Volume 45 Number 178,
July-September 2014
Is Brazil Becoming Deindustrialized?
Fernando Mattos and Bruno Fevereiro *
Date received: August 7, 2013. Date accepted: January 10, 2014

This study looks at some specific questions involved in the discussion surrounding deindustrialization in Brazil and delves into aspects of the theoretical-historical debate on deindustrialization found in literature on economic development. The first part emphasizes some elements of this debate in specialized literature, discussing theoretical aspects of the Kaldorian formulation and stressing the role of industrial growth and structural changes in the growth of capitalist economies. The second organizes and analyzes some recent performance indicators for the Brazilian industrial sector. Finally, this study offers some brief conclusions.

Keywords: Brazil, industrial sector, deindustrialization, economic development, productivity, commercial balance.

The debate surrounding the relatively reduced share of industrial activities in the GDP (gross domestic product) of developed countries has been referred to as deindustrialization, and has taken on various connotations in recent decades, although the term has not always been negative.

Recently, however, many developing countries (or “middle income countries”) have undergone deindustrialization in various forms, whose causes and effects have invariably differed from those of developed countries, notable for their mature economies, high per capita income and dynamic presence in international trade.

Brazil, in particular, has been the subject of growing and intense discussion recently, especially because various industrial sectors have started to feel the effects of competing and cheaper imports, and, to be perfectly explicit, have begun to lose market share to foreign markets that previously demanded their products. This situation has led to job losses in diverse areas of industrial activities in different regions.

The dialogue on deindustrialization in Brazil is therefore current and compelling, because the Brazilian industrial sector is (still) fairly complex and structured, and plays a key role in labor market dynamics.

This study aims to highlight some aspects of the debate on deindustrialization in Brazil and discuss some points in the theoretical-historical discussion and literature regarding economic development.

The text is therefore divided into two sections. The first describes some elements of the debate found in specialized literature. The second organizes and analyzes some recent indicators of Brazilian industrial sector performance, taking into account both the theoretical discussion and economic indicators. Finally, we provide some brief conclusions.


There are many ways to measure the complex phenomenon of deindustrialization. One is to evaluate it from the perspective of how the manufacturing sector GDP or its share in overall GDP has evolved. Another is to determine the share of high and medium technology manufactures in exports and compare the share of products of a certain country in the set of high technology products exported to the global economy. There are authors, like Rowthorn and Ramaswamy (1999),1 who believe that deindustrialization is simply the reduction of industrial employment relative to overall employment. Tregenna (2011), however, has a more rigorous interpretation, which says that deindustrialization should be defined as a situation in which the weight of industrial employment in overall employment falls, as well as the share of added value of the manufacturing sector in the GDP of the country or region in question. Dasgupta and Singh (2006) believe that analyzing how industrial employment has evolved is important, taking into account both formal and informal activities, especially when interpreting the economic growth of developing countries.

All of these criteria present different methodological aspects worthy of careful analysis to measure the deindustrialization process, because they depend on and derive from countless causes2 and have many implications. As such, it could be concluded, after a bibliographic review of the specialized literature, that taking any of these approaches on their own runs the risk of imprecise or incorrect analysis, which could lead to adopting mistaken industrial policies. All of these indicators are important and should be evaluated together to map out a more accurate diagnosis of the state of the industrial sector. The economic authorities of each country must evaluate these factors together to prevent deindustrialization that would lead to loss of economic growth and increasing external fragility of an economy.

Before succinctly analyzing some of these indicators in Brazil, it is useful to wonder: why is industrial (manufacturing) activity so important for the economic development of a nation? The answer is that, fundamentally, the importance of industry in the productive structure of a country lies in the fact that its activities lead to productivity gains, which are later spread throughout the entire economy, not only of the industrial structure itself, but also in primary sector activities (for example, the entrance of machines and equipment for agricultural and livestock activities and/or extractive activities also leads to greater productivity for both) and in the services sector and commerce.3 Moreover, when the manufacturing sector produces new goods, it creates activities in the tertiary sector, such as the simple commercialization of these goods and the rise of services to support industrial production. In this way, by promoting productivity (producing more per unit of labor or time, or a combination of the two), the manufacturing sector, already dynamic, generates jobs in other areas of the industrial sector and in primary and tertiary areas of the economy.

The increasingly fast rise of income also allows societies to progress in material terms. It is through productivity that the real salaries of employees in the entire economy can grow without generating inflation, promoting greater consumption of goods and services. Social welfare, created by this continuous process of gains due to productivity, will, in general, depend on the capacity of economic activity to maintain or increase these gains, and will also depend on how they are distributed socially. The distribution of economic gains generated by productive activities will be the result of the sociopolitical factors present in each society at a specific moment in time.

Historical experience demonstrates that, precisely due to this social/distributional aspect, the presence of industry is fundamental, because unions generally organize more strongly around industrial activities to ensure that the gains of social productivity generated by economic growth are distributed. Finally, and no less important, we must remember that manufacturing activities, especially gains derived from economies of scale and the increasing productivity that results, are decisive factors to increase the competitiveness of economies abroad and in this way, take over (or generate) markets on the international stage, by increasing national participation in global exports. Nikolas Kaldor synthesized this idea in a theoretical framework, and his ideas would later come to be known as Kaldor's laws, which were initially based on three observations he made in a cross-section analysis of 12 developed countries, comparing the time periods 1952-1954 and 1962-1963, and explaining the varied growth rates found among them.

The first law suggests a positive relationship between the growth rate of the industrial sector and the GDP, which would not be caused by the fact that the industrial product constitutes a large part of the total product, because the results of the tested coefficients and their statistical significance were maintained in a regression between the industrial sector growth rate and initial participation of the non-industrial sector, whose series were not correlated.4 In other words, the greatest GDP growth rates occurred in situations where industry accounted for a greater share of the gross domestic product.

The second law describes the positive relationship between the growth of labor productivity in the economy and the increase in industrial production (Kaldor-Verdoorn law), where an increase in industrial production sets off growth in productivity. The foundation of this law is based on the idea of growing returns of scale in industry, with both a static (like the reduction of unit production costs) and dynamic (through the effects of growth on production in capital accumulation and technology progress) nature.

The third law associates the growth in productivity of the non-industrial sector to the increased rate of industrial sector growth. The increase in industrial production attracts a surplus of labor employed in other economic sectors, reducing unemployment that may have been hidden in other sectors, thereby increasing their productivity.

Moreover, in terms of the external sector, it should be noted that a competitive exporting country enjoys efficient capture of international currency, because remittances in foreign currency from exports to not require counterparts in later expenditures, as is the case for foreign direct investments or loans taken from abroad, which increase the external liabilities of a country and require future remuneration (naturally, also in foreign currency). As such, in terms of external competition, it is also decisive that any country have an industrial sector that is continuously transforming itself with technology updates and incorporating technical advances. At the same time, it is worth mentioning that exports, besides their direct effect in terms of demand on economic activity, also produce an indirect result, which is to allow all other components of demand to grow faster than they would if exports were not rapidly growing and relaxing external restrictions.5

Historically, the economic development of humanity sped up and grew stronger starting with the first, and even more so, with the second, Industrial Revolution. Following the second, gains in productivity and productive activities taken as a whole happened faster, and per capita income differences between countries widened. This was precisely because some were more successful than others in how they incorporated and promoted the benefits of technical advances and increased manufacturing. Starting with the concept that development could be understood as “an inter-related set of long-term processes of structural transformation that accompany growth” (Syrquin, 1988: 205), in which one of these transformations is a change in the sectoral composition of GDP (or jobs), measured by the share of the primary, secondary and tertiary sectors in overall production (or employment), deindustrialization can be understood as the process of structural change that accompanies economic development, and may result from a change in relative prices or from an evolving demand structure.

Considering that technical advances and productivity evolve at different growth rates, the cost of production for one sector relative to the rest constantly changes. As such, the relative price structure also varies. This problem was addressed by a model introduced by Baumol (1967), considering an economy with two sectors, one dynamic and one static. Growth in productivity in the dynamic sector (e.g. industry), allows salary increases without generating pressure on the costs of production and, therefore, price increases. Meanwhile, salary growth in the dynamic sector would generate a spillover effect to salaries in the static (e.g. services) sector, which would pressure costs and, without corresponding growth in productivity, would result in price increases for that sector. There may then be a change in the composition of GDP and employment, while physical production in the two sectors makes the growth rate increase as well.

Moreover, as Pasinetti (1981) argues, income growth, when it accompanies increased productivity, also generates changes to the structure of demand. These shifts determine how the productive structure will evolve.6 According to this author, relative price variations could either slow down or set off changes that would occur anyways with income growth. As such, in the long term, a change in the composition of the productive structure is determined by demand, supported by an extended Engel's law, in which demand for certain products (agriculture and manufacturing) is saturated for the lower income of other products (services). In this way, and from this theoretical perspective, it is possible to understand the deindustrialization that has taken place in so-called developed countries7 as a result of changes in relative prices and, essentially, the structure of demand.

However, in many cases, the manufacturing sector may start to lose share before this stage of capitalist economic maturity has been reached. That is, there may be a reduced relative industrial weight on a per capita level, lower than what developed economies would have when deindustrialization begins. The loss of industry share of GDP has been a recurring phenomenon in so-called developing economies, while in developed economies, per capita income was around 10,000 USD when deindustrialization began, and in Brazil, the reduced share of industry in national income began around the 1980s, when per capita income was close to 5,000 USD. Since then, it has grown very little, in contrast to what has occurred in the US, Germany and the United Kingdom, after the onset of deindustrialization in the 1970s, when per capita income continued to grow. Between 1970 and 2010, per capita GDP in Germany grew by about 87% (equivalent to an average of 1.6% per year in the time period), while in the United Kingdom and the US, income grew by approximately 99% (close to 1.7% per year, on average, in that same forty-year interval).8

What is worrisome in a capitalist economy is precisely the risk that the share of the manufacturing sector will fall before its per capita income is high enough that its economy has reached a certain degree of maturity when deindustrialization begins. This can also be measured by the level of external insertion of an economy in international commerce, as will be shown below.


Before analyzing some data on how the commercial balance has evolved, it is useful to note how the share of industry in GDP and overall employment has evolved, to obtain a similar parameter with respect to evaluations by authors such as Rowthorn and Ramaswamy (1999) and Tregenna (2011), principally.

Figure 1 shows how the share of industry in the Brazilian GDP fell sharply in the mid-1980s, following a significant growth period from 1940 to 1980, the result of industrialization efforts driven by the Brazilian economy in that time period. However, the external debt crisis and hyperinflation faced by the Brazilian economy put an end to these industrialization efforts as the financial capacity of the State was exhausted and the nation faced hyperinflation, as well as the inability to predict prices and the scarcity of long-term credit, which all interrupted investments in industry. The abrupt commercial opening of the 1990s, rather than providing the minimum conditions required by industrial activities to face external competition that arose in the time period, only accelerated the falling share of the sector in GDP. In the years to come, the way in which the economy was handled following the successful monetary transition of the Real Plan, as well as the concept of transformations pushed forward by the stabilization plan,9 ended up accentuating the relative losses of the industrial sector in transforming national income.


Figure 1. Share of Transformation Industry in GDP, at Basic Prices

Source: IBGE (Brazil), prepared by the authors. Methodological note: National Consolidated Account System. Obs.: Concept used for 1947-1989 at cost of factors. For 1990-1994, the National Reference Accounts System 1985. For 1995-2010, the National Reference Accounts System 2000. For 2000-2012, preliminary estimated results based on the Quarterly Reference National Accounts 2000. Obs. Concept used starting in 1990, at basic prices.


In recent years, there has been some recovery, namely between 1999 and 2004, likely linked to the devaluation of the exchange rate that took place in that time period, followed by another period of decline by 2004/2005, but without the intensity of the 1990s. However, it is important to remember that the most recent drop took place off of a far reduced foundation (in terms of the share of industry in GDP), as compared to earlier years. Even so, it is clear that the movements that began in 2004/2005 must be studied, especially from the perspective of the most affected industrial sectors and aspects related to evolving export and import profiles. However, the topic should be evaluated in terms of industrial employment.

Official data10 taken from the Monthly Job Survey (PME) reveals that the total number of people employed in industrial transformation industries has grown since (at least) 2003, except for a brief time in 2009, when the most intense effects of the international crisis resulting from the collapse of the United States credit system were felt. Throughout 2010, the percentage of industry employment grew again, but in 2011, there was a strong slowdown, which continued until the end of the most recent data in 2013. The data in this survey – which has limitations, as it is restricted to the six major metropolitan regions of Brazil and does not consider other metropolitan regions or the vast interior areas – suggests that the current moment is crucial, as it likely represents a point of inflection. In terms of the share of industrial employment in the overall population, there is also clear evidence that it has fallen,11 at least since 2003, a situation which was not necessarily cause for worry prior to 2010, because the percentage of industrial jobs in absolute terms was still rising. Meanwhile, data from 2011 and the fourth quarter of 2012 reveal that these indicators are worthy of more careful, constant and systematic evaluation from now on, because they could also serve as an additional way to evaluate the effects of recent measures implemented by economic authorities, which should continue in the coming years.

As mentioned in the first section, the falling share of industry in general and, specifically, manufacturing, in jobs and the composition of aggregate value are two of the main aspects that constitute so-called deindustrialization. However, this phenomenon in developed countries has no acquire the same bad connotation, and can be understood merely as a natural process that accompanies a country’s economic development. The changing structure of demand that goes along with increasing per capita income, as well as shifting relative prices in favor of the services sector, the result of an industry trend of obtaining the highest growth rates of productivity, explain the idea of deindustrialization as the consequence of an economic growth feedback loop. Official data on how the share of Brazilian industry in jobs and the GDP has evolved reveals that Brazil is undergoing deindustrialization. However, it is of fundamental importance to determine whether it is the result of a feedback loop generated by the country’s development, on a path similar to developed countries decades ago, or if this is temporary, the consequence of factors unlike those seen in developed countries.

The first argument to support the thesis that industrialization has arrived before its time is that it began in a moment (the mid-1980s) when per capita income in the country was much lower than the levels recorded at the beginning of deindustrialization in developed countries. While in developed countries the weight of industry in GDP began to fall when there was a per capita income of around 10,000 USD, in Brazil, it began when per capita income was around 5,000 USD, where it remained for the next decades, unlike what happened in developed countries, where per capita income kept growing. It is therefore difficult to attribute Brazilian deindustrialization to a change in the demand structure as the result of significant and continuous income increases, because Brazil never reached a high-enough per capita income to say that the demand for manufactured goods was saturated.

However, the thesis that deindustrialization resulted due to a change in relative prices following different labor productivity growth rates among sectors must be further studied. As discussed in the first section, based on Baumol (1967), we expect that the weight of the manufacturing industry in GDP fell as a result of productivity growth in this sector as compared to others. Salary increases, derived from greater productivity, would not exert pressure on the prices of goods produced by the manufacturing industry. However, increasing salaries in this sector would generate a spillover effect for the salaries of other more stagnated sectors, like services, which, without a corresponding increase in productivity, would lead to higher costs, and therefore an increase in final prices in the services sector. In this way, relative prices would shift in this sector’s favor.

In fact, Brazil does seem to be undergoing a shift in relative prices in favor of services and other sectors. Even so, it has not been caused by an increase in productivity higher for the industrial sector as compared to others. Manufacturing industry productivity grew in the years following trade liberalization (the 1990s), but in recent years, it has behaved erratically. Between 2000 and 2004, labor productivity in this area grew about 8.83%, while in 2004-2008, the same indicator fell by 8.96%. The services sector saw the inverse trend, with productivity growing 13% between 2004 and 2008, after having fallen 8.89% in the previous period. Agriculture and extractivist industries are the only sectors that have seen continuous (and significant) productivity growth, but their share of the productive structure and GDP are small.12

These results call attention to a contradiction inherent to Kaldor’s laws, which helps explain the deindustrialization of developed countries, where industrial activities reached a high degree of maturity, but does not seem applicable in the case of underdeveloped countries recently hit with deindustrialization, especially Brazil, with its peculiar income concentration and highly informal labor market, as will be shown below. In fact, high growth of aggregate value and employment between 2000 and 2004 in manufacturing industry activities has translated into increased productivity for the sector (not to mention growing performance of scale inherent to these activities and implicit in Kaldor's second law), but that industrial sector growth did not translate into greater productivity for the services sector, contradicting what Kaldor's third law proposes.

Between 2004 and 2008, the employed population (EP) grew significantly in the manufacturing sector, which may help explain the greater productivity experienced in other sectors, in accordance with what the third law proposes. However, productivity fell for the transformation industries in this time period, and there was no growing performance of scale for the industrial sector, as the second law proposes. However, rather than rejecting the validity of Kaldor's relationships, it should be pointed out that there was no increase in productivity so that the transformation industry could help explain the current difficulties found in the Brazilian economy in maintaining high economic growth rates following recovery from the 2008 crisis.13

As seen earlier, the change in relative prices in favor of the services sector, and the consequent increase in its share in GDP in recent years, has not come about because the transformation industry has had greater productivity than the services sector. This change appears, in part, to derive from public policies adopted by the Workers' Party (PT)14 to reduce inequality and furnish real growth for the value of the minimum wage, which favors – of course – people located at the base of the labor market. This lower section of the market essentially consists of many people employed in traditional tertiary sector, low-income activities, as shown clearly in Table 2 data. The weight of service sector activities distributed among the poorest 25% of the labor market is greater than the percentage constituting overall labor market occupation, revealing a concentration in service sector activities in what is known as the base of the labor market. Moreover, service activities, where the poorest 25% of the labor market is situated, have seen falling productivity and, consequently, low yields, for domestic and collective, as well as social and personal, services. The return on these activities is close to the minimum wage, and they therefore benefitted in the most in terms of increasing returns in recent years due to the continuous real valuation policy instituted by the federal government in 2003.

It is therefore worthwhile to note how peculiar the sectoral makeup of the lowest yield strata of the distributional pyramid is, and how this is precisely due to the fact that the labor market is significantly unstructured and highly informal. These lower-income strata, in reality, aim to survive by selling low value added goods and services (also made up of people who work in domestic employment, whose incomes are among the lowest in Brazil). Manufacturing sector employees are also having trouble in negotiating real salary increases, and, when they do, the increases are generally lower than they are for the tertiary sector, principally because these workers are less subject to external competition and because they are favored by real minimum wage salary adjustment policies that were implemented following Lula's election.15

In Brazil, it is exactly because income is so strongly concentrated that any small improvement in the distributional profile (as has happened in recent years) leads to massive changes in the demand structure, which is expressed in higher sales of industrialized goods and low value added service sales, creating a feedback loop for income increases among members of the tertiary economic sector. However, in terms of demand for industrialized products, driven by increasing economic revenue, manufacturing production stands out in other countries, where there is a lack of structural governmental policies to support greater competition among manufacturing activities. This situation is starting to show signs of deindustrialization in Brazil, and will be analyzed further below, in terms of how Brazil's foreign trade profile has evolved in recent years. Table 3 displays how the Brazilian export profile has evolved by aggregate value, separated into three categories: basic, semi-finished and manufactured exports. Looking at a relatively long time period beginning in the 1970s, we can obtain a better overview of the process. According to this data, the profile evolved significantly in the 1970s and even the 1980s (despite the crisis plaguing the economy in those years) in favor of manufactured products. We can also note that in the 1990s, this share was relevant to exports, although it grew little, but only accounted for an extremely low share of industry in Brazilian GDP. However, starting in 2000, the path reverses, and the relative weight of manufactured and semi-finished products began to fall (strictly speaking, they began to lose share starting in 1995), benefitting the relative growth of basic products.16 This shift was especially accentuated starting in the mid-2000s due to the effect of Chinese demand for commodities, which increased prices. Meanwhile, Chinese insertion in manufactured goods produced an inverse effect on industrial prices.

Table 4 helps us evaluate how the Brazilian export profile has evolved recently in greater detail, as it shows data in absolute values (as well as the relative share) of more disaggregated aspects of overall exports. The data reveals an unequivocal increase in the share of primary products (non-industrial) between 1996 and 2010, and, among industrial products, reduced relative weight of higher technology items.

The same source of data was used in Table 5 for imports, also indicating deterioration in the Brazilian foreign commerce profile, in this case manifest in the growing relative weight and absolute values (USD) of high and medium technology goods, in particular.

As a result of the changing foreign commerce profile described above, the penetration of imports in industry17 has grown in all sectors, from labor-intensive activities (such as textiles) to the production of high added value goods (automobiles) to capital goods production. Here, what is important are imports, with respect to internal demand, which grew from 22.2% in the last quarter of 2004 to 42.9% in the last quarter of 2012.18

The data in Table 6 not only shows how the share of imports in consumption of various segments of Brazilian industrial activity has grown (categorized by degree of technology in production), but also how China has contributed to this phenomenon. According to this data, the sectors that have been most affected are highly labor and technology-intensive (evidently including capital goods), therefore corroborating other evidence mentioned in this article. For labor-intensive activities, Chinese imports doubled between 2005 and 2010, going from 5.4 to 10.8%. For technology-intensive sectors, the share of imports, already high in 2005 (29.0%), jumped to 37.5% in 2010, and two-thirds of this 8.5% was of Chinese origin.

Finally, the data in Figure 2, prepared using export profile information by technology level as defined by the UNCTAD, also suggests deterioration in manufactured goods exports, evaluated by category in overall global exports. This data reveals that Brazilian industrial exports have fallen over the past 15 years. Specifically, there was a big drop between 2008 and 2010 in the share of low technology intensity product exports as compared to overall global exports. The same is true of the other extreme of the profile, that is, high technology intensity products used in production.

The data in this section demonstrates that the Brazilian economy is becoming less dynamic on the global stage, which puts its economic development at risk, as Dasgupta and Singh (2011) and Palma (2005) have warned. Also, given the variety of evidence in this study, the falling dynamics of the external sector suggests that Brazilian deindustrialization has arrived early.


Figure 2. Manufacturing Exports by Degree of Technology
(Share of Brazilian Exports in Total Global Exports)

Source: UNCTAD, prepared by the authors.



Deindustrialization in developed countries was the result of unequal technical progress (and productivity) among sectors, which shifted prices in favor of the services sector. Moreover, the way in which per capita income evolved promoted a change in the demand structure in those countries, as demonstrated by various historical examples recorded in international literature on the topic (some of which were mentioned in the first section of this article). This is mainly due to the decreasing share of the manufacturing sector in GDP and the makeup of employment starting in the 1970s in those countries. On the other hand, in Brazil, the decreasing share of industry in employment and the GDP cannot be explained by the same causes seen in developed countries.

In Brazil, deindustrialization is taking place early. It began at a time when per capita income was much lower than it was in developed countries. In addition, the recent decreasing share of industry in GDP can be explained by a shift in relative prices in favor of the services sector, because this is not the result of increasing lower productivity in sectors such as the manufacturing sector, which, analogous to Baumol's (1967) description in the first section, could explain the falling share of non-industrial sector employment and, through a change in relative prices, the non-existent share in aggregate value.

However, in Brazil, the increasing share of the services sector in the GDP in recent years has taken place for a different reason than what has occurred in developed countries. This peculiarity can be explained, in part, by the nature of the Brazilian labor market, whose foundation (i.e. lowest income occupations) rests, in large measure, on service sector jobs, such as domestic employment, personal services, traveling salespeople, etc. Their income rose in the 1990s due to a change in relative prices driven by currency appreciation. In recent years, in addition to the new cycle of currency appreciation, the minimum wage has grown, there has been strong credit growth and laborers have had greater negotiating power as a result of reduced unemployment coming from greater growth between 2004 and 2010. In an environment of currency appreciation, the services sector, which produces non-commercial goods, does not compete with imported products and can therefore pass on the pressures on costs derived from salary increases to the final price, a situation not true of the transformation industry, which produces commercial goods, and therefore suffers from competition against imported products.

In addition, industrial competitiveness has been affected by the lack of policy to promote more active and sovereign participation in a new environment of international inter-capitalist competition, as well as taking a role in international commerce. The way in which the share of the transformation industry in Brazilian GDP has evolved reveals a resounding decrease in the 1990s. More recently, however, between 2004 and 2010, industry has been the object of intense debate and analysis in the Brazilian press and academic world. The current issue is that there is relatively low share of industries of transformation in the GDP. Moreover, current Brazilian GDP per capita, as well as the way in which the foreign trade profile has evolved recently,19 indicates early deindustrialization. This situation is even more worrisome if we take into account recent changes in the international division of labor, which have happened even faster than in the 1990s,20 which, naturally, generates additional challenges and risks to reduce the share of manufacturing activities, regardless of the analytical criteria applied.

The Brazilian commercial balance is increasingly dependent on the price and volume of commodities, a situation that – as Brazilian and Latin American history have shown – could be quickly reversed, although this does not seem to appear on the immediate horizon. We must also remember that as internal revenue increases, we might imagine that the demand for industrialized products grows faster than average income, precisely because income elasticity of demand for manufactured products is greater than for primary. In this way, given that the share of imports for consumption has grown in various industrial sectors – and it seems to be true, as was demonstrated in the study – there is a clear trend towards the deterioration of the manufactured products commercial balance.


Visiting researcher at the IPEA, where most of the research for this study was conducted.


Baumol W. (1967), “Macroeconomics of unbalanced growth: The anatomy of an urban crisis”, American Economic Review, June, available at: http:// piketty.pse.ens.fr/files/Baumol1967.pdf.

Dasgupta, S., and A. Singh (2006), “Manufacturing, services and premature deindustrialization in developing countries - a Kaldorian analysis”, working paper no. 2006:49. United Nations University (UNU–WIDER World Institute for Development Economics Research), May.

Kaldor, N. (1966), “Causes of the slow rate of economic growth of the United Kingdom”, Cambridge: Cambridge University Press.

______ (1978), “Further essays on economic theory”, London, Duckworth.

Kollmeyer, C. (2009), “Explaining deindustrialization: how affluence, productivity growth, and globalization diminish manufacturing employment”, American Journal of Sociology, vol. 114, no. 6, May.

Mattos, F. A. M., and M. D. Carcanholo (2012), Amenazas y oportunidades del comercio brasileño con China: lecciones para Brasil, Problemas del Desarrollo, vol. 43 no.168.

Nickell, S.; S. Redding, and J. Swaffield (2008), “The uneven pace of deindustrialization in the OECD”, London School (ESRC), May.

Oreiro, J. L., and C. Feijó (2010), Desindustrialização: conceituação, causas, efeitos e o caso brasileiro, Revista de Economia Política, vol. 30, no. 2 (118), April-June.

Palma, G. (2005), “Quatro fontes de desindustrialização e um novo conceito de doença holandesa”, Conferência de Industrialização, Desindustrialização e Desenvolvimento, FIESP, August.

Pasinetti, L. L. (1981), “Structural change and economic growth: A theoretical essay of the dynamics of the wealth of nations”, Cambridge University Press.

Puga, F., and M. O. Nascimento (2010), “Efeito China sobre as importações brasileiras”, BNDES, Visão do Desenvolvimento, no. 89.

Rowthorn, R., and R. Ramaswamy, (1997), “Deindustrialization – its causes and implications”, Economic Issues (10), FMI, Washington.

______ (1999), “Growth, trade and desindustrialization”, FMI Staff Papers, vol. 46, no. 1, FMI, March.

Syrquin, M. (1988), “Patterns of structural change”, in H. Chenery, and T. Srinivasan, Handbook of Development Economics, Elsevier.

Tregenna, F. (2011), “Manufacturing productivity, deindustrialization and reindustrialization”, working paper, no. 57, United Nations University (UNU–WIDER World Institute for Development Economics Research), September.

Thirlwall, A.P. (1983), “A plain man’s guide to Kaldor´s laws”, Journal of Post Keynesian Economics, vol. 5, no. 3.

______ (2000), “The nature of economic growth: an alternative framework for understanding the performance of nations”, Edward Elgar, Cheltenham.

UNCTAD-ONU (2003), “La acumulación de capital, el desarrollo económico y el cambio estructural”, New York.

UNIDO-Industrial Development Report (2009), Breaking in and moving up: New industrial challenges for the bottom billion and the middle-income countries, United Nations, Geneva.

* Faculty of Economics at the Universidad Federal Fluminense (UFF), fermatt@uol.com.br and Universidad Federal de Río de Janeiro, Brazil, jbfevereiro@gmail.com, respectively.

1 Also see Oreiro and Feijó (2010). Rowthorn and Ramaswamy (1999) define deindustrialization as the persistent decrease in the share of industrial employment in total country employment. In a previous work, Rowthorn and Ramaswamy (1997) postulate the deindustrialization is not necessarily damaging to an economy, and ascertain that the main reason why a capitalist economy goes through deindustrialization – which the authors see as a “natural consequence” of the growth of advanced economies – is that earnings on productivity in industrial activities supplant service sector earnings.

2 Rowthorn and Ramaswamy (1997) sought to systematize the causes and implications of deindustrialization, and claim that interpreting this phenomenon reveals how to interpret the economic development of developed countries or nations going through important structural changes, as has happened recently, for example, with Asian countries, rather than revealing a problematic phenomenon.

3 This situation is known in Kaldorian literature as Kaldor's third law, which proposes a strong positive causal correlation between the speed of manufacturing sector growth and the productivity earnings of the other sectors. See Thirlwall (2002).

4 For a more detailed discussion on possible issues related to endogeneity for Kaldor’s first law, see Thirlwall (1983).

5 Imports made with currency brought by exports can, frequently, bring supply components that are much more important for the continuity of economic growth (for example, capital goods, or intermediate goods whose internal market production is insufficient), promoting new ways to generate productivity earnings for all economic activities.

6 This argument is based on a stylized idea proven by empirical works going back to Ernst Engel, in the nineteenth century, who, in a generally accepted statement, stipulates that "when income increases […] per capita demand for each sector does not expand proportionally” (Pasinetti, 1981: 69).

7 See Kollmeyer (2009), Nickell, Redding and Swaffield (2008), Rowthorn and Ramaswamy (1997 and 1999), Tregenna (2011) and Dasgupta and Singh (2006), among other authors that, using broad databases, have empirically and theoretically analyzed the reduced participation of industry, both in employment and in income, in a variety of developed countries.

8 The data detailed and the methodology used are available to readers by writing to the authors by e-mail.

9 The premises of the Real Plan strategy, based on a liberal analysis and concept, can be defined as follows: 1) price stability improves the future of capitalist calculations and, therefore, stimulates productive investment; 2) commercial opening and an appreciated exchange rate oblige companies to obtain earnings from productivity, therefore becoming more competitive internationally; 3) privatization and FDI remove bottlenecks in the industrial supply and improve infrastructure, reducing costs for all stakeholders; 4) liberalizing the exchange rate will attract foreign financial capital, helping to finance deficits and also stimulating direct investment; 5) stabilization will promote a better distributional profile in Brazil.

10 Due to space limits in this article, the data in tables and figures provided in an earlier version of this study will be available to readers upon e-mail request: jbfevereiro@hotmail.com and fermatt@uol.com.br.

11 This aspect should also be evaluated further and is worthy of detailed research in the future. The increase in employment in tertiary sector activities could, in large measure, be linked to an intense process of change in relative prices. This hypothesis has also been widely discussed in literature on deindustrialization, as we saw in the first part of this article. This would bring about growing demand for services and merchandise sold by people employed in tertiary sector activities.

12 Unfortunately, the national consolidated accounts system, reference 2000, is restricted for 2009. The IBGE is working on a new series starting in 2010, based on an updated classification of sectors, but it will not be available until early 2015. The year 2009 was eliminated following analysis due to the cyclical effects of the crisis, which could make understanding the structural component difficult.

13 After growing 7.5% in 2010, the country had average growth of 2.06% per year between 2011 and 2013, IBGE, National Quarterly Accounts.

14 Starting in 2003.

15 Between 2002 and 2008, there was real growth of 9% for service sector salaries, and the low-income activities that most benefitted were domestic services, with real growth of 16%, public administration salaries, which grew 11%, together with social security spending, where the lowest pensions were corrected with the minimum wage, and housing and food saw a real increase of 12% between 2002 and 2008. Microdata from the PNAD (IBGE), prepared by the authors.

16 These statistics are measured in value, so they do not reflect the change in relative prices in favor of agricultural and mineral commodities starting in the mid-2000s.

17 As Puga and Nascimento (2010) wrote, "the import coefficient measures the participation of imports in apparent domestic consumption of each sector[…]” Apparent consumption, also defined as internal supply, is measured by taking total domestic production, subtracting foreign sales (exports) and imports. It could also be said that the import penetration coefficient refers to the amount of internal supply met by imports.

18 Detailed data available to readers upon request to the authors.

19 Figure 2 data and arguments by Mattos and Carcanholo (2012), in a recent edition of this academic journal, reveal that the insertion profile of Brazilian products in international trade has continued to deteriorate, indicating a loss of share for manufactures, and, in general, higher added value products, especially among global exports.

20 A recent report by the United Nations (UNIDO, 2009) studying global industrial development revealed, for example, that in the 2000s, the participation of Asian countries in overall global industrial employment has grown significantly. Similarly, this document ascertains, among other data, that the share of high technology exports in global manufacturing exports grew from 19.6% in 1990 to 22.6% in 2005. Likewise, data from the United Nations study (2009) demonstrated that the share of developing countries in total manufacturing exports in global commerce is increasing. This result is not merely due to China or to growing oil-derived exports. The study also reveals that Latin America's industrial performance has, for various reasons, been relegated to the performance of Asian countries since the 1970s (a situation further aggravated by differences in recent years). This reality, described in the UNIDO study, when compared with relevant outcomes in this article shows that Brazil's position has indeed deteriorated in recent years, falling to average performance as compared to other developing countries.

Licencia de Creative Commons  Problemas del Desarrollo. Revista Latinoamericana de Economía by Instituto de Investigaciones Económicas, Universidad Nacional Autónoma de México
is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

Published in Mexico, 2012-2018 © D.R. Universidad Nacional Autónoma de México (UNAM).
PROBLEMAS DEL DESARROLLO. REVISTA LATINOAMERICANA DE ECONOMÍA, Volume 49, Number 194 July-September 2018 is a quarterly publication by the Universidad Nacional Autónoma de México, Ciudad Universitaria, Coyoacán, CP 04510, México, D.F. by Instituto de Investigaciones Económicas, Circuito Mario de la Cueva, Ciudad Universitaria, Coyoacán,
CP 04510, México, D.F. Tel (52 55) 56 23 01 05 and (52 55) 56 24 23 39, fax (52 55) 56 23 00 97, www.probdes.iiec.unam.mx, revprode@unam.mx. Journal Editor: Moritz Cruz. Reservation of rights to exclusive use of the title: 04-2012-070613560300-203, ISSN: pending. Person responsible for the latest update of this issue: Minerva García, Circuito Maestro Mario de la Cueva s/n, Ciudad Universitaria, Coyoacán, CP 04510, México D.F., latest update: August 29th, 2018.
The opinions expressed by authors do not necessarily reflect those of the editor of the publication.
Permission to reproduce all or part of the published texts is granted provided the source is cited in full including the web address.
Credits | Contact

The online journal Problemas del Desarrollo. Revista Latinoamericana de Economía corresponds to the printed edition of the same title with ISSN 0301-7036