Chapter 3. Why did Nutrition Improve?

Income-Related Variables
Entitlements

In the previous section we have shown that impressive nutritional improvements occurred during the seventies and eighties. These results were obtained despite a poor economic performance during most of the past decade.

Three points have to be re-emphasized:

i) the sharp decline in infant mortality started when the economy was losing its momentum, that is, after the peak years of the first half of the 1970's;

ii) the descending trend continued even during the stagnating years of the 1980's; and

iii) the gains were generally greater in the rich Southeast region that in the poor Northeast, and for the families with relatively higher incomes than for the poorest.

Income-Related Variables

Poverty
Inequality

Poverty

Malnutrition is unquestionably a phenomenon of poverty. In this sense, it becomes an important point of analysis in this paper the fact that the income data available for Brazil indicate that the poor also gained from the rapid expansion of the Brazilian economy during the 1970's.

Table 12 displays the published figures for absolute poverty in Brazil and the Southeast and Northeast regions.¹ It shows a substantial reduction in the proportion of Brazilians living below the poverty line during the two decades²: while in 1970 almost 59% of the Brazilian population was in extreme poverty, in 1989 these numbers were reduced to 17%. The gains were considerably greater in the seventies, particularly in the first half of the decade, than in the eighties. From 1970 to 1980 the proportion of poor people fell from 58.7% to 25.7%, a reduction of 56%, of which 94% (or a 53% reduction), occurred between 1970 and 1976: if the average annual reduction in the proportion of poor for the entire decade of the seventies was of about 7.9%, for the 1970-1976 period alone this average was of 11.7%. During the 1980-1989 period the proportion of the Brazilian population living below the poverty line was reduced from 25.7% to 16.9%, a 34% decline (annual average of 4.55%).

¹ The figures for 1976 and 1989 are non-published estimates that we calculated based on secondary data. They are included in the table to complete the series.

² The reference for Brazil is to consider below the poverty line those individuals from families in which the per capita income is less than one-quarter of the legal minimum wage (a value of about US$25 to US$30). This criterion seems to satisfy the parameters of scarcity that characterize extreme poverty and has been used in works such as Pastore et al. (1983), Jaguaribe (1989), and Tolosa (1991).

Table 12
Proportion (%) of People and Families Below the Poverty Line
Brazil, Northeast and Southeast
1970, 1976, 1980, 1981, 1983, 1986, 1987, 1989

	1970	1976	1980	1981	1983	1986	1987	1989
Brazil
People	58.7	27.8	25.7	21.0	20.8	13.1	14.4	16.9
Families	48.2	22.0	17.7	15.8	15.4	9.7	10.8	14.2
Northeast
People	79.2	50.3	49.6	41.5	40.2	28.6	31.3	34.0
Families	72.9	41.3	32.5	34.2	32.5	22.9	25.0	29.6
Southeast
People	-	17.8	12.0	10.5	10.6	5.9	6.3	8.8
Families	32.6	12.4	10.1	7.6	7.8	4.3	4.7	8.0

Sources: Tolosa (1991); IBGE - Indicadores Sociais, 1979, 1984; IBGE - Perfil Estatístico de Crianças e Mães no Brasil, Several years; Pastore (1983); PNADs, several years.

These published figures should be viewed with caution, though, for since they refer to the proportion of people and families with a per capita family income of less than 25% of the prevailing legal minimum wage, they are the result of a ratio between two variables. In this sense, if the denominator (the value of the minimum wage, an exogenous variable defined outside the market by the government) increases relatively less than the numerator (the revenues and wages obtained by the economic agents in the market), reflecting, for instance, a policy of adjusting the minimum wage below the inflation of the period, many people and families might be artificially placed above the poverty line.³

³ Obviously the opposite is also true. New people may be incorporated inside the limits of absolute poverty if the minimum wage is adjusted at levels above those observed in the market.

⁴ Figures from Tolosa (1991), using a similar procedure, are also displayed in Table 13 as control.

⁵ We do not have an estimate for 1976 for the country as whole. However, from the data available in Table 13 for the Northeast and Southeast regions, and from the fact that the minimum wage in 1976 was greater than in 1980 in real terms (meaning that the figure present in Table 12 actually overestimates the proportion of poor), it is safe to say that the gains of the 1970's were, to a greater extent, obtained during the first half of the decade.

⁶ Tolosa shows an increase from 26% in 1980 to 32.5% in 1988.

⁷ Please note that even though not all years are present in Figures 8 and 9 we have maintained a line joining the points as a visual aid.

⁸ With the increase in inflation in 1979 wages began to be corrected twice a year for the increase in prices. Wages of workers receiving less than three minimum wages were corrected with 110% of the accumulated inflation of the previous six months; those workers receiving between three and ten minimum wages had their earnings adjusted in 100% of the inflation rate of the period; the adjustment for those receiving more than ten minimum wages was of 80% of the inflation rate. In 1980 the distributive aspect of the policy was reinforced with the limit of 50% of the inflation rate for wage-earners between fifteen and twenty minimum wages and free negotiations for employees receiving more than twenty minimum wages.

Table 13
Proportion (%) of People Below the Poverty Line
Minimum Wage of May 1980 as Basis
Brazil, Northeast and Southeast Regions
1970, 1976, 1980, 1981, 1983, 1985, 1986, 1987, 1988, 1989

Year	Brazil	Northeast	Southeast
1970	58.2(54.1)	(76.8)	(37.7)
1976	-	49.7	17.6
1980	27.0(25.7)	(49.6)	(12.0)
1981	24.7	-	-
1983	25.2	-	-
1985	-	43.8	12.8
1986	24.3	-	-
1987	34.5	-	-
1988	38.2(32.5)	(58.8)	(18.5)
1989	35.4	60.3	23.5

Sources: Tolosa (1991); PNADs several years.
Note: Values in parentheses are from Tolosa.

Figure 9. ABSOLUTE POVERTY & ECONOMIC GROWTH - BRAZIL, 1980-1989

Figure 10. REAL WAGES: SEMI-ANNUAL VALUES - BRAZIL, 1979-1983

The regional data for the Southeast and the Northeast show the same pattern previously described, with the gains from the seventies, however, accruing to a much greater extent to the Southeast region: the reduction in poverty in this region is of about 68% between 1970 and 1980, while in the poor Northeast region the decline in the proportion of people below the poverty line, during the same period, is of 35%. Moreover, as a result of the deteriorating conditions of the 1980's the absolute number of extreme poor people in the Northeast is greater in 1989 than in 1970: about 25.2 million compared to 23.2 million.⁹

⁹ For the Southeast region the numbers remain about the same: 14.7 million people below the poverty line in 1989 and 15.4 million in 1970.

Inequality

From the value of income we now turn to the issue of the distribution of income. Table 14 shows the values of the Gini coefficient for several years. Increases in inequalities endure throughout the sixties and first half of the seventies: 0.497 in 1960, 0.571 in 1970, and 0.589 in 1976. The data available describe a stabilization of the coefficient during the second half of the decade, between 1976 and 1980. The 1980's presents an upward trend that is broken only in 1981 and 1986: in 1981, probably as a result of the wage policy discussed above (see note 19), and in 1986 due to the distributive effect brought by the already mentioned "Cruzado Plan". Once again the impact of inflation should be noted: the Gini coefficient displays a raise from 0.588 in 1986 to 0.595 in 1987 followed by two sharp and consecutive increases in 1988 and 1989 when the annual inflation rate reached four digits. Figure 11 presents this direct association between inflation and income distribution: for the 1980's both variables follow an markedly similar tendency, substantiating our conjecture that high inflation had a very damaging impact on the poor.¹⁰

¹⁰ We could not find as straight relationship between the Gini coefficient and economic growth. A clearly opposite correlation was only obtained using a moving average of the rate of growth of the preceding three years, the reason for such period is not clear though, except for the fact that we would expect the patterns of income distribution to take longer to respond to variations in economic growth (considering that the structure of such growth is inducing changes in the distribution of income), than the values of earnings.

Table 14
Gini Coefficient
Brazil: 1960, 1970, 1976, 1980, 1981, 1983, 1985-1989

Year	Gini Coefficient
1960	0.497 (1)
1970	0.571 (2)
1976	0.589 (1)
1980	0.590 (1)
1981	0.573 (2)
1983	0.592 (2)
1985	0.600 (2)
1986	0.588 (1)
1987	0.595 (2)
1988	0.619 (2)
1989	0.636 (2)

Sources:

(1) Reis, Rodriguez e Barros (1991);
(2) IBGE - Indicadores Sociais, 1979, 1984

In the past decade we continue to observe an overall tendency of decline in the infant mortality rate. Our income indicators, however, show increases in the proportion of people below the poverty line and a deterioration in the distribution of income, particularly in the last years of the decade. In this sense, the explanation for the continued improvement must lie with other factors than the income variables. It is to the analysis of these elements that we now move to.

Entitlements

Nutrition Programs and Food Prices
Other Social Areas

Drèze and Sen (1989, p. 23) define the entitlement of a person as "the set of alternative commodity bundles that can be acquired through the use of the various legal channels of acquirement open to that person." Within this framework we have to stress the importance of those goods provided through the channels set by the social policies.

Sen has shown in several works that entitlement failures have been the major causes of famine and chronic malnutrition.¹¹ In this subsection we will apply this framework of analysis to help explain the improvements in health and nutritional status observed in the past two decades.

¹¹ See for instance: Sen, 1977, and 1981.

Nutrition Programs and Food Prices

Brazil has had a diversified experience with food and nutrition programs: from food subsidies to the direct distribution of foodstuffs, through market channels or from public facilities. As Musgrove (1990) puts it: "Among Latin American countries, Brazil has undoubtedly had the most varied and extensive experience with food and nutrition programs meant to improve food consumption and nutritional status of the children of poor families." Another important characteristic of the Brazilian case is that such programs have been almost entirely supported by national, i.e. non-aid, resources.

The monetary values and volumes involved in such programs are substantial: a single program of food supplementation for pre-school children distributed more than 1.5 million tons of food between 1976 and 1987; expenditures with food and nutrition programs have increased from 0.06% of the GDP in 1980 to 0.21% in 1989 (with a peak of 0.25% of the GDP in 1987). Table 15 displays the growth of expenditures with food and nutrition: it shows an increase of almost 300% between 1980 and 1989, from about 210 million dollars (US dollars of 1989) to about US$ 805 million. During the peak year of 1987 the amount destined for the area was 4.6 times that of 1980, reaching almost 1 billion dollars. The share of food and nutrition of the total expenditures of the social sector increased from the 0.6% in 1980 to more than 2.5% during 1986 and 1987, and ending the decade with 1.9%. From Figure 12 it is clear that the expenditures with food and nutrition constituted the fastest growing area of the social sector during the last decade.

Despite some impressive figures and ambitious goals, the Brazilian food and nutrition programs have presented several problems (see Peliano, 1992), the two most important are:

i) the programs have not reached the poorest regions and income groups and the most vulnerable age groups; and,

ii) the dissociation of the programs from health and educational efforts.

Table 15
Expenditures with Food and Nutrition
Brazil, 1980-1989

Year	Expenditure (US$ million of 1989)	Index 1980 = 100	% GDP	% Social Expenditure
1980	210.34	100.00	0.06	0.63
1981	219.79	104.49	0.06	0.64
1982	276.24	131.33	0.08	0.77
1983	306.91	145.91	0.09	1.00
1984	332.35	158.01	0.09	1.25
1985	609.61	289.82	0.16	1.96
1986	903.11	429.36	0.22	2.57
1987	967.80	460.11	0.23	2.58
1988	889.13	422.71	0.21	2.28
1989	805.67	383.03	0.18	1.94

Sources: Piola, Vianna and Camargo (1992).

With respect to the first limitation we could cite two examples:

i) the worker feeding program (PAT, Programa de Alimentação do Trabalhador), destines the largest amounts per assisted person of all area programs, US$ 45 (1986 value. World Bank, 1988), but benefits workers from the urban modern sector of the economy, therefore mainly reaching areas less affected by malnutrition. Under the PAT companies distribute coupons with monetary values that can be used for the payment of meals in any participating establishment. The amount spent by the company with such coupons are tax deductible;

ii) the school meal program is targeted at public primary students. As with similar experiences of other countries, this program has a large rate of mistargeting, for it reaches basically older, thus less vulnerable, urban children and many of its beneficiaries are children without nutritional deficits. Moreover, it is not known to what extent it substitutes for food at home rather than supplementing it. The fact that most programs are poorly linked to health and educational efforts is one of the most important weaknesses of the Brazilian intervention in the area of food and nutrition for some of the potential that these programs would bring are lost either by poor health conditions or household behavior. As a result of these limitations the direct assessments of the Brazilian programs show their lack of effectiveness, with results that do not correspond to the goals and amount of resources invested.¹²

¹² See Musgrove (1990), and also Peliano (1992).

¹³ As Sen (1977) has shown, food scarcity is not a necessary condition for hunger.

¹⁴ The basket is constituted by: meat, milk, beans, rice, wheat flour, potato, tomato, bread, coffee, banana, sugar, grease/oil, and butter. These products are used in the calculation of a Cost of Living Index for the city of Sao Paulo by a confederation of Labor Unions.

¹⁵ Note that the price index used to measure the inflation rate is national and the cost of the basket is for the city of Sao Paulo.

¹⁶ The fact that the real cost of the food basket decreased during four years, from 1974 to 1977, reinforces this conclusion.

Figure 14. INFLATION RATE AND FOOD PRICES - BRAZIL, 1970-1989

Figure 15. ANNUAL REAL CHANGE: WAGES & FOOD PRICES - BRAZIL, 1971-1983

Other Social Areas

There is no consistent source of data on social expenditures that cover the twenty years of interest of this paper, as a result we mainly used three different types of information:

i) the Institute of Applied Economic Research (IPEA) from the Ministry of Economy has started an accounting of the Federal Government's social expenditures in 1980. It is, in our opinion, the most reliable and methodologically correct data set available, though unfortunately it only covers the last decade (see Piola, Vianna and Camargo, 1992);

ii) FUNDAP, an administrative supporting agency to the state government of São Paulo, has done an analysis of social expenditures covering the 1975-1986 period. We have found some problems in specific areas in their data, even though the aggregate estimates seem adequate (see FUNDAP, 1991);

iii) finally, we have used data published by the Federal Revenue Service on budgetary expenditures for the 1970-1985 period. The problem with this source is that for many social areas non-Treasury resources account for the bulk of expenditures. Due to these limitations, we will use combinations of these and other sources in our analysis.

¹⁷ This practice is in fact a mechanism of self-defense against budgetary cuts.

water and sanitation: One of the most important entitlements in determining health and nutritional status is access to clean water and a sewage system. It is estimated that, as of 1967, 65% of the Brazilian population did not have access to an adequate water supply, and even less so to the sewerage system. The political and social pressures that were emerging with a growing urban population has led to the creation in 1970 the PLANASA (National Sanitation Plan), an ambitious program with the aim of delivering potable water to 80%, and sewage systems to 50% of the urban population by 1980.

Figure 16 displays the expansion of the water coverage brought about by the PLANASA for the 1970-1984 period, for the country and the Southeast and Northeast regions. For the country as a whole it is clear that the rate in which the program developed increased during the second-half of the seventies: for the 1970-75 period the population served by the program rose by 91.3%, averaging an annual increase of 13.85%; the expansion of the second-half of the 1970's (1975-80) was of about 118.1%, averaging 16.9% per year; despite the economic crisis, the program also expanded during the 1980-84 period almost 28%, or 6.35% per year on average. By the end of 1984 more than 63 million people had benefited from the program, reflecting a cumulative growth of 434% (about 12.7% per year). Table 16 displays the percentage of the urban population with access to a water supply in 1970,1980 and 1984. It shows that the proportion of the Brazilian urban population receiving clean water increased from 51% in 1970 to 84.5% in 1984.¹⁸ The reduction in the number of people without any water source in their households during the 1970's should also be noted: from more than one-fifth of the urban population (21.6%) in 1970 to less than one-tenth (8.3%) in 1980. With respect to later years of the 1980's, the evidence available indicates an average annual increase of about 3.7%, from 1980 to 1988, in the proportion of total households, that is urban and rural, having access to potable water, thus confirming that the expansion continued throughout the eighties.

¹⁸ 81% of the 84.5% of the urban population receiving clean water (or 68.3% of the total), were beneficiaries of the PLANASA.

¹⁹ As of 1984 about 52% of the population served in terms of water by the program were from the Southeast and 22% from the Northeast region, exactly matching each region's participation in the Brazilian urban population. In this sense, it can be said that if, with respect to water, the program did not present a tendency for equity, it did not operate in a regressive manner either.

Table 16
Urban Population: Access to the Water Supply
Brazil: 1970, 1980, and 1984

Urban Population	1970	1980	1984
% With access to the water supply	51.2	73.9	84.5
% Without any water source inside the household	21.6	8.3	5.4

Source: MPAS/CEPAL (1990).

Table 17
Urban Population: Access to the Sewage System
Brazil: 1970, 1980, and 1984

Urban Population	1970	1980	1984
% With access to the sewage system	20.2	33.3	36.0
% Without any sanitary installation inside the household	13.4	7.2	5.5

Source: MPAS/CEPAL (1990).

Of the total invested in the country in the areas of water and sanitation between 1968 and 1984, 72% were oriented for the former. Of this total invested in the water system 25.6% were destined to the Northeast and 55.1% to the Southeast regions. With respect to the investments in sanitation the Northeast region received only 11.5% and the Southeast the lion's share: 75.6%.

health: With respect to the health sector we could only construct a series on per capita health expenditures from 1975 to 1989 (see Figure 18). The data from the FUNDAP indicates an upward trend from 1976 to 1982. It should be noted, though, that there are large discrepancies between the FUNDAP data and the IPEA data set for 1980 and 1981. Both sources show, however, lower levels of expenses for 1983 and 1984 as a result of the economic crisis. A new and important upward trend appears from 1985 until 1989, with the peak being reached in this last year with an expenditure of US$68.73 per capita, an amount 2.4 times greater than the quantity spent in 1975.

Table 18 displays the availability of physicians and pediatric beds, the number of medical visits paid by IN AMPS²⁰, and the immunization coverage of public health centers. Particularly with respect to the latter the improvements are impressive: the proportion of children immunized against measles increased more than three times, from 16% in 1975 to 59% in 1988, and the proportion of children receiving all three doses of the vaccine against diphtheria, tetanus, and whooping cough was raised from 20% to 56% in the same period. The number of physicians per thousand of the population doubled between 1970 and 1986, while the availability of pediatric beds increased by 80% during the period. For both indicators the growth observed during the second half of the 1970's was greater than in the other period. The number of medical visits paid by INAMPS per year between 1975 and 1986 increased in about 78%, with almost 94% of such growth occurring between 1975 and 1980. Therefore, the figures from Table 18 suggest an improvement in the access of the poor to health services during the two decades, particularly because, with the exception of the number of physicians, all of the estimates refer to the public sector.

²⁰ INAMPS is the single largest source of health care financing in the country, with expenditures of about one-third of the country's total (public and private) health expenditures. Its funds come from payroll taxes levied on workers and employers.

Table 18
Health Care Coverage Indicators
Brazil: 1970, 1975, 1980, and 1988

Coverage	1970	1975	1980	1988
Immunization %
DPT(1)	-	20.0	37.0	56.0
Measles	-	16.0	56.0	59.0
Physicians(2)	0.7	0.8	1.1(5)	1.4(6)
Pediatric Beds(3)	2.0	2.5	3.5(5)	3.6(6)
Medical visits(4)	-	0.9	1.5	1.6(6)

Source: Monteiro et al. (1992).

Notes:

(1) Diphtheria, Pertussis, and Tetanus
(2) Per thousand inhabitants.
(3) Per thousand children with less than 5 years of age.
(4) Per thousand inhabitants per year.
(5) 1979 data.
(6) 1986 data.

Table 19
Literacy and Basic Education
Brazil: 1970, 1977, 1980, and 1988

	1970	1977	1980	1988
Literacy rate (%)
Population age 10 or more	65.9	74.2	74.7	81.5
Women 25-29 years of age	67.7	80.6	81.4	90.3
Basic education* (%)
Population age 10 or more	31.7	48.6	47.9	57.1
Women 25-29 years of age	35.5	56.3	60.2	72.3

Source: Monteiro et al. (1992).
Note: Refer to having 4 or more years of education

²¹ In the case of education, Treasury resources represent over 90% of the sector's expenditures.

It should be noted that in the context of this paper the benefits of the expenditures in education materialize with a considerable lag in time. The trend of the applications in the area is reflecting the commitment to the social policies. In addition, they provide an indication for this and the coming decades.

In this subsection we have shown that education, health, sanitation, and particularly water all show important developments and a pattern of progress closely related to the improvements described in Section III. Moreover, these factors have presented increases throughout the 1980's, allowing for the continued improvements observed in that decade.

Food entitlements acquired through non-income channels apparently have had a limited impact on the improvements observed and described in the previous section due to the low effectiveness of the Brazilian nutritional programs. The behavior of food prices, in aligning very closely to the overall inflation schedule, while thus not a contributing factor to the improvements, did not represent an obstacle either.