|Draft for comments
Pattern of Economic Growth and its Implication for Employment
While a number of countries of the developing world have witnessed significant progress in reducing poverty, the stubborn persistence of poverty remains a major challenge in many countries. Progress in poverty reduction has been quite impressive in certain parts of Asia, e.g., in the countries of East and South East Asia (ESEA). But even in some of those countries, it has not been possible to sustain the rate of poverty reduction when they were hit by shocks like the economic crisis of 1997-98, the food crisis of 2008 and the global economic crisis that is currently (i.e., during 2008-09) unfolding. On the other hand, in countries of South Asia, the rate of poverty reduction has been generally lower than that in countries of ESEA. Countries of sub-Saharan Africa fare much worse in terms of poverty reduction, and are less likely than others to achieve the Millennium Development Goal of poverty reduction (viz., one of cutting poverty by half in 2015 compared to the level in 1990).
Based on some projections (e.g., those by World Bank, 2003), it seems that the MDG of poverty reduction would be attained at the global level; but doubts remain about certain regions and countries. If China is excluded from the global calculations, the percentage below the official poverty line of US$1 per capita per day for the rest of the world is projected to decline from 28.5 per cent in 1990 to 15.7 per cent in 2015 –thus remaining over half of the 1990 level. And if one broadens the definition of poverty beyond income poverty to accommodate the multidimensional concept, the situation looks even less rosy.
The above background has led development practitioners and policy makers at various levels to look for all possible mechanisms for speeding up the rate of poverty reduction. And productive employment is recognized as critical in that respect. The experience of countries that succeeded in reducing poverty significantly and at high rates indicates the importance of sustained high rates of economic growth in achieving the result. However, some recent studies (Islam, 2006b; Khan, 2001, and 2007) have come up with an equally important finding that high growth is not enough; the pattern and sources of growth and the manner in which its benefits are distributed are extremely important from the point of view of achieving the goal of poverty reduction. There are country experiences demonstrating that there is no invariant relationship between economic growth and poverty reduction, and that variables relating to employment and labour markets are critical in determining the poverty reducing outcome of growth2. Empirical exercise based on cross-country data (Islam, 2006a) demonstrates that the employment intensity of economic growth has a significant influence on the rate of poverty reduction.
One important question in the context of the nexus (mentioned above) between economic growth, employment and poverty reduction is what is really meant by employment-intensive growth and whether such growth implies any compromise on productivity and efficiency through the adoption of primitive/outmoded technology. Of course, technology is an important element influencing the quantity of employment associated with a given amount of output, but not the only one. The sector composition of output is also important in determining the employment outcome of growth. Higher growth of sectors that are labour-intensive could lead to an increase in the employment intensity of growth.
Conceptually, employment intensity of economic growth (as measured by the elasticity of employment with respect to output) is inversely related to labour productivity. However, if employment-intensive growth is achieved through higher growth of labour-intensive sectors, it would not necessarily imply sacrificing labour productivity altogether. Indeed, it is possible to combine high elasticity of employment with some increase in labour productivity.
The basic objective of the present paper is to empirically test the proposition that countries where the pattern of growth was characterized by high growth of labour-intensive sectors were the ones that achieved more employment-intensive growth. Conversely, the countries where the pattern of growth does not exhibit such a character demonstrate lower employment intensity of growth.
The mirror image of the above relationship between the pattern and employment intensity of growth is the relative contribution of labour productivity and quantity of labour in output growth. The countries which exhibit a lower contribution of labour productivity to output growth are the ones that are characterized by higher employment intensity of growth. In quest of a pattern of growth that is conducive to a high rate of employment growth, the present paper would examine whether there are countries that exhibited high share of employment in output growth and yet were able to combine employment growth with growth in labour productivity.
The paper is organized as follows. Section 2 attempts to clarify some of the conceptual issues involved in the empirical exercises presented in the paper. Section 3 is devoted to an examination of the sector composition output in selected developing countries. That would be done at broad sector levels as well as at the sub-sector level for the manufacturing sector. The basic purpose of the empirical analysis of that section would be to examine if the experience of employment-intensive growth has a systematic relationship with the pattern of growth. Section 4 is devoted to an examination of the relative contribution of labour productivity and quantity of labour input to output growth. The attempt there would be to see how countries with different degrees of employment intensity of growth have performed in terms of combining labour productivity and employment in achieving output growth. Section 5 of the paper presents a few concluding observations based on its empirical findings.
2. The Conceptual and Analytical Issues Involved
2.1. Measuring employment intensity of output growth through employment elasticity
In order to measure the employment outcome of economic growth, a summary indicator is needed of the degree of employment growth that is associated with a given output growth. The term employment intensity of growth is increasingly being used in that context. A measure of the employment intensity of growth can be provided by the elasticity of employment with respect to output growth (which is often referred to as employment elasticity). Quantitative estimates of employment elasticity are based on the assumption that employment is primarily a function of output. Based on this assumption, employment elasticity can be measured either arithmetically (i.e., by dividing the proportionate change in employment by the proportionate change in output during a given period) or by applying the econometric method of regression analysis where a functional relationship between employment and output is postulated and estimated.
But in many developing countries, available data on employment, especially for the informal segment of the economy, often does not reflect the real demand for labour. In such economies, estimates of employment elasticity for the economy as a whole may not be very meaningful because a high figure may simply reflect the labour force taking refuge into low productivity employment of a residual character in the absence of unemployment benefits. For such countries, the concept and its empirical estimates may be more meaningful with respect to formal sectors where the assumption of employment data reflecting the demand for labour labour may be more valid.
When defined in the manner described above, it can be shown that employment elasticity is inversely related to labour productivity, and an employment elasticity of greater than one implies a decline in labour productivity (Islam, 2006a; Kapsos, 2005). In order to allow for growth in labour productivity, employment growth must be lower than output growth, which in turn implies that employment elasticity has to be lower than unity.
Employment elasticity of a sector reflects two different elements: (i) the composition of the sector, and (ii) technology used in various lines of production within the sector3. So, even if there is no change in technology, employment elasticity may change if the composition of the sector changes because employment elasticity of a sector is the weighted average of elasticities of the component sub-sectors (Khan, 2001). Take the case of the manufacturing sector. If the more labour intensive components have a greater weight in the sector, the overall elasticity would be correspondingly higher than in a situation where more capital-intensive components dominate the manufacturing sector. The same argument would apply for an economy as a whole. If, for example, manufacturing and service sectors of the economy exhibit higher employment elasticity compared to agriculture, an increase in the weights of the former in the GDP would lead to a rise in the overall employment elasticity.
2.2. The sectoral pattern of growth
It follows from the above that if a country’s policy environment is such that its labour-intensive industries (or sectors) have a greater incentive to grow at faster rates, it is quite possible for such industries (or sectors) to assume a greater weight in the manufacturing sector (or the economy as a whole). And hence, the overall employment elasticity for the manufacturing sector (or the economy) may rise without requiring any individual sub-sector (or sector) to adopt more labour-intensive technology than it is employing at present. All that would be needed to achieve such a result would be to ensure that the policy environment is conducive to (or at least does not discriminate against) the growth of the relatively more labour-intensive sub-sectors.
The above discussion brings one to the issue of the sectoral pattern of economic growth and development strategies pursued by developing countries. A couple of points may be made in this regard. First, the historical experience of the present-day developed countries would show that in the course of their development they generally went through a particular pattern of structural change in their economies, viz., diminishing share of agriculture in GDP associated with, first, increasing share of industries followed by an increasing share of services. Some of the late developers, especially some countries of ESEA (e.g., the Republic of Korea, Malaysia, Thailand, etc.), also followed very similar pattern.
Different economists have offered different explanations for the observed pattern of structural changes in an economy as mentioned above; notable amongst them are Fisher (1939), Clark (1951), Kuznets (1966, 1971) and Kaldor (1966, 1967). The major differences in their explanations lie in their relative emphasis on demand and supply side factors. Fisher, Clark and Kuznets focus mainly on the demand side and base their arguments on changes in the income elasticity of demand for products and services of different categories. They refer to the famous Engel’s law that income elasticity of demand for agricultural products is lower than that for manufactured goods. Likewise, the income elasticity of demand for services is higher than that for manufactured goods. Hence, with increases in incomes, demand for agricultural goods would decline and that for manufactured goods would increase. When production responds to this pattern of changes in demand, the share of agriculture in GDP starts declining and that of industry starts increasing. At some level of income, the income elasticity fo demand for services would exceed that for industrial goods, leading to rapid increases in the share of services in GDP. At that point, the share of industry in GDP starts declining.
Kaldor brings in some supply side arguments in addition to the above mentioned demand factor. According to him, agriculture is subject to diminishing returns to land (which is the major factor of production in that sector), and is thus unable to sustain its growth beyond a point. Industry, on the other hand, is not subject to diminishing returns to its major factor, and does not face the same constraint as agriculture. This, coupled with higher income elasticity of demand for industrial products, enables the industrial sector to act as the engine of growth. And growth in that sector pulls overall GDP growth through linkage effects. Moreover, as productivity, especially of labour, is usually higher in manufacturing, higher growth of that sector leads to higher productivity in the economy as a whole. And unless the pattern of growth in that sector is very capital-intensive, it would absorb labour transferred from agriculture which normally has lower productivity. Thus a process of growth starts which is associated with a gradual transfer of labour from low productivity sectors to higher productivity sectors4. Critical to the success of this process is the rate of growth of the manufacturing sector relative to that of agriculture and to that of the overall economy.
It is of course possible to criticise the demand-based explanation of structural change by bringing in the possibility of trade. In an open economy, demand for manufactured goods could be met, at least partly, through imports. Hence the manufacturing sector does not need to grow at the same rate as implied by the growth of demand for such goods. However, when trade is introduced into the system, one has to also consider the possibility of exports. And that would bring one to the second point concerning structural changes that could be envisaged for developing economies. The nature of the trade regime can actually influence the pattern of industrialization in terms of its sector composition.
One prediction which follows from the standard theory of international trade is that greater trade openness should lead labour-abundant countries to specialize in the production and export of goods which intensively utilize their abundant factor, viz., labour5. Thus, greater openness should lead to high growth of labour-intensive industries; and as the weight of such industries in the overall sector increases, there should be a rise in the employment intensity of output in the manufacturing sector as a whole. This process should continue until surplus labour gets exhausted and wage rates start rising, at which point employment elasticity may also start declining.
Neither Kaldor nor the earlier economists (e.g., Clark and Fisher) theorizing about structural changes in an economy mentioned anything about the composition of the manufacturing sector and its possible implication for employment. But from the point of engendering a process of quick transfer of labour from low productivity agriculture to industry, it is important not only to have high growth of manufacturing but also to have high growth of employment in that sector. That would be possible when labour intensive industries have a higher weight in the growth of the manufacturing sector. Such a pattern of industrialization would, of course, depend on a variety of factors, e.g., the pattern of demand in the domestic as well as external market, the policy environment prevailing in an economy, etc.
2.3. Employment-intensive growth and labour productivity6
Mention has already been made above of the inverse relationship between employment elasticity and labour productivity which implies the possibility of a trade-off between employment growth and labour productivity. In reality, however, this trade-off does not have to be very serious. One can see this easily if one remembers that in an accounting framework, both the quantity of labour input and labour productivity contribute to output growth. Depending on the policies pursued, a country may be able to achieve a balanced contribution of both these elements towards output growth. This proposition is explained further below.
For an economy as a whole, output is equal to the product of the labour force employed and labour productivity. This can be expressed through the following identity:
Y ≡ L × Y/L, (1)
where Y and L stand respectively for output and employment.
For small changes, one can write the above as
ΔY = ΔL + ΔY/L, (2)
where Δ indicates growth rate.
Expression (2) implies that growth in output is the sum of the growth of employed labour force and growth of labour productivity. Thus, both employment in quantitative terms and labour productivity can potentially contribute to output growth. Indeed, if output growth is sufficiently high, there could be scope for substantial increases in both employment and productivity growth. And that has been the experience of East and South East Asian economies like those of Rep of Korea, Taiwan, China, Malaysia, and to a lesser extent in Indonesia and Thailand (especially before they were hit by the East Asian economic crisis in 1997-98).
3. Employment Intensity and Sectoral Pattern of Economic Growth: The Asian Experience
Recent studies (e.g., Islam, 2002, 2008; Khan, 2001, 2007) have shown that in general the countries of ESEA region achieved more employment-intensive growth (demonstrated by higher employment elasticity _both at the aggregate level and for manufacturing) compared to countries of South Asia. Not only was the degree of employment intensity lower, but there has been a tendency for that to decline in the latter set of countries (Islam, 2008). How does this experience look in relation to their experience with regard to the sectoral pattern of growth? Data presented in Table 1 are intended to contribute towards an answer to this question. Following up on the discussion in section 2.2, data on growth of GDP and of manufacturing output for selected countries of Asia are presented in Table 1.
What comes out clearly from Table 1 is the higher growth of manufacturing output in relation to GDP growth achieved by the countries of ESEA compared to those of South Asia. In the Rep of Korea, for example, the elasticity of growth of manufacturing output with respect to GDP growth was over 2 in the 1960s and dropped to just below 2 in the 1970s. The figure dropped below 1.5 only during the 1980s. In Malaysia, the figure was between 1.5 and 2 for almost three decades (1970-96). Indonesia and Thailand also had a similar experience. In contrast, this figure has been in the range of 1.3 to 1.4 in India, and lower in Pakistan. In Bangladesh and Sri Lanka, there has been a considerable degree of fluctuation in the elasticity of manufacturing growth with respect to GDP growth. On the whole, not only has overall economic growth been higher in the countries of ESEA (except Philippines) than in South Asia, the manufacturing sector has been the major driver of growth in the former. It thus appears that only those countries were able to achieve a sectoral pattern of growth outlined by Kaldor (and others like Clark, Fisher and Kuznets as mentioned in section 2.2). Not surprisingly, they are also the countries that were able to achieve more employment intensive growth and Lewis type transformation of their labour markets _ albeit at varying speed7.
In order to get further insight into the sectoral pattern of economic growth, it would be useful to look at the composition of the manufacturing sector. Using UNIDO data (Indstat 3, 2005)8, an attempt has been made to rank sub-sectors of manufacturing industries (at the three-digit level) of various countries according to their capital-labour ratios. Based on such rankings, five most labour-intensive and five most capital-intensive industries for each country were identified for 1980, 1990 and 2002. The shares of the five most labour-intensive and five most capital-intensive industries in total manufacturing value added were then calculated. These shares are presented in Table 2. The idea behind this exercise is to see whether there has been a systematic change in the sector composition of the manufacturing sector in either direction.
Data presented in Table 2 point out a few interesting aspects of the sectoral pattern of industrialization in the selected countries of Asia. First, both Rep of Korea and Malaysia show an increase in the share of labour intensive industries up to 1990 and a decline thereafter. Korea also experienced a fall in the share of capital-intensive industries till 1990 and a rise thereafter. In Malaysia, on the other hand, the share of such industries in 1990 was already higher than in 1980, although there was a slight decline after that. The above figures indicate that while Korea provides a classic example of labour-intensive industrialization in its early phase of development, Malaysia comes close to that. Thailand also witnessed a rise in the share of labour intensive industries between 1990 and 2002. In contrast, India witnessed a gradual decline in the share of labour intensive industries and a rise in the share of capital-intensive industries. Pakistan and Sri Lanka also show similar trends, although not so clearly.
The figures for Bangladesh presented in Table 2 need to be interpreted with caution. They indicate a substantial rise in the shares of both top five labour-intensive and top five capital-intensive industries. These figures themselves are not implausible, and might indicate growth taking place at two ends of the spectrum. Indeed, there has been very rapid growth of one labour-intensive industry, viz., ready made garments, which may be reflected in the sharp increase in the share of the top five labour-intensive industries. It should, however, be noted that apart from this single industry, there has not been a similar growth in any other labour-intensive industry. In fact, the performance of another major labour-intensive industry of the country, viz., leather and leather products has been rather disappointing, resulting in a decline in its share in total manufacturing value added (Ahmed, et al., 2008).
In the case of India, data from national sources (viz., the Annual Survey of Industries) corroborate the findings based on UNIDO data. Using the ASI data, one recent study (Palit, 2008) concludes that the composition of Indian manufacturing has not undergone very substantive changes during the 1990s. That study actually shows that the overall share of the labour-intensive industries in total manufacturing output has declined significantly between 1990-91 and 2003-04. In 1990-91, the top five labour-intensive industries (viz. food and other food products, beverages and tobacco, wood products and furniture, other textiles, leather and fur products) accounted for nearly 41% of manufacturing output. In 2003-04, the share of the top five (which in that year were other textiles, leather and fur, beverages and tobacco, food products, and other manufacturing) dropped to 28.32% (Palit, 2008).
On the whole, data on the sector composition of the manufacturing sector clearly demonstrates that it is the countries of ESEA (with the exceptions of Indonesia and the Philippines) that were able to achieve a structural change towards labour-intensive industries during the early phases of their development when they had surplus labour. Countries of South Asia have not been able to achieve that. This is quite striking as far as India is concerned because that country marked a significant change in the strategy of development since 1991 when far-reaching reforms were introduced in the areas of both domestic economic policies and trade policies. And that implies that trade liberalization alone does not automatically enable a country to attain labour-intensive industrialization. There are other factors that influence the policy environment and hence the pattern of industrialization of a particular country that need to be addressed if a country intends to embark on a process of development that would rapidly absorb its surplus labour in productive employment9.