despite these limitations, some mechanisms, such as development plans, which are intended to define the use of public budgets, have managed to establish themselves and are now successful examples of participatory practice.

210) The regional division of the health sector, for example, differs from the one defined by law. 211) Federacion metropolitana de Municipios de San Jos. 212) The creation of a metropolitan structure, proposed by the national municipal congress back in 1976, had been blocked by the refusal of central government. 213) FEMETROMs main project here concerns the metropolitan areas rivers. 214) This involves producing common directives for urban development and protection of the environment and harmonising the application of standards among the municipalities. 215) Until 2002, mayors were elected by municipal councils and were appointed to be ejecutivo municipal or gerente. 216) Regidores are not salaried but receive attendance fees for council and committee meetings. 217) Five councillors for municipalities with less than one per cent of the national population, 7 for those from 1% to 2%, 9 for 2% to 4%, 11 for 4% to 8%, and 13 for municipalities with more than 8% of the national population. 218) Articles 14 and 33 of the Municipal Code.
With its 1998 Municipal Code, Costa Rica adopted a system of government based on a strict separation of powers at local level. This is a radical change, since the mayor was previously appointed by the municipal council. It also contrasts with the situation in other Central American countries, where the executive and deliberative functions work together. The municipal council elects a chair from among its number every two years. The mayor attends council meetings but may not vote. This pattern led in some municipalities to tensions between the two powers after the first elections of mayors, intendentes and district council members held in 2002. Municipal elections are held at the same time as presidential and legislative elections, and local and national terms of office are identical. This simultaneity tends to reduce the importance of local elections to the advantage of national party-political issues, a situation exacerbated by political parties monopoly over the choice of candidates. Many attempts have been made to move local elections to the mid-term of the national assembly, ever since the first national municipal conference in 1940. This movement, now represented by UNGL, has so far been unsuccessful. Districts are usually a mere administrative division of cantons. They have two representatives, a sindico and substitute, who sit on the canton municipal council but may not vote. In particular cases, districts may also have an assembly and extended powers. The Constitution219 authorises municipalities to create in districts far from the main town, district municipal councils, elected like municipal councils and enjoying operational autonomy.220 Executive authority is given to an intendente, elected in the same way as the mayor of the canton. At present, eight districts have this system.

2. Support services

To accompany the financial transfers and legal changes introduced in the first municipal code, passed in 1970, Costa Rica created a public institute to provide municipalities with training and technical and financial assistance; the municipal institute for development and advisory services (IFAM).245 IFAM is an autonomous entity with legal status and its own assets.246 It covers all municipalities, district councils and municipal federations, which belong on a voluntary basis, and works more generally for all municipal communities. Its mission of support to municipalities involves a large number of functions, whether financial, training, consultancy, research or promotion of municipal cooperation. In the financial field, IFAM provides municipalities with low-interest loans and acts as an intermediary when they wish to contract loans from national or international bodies. Since it was created, IFAM has financed more than 1,000 projects for a total value of over 12 billion (CRC). The projects involved modernisation of the administration,247 provision of services248 and infrastructure.249 The Institute also offers a permanent curriculum of training courses for elected officials and municipal employees, and those members of civil society who take part in local management. The main topics in the curriculum are the following: municipal management for mayors; participatory planning and negotiation for members of district councils; budget management; municipal legislation; and solid waste management. Last year, IFAM ran 25 training courses for municipal civil servants, 16% of the national total. IFAM provides technical assistance in fields relating to municipal finances,250 legal matters,251 computer systems, solid waste management and infrastructure.252 It carries out research and studies of administrative organisation and the operation of local public services in order to achieve continuous improvement. It promotes cooperation between municipalities,253 with other public bodies, as part of institutional coordination, with international bodies,254 the private sector and NGOs. IFAM collects, manages and controls two taxes, the spirits tax and the national beer tax.255 It transfers half the revenue from the former to municipalities, and the other half is its own main source of funding. Since 1998,257 it has transferred 25% of the revenues of both taxes to the Costa Rican institute of sport and leisure (ICODER).258 IFAM may also, at the request of a municipality, manage public works and services and act as a central equipment-purchasing agency. IFAMs main sources of funding are the following: a percentage of the alcohol tax,259 which it administers and distributes, fees and loan interest payments from municipalities, and interests on the investment of its cash flow.

Municipal autonomy is a constitutional principle,363 amply repeated in the law on municipalities.364 The principle of general responsibility, contained in the Constitution,365 is complemented in this law by a principle of subsidiarity, since any function that may be carried out efficiently under their jurisdiction, or which require a close relationship with their own community, must be reserved to the scope of responsibility of municipalities.366 They are required to assume these responsibilities,367 particularly for municipal services.368 Municipalities also have general responsibility for matters concerning the preservation of the environment and natural resources.369 The law on municipalities does not merely declare these principles, but details the areas of municipal responsibility.370 These may be categorised as follows: basic services, transport, public health, land use planning, management of the environment, and culture, sport and leisure. Although it is not explicitly mentioned, economic development is an integral part of municipal responsibilities. This does not appear to be the case for education, which is not mentioned anywhere. Municipal basic services involve the water and electricity supply, and the evacuation of wastewater. The municipality may build, maintain and manage municipal aqueducts, sewer networks and the wastewater treatment system, and the power network. Solid waste management, rainwater drainage, and market supervision are part of its functions for promoting public hygiene. To these are added responsibility for building, maintaining and managing public cemeteries, which may be operated under a concession. Municipalities are also required to develop transport and highways. They may build and maintain bridges and dirt roads, regulate and control the public transport service
358) Article 17. 359) Articles 18 and 19. 360) articles 20 and 21. 361) Article 22. 362) Article 40. 363) Constitution, Article 177. 364) Law 40 of 1988 on municipalities, reformed in August 1997. 365) Constitution, Article 177: Municipal authorities are competent for matters affecting the socio-economic development of their constituency. 366) Law 40 of 1988 on municipalities, reformed in August 1997, Article 2. Authors emphasis. 367) Law 40 of 1988 on municipalities, reformed in August 1997, Article 2. 368) Law 40 of 1988 on municipalities, reformed in August 1997, Article 6 369) Law 40 of 1988 on municipalities, reformed in August 1997, Article 6. 370) Law 40 of 1988 on municipalities, reformed in August 1997, Article 7.
and manage bus stations, river and lake harbours, if necessary with the help of the relevant central government body. They may design and plan urban and rural signage. In public health, municipalities must take part in building and maintaining urban and rural health units and centres and support hygiene and preventive health campaigns. In land use planning, they are responsible for planning, standardising and monitoring land use and urban and rural development. They define urban development plans and master plans, regulate and monitor the use of urban land, supervise the use of the subsurface, together with the relevant central bodies, monitor building standards, build and maintain streets, pavements, parks and squares. Their environmental responsibilities are extensive. They ensure the rational use of natural resources. To that end, central government is required to consult them before granting concessions to explore or exploit the natural resources on their territory, and must transfer to them at least 25% of the tax revenues from these concessions. Central government must also involve them in producing environmental impact studies. Municipalities are allowed to create municipal ecology parks. They belong to the natural disaster response system via municipal emergency committees linked to the national committee. In culture, sport and leisure, municipalities must primarily ensure that local heritage is protected, by building, maintaining and managing infrastructure.372 Municipalities are required to promote human rights, especially those of women and children, and to encourage citizens participation. Municipalities may delegate the provision of services to municipal enterprises they create,372 or other public sector institutions more qualified for the sector concerned, while monitoring execution. They may also grant concessions, after competitive bidding, to private persons, while continuing to exercise their regulatory and monitoring functions.373

383) These revenues may come from municipal taxes, duties and special contributions. Law 40 of 1988 on municipalities, reformed in August 1997, Article 49. 384) Law 40 of 1988 on municipalities, reformed in August 1997, Article 47. 385) Law 40 of 1988 on municipalities, reformed in August 1997, Articles 28 and 48. Until a municipal tax law is approved, municipal taxation is covered by Decree 10-91, Plan de Arbitrios del Municipio de Managua, published in the official bulletin La Gaceta on 12 February 1991 for the capital, and Decree 455 Plan de Arbitrios Municipal, published on 31 July 1989, for the other municipalities (Article 70).386) The data on the financing structure of all Nicaraguan municipalities are currently unavailable. We give those for Managua, with no claim that they are representative. See website of the Nicaraguan statistical institute: http://www.inec.gob.ni.387) Daprs Presupuesto General de ingresos y egresos, ano 2006, Municipalidad de Managua, 2005. 387) According to Presupuesto General de Ingresos y Egresos, ao 2006, Municipalidad de Managua, 2005. 388) Article 77 obliges central government to pay municipalities a sufficient percentage of its general budget, with the advantage given to those that have least resources. 389) Article 6 stipulates that the economic resources for the exercise of these responsibilities will come from own revenues and those transferred by the government, either as tax transfers or as financial resources. This principle is reasserted in Article 11. 390) This law was backed by the following bodies: AMUNIC, INIFOM, Nicaraguan network for democracy and local development, FISE and the Committee on Municipal Affairs. 391) In 2003, 46% of the budget went on servicing the debt. 392) Fondo de Inversin Social de Emergencia. 393) Instituto de Desarrollo Rural.
FISE finances small and medium scale projects intended to raise the living standards of the poorest communities, particularly by bringing them into the labour market. The IDR supports bodies, including municipalities, that commit themselves to projects that promote micro-enterprises. Municipalities are also able to contract short or medium term loans from public and private banks for public works or to raise the standard of public service provision.394 Long-term loans must be approved by the National Assembly and must concern large-scale public works. These loans must be repaid within the term of the elected officials who contracted them, except for long-term loans.395 Mayors and municipal councils may not leave debts to their successors. Nicaraguan municipalities autonomy of expenditure is relatively high. Municipal financial autonomy is a constitutional principle,396 reasserted in the Law on municipalities.397 Municipalities plan their budgets on their own and are only subject to monitoring by the public auditing body, Contralora General de la Repblica.398 They must reserve a certain percentage of their budget for investment, as laid down in the law on municipal budget regimes, according to the category of municipality.399 The annual municipal salary bill may not, at all events, exceed 30% of a municipalitys ordinary annual resources400.

415) Asociacin de Municipios de Nicaragua. 416) Federacin de Municipios del Istmo Centroamericano. 417) La DANIDA gre au Nicaragua un programme de renforcement de la gestion municipale et du dveloppement local (progestion), (APDEL). In Nicaragua, DANIDA runs a municipal management and development strengthening programme (Progestion), APDEL. 418) As part of its programme for local development and fiscal transparency (PRODELFIS). 419) This federation has already created a cooperation programme with Nicaragua: the Alianza Madriz-Madrid (AMM). 420) Diputacin de Barcelona. 421) InWEnt, Internationale Weiterbildung und Entwicklung (Capacity Building International, Germany), is funded by the German federal and Lnder governments. It offers courses for specialists and leaders in the fields of industry, politics, administration and civil society. 422) Municipal capacity-building focuses on management, planning, administration and project implementation, monitoring and evaluation. It requires lobbying for financial transfers, at present the adoption of a new municipal budget code and promotion of municipal association activities, particularly among municipalities 423) Law 40 of 1988 on municipalities, reformed in August 1997, Article 12. The creation of a mancomunidad requires the approval of the National Assembly. 424) Loi N40 de 1988 sur les municipalits, rforme en Aot 1997, Article 28. 425) Source: Progestion programme paper, Finnish Embassy in Nicaragua, December 2003.
($3.7m), European Union ($3.4m) and UNDP ($2.9m). The bilateral donors were the following:
Germany ($18.6m), Finland ($7.1m), Sweden ($3m), Canada ($5m) and the Netherlands ($0.8m).425
The first major steps to promote citizens participation were taken during the Sandinista period. In its first years in power, the FSLN established peoples municipal development councils. Some municipalities held peoples assemblies on the model of the cabildo abierto, which were often seen as counter-productive, since municipalities did not have the necessary autonomy and financial resources to meet the popular demand expressed at those assemblies. A number of municipalities called on voluntary work to compensate for their inadequate financial resources and the unreliability of government transfers. They also mediated between central agencies and the mass organisation that supported the FSLN. Citizens participation was finally formalised in 1988 by the law on municipalities. This originally required the mayor to hold two cabildos426 a year, and allowed the mayor to create consultative bodies, or municipal peoples councils.427 In 1991, a survey of 51 municipalities revealed that more than half of them had not held a cabildo and 60% had not held any meetings of a municipal peoples council. The reforms to the law increased municipal responsibility for participation, since municipalities were defined as guarantors of participatory democracy.428 To that end, they have a number of instruments. They may, for example, create elected territorial delegations429 and participatory associations and sectoral bodies.430 Citizens were given a right of initiative both collective and individual, enabling them to present draft ordinances and motions to the municipal council.431 They may also attend those sessions of the municipal council that are public.432 The municipal requirement to hold cabildos was extended.433 They are to be chaired by the mayor and municipal council, and held in ordinary session at least 12 times a year, at the initiative of either the municipal authorities or the citizens, to discuss the draft municipal budget and its implementation, and to be informed of the municipal development plan. This participatory design of the budget is limited, however, to a consultative role and is not binding on the municipal authorities. Cabildos may also be convened in extraordinary session by the municipal council or citizens. Municipalities in the autonomous regions that comprise indigenous communities in their territories are also required to recognise the formal traditional authorities that represent them, and to take them into account in municipal development plans and programmes, and in decisions that affect them.434 The law also opens up the possibility of referenda and plebiscites. Participation is one criterion for selecting the projects applying for finance from the emergency social investment fund (FISE). The citizens participation law, passed in 2003, mainly involves central administration, and hardly at all the local level. It offers citizens, for example, the right to initiate legislation and imposes on public authorities a duty of transparency.