Cash Transfers as the Silver Bullet for Poverty Reduction: A Sceptical Note by Jayati Ghosh
The current perception that cash transfers can replace public provision of basic goods and services and become a catch-all solution for poverty reduction is false. Where cash transfers have helped to reduce poverty, they have added to public provision, not replaced it. For crucial items like food, direct provision protects poor consumers from rising prices and is part of a broader strategy to ensure domestic supply. Problems like targeting errors and diversion from deserving recipients are likely to be even more pronounced with cash transfers and cannot be eliminated through technological fixes like the UID.
[Jayati Ghosh (firstname.lastname@example.org) is at the Centre for Economic Studies and Planning, Jawaharlal Nehru University, New Delhi.]
Over the past decade, cash transfers have become the latest fad of the international development industry, as the preferred strategy for poverty reduction. They are now being cited in many places as the solution to the problem of poverty. There has been a veritable deluge of books and articles hailing cash transfers as the most necessary, obvious and imperative strategy for poverty alleviation, even to the extent of suggesting that these should replace a great deal of other government activity. The title of a recent book says it all: Just Give Money to the Poor! (Hanlon et al 2010).
So what exactly is this strategy all about? In the recent international experience, cash transfers have been conditional (subject to the households meeting certain demands such as children attending school) or unconditional; targeted (given only to households or individuals meeting particular criteria) or universal. But essentially they amount to just what they sound like – the transfer of money to people by governments, rather than the provision of goods and services.
Some proponents of cash transfers tend to see this as a radically new idea newly thought up by governments in the South, which the international aid industry first fought and has now only grudgingly accepted. Thus Hanlon et al (2010) present the idea of cash transfers as “a development revolution from the South”: “an elegant southern alternative” that bypasses donors, governments and non-governmental organisations (NGOs) by empowering poor people to make their own decisions on what to spend the money on.
Despite claims of novelty, in fact this idea has a long history. Kautilya’s Arthasastra specified a system of taxation payments from the rich in order to enable transfer payments to the poor, including not only financial assistance during calamities but welfare payments to the chronically indigent and those unable to earn their own livelihood (Sen and Basu 2006). Islamic rulers in the Middle Ages were required to follow the tenets of zakat, using state revenues to provide income transfers for the poor, the elderly, orphans, widows and the disabled (Hamid 2003). Other historical examples abound.
The purpose of cash transfer schemes is to provide poor people with money and give them the freedom to choose what to do with it. Of course, this then generates other choices that have to be made: Who gets the transfers? How much do they get? In most countries that employ them, cash transfer schemes typically account for between 1% and 2% of GDP. If they are universal that usually spreads the money around rather thinly, so they account for very little. But if they are targeted, then the familiar problems of targeting (unfair exclusion, unjustified inclusion, large administration costs, possibilities of leakage and diversion) all become significant. If they are to be effective at all, cash transfers have to be assured, relatively easy to deliver and monitor and large enough to affect household income. But this also means that they have to be reasonably significant chunks of public spending. And this begs the question of what expenditures they are replacing.
Several of the more well-known recent “success stories” involve targeting and conditions on recipients that range from light to onerous. Brazil’s Bolsa Familia is a grant provided to families with less than a threshold monthly income, with the requirement of attendance at government clinics and 85% school attendance. The Oportunidades programme in Mexico is a highly conditional cash transfer system based on a complex system of eligibility (age, gender and level of education of each family member, access to electricity and tap water, household assets) and requiring family members, especially mothers, to meet various time-intensive conditions like attending meetings and providing “voluntary” community labour. In both cases, as noted below, they have added to pre-existing public delivery systems for health, nutrition and education, which have also expanded over the same period.
There can be no argument about whether progressive redistributive transfers are desirable. Indeed, redistribution is a major, even critical element of any fiscal system of taxation and public expenditure. Minimum income schemes for the destitute, pension payments for the elderly, child support grants, unemployment benefits and other forms of social protection are obviously desirable in themselves and constitute requirements for any civilised society, even the poorest one. They also contribute in the short term to more effective demand and therefore have positive multiplier effects, and in the long term to healthier, better educated and more secure populations.
So the question then is not whether or not to oppose cash transfers in general, but rather what specific importance to give them in an overall strategy of development and poverty reduction. Cash transfers cannot and should not replace the public provision of essential goods and services, but rather supplement them. In other words, cash transfers are desirable and can play a positive redistributive role if they are additional to other – and essential – public expenditure on essential goods and services that are required by all citizens, including the poor.
However, the current tendency is to see this as a further excuse for the reduction of publicly provided services, and replace them with the administratively easier option of doling out money. In many countries, the argument has become one of encouraging governments to give the poor cash transfers that will allow them to access whatever goods and services they want that are generated by private markets, rather than struggling to ensure public provision.
Such a position completely misses the point, which is that all the success stories of cash transfer programmes have been where they have come as additions to existing systems and structures of public provision. Indeed, many of them have also been predicated on the improvement of such public delivery systems.
In Brazil, for example, Bolsa Familia can be based on minimum school attendance only because there are enough public (and free) schools of reasonable quality that children of poor households can attend, which, in turn, means prior and continuing public investment in quality schooling and teacher education. Most reviews of the programme note that its success in increasing enrolment and attendance would not have been possible – or even meaningful – without educational policies designed to ensure concomitant improvements in the performance of children in schools. Similarly the programme can require compulsory attendance at health clinics because such clinics exist, are accessible and of reasonable quality, based on public health expenditure that runs at around 8% of GDP. This is important, because providing small amounts of cash to allow people to visit local private quacks will hardly compensate for the absence of a reasonably well-funded public health system that provides access to preventive and curative services. Also, cash transfers are obviously less effective in periods of rising prices of essential goods and services.
As Son (2008: 3) notes,
A presumption embedded in the CCT (conditional cash transfer) approach is that the supply of social services for education and health is in place and that stimulating demand through income transfers is necessary to induce major changes in human capital investment. This explains why the success of CCT programmes in some countries is no guarantee that they can be reproduced in others with the same performance. For instance, in many developing countries, children, particularly in rural areas, face supply-related problems, i e, there are not enough schools, classrooms, or teachers to offer adequate education to those who need or want them. In such circumstances, pouring resources into a CCT programmes may not be able to achieve the educational objective.
This is important, because ultimately social and economic policies are all about choices, and this is most starkly evident in the allocations of public expenditure. Governments typically do not have the luxury of being able to ensure enough spending to provide good quality public services and provide cash transfers that are large enough to be at all meaningful. So the decision to provide cash to designated beneficiaries may effectively imply a reduction of some other expenditure, even though this need not be the case and is clearly not what is desirable. Cash transfers are obviously advantageous in themselves, but insofar as they imply an explicit or implicit reduction of other public expenditure, these trade-offs need to be considered very seriously.
It should be further noted that cash transfers cannot be a solution to the problem of persistent poverty, simply because they do not in any way address the basic causes of poverty. The poor are defined not just by low income but more fundamentally by their lack of assets, which in turn is related to the concentration of assets somewhere else in the same society. Similarly their economic position or occupation is crucial, which is why they are usually those who are workers engaged in low paying occupations (whether as wage workers or self-employed) or those who are unable to find paid jobs or those whose existence and livelihood depend upon fragile ecologies in which survival is insecure and fraught with difficulty. Addressing poverty therefore requires structural changes: in asset distribution, in access to productive employment, in reduced vulnerability to environmental shocks, and so on. And so governments that are genuinely concerned with poverty reduction have to orient their policies and expenditure to such structural changes. Cash transfers may appear as (welcome) temporary palliatives in such a context, but they simply cannot be treated as a development panacea.
The Indian Discussion
Indian policymakers only recently woke up to this latest silver bullet for development, but now they are busy catching up. The idea was mooted in the Government of India’s Economic Survey for 2009-10. There have been subsequent arguments that the public distribution scheme for basic food items should be replaced with a system of coupons or cash transfers (Basu 2010). It is increasingly evident that the agenda of the United Progressive Alliance (UPA) government is to bring in cash transfers to replace public distribution of various essential items, including food. To begin with, Finance Minister Pranab Mukherjee has proposed that the existing system of subsidies for kerosene and fertilisers be done away with and replaced by direct cash transfers to chosen beneficiaries. In his Budget Speech for 2011-12, he noted,
The Government provides subsidies, notably on fuel and foodgrains, to enable the common man to have access to these basic necessities at affordable prices. A significant proportion of subsidised fuel does not reach the targeted beneficiaries and there is large scale diversion of subsidised kerosene oil... To ensure greater efficiency, cost effectiveness and better delivery for both kerosene and fertilisers, the Government will move towards direct transfer of cash subsidy to people living below poverty line in a phased manner.
This follows on other moves, for example by the Government of Delhi, which has embarked on a pilot project to replace foodgrain distribution under the public distribution system (PDS) with cash transfers.
There are two immediate problems that are evident in this approach. First, what ensures that the amount of the transfer will be sufficient to fully compensate for any price increases in the newly deregulated markets of these goods? Second, how will the government ensure that the cash transfer actually goes to those who are the intended beneficiaries?
Cash versus Kind
In the case of choice between direct public provision of some essential goods (like food and fuel) and cash transfers to consumers, the most immediate threat is that the rising prices in these deregulated markets will make such goods unaffordable for those who need them most. In situations of volatile and rising prices, the real value of cash transfers can get quickly eroded, and the systems of price indexation of such transfers are typically slow and inadequate to cover the price increases. This may explain why poor people in general prefer public provision of the good or service in question at a defined price, when it is of reasonable quality.
It is often argued that cash transfer systems can simply be indexed to price indices (for example in the case of food items, to the price index for the foods in question) to get around this problem. But anyone familiar with the lags in public response to price changes – for example in the setting of minimum wages, and more recently in the determination of the national wages under the Mahatma Gandhi National Rural Employment Guarantee Scheme (MNREGS) – will immediately understand that this is an excessively optimistic, even utopian, assumption in the Indian context. To revert to the food example once again, in a context of the high food inflation rates of between 15% and 20% per year experienced in India in the recent past, even within half-yearly revisions to the amount of a predefined transfer, there would be considerable effective loss to the designated beneficiaries.
It is to be noted that the poor are more likely to prefer provision in kind than those who are better off. An International Food Policy Research Institute study comparing various cash and kind (food) transfer programmes in Bangladesh found that “as income increases, beneficiaries’ preference for food declines, indicating that the poorest households prefer only food as the transfer. Conversely, relatively better-off beneficiaries tend to prefer only cash” (Ahmed et al 2009: 72, emphasis added).
A recent survey in Delhi found something similar: that slum women overwhelmingly (99%) preferred receiving food rations rather than cash transfers or coupons for a specified value, at least partly because of the fear of food price inflation that would erode the value of the cash transfer (Parsai 2011). The 593 women from 14 slums interviewed in this survey conducted by a group of NGOs including the Right to Food Campaign also feared that cash receipts could get spent on other household priorities or immediate needs of the moment, be it for a health emergency or a celebration, if not on liquor, etc. They therefore wanted a strengthened PDS that functions well, preferably run by self-help groups or cooperatives. So their desire was for a strengthened, improved and accountable system of distribution of the actual food commodity.
The Delhi survey also highlights a gender dimension with respect to the utilisation of cash within a household that is inadequately understood. Structures of power within households as well as social constructions of gender behaviour can affect decisions about how the money is spent, in ways that are not always expected or desired. It has been argued that handing over the cash payments directly to women will solve this problem, but this is not necessarily the case. Especially with respect to food, it has been found (particularly in south Asia) that women and girls are guilty of voluntary self-denial rather than being forced into choices that reduce their own consumption. So the possibility of cash transfers being used for other forms of expenditure that do not meet the intended purpose clearly exists. This is one reason why a number of women’s organisations in India have actively campaigned against cash transfers and argued instead for expanded and improved direct provision of goods and services by more accountable public delivery systems.
Posing the problem only in terms of whether the poor prefer cash or kind can also be misleading, however, because it completely leaves out the feasible and much more just alternative of universal provision of some essential items. This relates to the difficulties inherent in targeting, which are considered below.
The Issue of Targeting
Targeting of any particular scheme involves various costs and difficulties. To begin with, there are the administrative costs of targeting versus universal provision. These involve the costs of identifying the targeted beneficiaries within the population and ensuring that this targeting remains correct over time. They involve further costs of administering the benefits in such a way that all and only the chosen beneficiaries receive the benefits. There are many case studies that suggest that where “targeting” of the poor is successful, it involves systems of such administrative complexity that they involve huge expenditures of funds and personnel simply to enable the targeting. That is why schemes that are inherently self-selecting, such as employment programmes that the better-off will not be interested in, involve much lower costs and are actually more effectively targeted.
The very process of identification of the target group – say, “the poor” – can generate some well-known errors: Type I errors of unjustified exclusion of the genuinely poor and Type II errors of unwarranted inclusion of the non-poor. These are not simply mistakes that can occur in any administrative scheme, they are built into systems that try to provide scarce goods or services to one section of any population. In the case of cash transfers, of course, the possibility of errors can be even higher than in the case of delivery of a particular good (such as food) or service (such as health facilities), since there is little incentive for self-selection out of such a programme.
Prevalence of Exclusion/Inclusion Errors
Conditionalities imposed on beneficiaries are sometimes used as a substitute, or an implicit means of ensuring that the targeting has been correct, but this has had at best mixed results. For example, in the most widely quoted success stories of cash transfer described above, Oportunidades in Mexico and Bolsa Família in Brazil, both types of error remain high. An evaluation by the International Poverty Centre of the United Nations Development Programme (UNDP) showed that 70% of the poor were excluded from the benefits of Oportunidades, and 59% of the poor were excluded in Bolsa Família. In contrast, Bolsa Famíliawas found to have a higher inclusion error than Oportunidades: 49% of all beneficiaries are non-poor in the former programme but only 36% are non-poor in the latter (Soares, Ribas, Veras and Osorio 2007: 2-3).
Most significant of all, such errors are likely to be especially high where structures of power and inequity already deprive certain sections of the population of basic goods and allow powerful elites to appropriate goods and services that are provided in limited quantity. In hierarchical and discriminatory societies like India, where social and economic power is unequally distributed, it requires no imagination to realise that making a scarce good (for example, cheap food) supposedly available only to the poor is one of the easiest ways to reduce their access instead of increasing it. The problem is well known in India, where all public delivery systems have some element of leakage and diversion. It is absurd to imagine that providing cash rather than goods will solve this problem: how much simpler and easier it will be for those who benefit from such leakages to divert cash, rather than goods that have to be stored and resold!
In the case of essential commodities like food, there is an additional problem, which is that of establishing appropriate criteria for identifying beneficiaries. The current system of targeting in India is based on income or poverty or some other set of variables that seek to define poor households. But there is a distinction between food insecurity and poverty as currently defined (either income or other terms). It is evident from the National Sample Survey Organisation and National Family Health Survey information that the proportion of the population that is nutritionally deprived is significantly larger than the “poor” population, and in many states they are not completely overlapping categories either. To deal with food insecurity in an effective manner, it is counterproductive to base public food provision on a predefined group of the “poor”, which would deprive a large number of others who are also food-insecure. Given that entitlements are household based, the gender dimension of the problem is hugely important as well, since limiting provision to poor households is likely to deprive particularly women and girls in so-called “non-poor” households of access to adequate nutrition.
This is related to a third problem, the absence of any notion of dynamics in a rigid law that defines “poor” and “vulnerable” households in a static sense and changes the group only at infrequent intervals. Households – and people within them – can fall in or out of poverty, however defined, because of changing material circumstances. Similarly they can also go from being food-secure to food-insecure in a short time. The reasons can vary: crop failures, sharp rises in the price of food, employment collapses, health issues that divert household spending, the accumulation of debt, and so on. Monitoring each and every household on a regular basis to check whether any of these or other features has caused it to become food-insecure is not just administratively difficult, it is actually impossible.
Universal Access Crucial
This is why all successful programmes of public food distribution, across societies, have been those that have gone in for universal or near universal access. This provides economies of scale; it reduces the transaction costs and administrative hassles involved in ascertaining the target group and making sure it reaches them; it allows for better public provision because even the better off groups with more political voice have a stake in making sure it works well; it generates greater stability in government plans for ensuring food production and procurement. Even among the states of India, those that have a better record of public food distribution are those that have gone in for near-universal access (Himanshu and Sen 2011). Kerala, Tamil Nadu and Andhra Pradesh all have defined the poor in such an inclusive way that the vast majority of the population is included, which makes their schemes universal or close to universal.
Indeed, comparative studies of many countries have found that in general, universal provision of goods and services is much more likely to provide access to the poor (as noted in Mkandawire ed 2006). One important reason is that universal provision generates a middle class constituency that is concerned with ensuring good quality provision. Since the middles classes tend to have greater social and political voice than the poor (especially in countries like India) this acts as a powerful push for accountability.
Moreover, when it comes to basics such as provisioning of minimum food and fuel to very large populations, the problem itself is universal. (This is evident in the way all recent episodes of global inflation in these items have been mistakenly attributed to the growth of China and India.) The issue of domestic supply becomes absolutely crucial in the context of increasingly volatile and rising global prices of these essential items, as even periodic reliance on imports is problematic. This means that we need to ensure basic national self-sufficiency in food, while conserving natural resources ensuring that production occurs in a sustainable manner. Obviously, to achieve this more important goal, it is simply not good enough just to transfer cash to poorer consumers and hope that markets will take care of supply. That is why the producer incentives contained in food and fertiliser subsidies, the costs of buffer stocks and implications of demand for firewood if fuel prices rise too high, will all need to remain areas of public policy and provisioning. There is clearly a huge amount of public intervention required to manage these supply-side issues. Thinking of cash transfers as a magic bullet is likely to detract from necessary efforts in this direction, rather than provide any solution.
Targeting and the Technological Solution
Despite such overwhelming arguments and evidence in favour of universal provision and against targeting, the current tendency in India is to move away from universal provision as citizen’s right, to targeted systems based on criteria of often alarming complexity. At present, Indian policymakers appear to be veering around to the conclusion that such problems related to can be solved by the technological fix of providing all citizens with unique (and biometrically based) identification numbers, in the proposed unique identification (UID) scheme. It is now regularly argued by various official spokesmen that problems of corruption and leakage will be reduced or eliminated by utilising biometric UIDs for delivery. This is a chimera, and it is likely to become a very expensive chimera as well.
The UID – whatever other benefits it may provide – is at best a determinant of identity (though there are questions about its viability and the robustness of biometric indicators in very large and diverse populations like that of India). But unique identity is only the first step. The determination of the targeted beneficiaries – on the basis of poverty or any other indicator – still has to be done and this process still remains subject to all the errors of unfair exclusion and unjustified inclusion that have been described. These are socio-economic decisions that are affected by a complex set of political and social forces as well as power relations. Further, even after identification, the actual delivery to the targeted group is an area that is heavily fraught and prone to diversion of resources. Since it is so much easier to divert cash than a physical good or a social service, the power equations that enabled such diversion are likely to ensure that such diversion is even greater in extent.
Technology simply cannot address this fundamental problem. For example, we know that NREGS workers in certain districts of some states have been found to have handed their job cards (and thereby at least part of their wage payments) to contractors or locally powerful agents. This problem has persisted even in situations where the workers are meant to be receiving their wages through bank accounts that supposedly ensure against such diversion. So we can imagine that forcing poor people to put down their fingerprints in return for only a part (if that) of their cash transfer payment would be a relatively easy thing to ensure, at least in those areas. This entire structure requires very different responses that are much more based on social mobilisation and public demands for accountability. This is a messy and dangerous process, as the many brave social activists engaged in such action know to their cost, and lives have been lost in this endeavour. To think that technology will allow such power relations to vanish into thin air is ludicrous.
All this should not take away from the positive role that cash transfers can play in income redistribution, especially when they are additional to other necessary public spending. Clearly, as was recognised long ago in the Arthasastra and in the practice of zakat, cash transfers may in any case be socially necessary and can play an important role for those like the elderly, the disabled or children who are not in a position to earn incomes and must depend on others for material survival. Moreover, there is a logical core to the more modern idea that in certain cases it may be necessary to increase user charges for publicly provided goods and services and make universal compensatory cash transfers, rather than to keep prices indefinitely repressed at the expense of quality when the costs of public provisioning increase. But such a strategy cannot and should not be at the cost of universal public provisioning at generally affordable rates, whether in the case of health or education or in case of food or fuel security.
Nor should any of this make us lose our focus on the primary causes of poverty: asset inequality both nationally and internationally and the absence of positive structural change that provides more productive employment opportunities to the population in general. Dealing with those should still remain the primary concern of those who are genuinely concerned about poverty reduction and a life of dignity for all.
Ahmed, Akhter U, Agnes R Quisumbing, Mahbuba Nasreen, John F Hoddinott and Elizabeth Bryan (2009): Comparing Food and Cash Transfers to the Ultra Poor in Bangladesh, IFPRI Research Monograph No 163 (Washington DC: IFPRI).
Basu, Kaushik (2010): “The Economics of Foodgrain Management in India”, Ministry of Finance Working Paper no 2/1010, http://www.finmin.nic.in/WorkingPaper/index.asp accessed on 2 March 2011.
Hamid, Shadi (2003): “An Islamic Alternative? Equality, Redistributive Justice, and the Welfare State in the Caliphate of Umar”, Renaissance: Monthly Islamic Journal, Vol 13:8.
Hanlon, Joseph, Armando Barrientos and David Hulme (2010): Just Give Money to the Poor: The Development Revolution from the Global South (Sterling, Virginia, USA: Kumarian Press).
Himanshu and Abhijit Sen (2011): “Why Not a Universal Food Security Legislation?”, Economic & Political Weekly, Vol XLVI, No 12: 36-47.
Mkandawire, Thandika, ed. (2006): Social Policy in a Development Context (Palgrave: Macmillan and UNRISD).
Parsai, Gargi (2011): “Survey in Delhi Finds Slum Women Prefer Rations to Cash Transfers”, The Hindu newspaper, 7 April, http://www.hindu.com/2011/04/17/stories/2011041764571000.htm accessed on 19 April 2011.
Sen, Raj Kumar and Ratan Lal Basu, ed. (2006): Economics in Arthasastra (New Delhi: Deep and Deep Publications).
Soares, Fábio Veras, Rafael Perez Ribas, Fábio Veras and Rafael Guerreiro Osório (2007): “Evaluating the Impact of Brazil’s Bolsa Família: Cash Transfer Programmes in Comparative Perspective”, IPC Evaluation Note, No 1, December (Brasilia: International Poverty Centre).
Son, Hyun H (2008): “Conditional Cash Transfer Programs: An Effective Tool for Poverty Alleviation?”, ERD Policy Brief Series, No 51 (Manila: Asian Development Bank).
Tagged with: Cash Transfers