Is Manufacturing Still the Main Engine of Growth in Developing Countries?

Adam Szirmai

Since the late 18th century, the manufacturing sector has been the main engine of growth and catch up. Presently, however, service sector value added accounts for over 70 per cent of GDP in advanced economies. In addition, ICT services have become important sources of growth in a great many developing countries, in particular India. This raises the question whether the manufacturing sector will continue to be the most important engine of growth, development and catch up for developing countries in the 21st century.

Historical Patterns of Manufacturing Development

By and large the global diffusion of manufacturing had largely bypassed developing countries until around 1950. Japan was the only non-Western economy that realised large scale industrialisation prior to 1950. In other developing countries, such as Brazil, Mexico, India and China, there were early beginnings, but for a variety of reasons, industrialisation did not really take off. After 1950 this changed. Most developing countries experienced rapid structural change with large declines in the share of agriculture, and increases in both manufacturing and services. The pattern of structural change differed from that in the past. The service sector was already large in 1950 and its share increased in parallel with that of manufacturing.

In 1950, the share of manufacturing in a sample of 29 of the largest developing countries was 11 per cent, compared to 31 per cent in 16 advanced economies. The average share of manufacturing in these developing countries increased till 1980, when it reached 20 per cent. It then declined to 18 per cent in 2005. The Asian economies bucked the trend and continued to industrialise after 1980, reaching an average manufacturing share of 24 per cent in 2005. The highest shares for 2005 were found in China (34%) and Thailand (30%). In the same period, the share of manufacturing in the 16 advanced economies declined from 24 in 1980 to 17 per cent in 2005.

Some interesting observations emerge from the empirical record. First, even at the peak of industrialisation, manufacturing never accounted for much more than 30 per cent in the advanced economies. Next, the share of manufacturing in developing countries in the fifties was higher than one would expect on the basis of the literature. This is probably due to shortcomings in the early national accounts which focused disproportionately on the formal sector, thus exaggerating the share of manufacturing. Finally, in the period 1950-2005, a limited number of developing countries succeeded in become major industrial exporters to Europe, Japan and the USA. They became the workshops of the global economy. When one examines the successful cases of catch up, they were invariably countries which were also successful in industrialisation.

Manufacturing and Economic Growth

In the following, we briefly examine nine arguments for the importance of manufacturing for growth.

First, there is an empirical correlation between the degree of industrialisation and per capita income in developing countries. If one plots the share of manufacturing in commodity production against per capita incomes, there is a clear relationship between the two. The correlations are not perfect, but the poorest countries are invariably the least industrialised ones and the more successful developing countries are at the upper end of the scale.

Second, productivity is higher in the manufacturing sector than in the agricultural sector. The transfer of resources from agriculture to manufacturing provides a structural change bonus. We have examined sectoral productivity levels in 19 Latin American and Asian economies and found that between 1950 and 2005, value added in manufacturing was consistently much higher than in agriculture (Szirmai, 2008, table 5). A puzzling finding was that in post-war Latin America, value added per worker in services was higher than in manufacturing. This suggests that the structural change bonus for services might have been even higher than that for manufacturing. This effect disappeared after 1980, when productivity levels in manufacturing exceeded those in services.

Third, a further aspect of the structural change bonus argument focuses on the dynamics of sectors. Manufacturing is assumed to be more dynamic than other sectors. A transfer of productive resources to more dynamic sectors contributes to growth. Here the evidence turned out to be somewhat mixed (Szirmai, 2008, table 6). Between 1950 and 1973, productivity growth in manufacturing was indeed much higher than in agriculture. But after 1973, this was reversed. As in the advanced economies, productivity growth in agriculture in developing countries tends to be higher than in manufacturing. In terms of output growth, manufacturing continues to outperform agriculture in both advanced and developing economies, because the share of manufacturing in the total economy is shrinking everywhere.

Fourth, the transfer of resources from manufacturing to services provides a structural change burden in the form of Baumol’s disease. As the share of the service sector increases, aggregate per capita growth will tend to slow down. Developing countries with higher shares of manufacturing and lower shares of services will show faster growth than the advanced service economies. In recent years, this general argument has been qualified. Baumol’s law may still obtain in sectors like education, health care or personal services, but other services sectors such as financial services, computer and software services, transport, and distribution have become very dynamic.

Fifth, compared to agriculture, the argument runs that the manufacturing sector offers special opportunities for capital accumulation. Capital accumulation can be more easily realised in spatially concentrated manufacturing than in spatially dispersed agriculture. This is one of the reasons why the emergence of manufacturing has been so important in growth and development. Sectoral capital stock estimates for developing countries are still scarce, but what data there are indicate that after 1950 manufacturing is indeed far more capital intensive than other sectors (Szirmai, 2008, table 8). This confirms the importance of manufacturing for accumulation. However, the relative capital intensity of manufacturing does decline over time. By 1990, capital intensity in manufacturing is about the same as that in the total economy. In the advanced economies, manufacturing has lost its role as locus of capital accumulation. Since 1970, capital intensity in manufacturing is much lower than that in the economy as a whole.  It is even lower than in agriculture.

The secondary literature provides four further arguments for the potential importance of manufacturing: economies of scale, technological progress, linkage effects and Engel’s law.

Six, the manufacturing sector offers special opportunities foreconomies of scale, which are less available in agriculture or services. Scale effects in manufacturing have undoubtedly been important, but due to the rise of new ICT technologies, they are no longer limited to manufacturing. In certain ICT based service sectors, scale economies have become very important as the marginal costs of additional units of service approach zero.

Seven, the manufacturing sector offers special opportunities for both embodied and disembodied technological progress(Cornwall, 1977). Technological advance is concentrated in the manufacturing sector and diffuses from there to other economic sectors such as the service sector. Though technological advance in service sectors continues to accelerate, this aspect of manufacturing continues to be very important for developing countries engaged in catch up.

Eight, linkage and spillover effects are stronger for manufacturing than for agriculture or mining. Linkage effects refer to the direct backward and forward linkages between different sectors. Spillover effects refer to the disembodied knowledge flows between sectors. Linkage and spillover effects are presumed to be stronger within manufacturing than within other sectors. Linkage and spillover effects between manufacturing and other sectors such as services or agriculture are also very powerful.

Finally, as per capita incomes rise, the share of agricultural expenditure in total expenditure declines and the share of expenditure on manufactured goods increases (Engel’s law). Countries specialising in agricultural and primary production will not profit from expanding world markets for manufacturing goods.

Concluding Remarks

Some of the recent secondary empirical literature (e.g. Fagerberg and Verspagen, 1999, 2002; Timmer and de Vries, 2007), confirms the important role of manufacturing for growth and catch up in developing countries, but does indicates that this role is weakening over time. The strongest contributions are found in the period 1950-1973, when there were special opportunities for absorption of mass production techniques by developing countries. In the advanced economies, the role of services has become much more important.

A reading of our own empirical evidence confirms that the role of manufacturing was most important in the first decades after 1950. However, an analysis of countries experiencing rapid catch up after 1973, points to the continued importance of manufacturing as an important engine of growth and catch up.

WIDER Angle newsletter, May 2009
ISSN 1238-9544