Continuing the debate on direct cash transfers, the authors of the article "The Case for Direct Cash Transfers to the Poor" (12 April 2008) respond to Mihir Shah's criticism (23 August 2008). The six points of contestation by Mihir Shah - including those on the public distribution system and the National Rural Employment Guarantee Scheme - are refuted. The argument in essence is that seeing the problems with anti-poverty programmes as faulty design and limited availability of resources does not recognise the culture of immunity in public administration and the weak capabilities of local governments.
DISCUSSIONEconomic & Political Weekly EPW november 22, 200885More on Direct Cash Transfers Devesh Kapur, Partha Mukhopadhyay, Arvind SubramanianDevesh Kapur (dkapur@sas.upenn.edu) is with the Centre for the Advanced Study of India, University of Pennsylvania, US; Partha Mukhopadhyay (partha@cprindia.org) is at the Centre for Policy Research, New Delhi; Arvind Subramanian (asubramanian@petersoninstitute.org) is with the Perterson Institute for International Economics, the Centre for Global Development, and Johns Hopkins University, US.Continuing the debate on direct cash transfers, the authors of the article “The Case for Direct Cash Transfers to the Poor” (12 April 2008) respond to Mihir Shah’s criticism (23 August 2008). The six points of contestation by Mihir Shah – including those on the public distribution system and the National Rural Employment Guarantee Scheme – are refuted. The argument in essence is that seeing the problems with anti-poverty programmes as faulty design and limited availability of resources does not recognise the culture of immunity in public administration and the weak capabilities of local governments.In his response to our paper, Kapur, Mukhopadhyay and Subramanian (12 April 2008; pp 37-43, henceforthKMS), Mihir Shah (23 August 2008, pp 77-79, henceforthMS) makes several claims. One is thatKMS is hitherto the best articulation of the idea of direct cash transfers (DCTs). This we can find no fault with! His other claims can be summarised asfollows: (a) existing centrally-sponsoredpoverty programmes withDCT features suchasthe Indira Awas Yojana (IAY)and the Swarna-jayanti Gram Swarozgar Yojana(SGSY) have not worked; (b) making DCT effective requires complementary inputs; (c) panchayati raj institutions (PRIs) are weak and need support; (d) the public dis-tribution system (PDS) needs to be reformed instead ofDCTs that help the poor tobuy grain in the market; and finally, (e) viewing the National Rural Employment Guarantee Scheme (NREGS) as a cash transfer pro-gramme is a mistake.MS’ overall message is thatDCTs are “no magic bullet”. Before we address each of these specific issues, we would make two broad obser-vations. First, any DCT scheme of the sort that we proposed is, almost by definition, not intended to be a “development” project in the Amartya Sen sense of allowing people to attain a broad set of economic and other capabilities. It has a much narrower scope in providing the income to cross the (admittedly arbitrary) poverty threshold and is hence closer to a welfare scheme. Our main contention is that direct transfers deserve to be tried as an alter-native way of meeting this objective given the demonstrated failure of so many poverty schemes to do so. Second, it is striking that while commen-tators such as Mihir Shah acknowledge the many failures of government efforts over long periods of time to provide the inputs required to generate sustainable develop-ment, they continue to want to rely on the same, likely deteriorating, state capacity to provide these inputs without offering any argument as to why future efforts are likely to succeed while past ones have failed. The case for DCTs is precisely to minimise on such capacity so that there is a better chance (than the failed status quo) of meeting the objective of helping poor people. The onus is very much on the critics of cash transfers to make the case for the continuation of numerous existing poverty programmes which have failed for nearly half-century or so.(a) IAY and SGSY Are Cash Transfers That Do Not Work: MS claims that the “starkest way of illustrating the point is the IAY. KMS would have us believe that IAY would work if direct cash transfers were to be adopted. But the IAY is already based on direct cash transfers.” MS’ use of these examples illustrates a basic misconception about the KMS approach. In essence, we are not about making IAY or any of the other schemes work. These candidate schemes mentioned in KMS and by MS were selected because all of them are about transferring private goods to identifiedbeneficiaries. They are examples to illustrate the extent of possible replacement and emphasise that DCT should not be in addition to the existing programmes. Indeed, MS’ comments on IAY reinforce our point – households cannot be forced to use money in a manner that reflects the priorities of planners over their own preferences. Poor households may well, as MS notes, “have other needs that gain priority over housing”. This is precisely our argument for DCT, viz, whether the person wishes to use the cash transfer to build a house or engage in a livelihood should be that person’s choice. Likewise, ifSGSY were a cash transfer, all bank loans could be cash transfers. Of course, it is axiomatic that if the loan does not have to be repaid, it does become a cash transfer. But ex ante only the power-ful and connected will know that it is a cash transfer and treat it as such, while others will treat it as a loan. For the poor, SGSY, as currently structured, can have the same negative consequences as loan pushing. We contend that an objective that cannot, ex ante,be universal in prin-ciple, should not be a part of a categorical equity plan.
DISCUSSIONnovember 22, 2008 EPW Economic & Political Weekly86(b) Making Cash Transfers Effective Requires Complementary Inputs: MS says:What KMS forget is that any anti-poverty programme will work only if it leads to an end to dependence on doles (what they term in more glorified terms as direct cash trans-fers). To end this dependence requires cre-ating sustainable livelihoods for the poor. And these presuppose much more than transfer of cash. They demand skills, mar-kets, technology, material inputs, infra-structure and institutions. It is these that are in deficit in rural India, more than cash. And it is their lack that accounts for failure of anti-poverty programmes, as much as leakages of cash (the exclusive concern of KMS)…It is much more about ensuring ef-fective utilisation of this cash, which needs both developmental inputs (markets, technologies, skills and materials) and political ones (social mobilisation to set up monitoring mechanisms/institutions).It is gratifying to note that MS shares KMS’ assessment that existing anti-poverty programmes have failed. The six argu-ments inKMS for DCT were that it would (a) “expand their choices and eliminate… paternalism”; (b) “relieve financial con-straints faced by the poor”; (c) reduce “administrative costs”; (d) arrest the “grow-ingimmunity in public administration”; (e) remove the “inherent inequity in some of the subsidies”; and finally, (f) “clien-telism, patronage and corruption that attendCSS would be reduced”. We clearly stated that “Cash transfers work only with a well-functioning private distribu-tion system, which is not present for a number of services. These are functions that could be done by the local govern-ment and is their responsibility in many countries.” Surely, this does not support MS’ contention that KMS is exclusively concerned with “leakages of cash”, and that it forgets that anti-poverty pro-grammes will require “developmental… and political” inputs.The issue about moving away from transfers to identified beneficiaries to investing in physical and social infra-structure arises in the context of a budget constraint. However, to the extent that such transfers continue, our point is simple: compared to the scores of anti-poverty programmes that India has tried, DCTs are likely to improve the welfare of the poor to a greater extent by increasing their purchasing power to access goods and services, many of which the state should be providing, but is not doing. Yes, markets in rural India are imperfect – but are less so than a few decades earlier and the private sector has demonstrated considerable supply response. Markets, technology, and material inputs already exist, albeit with a huge variance in qual-ity. Connectedness to these inputs is pro-vided by infrastructure, i e, roads, com-munication and transport and many of these are provided by the private sector, e g, local para-transit services and cell-phones, though not always at prices affordable to the poor.Instead of belabouring the need for complementary inputs, we need to focus on the mechanisms and the institutions that will provide them. We are not advo-cating that the government get out of areas where it should provide public goods, e g, rural roads. However, the level of government is a question that needs to be asked – whether rural roads should be financed through a central scheme or through an appropriately fundedPRI’s budget. Whether a decentralised public sector (in the form of cash empowered PRIs) can deliver has yet to be established, but surely they need to be given the chance to try for themselves, where centralised mechanisms have failed.KMS addressed this in an admittedly briefer part of the paper, and it is to this we now turn.(c) PRIs Need Support: MS considers “PRIs in large parts of India today” to be “nothing more than work-in-progress. They have a very long way to go before they can become instruments of democracy and development at the grass roots. They need massive support from the state to be able to realise their potential. This is the whole unfinished agenda of reform of rural governance, the reform of the public sector in rural development…It is patently unfair to burdenPRIs with massive tasks of development without providing them the requisite support.”It appears that MS continues to believe in a top-down approach where well-intentioned agents of the state, with the requisite skills and knowledge, are ready to selflessly work to improve PRIs. Of course,PRIs are weak. But that is akin to saying that because a baby will fall the first time she walks, she should not walk. What is the form of “requisite support” needed by thePRIs now? Elsewhere MS complains that “social mobilisers and technical personnel…have not been supplied” toPRIs. But why not instead give PRIs control over the funds needed to hire the requisite functionaries for the purpose? Hence,KMS argue for an increase in transfers toPRIs “in a few functional block grants” with “an effec-tive system of financial accountability” commensurate with their responsibilities. As KMS argued, the least bad way of supporting PRIs is, provide them with “the resources and capacity to deliver good operational performance” and thereby transform them into places of real political contestation.We now turn to two specific schemes discussed byMS, viz,PDS andNREGS. (d) The Reach of PDS:MS says that it is not clear how transfers of cash to the poor would allow them to buy grain from the open market at a time of steep inflation. The problem is that India’s PDS is characterised by a whole range of inequities – its coverage is the weak-est in the neediest regions and it fails to cover crops grown and eaten by the poorest. The way forward is to reform the PDS and extend its reach to and density in the poorest parts of the country, where the need is the greatest.Given the long and egregious mis-management ofPDS, MS’ call to “reform thePDS” is puzzling. In contrast, we believe that if PDS is to be at all reformed, a smaller PDS focused on its core mission has a better chance. For the vast majority of locations, we maintain that the open market is a viable distribution channel. In this, the price of grain is immaterial. If open market prices rise, the food subsidy should rise concomitantly. Replicating this in aDCT would involve indexing a part of it to food prices. The critical benefit of moving to DCT is the reduction in logistics and distribution costs, which can be then used either to increase theDCT or on “developmental inputs” dear toMS, or even to reduce the fiscal deficit, as appropriate at the time. As for choice of grain, if the poor really have different preferences (and are not being forced by lack of resources), will notDCTs increase their purchasing power in the market, and thereby increase the supply of such grain?
DISCUSSIONEconomic & Political Weekly EPW november 22, 200885More on Direct Cash Transfers Devesh Kapur, Partha Mukhopadhyay, Arvind SubramanianDevesh Kapur (dkapur@sas.upenn.edu) is with the Centre for the Advanced Study of India, University of Pennsylvania, US; Partha Mukhopadhyay (partha@cprindia.org) is at the Centre for Policy Research, New Delhi; Arvind Subramanian (asubramanian@petersoninstitute.org) is with the Perterson Institute for International Economics, the Centre for Global Development, and Johns Hopkins University, US.Continuing the debate on direct cash transfers, the authors of the article “The Case for Direct Cash Transfers to the Poor” (12 April 2008) respond to Mihir Shah’s criticism (23 August 2008). The six points of contestation by Mihir Shah – including those on the public distribution system and the National Rural Employment Guarantee Scheme – are refuted. The argument in essence is that seeing the problems with anti-poverty programmes as faulty design and limited availability of resources does not recognise the culture of immunity in public administration and the weak capabilities of local governments.In his response to our paper, Kapur, Mukhopadhyay and Subramanian (12 April 2008; pp 37-43, henceforthKMS), Mihir Shah (23 August 2008, pp 77-79, henceforthMS) makes several claims. One is thatKMS is hitherto the best articulation of the idea of direct cash transfers (DCTs). This we can find no fault with! His other claims can be summarised asfollows: (a) existing centrally-sponsoredpoverty programmes withDCT features suchasthe Indira Awas Yojana (IAY)and the Swarna-jayanti Gram Swarozgar Yojana(SGSY) have not worked; (b) making DCT effective requires complementary inputs; (c) panchayati raj institutions (PRIs) are weak and need support; (d) the public dis-tribution system (PDS) needs to be reformed instead ofDCTs that help the poor tobuy grain in the market; and finally, (e) viewing the National Rural Employment Guarantee Scheme (NREGS) as a cash transfer pro-gramme is a mistake.MS’ overall message is thatDCTs are “no magic bullet”. Before we address each of these specific issues, we would make two broad obser-vations. First, any DCT scheme of the sort that we proposed is, almost by definition, not intended to be a “development” project in the Amartya Sen sense of allowing people to attain a broad set of economic and other capabilities. It has a much narrower scope in providing the income to cross the (admittedly arbitrary) poverty threshold and is hence closer to a welfare scheme. Our main contention is that direct transfers deserve to be tried as an alter-native way of meeting this objective given the demonstrated failure of so many poverty schemes to do so. Second, it is striking that while commen-tators such as Mihir Shah acknowledge the many failures of government efforts over long periods of time to provide the inputs required to generate sustainable develop-ment, they continue to want to rely on the same, likely deteriorating, state capacity to provide these inputs without offering any argument as to why future efforts are likely to succeed while past ones have failed. The case for DCTs is precisely to minimise on such capacity so that there is a better chance (than the failed status quo) of meeting the objective of helping poor people. The onus is very much on the critics of cash transfers to make the case for the continuation of numerous existing poverty programmes which have failed for nearly half-century or so.(a) IAY and SGSY Are Cash Transfers That Do Not Work: MS claims that the “starkest way of illustrating the point is the IAY. KMS would have us believe that IAY would work if direct cash transfers were to be adopted. But the IAY is already based on direct cash transfers.” MS’ use of these examples illustrates a basic misconception about the KMS approach. In essence, we are not about making IAY or any of the other schemes work. These candidate schemes mentioned in KMS and by MS were selected because all of them are about transferring private goods to identifiedbeneficiaries. They are examples to illustrate the extent of possible replacement and emphasise that DCT should not be in addition to the existing programmes. Indeed, MS’ comments on IAY reinforce our point – households cannot be forced to use money in a manner that reflects the priorities of planners over their own preferences. Poor households may well, as MS notes, “have other needs that gain priority over housing”. This is precisely our argument for DCT, viz, whether the person wishes to use the cash transfer to build a house or engage in a livelihood should be that person’s choice. Likewise, ifSGSY were a cash transfer, all bank loans could be cash transfers. Of course, it is axiomatic that if the loan does not have to be repaid, it does become a cash transfer. But ex ante only the power-ful and connected will know that it is a cash transfer and treat it as such, while others will treat it as a loan. For the poor, SGSY, as currently structured, can have the same negative consequences as loan pushing. We contend that an objective that cannot, ex ante,be universal in prin-ciple, should not be a part of a categorical equity plan.
DISCUSSIONnovember 22, 2008 EPW Economic & Political Weekly86(b) Making Cash Transfers Effective Requires Complementary Inputs: MS says:What KMS forget is that any anti-poverty programme will work only if it leads to an end to dependence on doles (what they term in more glorified terms as direct cash trans-fers). To end this dependence requires cre-ating sustainable livelihoods for the poor. And these presuppose much more than transfer of cash. They demand skills, mar-kets, technology, material inputs, infra-structure and institutions. It is these that are in deficit in rural India, more than cash. And it is their lack that accounts for failure of anti-poverty programmes, as much as leakages of cash (the exclusive concern of KMS)…It is much more about ensuring ef-fective utilisation of this cash, which needs both developmental inputs (markets, technologies, skills and materials) and political ones (social mobilisation to set up monitoring mechanisms/institutions).It is gratifying to note that MS shares KMS’ assessment that existing anti-poverty programmes have failed. The six argu-ments inKMS for DCT were that it would (a) “expand their choices and eliminate… paternalism”; (b) “relieve financial con-straints faced by the poor”; (c) reduce “administrative costs”; (d) arrest the “grow-ingimmunity in public administration”; (e) remove the “inherent inequity in some of the subsidies”; and finally, (f) “clien-telism, patronage and corruption that attendCSS would be reduced”. We clearly stated that “Cash transfers work only with a well-functioning private distribu-tion system, which is not present for a number of services. These are functions that could be done by the local govern-ment and is their responsibility in many countries.” Surely, this does not support MS’ contention that KMS is exclusively concerned with “leakages of cash”, and that it forgets that anti-poverty pro-grammes will require “developmental… and political” inputs.The issue about moving away from transfers to identified beneficiaries to investing in physical and social infra-structure arises in the context of a budget constraint. However, to the extent that such transfers continue, our point is simple: compared to the scores of anti-poverty programmes that India has tried, DCTs are likely to improve the welfare of the poor to a greater extent by increasing their purchasing power to access goods and services, many of which the state should be providing, but is not doing. Yes, markets in rural India are imperfect – but are less so than a few decades earlier and the private sector has demonstrated considerable supply response. Markets, technology, and material inputs already exist, albeit with a huge variance in qual-ity. Connectedness to these inputs is pro-vided by infrastructure, i e, roads, com-munication and transport and many of these are provided by the private sector, e g, local para-transit services and cell-phones, though not always at prices affordable to the poor.Instead of belabouring the need for complementary inputs, we need to focus on the mechanisms and the institutions that will provide them. We are not advo-cating that the government get out of areas where it should provide public goods, e g, rural roads. However, the level of government is a question that needs to be asked – whether rural roads should be financed through a central scheme or through an appropriately fundedPRI’s budget. Whether a decentralised public sector (in the form of cash empowered PRIs) can deliver has yet to be established, but surely they need to be given the chance to try for themselves, where centralised mechanisms have failed.KMS addressed this in an admittedly briefer part of the paper, and it is to this we now turn.(c) PRIs Need Support: MS considers “PRIs in large parts of India today” to be “nothing more than work-in-progress. They have a very long way to go before they can become instruments of democracy and development at the grass roots. They need massive support from the state to be able to realise their potential. This is the whole unfinished agenda of reform of rural governance, the reform of the public sector in rural development…It is patently unfair to burdenPRIs with massive tasks of development without providing them the requisite support.”It appears that MS continues to believe in a top-down approach where well-intentioned agents of the state, with the requisite skills and knowledge, are ready to selflessly work to improve PRIs. Of course,PRIs are weak. But that is akin to saying that because a baby will fall the first time she walks, she should not walk. What is the form of “requisite support” needed by thePRIs now? Elsewhere MS complains that “social mobilisers and technical personnel…have not been supplied” toPRIs. But why not instead give PRIs control over the funds needed to hire the requisite functionaries for the purpose? Hence,KMS argue for an increase in transfers toPRIs “in a few functional block grants” with “an effec-tive system of financial accountability” commensurate with their responsibilities. As KMS argued, the least bad way of supporting PRIs is, provide them with “the resources and capacity to deliver good operational performance” and thereby transform them into places of real political contestation.We now turn to two specific schemes discussed byMS, viz,PDS andNREGS. (d) The Reach of PDS:MS says that it is not clear how transfers of cash to the poor would allow them to buy grain from the open market at a time of steep inflation. The problem is that India’s PDS is characterised by a whole range of inequities – its coverage is the weak-est in the neediest regions and it fails to cover crops grown and eaten by the poorest. The way forward is to reform the PDS and extend its reach to and density in the poorest parts of the country, where the need is the greatest.Given the long and egregious mis-management ofPDS, MS’ call to “reform thePDS” is puzzling. In contrast, we believe that if PDS is to be at all reformed, a smaller PDS focused on its core mission has a better chance. For the vast majority of locations, we maintain that the open market is a viable distribution channel. In this, the price of grain is immaterial. If open market prices rise, the food subsidy should rise concomitantly. Replicating this in aDCT would involve indexing a part of it to food prices. The critical benefit of moving to DCT is the reduction in logistics and distribution costs, which can be then used either to increase theDCT or on “developmental inputs” dear toMS, or even to reduce the fiscal deficit, as appropriate at the time. As for choice of grain, if the poor really have different preferences (and are not being forced by lack of resources), will notDCTs increase their purchasing power in the market, and thereby increase the supply of such grain? 
