Addressing Vulnerability and Enhancing Productivity in South Asia Naila Kabeer∗

Development Policy Review, 2002, 20 (5): 589-614

Safety Nets and Opportunity Ladders: Addressing Vulnerability and Enhancing Productivity in South Asia Naila Kabeer∗

As patterns of poverty and vulnerability in South Asia change, households have to balance immediate needs and long-term goals. For the poor, these choices, and the costs of precautionary measures, are particularly acute and call for suitable government policies. While policy-makers face a number of trade-offs between promotion, prevention and protection goals, careful design can maximise the potential to reconcile these objectives. A review of experience suggests a number of lessons regarding the relative benefits of targeted and universal programmes; the need to differentiate microfinance products for different groups amongst the poor; ways of basing the self-targeting of public works on rights rather than stigma; and the influence of political processes (such as decentralisation) for the overall effectiveness of social protection.

Increased attention to social protection on the part of international agencies1 reflects concerns about the increasing vulnerability to risks associated with globalisation, structural adjustment and economic reform, regional and international financial crises, environmental degradation, and demographic transition. Vulnerability varies across the world, as do social protection measures designed to deal with it. Within Asia alone, countries vary considerably by income, growth rates, incidence of poverty, inequality, forms of government, political economy and extent of integration into global markets. For countries undergoing transition from plan to market, attention has focused on the emergence of newly excluded groups as state-guaranteed social welfare systems are dismantled (see Cook in this collection). In South-east Asian countries which opened up to international market forces and bore the brunt of the Asian economic crisis, globalisation is seen as a major cause of insecurity (Suwannarat, 2000). Older forms of vulnerability persist in South Asia, although they may have been exacerbated by some of the changes mentioned above. Economic liberalisation has been relatively slow compared with Latin America and East Asia. South Asian countries remain more or less characterised by a dual economy: the formal sector (which offers higher remuneration, better working conditions and greater social protection) constitutes a very small part of the total (only about 15% of India’s workforce in 1987-8 was made

∗ Professorial Fellow at the Institute of Development Studies at the University of Sussex. She wishes to thank Ramya Subrahmanian for valuable assistance and feedback on earlier drafts of this report, and Gary Suwannarat and Sarah Cook for their comments. She would also like to acknowledge the support of the Ford Foundation and the UK Department for International Development for her research. 1. All major international agencies have had conferences and/or produced publications on this topic in recent years: see, for example, World Bank (2000); and Holzmann and Jorgensen (1999).  Overseas Development Institute, 2002. Published by Blackwell Publishers, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA. 590 Naila Kabeer up of regular salaried employees). Most economic activity take place in unregulated, informal sectors which straddle urban and rural areas. Consequently, it is measures aimed at protecting the workers in the informal sector, rather than those operating in formal labour markets, which are most relevant to discussions about risk and vulnerability in the region. The long history of dealing with such measures in the Indian sub-continent provides a rich repertoire of lessons regarding what to emulate and what to avoid. At the same time, the pervasive problem of poverty in the sub-continent means that it is beyond the fiscal capacity of any state to devise measures capable of dealing with vulnerability in all its manifestations. Social protection measures have to be integrated into the overall development strategy of the country rather than treated as stand-alone programmes (Suwannarat, 2000; Guhan, 1994). As Jimenez (1999) suggests in his discussion on ‘modernising safety nets’, the challenge is to devise safety-net measures which are investments in production rather than simply protection of consumption. This article explores some of these issues through a review of the literature on social protection in India and Bangladesh. It brings together lessons from responses to vulnerability, exploring the extent to which these responses lend themselves to ‘investment’ in broader development goals. The article is structured as follows. The next section provides an overview of poverty and vulnerability in India and Bangladesh. The third section describes the main social protection measures in the two countries, while the following section draws out key lessons from this experience. The final section offers some concluding comments.

Poverty, vulnerability and social protection in South Asia

Trends in poverty

Poverty has declined in India and Bangladesh over recent decades but at differing rates and with considerable fluctuations. The poverty rate in India fell from 45% in 1950 to 36% in 1993-4 (World Bank, 1998a), with urban-rural differences narrowing over this period; but population growth has resulted in an increase in the absolute numbers in poverty (from around 165 million to 320 million). The poverty rate fluctuated until the mid-1970s and declined sharply between the mid-1970s and mid-1980s, and then gradually up to 1990. Rural poverty rose particularly sharply immediately after stabilisation measures were implemented (from 37% in 1991 to 44% in 1992), primarily as a result of cuts in public spending: this halted growth in rural non-agricultural employment, which had helped to drive poverty reduction throughout the 1980s. Falls in rural employment rates led to a reversion to casualised agricultural employment under ‘distress’ circumstances (Ghosh, 1998). Economic recovery has been rapid since the early 1990s, but trends in poverty remain unclear because of fluctuations. The decade following Bangladesh’s independence in 1971 was crisis-ridden: as a result of a massive cyclone plus war and its turbulent aftermath, which culminated in famine in 1974, poverty shot up to around 80%. The poverty rate fell in the first half of the 1980s, then rose again in the second half. The 1990s saw a significant and sustained decline in Bangladeshi poverty, particularly in the urban areas: while rural poverty was estimated at around 57% in the mid-1990s, urban poverty declined more systematically, Addressing Vulnerability and Enhancing Productivity in South Asia 591 to around 35%. Urban-rural differences have thus widened over this period (World Bank, 1998b; H. Z. Rahman et al., 1998).

Patterns of poverty and vulnerability

Poverty and vulnerability in South Asia are strongly rural in nature and particularly severe in remote regions2 and in areas prone to natural disaster or difficult to irrigate. Poverty has fallen fastest where agricultural growth and improvements in human development have gone hand in hand. Indian states in which the poor made up around 40% of the population in the early 1990s include Bihar, Madhya Pradesh and Orissa: in terms of absolute numbers, Uttar Pradesh and Bihar account for around 27% of the country’s population but over 32% of its poor. In Bangladesh, regions with large urban areas have fared best, reporting lower levels of poverty than other regions. The poor tend to be those with few productive resources; in a largely agrarian economy, they tend to be landless and to rely on physical labour to earn a living. However, poor nutrition and health, and lack of education and skills undermine returns to labour. In addition, lack of assets deprives the poor of an adequate buffer against crisis, leaving them vulnerable to declines in their standard of living. Poverty and vulnerability in South Asia are not generally associated with open unemployment: instead, the problems reflect the nature of the activities that poor and vulnerable groups engage in. Those who lack assets, education and social networks tend to be in the worst paid and most casualised segments of the labour and commodity markets, engaging in a multiplicity of intermittent, seasonal and/or poorly paid activities to survive (see also Harriss, 1992: 342). Casual agricultural waged work is the most poorly paid form of employment available to the poor in India and Bangladesh (World Bank, 1998b; NIRD, 1999). Poverty also has a social dimension in South Asia, characterised by entrenched social inequalities, particularly those associated with caste and gender. Gender cuts across class, leading to deprivations and vulnerabilities not necessarily associated with household income. Female participation in the labour force is much lower than male across India, particularly in the northern plains, as well as in Bangladesh. Caste exacerbates poverty for some, confining those from lower castes to poorly paid, often stigmatised occupational niches and exclusion from welfare and poverty-reduction services. Escape is only possible by migrating to areas where their caste is unknown. Age is also emerging as a marker of vulnerability. As social norms change and family structures shift from joint or extended family systems to nuclear families, the care of the aged and infirm is increasingly viewed as a burden among poorer households, although ‘… aged persons who own assets are more likely to be well looked after’ (Shariff, 1999: 214). The very young too tend to be vulnerable: their welfare may be sacrificed in times of scarcity in order to protect economically active household members. Finally, a concern with vulnerability draws attention to the ‘vulnerable non-poor’ or ‘tomorrow’s poor’, as Rahman describes them. These are households above the poverty

2. For instance, 76% of the poor in India and 89% of the poor in Bangladesh live in rural areas. 592 Naila Kabeer line, but whose livelihoods are so precarious that any downward mobility in the event of a crisis will push them into poverty.3

The changing nature of vulnerability: the erosion of informal safety nets

Social security in South Asia was traditionally embedded in a variety of social and production relations: the caste-based jajmani system under which landlords extended food security to labourers and craftsmen; common property and customary rights; and institutionalised feeding in temples, maths and langurs (Guhan, 1992). These older forms of security have been weakened: population growth and the emergence of a labour surplus slowly negated the need for landlords to retain the loyalty of labour through crisis insurance (Vasavi, 1999a). The disintegration of jajmani is linked with the opening up of village economies (Breman, 1995). Such changes were not necessarily negative if, as in the more accessible regions, they opened up opportunities for the poor (Breman, 1996: 238). Gough (1981) also notes how ease of access to the market economy affected village occupational and social structure and hence social relationships between privileged and poor. In Thanjavur, Tamil Nadu, coastal regions which had been more ‘disturbed’ by external change had also broken more easily with traditional caste-based occupational structures. Where an area’s comparative advantage remained in agriculture, however, the traditional social order, with its hierarchy and caste-based occupation, showed less evidence of change (Epstein, 1973). And where changes leading to the erosion of informal safety nets were not accompanied by the opening of key local markets to the poor, households often found themselves deprived of older forms of protection while being exposed to new risks. This is highlighted in Vasavi’s analysis of a spate of suicides by farmers in southern India in 1998. While there was a generalised increase in insecurity in agriculture as a result of the substitution of high-productivity, high-risk crops for low-productivity, low-risk crops, small farmers proved most vulnerable. They had neither state support nor capital and insurance markets to underwrite risks which might previously have been shared across caste- and kin-based support networks. There is evidence from Bangladesh, too, that older forms of protection are declining. Adnan’s review of village studies between 1942 and 1988 points to the decline of traditional patrilineage and kin–based networks as sources of security, suggesting that deteriorating economic conditions resulted in a breakdown in social organisational forms during the crises of the 1970s, most clearly in the 1974-5 famine (1990: 165). Informal safety nets have not disappeared (see Rahman, 1995), but Adnan suggests they have shifted from kin ties to employer patronage, and shrunk considerably in scope.

Coping with vulnerability: household responses

Vulnerability adds a concern with fluctuations to the concern with levels of income, consumption and well-being of conventional poverty analysis. As Drèze and Sen (1991: 10) point out: ‘The average experience of the poorer populations understates the precarious nature of their existence, since a certain proportion of them undergo severe –

3. They have been estimated to make up 21% of rural households in the Bangladesh context. Addressing Vulnerability and Enhancing Productivity in South Asia 593 and often sudden – dispossession, and the threat of such a thing happening is ever- present in the lives of many more’. Burgess and Stern (1991: 40) note that chronic severe deprivation in developing countries is compounded by the uncertainty with respect to livelihood and life which threatens a wider section of the population. Fluctuations in the incidence of poverty thus provide us with an aggregate indicator of vulnerability, since they track the movement of households into, and out of, poverty. Vulnerability takes different forms because it arises from different causes. Like poverty, it can be structural or transitory, reflecting risk from routine or predictable phenomena or unanticipated risks. Seasonal deficits – food prices and availability or slackness in the labour market – are the most common ‘anticipated risks’ in South Asia. The implications are usually most severe for those reliant on agricultural labour. The most common unanticipated risks in Bangladesh include natural disasters, illness-related expenditures and various kinds of social and economic ‘insecurity’ (H. Z. Rahman et al., 1998); in India they tend to be natural calamities, the death of the breadwinner, the need to meet customary obligations, political instability, and market shocks. It is not exposure to risk per se which defines vulnerability, but the ability of households to deal with it in ways that do not jeopardise current well-being or future prosperity: fluctuations in income due to these risk events impinge on households regardless of economic status, but are devastating for the poorest (Vasavi, 1999a: 70). Consequently, while household livelihood strategies are organised around immediate as well as long-term goals, security and the capacity to cope with crisis are likely to be important among them. Figure 1 points to some of the goals of poorer households identified by the literature, including material goals – survival, security and accumulation – and intangible ones, including dignity, self-respect and social standing.

Figure 1: Household livelihood strategies: a continuum of goals and means

Expanding Enhancing and Reducing productivity diversifying Long- Immediate fluctuations in of current access to term needs: consumption assets institutions goals: Survival Self- and Meeting Reducing Expanding, Improving reliance/ security basic fluctuations diversifying terms of accumul- needs in income and access to ation flows strengthening institutions asset base

The extent to which households are able to take precautionary or ex ante measures against crisis is one indicator of vulnerability. Precautionary measures can, however, carry costs in the form of ‘diminished lives’ (Drèze and Sen, 1991: 11): households in risk-prone rural India can sacrifice up to 25% of average income to reduce exposure to shocks (Matin et al., 1999). But the costs of crisis are likely to be higher for those 594 Naila Kabeer households which lack the resources to take anticipatory action.4 During crises, households which have been unable to prepare themselves in advance are at their most vulnerable: unfavourably positioned to bargain for higher wages, better prices and favourable credit terms, they are often forced to cope in ways that undermine their capacity to recover or to protect themselves against the next crisis. The extent to which responses are reversible is another indicator of vulnerability. Figure 2 offers an illustrative sequence of crisis-coping responses. Households which rely on strategies at the ‘irreversible’ end of the spectrum are likely to be the most vulnerable or exposed to crisis for longer. Where possible, households will avoid irreversible strategies because these leave them less able to recover and more vulnerable to the next crisis.

Figure 2: Gradation of crisis coping strategies by reversibility: an illustrative classification

• Cut down on consumption • Draw down savings • Borrow from kin/neighbours • Borrow from moneylenders • Intensify labour inputs Reversibility of • Sell off consumer assets strategy • Temporary migration • Sell off producer assets • Permanent migration • Begging, prostitution, crime • Disintegration of family unit

Time Source: Based on Maxwell and Frankenberg (1992) and Lariviére et al. (1999).

Households may face trade-offs between security and the value they accord dignity and self-respect. Migrant labourers described by Breman chose to forgo the security of permanent employment with high-caste landlords for riskier options of urban wage labour, because it allowed them to escape the stigma of caste. Research in Bangladesh suggests that many households prefer to cut basic consumption rather than allowing female members to go out to work, again a relatively reversible strategy (Kabeer, 2001a). A disaggregated analysis of household attempts to deal with crises provides some insights into the vulnerability of vulnerable individuals within the household. Property owned by women is often sold off first (Agarwal, 1991; Kabeer, 1991), and the consumption of women and children may be cut back first, or they may be abandoned to protect the consumption of adult males5 (Greenough, 1982; Alamgir, 1980; Adnan,

4. According to Rahman (1995), rural households in Bangladesh experienced losses of nearly 20% of average annual income in 1992 as a result of crisis events: 11% as a result of the crisis itself (e.g. failure of crops) and 9% as a result of attempts to cope. 5. The rise in excess female child mortality after the 1974 famine in Bangladesh suggests that such a strategy can be fatal. Addressing Vulnerability and Enhancing Productivity in South Asia 595

1990). Finally, reliance on child labour or withdrawing children from school are common means of anticipating or dealing with crisis (Kabeer, 2001a).

Coping with vulnerability: state responses

The inadequacy of remaining traditional community safety nets has focused attention on the state as the obvious alternative provider of social security to the poor. However, the pervasiveness of poverty and vulnerability in the region significantly constrains the scope of formal protection measures. Private credit and insurance markets are underdeveloped, as is social insurance, given high levels of self-employment, unstable and irregular wage employment and widespread underemployment. The resources available for formal social protection measures are limited, given low tax revenue and competing demands on the budget; and means-testing, which might allow for more targeted transfers to the needy, is hard to apply, given the irregularity of incomes and diversity of income sources. It is particularly difficult to reach rural (and to a lesser extent urban informal sector) populations which are spatially scattered, occupationally diverse and, administratively complicated to serve. Guhan concludes that ‘neither Beveridge nor Bismarck nor Roosevelt can provide a model for social security in developing countries’ (1994: 38). Instead, social protection in poor countries has to be viewed ‘as part of, and fully integrated with, anti-poverty policies, [which are] broadly conceived in view of the complex, multi-dimensional nature of poverty and deprivation’ (ibid.). In order to develop some conceptual clarity about the notion of social security – one that is neither excessively specific and hence limited in scope, nor so general as to encompass development itself – Guhan distinguishes between:

• promotional measures to improve real incomes and capabilities; • preventive measures, to avert deprivation in specific ways; • and protective measures, which more specifically aim to guarantee relief.

These may overlap, but they suggest a gradation of approaches (see Figure 3). An outer circle of promotional measures would include macroeconomic, sectoral and institutional measures relevant to poverty reduction. Preventive measures, the middle circle, are direct measures for poverty alleviation, which we shall be discussing later. Finally, the inner circle of protective measures is narrowly targeted safety-net measures in the conventional sense, aiming to provide relief from poverty and deprivation to the extent that promotional and preventive approaches have failed to do so.6 This conceptualisation points to the need for multiple approaches to social protection where poverty and vulnerability are widespread and highly differentiated. It also corresponds to the needs of different vulnerable groups within the population: those whose livelihoods are too precarious to guarantee basic survival needs and who need

6. The World Bank suggests that improving primary education, reducing communicable diseases, improving water and sanitation and reducing household insecurity through public works programmes would do most to reduce poverty. Of these, the first three can be regarded as promotional measures, while the fourth is a prevention or protection measure, depending on its design. 596 Naila Kabeer

Figure 3: Conceptualising approaches to poverty and vulnerability

PROMOTIVE MEASURES

PREVENTIVE MEASURES

PROTECTIVE MEASURES straightforward assistance; those who are able to meet basic needs but need greater security of livelihoods; and those who would benefit from improved overall opportunity. Safety nets designed to protect are a last resort, reserved for groups unable to benefit from other measures, or for situations when those prove inappropriate.

Lessons from the South Asian experience

The most important measures for protection and prevention in South Asia are self- employment, promoted through subsidised credit or assets; wage employment through public works; and the distribution of basic goods at subsidised prices. In both India and Bangladesh, public expenditures on anti-poverty and safety-net measures rose considerably during the 1990s. NGOs have also become involved in the delivery of social protection cover in a variety of ways, with the range and sophistication of approach particularly pronounced in Bangladesh.

Targeting and trade-offs: universal versus selective provision

The ‘politics of targeting’ is an issue whenever one group finances a subsidy to another, but is most pronounced with safety nets, which tend to be straight transfers of one sort or another. At the heart of the issue are the trade-offs associated with different methods of targeting: the concentration of transfers on the poor achieved through finely tuned targeting has to be offset by the costs of such an approach (higher administrative costs, more invasive information requirements and greater power assigned to administrators). Because no targeting mechanism is perfect, there is a trade-off between E-errors, the inclusion of non-eligible participants, and F-errors, the exclusion of the eligible. Which is less desirable will depend on whether priority is given to the ‘inclusion’ of the eligible, or to excluding the ineligible: the decision about which set of errors is more acceptable is as much political as it is technocratic. Estimation of the marginal odds of participation by different quintile groups in India’s three main poverty alleviation programmes found that self-targeting public works programmes performed best in reaching the poorest quintile. IRDP credit was more effective within the three middle quintiles (including those living at the poverty Addressing Vulnerability and Enhancing Productivity in South Asia 597 line – roughly the 40th percentile), but not the extreme poor. And the food subsidy via the Public Distribution System (PDS) flowed disproportionately to those least in need (Lanjouw and Ravallion, 1998, cited in World Bank, 1998a; also Subbarao et al., 1997). The failure of the PDS to reach poorer sections of the population resulted from the fact that the allocation of food grains to states did not necessarily reflect the incidence of poverty, and from the influence of aspects of state-level political economy, including the presence of surplus food grain production, populist government/active civil society favouring pro-poor distribution, and efforts by the state to supplement central allocation (Mooij, 2000). Thus, while aggregate results suggest that the PDS disproportionately benefits the better-off sections, its ability to reach the poor varies by state. It appears to have been most effective in reducing poverty in some southern states, with estimated declines of between 5.1 and 6.3% in Andhra Pradesh, Kerala and Karnataka in the mid- 1980s, and to have been least effective in the poorer states of Uttar Pradesh, Orissa and Bihar where the decline in poverty attributed to the PDS was around 0.4% (see Mooij, 2000, on Bihar ). One conclusion (advanced by the World Bank and adopted by the government under economic reform) is that the PDS needs more selective targeting so that benefits flow disproportionately to the poor (World Bank, 1991; Suryanarayana, 1995).7 Others disagree. Ghosh (1998) notes that targeting is not costless, and that, given the importance of the PDS in making the minimum dietary needs of the poor affordable, it should be made more genuinely universal through further improvements in coverage, rather than restricted. That some states report greater success in reaching the poor through the PDS supports Ghosh’s argument: as Patnaik (1998) points out, these tend to be populist governments with the political will to ensure greater distributional equity. Harriss-White (1999: 310) is more ambivalent, citing a study from rural Tamil Nadu which confirmed that the poorest deciles had defective access, which supported arguments for the removal or extreme streamlining of the PDS. At the same time, she reminds us of the Scheme’s socially redistributive impact: in the Tamil Nadu study it contributed 35% on average to rice calories consumed, but 40% to the diets of poor peasant households and up to 72% to those of landless labourers. Differing positions around the future of the PDS reflect differing values as much as differing assessments of its impact. Those predisposed in favour of market forces and against state intervention tend to favour lower budgets and taxes, and hence a more selective PDS (Gelbach and Pritchett, 1997). They view support for a more universalist PDS as ‘distributivist’: that is, premised on socialism and egalitarianism (Lal, 1994). However, there are pragmatic as well as ideological arguments against targeting, including higher administrative costs and the possibility of corruption. Indeed, economists not usually associated with redistributivist ideals have suggested that ‘leakier may be better’ if leakage provides those most able to contribute financially to a programme with some stake in its continuation: the budgetary support that a programme is able to command within the population is responsive to who benefits from it (Gelbach and Pritchett, 1997).

7. Food rationing under the Bangladesh Public Distribution was also found to disproportionately benefit the relatively better-off (Abdullah and Murshid, 1986). Food rationing was abolished by the late 1980s under structural adjustment and food distribution increasingly shifted to food-for-work and other safety net programmes. 598 Naila Kabeer

The ‘depth’ of outreach: microfinance and the extremely poor

Access to financial services for the poor helps smooth income and consumption, enables the purchase of inputs and productive assets, and provides protection against crises. Government efforts to compensate for the exclusion of the poor from formal financial institutions such as the IRDP in India have failed to reach the very poor and socially excluded and have experienced widespread loan default – largely because the bureaucratic, top-down identification of an internally heterogeneous category of poor people used in these programmes has resulted in targeting errors despite compliance with official rules (Drèze, 1990). In Bangladesh too, the delivery of subsidised credit and inputs to farmers’ co-operatives through the BRDB was also found to benefit better- off farmers disproportionately.8 The proliferation of NGO-led microfinance programmes, many of which deliberately target women from poorer households, was partly in response to the perceived failure of government efforts. The Grameen Bank in Bangladesh is internationally renowned for lending to poor women in groups responsible for ensuring weekly repayment. Also in Bangladesh, the Bangladesh Rural Advancement Committee (BRAC) and Proshika have combined lending to the poor with some social mobilisation, while others, like the Association for Social Advancement (ASA), have modified the Grameen approach to offer a more diversified range of financial services. Less well-researched is the self-help model pioneered in India by organisations like PRADAN. Most of these programmes report high repayment rates and have benefited large numbers of households below the poverty line. However, they too fail to reach the extremely poor (Wood and Sharif, 1997; Hulme and Mosley, 1997; Halder and Husain, 1999; A. Rahman et al., 1998; Rahman and Hossain, 1995) who operate in the ‘mini- economy’ in which production, consumption, exchange, trade, savings, borrowings and income-earning occur in very small amounts (Matin et al., 1999). These small units of transaction make it difficult for financial institutions to charge standardised administrative costs. The constraints posed by the high transaction costs of dealing with the extremely poor have been exacerbated by increased emphasis within the donor community on the ‘sustainable’, as opposed to the ‘subsidised’, transfer of resources to the poor. Increased stress on loan repayment affects organisations’ ability to be responsive to the fluctuating income flows of the very poor, while the emphasis upon weekly repayments generates additional pressures on groups to exclude the very poor (Montgomery, 1996) – and indeed encourages self-exclusion from credit programmes by the risk-averse extremely poor (Evans et al., 1995; Hashemi, 1997). Recognition of the failure of micro-credit programmes to reach the extremely poor has elicited a number of responses: that micro-credit is irrelevant for those who need the security of wage labour (Khandker, 1998; Drèze, 1990; Rath, 1985); that it merely replaces older dependencies on moneylenders with dependency on NGOs (see Kabeer, 2001b, on Nijera Kori); or that the problem is not with micro-finance per se, but with the objectives for which and terms on which it is provided. Hulme and Mosley (1997)

8. The DWCRA programme set up in response to such failures has fared little better (Kabeer and Murthy, 1997). Addressing Vulnerability and Enhancing Productivity in South Asia 599 trace failures of outreach to the shift to the ‘promotional’, enterprise-focused model of credit provision for poverty alleviation. Such a model benefits moderately poor households with the capacity for enterprise and risk, but it excludes the poorest whose livelihoods are insecure, and who are in need of preventive and protective support. They argue for financial services for the poor in terms of their intrinsic value in helping poor people to meet their needs rather than their instrumental value in promoting development goals. This means moving beyond the ‘micro-credit template’ to a more flexible range of financial services. Some NGOs in Bangladesh have redesigned interventions to reach the extremely poor. BRAC’s Income Generation for Vulnerable Group Development (IGVGD) programme targets the poorest women, combining some assurance of household food security with assistance in enterprise development over a longer than usual timeframe. It targets the poorest women: members of assetless households, women with irregular or no household income, women who work as casual wage labour, and female household heads. Members are provided with wheat rations for two years, during which time they are expected to form savings groups and train in income-generating activities, for which they are provided with credit. Surveys have found the IGVGD to be successful in targeting destitute women who are excluded or exclude themselves from NGO activities, and in enabling them to reach a stage where they can become members of NGO micro-finance activities. BRAC proposes to extend this programme to other sections of the excluded poor over the next five years and to complement asset grants and credit with inputs and the provision of necessary technical services and supervision. The Grameen Bank changed its approach when it found that recipients of loans in Rangpur were behind on repayments. Offering loans in Rangpur, one of the most economically depressed areas in Bangladesh, was likely to exacerbate problems, as poor recipients would use the loans for consumption or to repay other loans. The bank embarked on a goat-leasing programme, providing defaulting borrowers with a goat which they could raise and then repay in the form of a kid from the first and second litters, with no cash repayments required. Since goats are hardy animals, the women had repaid their loans by the end of the first year. The programme has successfully brought many of the poorest back into its micro-credit programme. NGOs have also sought to improve their outreach to the poor by means of experimentation and diversification. ASA started a flexible ‘open-access’ savings service for its core poor group, setting a quota for them in its entrepreneurship development programme. Associated membership status allows them to benefit from group activities without savings and other responsibilities. Proshika gives its very poor borrowers access to risk-reducing technologies from its Employment and Income Generation Programme profiles, so that borrowers feel less threatened by actual or potential income losses. Provisions which allow the very poor greater security for their savings have proved a major innovation. This success suggests that the priority for many poor people may lie less in the ability to access new sources of credit than in their ability to secure their savings while also ensuring flexible access. ‘Self-help groups’ to reach the very poor and socially excluded are important innovations in the Indian context. These groups are initiated around self-generated savings and credit activities without external financial assistance. Members agree on how much to save and the size of loans to each other. PRADAN’s ability to reach poorer groups has been due in large part to tailoring the scale of its activities to the 600 Naila Kabeer capabilities of the poor, focusing on ‘tiny’ irrigation schemes and on adapting production technology for poultry and mushroom cultivation to the production systems of poor people. Over the longer run, PRADAN seeks to develop linkages between self- help groups and mainstream financial institutions to ensure that access to external credit is sustained. Finally, it helps groups to invest their borrowed funds in viable income- generation activities, often mobilising funds from government programmes for infrastructure for livelihood-generation activities. At the other end of the poverty spectrum are ‘tomorrow’s poor’, households excluded from financial services for the poor because they own land or are above the poverty line, but which are also too asset-poor to access the formal banking system. They too have tended to lose out from the ‘micro-finance revolution’. This group includes many insecure households which would benefit considerably from access to financial services. Financial services, often small-scale rather than micro-enterprise for this group, require a change in orientation from ‘asymmetrical, and paternalistic’ approaches, characteristic of many micro-credit organisations, to ‘more robustly businesslike’ and demand-driven services (A. Rahman et al., 1998; see Kabeer, 2001a, for an example from Bangladesh).

Seasonality, stigma and self-targeting: public works and employment generation

Public works programmes tend to be more effective than other forms of prevention and protection in reaching the poorest because they are self-targeting: they offer benefits (inferior foods or lower-than-market wages) or impose requirements (unskilled manual labour) which are likely to attract only the poorest. They enjoy low administrative costs (unlike means-tested programmes) and fewer errors of inclusion (unlike untargeted transfers). However, public works programmes raise questions about the terms on which public assistance is provided to the poor. What makes self-targeting effective is the social stigma attached to participation. It could therefore be argued that while such programmes may have lower administrative costs than other targeted programmes, they have higher social costs for intended beneficiaries. Stigma can, however, be reduced, though not completely eliminated, through some sensitivity in programme design. Vasavi’s study of famine relief in Karnataka shows how early colonial efforts to provide famine relief failed to achieve the participation hoped for because they were designed without attention to local values; and how gradual improvement in uptake was due in part to a worsening situation, but in part to a redesign of relief efforts which made them more acceptable (1999a). Many of these lessons have now been widely accepted. The Maharashtra Employment Guarantee Scheme (EGS) in India is a particularly successful public works programme which exemplifies many of these features. As normal, it is self-targeted, offering unskilled wage labour at wages below the agricultural labour market rate. What sets the EGS apart, however, is its legally enshrined ‘guarantee’ of employment (Moore and Joshi, 1999): work is a right rather than a welfare measure. The guarantee of work has strengthened the bargaining position of landless labourers vis-à-vis employers (Echeverri-Gent, 1988), allowing relatively marginalised groups to become less dependent on landlords for economic support (Drèze, 1990). Moreover, as Moore and Joshi point out, the right to employment embodied in the EGS legislation guarantees Addressing Vulnerability and Enhancing Productivity in South Asia 601 other enabling benefits (for example, payment if workers use their own tools; plastic spectacles for those employed in stone breaking, etc.). The right to childcare and the guarantee of work within eight miles of home help to explain the high female participation rates (Engkvist, 1995). The legislative status of both the primary right to work and these subsidiary rights has provided the basis for political activism focusing on their fulfilment. In this, the EGS stands in marked contrast to the Employment Assurance Scheme (EAS), a nation- wide scheme implemented by the central government, which is identical to the EGS, ‘except for the lack of a guarantee of work’. This is a crucial difference: ‘there is no client mobilisation around EAS, because the lever of rights is lacking’ (Moore and Joshi, 1999: 29). Recent research suggests the ‘guarantee’ element of the EGS has weakened and with it its targeting efficiency, as a result of the decision in 1988 to double piece rates in line with the doubling of statutory minimum wage rates in agriculture. De facto rationing occurred because of the failure to increase the budget commensurately with the wage increase: this undermined the ability of the EGS to provide employment for all who needed it, and thus to underpin the minimum wage (Ravallion et al., 1993: 269). Such findings underline Sen’s (1998) point regarding the need for clear objectives in the design of safety nets. He identifies two trade-offs: firstly between generating unskilled employment opportunities for poor households and creating community infrastructure (which may require more skills and higher non-labour costs) and, secondly, between paying minimum wages (and thereby improving the living standards of those who gain such employment) and maximising the amount of new employment that can be generated. He suggests that, while asset creation and income redistribution are both laudable objectives, they should not be confused with the primary goal of a safety-net measure – to generate employment for vulnerable households in time of need. The employment guarantee element of such schemes can only be upheld if wages are set at levels where the demand for employment matches the funds available. The timing of public works may also be important. A. Rahman et al. (1998) have pointed out that, as new forms of irrigation have changed the seasonality of labour demand, food-for-work (FFW) schemes that historically provided work during the slack season are in effect now operating during periods of peak labour demand.

Anticipating crisis: the importance of ex ante measures

Precautionary measures against crisis may diminish a household’s standard of living, but ex post responses are often more costly. Subbarao et al. suggest that the same principle may apply to public safety-net measures and that differences in the nature and effectiveness of public administration may explain the harsher consequences of drought in sub-Saharan Africa compared with South Asia. Factors behind the better outcomes in South Asia include ‘the prevalence of consumption-smoothing public works programs, an administrative and institutional framework that is well-adapted (over time) to drought situations, and a free press’ (Subbarao et al., 1997: 14); and the willingness and ability of government to scale up public works efforts (Mundle, 1994). A comparison of the 1974-5 famine in Bangladesh with the extended drought in Maharashtra in the early 1970s underlines the importance of ex ante measures. The more serious outcomes in Bangladesh (including much higher rates of distress land sales) may have reflected the 602 Naila Kabeer harsher risk environment in the country, but the timely provision of public works employment in Maharashtra (which was later to evolve into the EGS) seems to have helped (Cain and Lieberman, 1982). A comparable relief effort by the Bangladesh Government would have saved many people from starvation and impoverishment (Ravallion et al., 1993: 166). The response to the 1998 floods in Bangladesh provides a barometer of change in the country.9 This response clearly benefited from the lessons of the past and from the institutionalisation of safety-net and other programmes. A Vulnerable Group Feeding programme began in flood-affected thanas soon after the onset of the floods in August. The government removed tariffs on imported rice after a production shortfall in early 1998: the private sector imported 1 million tons of rice through official channels (more than 200,000 tons by March 1999), keeping prices at import parity levels. Government and NGO efforts and private sector supplies reduced consumption fluctuations and helped distribute food and oral salines. The provision of information about relief entitlements also ensured local-level pressure on officials to ensure proper distribution. Some important lessons can also be drawn from the experience of the 1998 floods with regard to the role of the micro-finance sector in countering risk, which in Bangladesh in 1998 demonstrated ‘a new level of institutionalisation of credit’ (Johnson, 2000: 4; Development Alternatives Inc., 2000). The largest organisations provided relief and reconstruction assistance immediately from reserves, and survived the shock by running these down. The Grameen Bank also benefited from linkages with the formal banking system, borrowing $50m. to assist recapitalisation. While the smaller microfinance organisations suffered from a liquidity squeeze,

[i]n contrast to the floods of 1988, there were no calls for loan write-off, either from government or borrowers. Instead the sector as a whole sought to promote standards of good practice which included not writing off loans; rapid assessment of the impact on the MFO’s portfolio; the development of new products and responses in order to create safety nets for vulnerable groups; and ensuring that relief work did not undermine the financial management of the MFO (Development Alternatives Inc., 2002: 4).

This willingness to continue loan repayments signalled an acceptance of NGOs as credit providers rather than merely conduits for relief. Household responses also provide insights into the significance of micro-finance: most preferred to reduce consumption or borrow from other sources rather than suspend loan repayments which might have jeopardised future borrowing (Zaman, 1999).

Climbing out of poverty: the developmental potential of protection measures

In addressing whether protection and promotion of livelihoods can be linked, it is necessary to note the varying nature of vulnerability. Those physically unable to take advantage of economic opportunities are least likely to be able to enter public works programmes or contribute to social insurance schemes: for these groups, non-

9. The speed and effectiveness of the response in Bangladesh is in marked contrast to the response to the cyclone in neighbouring Orissa in 1999, where there were few NGOs on the ground and the government response was both belated and badly managed. Addressing Vulnerability and Enhancing Productivity in South Asia 603 contributory social assistance may be most appropriate. For the extremely poor who are physically able but lack economic assets, the priority may be to reduce fluctuations and build up assets. For those who face social discrimination as well as economic deprivation, voice and agency may be as important as social protection. Moderately poor households and the vulnerable non-poor are most likely to benefit from some combination of protection and promotion which gives them a more secure base from which to diversify into higher-return, higher-risk activities. Policy measures for vulnerable groups must account for different capabilities and different needs. Protection measures can evolve into, or link up with, promotion by starting with the resources and activities that households currently manage and building up their capacity to manage more resources and higher value activities. PRADAN’s experience, for instance, suggests the existence of ‘livestock ladders’ in rural India: the very poor may at first only be able to manage small animals, but with experience and resources may graduate to larger livestock.10 There may also be ‘human capital ladders’: health needs are likely to be a priority for the labour-dependent poor, but once health needs are assured, other forms of human capital (literacy and numeracy) may become significant, leading on to other business skills. The literature on safety-net and poverty-reduction measures draws attention to different impacts of different measures. Sen (1995) has suggested that because public works programmes are better targeted than most anti-poverty measures they tend to be more effective in moderating the severity, rather than the incidence, of poverty. It is also important to note that the protection provided by the EGS during the major drought in Maharashtra in 1988 reduced the need for ‘irreversible’ household coping strategies such as selling land. The broader development role of public works has, however, been uneven. Lack of planning and co-ordination by local authorities has often prevented such programmes from generating infrastructure for the community. With reference to Sen’s point about the trade-off between creating work and creating infrastructure, Papola and Sharma express concern that the 60:40 ratio of wages to materials in the Jawahar Rozgar Yojana (JRY) public works programme in India limits the achievement of the secondary objective of asset generation, particularly as no allowance has been made for maintaining infrastructure once constructed (Papola and Sharma, 1996; Nayar, 1995). Public works projects were often ad hoc and poorly prepared, creating unsuitable or impermanent assets (Papola and Sharma, 1996: 602). Similar criticisms were levelled against FFW programmes in Bangladesh. Unclear resource allocation and recording, haphazard implementation under political pressure, and the desire to spread food aid as widely as possible, all ensured that FFW funds were put to uneconomic use. However, as a result of growth in food aid resources available through FFW programmes – WFP food aid had increased tenfold by 1984/85 – developmental objectives were built into the programme. A committee set up to explore this new orientation recommended integrating food aid resources into national and sectoral planning processes, and diversifying FFW activities to fisheries and forestry (SIFAD, 1989). The decision to combine relief with development in FFW activities

10. Narendranath, pers. comm. See Whitehead (1998) on livestock ladders in Ghana. 604 Naila Kabeer explained the success of the government’s response to the 1991 cyclone.11 CARE-Bangladesh, which executes the WFP project and contributes to FFW projects in its own right, also decided to shift to an Integrated Food-for-Development project in order to achieve a greater focus on development (Hyder, 1996: 30). Ten women from each administrative unit are selected to maintain rural roads for four-year periods. Their wages are saved in a bank and at the end they receive nine months of training and use the savings to start up businesses. By June 1997 the programme had provided nearly 60,000 women with training and contributed to rural development by increasing the year-round utilisation of roads (Langworthy, 1999; Ghani, 1999). The reduction in transport and travel costs facilitated an increase in roadside businesses and faster growth in agricultural incomes, based on easier access to agricultural input and output markets. Agricultural intensification had not generated wage employment: rather, labourers have moved into non-agricultural activities (commercial transport, small businesses) supported by roads. Finally, while men continue to make up 80% of users, rates of increase in use by women were much higher (29% compared with 18%). Use of roads to obtain access to education went up by 58% for female users and by 15% for male, and women and poorer groups reported increased access to NGO programmes and social services. The literature on financial interventions for the poor deals extensively with impacts relating to promotion as well as protection, with micro-finance organisations diversifying ‘products’ to address different groups and goals. There has been some debate about the extent to which financial services for the poor have reduced poverty. In India, estimates for 1977-9 and 1987-8 suggest that IRDP benefits have primarily benefited the moderately poor and near poor; only 19% of households that crossed the poverty line had benefited from IRDP assistance (Papola and Sharma, 1996). In Bangladesh too, Morduch (1998) found little evidence of increases in household income/consumption which might have resulted in a reduction in poverty from micro- credit. Such conclusions need careful interpretation, however. Findings from seven developing countries suggest that the impact of lending on the household income and asset position tends to increase at a declining rate as the income and asset position improves, reflecting the greater preference of the very poor for consumption loans, their greater vulnerability to distress sales of assets, and their limited range of investment opportunities (Mosley and Hulme, 1998). There also appears to be a threshold cumulative loan size beyond which micro-credit begins to make a significant impact on household poverty (Zaman, 1999; Hulme and Mosley, 1997). There is greater agreement about the impact of micro-credit on household vulnerability. Morduch estimated that Bangladeshi households with access to micro- credit experienced 47-54% less consumption variability than households without access. Other studies found that households with access to micro-credit were more likely to report land and/or non-land assets than those without (Hashemi et al., 1996; Pitt and

11. ‘Within days, the Bangladesh Water Development Board, with assistance from the WFP, took up the tasks of emergency relief and rehabilitation. Temporary low embankments were constructed … In an unprecedented action, the work was carried out during the monsoon, from May to September 1991. Wheat was disbursed as food-for-work wages, providing much needed food and employment to over 55,000 men and women … [and] effectively stopped further inundation … an investment of 35,000 tons of wheat, provided as wages, enabled the farmers to produce 170,000 tons of rice’ (Hyder, 1996: 21). Addressing Vulnerability and Enhancing Productivity in South Asia 605

Khandker, 1995; Bruntrup et al., 1997; Zaman, 1999), while others found that longer- term borrowing households had higher levels of productive assets than those that had recently joined (Montgomery, 1996). Many poorer women borrowers invested their loans in poultry and goats which acted simultaneously as savings and an investment: they were able to sell the eggs or kids to tide them over in times of need, while also retaining the original asset as a source of future income (Kabeer, 1991, 2001a). Access to credit allows households to diversify livelihoods so that irregular returns from trade or wage labour can be supplemented by lower paid but more regular forms of self- employment (Kabeer, 1998; Khandker et al., 1998). In some cases, the decision to take up self-employment in place of wage employment represents a decision to move out of a badly paid form of labour (Kabeer, 2001b). Finally, access to NGO credit may also represent a diversification of credit sources: a longitudinal study from India reported that NGO credit protected poor women from the worst effects of major crises, not removing reliance on moneylenders but allowing past debts to be paid off so that onerous interest payments were removed (Noponen, 1990: 250). A third impact relates to intra-household vulnerabilities. There is increasing evidence to suggest that the practice of targeting women in financial interventions has reduced their vulnerability by enhancing their sense of self-worth and the value accorded them by other household members, giving them a greater role in household decision-making, enabling savings and the acquisition of own or joint assets, and providing them with greater public mobility and increased awareness of rights (Kabeer, 2001b; Naved, 1994; Zaman, 1999; Schuler et al., 1996; Amin and Pebley, 1994; Hashemi et al., 1996; Noponen, 1990; Pitt and Khandker, 1995). Less well-researched but of equal importance is the apparent impact of credit on children, who are more likely to go to school, particularly if loans are made to women (Pitt and Khandker, 1995; Ridao-Cano, 1999 and forthcoming). Household receipt of credit may also help close gender gaps in education (Kabeer, 2001a). More research is needed on the wider impacts of these programmes on those who are not direct beneficiaries, and on the wider local economy. It has been suggested that tying credit to the purchase of a limited range of assets (as under the IRDP) or the confinement of the poor to certain activities can lead to saturated markets, particularly with female entrepreneurs who typically operate at the poorer end of highly segmented commodity markets. An evaluation of Nari Nidhi, an organisation in Bihar, found that when a large batch of loans were given to female fish vendors in particular villages, demand for fish was unable to absorb the sudden increase in supply, leading to falling prices and incomes. Evidence from Bangladesh, by contrast, indicates that the long-term presence of a micro-credit programme increased rural wages through a tightening of the labour market (Khandker et al., 1998). Anecdotal evidence suggests a similar effect as a result of SHARE’s activities in Andhra Pradesh, while PRADAN’s activities in Bihar and Rajasthan may have led to a fall in the interest rates charged by moneylenders (Narendranath, pers. comm.). Experimentation and innovation to enhance the pro-poor growth potential of micro- finance continue. New opportunities to help poor and disadvantaged groups overcome the constraints arising from segmented markets include the provision of mobile phones to village women by the Grameen Bank, which gives village women a new means of livelihood (selling time on the phone to other villagers). This initiative also, by 606 Naila Kabeer providing villagers with improved access to information about market prices, holds out the promise of greater integration of national markets.12

Rights, collective action and the politics of poverty reduction.

Apart from design and administrative weaknesses, social protection measures may suffer from problems arising from the political environment in which they are implemented, including ‘elite’ capture, the use of programme resources as patronage, corruption, revenge,13 and the vagaries of electoral politics. Populist politicians have often used government programmes to enhance their own popularity in ways that undermine programme effectiveness: waiving loans arrears or increasing wages on public works programmes without commensurate budget increases. Governments have sought to defuse opposition to economic reform through welfare schemes designed to ‘benefit groups whose clout might pose a serious threat to reform’ (Jenkins, 1999: 185). Indeed, the leakages of resources from poverty alleviation and safety-net programmes can serve political purposes by contributing to their sustainability during adjustment (Nelson, cited in Jenkins). Given these governance problems, there is an emerging stress on the importance of mobilising users to demand greater transparency and accountability from public providers. However, this mobilisation is itself not immune from manipulation: government initiatives to organise communities into user groups are not always inclusive or responsive to the needs of the poorest or the socially excluded (Saxena, 1994; Vasavi et al., 1997; Sarin, 1999). State-led corporatism targeted at the poor is open to charges of co-optation, where ‘the state mobilises and organises socio-economic groups in its own interests, and in ways that closely circumscribe the autonomy of the organisations and their capacity genuinely to represent societal interests’ (Moore and Joshi, 1999: 52). In theory, decentralisation expands opportunities for local communities to voice concerns to elected representatives, fostering transparency and accountability in decision-making. Decentralisation is no magic bullet, however. In much of India, it suffers from insufficient financial allocation to local levels, underdeveloped planning capacity and inadequate dissemination of information. Jenkins reports that government evaluations of JRY found that roughly one-third of village panchayats implementing JRY schemes did not possess the guidelines. In Bangladesh, H. Z. Rahman et al. (1998) comment that the potential of local governance, essential for tackling poverty, has so far ‘been restricted to only concerns of electoral rights and vague notions of local planning’. While local democratic elections have been a positive outcome of the decentralisation of the past decade, the results have been uneven. In areas where NGOs and social movements had been active prior to serious efforts at decentralisation – and/or where state policies have been supportive, as in Kerala and Karnataka – the

12. In Sri Lanka, the introduction of telephone services in rural areas increased farmers’ share of prices received for crops sold in cities by between 50 and 80%. 13. Gunatilaka (1999) cites a judge in Sri Lanka who observed ‘it is remarkable how often over the years it has turned out by some extraordinary coincidence that the public interest appeared to require the acquisition of lands belonging to persons politically opposed to the party in power at the time’ (p. 8). Addressing Vulnerability and Enhancing Productivity in South Asia 607 results have been positive.14 NGOs played an important role, for example, in the 1998 Bangladesh local government elections, when voter mobilisation through a network of NGOs led to high turnout levels, particularly among women, and in which some NGO group members from poor and landless sections contested and won Union Council posts. Other experiences with decentralisation have led to fears that local-level politics enable powerful groups to lobby more effectively for resources, capture more of the political space and create conditions under which the participation of the poor and excluded is even harder to achieve. Although Indian state policy mandates reservation of a fixed proportion of panchayat seats and leadership positions for traditionally excluded groups (Scheduled Castes, Scheduled Tribes and women), this entitlement to participation is yet to be fully accepted. While the evidence on decentralisation overwhelmingly documents the problems of equal participation of excluded groups in agenda-setting at the local level, there are still indications that there has been some positive impact in some places in terms of improving developmental results for the poor. Whether these outcomes are merely related to an overall improvement in efficiency as a result of decentralised administration (as Crook and Manor, 1994, note for Karnataka), rather than long-term redistribution of resources and power, remains to be seen. By far the greatest impact on social exclusion appears to have been through collective action by social movements that lobby to challenge institutional failures. In India, identity-based political formations have had considerable success in challenging caste-based discrimination, particularly in education (Velaskar, 1998; Heller, 1996). The right-to-information movement is another example of collective action tackling governance and corruption in government anti-poverty interventions (public works and the PDS). In Bangladesh, a DFID-funded review of ‘big NGOs’ found that their adoption of a service delivery role had undermined their ability to demand greater accountability and responsiveness from public sector services. The few organisations which had retained a focus on social mobilisation showed the greatest willingness to pressurise the state – and increasingly, other NGOs – for greater transparency and accountability. In the absence of any incentives within the public sector to observe these principles, it is collective action by grassroots movements (such as Nijera Kori and Proshika in Bangladesh) that may prove most effective in ensuring access to public goods and services (Wood, 1994; Kabeer, 2001b). NK is the only NGO of any significance in Bangladesh which has eschewed direct provision of services on the grounds that it undermines effectiveness in building the organisational capacity of the poor to demand their rights.

Conclusion

A review of vulnerability in South Asia and the various attempts to address it can draw out a number of key lessons.

14. For studies on effective decentralisation policies and experiences, see Franke and Chasin (1992) on Kerala; Crook and Manor (1994) on Karnataka. 608 Naila Kabeer

The first point relates to the issue of trade-offs and the challenges these present for the design of social protection policies. One set of trade-offs relates to household livelihood strategies for achieving various goals, including survival, security, status and self-worth, and accumulation. While all households face resource constraints on their ability to achieve these goals, poorer households clearly face ‘harder’ constraints and more limited options. Thus, while better-off households may face trade-offs between different kinds of ‘goods’, for example, better food or better clothes, poorer household tend to face trade-offs between different kinds of privations or ‘bads’, for example, demeaning forms of work which give some security or forms of work which offer independence but carry greater risks. Social protection measures can feed into the harshness of these trade-offs by stigmatising the poor. As Sen points out, any form of protection ‘that requires people to be identified as poor and that is seen as a special benefaction for those who cannot fend for themselves would tend to have some effects on their self-respect as well as on the respect accorded to them by others’ (Sen, 1998: 13). However, there are forms of protection which can actively mitigate the harshness of these trade-offs, for instance by seeking to transform employment on public works programmes from ‘special benefactions’ into contributions to community assets or guaranteed livelihood rights. Organisations which mobilise the poor to fight for their rights can play an important role in ensuring that such a transformation occurs from the bottom up. A second set of trade-offs occurs at the policy level between protection and promotion. To some extent, this trade-off may be inevitable. Relief requires the generation of employment (generally unskilled), while development requires the creation of assets and possibly a greater use of materials and skilled labour. Ravallion et al. (1993) and Sen (1995) point out one basic principle which has to be in place if public works are to be primarily a form of safety net, which is to ensure that wages are set at levels which maximise the level of employment compatible with the budget available: too low and it is unlikely to attract all those who need protection; too high and employment has to be rationed, and may fail to reach those who need it most. Others have emphasised the idea of safety nets which lend themselves to rapid evolution from relief to development, integrating the design of public works into sectoral planning processes, rather than treating them as stand-alone efforts. It may also require that project planning is decentralised to reflect local needs and constraints. Another version of the policy trade-off between protection and promotion is in the microfinance sector. Here the trade-off reflects the contrast between the need for microfinance organisations to achieve sustainability (with an associated bias towards better-off households which are more likely to repay their loans), and their social protection objectives, which require them to reach poorer households. The policy lesson here is to design microfinance interventions to meet the heterogeneous financial needs of the poor. A second set of points relates to thinking about synergies, complementarities and secondary effects, recognising that certain assets are far more productive when used in combination rather than on their own. For Bangladesh, Zohir (2001) points to the need to bring about a closer interaction between microfinance, agricultural extension and market channels to realise the full potential of each. Secondary effects refer to the wider impacts of particular kinds of safety nets: using public works programmes to link isolated households to markets or social services Addressing Vulnerability and Enhancing Productivity in South Asia 609 generates employment as well as improving future welfare and opportunities. Governments and donors are more likely to finance safety-net programmes if these wider developmental effects can be demonstrated. Finally, the importance of improved governance for various aspects of social protection measures is clear. Poor governance results in leakages, adds to the cost of safety nets, and transforms them into resources of political patronage and revenge, perpetuating pre-existing inequalities. Analysis points to the importance of efforts ‘from below’ to pressure for such reform. The South Asian context provides a variety of examples of collective bottom-up action which seeks to improve the terms of participation in markets and social life for the poor, as well as creating a political climate in which corrupt governance can be exposed.

References

Abdullah, A. and Murshid, K. A. S. (1986) The Distribution of Benefits from the Public Food Distribution System. BIDS Research Report No. 60. Dhaka: Bangladesh Institute of Development Studies. Adnan, S. (1990) Annotation of Village Studies in Bangladesh and West Bengal. A Review of Socio-economic Trends over 1942-88. Comilla: Bangladesh Academy for Rural Development. Agarwal, B. (1991) ‘Social Security and the Family: Coping with Seasonality and Calamity in Rural India’, in E. Ahmad, J. Drèze, J. Hills and A. Sen (eds), Social Security in Developing Countries. Oxford: Clarendon Press. Alamgir, M. (1980) Famine in South Asia: Political Economy of Mass Starvation. Cambridge, MA: Oelgeschlager and Hain Publishers. Amin, S. and Pebley, A. (1994) ‘Gender Inequality Within Households: The Impacts of a Women’s Development Programme in 36 Bangladeshi Villages’, Bangladesh Development Studies XXII (2 & 3): 121-55. Breman, J. (1995) ‘The Informal Sector Reconsidered’, Indian Journal of Labour Economics, 38 (3). Breman, J. (1996) Footloose Labour: Working in India’s Informal Economy. Cambridge: Cambridge University Press. Bruntrup, M., Alauddin, S. M., Huda, A. and Rahman, M. (1997) Impact Assessment of the Association for Social Advancement (ASA). Dhaka: ASA. Burgess, R. and Stern, N. (1991) ‘Social Security in Developing Countries: What, Why, Who and How?’, in E. Ahmed, J. Drèze, A. K. Sen and J. Hills (eds), Social Security in Developing Countries. Oxford: Clarendon Press. Cain, M. and Lieberman, S. S. (1982) Development Policy and the Prospects for Fertility Decline in Bangladesh. Centre for Policy Studies Working Paper No. 91. New York: Population Council. Crook, R and Manor, J. (1994) ‘Enhancing Participation and Institutional Performance: Democratic Decentralisation in South Asia and West Africa’. Report to ESCOR, ODA. London: Overseas Development Administration. Dev, M. and Suryanarayanan, M. H. (1991) ‘Is the PDS Urban-biased and Pro-rich?’, Economic and Political Weekly, 12 October. Development Alternatives Inc. (2000) Bangladeshi Experience in Adapting Financial Services to Cope with Floods: Implications for the Microfinance Industry. Bethesda, MD: Development Alternatives Inc. 610 Naila Kabeer

Drèze, J. (1990) ‘Poverty in India and the IRDP Delusion’, Economic and Political Weekly 25 (39): A95-104. Drèze, J. and Sen, A. (1991) ‘Public Action for Social Security: Foundations and Strategy’, in E. Ahmad, J. Drèze, J. Hills and A. Sen (eds), Social Security in Developing Countries. Oxford: Clarendon Press. Echeverri-Gent, J. (1988) ‘Guaranteed Employment in an Indian State. The Maharashtra Experience’, Asian Survey 28 (12): 1294-1310. Engkvist, R. (1995) Poverty Alleviation and Rural Development through Public Works: The Case of the Employment Guarantee Scheme in Maharashtra. Field Study Series No. 57. Department of Economics, University of Lund. Epstein, T. S. (1973) South India – Yesterday, Today and Tomorrow: Mysore Villages Revisited. London: Macmillan Press. Evans, T. G., Rafi, M., Adams, A. M. and Chowdhury, M. (1995) ‘Barriers to Participation in BRAC’. Dhaka: BRAC Research and Evaluation Division and the Harvard Centre for Population and Development Studies (mimeo). Franke, R. W. and Chasin, B. H. (1992) Kerala: Development through Radical Reform. San Francisco: Institute for Food and Development Policy. Ghani, A. (1999) Impact of Roads: A Qualitative Assessment. Dhaka: Care-Bangladesh. Gelbach, J. B. and Pritchett, L. (1997) ‘Indicator Targeting in a Political Economy: Leakier Can Be Better’. Washington, DC: World Bank (mimeo). Ghosh, Jayati (1998) Assessing Poverty Alleviation Strategies for their Impact on Poor Women: A Study with Special Reference to India. Discussion Paper No. 97. Geneva: UNRISD. Gough, K. (1981) Rural Society in Southeast India. Cambridge: Cambridge University Press. Greenough, P. R. F. (1982) Prosperity and Misery in Modern Bengal: The Famine of 1943-44. Oxford: Oxford University Press. Guhan, S. (1992) ‘Social Security in India: Looking One Step Ahead’, in B. Harriss, S. Guhan and R. H. Cassen (eds), Poverty in India. Research and Policy. New Delhi: Oxford University Press. Guhan, S. (1994) ‘Social Security Options for Developing Countries’, International Labour Review 133 (1): 35-53. Gunatilaka, R. (1999) ‘Rural Infrastructure Programmes for Poverty Reduction: Policy Issues from the Sri Lankan Experience’. Paper presented at the Regional Consultation on WDR 2001 for South Asia, Rajendrapur, Dhaka. Halder, S. and Husain, A. M. (1999) ‘Identification of the Poorest and the Impact of Credit on Them: The Case of BRAC’. Dhaka: BRAC Research and Evaluation Division (mimeo). Harriss, B. (1992) ‘Rural Poverty in India. Micro-level Evidence’, in B. Harriss, S. Guhan and R. H. Cassen (eds), Poverty in India. Research and Policy. New Delhi: Oxford University Press. Harriss-White, B (1999) ‘On to a Loser: Disability in India’, in Harriss-White and Subramanian. Harriss-White, B and Subramanian, S. (eds) (1999) Illfare in India: Essays on India’s Social Sector in Honour of S. Guhan. New Delhi, London and Thousand Oaks: Sage Publications. Hashemi, S. M. (1997) ‘Those Left Behind: A Note on Targeting the Hard-core Poor’, in Wood and Sharif. Addressing Vulnerability and Enhancing Productivity in South Asia 611

Hashemi, S. M., Schuler, S. R., and Riley, A. P. (1996) ‘Rural Credit Programs and Women’s Empowerment in Bangladesh’, World Development 24 (4): 635-53. Heller, P. (1996) ‘Social Capital as a Product of Class Mobilisation and State Interventions: Industrial Workers in Kerala and India’, World Development 24 (6): 1055-71. Holzmann, R and Jorgensen, S. (1999) ‘Social Protection as Social Risk Management: Conceptual Underpinnings for the Social Protection Sector Strategy’, Journal of International Development 11 (7): 1005-27. Hulme, D. and Mosley, P. (1996) Finance Against Poverty. London: Routledge. Hulme, D. and Mosley, P. (1997) ‘Finance for the Poor or Poorest? Financial Innovation, Poverty and Vulnerability’, in Wood and Sharif. Hyder, M. (1996) ‘From Relief to Development: Food-for-work in Bangladesh’, Disasters 20 (1): 21-33. Jenkins, R. (1999) Democratic Politics and Economic Reforms in India. Cambridge: Cambridge University Press. Jimenez, E. (1999) ‘Issues in Modernising Social Safety Nets’. Paper presented at the Manila Social Forum, Philippines, 9-12 November. Johnson, S. (2000) ‘Microfinance and Instability’. Centre for Development Studies, University of Bath (mimeo). Kabeer, N. (1991) ‘Gender and Rural Poverty: Analysis from Bangladesh’, Journal of Peasant Studies 18 (2): 241-62. Kabeer, N. (1998) Money Can’t Buy Me Love? Re-evaluating Gender, Credit and Empowerment in Rural Bangladesh. Brighton: Institute of Development Studies at the University of Sussex. Kabeer, N. (2001a) ‘Conflicts Over Credit: Re-evaluating the Empowerment Potential of Loans to Women in Rural Bangladesh’, World Development 29 (1): 63-84. Kabeer, N. (2001b) Making Rights Real: Nijera Kori, Social Mobilisation and the Empowerment of the Poor in Rural Bangladesh. IDS Working Paper. Brighton: Institute of Development Studies at the University of Sussex. Kabeer, N. and Murthy, R. K. (1997) Compensating for Institutional Exclusion? Lessons from Indian Government and Non-government Credit Interventions for the Poor. IDS Discussion Paper No. 356. Brighton: Institute of Development Studies at the University of Sussex. Khandker, S. R. (1998) ‘Micro-credit Programme Evaluation: A Critical Review’, IDS Bulletin 29 (4): 11-18. Khandker, S. R., Samad, H. A. and Khan, Z. H. (1998) ‘Income and Employment Effects of Micro-credit Programmes: Village Level Evidence from Bangladesh’, Journal of Development Studies 35 (2): 96-124. Lal, D. (1994) Labour Market Insurance and Social Safety Nets. HROWP 41. Washington, DC: World Bank. Lanjouw, P. and Ravallion, M. (1998) ‘Benefit Incidence, Public Spending Reforms and the Timing of Programme Capture’, World Bank Economic Review 13 (2): 257-74. Langworthy, M. (1999) Final Impact Assessment Report of IFFD Project. Dhaka: Care- Bangladesh. Lariviére, S., Matin, F., and Boily, M.-H. (1999) How the Poor Perceive Poverty in sub- Saharan Africa. Background Paper 1(a) for the Poverty Status Report for Africa. Matin, I., Hulme, D. and Rutherford, S. (1999) Financial Services for the Poor and Poorest: Deepening Understanding to Improve Provision. Finance and 612 Naila Kabeer

Development Research Programme Working Paper Series No. 9. Manchester: IDPM, University of Manchester. Maxwell, S. and Frankenberg, T. R. (1992) Household Food Security: Concepts, Indicators, Measurements. A Technical Review. New York: UNICEF and Rome: IFAD. Montgomery, R. (1996) ‘Disciplining or Protecting the Poor? Avoiding the Social Costs of Peer Pressure in Solidarity Group Micro-credit Schemes’, Journal of International Development 8 (2): 289-305. Mooij, J. (2000) Food and Power. The Political Economy of the Public Distribution System in Bihar. Indo-Dutch Programme on Alternatives in Development Occasional Papers and Reprints 2000-1. New Delhi: Indian Council of Social Science Research and The Hague: Netherlands Foundation for the Advancement of Tropical Research. Moore, M. and Joshi, A. (1999) ‘Between Cant and Corporatism: Creating an Enabling Political Environment for the Poor’, IDS Bulletin 30 (4): 50-59. Morduch, J. (1998) ‘Does Microfinance Really Help the Poor? New Evidence from Flagship Programs in Bangladesh’. Cambridge, MA: Harvard University and Stanford, CA: Hoover Institution, Stanford University (mimeo). Mosley, P. and Hulme, D. (1998) ‘Micro-enterprise Finance: Is there a Trade-off between Growth and Poverty Alleviation?’, in I. Matin and S. Sinha (eds), Recent Research on Microfinance: Implication for Policy. Brighton: Poverty Research Unit at the University of Sussex. Mundle, S. (1994) Deprivation and Public Policy: Redefining the Development State. The Wertheim Lecture 1994. Amsterdam: Centre for Asian Studies. Naved, R. (1994) ‘Empowerment of Women: Listening to the Voices of Women’, Bangladesh Development Studies XXII (2 & 3): 121-55. Nawani, N. (1994) ‘Indian Experience on Household Food and Nutrition Security’, in Regional Expert Consultation. Bangkok: FAO/UN. Nayar, R. (1995) ‘An Assessment of Wage-employment Programmes: JRY and EAS’, Indian Journal of Labour Economics 38 (3): 539-47. NIRD (1999) India Rural Development Report 1999: Regional Disparities in Development and Poverty. Hyderabad: National Institute of Rural Development. Noponen, H. (1990) ‘Loans to the Working Poor: A Longitudinal Study of Credit and the Household Economy’, International Journal of Urban and Regional Research 16 (2): 235-51. Papola, T. S. and Sharma, A. (1996) ‘Structural Adjustment, Poverty, Employment and Safety Nets: The Indian Experience’, The Indian Journal of Labour Economics 39 (3): 591-624. Patnaik, U. (1998) ‘Globalisation, Poverty and Food Security: Towards the New Millennium’. Paper presented at the conference of the Indian Association of Women’s Studies, Pune. Pitt, M. and Khandker, S. (1995) ‘Household and Intrahousehold Impacts of the Grameen Bank and Similar Targetted Credit Programs in Bangladesh’. Paper presented at workshop on ‘Credit Programs for the Poor: Household and Intrahousehold Impacts and Program Sustainability’ organised by the World Bank’s Poverty and Social Policy Department, Washington, DC and the Bangladesh Institute of Development Studies, Dhaka. Addressing Vulnerability and Enhancing Productivity in South Asia 613

Rahman, A., Razzaque, M. A., Rahman, M. M. S., Wadood, S. N., Yusuf, M. A. and Mondol, S. H. (1998) ‘On Reaching the Hard Core Poor: Redefining the NGO Strategy’. Dhaka: Bangladesh Institute of Development Studies (mimeo). Rahman, H. Z. (1995) ‘Crisis and Insecurity: The “Other” Face of Poverty’, in Rahman and Hossain. Rahman, H. Z. (1999) Poverty: The Challenges of Graduation. Dhaka: Bangladesh Institute of Development Studies. Rahman, H. Z. with Hashemi, S., Siddiqui, K. and Sen, B. (1998) ‘Poverty Issues in Bangladesh: A Strategic Review’. Dhaka: DFID (mimeo). Rahman, H. Z. and Hossain, M. (eds) (1995) Rethinking Rural Poverty: Bangladesh as a Case-study. London, New Delhi and Thousand Oaks: Sage Publications. Rahman, R. I. (1986) Impact of Grameen Bank on the Situation of Poor Rural Women. BIDS Working Paper No. 1: Dhaka: Grameen Evaluation Project, Bangladesh Institute for Development Studies. Rath, N. (1985) ‘Garibi Hatao: Can IRDP Do It?’, Economic and Political Weekly XX (6). Ravallion, M., Datt, G. and Chaudhuri, S. (1993) ‘Does Maharashtra’s Employment Guarantee Scheme Guarantee Employment? Effects of the 1988 Wage Increase’, Economic Development and Cultural Change 41 (2): 251-75. Ridao-Cano, C. (1999) ‘Child Labour in Bangladesh. Preliminary Results’. Department of Economics and Population Program, University of Colarado (mimeo). Ridao-Cano, C. (forthcoming) ‘Child Labour and Schooling in a Low-income Rural Economy’, Journal of Economic Literature. Sarin, M. (1999) ‘Should I Use My Hands for Fuel? Gender Conflicts in Joint Forestry Management’, in N. Kabeer and R. Subrahmanian (eds), Institutions, Relations and Outcomes: A Framework and Case Studies in Gender Aware Planning. London: Zed Press and New Delhi: Kali. Saxena, N. C. (1994) ‘Panchayats and Common Land Afforestation in India’, in A. Mukherjee (ed.), Decentralisation: Panchayats in the Nineties. New Delhi: Vikas Publishing House Pvt. Sen, A. (1995) ‘The Political Economy of Targetting’, in D. van de Walle and K. Nead (eds), Public Spending and the Poor: Theory and Evidence. Baltimore, MD: Johns Hopkins University Press. Sen, A. (1998) ‘Economic Reforms, Employment and Poverty. Trends and Options’. Paper presented at conference on Poverty: Emerging Challenges, Dhaka. Shariff, A. (1999) India Human Development Report: A Profile of Indian States in the 1990s. Delhi: Oxford University Press for the National Council of Applied Economic Research. Schuler, S. R., Hashemi, S. M., Riley, A. P. and Akhter, A. (1996) ‘Credit Programs, Patriarchy and Men’s Violence against Women in Rural Bangladesh’, Social Science and Medicine 43 (12): 1729-42. Shukla, S. C. and Kaul, R. (eds) (1998) Education, Development and Underdevelopment. New Delhi and London: Sage Publications. SIFAD (1989) Joint Government of Bangladesh/Donor Task Force on Strengthening the Institutions of Food Assisted Development Final Report (Vols. 1 and 2). Dhaka: Ministry of Planning, July. Subbarao, K., Bonnerjee, A., Braithwaite, J., Carvalho, S., Ezemenari, K., Graham, C. and Thompson, A. (1997) Safety Net Programs and Poverty Reduction. Lessons from Cross Country Experience. Washington, DC: World Bank. 614 Naila Kabeer

Suryanarayana, M. H. (1995) ‘PDS: Beyond Implicit Subsidy and Urban Bias – The Indian Experience’, Food Policy 20 (4): 259-78. Suwannarat, G. (2000) ‘Safety Nets in Asia: Addressing Exclusion, Challenging Orthodoxy’. Report to the Ford Foundation workshop on Safety Nets in Asia. Bangkok. Vasavi, A. R. (1999a) Harbingers of Rain. Land and Life in South India. Oxford: Oxford University Press. Vasavi, A. R (1999b) Agrarian Distress in Bihar. NIAS Report R5-99. Bangalore: National Institute of Advanced Studies. Vasavi, A. R, Vijaya, P.G., Chand, Sherry and Shukla, S. R. (1997) ‘Blueprint for Rural Primary Education: How Viable?’, Economic and Political Weekly, 13 December. Velaskar, P. (1998) ‘Ideology, Education and the Political Struggle for Liberation: Change and Challenge among the Dalits of Maharashtra’, in Shukla and Kaul. Whitehead, A. (1998) ‘Gender, Poverty and Intra-household Relations in Sub-Saharan Africa Smallholder Households: Some Lessons from Two Case Examples’. Background paper for 1998 Status of Poverty in Africa Report. Wood, G. D. (1994) Bangladesh: Whose Ideas, Whose Interests? London: Intermediate Technology Publications. Wood, G. D. and Sharif, I. A. (eds) (1997) Who Needs Credit? Poverty and Finance in Bangladesh. Dhaka: Dhaka University Press Ltd. World Bank (1991) India 1991: Country Economic Memorandum Vol. II. Washington, DC: World Bank, April. World Bank (1998a) Reducing Poverty in India. Options for More Effective Public Services. Washington, DC: World Bank. World Bank (1998b) Bangladesh: From Counting the Poor to Making the Poor Count. Washington, DC: World Bank. World Bank (1999) The India Poverty Project: Poverty and Growth in India, 1951-94. Washington, DC: World Bank. World Bank (2000) World Development Report 2000/2001. Washington, DC: Oxford University Press for the World Bank. Zaman, H. (1999) Assessing the Impact of Micro-credit on Poverty and Vulnerability in Bangladesh. Policy Research Working Paper 2145. Washington, DC: World Bank. Zohir, S. (2001) ‘Strengthening Linkages between Rural Credit, Agricultural Extension and Marketing: Bangladesh’, in ESCAP, Grass Roots Synergy in Rural Poverty Alleviation: Why and How? New York: United Nations.