38  Social Protection for the Poor and Poorest <
basis in the preceding tradition of natural law by insisting on 'the principle of utility' as the primary objective of public policy. The question (asked by Viner but not by Bentham) was why anyone should expect utility to be the primary objective of public policy. One possible answer is that people have j a right to utility, and that it is the duty of the state to uphold rights. If this / is the case, then the welfare losses, measured in utilitarian terms by the risk-based school of economists, also become rights violations.
Rights
The rights-based supporters of social protection think of poverty as a viola-J tion of human rights. Chronic poverty is then an ongoing violation of / rights. One can infer that, if poverty is a violation of rights, then there ' must be a right to get out of poverty. Indeed, in 1986, the UN General Assembly declared that human beings have 'a right to development', and this right has been reiterated in numerous UN declarations since. Poverty reduction is thus not a policy option that a state might take on or not, as it sees fit, but is a duty of states to their citizens, since rights are binding legal obligations.
When they are confronted with rights-based arguments, the economists -either from the risk-driven school or the basic netjjds school - view the i rights-based claims with great scepticism.14 They point out that no amount 1 of legal obligation can get around the tough questions of trade-offs in economic and social policy-making in a world characterised by material scarcity and infinite wants. More of one thing means less of another, and more of one thing today may mean less of everything tomorrow. The absolutist 1 language of the rights discourse - for example, 'there are no small rights' | (UNICEF 2000a: 4), 'all rights for all children' (UNIGEF 2005) - is uncon-L-vincing to those reared on the doctrine of scarcity, the production possibility curve and associated opportunity costs. When economists point these things out, the rights-based thinkers can only respond with the doctrine of the progressive realisation of rights, which has its own problems, as seen above.
The legalistic tradition in human rights thinking views the binding legal obligations implied by rights as its trump card. The legalistic rights-based thinkers thus tend to denigrate needs-based arguments as being paternalistic, based on charity and moralism - soft ground compared to legal obligations. 'Benevolent and charitable actions, while good in themselves, are insufficient from a human rights perspective' (UNICEF 2000b: 9-10; see also UNICEF 2004: 11-12.) The legalistic rights-based thinkers \tend to ignore the risk-based economic arguments of neo-classical eco-■nornics, or dismiss them as 'welfarist' and thus, again, as being weaker than arguments based on law.
vThe only ones who address the link between rights and needs systematically are those who think that rights must be based not in natural law or the
Lauchlan T. Munro 39
whims of nation states when they make international law, but on a theory \ of human needs. Like the risk-based economists, they also address the prob- ' lem of moral hazard in social protection. Economists and moralists have often stressed moral hazard as a potential problem in social protection. The I existence of a social safety net may distort patterns of work and levels of 1 savings and effort, and social protection schemes incorporate mechanisms intended to avoid such problems. Many of these mechanisms, however, have the effect of denying coverage to those who should be getting benefits from the social protection scheme (Munro 2002). Economists have explored the welfare losses involved in both failing to cover the whole target group ofI a social protection scheme and in providing benefits to too many (ineligi- ; ble) people (Cornia and Stewart 1993). Rights-based thinkers usually ignore i the moral hazard issue in discussions of the design of social protection measures. A few, however, have grappled with the issue, most notably Peter Taylor-Gooby (1991, ch. 8). Admitting that moral hazard is a problem, Taylor-Gooby insists that a rights-based framework can cope with moral hazard. Since rights come with correlative duties, one of the duties of the rights holder is not to abuse his/her rights. This implies not only the usual injunction not to invoke one's own rights to the detriment of the rights of others, but also the injunction to invoke one's right to social protection (and , hence invoke the state's duty to provide it) if and only if one has made a good-faith effort to support oneself to the best of one's ability. By implica- ' tion, it is reasonable for the state to design social protection measures with > this in mind. Taylor-Gooby's argument opens a terrain for fruitful conversations between the three schools on how social protection mechanisms should be designed.
Needs
The needs-based group is the least theoretically sophisticated of the three schools considered here. The definition of 'needs' is largely intuitive, or biomedical. The satisfaction of needs is taken as a strong moral imperative, albeit one that also has the benefit of being useful for instrumental purposes as well. Initially seen as a progressive and humanistic approach to development in general, and to economics in particular, the needs-based , school is now denigrated as being paternalistic or welfarist by the rights- / based thinkers. The needs-based approach has been called 'the cattle rear-; 7 ing approach to development': make sure they are adequately nourished, > fully vaccinated and well housed, and everything will be all right.15 Somewhat unfairly, the basic needs school has been accused of ignoring the j rights and dignity of the poor, and of avoiding issues of agency and partici- i pation.16
Coming from economics, albeit from a somewhat heterodox end of the profession, the needs-based thinkers speak a common language with the / risk-based neo-classical economists, at least up to a point. Both can talk of [
40  Social Protection for the Poor and Poorest
Lauchlan T. Munro 41
returns to investment in human capital, for example, and both see that welfare losses resulting from market failures can have both static and Jong-term effects. Both schools acknowledge the problem of scarcity in the lace of unlimited wants, and so both feel comfortable addressing issues such as trade-offs, rates of return to various types of investment, and the long-term effects of resource allocations.
The two schools may diverge, however, over the issue of Pareto optimally. Pareto optimality is politically conservative in that it suggests that welfare improvements should be sought (or can only be proved to have taken place) if the welfare of one party can be increased without reducing [he welfare of another. Risk-based thinking on social protection tends to rely \_ heavily on Pareto optimality as a benchmark. The needs-based thinkers have often had a strong leaning towards redistributionist policies to increase the welfare of the many, even if it is to the (temporary) detriment of the few. Furthermore, some close allies of the basic needs school have shown Ihal, under certain plausible assumptions of unemployment and landlessness, a Pareto optimal equilibrium can reduce some citizens to below starvation levels (Coles and Hammond 1995).
The needs-based thinkers can also communicate with the rights-based thinkers. For example, both the needs-based and theri|;hts-based approaches call attention to the links between those in need and those who already have enough. The needs-based thinkers assert, for example, that even the materially well off in developing countries would be still better off in several ways if their fellow citizens were better educated, healthier and more secure. Rights-based thinkers also sometimes stress the common interest that rights holders and duty bearers have in realising everyone's rights (see, for example, UNICEF 2004; 11-12), though many thjnk this is politically naive (Table 2.1).
Conclusions
It is important to understand why we support what we support. The fuel that the ends are the same (or very similar) does not mean that the pathways /taken to get there were the same. The three pathways to social protection J and poverty eradication - risks, needs and rights - twist and turn. At times they approach each other, occasionally they intersect, but then they diverge again. Opponents of social protection may wish to take advantage of the diverse paths taken by its proponents to sow confusion and to paint a portrait of intellectual confusion surrounding social protection. Those who believe in social protection need to know how to defend it. To do so, they need to understand how risks, needs and rights support and contradict each other as justifications for social protection. This has been the main object ive of this chapter.
Table 2.1 schools
Compatibility and contradiction between the risks, needs and rights
Areas of compatibility  Areas of contradiction
Risks and rights
Everyone has the right to utility
Rights and needs
(1) Development is about people, not commodities
(2) Both rights and needs involve social relations
(3) One theory of human rights is based on human needs, not international or natural law
(4) Rights and needs both have a strong moral dimension
(1) Market failures provide no legal grounding for social protection
(2) Risk-based economists separate the ends from the means of policy; rights-based theorists find this ends-means separation unacceptable
(1) Natural law and legalistic theories of rights reject needs-based arguments as charity, paternalism, welfarism and/or moralism
(2) The rights-based school's doctrine of 'progressive realization' of rights is poorly specified compared to the economists' language of scarcity, trade-offs and opportunity costs
Sonne: Author.
A nother objective of the chapter has been to point out how and where the proponents of these three views of the bases of social protection might I begin to talk to each other. A striking feature of the three schools is the dif-' fcrent levels at which they operate. The risk-based school is highly technical and operational; the rights-based school is highly theoretical and not much concerned with the details of design and implementation of social protection schemes.17 Debate on the design of social protection schemes has traditionally been dominated by the risk-based school - that is, by economists concerned with moral hazard - or by the needs-based school, who have looked at how needs are met or unmet under various schemes. What has been missing is the human rights element. Advocates of the rights-based approach to social protection need to leave their ivory towers, roll up their sleeves and contribute ideas for a rights-based design of social protection schemes.
Second, the underlying basis of utility, essential to risk-based analysis, is philosophically underdeveloped, as is the underlying basis of the basic human needs school. The rights-based school has a comparative advantage
42  Social Protection for the Poor and Poorest
Lauchlan T. Munro 43
here, though its tendency to become hyper-legalistic, literalist and pedantic \ should not be underestimated.18 A needs-based theory of human rights , I seems to offer a fruitful basis for dialogue. Any takers? \ Third, the needs and rights-based schools need to develop (better) ana- ' lytical and design tools. The needs- and rights-based schools do a good job of proving the desirability for social protection in principle, but do a much [ less convincing job in analysing individual social protection situations. ; [ Where they do provide tools for the analysis of such situations, these are -i usually in the form of a needs- or rights-based critique. Furthermore, the two schools provide little, if any, concrete guidance on, for example, the role of social protection in public policy more generally. Finally, the attentive reader will have noted that the treatment of chronic ' poverty, a core theme of this book and part of its title, has been uneven in i the course of this chapter. Far from being an accident, this is a reflection of l the state of the literature. Chronic poverty is a relatively new concern (or at / least a new terminology), and it has entered the mainstream of public policymaking only after many of the classics of the literature on risks, needs and , rights and social protection have been written. An important task for scholars today is to reconsider, and to rewrite, the classics of the field using chronic poverty as a starting point. |;
Notes
1. The views expressed in this chapter are personal, and should not be attributed to any organisation with which the author is or has been affiliated.
2. Witness the recent debates about insurance coverage against terrorist attacks, especially for airlines.
3. Asymmetrical information about the insured person's intentions is the other reason why the private sector does not provide insurance against unemployment (or pregnancy) (Cullis and Jones,1998: 224).
4. An exception is commercial lending to students in professional and graduate schools of, say, medicine, law and business administration. Here future earnings are likely to be high, and previous high performance in undergraduate studies serves as an indicator of future performance.
5. And, in some cases such as refugees, to the citizens of other states as well.
6. The US government is one of the main proponents of the view that there are no economic, social or cultural rights. Of course, various stripes of economic liberals and social conservatives also support this view.
7. Among the other important elements of the corpus of international human rights law are the International Covenant on Civil and Political Rights, the Geneva Conventions, the Convention on the Elimination of All Forms of Discrimination against Women, and the Convention on the Rights of the Child.
8. William O'Neill and Vegard Bye (2002: 17), consultants to the UN Office of the High Commissioner for Human Rights, refer to a 'legalistic approach to human rights [that relies] ...on legal jargon and obscure references to arcane procedures that no one has the time to remember'. Sarah Forti (2005) calls it the 'legal centralist approach to human rights'.
9. The religious origins of the natural law tradition are obvious in the naming of two of the main UN human rights treaties, which are called 'covenants' (UN 1966a; UN 1966b).
10. And all of these things are found, albeit under different names, in the Universal Declaration of Human Rights.
11. (Also frequently known simply as 'basic needs'.
12. 'Subsequent research has shown that the relationship between growth and the satisfaction of needs is varied and complex (Afxentiou 1990; Mehrotra and Jolly 1997).
13. In the short term, some redistribution might be needed via the fiscal system.
14. Or from just about any other school of modern economics, for that matter.
15. David Hulme anecdote told at the opening of the Conference on 'Winners and Losers from Rights-Based Approaches to Development', University of Manchester, 21 February 2005.
16. This is actually a caricature of the basic needs school. The 1970s basic needs school (for example, Streeten etal. 1981) dealt systematically with dignity, security, agency and participation.
17. Or of any other public policy mechanism outside of the realm of judicial and quasi-judicial institutions.
18. See endnote 7.
References
Afxentiou, Panayiotis (1990) 'Basic needs: A survey of the literature', Canadian Journal of Development Studies, XI(2), 241-257.
Arbour, Louise (2005) 'Introduction by the High Commissioner' in Office of the High Commissioner for Human Rights Annual Appeal 2005: Overview of Activities and Financial Requirements. Geneva, OHCHR.
Barr, Nicholas (1992) 'Economic theory and the welfare state: A survey and interpretation', Journal of Economic Literature, XXX(2), 741-803.
Barr, Nicholas (1998) The Economics of the Welfare State, 3rd edition. Oxford: Oxford University Press.
Barry, Norman (1990) Welfare Concepts in the Social Sciences Series, Milton Keynes, UK: Open University Press.
Bentham, Jeremy (1789) An Introduction to the Principles of Morals and Legislation. Oxford: Clarendon Press (many modern editions).
Bentham, Jeremy (1998) Anarchical Fallacies, Headline Series, Foreign Policy Association 318, 56-68.
Brouwer, Marjolein, (2005) 'RBA: A duty to cooperate?' Paper presented to the Conference on 'Winners and Losers in the Rights-based Approach to Development, University of Manchester, 21 February.
Coles, Jeffrey L. and Peter J. Hammond (1995) 'Walrasian equilibrium without survival: Existence, efficiency and remedial policy' in K. Basu etal. (eds) Choice, Welfare and Development: Festschrift for Amartya Sen. Oxford: Clarendon Press.
Cornia, Giovanni Andrea and Frances Stewart (1993) Two Errors of Targeting, Innocenti Occasional Papers, Economic Policy Series, No. 36, Fiscal Policy and the Poor Subsidies, UNICEF, Florence. Reprinted in Dominique van de Walle and Kimberly Nead (eds) (1995) Public Spending and the. Poor: Theory and Evidence. Baltimore and London: Johns Hopkins University Press for the World Bank.
Cornia, Giovanni Andrea, Richard Jolly and Frances Stewart (1987) Adjustment with a Human Face, two volumes. Oxford: Clarendon Press.
44   Social Protection for the Poor and Poorest
Lauchlan T. Munro 45
Courdesse, Laure-Anne and Sarah Hemingway (2005), 'A human rights-ba approach to development in practice: Some lessons learned from the rights-based I municipal assessment and planning project in Bosnia and Herzegovina' Paper r. sented to the Conference on 'Winners and Losers in the Rights-based Approach lo ; Development', University of Manchester, 21 February. \
Cullis, John and Phi lip Jones (1998) Public Finance and Public Choice. Oxford and New \ York: Oxford University Press.
Doyal, Len and !an Gough (1991) A Theory of Human Needs. London: Macmillan.
Engels, Friedrich (1845) Die Lage der Arbeitenden Klasse in England, O. Wigand, Leipzig, \ (English translation 1892 The Condition of the Working Class in England in 1844. \ London: Allen and Unwin).
Forti, Sarah (2005) 'Challenges in the implementation of women's human rights: "_; Field perspectives on domestic violence and HIV/AIDS' Paper presented to the Conference on 'Winners and Losers in the Rights-based Approach to Development', University of Manchester, 21 February.
Goulden, Jay, Brigitte Bode and Elsa Martinez (2005) 'Putting rights-based development into context: CARE's programming approaches in Malawi and Bangladesh' Paper presented to the Conference on 'Winners and Losers in the Rights-based ; Approach to Development', University of Manchester, 21 February.
Hobbes, Thomas (1651) Leviathan. London: Oxford University Press. (Many modern editions.)
International Labour Office (ILO) (1942) Approaches to Social security: An International Survey, Scries No. 18. Montreal: McGill University. J;
Jolly, R. and G. A. Cornia (1984) The Impact of World Recession on Children: A UNICEF . Report. New York: Pergamon Press.
Ljungman, Ceclia and Sarah Forti (2005) 'Applying a rights-based approach to development: Concepts and principles' Paper presented to the Conference on 'Winners and Losers in the Rights-based Approach to Development', University of Manchester, 21 February.
Maslov, A. H. (1943) 'A theory of human motivation'. Psychological Review, 50(4), 370-396.
McHale, John and Magda Cordell McHale (1977) Basic Human Needs: A Framework for Action, Report to the UN Environment Programme. New Brunswick, NJ: Transaction Books.
Mehrotra, Santosh and Richard Jolly (eds) (1997) Development with a Human Face: Experiences in Social Achievement and Economic. Growth. Oxford: Clarendon Press.
Munro, lauchlan T. (2000) 'Is there a liberal case for the comprehensive welfare state?' Discussion Paper No. 61, Institute for Development Policy and Management, University of Manchester, May.
Munro, Lauchlan T. (2002) 'Social protection', in Colin Kirkpatrick, Ron Clarke and Charles Polidano (eds), Handbook on Development Policy and Management. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 182-192.
Munro, Lauchlan T. (forthcoming) 'The "human rights-based approach to programming": A contradiction in terms?', in Sam Hlckey and Diana Mitlin (eds), Rights, Wrongs and Realities: Reflections on the Rights-Based Approach.
NORAD (2001) Handbook in Human Rights Assessment: State Obligations, Awareness and Empowerment. Oslo: Norwegian Agency for Development Cooperation.
O'Neill, William and Vegard Bye (2002) 'From Lligh Principles to Operational Practice: Strengthening OHCHR Capacity to support UN Country Teams to Integrate Human Rights in Development Programming', Consultancy report to the UN Office of the High Commissioner for Human Rights, Geneva.
Bights and Democracy (2003) 'Economic and social rights: Canada's Supreme Court
disappoints' Libertas, Issue 9, International Centre for Human Rights and
Democratic Development, Montreal, January. Riis, Jacob A. (1890) How the Other Half Lives: Studies among the Tenements of New York.
iNew York: Charles Scribner. gpbertson, Robert E. (1994) 'Measuring state compliance with the obligation to
devote "the maximum available resources" to realizing economic, social and cul-- tural rights', Human Rights Quarterly, 16, 693-714. Rowntree, S. (1901) Poverty: A Study in Town Life. London: Macmillan. Seers, Dudley (1969) 'The meaning of development', International Development Review,
11(4), 2-6.
Sen, Amartya (2004) 'Elements of a theory of human rights', Philosophy and Public
Affairs, 32(4), 315-357. Stiglitz, Joseph E. (2000) Economics of the Public Sector, 3rd edition. New York and
London: W.W. Norton and Co. Streeten, Paul, Shahid Javed Burki, Mahbub ul Haq, Norman Hicks and Frances
Stewart (1981) First Things First: Meeting Basic Human Needs in Developing Countries.
Oxford and New York: Oxford University Press. Taylor-Gooby, Peter (1991) Social Change, Social Welfare and Social Science. London and
New York: Llarvester-Wheatsheaf. UN (1948) Universal Declaration of Human Rights. New York and Geneva: United
Nations.
UN (1966a) International Covenant on Economic, Social and Cultural Rights. New York
and Geneva: United Nations. UN (1966b) International Covenant on Civil and Political Rights. New York and Geneva:
United Nations.
UN (1989) Convention on the Rights of the Child. New York and Geneva: United Nations.
(Downloaded from http://www.ohchr.org/english/law/crc.htm, on February 2005). UNDG (2003) The Human Rights Based Approach to Development Cooperation: Towards
a Common Understanding among Un Agencies. New York: UN Development Group
Office. (Downloaded from www.undp.org/governance/docshurist/030616Commo
nUnderstanding.doc, on 27 February 2005). UNDP (1990) Human Development Report 1990. New York: Oxford University Press. UNICEF (1.987) Children on the Front Line: The Impact of Apartheid, Destabilisation
and Warfare on Children in Southern and South Africa. Geneva and New York:
UNICEF.
UNICEF (1998) Guidelines for Human Rights-Based Programming Approach - A Human Rights Approach to UNICEF Programming for Children and Women: What It Is, and Some Changes It Will Bring, CF/EXD/1998-004. New York: UNICEF, 17 April.
UNICEF (2000a) Poverty Reduction Begins with Children, Division of Evaluation, Policy
and Planning. New York: UNICEF, March. UNICEF (2000b) Programme Policy and Procedures Manual: Programme Operations,
Programme Division. New York: UNICEF, January. UNICEF (2004) Programme Policy and Procedures Manual: Programme Operations,
Revised Edition, Division of Policy and Planning. New York: UNICEF, June. UNICEF (2005) All Rights for All Children: UNICEF in Central and Eastern Europe and the
Commonwealth of Independent States, Communication and Partnership Section,
UNICEF Regional Office for CEE/CIS, Geneva, (Downloaded from http://www.
unicef.org/publications/index_25088.html on 22 April 2005).
46  Social Protection for the Poor and Poorest
Vandemoortele, Jan (2000) Absorbing Social Shocks, Protecting Children and Reducing -I
Poverty: The Role of Basic Social Services, Staff Working Papers, Evaluation, Polle ! and Planning Series No. EPP-00-001. New York: UNICEF.
Viner, Jacob (1949) 'Bentham and J.S. Mill: The utilitarian background', American 1
Economic Review, XXXIX(2), 360-382. i
Insurance for the Poor?
Stefan Dercon, Tessa Bold and Cesar Calvo1
Introduction
Households in developing countries are exposed to high risks, with important consequences for their welfare. It has long been acknowledged thai shocks, ranging from individual-specific (such as illness, theft or unemployment) to economy-wide (such as droughts or recessions), have impor-laril implications for consumption and nutrition, not least among the poor. Policy responses have mainly focused on safety nets or other social security mechanisms. This chapter goes beyond this view by arguing, first, that the costs associated with these risks are much higher than those estimated from thoii short-term impact and, second, that expanding insurance provision lor the poor could result in substantial long-term welfare benefits. To illus-li.ife this possibility, we use examples, mainly from Latin America, starting from a consideration of how risk affects the poor and the ways in which lhey respond to it. The chapter assesses the most promising insurance inst ruments, while emphasizing that expanding insurance provision should nol be seen as a panacea, but instead be viewed as part of a comprehensive extension of protection of the poor.
In addressing the case for extending insurance to the poor, a number of key questions need to be answered. Is risk prevalent? What are the main risks laced by the poor? Thinking about the design and promotion of specific insurance products requires a careful understanding of the risks the poor lace, and their consequences. Recent surveys have highlighted the variety of i isks faced by the poor (Morduch, 1995; Townsend, 1995; World Bank, 2000; JXMton, 2002; Fafchamps and Lund, 2003). Some of these risks are relatively Miaightforward to insure, such as funeral or health costs, while in the case of oihcis - say, country-wide recessions or crime - several factors come into play. Any discussion about insurance for the poor will need to acknowledge the shortcomings of an insurance-related approach and the need for alternative mechanisms to deal with the implications of particular shocks. We argue here that there is a need to think in terms of complementarities
47
48  Social Protection for the Poor and Poorest
Stefan Dercon, Tessa Bold and Cesar Calvo 49
with other mechanisms to reduce and to cope with risk, including safety * nets. The first section will take up these issues in more detail.
The poor try actively to manage risk and to cope with its consequences. I
A study of the strategies employed to manage and to cope with risk helps us \
to understand the implications of risk for welfare as well as for the design \
of policy responses, including insurance. Much research has been con- = ducted in recent years on these strategies and their implications, but n
work is definitely needed. These strategies typically involve households I ry- [
ing to shape the risk they face through managing their activity and asset J
portfolios. A typical example is diversification of activities. Combining ;
economic activities with returns that are imperfectly correlated reduces |
overall exposure to risk. This also implies accepting lower mean returns, ;
which could effectively perpetuate poverty in the long run. Other strafe- \
gies involve risk-coping mechanisms - for example, self-insuring through \
accumulating assets in good years to be drawn on in bad years, or entei ing I
into informal mutual assistance arrangements within families or commu- !
nities. Most evidence suggests, nevertheless, that risk management and I
coping is rather imperfect, and shocks result in substantial fluctuations in \
welfare outcomes, which also undermine the asset base of households lor \
future wealth creation, not just in terms of physical, and financial assels, |
but also nutrition and human capital (Morduch, '1995; Dercon, 2002; I Dercon and Hoddinott, 2004).
The impact of shocks and the limited range of risk strategies availabl
those in poverty in developing countries significantly undermine their abil- j
ity to grow out of poverty, and have important implications for their wel- t
fare. The overall result is not only fluctuations in welfare levels, but also a 1
loss of efficiency in that the poor are induced to use their assets less efti- |
ciently than the rich. Theoretical models such as the one developed hy ?
Banerjee and Newman (1993) build on this feature to show that risk may )
well result in poverty traps, a situation in which those who cannot escape [
poverty by their own means end up living in permanent poverty, even if [
other sectors of the economy are growing). The implication is also that there I
is no trade-off between equity and efficiency when measures are taken to {
avoid those poverty traps. In other words, there is a case for providing insur- \
ance at subsidized rates so that some do not slip into poverty.This point is \ expanded on further in the second section.
In the third section , the case for strengthening insurance and insurance \ substitutes is made, but it is also acknowledged that insurance products can
be costly and, more importantly, that they cannot solve all the problems ; facing the poor. Alternative (complementary) measures are discussed as well, not least in response to the realization that some economic shocks or
social and political risks usually cannot be insured by insurance markets, or \ at least that alternative measures may be more cost-effective. Still, the scope for insurance products for the poor remains strong.
The fourth section discusses general issues related to the design of insurance for the poor. A key issue to be considered is that one must ensure t-har the poor are reached effectively, which suggests the need to involve local and grass-roots organizations with established links to the poor. Equally important is that the system should be able to provide a cost-effective service and be sustainable. Insurance provision is a specialized service, and should involve private (and possibly public-sector) institutions wilh the experience and financial capacity to operate such schemes. A 'partner-agent' model is likely to be the most effective institutional arrangement. The final section notes the main conclusions.
Kisk and the poor
There are a number of ways of classifying risks faced by the poor. Two issues are relevant for our purposes: the extent to which the poor are affected by these sources of risk, and the extent to which developing more insurance is an appropriate response. To discuss the first point, the section discusses cova nance of risks across a population, and the frequency of risks over time. Insurance contracts are offered most easily if risks within the relevant population are not covariate - that is, risks do not affect a large proportion of the population at the same time. Insurance for rare and infrequent events is also typically more difficult to offer. Taken together, insurance contracts are less likely to be on offer for rare and covariate shocks.
Tu rning to the typology of risks, in recent years a number of studies have highlighted the risks facing the poor. The focus has largely been on data from South Asia and, more recently, from Africa (Morduch, 1995; Dercon, 2002). One key difference between these regions and Latin America and the Caribbean is the urban nature of poverty in the latter. Whereas poverty is mainly a rural phenomenon in Africa and Asia, in Latin America and the Caribbean urban areas account for most of the poor. There are risks that are specific to the urban poor but not to the rural poor, and vice versa, or at Icisl risks that have a different intensity from those in a rural setting.2 The discussion that follows attempts to give at least equal weight to both types of settings.
I;evv studies have tried to record systematically the sources of risk faced by the poor. As part of a more extensive World Bank study on poverty in (iiialemala, the analysis by Tesliuc and Lindert (2002) provides a unique insight into those sources and their consequences in that country, based on a specifically designed household survey combined with focus group interviews, conducted in the year 2000. It should be noted that Guatemala is one of i lie most rural economies in Latin America, presenting the lowest percentage of urban poor, so these findings need to be supplemented by other sources for a more complete picture. Although no serious economic crises or nalural disasters occurred in 2000, Tesliuc and Lindert found that about
50  Social Protection for the Poor and Poorest
Stefan Dercon, Tessa Bold and Cesar Calvo 51
Table 3.1   Incidence of reported shocks, Guatemala, 2000 Frequency reported   Types of shocks 15 or more Pests, lost harvest
6-14 per cent Income drop, accident of breadwinner, job losses, drought
2-5 per cent Loss of terms of trade, crime, floods
0-2 per cent Hurricane, bankruptcy, landslide, death of breadwinner,
enterprise closure, land dispute, fire, earthquake
Source: Tesliuc and Lindert (2002).
53 per cent of the households interviewed reported one or more shocks: 23 per cent mentioned a 'natural' shock (from pests to forest fires and floods); 17 per cent reported an economic or other human-related shock; and 13 per cent reported both. The most commonly reported shocks were related to agriculture, but many different types of shocks were identified (see Table 3.1). Some of these shocks were largely 'rural' - pests, lost harvest and drought are the obvious ones here, while others ^are common in urban areas - including crime and job losses, which had more than double the incidence in urban areas. Accidents and floods have a similar incidence in both rural and urban settings, with no apparent differences in the overall reported incidence of shocks across both areas. Other studies in the same period confirm the high incidence of shocks. For example, Gaviria and Pages (1999) report that, in the first half of 2000, 36 per cent of urban Guatemalans reported a shock causing a loss of income. A study on Peru (Chacaltana, 2002) reported about 30 per cent of households facing a significant shock to income or wealth in 2001. In short, all available evidence suggests that shocks are prevalent in both rural and urban households in Latin America and the Caribbean.
The study by Tesliuc and Lindert (2002) has other striking findings. First, in 2000, all these shocks were typically idiosyncratic; that is, only a sub-section of a local population was affected. It should nevertheless be recalled that this was largely a year without serious 'large' and covariant shocks. During the previous five years, some shocks had a much higher incidence, including hurricanes (Hurricane Mitch was included in this period), with 44 per cent reportedly being affected, while forest fires hit 17 per cent at least once over a five-year period. These types of risks are largely covariant, affecting typically whole communities or neighbourhoods. Second, there was a rather high incidence of households being affected by more multiple shocks - a phenomenon known as 'bunching' of risk, which may exacerbate the consequences of shocks. Agricultural shocks, such as drought and pests, tend to come together, as do economic shocks, such as job losses and accidents, or the death of a breadwinner. They also find that the poorest are more affected by
Table 3.2 Effects of natural disasters in Latin-American and Caribbean countries, ^70-2001
Caribbean
Central
America      South America
Total
..Affected (000s)
Killed
Injured . Homeless
Damage
<US$ 000s)
Annual damage
19,774 5 8 971 10,187,949
318,373
20,146 85 202 2,664 23,121,364
722,543
104,980 156 276 4,240 35,192,517
1,099,766
144,900 247 486 7,875 68,501,830
2,140,682
Source: From Chacaltana (2002), based on CR ED data.
shocks. In terms of asset or welfare loss, the poor are typically hit hardest. This is especially the case for shocks related to agricultural risks.
In Tesliuc and Lindert (2002), health shocks were not analysed in their survey, which does not mean that they are irrelevant. For example, in a rural sample in Mexico (reported in Ibarra, 2003), it was found that about 48 per cent of households reported a shock in wealth or welfare related to a drop in yields (largely a result of weather-related events), while the second main cause was illness of the farmer or a member of his/her family (reported by about 15 per cent of households). This is consistent with surveys from across the world where illness is typically the second most frequent risk in rural settings after crop failures (see Dercon, 2002), and before many other risks (such as loss of livestock, crime or fire).
In sum, a discussion of the risks faced by the poor in Latin-American and Caribbean countries will need to consider those risks with high incidence and with serious consequences. 'Natural' risks are clearly at the top of the list, but not only for rural settings. Some are obviously rural, such as those related to harvest losses caused by drought or pests, but floods or hurricane risks, and other large-scale natural disasters, are also relevant to urban settings. Data on the impact of large-scale natural disasters over a thirty-year period (see Table 3.2) suggest that their impact on households - ranging from death, injury, homelessness and physical damage - is substantial.
Market failures and household responses to risk
If these risks are substantial and the consequences are as serious as suggested above, one questions why insurance markets are not offering insurance contracts to the poor. There are a number of reasons why this may not occur. Information asymmetries undermine the provision of insurance. Insurance contracts are exposed to adverse selection (hidden information)
52 Social Protection for the Poor and Poorest
Stefan Dercon, Tessa Bold and Cesar Calvo 53
and moral hazard (hidden action). These have been held responsible for the failure of crop insurance systems (Binswanger, 1986; Braverman and Guasch, 1986), or the lack of insurance against natural disasters or economic recessions. However, it remains to be explained why these asymmetries might be considered more perverse when policyholders are poor.
Insurance providers mitigate information asymmetry by promoting group insurance (against adverse selection) and by requiring co-payments and deductibles (against moral hazard). Although insuring large groups is a feasible strategy, co-payments and deductibles may well discourage the poor from buying the product. In any case, these payments and deductibles will probably need to be lower than the values needed to separate 'good risks' from 'bad risks' (the so-called second-best separating-equilibrium values). As contracts will still allow for significant moral hazard, insurers will require high premiums and thus discourage the poor.
A related issue is that as the poor do not usually participate in the formal economy, formal insurers also face enforcement problems and/or the poor confront extra costs. For example, claiming for home insurance when there are no formal titles to land or homes imposes extra verification costs, which discourages firms from offering contracts to the poor or makes them less attractive. Similarly, the assets o\ the poor may be of relatively low value, so the transaction costs involved in valuation would be relatively high relative to the size of the contract. Costs related to providing birth and death certificates may also make insurance contracts less attractive to the poor.
Supplying the poor with insurance implies further high transaction costs. For example, micro-credit experiences suggest that the poor find it easier to deal with frequent repayment in small instalments.3 Tliis suggests that payments of premiums may also occur ideally in small instalments, adding transaction costs to insurance provision.
Furthermore, it has been suggested that the poor sometimes have difficulty in fully understanding their rights in insurance contracts. McCord etal. (2001) report several cases where the poor did not file their claims after being affected by an event covered by their policies. In other cases, some policyholders have expected coverage beyond the scope of their contract.
Finally, many of the most serious risks faced by the poor may well be cov-ariant and therefore not easily insured by an emerging insurance market. The fact that a sizeable part of the population is dependent on agriculture, and that macroeconomic instability is substantially higher in developing rather than in richer countries (resulting in serious covariate shocks in the economy) is bound to limit the emergence of private insurance focused on poorer segments of the population.
The lack of market-based insurance could in principle have been compensated for by 'social insurance' - or public-sector-based insurance provision as part of broader social security programmes. In practice, the coverage of
these programmes for the poor is minimal in most Latin-American and Caribbean countries. For example, a recent review of Guatemala's social insurance system concluded that the 'system provided minimal coverage of
;e population, risks financial crisis, faces allegations of corruption and is /regressive' (World Bank, 2003). The Instituto Guatemalteco de Seguridad Social (1GSS) covers workers in the formal private and public sectors only, and runs a number of programmes. Programmes analysed, such as the accident-maternity-sickness (IVS) programme, were shown not only to be in deficit but also regressive in terms of the incidence of benefits.
The lack of formal insurance or social insurance systems does not mean that the poor are passive towards the risks they face. Much of their livelihood is centred around ways to reduce, mitigate and cope with risks. The poor use risk management and risk coping strategies to alleviate risks. Table 3.3 describes these strategies and their shortcomings. By risk management we mean that they try to reduce their exposure to risk, or mitigate the risk of some income sources by combining them with others. Diversification of crops and other sources of income is one typical example. Other common strategies involve migration and relative specialization in low-risk activities, even at the cost of lower returns. Risk-coping strategies effectively try to smooth consumption given income fluctuations linked to risk. These
Table 3.3  Risk management and coping strategies
Strategy
Examples
Shortcomings
Ma naging and reducing risk    Crop diversification   Sacrifice of expected income resulting from changes in       Specialization in sources of income low-risk activities
Migration of some members of the household
Asset: management Savings as Lack of suitable saving
self-insurance assets (risky or bulky assets,
insecurity)
Focus on liquid, less productive assets
Long building-up time
Covariancc in asset price and income
Informal insurance Reciprocal gifts/        Incomplete protection
loans from friends/    vulnerability to covariant relatives risks
Market-based Insurance Typically not available
Source: Based on Dercon (2002), Holzmann and jorgensen (2000) and World Bank (2000).
54  Social Protection for the Poor and Poorest
strategies include self-insurance - that is, building up suitable liquid assets in good years that can be drawn on during a bad year. An alternative strat-egy is to enter into informal 'risk-sharing arrangements' - that is, informal insurance arrangements based on reciprocal gifts or contingent credit.
Risk management and coping strategies are always present in the life of the household. However, if a serious crisis occurs, households resort to moi extreme actions - that is, survival strategies: 'emergency' actions to be take when a reduction in income is unavoidable. See Table 3.4 for a summary of such strategies.4
Some striking conclusions emerge from the empirical literature on this issue. Households cope with risk mainly by using income-based strategies such as diversification of income sources - and assets for buffering consumption. Informal insurance and credit is also used, but only in a relatively limited number of cases. These strategies are not without cost. As has been widely documented, both income and asset-based strategies imply efficiency losses in the generation of income, and thus may lead to poverty traps (see, for example, Rosenzweig and Binswanger, 1993; Rosenzweig and Wolpin, 1993; Dercon, 2002).5 Lack of formal or informal insurance provision forces households to choose a safe portfolio of activities and'assets, which typically implies a lower mean return. Lack of insurance provision precludes the poor from taking risks to raise their income, thus perpetuating their poverty. The process is exacerbated by shocks as asset holdings are difficult to rebuild after they are depleted. In emergency cases, households are also forced to sacrifice human capital, as is the case when children drop out ot school and start working (Pizarro, 2001). In Peru, Jacoby (1994) finds thai 'children from households with lower income...and greater childcaie responsibilities begin withdrawing from school earlier'. In urban areas, issues of privacy in the household arise as families rent out rooms, or children come back to live in the parental house in order to rent out their own
Table 3.4  Survival strategies
Strategy	Examples	Shortcomings
Changes in sources of income	Child labour	Sacrifice of human capital
Asset management	Selling/pawning of real/productive assets	Long time to replace them
Informal insurance	Charity	Incomplete protection Vulnerability to covariant risks
Market-based	Bank loans for consumption credit	Usually not available
Source: Based on Dercon (2002), Hokmann and jorgensen (2000) and World Bank (2000).
Stefan Dercon, Tessa Bold and Cesar Calvo 55
houses (Zaffaroni, 1999). In fact, after analysing data on Guatemala, Tesliuc ind Lindert (2002) concluded that 'the poor have lower resilience than the -ich to the effects of shocks. The probability of restoring household income 0 the level that prevailed before the occurrence of the shock rises with I      jrra>rne'(p. 37).
In short, the restricted risk strategies available to the poor result in effi-( ciency losses (Rosenzweig and Binswanger, 1993). It also means that the ! welfare losses caused by lack of insurance are well beyond those arising [ from income fluctuations and other transient effects in consumption. They \ involve permanent or chronic effects on poverty, implying substantially \ higher welfare costs and lower efficiency. These efficiency losses also mean ; that effective interventions could be implemented without involving a i trade-off between efficiency and equity, since increasing equity (spending ; focused on the poor), also raises efficiency.6 If providing insurance could J enable the poor to take on riskier, but higher-return activities, then in prin-f ciple support for the poor could be financed by improvements in efficiency, j This makes a case for interventions to encourage insurance with public (and \     aid) money based on efficiency considerations (see Dercon, 2004).
St 11 pe for insurance provision to the poor
| Insl ruments that could provide incentives and means for the poor to protect | themselves ex ante against hardship include better insurance products, self-| insurance via savings, and access to credit in order to facilitate asset build-i ng and properly manage those risks that might affect income. Ex ante measures should also focus on reducing risk itself. Ex post measures would provide a genuine safety net, targeted appropriately at the poor but large enough in scale and coverage to provide broad social protection to assure a minimal and sustainable standard of living. Such measures could be part of a more general welfare support system, or be targeted specifically to respond to risk-relnted hardship.
The potential role of these complementary ex ante measures should be stressed here. One set of measures involves reducing directly the risks faced by I he poor - for example, policies supporting health prevention and sanitation. Better information systems on prices and weather conditions could have substantial benefits, while investments in technology could reduce certain types of risk, with irrigation systems and drought-resistant crops being very good examples of the latter. Indeed, these types of measures could make certain risks, at present too large or covariate to be offered viable insurance cover, more easily insurable in a cost-effective way. They clearly highlight the need for multisectoral approaches to deal with risk and insurance.
Other financial products can also play a part in coping with risk. Savings instruments, for example, have been largely undervalued as an effective
56  Social Protection for the Poor and Poorest
instrument for protection against hardship (Dercon, 2002; Morduch, 2004) While credit provision to the poor has received much attention, relatively little has been directed towards savings, even when theses do present man5 advantages as an area for subsidized intervention and regulation. For exam pie, they are not affected by the information or reinsurance problems con cerning credit and insurance, and transaction costs involved in thes< operations, while not negligible, are likely to be largely restricted to thi administrative handling of savings. One of the key issues is that insurana via financial assets may be risky, especially given the endemic risks of infla tion in developing countries. Typically, financial savings are not tailored t( the poor, offer low or negative returns, and involve extremely high transaction costs imposed on the saver. Savings should not be considered as fust; means of building a credit-worthy reputation or mobilize capital aside fron the normal economic activity of the household. Typical products are tai lored to long-term deposits, with highly punitive returns for those looking for flexible instruments to respond to unexpected hardship.
Credit products could also help to provide better protection against risk Credit can act as an insurance substitute, and products for this purpose should be part of the standard portfolio of financia [Instruments offered t( the poor. Furthermore, credit can help to diversify t|e source of income am build up assets. It can also increase income, reduce risk in income, anc enhance the ability to cope with shocks that might affect income. Financia products for the poor should be flexible and take into account the fact thai they face substantial risks. Linked credit and insurance contracts are on< option - for example, linking credit and health insurance. This form of insurance is not the focus of the present chapter, but there is a definite need for more research on such products. *
As part of a general system of protection against risk-induced poverty, there is clear scope for insurance targeted at the poor. Furthermore, evidence suggests that the best method for offering insurance to the poor is the partner-agent model, in which an established insurer, possibly with public-sector support, co-operates with local microfinance institutions. Existing informal institutions are also potential agents.
Insurance involves the pooling of risk over a large number of similar units and is most appropriate for uncertain and high losses, which are greater than the amount a household can save for, or repay. When the loss and the degree of uncertainty decrease, insurance loses out to credit and saving. Insurance therefore involves exchanging the uncertainty of large losses for the certainty of small regular payments. Policyholders pay for the losses incurred by others, while the costs and risk are assumed by the insurer. For less uncertain or smaller losses, savings or credit may be more appropriate.7
Brown and Churchill (2000) suggest that there is scope for insurance provision only when the following criteria are met: (i) a large number of similar units exposed to risk; (ii) limited policyholder control over the insured
Stefan Dercon, Tessa Bold and Cesar Calvo 57
event; (iii) the existence of insurable interest; (iv) losses can be identified ami measured; (v) losses should not be catastrophic: reinsurance becomes increasingly difficult with increasing covariance across people (such as a hurricane or a flood); (vi) availability of historical information on a sufficiently large number of people or property exposed to the same risk, so that probability of loss can be estimated; and (vii) premiums are affordable. They propose a rule of thumb by which, if the probability of a loss exceeds 40 per cent, premiums will definitely be too high to be affordable.
There are numerous examples of insurance schemes that have been introduced without meeting these criteria, one of the most infamous examples being the crop insurance programmes introduced in the early 1980s in different parts of the world. Many of the criteria included above apply to both poor and rich insurance clients. However, some of them make it particularly difficult to insure the poor profitably. The need for premiums to be economically affordable often means that the policy portfolio cannot in fact be covered by contributions, or that insured amounts are so small that they make little difference to the vulnerability of the poor. Self-Employed Workers Association (SEWA), an Indian health and life insurer, is a case in point, willi payouts so low that they only cover about 10 per cent of losses caused by i llness-related shocks. Insurance to the poor is traditionally fraught, with high per-unit transaction costs, because premiums need to be small and be collected frequently, while the total amounts of policies are also small. Problems such as moral hazard and adverse selection are not necessarily more damaging among the poor, but the higher transaction costs in dealing wilh them may mean that these issues make insurance unprofitable. None the less, a number of small-sized (microfinance) institutions already cater to Hie poor. Their successful experiences may help to develop some best-practice guidelines for potential entrants into the small-scale insurance market who want to target the poor.8 Some of these lessons are discussed in the fourth and fifth sections of this chapter.
A key lesson is that ex ante measures, in the form of a savings, credit and insurance system, may provide substantial protection to the poor, but ultimately they cannot fully insure individuals and families. In short, some ex post measures that entail transfers to those affected by uninsured risk would still be necessary as part of a comprehensive system to protect the poor against risk. Insurance products for the poor need to be simple, insuring only specific, highly observable risks with measurable losses, while high-risk groups may need to be excluded by design for the scheme to be sustainable.
All self-protection strategies require some outlay beforehand, and self-insurance fails if shocks occur in successive periods. Credit as a substitute for insurance may not be available either. Certain highly covariate and rare events are very difficult to insure. This means that some 'natural' risks, such as catastrophes, may not be covered easily by a pure insurance system. Other
58  Social Protection for the Poor and Poorest
risks require that other types of measures be applied, and market-based J insurance products are unlikely to be the most sensible or the only response: I 'social' risks such as crime or enforcement of property rights are examples. J. While it is possible to design products that insure against the consequences J of these risks, they only address part of the problem. j But even if there are clear limits to insurance provision for the poor as a f solution for their vulnerability to risk, insurance is definitely an option i worth focusing on. In particular, life and health insurance, as well as forms of property and asset insurance, are possibilities, and even insurance against some covariate risks, such as drought or, in general, weather insurance.
Implementing insurance for the poor
Some key issues pertaining to insurance management need to be addressed when implementing insurance programmes for the poor. This section addresses institutional arrangements and issues such as financial management, premium calculation, distribution of services, and reinsurance, with a particular focus on targeting the poor.
It is paramount that agents involved in insurance schemes have very close contact with the poor. This is unlikely to be achieve^ by either government agencies or by standard private insurance providers.' As such, institutions with close links to grassroots organizations or NGOs may be ideal agents - ■■ for example, microfinance institutions (MFIs), which are relatively widespread within developing countries. But, since purchasing insurance involves a payout only in the case of an adverse shock, it is critical thai insurance customers clearly know and understand the benefits to which they are entitled. This requires a simple and clearly stated policy, swifl processing of claims, and careful financial management of the insurance portfolio by the insurance provider. To inspire trust among the clientele, adequate reserves need to be held and financed through underwriting, reinsurance and investment. To be financially viable, insurers need to have e sufficiently diversified investment portfolio. This is something that MFIs oi other institutions working closely with the poor may often find hard to l achieve.
A partner-agent arrangement, in which a local institution or the MFI undertakes only the distribution of insurance services, linked with a private or possibly public-sector insurance provider, may therefore be more appropriate when targeting poor customers. One of its advantages is that it elimi- • nates agent risk and allows the institutions involved to focus on their particular strengths. It also allows local institutions and MFIs to offer greater benefits to policyholders at a similar cost. The most important drawback of this model is the limited availability of potential partners. Fostering these relationships is an issue that public policy should address, by providing a clear institutional and regulatory environment. This point is discussed in
Stefan Dercon, Tessa Bold and Cesar Calvo 59
further detail in the sixth section, below. Within the context of a partner-agent arrangement, mutual insurance funds may overcome some of the resistance to insurance, since they mimic features of informal insurance arrangements in which funds are often returned to members at regular intervals. Exiting informal arrangements may, however, become a part of an y$Vs established set of procedures.
Turning to the issue of premium-setting, most of the existing insurers surveyed by Brown and Churchill (2000) calculated their premiums either in house or by partnering with an established insurer to gain access to the required expertise. Brown and Churchill also found that MFIs that co-operate with established insurers are usually able to offer coverage at better prices. Instituto de Fomento a la Comercializacion Campesina (IFOCC) in Peru searched for partners with the actuarial expertise they lacked, but they were unable to find an established insurer willing to provide a product for the low-income market. Instead, they used their own simple calculations based on historical mortality statistics within their credit portfolio. ASA in Bangladesh, however, followed a different and far more risky approach, basing their premiums on customer demand, starting out with very high premiums on their mandatory insurance policy. Numerous complaints were received from their clients, however, so premiums were then lowered successively until complaints stopped. While this ensures that clients are able to afford premiums and are satisfied with the rates offered, it obviously entails a higher risk than the calculation of premiums based on actuarial principles.
As Rutherford (1999) points out, one of the most important demands the poor make on their financial services is easy access and regular small payments, which impose the necessary payment discipline. An agency employing home service distribution and collecting premiums on a weekly basis would be well suited to the needs of low-income households, although this may incur high transaction costs. Integrated distribution, as practiced by SEWA in India, where life insurance is distributed through already existing fixed deposit accounts, could help to curb these costs.
Reinsurance is one element that is almost completely absent in micro-insurance and similar insurance institutions focused on the poor. One of its many benefits is that it can improve the ability of insurers to grow, helps lo stabilize financial results, protects against catastrophic losses, and improves underwriting expertise. Reinsurance in low-income markets can also open up markets for some of the large-scale covariant risks such as many natural disasters (Skees etal., 2004). However, to attract reinsurance, it is of critical importance that primary insurers have sound pricing policies and controls to deal with abuse. According to respondents in the Brown and Churchill survey, all partner institutions in partner-agent arrangements and some co-operative insurers were likely to have reinsurance contracts. However, few of the MFIs and other smaller organizations in their study have reinsurance, which leaves them highly exposed to sudden
60 Social Protection for the Poor and Poorest
Stefan Dercon, Tessa Bold and Cesar Calvo 61
increases in claims, and prevents them from having access to a potential valuable source of expertise.
There is substantial scope for the government to effectively support insurance market serving the poor. A favourable policy environment >
support the proliferation of insurance services among the poor by faciliiat- e ing the establishment of local (micro) finance institutions, and mat
insurance provision to the low-income segment of the market more attrac- f
tive to established insurers. The strategy is unlikely to involve large-scale \ subsidies, but government spending will need to be directed towards eslal
lishing the necessary infrastructure, institutions and regulatory environ- \ ment to promote this segment of the market. Unfortunately, such a polit
environment does not exist in most parts of the world and, in fact, some of *
the existing regulations actually present a bias against the use of finance \ and insurance products targeted at the poor (see Dercon et ah, 2005) for further discussion).
"''ill
Conclusions
The poor face substantial risk, in the form of natural, health, social and |
economic risks, and are also more likely to be affected by these. In general, [
the poor use sophisticated strategies to cope with these risks, but (he I
measures are inadequate and welfare losses substantial. The strategies I
themselves come at an additional cost in terms of long-term welfare. Risk, [
and how the poor respond to it, contributes to the persistence of poverty, j
There is a clear need for policies to reduce risk and its consequences for [
the poorest. Indeed, there may well be an efficiency argument for provid- I
ing subsidized insurance and protection, given risk-induced poverty traps. I
This chapter has argued for the need to foster insurance provision, nol as \
panacea to solve all problems, but as part of a comprehensive system. The l
current focus on ex post measures in the form of some kind of safety net is i
not cost-effective or sufficient to reach the poor. Other components of such |
a system would be ex ante measures to stimulate and protect self-insurance \
through savings, reducing risk and fostering credit for the poor as a form of I insurance and to allow a stronger asset base to grow out of persistent vuh -
ability to risk. These efforts need to be supplemented by a careful and well- \
designed safety net, since some risks should not be addressed by 'ex ante' i
insurance-related mechanisms - examples are certain covariate economic or }
catastrophic risks. A high proportion of risk, including economic and social \
risk, is also largely human-made, and reducing the impact of this requires [
actions to address its causes - inflation, crime or waste-related risks are \
examples. Providing protection only against the consequences of such risks \ is unlikely to be cost-effective.
In terms of the basic institutional arrangement for insurance provision, [ the partner-agent model appears to be the most suitable, so that an
esfablished insurer (the partner, from the private sector, possibly in partnership with the public sector) links up with an institution with local financial connections, such as a microfinanee institution (MFI). The advantage is that this would include a mechanism to provide easy access and terins to the poor, while costs are reduced and sustainability is protected by rejnsUrance and contracting with an established insurer. Judging from case studies, it appears that a number of products might be suitable for promotion, including life, property, health, weather or price insurance, possibly linked to credit. One should be cautious about the likely success of these schemes. In terms of types of coverage, the experience of existing insurers catering to the poor shows that it is difficult to offer profitable comprehensive coverage to low-income households. In part, this can be explained by the financial capacity of clients and the lack of opportunities for diversification. Term life insurance is the most sustainable type of insurance, but the support of governments, donors and NGOs is necessary to branch out into other profitable products. Product features should include group policies, mandatory insurance and incentives to cope with moral hazard and adverse selection - for example, by rewarding members who do not submit any claims during the year.
While subsidized insurance for the poor can be an attractive option on efficiency grounds, an important role for the government would be to establish a more effective regulatory framework to foster the establishment of micro-insurers at the local level, while maintaining overall stability and credibility of the entire financial system. While relaxing entry requirements for MFIs to enter the insurance market may be beneficial, incentives should he also provided so that MFIs partner with established insurers, through the part uer-agent method.
Finally, objectives should be clearly defined when extending insurance to the poor. Uninsured risk means that poverty is perpetuated, with the possibility that a risk-induced poverty trap might occur. Extending insurance, as part of a credible comprehensive system of social protection, should allow the poor to sustain their assets and to enter into more profitable, riskier activities. In short, it would allow the poor to focus on long-term strategies lo get out of poverty.
Of critical importance are the credibility and sustainability of insurance provision as part of a broader social protection system. These cannot be iin tier estimated. Often, institutions in developing countries are not transparent or sustainable, and therefore well-intentioned measures may lack the credibility and public support to succeed. Credibility cannot be acquired easily, and governments face an uphill struggle in this regard. There is a key role for the government in the development and support of an appropriate regulatory and institutional framework for such programmes, and sustainable and transparent institutions to monitor these activities.
62 Social Protection for the Poor and Poorest
1
Stefan Dercon, Tessa Bold and Cesar Calvo 63
Notes
1. Thischapterisbasedonworkinitiallyfinancedbythelntcr-Am
Bank and by the World Institute of Development Economics Research (WIDER), ' We would like to thank Luis Tejerina for insightful comments. :
2. It should be pointed out here that the urban poor face different risks, not more (ot fewer) risks. For example, in an LSMS in Peru in 2001, 30.1 per cent of urban ■ households reported having suffered a shock. In the rural sample, it was 29.6 per -cent (Chacaltana, 2002).
3. In fact, Armendariz and Morduch (2000) argue that theoretical literature on micro-credit has exaggerated the focus on joint liability and dynamic incentives, and neglected the importance of the repayment schedule. This comment is bound to be relevant for insurance provision as well.
4. More information on this can be found in Dercon (2002), while the Social Risk Management Approach is discussed in Holzmann and Jorgensen (2000) and in World Bank (2000).
5. It is worth noting that most studies focus on poverty traps in the rural sector. Hence, the effect of risk exposure on urban investment decisions remains to he explored. This research is especially relevant for Latin America.
6. For a more detailed discussion, see Dercon (2004). For a theoretical discussion on poverty traps induced by risk, see Banerjee (2004).
7. This feature may also explain why the poor in Latin America may be unwilling to purchase some of the existing 'formal' insurance products available and instead prefer to rely on 'autarkic solutions', including self-insurance, since the lack of appropriately targeted and designed products would make existing products relatively too costly for the poor, possibly outweighing the benefits.
8. The survey by Brown and Churchill (2000) provides a number of examples.
References
Armendariz, B. and J. Morduch (2000) 'Microfinance beyond group lending', The
Economics of Transition, 8(2), 401-420. Banerjee, A. V. and A. F. Newman (1993) 'Occupational Choice and the Process of
Development', Journal of Political Economy, 101(2), 274-298. Banerjee, A. V. (2004) 'The Two Poverties', in S. Dercon (ed.), Insurance against poverty. ;
Oxford: Oxford University Press, pp. 59-75. Binswanger, H. (1986) 'Risk aversion, collateral requirements and the market for
credit and insurance in rural areas', in P. Hazell, C. Pomareda and A. Valdes (eds),
Crop Insurance for Agricultural Sectors: Issues and Experience, London: John Hopkins
University Press.
Bravcrman, A. and J. L. Guasch (1986) 'Rural Credit Markets and Institutions in
Developing Countries: Lessons for Policy Analysis from Practice and Modern
Theory', World Development, 14(10-11), 1253-1267. Brown, W. and C. Churchill (2000) Providing Insurance to Low-income Households.
Washington DC: USAID. Chacaltana, J. (2002) Social Funds and the Challenge of Social Protection for the
Poor in Latin America. (Conference on Social Protection), ADB-IADB. Dercon, S. (2002) 'Income Risks, Coping Strategies, and Safety Nets', World Bank
Research Observer, 17(2), 141-166. Dercon, S. (ed.) (2004) Insurance against Poverty. Oxford: Oxford University Press.
rjercon, S. and J. Hoddinott (2004) 'Health, Shocks and Poverty Persistence', in S. Dercon (ed.). Insurance against Poverty. Oxford: Oxford University Press, pp. 124-136.
Dercon, S., T. Bold and C. Calvo (2005) 'Insurance for the Poor', QEH Working Paper
125. Oxford: Department of International Development. Fafchamps, M. and S. Lund (2003) 'Risk-sharing networks in rural Philippines',
Journal of Development Economics, 31(1), 261-287. Gaviria, A. and C. Pages (1999) 'Patterns of Crime Victimization in Latin America - Research Department', Working Paper 408, Inter-American Development Bank. Holzmann, 11. and S. Jorgensen (2000) Social Protection as Social Risk Management:
A new conceptual framework for social protection and beyond, Social Protection
Discussion Paper 6. Washington, DC: The World Bank. Ibarra, H. (2003) 'Comments on Risk Managed Challenges in Rural Financial Markets'
by J. Skees. Mimeo. AGROASEMEX. Jacoby, Hanan G. (1994) 'Borrowing Constraints and Progress through School:
Evidence from Peru', The Review of Economics and Statistics, 76(1), 151-160. McCord, M. J., J. Isern and S. Hashemi (2001) Microinsurance: A Case Study Example of
theFull ServkeModelofMkroinsurance Provision, Self-Em ployed Women's Association.
MicroSave-Africa.
Morduch, J. (1995) 'Income Smoothing and Consumption Smoothing', Journal of
Economic Perspectives, 9 (Summer), 103-114. Morduch, J. (2004) 'Consumption Smoothing Across Space: Testing Theories of
Risk-Sharing in the ICRISAT Study Region of South India', in S. Dercon (ed.),
Insurance Against Poverty, Oxford: Oxford University Press, pp. 38-56. Plzarro, R. (2001) 'La vulnerabilidad social y sus desafios: una mirada desde America
Latina', Serie Estudws Estadistkos y Prospectivos, 6. ECLAC. Rosenzweig, M. and H. Binswanger (1993) 'Wealth, Weather Riskand the Composition
and Profitability of Agricultural Investments', Economic Journal, 103, 56-78. Rosenzweig, M. and K. Wolpin (1993) 'Credit Market Constraints, Consumption
Smoothing, and the Accumulation of Durable Production Assets in Low-Income
Countries: Investments in Bullocks in India', Journal of Political Economy, 1001(2),
223-244.
Rutherford, S. (1999) The Poor and Their Money. Institute for Development Policy and Management, University of Manchester.
Skees, J., P. Varangis, D. Larson and P. Siegel (2004) 'Can Financial Markets be Tapped to Help Poor People Cope with Weather Risks?', in S. Dercon (ed.), Insurance Against Poverty, Oxford: Oxford University Press, pp. 422-437.
Tesliuc, E. and K. Lindert (2002) Vulnerability: A Quantitative and Qualitative Assessment, Guatemala Poverty Assessment Program, The World Bank.
Townsend, R. M. (1995) 'Consumption insurance: An evaluation of risk bearing systems in low-income economies', Journal of Economic Perspectives, 9 (Summer), 83-102.
World Bank (2000) World Development Report 2000/1. Attacking Poverty,
Washington, DC: The World Bank. World Bank (2003) Guapa: The Guatemala Poverty Assessment. Washington, DC: The
World Bank.
Zaffaroni, C. (1999) 'Los recursos de las familias urbanas de bajos ingresos para enfrentar situaciones crfticas', in R. Kaztman (ed.), Activos y estructuras de oportuni-dades. Estudios sobre las rakes de la vulnerabilidad social en Uruguay, ECLAC and UNDP.
Rachel Sabates-W heeler and Stephen Devereux
4
Transformative Social Protection: The Currency of Social Justice
Rachel Sabates-Wheeler and Stephen Devereux
Introduction
Social protection emerged as a critical response to the 'safety nets' discourse of the late 1980s and early 1990s. In the World Development Report 1990, for example, safety nets were very much the third prong of the World Bank's three-pronged approach to 'attacking poverty' (World Bank, 1990), and were conceptualized as minimalist social assistance in countries too poor j { and too administratively weak to introduce comprehensive social welfare '•'   programmes. During the 1990s, as new thinking emerged in areas such as j 'rights-based approaches', 'sustainable livelihoods', and the multidimen-j sional nature of poverty and vulnerability, safety nets began to be criticized as residualist and paternalistic, and more sophisticated alternatives began to be proposed. As this agenda has evolved, the broader potential of social protection began to be recognized, and bigger claims are now being made for what social protection can and should strive to achieve. There are two interconnected strands in this response, both linked to a con-/ cern for long-term and sustainable poverty reduction. The first links risk man-/ agement explicitly with economic growth, and argues that reducing risk or protecting the poor against income and consumption variability will allow them to I invest and accumulate - a 'trampoline' out of poverty (World Bank, 2000). IDespite being promoted vigorously in international development publications, this link has not yet become a key component of anti-poverty programming in practice. In low-income countries, social protection continues to be perceived by governments and donors as comprising fiscally unsustainable 'consumption' y transfers to the economically inactive or unproductive poor, which diverts scarce public resources from 'productive' investment for economic growth, and 0 therefore deserves lower priority as a poverty reduction tool.
At the same time, most advocates of social protection do not make the second connection, which we argue is of fundamental importance to
long-term poverty reduction - namely, the positive relationship between^ livelihood security and enhanced autonomy or empowerment. While understandings of 'poverty' have moved to incorporate social dimensions^ of well-being together with rights-based approaches, social protection continues to be conceptualized by many development agencies mainly in i terms of public responses to livelihood shocks - the conventional eco-j nomic 'safety net' function. But this is 'economic protection', not 'social! protection', and it is hardly socially transformative. Largely missing from the World Bank's 'social risk management' framework, for example, is a concern for equity and social rights. We argue that an appreciation of this second linkage can help to create the policy conditions for a virtuous cycle of pro-poor growth, governance systems that are accountable and responsive to poorer as well as wealthier citizens, and an approach to development that is grounded in concerns for social justice.
In an attempt to challenge the negative perceptions and narrow preconceptions that still surround social protection, this chapter addresses both conceptual and policy issues in the social protection literature. In the first part of this chapter we locate the conceptual origins of social protection as a i response to risk and vulnerability, and argue for a broader conceptualization of vulnerability based on an appreciation of structural inequalities. Attempting to address structural vulnerabilities (together with other forms) requires taking a political approach to social protection, focusing on rights, duties, democracy and advocacy. We describe what we mean by the 'transformative' potential of social protection. The following section considers 1 'transformative social protection' in practice, by discussing several measures S that highlight the economic growth and social transformation functions and linkages of social protection, drawing on experiences in mainstreaming social protection in Uganda's poverty reduction strategy. The chapter concludes by reasserting the case for social protection as supporting social as well as economic goals of development.
Social protection as a response to risk and vulnerability
People across the world face a wide range of risks and hazards that impact on their livelihoods, both directly and indirectly. In order to minimize the likelihood of downward livelihood trajectories, and institute effective and sustainable social protection strategies, we must understand how house- ', holds, communities and countries attempt to manage potential impacts i from hazards as well as the hazards themselves.
To be a useful concept, vulnerability must be defined in relation to some other phenomenon, such as poverty, malnutrition, exclusion or neglect. Vulnerability is thus a multidimensional concept and corresponds to the complexity of the phenomenon it is defined against. For example, if we are
64
66  Social Protection for the Poor and Poorest
Rachel Sabates-Wheeler and Stephen Devereux 67
interested in measuring vulnerability to poverty and we know that poverty " is multidimensional, then vulnerability to one aspect of poverty (say, malnutrition) may not mean vulnerability to another aspect (say, lack of access to education). A common way of conceptualizing vulnerability is to view it as a product j | of two components: exposure to a hazard (a shock or process); and resil- I ! ience, or the ability to manage the hazard (Chambers, 1989; Bankoff et a}., j / 2004). Understanding vulnerability in two-dimensional space is importai it j as it illustrates the very different policy responses that need to be considered | in relation to what constitutes the vulnerability of any one person, house- J hold, community or 'vulnerable group'. It is particularly useful for acu situations requiring an immediate response. That is, at any one time it is possible to construct a static vulnerability profile that indicates whether the hazard or the ability to cope is the main determinant of vulnerability. Policies appropriate to the composite nature of the vulnerability can then be designed. However, to understand vulnerability fully it is not enough simply to take a one-period view. Vulnerability needs to be forward-looking, as it makes a prediction about future poverty (or other outcomes). Vulnerability does not simply refer to those who are likely to become poor in the future as a result of an unexpected shock, but also to those |yho will remain poor, those who might fall deeper into poverty and those who may move in and out of poverty because of predicable fluctuations such as seasonality (Dercon, 2001). This disaggregation is important, as the appropriate policy responses are very different for each distinct group.
An understanding of vulnerability is further complicated by the notion of 'ability to manage'. How do we measure this, and what are the implications of different measurement approaches? If we do not unpack what 'ability to cope' means then we cannot determine whether we are dealing with transitory or chronic poverty, or vulnerabilities related to structural inequalities and lack of access to rights and opportunities. 'Ability to manage' is often proxied by income or consumption poverty measures. For example, if a household falls below a specified poverty line it may be considered unable to cope. Poverty-gap measures are also used to estimate the severity and depth of poverty, and these measures could also be used to infer (in)ability to cope. More recent work focuses on asset profiles, or asset bundles, as a way of understanding poverty. The problem with using any one of these measurements for looking at vulnerability is that each measurement will only provide a partial story. For example, using a consumption measurement, we may conclude that in a certain environment households are able to smooth consumption using a range of risk management strategies. However, it is not possible to see from this what households have been doing with their income or assets in order to smooth consumption. In other words, we are unable to understand the dynamic relationships between consumption and income smoothing, or income smoothing and asset depletion.
These income, consumption and asset-based understandings of vulnerability underpin the majority of government and donor approaches to vulnerability. I'or this reason we see many agencies taking an instrumentalist approach to social protection policies, as a collection of measures to manage risk and thus improve or protect livelihoods by stabilizing income and consumption or building up assets. Consider these definitions from the International Labour Office (ILO) and the World Bank: '[Social protection is] the provision of benefits to households and individuals through public or collective arrangements to protect against low or declining living standards' (van Ginneken, 1999); 'Social protection interventions assist individuals, households, and communities to I setter manage the income risks that leave people vulnerable' (World Bank, 2004).1
While we do not dispute the fact that income, consumption and assets are crucial in helping to overcome poverty and minimize livelihood shocks, we would argue that 'ability to manage' is rather more complex than a simple focus on household income and asset portfolios. It is instead a complex function of existing behaviour, reflected in livelihood profiles that themselves represent long-term or structural adaptation to predictable shocks and stresses; crisis response behaviour (such as the ability to rely on formal and informal insurance and networks in times of crisis); and external (policy) responses to a predicted and actual crisis. Provision of consumption, income and asset insurance is only a partial response to vulnerability. An expanded view of social protection must incorporate responses to both chronic and structural vulnerability.
Kri onceptualizing vulnerability
Vulnerability can be conceptualized in a variety of ways, depending criti- \ cally on the unit of analysis and the source of risk. Within the Social Risk 1 Management framework of the World Bank (as in most literature on social protection) vulnerability is attributed to the characteristics of a person or group, an event affecting a person or group, or a point in a person's life-cycle. For example, in any given context, people living with disabilities can he characterized as being more or less vulnerable than people living without disabilities. This type of analysis tends to classify vulnerability according to , a range of risks or shocks that affect one or more of a variety of livelihood \ assets (World Bank, 2000: 136-8). This is reflected in the range of policy! instruments proposed, such as reception centres for orphans, shelters for women suffering domestic abuse, disability aids for farmers living with dis-abililies, foodgrain warehouses, and various social assistance programmes (World Bank, 2000: 141). However, if, rather than focusing on risk as an exogenously given factor to be managed, vulnerability is conceptualized as emerging from and embedded in the socio-political context, then our atten-; tion would no longer be focused on how to design a policy so that various
68  Social Protection for the Poor and Poorest
Rachel Sabates-Wheeier and Stephen Devereux 69
groups face less risk in a given context, but on how to transform this context to minimize risk for a range of vulnerable groups.2
Focusing on the vulnerabilities of the individuals or groups in any given space or at any given moment to enable them to better access and make use of that space is certainly important as a short-run agenda. However, if we re-focus our attention to include also the second conceptualization of vulnerability as presented above, space, context and time no longer bind, but instead become the crucial point of our understanding of constructions of vulnerability. The question no longer becomes how do we design a policy so that various groups face less risk in given spaces, but, how did this space, or context, emerge? Whose interests were served in the creation of the space and whose interests are served in maintaining the status quo! A focus on space and time necessarily leads to contextual socio-political analyses of vulnerability. (Sabales-Wheeler and Waite, 2003)
The dominant policy agenda around social protection is concerned almost exclusively with measures and programmes that stabilize expectations ol risk, without affecting the fundamental causes of Vulnerability that are embedded in social and political relations at all levels. For example, liic social risk management (SRM) framework mainly addresses economic risks to household incomes and assets (World Bank, 2000: 138). Absent from t his framework are 'social risks' that also contribute to poverty and the construction of vulnerability. This is evidenced by the way in which social inclusion, social cohesion and social stability are treated as being positive externalil ics of well-designed 'social risk management' interventions. Social risks may be categorized as 'structural' or 'contingent'. Structural risk refers to situations where groups or individuals are marginalized or discriminated against, and by its nature has longer-term implications for poverty and vulnerability than contingent risk, which is a function of environmental or economic factors, such as an earthquake, or hyper-inflation. Because the SRM framework is largely focused on income variability, with other (especially social) dimensions of vulnerability effectively being overlooked, we argue that I lie SRM approach does not incorporate a comprehensive understanding of vulnerability, and is therefore limited in its scope and purpose of social protection provisions.
Related to a limited conceptualization of vulnerability is the lack of attention paid to chronic poverty in standard approaches to social protection. The 'chronic poor' include people who have never recovered from a severe shock, such as a disabling illness or loss of assets. In their analysis of chronic poverty and social protection. Barrientos and Shepherd (2003: 7) state thai: 'Although risk and vulnerability are key factors in explaining the descenl into poverty, it is not clear... how important they are in maintaining people
|M poverty, transmitting poverty from one generation to the next, and in preventing the interruption of poverty/ Importantly, Barrientos and Shepherd (2003: 3) highlight structural reasons related to 'social, political and economic structures and relationships, and processes of exclusion and adverse incorporation', that prevent some of the chronic poor from benefiting from development policies and market changes. The chronic poor 'have fewer options, less freedom to take up available options, and so remain stuck in patterns of life which give them low returns to whatever few assets they have maintained' (Hulme et al, 2001: 8). Social, political and economic siiuciures are typically the defining characteristics of livelihood risk, with the possible exception of some natural disasters - though even in these cases, the contribution of socio-political factors has persistently been underappreciated (Bankoff etal, 2004).
Relocating an understanding of 'vulnerability' in socio-political space necessarily conjures up linkages to large literatures on social exclusion, rights-based approaches, citizenship and power (for example, Kabeer, 2002, 2005; Gaventa, 2004; Nyamu-Musembi and Cornwall, 2004). These literatures will not be reviewed here. However, in accordance with the general thrust of the literature, we would argue that, like the rights-based approach or an agenda for inclusive citizenship, a transformative approach holds little ruca ii ing if it is unable to achieve a positive change in power relations among various stakeholders - development actors, government agencies, differentiated socio-economic groups, different household members. Thus a vision of transformative social protection, 'must be interrogated for the extent to which it enables those whose lives are affected to articulate their priorities and claim genuine accountability' from different implementing and 'provisioning' stakeholders (Nyamu-Musembi and Cornwall, 2004).
Introducing the transformative i-lement to social protection
If a person's needs for social protection are defined in a narrow 'safety net' \ sense, as mechanisms for smoothing consumption in response to declining or; fluctuating incomes, then the focus of interventions will logically be on targeted income or consumption transfers to affected individuals. In our view, ttie range of social protection interventions should extend well beyond social transfers - and the resources transferred should be broader than cash or food, include redistribution of assets that will reduce dependency on handouts and enable some of the poor to achieve sustainable livelihoods. Targeted social j"' t lansfers provide 'economic protection' in response to economic risks and live- \i I i hood vulnerability. Other forms of 'social protection' would address distinct & problems of 'social vulnerability', not necessarily through resource transfers, hut by delivering appropriate social services and implementing measures to modify or regulate behaviour towards socially vulnerable groups.
70  Social Protection for the Poor and Poorest
Rachel Sahates-Wheeler and Stephen Devereux 71
Strategies to deal with problems of social vulnerability require a transformative element, where 'transformative' refers to the need to pursue policies that integrate individuals equally into society, allowing everyone to I take advantage of the benefits of growth and enabling excluded or mar; ; alized groups to claim their rights. For example, support to trade unions may enable socio-economically marginalized groups to claim their rights i0 decent working conditions; facilitation and creation of spaces for delibcra-/ tive democratic processes can increase citizen participation; sensitizal ion I and awareness-raising campaigns can transform public attitudes and behnv-' iour; and changes to the regulatory framework could protect vulnerable or minority groups against discrimination and abuse.
Another sphere where transformative social protection policies may be needed is the intra-household division of resource ownership, access and use. For example, many difficulties involved in the provision of social ] tection for women relate to socio-cultural values that leave women in pai-ticularly vulnerable positions. Clearly, social protection instruments designed for many categories of women must include a substantial 'transformative' element, in the sense that power relations between men and women become more balanced. Appropriate legislation is necessary, but I his goes only a small way towards changing socio-cultural values. Efforts could focus on educating men and women about their rights, and how to access them. Other political and institutional constraints facing women relate lo lack of access to the legal system; cultural resistance to changes in gender relations; and commonly held beliefs about women's role in land management and property ownership.
Bearing the above points in mind, we have devised our own conceptual and operational definitions of social protection. Om- conceptual definition incorporates but goes beyond transfer-based responses to economic risk and vulnerability:
Social protection describes all initiatives that transfer income or assets lo the poor, protect the vulnerable against livelihood risks, and enhance Hie social status and rights of the marginalized; with the overall objectives of extending the benefits of economic growth and reducing the economic or social vulnerability of poor, vulnerable and marginalized groups.
Our working definition adapts terminology first introduced by Guhan (1994), and elaborates on the mechanisms that deliver social protection. Social protection includes four categories of instruments: 'provision' measi which provide relief from deprivation; preventive measures, which attempt 1 o prevent deprivation; promotive measures, which aim to enhance incomes a i id capabilities; and transformative measures, which seek to address concerns of social justice and exclusion.
'Provision' measures provide relief from deprivation. Provision measures are narrowly targeted safety net measures in the conventional sense - they aim to provide relief from poverty and deprivation to the extent that promotional
and preventive measures have failed to do so. Provision measures include social assistance for the chronic poor, especially those who are unable to work arid earn their livelihood. This equates closely to mainstream 'social welfare'. Social assistance programmes include targeted resource transfers - disability benefit, single-parent grants, and social pensions that are financed publicly (out of t he tax base, with donor support, or through NGOs). Other provision measures can be classified as social services for poor individuals and groups who i ieed special care, including orphanages and reception centres for abandoned children, and the abolition of education and health charges to extend access to basic services to the very poor.
Preventive measures seek to prevent deprivation. Preventive measures deal directly with poverty alleviation. They include social insurance for 'economically vulnerable groups' - people who have fallen, or might fall, into pov-crtv, and may need support to help them manage their livelihood shocks. This is similar to 'social safety nets'. Social insurance programmes refer to formalized systems of pensions, health insurance, maternity benefit and unemployment benefits, often with tripartite financing between employers, employees and the state. They also include informal mechanisms, such as savings clubs and funeral societies. Strategies of risk diversification - such as crop or income diversification - can also be considered as preventive measures.
Promotive measures aim to enhance real incomes and capabilities, achieved through a range of livelihood-enhancing programmes targeted at both households and individuals, such as microfinance and school feeding. The inclusion of promotive measures as a category here is open to the criticism l hat it takes social protection too far beyond its original conceptualization. However, the intention is not to broaden the scope to include (potentially) all development initiatives, but to focus on promotive measures that have income stabilization at least as one objective. Examples include micro-credit invested in small enterprises in order to fulfil income stabilizing and consumption smoothing functions, or public works projects that transfer food rations or cash wages while simultaneously building economic infrastructure such as roads or irrigation.
transformative measures seek to address concerns of social justice and exclusion, such as the exploitation of workers, or discrimination against clhuic minorities. Transformative interventions include collective action for workers' rights, changes to the regulatory framework to protect 'socially vulnerable groups' (for example, people with disabilities, or victims of domestic violence) against abuse, as well as sensitization campaigns (such as the 'HIV/ AI1XS Anti-Stigma Campaign' discussed later in this chapter) and advocacy lo enhance social equity.
These categories may, of course, overlap, and many social protection interventions aim to achieve more than one objective. Public works projects, school feeding schemes and micro-credit programmes all strive towards
72  Social Protection for the Poor ami Poorest
Rachel Sabates-Wheeler and Stephen Devereux 73
promoting incomes in the long term as well as preventing deprivation in the short term. Similarly, a 'transformative' measure such as eradicating labour market discrimination against HIV-positive people is not only a vie tory for social justice, it also enhances the employment prospects of peopli living with HIV and AIDS.
Figure 4.1 illustrates the relationship between these categories and presents our suggested conceptual framework for thinking about social protection. Reading from the bottom, 'provision' policies are essentially old-style 'safety nets'. However, if designed well these instruments can have positive effects on the prevention of deprivation, livelihood promotion and even social transformation. The solid black lines in Figure 4.1 indicate a strong and direct relationship. For example, 'preventive' interventions, such as crop diversification to reduce agricultural risk, may also have 'promotive' outcomes in the sense that a wider crop portfolio might lead to a competitive market advantage. In fact, most preventive mechanisms could be argued as having promotive effects, in the sense that risk reduction enables people to take advantage of opportunities that they would otherwise have been unable to do.
The top half of the diagram in Figure 4.1 corresponds with instruments and policies that facilitate movements out of poverty, or 'springboards'. Broadly speaking, the left-hand side of the diagram corresponds to social protection interventions that have economic outcomes and direct growth effects, whereas the right-hand side represents social outcomes - the 'transformative' aspect of social protection. At the top of the figure the economic and social dimensions come together and are interrelated. That is, by pursuing activities that overcome structural inequalities and injustice, people are better able to engage in both society and the economy, which will have positive spin-offs not only for their livelihoods but also for economic growth. The dashed line connecting 'promotion' and 'transformation' illustrates both the interconnectedness of economic and social vulnerabilities,; and the potential for interventions in either sphere to reduce both sources of vulnerability. This is especially true of vulnerabilities that are structural and chronic, rather than contingent and acute.
The dashed lines indicate a less obvious or weaker relationship. For example, some preventive mechanisms can be transformative, and vice versa, but this relationship is neither strong nor inevitable. One example is microfi-nance schemes that provide both social insurance and economic opportunities, and often have positive knock-on effects by empowering individuals within their families, and households within their communities. Another weak relationship is illustrated by the dashed line from 'provision' to 'promotion', highlighting the possibility that certain safety net measures may build capabilities or assets, thereby enabling beneficiaries to grasp opportunities that otherwise would be denied to them. (In the next section of this chapter we draw on the example of school feeding schemes to illustrate this
Promotion
Economic opportunities \ ~ :---
Springboards
Minimum wage legis/ation
Labour market regulations
Transformation
Social justice
Crop diversification Migration Property rights Microcredit
Agricultural J extension ' Microfinance ' for women I
ill
School feeding Public works
Prevention
Insurance and diversification mechanisms (social security)
Safety nets
Provision
Social assistance and coping strategies (social assistance
formal and non-state)
Figure 4,1  A conceptual framework for social protection
linkage.) Similarly, some social protection instruments, such as minimum wage legislation (also discussed below), can be both promotive and transformative. Paying workers a fair wage enhances their incomes and capabilities, while the very process of bargaining with employers to raise and enforce the minimum wage - through trade unions or public campaigns - can be politically empowering, especially when supported by government.
Finally, the very thick dashed line linking provision and transformation indicates a potential negative relationship between the humanitarian objective of delivering social assistance, and social objectives such as human dignity and autonomy. Some social protection measures can have the unintended consequence of reinforcing established power hierarchies, or of being stigmatizing and exacerbating social polarization and exclusion. Examples include certain targeting mechanisms that are applied to social assistance programmes for 'vulnerable groups' (for example, singling out 'AIDS orphans' from other vulnerable children, or requiring adults to declare their HIV status), or paying participants carrying out public works with food
74  Social Protection for the Poor and Poorest
Rachel Sabates-Wheeler and Stephen Devereux 75
rations rather than cash wages. One advantage of drawing attention to the | social dimensions is that this confirms the need for thinking on social pro- ; tection to move beyond old-style safety nets, and to ensure that social pro- \ tection interventions have neutral or positive - rather than negative - social \ impacts. | It is important to emphasize that our expanded definition of social pro- \ tection does not extend to all policy measures that promote livelihoods and :' economic growth. Instead, we highlight the potential of certain social protection measures to contribute to economic growth and productivity as well _ as to social equity, either through achieving both objectives simultaneously -or through linkages with other interventions. A good example (discussed below) is school feeding schemes - a social protection intervention that stabilizes food consumption but also enhances access to education for poor \ and socially excluded children, thus building their human capital and improving their lifetime earning potential, through linkages with educa- ' tion facilities.
Transformative social protection in practice
This section discusses three interventions that reflecj a broader conceptu- * alization of what social protection is, or what it can achieve. The selected ? case studies each relate to different parts of our conceptual diagram (Figure 4.1). '4 The first intervention (school feeding) relates to the left-hand side of the -diagram, and illustrates potential linkages between safety nets or social assistance programmes and economic growth. The second intervention (anti-discrimination campaigns) relates to the right-hand side of the diagram, and identifies how social protection can be mobilized for social jus- : tice ends. The third intervention (labour rights and minimum wages) relates -to the top of the diagram, and illustrates how social protection can achieve growth-equity linkages. The actual case studies discussed in this section 1 draw on work done by the authors with the Government of Uganda's Social Protection Task Force during 2002-2003 (Devereux and Sabates-Wheeler, 2003), as part of an effort by the Ministry of Gender, Labour and Social -Development to 'mainstream' social protection across all sectors, in the process of revising the national Poverty Eradication Action Plan (PEAP).
Safety-net - growth linkages: school feeding
One example of a social protection intervention that can contribute to longer-term poverty reduction goals is school feeding schemes that provide learners with free meals at school. Because school meals serve two func- ; tions - providing an immediate consumption transfer to children who are often malnourished, and encouraging children from poor households to attend school even during difficult times - school feeding schemes can be _
characterized as serving both protective and promotive social protection objectives. However, making this case requires a demonstration that education generates higher incomes for school-leavers, that school feeding improves educational outcomes, and that school feeding is pro-poor.
Education for all is not only a basic right, it is also an investment in the future. In Uganda, the right to education is enshrined in the Constitution of 1995, and the case for education as a driver of poverty reduction has been proved empirically (Appleton, 2001; Lawson et ah, 2003). Poverty and lack of education are closely correlated, while returns to education are significant, and higher for primary than for secondary or tertiary education. Educated Ugandans benefited most from Uganda's strong recent growth and its poverty reduction performance. While the national poverty headcount fell by 39 per cent during the 1990s, it fell by only 28 per cent for Ugandans in households with uneducated heads (Appleton, 2001: 4). A large part of the explanation is that education enhances access to non-agricultural employment, and Uganda's agriculture sector grew more slowly than other sectors in the 1990s.
School feeding is controversial, and empirical evidence on its positive and negative impacts is limited and inconclusive. However, several case studies have confirmed that provision of free meals at school increases enrolment and attendance rates, and reduces drop-out rates, and that these effects are highest among the poorest families. School feeding even has 'safety net' effects: in difficult times such as during a drought or conflict, providing school meals (or take-home rations - 'food-for-education') encourages parents to leave their children in school, rather than withdrawing them to save costs and to assist with the search for food. Some school feeding and food-for-education projects have been found to reduce gender gaps in education access and attainment, and even to improve learners' concentration in class and their performance in examinations (Bennett, 2003; World Food Programme, 2006).
Uganda has implemented universal primary education (UPE) since 1997. Under UPE, fees were waived for state primary schools, which resulted in a doubling of primary enrolment. Given this success in terms of improved access to education, the case for introducing a school feeding scheme on top of UPE might seem rather weak. In fact, President Museveni declared in 2003 that there should be no free school meals in Uganda: the government's responsibility is to provide teachers, classrooms and textbooks; and it is the responsibility of parents to send their children to school and to ensure that they are adequately fed. On the other hand, large numbers of Ugandan children remain outside the formal education system, including several groups identified by the Ministry of Education as 'disadvantaged children': children who have never enrolled in school; those who have dropped out; orphans; refugees; geographically and culturally isolated children; street children; children who must work; children with disabilities; and abused
76  Social Protection for the Poor and Poorest
Rachel Sabates-Wheeler and Stephen Devereux 77
children (MoES, 2002). UPE has not reached these vulnerable categories of I
children, partly because fees are not the only education cost facing parents 1
and partly because the poorest families see no immediate benefit in educat- I
ing their children or the orphans in their care. In these circumstances, pro- |
viding a daily meal for a child from a food-insecure household can provide i the incentive needed, and it enhances equity of access to education.
Taking these considerations into account, the Uganda Social Protection \ Study Team argued that targeted school feeding should be considered for specific vulnerable groups (such as orphans) and in areas where poverty is
high and is recognized as a barrier to participation in education. This would \
include geographically inaccessible communities, areas with high conccn- \
trations of displaced or resettled people, and informal urban settlements j
where street children and destitute households are concentrated (Deveieux j
and Sabates-Wheeler, 2003). This recommendation supported the seen ml I
Participatory Poverty Appraisal (UPPAP2), which recommended providing | orphans with meals at school, to improve both their nutritional and educational status (MFPED, 2002: 145).
Campaigning to transform: anti-discrimination campaigns j,
One arena of social protection that attends directly to the 'social' needs of socially vulnerable groups is campaigning against various forms of discrimination - whether on the basis of ethnicity, gender, religion, disability or sexual orientation ~ as part of a broader emerging agenda around realizing economic, social and cultural rights. A good example is the Convenl ion on the Elimination of All Forms of Discrimination Against Women (CEDAW), a United Nations campaign during the 1990s that raised awareness aboul gender discrimination in countries across the world.
Recently, concerns have been raised about the nature of formal and non-formal responses to HIV/AIDS, specifically its 'social impacts' in terms of the burden of caring and attitudes towards people living with AIDS (PLWA). In Uganda, which was the epicentre of HIV/AIDS in Africa during the 1980s, the pandemic has had a terrible impact, causing over 800,000 estimated \/ deaths and creating over two million single - or double-parent orphans -|  almost one child in five. Approximately 25 per cent of Ugandan households ! now include one or more orphans. Many of these are elderly or widow-P 'headed households, under traditional arrangements of caring for orphans through extended family and kinship relationships, which tends to place , the burden of care on women. These informal social protection mechanisms I are coming under severe strain, as the costs of caring for the sick and provid-/ ing for the basic needs of large numbers of dependants is exceeding the limited resources of poor Ugandan families. Orphans who are not located in extended families or supportive communities, and are not taken into
reception, centres, often become street children, surviving by begging or petty crime.
[n the late 1990s, the Government of Uganda launched a 'crusade' against HIV/AIDS, with donor support, under the umbrella of a five-year National] Strategic Framework for reducing prevalence rates and strengthening! national response capacity (UNDP, 2000: 16). HIV prevalence did in fact decline between 1996 and 2000, from 10 per cent to 8.3 per cent, largely because of changes in sexual behaviour among the Ugandan population (republic of Uganda, 2001: 104). In this period, the government's response lo AIDS focused on creating awareness about the causes of HIV transmission, encouraging behavioural change - including challenging cultural prat tices such as the remarrying of widows to a brother of the deceased -and promoting the use and availability of condoms.
A parallel set of interventions has also emerged in many high HIV-, prevalence African countries since the 1980s, focusing more on the consequences of the disease than its prevention. Recognizing the costs that HIV/, AIDS imposes on the livelihoods and coping capacity of poor households, these initiatives aim to provide various forms of support to PLWA, 'AIDS orphans' and carers. These initiatives are enormously important to the people they assist, but they are also patchy and limited in impact, with incomplete coverage across the country and variable provision of support - from food and cash transfers to nutritional advice for orphan carers, to reception centres for orphans without support, to vocational training and microcredit for older orphans.
Most of these measures can be characterized as standard social assistance to meet the subsistence and livelihood needs of PLWA and orphans. These provisions address the reality that HIV/AIDS has deepened poverty in \ Ugandan households, as affected families lose productive labour, sell off | assets to care for terminally ill members, and pay burial costs after a death, j In short, these measures address the economic costs of HIV and AIDS at the household level. Until recently, relatively little attention has been given to the 'social costs' of AIDS, such as the problems faced by PLWA in terms of securing or retaining employment, the stigma and social exclusion faced by bereaved relatives, the vulnerability of widows to losing property and being remarried against their will, and the risks faced by orphans of being abused or neglected.
In a special session on AIDS in June 2001, the United Nations General Assembly passed a declaration of commitment to fight AIDS, which included a global target for all UN member states to enact or strengthen antidiscrimination protection for people living with HIV/AIDS. In 2002, the ^_ International Federation of Red Cross and Red Crescent societies (IFRC) ■/ launched its own 'HIV/AIDS Anti-Stigma and Discrimination Campaign', in 1 support of the United Nations declaration. Under this global initiative, the 0 Uganda Red Cross Society launched a two-year campaign in July 2002,
78  Social Protection for the Poor and Poorest
Rachel Sabates-Wheeier and Stephen Devereux 79
aimed at eliminating stigma and discrimination against Ugandans living [ with HIV/AIDS, in accordance with the commitment of the IFRC to 'making a difference in the lives of the vulnerable' (URCS, 2002). A key message of the campaign was: 'AIDS is a disease and not a disgrace'. The intention was to challenge attitudes and practices that discriminate against PLWA, such as [employers insisting on job applicants taking pre-employment HIV tests, and /rejecting all HIV-positive applicants. The overarching objective of the campaign was: 'to contribute to the change in perceptions, attitudes, policies -arid behaviour towards PLWA, in order to ensure that those people who are already HIV+ or have AIDS are able to receive the appropriate care, have access to affordable drugs and can live full and useful lives within their communities' (URCS, 2002).
This campaign can be characterized as a 'transformative social protection' measure. Unlike conventional social safety net or 'protective social protection' measures, anti-discrimination measures address the social rather than economic needs of a socially vulnerable group. Like minimum wages, this is an affordable intervention, since it has negligible implications for public spending or donor budgets. Anti-discrimination campaigns also have the potential to be 'protective' and 'promotive' as well as- 'transformative'. It is well documented that discrimination reduces the livelihood opportunities of affected groups. Affirmative action campaigns to promote disadvantaged groups, such as black South Africans or low-caste Indians, recognize that social exclusion carries economic costs, and attempt to intervene in the labour market to correct for this socio-economic discrimination. Similarly, the anti-stigma campaign in Uganda attempts to ensure that hostility towards people living with HIV/AIDS does not undermine their ability to earn a living. *
Growth - equity linkages: labour rights and minimum wages
Unskilled and semi-skilled workers are acutely vulnerable to exploitation by employers, especially in poor countries with widespread chronic poverty, limited employment opportunities and low levels of worker mobilization and unionization. In such conditions, it is often argued that low-paid workers need legal protection, through the introduction and enforcement of labour legislation that secures decent working conditions (reasonable working hours, maternity leave, job security and so on) and minimum wages. Labour market interventions are an attractive social protection policy arena, for several reasons: they reach people who live close to the poverty line, they cost government or donors very little to implement (since the costs are borne by private-sector employers), they have direct poverty-reducing
impacts, and they are 'transformative' in that they empower categories of porkers who are typically not unionized and have little bargaining power with their employers.
Despite these benefits, labour market interventions (especially mandated minimum wages) are controversial. The debate is often characterized as one between opponents of 'labour market rigidities' and supporters of 'fair labour standards'. Economists argue that legislation that raises wage rates and other employment costs above the market-clearing equilibrium will destroy jobs among the very group who were intended to gain. In a competitive labour market, employers faced with rising labour bills will respond by laying off some workers and shifting towards 'actualized labour' arrangements (daily or part-time contracts to which employment legislation does not apply) or mechanized forms of production. Those workers whose relative employment costs rise most - the lowest paid unskilled workers - are most vulnerable to these 'disemployment' and 'casualization' effects. A pragmatic argument against minimum wages is that many poor workers fall outside waged employment altogether, in 'uncovered' sectors such as subsistence agriculture and informal (self)-employment. Even if minimum wages are legislated for, compliance is difficult to enforce in these sectors, so the poorest segments of the working population are rarely reached (Alatas and Cameron, 2003).
The theoretical case against labour market interventions has been challenged by the '"new economics" of the minimum wage' (Lustig and McLeod, 1996), which argues that the critique is built on three 'textbook' assumptions - that labour markets are perfectly competitive, that workers are perfectly substitutable, and that perfect information exists in the economy - which do not hold in the real world. For example, where labour markets are monopsonistic - a realistic scenario in many developing countries - a higher wage rate can in fact increase aggregate employment. The reason is that monopsony employers pay workers less than their marginal value product (this is known as 'Pigovian exploitation') and hire to a point where the marginal cost of labour exceeds labour supply; hence, forcing the monopsonist to pay a competitive wage raises the supply of labour towards a market-clearing equilibrium (Jones, 1997: 3). Put another way, employers who are extracting excess profits by under-remunerating labour can afford to raise wage rates without laying off workers, and might even increase profitability by hiring more workers at these higher rates.
Empirical evidence can be found to support both sides of the debate. A cross-country regression analysis covering twenty-two countries found statistically significant evidence that minimum wage legislation can contribute to poverty reduction: 'minimum wages and poverty are inversely related: i.e., an increase (decline) in real minimum wages is accompanied by a fall (rise) in poverty' (Lustig and McLeod, 1996: 1). Another review of several
80  Social Protection for the Poor and Poorest
1990s studies found 'little, if any, disemployment effect' from minimum i wage policies (Saget, 2001: 21). Conversely, a study of the impact of a mandated " ■ increase in Ghana's minimum wage in 1991 found that it had had the unintended effect of reducing manufacturing employment by 5-6 per cent displacing workers who lost their jobs into the informal sector, and depressing informal wages (Jones, 1997). In South Africa, a simulation exert ise found that legislating wage increases for domestic workers and farm workers would result in significant job losses in both sectors (Bhorat, 2000). \
In Uganda, the debate about minimum wages is ongoing, and unresolved \ (Devereux, 2005). Although a Minimum Wages Board was established in \ 1935, the statutory minimum wage has not been raised since 1984, and its : value has collapsed to such an extent that it has no real effect. In some sec- \ tors, such as the agricultural estates, wages and conditions of employment are so bad that workers are trapped in poverty and are at the mercy of employers who, inter alia, lay off pregnant women, who have to reapply for their jobs after giving birth; continuously rotate casual workers so as to -, avoid employing them on a permanent basis; and force their workers to ~ work illegally long hours (MFPED, 2002). In the absence of effective trade \ unions, the Uganda Human Rights Commission has consistently called for government intervention to protect low-paid worker|,against exploitation and to raise the minimum wage to a 'living' level (UHRC, 1999).
During the process of revising Uganda's PEAP, the Social Protection Task Force recommended that the Government of Uganda intervene to provide social protection for vulnerable employees, not only by instituting a realistic minimum wage but also by insisting that working conditions are improved - that the laws on maximum working hours are not violated, that workers are provided with decent accommodation, that women are not dismissed for becoming pregnant, and that some form of job security is introduced for 'permanent casuals'. More effective labour inspections and enforcement mechanisms would be required to ensure compliance - the number of Labour Inspectors was cut drastically under Uganda's civil service reform programme - and adequate social assistance arrangements would need to be put in place to provide a transitional safety net for any adversely affected families (Devereux and Sabates-Wheeler, 2003: 35).
The PEAP Secretariat was initially sceptical about these proposals, arguing that 'it is difficult to design public interventions that improve the well-being of workers without simultaneously increasing the costs of employers and hence reducing employment'. However, the Secretariat did agree that labour market interventions such as a minimum wage might be appropriate in contexts of market failure. By the time the draft PEAP was published in March 2004, the Government of Uganda appeared to have accepted the case for introducing binding minimum wages, at least for specific sectors.
Rachel Sabates-Wheeler and Stephen Devereux 81
International evidence shows that blanket minimum wages tend to reduce employment opportunities. However, there is a case for some regulation of wages in sectors where employers have great market power, in which case the effects of a minimum wage on employment may not be harmful. Government, in consultation with the private sector will consider whether some regulation of wages or working conditions may be appropriate in certain sectors. (Government of Uganda, 2004: 103)
Conclusion
By arguing in this chapter for an approach to social protection that emphasizes social justice we are not arguing against the important 'safety net' role that social protection has conventionally played, in terms of safeguarding lives and livelihoods in contexts of chronic and acute economic risk and vulnerability. Instead, we are focusing attention on the relatively neglected area of social risk and vulnerability, and building a case for a stronger role for social protection in terms of empowering the poor and transforming the conditions in which they struggle to construct viable livelihoods.
In this chapter we have conceptualized vulnerability in social as well as economic terms, and argued for social protection as a comprehensive package of interventions that addresses both aspects. Although the socially vulnerable - orphans, people living with HIV/AIDS, ethnic minorities -often need income and consumption support, social protection, properly conceptualized, is not just 'economic protection'. Poverty and vulnerability are about social as well as economic deprivation, and our argument is for an elaborated understanding of social protection that concerns itself directly with addressing 'social risk' and non-economic vulnerability, such as social exclusion, discrimination and violations of minority rights.
Apart from establishing a more positive and proactive role for social protection that extends its scope beyond its roots in residualist and often stigmatizing social safety nets, there are other reasons for supporting the 'transformative' component of social protection. First, 'transformative social protection' is more fiscally affordable than 'economic social protection', which is unpopular with economists and governments because it implies expensive (and often recurrent) transfers of public resources to large numbers of people who are regarded as having low or zero productivity (McDonald eta!., 1999). In developing countries, policy-makers face binding fiscal constraints that limit their public spending choices, so the identification of low-cost interventions that can improve the livelihoods of the poor significantly - such as mandated minimum wages, whose cost is borne by employers rather than the state, or anti-discrimination campaigns that cost next to nothing - is doubly attractive.
Second, we have identified powerful synergies between the 'economic' (provision, prevention, promotion) and 'social' (transformation) functions