wfpmini.gif (539 bytes)

23-24 October 1998
Rome, Italy



This paper was written by Dr. Neville Edirisinghe
of the Strategy and Policy Division, WFP.


Table of Contents


Evidence from a Developed Country
Evidence from Developing Countries




Food has a special value that increases the effectiveness of assistance programmes meant for the poor. This is the "food factor".

A question frequently raised in relation to food assistance programmes is: is giving food more effective than providing cash in improving the nutritional welfare of the poor? There is evidence to show that more food is consumed as a result of food transfers than cash transfers. Where improvement of nutrition among poor households is a major objective of assistance programmes, the food factor helps achieve this more effectively than when a cash amount of equal value is given to such households.

Food-related assistance programmes are common. The food stamp programme in the United States is a prominent example of an income assistance programme directly linked to food. It is the primary source of nutrition assistance to low-income Americans. At least 30 other countries have some sort of food transfer programme. Schemes vary, but may be based on food stamps, quantity rations, food price subsidies or food for work (Subbarao et al., 1997).

Several evaluations of food assistance programmes have been undertaken. They are examined in this paper to discern empirical evidence of the existence of the food factor.

Food versus cash: the theory

It is often argued that monetary assistance may be equally effective or even better than providing food or other in-kind transfers to the needy. With cash the poor have the freedom to choose how much to spend on food and how much on other needs. However, since food is a basic need, the poor spend a high proportion of their additional income on food. Typically, poor households in developing countries are found to spend 65 to 70 percent of their incomes on food (World Bank, 1992). Thus, it is likely that cash assistance will increase the food consumption of the poor.

The counter-argument is that there is a special value in food - the food factor - which results in a greater impact on nutrition than cash assistance.

What is the position of economic theory on this issue? Traditional theory would expect households to treat food assistance as a form of income that adds to their pool of household income (Southworth, 1945). Therefore, conceptually it does not matter whether food or cash is given to help them.

Food versus cash: practical experience

Many studies of actual programmes indicate that the theory does not hold (Fraker, 1990). It has been found that poor households view food-related incomes differently from cash incomes. The result is enhanced food consumption due to food income.

How do we find out whether food income contributes more to food consumption than an equivalent of cash income?

What is specifically examined in this paper is the marginal propensity to consume food, that is, the proportion of an additional unit of income that would be allocated for food consumption. Would this proportion be the same whether the additional unit came from food income or cash income?

Evidence from a Developed Country

The food versus cash issue has been studied most extensively in relation to the food stamp programme in the United States. Food stamps are provided to enable the poor to improve their diet by increasing their food purchasing power. The question studied is: why is it necessary to link the income support to food rather than just giving cash to low-income families? If food stamps do not make "an additional" contribution to food consumption, then there would be no good economic reason to have a food stamp programme.

Many studies on the United States food stamp programme have found that food stamps lead to higher food consumption than an equivalent cash transfer would bring about. A review of 19 such studies shows that most estimates of the increase in food expenditures resulting from an additional dollar of food stamps coupons range from 17 to 47 cents, compared with estimates of 5 to 13 cents for an additional dollar of ordinary cash income (Fraker, 1990). Clearly, if cash were given instead of food stamps, food consumption of poor households would be significantly lower.

In fact, the United States food stamp programme conducted an experiment on the issue of cash versus food stamps. In a number of locations food stamp benefits were issued in the form of cheques rather than the traditional coupons. A careful study of this experiment found that giving cash instead of food stamps reduced food expenditures among recipients by 18 to 28 cents for each dollar of food stamps spent (Fraker et al., 1995).

These results defy the traditional theory. The overwhelming evidence is that the proportion of a marginal dollar from food stamps allocated for food is three to four times greater than the allocation from normal income. In other words, the presence of a food factor is a reality in the case of the United States programme.

Evidence from Developing Countries

Is the concept of the food factor also relevant to developing countries? Not many studies have looked into this issue. But the available evidence suggests that the answer is "yes". Most of this evidence comes from research conducted by the International Food Policy Research Institute (IFPRI).

One study examined India’s food subsidy programme as it operated in the State of Kerala (Kumar, 1979). One of the issues examined in the IFPRI study was how poor households used the increased "income" from the food subsidy to improve their diet. Households receiving the subsidy are given a ration of food at a lower price than they would pay elsewhere. The difference between the ration price and the open market price (rupees per unit) multiplied by the quantity of the ration food obtained is the subsidy income. Was the food subsidy income used any differently from other types of income earned by the households?

The impact of subsidy income and other incomes was evaluated in relation to four measures of household nutrition: calorie consumption, protein consumption, dietary quality and nutritional status of children at weaning age. The ration subsidy had a significant positive impact on all these measures. The improvements in these indicators of consumption and nutrition, a result of the incremental rupees from the food subsidy, were six to ten times greater than the improvements that resulted from increments of cash income.

Another IFPRI study reviewed a pilot, targeted food price subsidy scheme implemented in three provinces of the Philippines in 1983 (Garcia and Pinstrup-Andersen, 1987). The scheme provided subsidized rice rations and cooking oil to inhabitants of half of 14 villages selected for their high incidence of malnutrition and poverty. The remaining villages provided a control population to check on nutrition indicators. The scheme also had a nutrition education component. The evaluation of the scheme indicated that subsidizing food rations could be far more effective in improving food consumption and nutrition than providing an equivalent cash income to the poor households:

IFPRI also conducted a study on Sri Lanka’s food subsidy scheme (Edirisinghe, 1987). One of the issues examined was whether cash transfers or food transfers are more effective in increasing food consumption by poor households. Like India (and for similar reasons), Sri Lanka has also had a food subsidy scheme for over five decades. However, a major difference in Sri Lanka is that it replaced the subsidized rice rationing scheme and other general food price subsidies with a food stamp scheme in 1979.

Data for the study on rice rations come from a period (1978/79) when households identified as needy were eligible for a subsidized rice ration. They also benefited from a general price subsidy on wheat and sugar. Theoretically, they would allocate to food expenditure the same proportion of an additional rupee of subsidy income as the proportion they allocate from an additional rupee of their own cash income.

IFPRI results again show that the "theory" does not hold. Poor Sri Lankan households treat food subsidy income differently from their normal cash income:

The Sri Lanka study is especially important because it also looked at the difference between food stamp income and own cash income and how it affects food consumption. In 1979, Sri Lanka introduced a food stamp scheme to replace the rice rationing scheme. The IFPRI study found that the households treated food stamp income the same way they treated their own cash income when making food purchases. Under the new scheme, low-income households received a certain number of food stamps which they could use to purchase designated items (food and kerosene). They also had the option of depositing the food vouchers in a post office savings account and receiving credit for the value of the vouchers. In essence, although food stamps were food-related, they were clearly perceived as a form of cash transfer. This may be the major reason why the IFPRI study found no difference in the way a marginal rupee from food stamps and a rupee from normal income were allocated. A similar result was observed in the case of a short-lived food stamp scheme in Colombia (Pinstrup-Andersen, 1988).

Why a food factor?

How does one explain the existence of the food factor that makes a food transfer do better than cash transfers in increasing the food consumption of poor households?

A possible explanation that has received much attention relates to differences in tastes and preferences within households (Haddad et al., 1997). The fact that assistance comes in the form of food (or a coupon to buy food) could give some household member(s) who gives a higher priority to food or nutrition more control over the household budget.

The claim that household decisions are independent of who receives or who is entitled to income has been refuted in a number of settings. In the mid-seventies the British child allowance system was changed to make child benefits payable in cash to the mother. The reasoning was that children would benefit because mothers had control over the cash allowance (Alderman et al., 1995). In Gambian households, when the proportion of cereals produced by women increased, household consumption of calories also increased (von Braun, 1988). Raising of income accruing to wives in Philippine households increased acquisition of calories and protein (Garcia, 1990). In Côte d’Ivoire a doubling of women’s share of cash incomes raised the household budget share of food and lowered the budget shares of alcohol and cigarettes (Hoddinot and Haddad, 1991).

The evidence is strong that greater control of resources by women leads to better nutrition and health, especially among children in poor households. Even if food is not directly given to women, the fact that the assistance comes in the form of food creates an entitlement to this assistance by women because typically women shoulder the responsibility for household food consumption matters.2 This concept of entitlement makes food a relevant form of remuneration, for example in public works programmes, even if the workers are predominantly male (Haddad et al., 1997).


1 The average subsidy income received per person among poor households was about 21 rupees, which implies that the food transfer added roughly 588 calories to an individual’s consumption.

2 For an in-depth discussion of the impact of directing food resources to women, see "Women in Food Aid Interventions: Impacts and Issues", WFP, 1998.



Alderman, H., et al., 1995. Unitary versus collective models of the household: is it time to shift the burden of proof? The World Bank Research Observer 10 (1).

Braun, J. von., 1988. Effects of technological change in agriculture on food consumption and nutrition: Rice in a West African setting. World Development 16(9):1083-1098.

Edirisinghe, N., 1987. The Food Stamp Scheme in Sri Lanka: Costs, Benefits, and Options for Modification. Research Report No. 58. International Food Policy Research Institute, Washington, DC.

Fraker, T.M., 1990. The effects of food stamps on food consumption: A review of the literature. Food and Nutrition Service, USDA, Alexandria (Va), USA.

Fraker, T.M., Martini, A.P., and Ohls, J., 1995. The effects of food stamps cashout on food expenditures. The Journal of Human Resources 30(4).

Garcia, M., 1990. Resource allocation and household welfare: A study of personal sources of income on food consumption, nutrition and health in the Philippines. (PhD dissertation, Institute of Social Studies, The Hague, the Netherlands).

Garcia, M., and Pinstrup-Andersen, P., 1987. The Pilot Food Price Subsidy Scheme in the Philippines: Its Impact on Income, Food Consumption, and Nutritional Status. Research Report No. 61. International Food Policy Research Institute, Washington, DC.

Haddad, L., Hoddinot, J., and Alderman, H., 1997. Intrahousehold Resource Allocation in Developing Countries. International Food Policy Research Institute, Washington, DC.

Hoddinot, J., and Haddad, L., 1991. Household expenditures, child anthropometric status, and the intrahousehold division of income: evidence from Côte d’Ívoire. Research Program in Development Studies Discussion Paper 155. Woodrow Wilson School of Public and International Affairs, Princeton University, NJ.

Kumar, S.K., 1979. Impact of Subsidized Rice on Food Consumption and Nutrition in Kerala. Research Report No. 5. International Food Policy Research Institute, Washington, DC.

Pinstrup-Andersen, P., (ed.), 1988. Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options. International Food Policy Research Institute, Washington, DC.

Southworth, H.M., 1945. The economics of public measures to subsidize food consumption. Journal of Farm Economics 27:38-36.

Subbarao, K., et al., 1997. Safety net programmes and poverty reduction: Lessons from cross-country experience. The World Bank, Washington, DC.

World Bank. 1992. Poverty Reduction Handbook. The World Bank, Washington, DC.