Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

The feasibility of a food-coupon programme in Brazil

Antonio Carlos Campino

The nature of undernutrition in Brazil

Undernutrition in Brazil is largely an urban phenomenon; its prevalence increases with city size, and it is more acute in the urban slums and outlying areas where the majority of the affected population lives. Protein-calorie deficiency is the most significant nutrition problem among the Brazilian population. Calorie deficiency is far more important than protein deficiency, however, except among children under 1 year old. Because of its relative importance, the nutrition problem in Brazil is strongly correlated with the lack of sufficient food and presents only a weak association with the need to correct bad nutrition habits.

The ideal solution to this problem would require a significant process of income redistribution. The word significant is used to stress the size of the undernourished population and the magnitude of the income transfer that would be required. Campino and Farina [3] estimated that 34% of the population of Brazil had a caloric intake insufficient to avoid health problems; that is, equal to or below 80% of the minimum requirements. (For purposes of comparison, one should consider that Reutlinger [4] estimated that 6% of the population of Latin America had a calorie intake insufficient to prevent health problems.) It would be necessary for the real income of the poor population to increase by 70% to cover a caloric gap of 20%. These characteristics imply that income redistribution, although essential for the country, is a long-term measure, and it will be necessary therefore to examine the feasibility of other, short-term measures to alleviate the problem.

One possible measure is to reduce food prices, but to promote sustained reductions also requires a long-term time-frame. Therefore, in spite of their stop-gap nature, food and nutrition programmes have an important role in Brazil and should be examined in greater detail in relation to their targeting, efficiency, and effectiveness.


Overview of the nationwide food and nutrition programmes

The last year for which a complete set of data was available that could be used to assess the coverage and the cost per beneficiary of the food and nutrition programmes in Brazil was 1986. The programmes comprised in PRONAN (National Food and Nutrition Programme) served a population estimated at 39.8 million, corresponding to 28% of the total population. However, there may be some overlapping; for example, a child could be counted twice, as a beneficiary of the food-supplementation programme (PSA) and also as a beneficiary of the school-lunch programme (PNAE). To cover these 39.8 million people, Cz$8,792 million (US$644.1 million) was spent in 1986 by the government sector alone, as follows: 57% on school lunches (PNAE), 32% on food supplementation for children and pregnant and lactating mothers (PSA and PCA), 5% on the national milk programme (PNL), 4% on food for workers (PAT), and 2% on food subsidies through retail stores (PROAB). The average government expenditure per person served during 1986 was Cz$220.9 (US$16.18).

These costs are only those directly attributed to the programme. However, the PSA is integrated with health actions, and the population benefits from the joint action of food availability and health services, although this is not reflected in the figures. This is also the case, but to a lesser extent, with the PCA programme.

Only federal government funds were assessed, although states and counties also allocate monies to the various programmes, especially school lunches. If these funds had been taken into account, the total volume of resources would have changed, as would the relative importance of the programmes; for example, the school-lunch programme would probably have a better relative position.

The total expenditure on these programmes increased from US$644.1 million in 1986, corresponding to US$765.5 million in 1989, to over US$1.028 billion in 1989. During this period there was also a significant reallocation of resources. The largest beneficiary was the non-targeted national milk programme, whose resources increased by a multiple of 15 from 1986 to 1989 (from US$31.4 million to US$467.3 million), followed by the complementary feeding programme, whose resources were increased by a multiple of 4 (from US$33.4 million to US$147.5 million) 2

TABLE 1. Characteristics of the PRONAN nutrition programmes, 1986

  Target populationa Foods Beneficianes (millions) Cost per beneficiary
PSA—supplementary feeding programme P/N, C < 7 FY < 2MW basic food basket 8.6 Cz$277.0 (US$20.30)
PCA—complementary feeding programme P/N, C < 3 foods formulated (US$16.21) 2.06 Cz$221.4
PNAE—school feeding programme K/E formulated foods 21.72 Cz$229.7 (US$16.83)
PAT—workers' feeding programme Formal-sector workers   2.5b Cz$ 141.1c (US$10.30)
PROAB—basic foods supply programme FY < 2MW staples 11 basic food (US$4.55) 2.9 Cz$62.2
PNL—national milk programme FY < 2MW milk 1.5 Cz$286.24 (US$20.97)
Total     39.8 I US$ 6.18 1

Source: INAN, and Lucia Pontes de Miranda Baptista of IPEA.
Total cost of all programmes: Cz$8,792 million.
a. P/N = pregnant and nursing women; C<7, C<3 = children under 7 and 3 years old respectively; K/E = kindergarten and elementary-school children; FP < 2MW = families with income less than two minimum wages.
b. 1983.
c. 1986 estimate.

Food stamps as an alternative

The data in table 1 indicate that by 1986 the political decision to spend a significant amount of resources on food and nutrition programmes in Brazil had already been made. However, each project had significant shortcomings, such as overlapping, duplication of administrative structures, and food waste in the health centres (since they were not equipped for food operations). Food stamps appear to present a workable alternative to the existing programmes, since its operation does not involve some of the shortcomings of direct food distribution.

Because a food-stamp programme specific for milk (the fluid-milk programme, or FMP) has been in operation in the state of São Paulo since 1983, this article reviews the lessons learned from this experience and assesses its applicability to the whole country. The literature on how such programmes originated and evolved in the United States and the published evaluations of the effects of similar programmes in other countries indicate the following:

  1. They benefit low-income families but, as has been shown in Mexico, not necessarily the poorest families [6].
  2. They have an impact on total food expenditures; in the United States each dollar provided by the programme increases aggregate food spending by US$0.35-US$0.60 [7].
  3. Anthropometric evaluations show no perceptible nutrition effects, but such effects are evident in terms of increased intakes of calories, protein, and other nutrients [7-9].
  4. The programme affects the level of economic activity [10].
  5. There is apparently no significant difference in the distribution of food stamps in the form of coupons and in the form of checks [11].

These conclusions have implications for the study of the FMP in São Paulo. They lead one to expect the programme to benefit poor families, since only those with undernourished children or with children at risk of being undernourished would be able to enrol. Another expected outcome is a positive impact on food spending. A measurable nutrition effect in terms of anthropometric indicators is not to be expected, but the programme should produce measurable effects in terms of increased caloric intake. Finally, the idea that the programme would help improve the performance of some local economies (milk-producing counties) seems to find support in data on the American food-stamp programme, which presented a gross multiplier on the order of 3 [12].


Evaluation of the fluid-milk programme in São Paulo

The FMP began in 1983, with three main changes in relation to the former programme. First, the product distributed was changed from powdered to fluid milk. The main purposes of this change, according to the state Secretary of Health, were (1) to avoid the oligopolistic practices of the large powdered milk industries, since the cost of powdered milk distribution accounted for more than 50% of the expenditures of the health centres, and (2) to stimulate local production of milk at the county level Second, the quantity distributed was increased from 2 kg of powdered milk (equivalent to 14 litres of fluid milk) per child per month in 1983 to 30 litres of fluid milk per child per month in 1986.4 Finally, the relevant age-group for inclusion in the target population was increased from less than 1 year to less than 5 years.


Definition of the target population

The target group for this nutrition programme was made up of children under 5 years of age in families of low socio-economic status who either were below the tenth percentile of the Santo André standard weight for age (hereafter referred to as the undernourished) or, although they were above that percentile, had been losing weight consistently over a three-month period (hereafter referred to as being at risk of undernutrition).5 All the children were examined and weighed once a month for three consecutive months at the health centre before it was decided whether they were entitled to receive milk.


Forms of milk distribution

The method of milk distribution differed between the metropolitan area of São Paulo and other areas. Because the large dealers in fluid milk in the metropolitan area declined to cooperate with the programme. the tickets distributed by the health centres there (each mother received 30 tickets at the health centre at the child's regular monthly visit to the physician) could be traded only at a specific counter operated by the county authority. Outside the metropolitan area. they could be traded at a local vendor, usually a small neighbourhood bakery where people bought bread and milk on a daily basis. The vendor would then trade the tickets in at the local branch of the state-owned bank (Banco do Estado de São Paulo) on certain pre-established days. This mechanism could be improved by following the example of the federal government's programme, in which the tickets are traded daily for milk delivered by the company in charge of distribution.


Coverage and representativeness

The increase in geographical coverage occurred rapidly in the inland counties but it took some time to achieve an agreement with milk distributors in the metropolitan areas. In the inland counties, the FMP had already achieved significant coverage by 1984, whereas it did not even begin in the metropolitan area of São Paulo until August 1986.

When this study was conducted, the coverage in geographical terms was already very good; milk was being distributed to 401 out of a total of 572 counties in the state of São Paulo (374 of 537 inland counties, and 27 of 35 metropolitan-area counties). The quantity increased tenfold between January 1984 and December 1986.

The programme aided 129,000 children from birth to 4 years of age, or 3.27% of the estimated 4 million children in that age bracket. However, with a well-targeted programme, coverage of the relevant undernourished population could be significant.

Since there are no data on the total target population, data from the IMPEP/IPE study [14] of lowincome families were used. According to that study the population earning less than 0.5 of a minimum wage per capita (a rough measure of poverty) constituted 9% of the total population; in this income group, 46% of the children aged 6-60 months were undernourished. If these coefficients still hold true, the total number of poor, undernourished children from birth to 4 years would be 165,000, and programme coverage, if properly targeted, would correspond to 78% .6



Two samples were taken, in the counties of Santa Isabel and Sorocaba, to evaluate the efficiency and the effectiveness of the FMP. The interviews were carried out between 22 June and 21 August 1987.

The population sampled comprised children enrolled in the health centres of these two counties. These children either qualified for the FMP or were well nourished. Therefore, it should be clear that the data do not refer to all the undernourished children living in the two counties but only to those identified as such by the health centre staff. The same holds true for the sampled population of well-nourished children. Table 2 presents the distribution of children.

Santa Isabel and Sorocaba were chosen both for practical reasons and to ensure that the data would come from counties representing diverse realities within the state of São Paulo. First, support was given to the research team by the officials of the state Secretary of Health in charge of the health centres in these two counties. Second, these counties have characteristics that make them useful as objects of such a study. Sorocaba (population 328,787 in 1985) is 100 km from the city of São Paulo, and is a highly industrialized county with a significant proportion of its population made up of blue-collar workers. Santa Isabel (population 36,404 in 1985) lies 40 km from the city of São Paulo, and is a "sleeper," which means that a significant part of its population works in neighbouring counties; those who make a living from rural activities are the second most numerous group.

TABLE 2. Distribution of undernourished and well-nourished children in Sorocaba and Santa Isabel counties

  Total under-nourisheda Sampling fraction (%) Sample
Under nourished or at risk Well nourished
  V. Barcelona 404 21.8 88 64
  V. Fiori 312 26.9 84 95
  subtotal 716 24.0 172 159
Santa Isabel 179 26.3 47 35
Total 895 24.5 219 194

a. Total number of undernourished children registered at the health centre.

Contents - Previous - Next