Contents - Previous - Next

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

Assessment of selected programmes

PSA (PNS) supplementary feeding programme

The PNS was planned in 1974 and initiated in 1975. Its philosophy is to integrate health and nutrition activities with the goal of multiplying the individual effects of each activity. Food is distributed free of charge through public health centres to low-income families with pregnant and nursing women and children under 7 years of age. The provision of food supplements was expected to be an additional incentive for pregnant and nursing women to come regularly to health centres.

It was assumed that the food would reach low-income families since medium- and high-income families generally use private health facilities. However, the method of estimating the target population to determine the quantities of food to be allocated to each centre was faulty. It was estimated that pregnant women represent about 5%, nursing women 1.5%, and children under 6 years of age 20.92% of the total population. Assuming the upper poverty limit to be a family income of two minimum wages, it was estimated that 57% of the Brazilian population had an income below the limit. The target population was thus defined as 15.62% of the population in each area (15.62= 27.42 x 57). Of this number, it was assumed that 21% would be pregnant mothers, 7% nursing mothers, 7% children from 6 to 11 months of age, and 65% children from 1 to 6 years of age. The strict application of this rule created significant problems, because, once the target population quota had been met, no additional families could be added to the programme regardless of whether or not they qualified. For example, in low-income areas the proportion of poor families is obviously much higher than 57%. Also, family composition is not the same for all income brackets, as was assumed [10].

The food is supplied by the Brazilian Food Company (COBAL) to the Secretariat of Health in each state; the Secretariat is then in charge of administering the programme, including distributing food to each health centre. INAN is responsible for supervising and financing COBAL.

Initially the food distributed supplied individual recipients with 25% of their daily caloric and protein requirements, but the proportion was later increased to 40%. At first each family received rations of four foods, including dried non-fat milk, sugar, and two kinds of cornmeal (wheat flour was also distributed in two southern states). Later, rice (from 1976 on), beans, and manioc flour (from 1977 on) were added. As the supply of these products varied over time, the mix of foods distributed also changed.

TABLE 4. PSA (PNS) coverage, 1975-1987

  Participants (thousands)
1975 252
1976 1,022
1977 1,806
1978 2,923
1979 2,996
1980 2,883
1981 2,510
1982 2,510
1983 4,191
1984 4,162
1985 3,697
1986 8,692
1987 9,303

Source: Refs.10,16, and 26.

Table 4 presents data on the number of programme beneficiaries during 1975-1987. The totals overestimate the coverage since individuals can be counted twice (e.g., a pregnant woman can also be counted as a nursing mother a few months later) [10]. Also, the duration of each family's participation is not known. Turnover and drop-out rates can be very important. In addition, the programme calculated food needs on the basis of individual requirements, taking into account only children 3-11 months and 1-6 years old and pregnant and nursing women, with no consideration given to intra-family leakage [27]. Although a number of evaluations of the PNS were carried out, none calculated increases in participant food consumption or considered intra-family leakages [27].

TABLE 5. Effect of participation in the PNS on the nutrition status (Gomez) of children under 3 years of age Salvador, Bahia,1976-1980

  Length of participation
Beginning 6-24 mo 24-48 mo
N % N % N %
Normal 1,512 32.4 912 45.3 844 41.6
  grade I 1,698 42.0 860 42.7 951 46.9
  grade II 678 16.8 218 10.8 219 10.8
  grade III 153 3.8 22 1.1 15 0.7
Total 4 041 100 2.012 100 2.029 100

Source: Ref. 26, table 29, p. 111.

An evaluation carried out in the state of Bahia found a significant decrease in the percentage of children having low birth weight. The proportion of participating women giving birth to infants weighing less than 2,500 g decreased from 15% for all participants to 11% for the group remaining at least three months in the programme; the figures were 37% and 17% respectively for women whose babies weighed between 2,500 and 3,000 g [24; 28-30]. Results of the most complete evaluation that analysed data collected in Bahia during 1976-1980 on 4,041 children under 3 years of age are shown in table 5 [31]. The programme is responsible for a decrease in malnourished children, but the improvement in nutrition status is concentrated among those over 1 year of age. Other evaluations, such as one carried out in São Paulo, showed no clear reduction in the frequency of malnutrition attributable to the programme [24].


Deficiencies in targeting and in operational procedures

The target population of the PSA (PNS) was women and children who attended health centres. It was assumed that such attendance was sufficient proof of a family's low income which, according to the programme's definition, was lower than 2.5 minimum wages. This criterion, however, does not take the following considerations into account: (1) not all poor people go to health centres. either because of the location or because there are none in their district; and (2) the limit of poverty and nutritional risk was set at two minimum wages with no plausible explanation given. Other studies indicate that from 3 to 3.5 minimum wages is a more appropriate poverty line [23].

There is little question that the beneficiary population would be classified as needy. The actual number of beneficiaries, however, is questionable, as the method used for adding new enrolments to the number in the previous period leads to double counting. Also, the beneficiaries are not weighted by duration of participation, which leads to overcounting, since a single visit makes a person a beneficiary. Therefore, strictly speaking, the population benefiting from the PNS was an unknown quantity, but it should be made clear that this last criticism is valid for all food and nutrition programmes in Brazil.

The actual quantities of food distributed were less than previously deemed necessary by INAN, and were reduced even further at the municipal level. The reasons included favouritism to non-eligible persons, and pressure to enrol persons who did not fulfil requirements.

Resources used in the programme tended to decline, year by year, up to 1983, the last year of analysis [23]. To maintain the 1975 quality standard, the 1983 allocation should have been at least 2.74 times greater than it was [23].


PROAB, supply of basic food in low-income areas

PROAB, begun in 1979, has as its objective "to improve the health conditions of low-income families by increasing consumption of basic food through subsidies." It started in the Recife area and later expanded to other state capitals in the north-east. Data on the potential beneficiaries are presented in table 6.

COBAL is in charge of buying and distributing food to selected retailers. It also selects participant retailers located in low-income districts, under the supervision of INAN. The retail price paid by consumers is determined by the following formula: cost of commodities purchased from producers by COBAL + 7% margin for COBAL - 20% subsidy + 11% margin for retailers (cost + 7% - 20% + 11 %) = retail price. Eleven basic products were selected: rice, sugar, dry beans, manioc flour, cornmeal, dried milk, eggs, dried meat, macaroni, fish, and edible oils.

A basic hypothesis of the programme is that undernourished families can best be reached through geographical targeting, that is, distribution of food exclusively in poor districts. Although some income-group leakages occur, their magnitude is not known.

The amount of food per capita distributed to beneficiaries decreased during 1980-1987, raising doubts about the reported coverage and about whether the family calorie and protein intake targets were being met. Reported irregularities in the supply of food by COBAL may have contributed to the problem [34]. In particular, during 1986, a year of food shortages due to increased demand after implementation of the Cruzado Plan, some foods were not regularly distributed for several months.

TABLE 6. Potential beneficiaries of PROAB (in thousands)

State capital 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
Recife 60 100 160 160 300 516 672 696 770 770
Teresina - - 120 120 200 325 325 325 325 325
Maceió - - 50 50 140 275 236 325 325 325
Joao Pessoa - - 60 60 210 260 183 183 183 183
Natal - - 70 70 270 340 333 435 519 527
Fortaleza - - - - 200 362 220 362 425 425
Säo Luís - - - - 100 165 259 256 306 306
Aracaju - - - - - - - - 32 32
Salvador - - - - - - - 212 540 540
Total 60 100 460 460 1,420 2,243 2,228 2,794 3,425 3,433

Sources: Refs. 6, 26, 32 and 33.

The subsidy represents, for all practical purposes, the only cost of the programme. Administrative costs are low, since no additional employees were hired either at INAN or COBAL. COBAL charges 7% to cover its costs and supposedly has no operational deficit.



In any subsidy programme the greatest additional consumption is obtained using products of high price elasticity, whether or not they are cheap in terms of cost per nutrient [26]. This would justify the inclusion of relatively expensive foods such as meat and fish, since the clients seldom consume such foods and would respond with greater consumption at a lower price. However, the choice of these products is questionable, given the fact that they are basically protein, whereas the fundamental problem diagnosed is a calorie deficiency. In addition, by including sugar, INAN shows that it is not concerned with the harmful effects of sugar on dental health [26].

The programme's operational efficiency depends to a large extent on COBAL's efficiency, which in turn is largely dependent on programme financing. COBAL is reasonably efficient as far as costs and food losses are concerned.

The following conclusions can be drawn from the Joaquim Nabuco Foundation evaluation [32] carried out in Recife by means of a follow-up that took place from 5 to 12 months after initial implementation:

  1. PROAB offers greater consumer freedom through respect for consumer preferences.
  2. Retailers report irregularities in product measurement and irregular supply, and difficulty returning food refused by consumers because of its poor quality.
  3. COBAL's selling prices to retailers have risen in relation to the market prices, especially those of rice, sugar, and oil and, to a lesser extent, beans, milk, and manioc flour.
  4. In the majority of cases, the prices without a subsidy would be higher than those at the supermarket. This means that the subsidy is not being passed on entirely to the clients. The same is true when compared with average and minimum prices of the mini-markets.

No evaluations of PROAB have analysed real coverage, leakages, effects on total food expenditures, expenditures and consumption of particular foods, or the nutrition status of selected age-groups.


PAP-a PROAB with no subsidy

TABLE 7. Coverage and volume of food marketed by the PAP, 1985

Cities 17
Retailers 2,828
Potential beneficiaries (thousands) 3,932
Food marketed (tons) 16,495

Source: Ref. 36.

The objective of the PAP is to supply food to low-income areas at reduced prices. It does not subsidize food prices directly. Instead, the hypothesis is that COBAL can supply private retailers at a lower price than local wholesalers and supermarkets, since it deals with large volumes and in theory acquires food directly from producers. It is expected that retailers who are able to buy stock at lower prices will also lower their selling prices.

The strategy of the PAP is similar to that of PROAB. COBAL is in charge of selecting the retailers, determining the quantities to be distributed in accordance with the number of potential beneficiaries, and buying and selling the food. The final price is the sum of the cost of acquisition by COBAL plus 7% to cover administrative costs and a margin of 11% for retailers.

The efficiency of COBAL is controversial [35]. Some have found that the prices it charges are similar to and sometimes even higher than those charged by wholesalers and supermarkets in the area [26; 32]. In other accounts, prices have been found to be lower than the retail average [36]. It has also been said that COBAL's presence is important as it may be a leader in setting prices that are then followed by the private wholesalers and supermarkets. In this case, the efficiency resulting from COBAL's price leadership behaviour is observed.

In the absence of any formal evaluation of the PAP, the data presented in table 7 give a general idea of its coverage and costs.


PNAE, the school lunch programme

The PNAE is the most traditional nutrition programme in Brazil and therefore data collection has been better, including the collection of data obtained for the beginning of the period under analysis.

A tendency toward reduction in its resources in real terms can be observed until 1985, when the total budget was increased by 121%. In 1986 there was an additional increase, in real terms, of 81% in relation to the preceding year. The data in table 8 show that the proportion of the target population participating in the programme increased during the entire 19751985 period, with a sharp reduction in the last year; the reason for this reduction has not been discovered.

During 1975-1979 the cost per kilogram tended to increase, and the quantity per recipient increased until 1977 and then decreased in 1978 and 1979; however, in 1979 it was still substantially (35%) above the level at the beginning of the period under analysis. From 1979 until 1984 the cost per kilogram of food showed a tendency to decline (with one exception in 1983), as did the quantity distributed per child; thus the cost per child also declined. In 1985 there was a significant increase in the quantity of food distributed per child from 5.7 kg the previous year to 11.1 kg, a 94% increase. The cost per kilogram also increased, but only by 10%. Therefore, although the total amount spent per child increased by 112.8%, this was accomplished partially by a reduction in the proportion of the target population served.

No specific evaluations of the school-lunch programme are available in the literature, with one exception [10]. However, since the approach of that report is different from that presented here, no evaluation is offered.

Table 8. Some performance indicators for the school-lunch programme (PNAE)

  Recipients (thousands) Population served as % of target population Food distributed (kg per recipient) Total cost per recipient (1986 Cz$) Food cost per kg (1986 Cz$)
1975 11,138 54.2 6.0 77.2 12.9
1976 11,769 55.7 10.9 119.2 10.9
1977 12,970 60.4 10.3 128.0 12.5
1978 14,070 64.8 9.7 120.4 13.4
1979 14,004 63.6 8.1 123.4 15.3
1980 15,051 68.2 6.9 99.6 14.5
1981 15,623 72.2 8.1 94.3 11.7
1982 18,720 76.6 6.8 73.7 10.8
1983 19,542 80.3 4.9 75.8 15.3
1984 20,838 83.1 5.7 63.0 11.0
1985 9.1 733 73.2 11.1 134.1 12.1

Source: Ref. 37

PAT, the workers' food programme

To date, it has not been possible to obtain the data necessary to analyse the performance indicators of the PAT. This programme, which is governed by Law 6.321 of 14 April 1976. permits the deduction of up to 5% of the taxable profit of the fiscal year for purposes of income tax owed by legal entities. This deduction is intended to cover proved expenses incurred in programmes of food for workers.

The law imposes an upper limit of 20% of the direct cost of a meal that can be charged to the employee. It also forbids the payment of salary by food, qualifies the content of the meals by fixing a minimum amount of calories that must be supplied for each type of meal (lunch, snacks), and establishes standards for responsibilities, inspection, and penalties.

The programme is aimed at feeding low-income workers to improve their state of health, increase productivity, and reduce absenteeism and work-related accidents. It began in March 1977, and by 1980 3,091 programmes had been approved, benefiting 1.7 million workers. The authorized companies were situated mainly in São Paulo (46% ), Rio de Janeiro (18%), and Rio Grande do Sul (13%), with employees in these three cities constituting 7]% of the total beneficiaries. Given an economically active population of 46 million, with 17 million salaried workers, there is room for the programme to broaden its scope.



Although three evaluations of the PAT were carried out [38-40] only the 1982 FIPE one [39] will be reviewed. It was conducted in São Paulo, where a significant number of participant companies, beneficiaries, and target populations are concentrated. The sample consisted of 77 participant companies, 50 non-participating companies, 261 families of workers, and 64 children up to 5 years of age (sons and daughters of workers). Of the participant companies, 51% had already been supplying food services before joining the PAT or before the advent of the programme, and 85% of these firms had already been doing so prior to 1975. Extension of the benefit to a greater number of workers occurred basically in the service and commercial sectors in small national firms through expansion of the ticket system. The PAT was largely unknown. In fact, 23% of the enterprises that did not participate but provided food, and 58% of those that did not participate and did not provide food had not been aware of the programme. The availability of a tax incentive was the prime motivation for participation. The nutritional benefits to the workers were no longer included as important. This held true regardless of the sector of activity or the system adopted.

Average costs per meal for the workers were lower in the participating companies than in nonparticipating ones: around 70% less in companies with their own food-supplying system and 27% less in companies using external suppliers. It was not possible to pinpoint any clear influence of company participation in terms of rates of absenteeism, dismissals, and accidents. There were indications of a positive impact only in terms of reduced turnover.

The basic reason for abandoning participation in the PAT was the low value of the incentive, which was less than the opportunity cost of using these financial assets in areas of alternative investment. Other factors were the fear of greater tax inspection, and the fact that the PAT had established a maximum cost per meal that was lower than the cost of the meal supplied. Companies that did not participate but that supplied food indicated as reasons for nonparticipation the small profit, inadequate tax incentive, and preference for other investments.

Classification of the nutrition status of the 64 children examined using the Waterlow criterion was 65% eutrophic, 26.8% with mild malnutrition, 4.7% with stunted growth, and 3.1% obese. This study found a prevalence of mild malnutrition less serious than that found in other nutrition studies carried out in São Paulo.

Comparison of two subsamples, one consisting of families whose head benefited from subsidized food (PAT families), and the other consisting of families whose head ate at home, showed that the mass of calories available was greater in the former, but the same did not hold true for proteins. Beneficiaries considered the quality of meals as average to very good and the quantity sufficient. Nevertheless, most suggestions for improvement had to do with quality.


PCA, the complementary feeding programme

The PCA is a programme of free distribution of foods, operated by the Brazilian Foundation League of Assistance (FLBA), an agency of the Ministry of Social Security and Welfare. Its main objective is to increase the food consumption of persons who suffer from, or are at increased risk of suffering from, malnutrition; its secondary objective is to stimulate the food industry [41].

Food is distributed once a month by means of existing installations loaned by the community and operated by residents under FLBA supervision. This distribution supplies 20% to 25% of the calorie and protein requirements, and is restricted to urban areas. Beneficiaries are registered individually and selected according to family income (less than two minimum wages) and biological vulnerability (small children and pregnant women).

The programme distributes formulated foods on the assumption that these are more nutritious than traditional food and that they present less risk of intra-family dilution. Since 1983 the items distributed have been gruel or sweet pap for children 6-36 months old, and a cream soup and another sweet pap for pregnant women or nursing mothers. The sweet formula uses cornmeal, corn, and soybeans; in the soups, corn is replaced by beans. Distribution allotments consist of 2 kg per month per child and 4 kg per month (2 kg of each product) per adult woman.

The programme went into operation in 1977 in the metropolitan areas of Rio de Janeiro and Belo Horizonte, and expanded over the course of the years. By 1984 it was operating in the north, northeast, federal district, Minas Gerais, Espirito Santo, and Rio de Janeiro.


Programme performance indicators

Performance indicators for this programme, presented in table 9, show a continuous increase in the number of recipients between 1977 and 1986. By the last year the number had risen to 8.5 times that in 1977. The cost per recipient had declined significantly since 1980. In 1986, in real terms, it amounted to only 14% of the 1980 level. The programme's available resources are very small. On a monthly basis, they amounted to Cz$20.4 a month per person in 1986, which equals 2.53% of the minimum wage of Cz$804 a month per person.

The cost per kilogram of food distributed through the programme decreased sharply between 1981 and 1985. The amount of food distributed per recipient also decreased, from 24.8 kg in 1978 to 17.6 kg in 1984. Therefore the programme was able to serve an increasing number of persons with a reduction in food costs, but this could mean a reduction in the quality of the food distributed.



There are two significant evaluations of the PCA programme. In terms of the choice of beneficiaries, Musgrove [26] pointed out that no criterion seems to be completely fair and acceptable for resolving the need to extend the benefit to some persons and not to others, and that, therefore, biological vulnerability (small children and pregnant women) makes about as much sense as any other. This does not mean that in addressing these groups the PCA is reaching a proportion that reflects the real composition of the poor population, since coverage of children has been observed to be much greater than that of adult women [26].

Regarding the selection of food, Musgrove compared the cost per 100 calories of formulated foods distributed by the PCA with that of the natural food staples distributed by the PSA and other programmes [26]. He determined that salted formulated foods cost more in cruzados per calorie than do bread, noodles, cornflour, rice, or the rice-bean-egg and noodle-egg-soybean mixtures, and less than milk, eggs, or the milk-cornflour-sugar mixture. Sweet formulated foods cost more than any other alternative except the milk-cornflour-sugar mixture.

With regard to total costs and their composition, the estimates for one unit in Rio de Janeiro (1980) presented a total cost of Cz$83.74 (US$1.68) per kilogram. This comprised the costs of food, transportation, storage, personnel and "other" items. Food alone corresponded to 83% of the total cost.

The FLBA/UNICEF evaluation [42] consists of a research project carried out in the federal district in March 1978 prior to implementation of the PCA, and another in September 1978 six months after the PCA went into full operation.

The main results were as follows. First, the real expenditures for food rose approximately 8%. Second. the programme was considered a success because low intra-family dilution of the formulated foods was found. It had little or no effect on the height of the children served, however, which seems to indicate that the quantity of additional foods consumed and/or the interval of availability were insufficient to achieve more significant results. Finally, the worsening of the health of some children reflected conditions that are not directly nutritional (e.g., diarrhoea and infections), which would seem to indicate inadequate integration of food aid and health service.

An FLBA evaluation of birth weights in 1984 studied 36,105 births in Ceara and Rio de Janeiro [26; 41] In all, it appears that the PCA reduced the frequency of low birth weight (< 2,500 g) by 1.6%, increased the frequency of birth weight over 3,000 g by 2.2%, and reduced the frequency of deficient birth weight (2,500-3,000 g) by only 0.6%.

An anthropometric analysis of 1,086 children under 3 years of age was conducted in the federal district in 1978 at the beginning of the programme and after six months. The only information available is weight for age. Over half of 651 children began at a normal weight for their age, and almost all maintained this status; only 62 worsened over the six-month interval. Of the 435 children who initially had low weight, 48% remained in this category, 49% improved (almost all passed from degree 1 malnutrition to normal), and only 3% worsened.

The total number of seriously undernourished children was reduced to almost half (largest relative change), but in absolute terms the most frequent change was from slight malnutrition to normal. In all, those who improved were almost three times more numerous than those who worsened.

The typical cost was calculated using food quantities that would ensure that a child would gain weight normally or have improved nutrition status. It was calculated that a total of US$151.80 was required to maintain normal weight for six months, while US$398.50 was required to maintain normal weight for one year. To ensure an improvement of the nutrition status in six months, US$271.35 would be required. In the face of such high costs, the evaluators recommended restrictions on food distribution and an anthropometric criterion for enrolment in the programme. However, this idea of excluding the undernourished would be equivalent to changing from a preventive and non-clinical programme to one with a basically curative emphasis.

TABLE 9. Some performance indicators for the complementary feeding programme (PCA)

  Number served Food (tons) Cost Food per person (kg) Cost per person (Cz$)
Cz$ thousands Cz$ per ton
1977 242,041 7,014 - - 28.98 -
1978 285,862 7,094 156.2 22.02 24.82 0.546
1979 292,190 9,216 211.8 22.98 31.54 0.725
1980 289,707 9,000 236.4 26.27 31.07 0.816
1981 350,000 9,023 231.2 25.62 25.78 0.661
1982 388,590 10,683 211.5 19.80 27.49 0.544
1983 700,000 12,990 158.9 12.63 18.56 0.227
1984 750,000 13,203 121.5 9.20 17.60 0.162
1985 997,000 - 112.3 - - 0.113
1986 2.060.491 - 245.0 - - 0.119



A synthesis of recent experience with nutrition interventions in Brazil shows patterns similar to those found in other countries, as would be expected. First, multisectoral planning, which embodies the idea of a global attack on nutrition deficiencies and poverty, began in the early 1970s and received support from policy makers throughout the decade. But the approach had disappointing results in the early 1980s owing to political limitations. After the end of a decade of multisectoral planning, interventions were more supplemental, targeting high-risk population groups and sometimes coupled with health interventions. Therefore, reviews of evaluations of the interventions [4; 26; 43] indicate that improvements in health-care practices resulted in significant decreases in the percentage of low-birth-weight children and in morbidity and mortality rates, compared to the meagre results obtained through food- and income-transfer programmes [44: 45]. In addition, nutrition interventions, in isolation from health actions, can at best assume a preventive character, given the worsening social conditions associated with the problems of external debt and structural adjustments.

Finally, current nutrition interventions have been expanded and new programmes defined in the past two years, but no evaluations have been conducted to measure their impact. Some reports highlighted certain problems and pointed to overlapping coverage in certain areas [26; 46]. Thus, there is an urgent need for evaluation and revision of some programmes and, perhaps, a reduction in the number of programmes to simplify central administration and budgeting.

Contents - Previous - Next