Punjab government announced, at the start of Ramazan, that atta bags will be available, across the province, for Rs 300 per bag. The provincial government is subsidising atta for all residents of the province. It is obvious that not all the people living in the province need the subsidy. There are those, at higher incomes or with more endowments, who can afford to pay Rs 600 per bag or even more without any problem. But the problem faced by the province is simple: it has no way of separating out those who can afford to pay more and those who cannot. So by default, if it intends to reach the sizeable number of those who cannot afford to pay market prices, it has to offer a subsidy to even those who do not need it. There are two consequences of non-targeted subsidies. First, if a desired level of subsidy needs to reach the poor, the overall size of the subsidy will have to be bigger since it will include those people as well who do not need the subsidy. Second, if a set amount has to be distributed over the entire population, and resources are finite and limited for even governments, the per capita size of the subsidy will have to be reduced. This dilution of the subsidy might imply that those who deserve the subsidy and for whom the programme was designed might not get enough of the subsidy. In the case of atta, this is clearly the case. All people buying at Rs 300 per bag are getting the subsidy, irrespective of whether they need it or not. But if the government could have targeted the amount of the subsidy to the poor, could they have brought the prices, for the poor, even below Rs 300? Clearly that would have been possible. But targeting is not easy and it might not be cheap either. Targeting requires the ability to separate those who belong to a certain set or category from those who do not. For example, in the case in hand, targeting requires the ability to find out, based on household or individual incomes, if the household or person deserves to be in the group that requires atta at Rs 300 per bag. In other cases, the criteria will be different. For access to schooling, the criteria might include whether a house has school going children or not and whether household income is above or below a certain threshold and so on. But the ability to do such means testing requires a lot of information on every household and getting this, to say the least, is not easy. Suppose the Pakistani government wanted to provide the poor households in the country with Rs 1000 per month as cash support (Benazir card has this principle). How will we find all the right households? One way of course would be to give every household in the country Rs 1000 a month. This would definitely include the poor as well, but it would just be too costly and clearly not feasible. The other option is to find the deserving households. Suppose you fix a household income level below which you want to help all households. Say that is Rs 6,000 a month. Then we need: i) an accessible application process through which households can apply for this support, ii) a system at the local level (at the level of villages in rural areas and ward level in urban) that can identify candidate households (as we should not rely on applications only), iii) a verification arm that can do household level checks before enrolment of the people, iv) a mechanism for transfer of resources (money in this case) that is transparent and beyond abuse, and v) mechanisms for regular independent checks on the selection of the households and delivery of benefits to the target group. All of the above are not easy to accomplish and they are not cheap either, but this is the task that has been set by the federal government for the new head of the Benazir Card programme. If we cannot develop such a system, we will always have leakage's in targeted subsidy programmes and charges of nepotism and corruption, true or not, will always be made and will be difficult to refute. Or we will be forced to have non-targeted programmes. The federal government, even before Benazir Card, did have a couple of targeted programmes. Zakat, run by the Ministry of Religious Affairs and Pakistan Bait ul Maal, run by the Ministry of Social Welfare, are conditional cash transfer programmes. But despite the fact that the previous government expanded Bait ul Maal a bit, the government never really invested in creating the proper delivery system for these conditional grants. Bait ul Maal, now being run by a PPP MNA, always had a large client database, and disbursed billions of rupees, but it never developed systematic means for i) universal data and enrolment across Pakistan, ii) proper means testing and iii) capability for credible independent verification. There were always allegations, against Bait ul Maal, of nepotism and corruption, and clearly, without the proper systems, these will not only continue, they will be hard to counter as well. It is also clearly the case that the Government of Pakistan needs the organisational and institutional capacity to be able to roll out conditional cash grant and other social protection systems for the poor and deserving of the country, and across the entire country. Even though the setting up of such capacity is not costless, it makes economic sense to bear that cost once rather than live with leakage in all systems and bear the stigma of nepotism and corruption again and again. Rough estimates suggest that Rs 15-20 billion could help set-up such an organisation across the country. Given Pakistan government and the governments of the provinces spend a lot more than this amount in social protection programmes and subsidies every year even now, this cost does not appear to be much at all. And the benefits, in terms of reduced leakage's as well as gain in reputation, would allow us to recoup the expense very quickly. In addition, once we have such an organisation and system in place, it would be of benefit to not only the federal government or the provincial governments, it could be used by local governments for delivering local benefit programmes as well, and it could also be used by NGOs and/or philanthropists for delivering their programmes too. The organisation could provide the 'back-bone' or skeleton on which the society could roll out most of its targeted programmes. Say, Punjab decides to give subsidy to the poor, in addition to the federal subsidy. All it would have to do would be to i) set the criteria for selection, ii) run the data base that this organisation will have, and iii) ask the organisation to start delivering the requisite benefit to the identified households across Punjab. Similarly, if Sialkot wanted to run a conditional cash transfer programme to encourage enrolment of females in schools, it could do this through the same organisation. The benefits of having such a backbone could be very large. The economic situation of the country is not good. Poverty is increasing as output and employment growth has slowed down, inflation has increased, income growth has slowed and inequality is continuing to increase. As more people become poor or become vulnerable to poverty, we need to ensure we have a way of reaching them to be able to ensure that they can get the needed help in making it through. Since privatisation, liberalisation and de-regulation has limited our ability to have alternative channels (subsidised education, health and/or infrastructure), we need to develop direct channels for targeting help. Otherwise, we will be limiting ourselves to non-targeted subsidies that waste a lot of resources on the non-deserving. Even if the economic situation were to pick up, there will always be people who are not able to participate in the mainstream of economic activity and will need help from the society, and help as a right even. So, even in good times we will need this 'backbone'. We should start investing in creating it right away. The writer is an associate professor and Head of Department of Economics, LUMS and senior economic analyst E-mail: faisal@nation.com.pk