Promotion of Self-Employment for the Poor. Another set of tar¬geted programs aimed at directly generating additional incomes for the poor are those focusing on self-employment possibilities. The basic idea is that members of poor households can be assisted to set up as independent self-employed producers engaged in commodity production or provision of services with a fairly small capital investment. These ventures can be financed by provision of subsidized credit, which is made available on special terms and through special procedures (e.g., without collateral security) in order to overcome some of the problems which limit access of the poor to institutional credit. The administering agency normally also undertakes to examine the technical and economic viability of the project to ensure that it is capable of becoming a self-sustaining source of addition¬al income. In this process, it also provides marketing advice. Typically, the self-employment ventures are on a very small scale, employing mainly family labour.
Such programs are part of the strategy for poverty alleviation in most countries in South Asia. In India there is the Integrated Rural Develop¬ment Programme (IRDP). In Bangladesh there is the Grameen Bank (GB) and also similar schemes run by the Bureau of Manpower Employment and Training (BMET), Bangladesh Small and Cottage Industries Corporation (BSCIC) and the Bangladesh Rural Advancement Committee (BRAC). In Sri Lanka there are programs of the National Youth Service Council (NYSCO) and others by the Central Bank and the Women's Bureau.
The scale of the programs varies from country to country. The Indian IRDP covers approximately 3 million beneficiaries per year with the avowed objective of taking each beneficiary household above the poverty line. Since the total number of households below the poverty line in rural India is estimated at about 55 million, coverage of 3 million beneficiaries per year could make a significant impact over time if the program were as effective as expected. By contrast, the coverage of the Bangladesh and Sri Lanka programs is much more modest compared with the scale of the problem.
The experience with these programs in several countries has been reviewed and the Indian experience has been extensively studied. There are several issues of concern.
One of the issues is whether these programs, which involve a substan¬tial subsidy element, can be successfully targeted so that only those who are eligible become beneficiaries. The experience in this regard is mixed. The Sri Lankan experience shows that most of the beneficiaries were not from the target group. In Bangladesh, the experience varies, with beneficiaries of the Grameen Bank schemes being mostly from the target group, but other schemes less so. India's experience with the IRDP has been fairly good in this respect, with 70 per cent of the beneficiaries belonging to the target group. Considering that further improvements in implementation are pos¬sible, we can confidently conclude that effective targeting is possible.
Another issue relates to the effectiveness of these programs in generating additional income for the beneficiaries. There are bound to be many cases of clear failure reflecting situations where either there has been outright misappropriation of funds or more likely the assets provided for the self-employment venture (milch cattle, goats, bullock carts, pumpsets, sewing machines, etc.) have been sold in a relatively short time and the proceeds consumed. Some of these cases where the asset is lost may reflect poor viability of the project to begin with. Some may arise from the fact that a poor household may be compelled to sell even an asset which is earning a reasonably high rate of return in times of distress because it cannot get access to credit on reasonable terms to tide over temporary difficulties. On balance, the Indian experience shows that the number of such cases of loss of assets is not unduly large. A survey of beneficiaries in 1987 showed that in fully 72 per cent of the cases the assets were found to be intact and operational after two years.
From the fact that assets are retained in operational order, we can conclude that they are found to be productive, but this does not indicate the quantum of additional income generated from the household. Such quantification of the additional income due to these assets is extremely difficult in practice. Base line estimates of income are typically unreliable and it is also difficult to take account of the effects of inflation over time. However, available studies suggest that the IRDP has contributed to raising incomes of beneficiary families subject to two qualifications. One qualification relates to the criticism that in many cases the increase in income is insufficient to achieve the objective of taking the household above the poverty line. This is not however a particularly serious objection since it may only reflect the fact that the investment undertaken was not large enough given the income level of the beneficiary and the gap between this level and the poverty line. A more serious problem noticed in some studies is that the initial increase in income is often not sustained over a longer period. In other words, the self-employment activity often deteriorates over time instead of becoming a self-sustaining viable enterprise. An interesting hypothesis is that this probably happens because the self-employed producers assisted under this program are not assured of continued preferential access to credit. For example, a unit may get its initial credit requirement under the special dispensation of the self-employment scheme, but if no similar provision is made to cater to its subsequent credit needs, it may be difficult for the unit to survive over time.
The key issue in assessing the economic viability of such self-employ¬ment programs must relate to the rate of return earned in such invest¬ments. Alam reports a very low figure of 4.9 per cent for profits earned as a percentage of total investment in the Grameen Bank schemes. However, it is possible that profit in this calculation is understated because much depends on the imputed value of own labour. An alternative approach is to look at the incremental capital output ratio (ICOR), which is perhaps easier to measure than the rate of return and may in any case be more appropriate where there is significant underemployment. On this count, the results for the Indian programs are quite encouraging. A recent longitudinal study of IRDP in one of the Indian States, Uttar Pradesh, shows that the average ICORs obtained were as low as 1.5 if the accounting is over a two-year period and 2.0 over a four-year period for those households which retained their assets. This excludes investments made in the failure cases where the assets were lost. If we include the investments of households which had lost their assets after four years, the ICOR rises to about 3.0. This is higher than is typically thought to be the case for self-employment schemes, but it is certainly significantly lower than the average ICOR of about 4.6 for the economy as a whole.
One dimension in which the IRDP has not performed well thus far is that of repayment of loans. In the study quoted above, whereas 72 per cent of all beneficiaries had the assets intact after two years, only 28 per cent had no credit overdues. In other words, over 60 per cent of the beneficiaries who still retained their assets after two years had not made timely repay¬ments. This is undoubtedly a serious problem since a credit-based scheme can only be viable if credit is repaid. However, it is possible that this percentage can be significantly improved through better monitoring and implementation. One of the problems in implementing these programs is that an impression is often created among beneficiaries that the credit does not need to be repaid, leading to avoidable overdues. Greater involvement and follow up by the banks could lead to much better results. In this respect, the performance of the Grameen Bank in Bangladesh is very much better with much lower overdues.
To summarize, the experience gained with self-employment schemes suggests that there are many problems which need to be resolved if these schemes are to become a truly effective instrument for poverty alleviation. It is certainly not easy to set up poor households as independent producers. Simply providing them with cheap credit or cheap credit plus a capital subsidy will not ensure the establishment of a viable source of additional income. At the same time, it cannot be denied that there is a substantial scope for promoting household-based production activities catering to the local market and requiring only a minimum of skills. Well-designed systems for providing credit, not just initially on a one-shot basis but also on a continuing basis to meet the economically justifiable needs of the enterprise, could enable large numbers of poor households to expand their income-earning capability. If this is combined with an element of training and marketing advice, and also supported by possible cooperative market¬ing arrangements, where appropriate, it could provide the basis for self-sustaining viable production which could help in poverty alleviation.
Conclusion
The conclusion which emerges from this discussion is eclectic. As Bhagwati points out, the optimal strategy for poverty alleviation is bound to be a mixed strategy involving some combination of the direct and indirect approach. The indirect approach has the advantage that since it focuses on the operation of the whole economy, successful indirect strategies are likely quantitatively to make a much larger impact than any possible direct intervention programs, which are bound to be severely constrained by budgetary considerations. On the other hand, since the approach depends on the overall pace and pattern of growth, it may not be able to provide adequate income support to all groups. In the longer run, the impact of a labour absorbing, skill-based growth process is bound to be the decisive factor in removing poverty, but in the short run it may not provide a solution for particular groups bypassed or marginalized by development.
The direct approach has the advantage of responding immediately and visibly in support of identified target groups. Its main disadvantage is that it necessarily involves a heavy draft on budgetary resources which most low- income countries can ill afford. The resource requirement of the direct approach brings in the possibility of a trade-off between the two approaches since resources used for directly targeted interventions can also be used to support particular aspects of the overall growth strategy, especially those which are most likely to have the maximum indirect effect on poverty. Needless to say such trade-offs can be quantified only through the use of general equilibrium models. All such quantifications are subject to numerous well known limitations, but where they are feasible they can throw useful light on the cost effectiveness of direct targeted programs.
An interesting example of explicit quantification of trade-offs is a recent study which examines the cost effectiveness of large-scale rural works programs in India as instruments for raising incomes of the rural poor. The study finds that even if the program is financed by a reduction in the general level of investment in the economy, the cost in terms of reduction in GDP growth over a 20-year horizon is quite modest, while the benefit in terms of income levels of the poorest classes is very substantial. For example, they estimate that a rural works program requiring invest¬ment of 6 per cent of GDP to begin with and a declining percentage over time would, on certain favourable but not impossible assumptions about the effectiveness of targeting and productivity of investment, involve a slow¬down in growth of only 0.25 per cent per year over the 20-year horizon. Terminal year GDP per capita would be lower by only 4.6 per cent but the income of the two poorest classes in the rural areas in the terminal year would be 40 per cent higher than in the base run. The volume of resources envisaged in these simulations is of course massive and feasible levels of investment in such employment programs are likely to be much lower in practice. The benefits would also be correspondingly smaller, but the simulations indicate that the cost of these benefits, in terms of growth foregone, may not be very large. What the simulations establish is that if a rapid increase in the incomes of the poor to some minimum level is a matter of priority, targeted programs on some suitable scale may well be an essential requirement.
It is important to recognize that while some elements of a direct approach are unavoidable as long as significant poverty persists, it is perhaps best to view this as a necessary safety net while relying on broader based growth processes to provide sustained growth in income levels for the bulk of the population. This underscores the importance of ensuring that the total policy framework is such as to ensure that growth is as broad based as possible. Failure to ensure this will lead to an inherently unstable situation in which development strategy choices are made which fail to make full use of the potential for reducing poverty through growth, and this in turn generates pressures to expand schemes for direct intervention. A successful strategy is surely one in which the need for the direct approach diminishes gradually.