, , ,

José Luis Oreiro*

         Among heterodox economists, mainly these one with a Kaldorian background, there is no doubt that manufacturing industry is the engine of economic development, since this sector is the source of increasing returns of scale (both internal and external to the firms), the sector with higher backward and forward linkages in the productive structure, the sector that produces tradeable goods that had a higher income elasticity of demand and hence softens the balance of payments constraint to growth and produce and spread new technologies through other sectors by means of new capital goods (Thirwall, 2002). The special role of manufacturing sector in economic development was one of the central theses of both classical development theory and Latin-American Structuralism (Ros, 2013).

         Nowadays, the central role of manufacturing industries to economic growth and the technological catch-up process are highlighted in Szirmai (2012), Thirlwall (2002), and Tregenna (2009), among others, through stylized facts and empirical analysis. They show that economic growth depends on the composition of productive structure and, especially for developing economies, industry.

         According to Rodrik (2016), manufacturing tends to experience relatively stronger productivity growth and technological progress over the medium to longer term. Therefore, premature deindustrialization closes off the main way to achieve fast economic convergence in low- and middle-income countries. It was the industrialization process that permitted catch up and convergence with the West by non-Western nations, such as Japan in the late 19th century, and South Korea, Taiwan, and China, among other countries, in the 20th century.

         Rodrik (2009, 2016) highlights that the rapid economic growth of developing economies since the 1960s is associated with the largest transfer of productive resources (labor and capital) to the most modern industries. The structural shift toward industrial activities drives economic growth.

         Szirmai (2012) presents a series of empirical and theoretical arguments about the role of industry as the “engine of growth” in developing economies. Basically, productivity in manufacturing is higher than in agriculture because the transfer of resources from this sector to the industry provides a “structural change bonus.” This “bonus” comes because of the transfer of labor from economic activities with low productivity to high productivity activities (Lewis, 1954). This automatically raises the overall productivity of the economy.

            For most mainstream economists, however, there is nothing special in the manufacturing sector. For then, economic growth is the result of capital (both physical and human) accumulation and technological progress. The composition of output is seen irrelevant for long-term growth or the result of factors endowments and the quality of institutions.  In this second interpretation the existence of a positive correlation between periods of growth acceleration and increasing manufacturing share in GDP – the so-called industrialization process – is just the result of a high pace of capital accumulation allowed by the combination of high saving rates and “good institutions”, which are generally defined as the institutions that induce growth, being so a non-falsified concept in the sense of Popper (1972). In their framework, that there is no need to adopt policies to develop the manufacturing industry, but only to improve the quality of institutions: good institutions will allow growth accelerations which are, in general, associated with industrialization, at least in the first stages of economic development.

            One of the basic propositions of the Classical Development Theory is that the combination of unlimited supply of labor due to the existence of a traditional or non-capitalist sector in developing economies with internal and/or external economies of scale can result in a poverty trap due to the paradox of underdevelopment: In developing economies a lower stock of capital per-worker is associated with low returns for capital accumulation, not with high profit rates as in the traditional neoclassical growth models (Ros, 2013, pp.154-159). This means that the main obstacle for growth take-off is not the shortage of savings, but the lack of incentives for capital accumulation by the private sector. Underdeveloped economies are poor because they had low levels of capital per-worker and, at the same time, the profit rate and the incentive to capital accumulation is low because of the low-level of the capital per-worker. If investment can be accelerated by some non-market mechanism than it will produce the savings required for the development process to continue since labor will be transferred from the low productive subsistence sector for the high productivity modern sector; but real wages will stay at a constant level (the subsistence wage plus some constant wage premium). This means that the share of profits in income will rise as the investment is increasing, and since the propensity to save out of profits is higher than propensity to save out of wages; than the saving rate will increase because of the increasing in the investment rate. Economic development always and elsewhere produced the saving rate required for its long-term sustainability.

            For classical development theory, a self-sustained process of economic development requires the achievement of a critical mass for capital per-worker, which can only be done by the State, so industrialization, at least in its first stages, had to be necessarily State-Led. Moreover, as claimed by Prebisch (1950) and the economists of ECLAC, due to low-income elasticity of demand for exports of underdeveloped countries, whose exports are mainly composed by primary products, then in the early stages of industrialization there is no option rather than impose some controls over imports, trade tariffs and exchange rate controls to soften the external constraint. These policies will induce a process of import substitution to reduce the income elasticity of imports and “save” the external currency required for import capital and intermediate goods required to industrialization. So, development process will be an Import-Substitution-State-Led Industrialization (ISSI, hereafter).  

            Latin-American Structuralism had not given to the real exchange rate any important role in the process of economic development. Because of the so-called “elasticity pessimism” – that is due to the fact the composition of exports and imports of underdeveloped economies do not allow the fulfilment of the Marshal-Lerner condition- an exchange rate devaluation will not either increase the trade surplus or induce the substitution of imports for domestic production, since it will produce an increase in the domestic prices of imported capital goods. A system of multiple exchange rates, where a more appreciated exchange rate is defined for capital goods imports and a more depreciated exchange rate for final goods imports combined with high import tariffs will both relief the external constraints and provide the incentives for import substitution. Capital accumulation led by state-owned enterprises will also be necessary for the “big push” required for the economy to scape the poverty trap.

            The first Latin-American economist to see a role for the exchange rate in the process of economic development was Diamand (1972). According to him, for countries specialized in exports of primary products, there is a problem of an unbalanced productive structure. In his words:

The essential feature of the new economic reality of the Primary exporting countries in the process of industrialization is the which we have dubbed an unbalanced productive structure. It is a productive structure. Composed of two sectors of different price levels:  the primary-agricultural sector in our case -, which works at international prices, and the industrial sector, which works at a level of costs and prices considerably higher than the international. This peculiar configuration, not even imagined by the generations dedicated to the elaboration of the economic theory that today is taught in universities, gives rise to a new economic model, characterized by the chronic limitation that the external sector exerts on economic growth. Indeed, while the growth of the economy – particularly industrial growth – always requires increasing amounts of foreign exchange, the high level of industrial prices that characterizes the unbalanced productive structure prevents industry from exporting. Thus, unlike in industrial countries, in which the industry self-finances the foreign exchange needs posed by its development, the Argentine industrial sector does not contribute to obtaining the foreign exchange it needs for its growth. Its supply is always in charge of the agricultural sector, limited either by lack of a greater production, or by problems of world demand or by both.

In the initial stage of this type of development a rapid replacement of imports makes the industry contribute to keeping the balance of payments balanced by saving foreign exchange. Subsequently, the replacement process becomes increasingly slow. Finally, it is reached that substitution at most can neutralize the increase in imports brought by technological progress by the incorporation of new products (cars, television, yarnssynthetics, etc.). From this moment on, a process of divergences begin between the growth of the industrial sector consuming foreign exchange, which it does not contribute to the production of them, and the provision of these currencies by the much slower growing agricultural sector. This diversity is responsible for the balance of payments crisis in Argentina and is the main growth limiter in the country. The expansion of domestic production, each time it occurs, increases imports. Once the reserves are exhausted, the country is forced into a devaluation. This occurs even without a prior increase in costs, which requires the restoration of parity. This is a devaluation of another kind, which is imposed by the imbalance that arises in the productive structure itself, as a result of the already marked divergence between the consumption and the supply of foreign exchange. (1972, p.2)

            The problem of the unbalanced productive structure emerges from setting the nominal exchange rate at a level compatible with the productivity of the primary sector, that is, in a level that makes the exports of such goods competitive in international markets, but that is too much appreciated for the manufacturing firms to be also competitive in such markets (1972, p.18). As a result of the overvalued exchange rate for manufacturing goods, the domestic manufacturing industries of these countries are uncapable to conquer international markets and thus increasing their share of manufacturing exports. Therefore, domestic manufacturing firms are confined to domestic markets, where the scale of production is not enough to provide then even the productivity to compete with imported goods unless high import tariffs are set to isolate the domestic market from external competition. Thus, tariff protection that should be only used in the phase of infant industry to promote the initial stages of industrialization through import substitution, became permanent due to the political pression of domestic industrial entrepreneurs. A permanent protection destroys the incentives for the domestic firms to increase its productivity over time by introducing labour-saving technologies and hence will increase the technological gap of domestic firms relative to foreign firms, reducing also their non-price competitiveness.

            Diamand´s arguments are very similar to the role of exchange rate in the process of economic development according to the Brazilian New-Developmentalist School. As I have already argued (New Developmentalism and Balance of Payments Constrained Growth Models: convergences and divergences | José Luis Oreiro (wordpress.com),), a competitive level for real exchange rate is required to compensate the technological backwardness of domestic firms in developing economies. Overvaluation of real exchange rate is a very important cause of premature deindustrialization of these economies, as it is showed by Oreiro et al (2020) for the Brazilian economy. Moreover, Gabriel et al (2020) had shown with a panel data econometric model for a broad sample of 84 countries for the period of 1990 to 2011 that the effect of real undervalued Real Exchange Rate is positive and statistically significant with a lag for all the technological gap levels considered, increasing their effect on the per capita income growth rate when the technological gap measure is higher (for each group of countries). The effect of undervalued RER on the per capita growth rate is conditional on the technological gap level considered: the greater the gap of the sample of countries in relation to the technological frontier, the greater the effect of the undervalued RER on per capita income growth rate.

            These findings shows that the common criticism made against the role of real exchange rate in economic development according to which manufacturing growth and per-capita income growth depends only on non-price competitiveness of the manufacturing sector is wrong. The higher is the technological gap the firms of a country face, more important is for then to have an undervalued exchange rate to compensate their technological backwardness relative of the firms of developed countries.

            This does not mean, for sure, that the only thing that can be done to foster economic development is to set the exchange rate at the “right” level, which is for new-developmentalist school the so-called industrial equilibrium level, recently redefined by Oreiro (2020) and calculated by a new methodology developed by Oreiro et al (2020). Science and Technology policies, as well as industrial policies, are required to reduce the technological backwardness of domestic firms relative of the firms of developed countries; but it is impossible for these policies to give any relevant result if real exchange rate is overvalued.

Brazil, the major economy of Latin America, had experienced a very long period of overvalued exchange rate since 1994, which is briefly reverted by a balance of payments crisis due to high current account deficit (1999), a sudden stop of capital flows (2002 and 2008) and the end of commodity boom (2015). During the governments of President Luis Inácio Lula da Silva and President Dilma Rouseff a lot of industrial policies are adopted, but all failed to prevent the second wage of premature deindustrialization since 2005. Manufacturing share in real GDP felt from more or less 17 % in 2005 to 11% in 2019. According to the estimates made by Oreiro et all (2020) more or less of 40% of the reduction of the manufacturing share can be directly attributed to exchange rate overvaluation. This means that to restart the industrialization of the Brazilian economy is necessary to keep the real exchange rate at a competitive and stable level in the medium to the long-term. This demands the adoption of a new macroeconomic policy regime in which capital controls will certainly have an important role to manage the exchange rate without jeopardizing the autonomy of monetary policy.

To sum up: having a good macroeconomic policy regime is not the only thing required to restart Brazilian economic development but is a crucial beginning.


Diamand, M. (1972). “La Estructura Productiva Desequilibrada Argentina y el Tipo de Cambio”. Desarrollo Económico Vol. 12 N° 45.

Gabriel L.F., Ribeiro L.C.S., Jayme Jr F.G. and Oreiro J.L. (2020), “Manufacturing, economic growth, and real exchange
rate: empirical evidence in panel date and input-output multipliers”, PSL Quarterly Review, 73 (292), pp. 51-75]

LEWIS, A. (1954). “Economic Development with unlimited supply of labor”. The Manchester Scholl 22: 139-91.

Oreiro, J.L. (2020). “New Developmentalism: beyond competitive exchange rate”. Brazilian Journal of Political Economy, vol. 40, nº 2, pp. 238-242.

Oreiro J.L., Manarin D’Agostini L.L., Gala P. (2020). ” Deindustrialization, economic complexity and exchange rate overvaluation: the case of Brazil (1998-2017)”, PSL Quarterly Review, 73 (295):313- 341.

POPPER, K. R. (1972) Conjecturas e refutações. Brasília: UNB.

PREBISCH, R. (1950). The Economic Development of Latin America and Its Principal Problems. New York: United Nations.

RODRIK, D. (2009). Growth after the crisis. Cambridge, MA.: Harvard Kennedy School.

RODRIK, D. (2016). Premature deindustrialization. Journal of Economic Growth. 21(1): 1–33.

Ros, J. (2013). Rethinking Economic Development, Growth and Institutions. Oxford Economic Press: Oxford.

SZIRMAI, A. 2012. Industrialization as an engine of growth in developing countries, 1950-2005. Structural Change and Economic Dynamics, 23: 406–420.

Thirlwall, A. (2002). The Nature of Economic Growth, Aldershot: Edward Elgar.

TREGENNA, F. (2009). “Characterizing deindustrialization: An analysis of changes in manufacturing employment and output internationally”. Cambridge Journal of Economics, 33: 433–466.


* Associate Professor at Economics Department of University of Brasília (UnB) and Level IB Researcher at the National Council for Scientific and technological Development (CNPq). E-mail: joreiro@unb.br.