Economic Growth, Poverty and Income Distribution in Brazil Carlos Aguiar de Medeiros

(Paper prepared for the Conference “New Ideas on Development Economics Growth and Distribution Under Financial Globalization”. Rio de Janeiro, January, 28 and 29, 2002)

What distinguishes Brazilian society today is not only the poverty of 31% of its population – 54 million inhabitants in 2000 – but the extraordinary income gap between the very rich and the poor people . According to the World Bank Indicators for 2001, amongst 152 countries only 5 – Central African Republic, Nicaragua, Sierra Leone, South Africa, Swaziland – have an income Gini index higher than Brazil. This economic contrast in a medium income level economy characterized by the diversity and sophistication of its markets is an indisputable source of a high social distress and growing urban violence. A huge fight against the poverty and the extreme income concentration is a challenge for any progressive government in Brazilian society.

Despite the complexity of this challenge there is nowadays in this country, as well as in most developing countries, a persistent mantra, strongly diffused by World Bank studies on poverty. According to this, investment in education is the most efficient and adequate structural action to achieve a higher equity in income distribution and lower poverty levels. In addition to this policy, the state, runs the dominant argument, has to provide a safety net to the very poor redressing its social policies as pensions, aids, educational and public investments from the richest to the very needy. Although economic growth is assumed to be a good thing for reducing the absolute poverty, it is not considered a central priority. In this approach economic growth is not a deliberate target for economic policy and with correct income distribution policies, higher poverty reduction can be achieved with lower rate of income growth.

The core theory underlying this message is the human capital. The strong correlation in the Brazilian labor market between income levels and formal education and the low years of schooling of the majority of employed population is considered an indisputable confirmation of this assumption. According to this theory, the high growth that occurred in the seventies was high-skilled demand biased. In consequence, the income inequities were increased. After the trade openness in the nineties and in presence of fast rate of industry productivity this mismatching in demand and supply of qualifications has improved.

It is not my purpose in this opportunity to criticize the inconsistencies of this wage theory based on demand and supply of labor in a competitive market or its accuracy in describing the evolution of Brazilian relative wages. (It is indeed a waste of time for the people that believe in this neoclassical theory there is a faith principle in market mechanism, and no word can challenge it, for the people that don’t , as I assume to be the case here, there is no need to say any additional word).

What I would like to stress is some curious coincidences of this message with some propositions developed from different perspectives. In fact, amongst the heterodox thought there is a kind of reasoning that professor James Galbraith denominated the “supply side progressives” . From this point of view the new technologies and globalization forces have sharply increased the demand for high skilled labors increasing the polarization in labor market. To acquire a competitive edge in global economy the country has to develop and adapt new technologies and unless a massive educational investment, this possibility can not be achieved or can result in great differences in economic opportunities among firms and workers.

Underlying to this same conclusion – although vested in different arguments- is the assumption that income distribution is essentially related to labor market microeconomics and can not be explained by macroeconomics forces.

From the left-wing political specter, by its turn, although there is a strong criticism of macroeconomic policies and growth stagnation that occurred in Brazilian economy in the last decade, a skeptical hope about the positive impact of economic growth in income distribution prevails. The main argument is based in Brazilian long run performance and assumes that economic growth and income concentration that took place in Brazil from the fifties through the eighties were parts of the same process. Although there is in this reasoning a defense of a macroeconomics oriented toward an employment generation, the emphasis is not on speed of the economic growth but in its quality. Economic growth is considered harmful or good for the poor people depending on qualitative and institutional aspects achieved by distributive policies.

Before I go any further along this line of thought its is important to say that the orthodox thought on income distribution has imposed an informational and data basis that is very difficult to escape in international comparisons. In fact all we discuss about income distribution is basically based on labor income, the rule of state on income transference is, by its turn, confined to the social budget. This suits very well the main concern of mainstream thought centered on labor market microeconomics but creates difficulties for other kind of analysis. In fact if we look for changes in the labor share or the impact of the real rate of interest in the income transference from the state to the asset owners, the data have not the same quality or can not be used for international comparisons or for long trends. This, I think is an important area for a network as IDEAS.

Before we go to Brazilian experience let us address our attention to the relationship between growth, poverty and income distribution. A change in poverty levels can be split in two components: a growth component, depicting the effect of a change in the income mean while the distribution (a Lorenz or Gini) remained constant; and a redistribution component when the change in poverty is due to a change in redistribution while keeping the mean income constant. When in an influential paper published by World Bank, David Dollar and Kraay (2000) argued that “Growth is Good for the Poor” they assumed a direct relation between the mean and the poorer income strata. Centered in a long span of time – 20 years- they considered a homogeneity in distribution along all the period sustaining the conclusion that the growth in lower income recipients is not very different from the mean growth.

This conclusion is very difficult to be sustained empirically or theoretically. Comparing the Brazilian and Indian experience during the eighties, Datt and Ravallion (1992) found that in Brazil’s case “The headcount index of poverty would have fallen by a 4.5% points over the period if only growth had been distributionally neutral. By contrast, distributional effects contributed to the alleviation of poverty in India between 1977 and 1988, though growth accounted for the bulk of the improvement.”(291)

Recently, Pedro Sainz and Mario la Fuente from ECLA showed (2001) that in the last two decades the low growth in Latin America was strongly uneven; during the eighties and the nineties, the poor household had a much lower income progress than the median income growth. Its is interesting to observe that if in the eighties the collapse of income growth affected badly the poor households, in the nineties, with the low growth that occurred in many countries, the poorer households did not recuperate what they have lost. As put by the authors, “there was an asymmetry in Latin America between crisis and growth: income concentration in the former and rigidity in the later” (170).

The evidences seem to show that there is no systematic relationship between the evolution of inequality and growth performance. This is true and if we are considering the wage share, this is what we expect from the classical political economy. In Smith, Ricardo or Marx, the rate of economic growth affects indirectly the income distribution through a positive influence on working class bargain power. But in the Brazilian economy as well as many populous developing countries, we have to consider the presence of a surplus labor as a distinguished aspect of our economic reality. This point of departure for income distribution studies was strongly developed in Latin-American structuralist school in sharp contrast with the mainstream analysis based on full employment assumptions. From this perspective decreasing structural unemployment was strongly necessary for a sustained and more equal growth.

In Brazilian society the poor people are major formed by working poor, the rate of activity of the poor households are not different from the richer ones. For the majority of people, to remain unemployed is a luxury living position. Because the poverty is much more intense in informal sector, the expansion of formal sector resulting from economic growth, has a direct and not only indirect effect on household income distribution. The faster is the rate of economic growth the faster will be this structural change. In presence of a large surplus labor, a sustained high rate of economic growth can bring about a better income distribution provided that the demographic expansion does not compensate this contraction on surplus labor.

How much growth is necessary to create enough jobs and which kind of jobs benefited more the poor people?

As we have seen, one critical point from Brazilian contemporary reality is the hypothesis that globalization and new technologies have increased the demand for high skilled workers, so, say neoclassical economists or “supply side progressives”, economic growth will be accompanied by a income concentration unless a strong supply of new skilled workers occurs. Other criticism comes from empirical studies in Brazil during the nineties. The conclusion is that employment elasticity to income growth has declined thanks to labor saving techniques of modern activities. The two arguments say that the aggregate employment creation associated to this modern trend in economic growth will be low in comparison with the past and will be skill intensive and biased to the rich.

To answer to this last objection I would like to say that I strongly agree with Professor Galbraith when he said that “We should not be surprised to see a relationship between technological change, unemployment, economic growth and inequality. In the slump, when unemployment rises, the wages of the lowest-paid workers suffer because they are the least well protected by their employer’s monopolistic positions. Inequality rises. As the recovery starts, moreover, the first new jobs fall to highly paid workers …and inequality rises again.” Only when “Existing capital is exploited more intensively, so additional employment come, not mainly in the investment sectors, but in the production of consumption goods and services. It is at such times that we should expect to observe decreasing inequality in the structure of wages” (pg. 63).

The main conclusion is: what is important for the low skilled workers is the persistence of high economic growth and not a short sequence of bust and boom that characterized the Brazilian and Latin-American economies in the nineties. This volatility in economic growth due to macroeconomic forces is the basis of what Pedro Sainz from ECLA has said about the asymmetric growth in Latin America and seems to play a much more important influence on income distribution than a skill intensive technologies.

Up to now, all we have said about the influence of economic growth on income distribution was confined to relative wages. A structural change in employment appeared in this analysis as a route of wage increases through worker mobility from low productivity to high productivity activities. But what distinguished the classical and macro structuralist analysis of income distribution was the assumption that the real wage rate is institutionally formed.

Given the existence of a surplus labor, the relevance of institutional minimum wage for income distribution is really effective and in Brazil, as well as in developing countries the living wages are strongly dependent on food prices. The influence of the institutional minimum wage on poverty levels is no more in dispute in Brazil, even the mainstream labor economists agree, based on empirically studies of the nineties, that increases in real minimum wage diminish the poverty and vice-versa. For non neoclassical labor economists there is not only an effect on poverty but in wage distribution as well: the range of wages is wider and the gap of skilled and non skilled wage is bigger when the minimum wage is low and vice-versa.

One important fact of Brazilian industrialization, a key aspect for the persistence of high income inequality, was the low real salary paid to non qualified labor. To understand this is necessary bring to the picture the unbalance growth between industry and agriculture, and the institutional and macroeconomic aspects that through relative prices shape our distributive reality. Those factors, in my point of view, are essential to explain the frustration in economic growth to provide a better income distribution as occurred in our “developmental” years. Let us now consider briefly the Brazilian experience.

As we said, most of criticism of the positive influence of economic growth on income distribution in Brazil between the fifties and eighties, was based in is fact: there was enough job creation, the informal sector diminished, despite that, a sharp income concentration took place along these decades. The point I would like to emphasize is that the years of high economic growth in Brazil do not fail to provide a huge fall in absolute poverty through a high labor mobility. A high speed of urbanization that occurred along these years was not, however , only the consequence of fast industrialization but the mixed result of demographic explosion with huge migration of millions of poor people from the countryside and small villages expelled from a miserable and low productivity agriculture. During this phase the modern agriculture was essentially export driven centered in few commodities employing relatively few (and underpaid) people, and land proprietory was highly concentrated . The macroeconomic regime of high growth was achieved by a stable but chronic rate of inflation, a high indexed rate of exchange and under indexed minimum wage. Thus our “golden age” of economic growth was accompanied by high rural poverty, intense migratory flux, permanent increases in food prices, declining minimum wages and labor class political repression in the big cities. In this context the wages paid to skilled labor, the rate of profits in modern industry and land rents in rural and urban propriety shaped a high concentrated income distribution.

The effects of income growth and the distributive effect associated to structural change were positive on reducing poverty but the evolution of relative prices and wage policies avoided a real income growth of urban and rural low skilled labors.

During the seventies some structural changes start to transform this perverse style of growth. In fact, the persistent industrial growth brought about a modern labor class with growing economic strength, a better wage policy created in the middle of this decade, provided a real gain in minimum wage. As a response of a steady urban food demand, new private investments in food production were realized and new public infra-structure was built. The trend on income concentration was halted and the wage share increased. All this facts signalized that Brazil was on the threshold of another pattern of growth.

The macroeconomic crisis of the eighties and the unstable and low growth of the nineties aborted this inflection in our unequal model of growth. From the beginning of the eighties until 1995, the high rate of inflation had powerful distributive effect over the real income of first deciles of the household income distribution. Besides this effect, caused by large swings in relative prices, the high volatility in economic growth and stagnation of per capita income was accompanied by a large change in labor market affecting mainly the formal sector, specially the low skilled workers. In the short economic upswing occurred during these fifteen years, as was showed in many studies, the employment fell to most skilled labors. This second effect amplified the observed income concentration. But there was something more: the owner of better protected assets from the ravage of high inflation as the indexed bank accounts, public bonds, and land had huge capital gains. Naturally its consequences on income concentration was undervalued in standard household statistical survey.

This inequity in income could be worst if some new social transference were not in action. By far the most important was the extension of pension funds to the rural labors and the increase of the retirement minimum payment. The effect of this progressive policy enacted in 1988 on poverty alleviation is above any question.

The 1994 Stabilization Plan that followed the large influx of external capital and financial and trade openness brought about different movements. First, the sharp fall in the rate of inflation ended the main previous income distributive factor. The faster drop in food prices than in other prices was in fact a “green anchor” of the Plano Real . This price fall as we will see in more detail, was due to structural and macroeconomic forces and had a positive effect in income distribution through a sharp reduction in absolute poverty. But, the unstable and low growth that succeeded the stabilization plan, exerted a contrary distributive effect. In fact, pressed by external competition and an unfriendly economic policy, the industrial sector suffered a huge employment contraction. This crisis in industrial jobs was accompanied by a stagnation in civil service employment. In less than two years the positive influence of the relative prices on income distribution was more than compensated in S. Paulo by a negative distributive effect caused by a systematic growth in unemployment and low wage occupations. Besides this fact, the high and stable real interest rate gave to the owners of financial assets large capital gains. Naturally this last trend was strongly underreported in statistical household data, but accordingly to National Accounts, the wage share sharply contracted in this decade.

The real question underlying this no job economic model was the extreme volatility in growth rate and stagnation over the last ten years, a straight consequence of a macroeconomics based on volatile capitals, financial deepening and external vulnerability. In Brazil as well in Latin America, the business cycle in the last ten years was governed by net financial external influx.

In countryside two different and opposite forces were in action. Thanks to the huge investments in infra-structure of the late seventies and technical innovations that open new agricultural lands in the last twenty years, the Brazilian agriculture suffered an intense modernization, specially in crops directed to food production for internal markets. This transformation decreased the huge structural heterogeneity that characterized the Brazilian economic development. The main positive consequence of this, were the cheapening of food benefiting the urban and rural wage earners and an increase in rural income in large areas of the country. The other side of this technical change was a vast rural unemployment and productivity polarization in agriculture sector. The rural unemployment was not only the consequence of technical progress but the result of low growth in the real income of urban poor. In this circumstance and without new technologies, credit and good prices, the agricultural land contracted in the last years and thousands of small producers downfallen .

In addition to this new phenomena we have the other specter of our agriculture. Millions of small establishments remained attached to subsistence activities, mainly located in Northeast countryside and small villages forming nowadays, as well as in the past, the hard nucleus of absolute poverty in the country. A nucleus excluded from any positive effect of economic growth and badly dependent on poverty alleviating policies. As we said before the most effective change that reached this very poor group was the inclusion of the old rural labor in a retirement public plan.

The low rate of growth, the stagnation of employment in the formal sector and the expansion of unemployment rate, exerted a negative and countervailing tendency over positive changes in relative prices brought about by the decline in food prices.

More than ever we can not exaggerate the importance of a high and sustained rate of growth for a betterment in income distribution. The unbalanced growth that characterized our years of high industrial growth has changed, the price of food has lowered accordingly, the path for a sustained increase in real minimum wages is open. Unfortunately, there is not enough jobs, the vulnerability to poverty has sharply increased in big cities. The demand for safety nets and income policies that are presently in debate on Brazilian society can not circumvent the necessity of a new model of development based on sustained rate of growth with steady expansion of the real minimum wage.

Given the pattern of consumption of the lower wage earners, intensive in food and mass production industrial products, and the present under utilization of productive capacity, this pattern can be led by household and government consumption and if not interrupted by external shocks and internal contractionary measures, can expand low skill jobs in a rate much more faster than the experience of the nineties. Investments in structures and on public utilities not only in big cities but on small villages can spread a more labor intense and regional balanced growth. Thanks to demographic forces, the rate of dependency in Brazil has declined steadily, lower rate of job creation generates today a higher household per capita income than in the years of high growth of the sixties and seventies.

In order to achieve a sustained rate of growth is necessary smaller dependency on financial capital and more capacity to export, and for this I don’t see any trade-off in the long run with the expansion of internal markets, but this is not our point in this session.
(This material is distributed without profit to those who have expressed a prior interest in receiving this information for research and educational purposes.)