Tuesday, June 12, 2007

Eliminating Poverty in India


What is the cost of eliminating poverty and hunger in India? That of course depends on the extent of poverty, which has been mired in academic debates about the measurement of poverty. There is however universal agreement that in the years from 1993-94 to 1999-2000 the poverty rate (HCR) was between 25% and 35%. We can therefore skirt the esoteric debate about the precise change in poverty between 1993-4 and 1999-2000 and its level in either year by considering three numbers. For each of these years we order the households/person by consumption level and identify the ones which are 25%, 30% and 35% from the bottom. That is we identify in each year the consumption level of the person(s) who would be just at the poverty line if the poverty rate was 25%, 30% and 35% respectively. Then we calculate the income transfer needed for every body below that level to be brought up to the level. This data is summarised in the table below.
Table 1: Consumption Expenditures and Expenditure Gap

In 1993-94 the Central government expenditure in the budget category “subsidies” was Rs. 12,682 crore of which Rs. 10,099 crore were for food and fertiliser subsidies. The latter would have been enough to bring all the poor to the consumption level of the person/household at the 25% level. During the same year the Central and State governments together spent another Rs. 14,160 crore on the budget categories ‘Rural development,’ ‘Welfare of SC, ST & OBCs’ and ‘Social Security and Welfare.’ This expenditure would have been enough to bring all the poor to the consumption level of the person/household at the 30% level. These two sets of expenditures (Rs. 25850) would have been more than sufficient to eliminate poverty in 1993, if transferred directly to the poor and disadvantaged (SC, ST, handicapped, old, poor farmers).
In 1999-2000 the total subsidies provided by the Central government were Rs. 25,690 crore of which Rs. 22,680 crore were for food and fertiliser. During the same year the Central and State governments together spent another Rs. 28,080 crore on ‘Rural development (RD),’ ‘Welfare of SC, ST & OBCs and ‘Social Security and Welfare.’ Either of these was sufficient to bring all the poor to the consumption level of the person/household at the 30% level. Given that poverty was between 26.1% and 28.6% either of these if transferred directly to the poor and disadvantaged (SC, ST, handicapped, old, poor farmers) would have eliminated poverty. Together these subsidies and poverty alleviation expenditures (Rs. 53,770 crore) would have been sufficient to eliminate poverty in 1999-2000, even if administrative costs and leakages used up half the allocation (and the small fraction of RD expenditures on water supply were excluded).
It can be argued that the ideal (most efficient) social welfare policy is a direct transfer of income to the poor through a negative income tax. In a developed country this would be very easy. How can we transfer these amounts directly to the poor, the needy and the disadvantaged in a poor country? The answer, by setting up an Indian version using a modern smart card system that delivers cash and/or subsidies to the poor based on their entitlements as per specified parameters and norms. Such a smart card could be programmed with identity (photo & biometric fingerprint), and have information on social (SC/ST) and personal/household characteristics. Each person/ households’ entitlements could be in the form of specified subsidies (per unit subsidy of si for up to qi units for all i in C) for the purchase of a set of items C. The set of items C could include food/cereals, kerosene, midday meals, nutrition supplements, drinking water, toilet/ sanitation services, basic drugs, schooling (primary/secondary), internet access, electricity and a host of other items reflecting the dozens of subsidies and programs currently in existence. The entitlement could be varied with and dependent on various economic and social handicaps such as SC-ST, age (infant or aged), mental handicap, physical disability, female head of household, lactating mother, chronic illness. In this way all the current stakeholders, special interest groups and social policies could be accommodated within a single integrated system.
These subsidies would have to be collected by the provider of the specified service from the government through the smart card system just as is done currently in a credit card system. Alternatively all these entitlements could be calculated and consolidated into a single cash value to be delivered to the beneficiary every month at his residential address, through the smart card system. Though on theoretical economic grounds the latter may be the preferred option, the former would also yield substantial gains and perhaps be more feasible at this stage.
If poverty could be eliminated so easily why has this not been tried before? There are many reasons, but the most fundamental is illustrated by the following experience: In the formulation of the tenth Plan for food policy/ PDS system there was a proposal to gradually introduce a credit /debit /smart card system to replace the existing PDS system characterised by enormous leakages and high administrative costs (see Virmani and Rajeev (2001)). In this system the entitled person could obtain the specified subsidy from any participating supplier of food/cereals. The person would pay the supplier the difference between the market price and the unit subsidy, and the supplier would collect the subsidy from the government. The formal proposal was to carry out an experiment (as a first step) to determine its effectiveness and to learn about and iron out any problems that may arise. Consequently funds were allocated in the tenth plan for introducing it in a sample of urban areas along with the introduction of food stamp system in a sample of rural areas. Not a single State govt agreed to undertake this experiment, as it has the potential of dramatically reducing leakages and administrative costs.
Smart Card System
The smart card would also constitute a national identity card. For instance the card could contain information on citizenship and voting eligibility (constituency for voting) as provided and checked by the home ministry and the election commission respectively. Secrecy and confidentiality clauses would have to be built into the national smart card system by law. For instance, any person who does not want to avail of any subsidies / entitlements from the government need not provide the information needed for calculating & monitoring the subsidy/entitlement. They would for instance only provide the information necessary to obtain a passport and voter registration card. Many agencies of government (e.g. CBEC, CBDT, and Home) have proposed identification cards. There are significant economies of scale in having one smart card system for all citizens, with different agencies having their own special modules (password protected access to memory segments) within the card for their specialised needs.
The setting up of a smart card system is somewhat distinct from running it even though there may be economies of scope. The former is very similar to carrying out a (special) census in which the data gathered would be entered into a smart card. There is however an additional, technically challenging component, the simultaneous recording of a photo and a biometric fingerprint so as to minimise fraud. The experience with a similar system used in SEBI MAPIN project suggests that it would be best to sub-contract it to private parties in each State/region.
The running of smart card system is on the other hand very much like the running of a credit card system. All the credit card companies, as well as companies that provide back office services to credit card issuers or marketers, would be interested in competing to obtain the contract for the running of such a system. As a credit card company has to incur a fixed cost in setting up its own credit card system, these companies may be willing to charge below cost if they can share the fixed costs of the public system with their private card systems. This could make a significant difference in the cost of spreading the system to the rural areas. Cash delivery through smart card would be akin to a modern version of the Post & Telegraph department’s money order system, already operational with specialised companies that intermediate international/national remittances. The cost of setting up and running a nationwide cash delivery system for the poor would probably be significantly less than that of a commodity related system. The total steady state cost of running this system (including depreciation and return on capital) should be of the same order as the current credit card systems (< 10%).
The identity of the households below the poverty line is not fixed from year to year. The largest turnover occurs because of health shocks followed by natural disasters (droughts and floods) that knock people below the poverty line, while others who have recovered from the shock or have improved their position move above the line. As a matter of abundant caution we could target the bottom half of the population for issue of smart cards (with complete entitlement related information). Annual updating of entitlement related information could be done for those below the poverty line and those up to half this percentage above the line (i.e. if poverty rate, HCR, is 24%, cover poorest 36%).
An independent authority including government officials and non-government organisations could be set up to monitor the integrity of the Poverty Elimination System. This supervisory authority would ensure that private operators are running the smart card system in a manner needed to ensure that the subsidy reaches the poor.
Poverty, which rose during 1950-1 to 1979-80 (Indian socialism), has been on a clear down trend during the Market reform period (1980-1 to current). The level of poverty in 1999-2000 was estimated by poverty experts to be between 26.1% and 28.5% as per the Planning Commission methodology. This level of poverty is to be expected in a low income country like India. India’s poverty ratio is relatively high because we are a relatively poor/ low income i.e. with low average income. 90% of the countries in the world had in 1999-2000 a higher per capita (average) income than India. Our Global poverty ranking is in fact better than our ranking by per capita income. The absolute number of poor is very high because our population is very large, the second highest in the world. Contrary to popular wisdom, the large number of poor has little to do with income distribution. Our income distribution as measured by the Gini co-efficient is better than 3/4th the countries of the World. Further our rank with respect to income distribution is even better, with the poorest 10% of the population having a consumption share that is the 6th highest in the World.
Poverty can be eliminated within the next five years if we are willing to radically change the approach to “poverty alleviation” which was started in the early 1980s. Smart cards anyone?

No comments: