All subsidies are implicit transfers; the distinction between direct and indirect transfers is, nevertheless, useful. Food subsidies are a common method of attempting to reduce food insecurity, although as noted above, because of fiscal constraints they have become less popular in recent years, with greater emphasis placed on employment and productivity-raising interventions. Where subsidies or direct interventions such as school-feeding programmes are pursued, the emphasis on targeting and cost-minimisation has become more pronounced with the introduction of economic reform programmes.
The costs of any food subsidy scheme fall into three main categories:
(i) Fixed costs. These overhead expenses will be greater per unit of subsidy, the more narrowly the programme is targeted, since targeting requires knowledge of household income and that knowledge requires consistent updating if it is to remain meaningful. The less developed the administrative infrastructure of the country, the higher the initial administrative costs are likely to be. Fixed costs not only include administrative expenses but also appropriate physical items such as warehousing facilities and buildings and equipment necessary for distribution. Obviously the latter are much greater when subsidised commodities are distributed as part of the scheme, such as through ration shops.
(ii) The variable cost of the system is the value of the subsidy to the household. This will vary, not only with the extent to which the government wishes to subsidise households, but also with the extent to which it is successful in targeting. Subsidising the non-poor, or those not included within target group households represents leakage from the system and should, if possible, be estimated as a separate item.
(iii)Apart from these explicit costs, account should be taken of the implicit costs which may accompany any food subsidy. Although not easily quantifiable, these can have important ramifications if their implications are not accounted for. A typical example is the case where food subsidies lead to an expansion of consumption of particular foods which are not able to be produced domestically, resulting in an expansion of imports and shift in consumption patterns away from the local staple. Policy makers in this context should attribute the implicit costs associated both with the diversion of foreign exchange and with the loss of producer income.
The incentive impact of a subsidy refers to the responses of the recipients. Governments will be concerned that receipt of a subsidy will not lead to a reduction in household effort with regard to the acquisition of income or food. Attention must also be paid to the political acceptability of subsidies. While the problem of subsidy reduction in the face of poverty, and the instability it can create, is well-known and apparent, the long-term political acceptability of subsidies may depend on them benefiting not only target group households, but also an adequate proportion of the non-poor population, since it is with these that political power ultimately resides. The commitment to subsidise the consumption of the poor may therefore be more forthcoming if it also benefits people who exert political influence. This may explain the examples of costly general subsidies which bestow the greatest benefits on the non-poor members of the population.
Finally, it cannot be emphasised enough that any well-designed subsidy scheme must reflect local constraints and possibilities. In this respect, adaptation is often a preferable policy to sudden change. In any case, the undiluted adoption of strategies which have been observed to work well elsewhere may prove disappointing. Design should take account, as far as possible, of local abilities, the sophistication of existing information networks, current consumer preferences (although not necessarily by reflecting them), the source of food supplies, as well as government priorities. It must also be borne in mind that the optimum system at any one point in time may soon become sub-optimal as the constituents of target groups change, and the administrative capacity of the country grows. In-built flexibility is consequently a desirable characteristic.
Food subsidies In Lesotho, a maize-meal, introduced originally to offset the acute food insecurity experienced though drought, is channelled thorough the three principal millers. It is assumed that the subsidy is passed on to consumers by wholesalers and retailers. A problem stems from the fact that the subsidy applies to maize-meal rather than maize grain or the offal, and is a subsidy on milling costs. This means that the more refined is the meal, the greater the volume of subsidy it contains. Since poorer consumers consume a higher proportion of cheaper, less refined meal, this implies that they receive less subsidy per unit of maize-meal that the less poor consumers: i.e. the subsidy is regressive in the distribution of its benefits (WB, 1995c). Any non-targeted subsidy will to some extent be regressive in its impact, since it will benefit most those who consume most, as did the general maize subsidy in Zambia prior to 1989. One reason why it was replaced by a coupon system was because of the escalating and unsustainable cost: a reason why general food subsidies around the world are being scaled down or targeted (WB, 1995b). |
General subsidies have been common throughout the Region. They are designed to reduce effective marketing margins. They may involve subsidies direct to wholesalers, or the acceptance by the treasury of parastatal losses. In addition, storage costs can be subsidised directly through the provision of low-rent storage facilities; the provision of cheap credit for storage purposes; subsidised fuel for transporting food commodities or low rates for transport. A further illustration of a general, untargeted subsidy is provided by food rations where the subsidised commodities are restricted in quantity but not in coverage, and by the untargeted issue of food stamps or coupons. Their advantages reside in the relative simplicity of general or non-discriminatory policies, and the limited demands on administrative resources. The main disadvantage lies in their inefficiency, since all groups benefit, not only those targeted by government objectives.
As noted, where consumer subsidies remain, there is a greater emphasis on targeting. Any form of targeting will reduce the variable costs, although this can only be considered worthwhile if any savings in this respect are not outweighed by higher administrative costs, and the process does not lead to the exclusion of groups which the programme is designed to help. There are several approaches to targeting subsidies: (i) self-targeting through geographical location; (ii) self-targeting through the subsidization of low quality or inferior commodities; and (iii) targeting by income.
An example of geographic targeting would be through the location of ration shops or distribution centres in poor districts. This is in effect a form of self-targeting by location. Self-targeting has the advantage of administrative simplicity, although a degree of leakage is inevitable and accepted. An alternative mechanism of self-targeting is by commodity: most empirical studies of consumption patterns confirm that lower income groups tend to consume a higher proportion of foods perceived as "poor peoples" foods", and that their income elasticity of demand for such foods is relatively high compared with other income groups. Typically, these foods include basic staples such as cassava, plantain, sorghum and millet. Maize is in a more intermediate category, but the less refined is the maize-meal, the more it is perceived as suitable for self-targeting.
Targeting An effective method of targeting is a general subsidy on commodities primarily or exclusively consumed by target groups. This has the advantage of administrative simplicity and a low fixed to variable cost ratio. This form of self-targeting, proved successful in a number of countries, is the ideal system for countries with a small information base and limited administrative capacity. Such a programme could also operate around the coarser maize-meal produced by small-scale hammer mills which are becoming more widespread in both Zambia and Zimbabwe. If necessary, the burden of the subsidy could be borne partly by consumers of the more refined grades. through the imposition of a tax on its consumption. Additionally, other commodities considered "poor peoples' foods" in non-maize growing regions of the country, such as sorghum, millet and cassava flour, could be subsidised in this way (Pearce, 1991). There are two main problems associated with self-targeting by commodity. First, the choice of an inferior commodity is not always straightforward, since a commodity which is regarded as inferior in one part of the country, may be the principal staple in another, e.g. cassava in some countries of the Region. Subsidising a commodity which is insufficiently inferior could prove almost as expensive as a general subsidy. Secondly, since any such scheme relies on self-targeting, it is likely to involve high costs in terms of leakage. Nevertheless, in terms of raising levels of food consumption among target group households, such a system could prove extremely effective. |
A common example of targeting on the basis of income is through the issue of food stamps or coupons. Problems posed by this form of subsidy arise from the data base from which they are administered. In most low-income countries, information on household income will be scarce or non-existent, making it difficult to locate the target recipients. The problem is compounded over time, as households move both physically, thus creating the danger of being lost to the administering authorities, and in and out of different income brackets. There is a danger, under these circumstances, of the numbers involved becoming cumulative. Effective and persistent monitoring is necessary for the success of any coupon system, not only for initial recipient identification, but also because the poorest households are often amongst the most itinerant.
These include programmes of supplementary feeding to special target groups, such as pregnant and lactating women, and young children; and school feeding programmes. Again, there are many examples of successful application. Generally, such programmes are successful in targeting small populations for a limited period. They are less ambitious in their scope than the types of system outlined above, but when funds are very limited they can prove highly effective in improving household food security amongst the sections who are the most vulnerable at any one time. Their success depends on their limited objectives and their ability to employ existing administrative infrastructures. An example of the latter is the use of health centres, for which the process brings secondary advantages in the form of encouraging higher attendance. The ability to use such infrastructures, however, provides a further reason for limiting the scope of such operations, since otherwise the symbiosis can prove mutually destructive.
Supplementary feeding programmes for schoolchildren are common in the Region, as are schemes to provide additional rations for pregnant and lactating mothers and the old and infirm. Usually such schemes are administered through schools or health centres. Apart from the obvious targeting towards potentially vulnerable groups, narrower definition of the target is often inadequate and is, in any case, difficult. A possibility is to identify children through nutritional criteria, and to provide supplementary food. The issue then is where to provide the additional food. Some schemes allow dry "take-home" rations, while others require the food to be eaten at school to ensure that it is consumed by the target individual. Similarly, food provisioning through health centres may be geographically targeted, especially during famine periods, but again this may be a blunt instrument. In all such schemes, the administrative costs can become a high proportion of the total, although the targeting focus is very narrow, being on individuals rather than households. Such projects are mainly associated with nutritional monitoring and other health-related programmes.
