Structural change and labour productivity trend in the non-agricultural sector: a study of Asia.

• Author: Roy, Amrita

1. Structural change and relative labour productivity in different sectors

   For most developed countries, the structures of production and employment have generally been seen to transform in a similar manner over the course of economic development. With development, the importance of the agricultural sector declines both in terms of its share in national income and employment and the non-agricultural sector becomes the driving force of the economy (Fuchs 1969; Kuznets 1979).

   Over the last few decades, many of the developing countries have transformed from one, based on agriculture to a service sector dominated economy. But contrary to the general trend noted in the developed countries, the gap between output and employment shares of the service sector has remained wide in many developing countries in Asia. For example, in 2010 in South Asia, average contribution of the service sector in national income was 54 percent in contrast to its 27 percent share in total employment (World Bank (2012) (1)). Considering the experience of India, it has been noted that the service sector has not been able to generate enough employment in India even though it accounts for the major share of national income (Bhattacharya and Mitra 1990; Gordon and Gupta 2004; Banga 2005; Dasgupta and Singh 2006; Joshi 2004, 2008). Similar to India, in case of China, there exists a considerable gap between the output share and employment share of the service sector but it is not as wide as observed in India.

   We also see that compared to the early industrialized countries of East Asia (e.g. Japan and South Korea), in the late industrialisers of East Asia (e.g., Philippines and Thailand) and also in India and China, the employment shares of the leading sectors (industry and service sector) are rising less than proportionately to their output shares. All these East Asian countries starting out as predominantly agrarian economies are now in different phases of development. This provides us the scope for comparing the development phases of these countries while looking at the relative productivity trends of their different economic activities.

   Main problem related to a continuous increase in the gap between output and employment shares (or disproportionate growth) in the non-agricultural sector is that it might result in worsening of income distribution. If this disproportionate growth in the non-agricultural sector is regarded as a short run phenomenon, which was also a part of the process of development in the presently developed countries, then in the long run ultimately the process will reverse. However, if such a rising trend signals a long-run tendency for the gap between output and employment shares of the non-agricultural sector to persist or widen over time then this is a matter of serious concern. With a rising gap between the output and employment shares, a large proportion of the population in the economy might remain trapped in a state of increasing relative poverty. At the same time, with sufficiently low rates of employment growth in the non-agricultural sector, a state of absolute poverty might exist as well.

   In many developing countries, agricultural sector acts as a residual sector and absorbs the surplus labour that do not get employed in the non-agricultural sector. In this situation, given a constant rate of growth of labour force in the economy, a declining labour productivity of the agricultural sector might indicate a disproportionate growth of different sectors. Further, the gap between the labour productivities of the agricultural and non-agricultural sectors might indicate the severity of the disproportionate growth. One way to look at the issue of disproportionate growth, therefore, is to study the change in the labour productivity ratio between these two sectors over time.

   In this paper, we develop a model that explains the different trends in labour productivity of the non-agricultural sector relative to the agricultural sector. We consider a simple two-sector model, involving a high productivity sector (non-agricultural sector) and low productivity sector (agricultural sector), of a dual economy. This model predicts that even though technological development accounts for the relationship between changes in sectoral output and employment shares in an economy, in the developing countries the importance of low productivity sector in absorbing surplus labour also affects this relationship. The long run labor productivity of non- agriculture relative to agriculture approaches unity depending on growth rate of labor force and productivity growth both outside and within the agricultural sector.

   To check for the validity of the postulations of the model, using data of eight East Asian countries over 1970-2014, we have estimated the relationship between labour productivity ratio of the agriculture and the non-agricultural sectors and labour absorption capacity of the non- agricultural sector, controlling for technological development of the country and growth of the sectors. In our estimation results, we find that controlling for other factors, non-agricultural employment relative to total labour force is positively related to the labour productivity ratio of the agricultural sector and the non-agricultural sector. At the same time, we also find that labour productivity in the non-agricultural sector has a significant and negative relation to the labour productivity ratio of the two sectors.

   The paper is organized as follows: In section 2, we consider the trends in labour productivity ratio between the non-agricultural sector and the agricultural sector for a set of East Asian countries for the period 1970 to 2014. In section 3, we present the historical evidence on the relation between changes in employment structure and changes in production in the course of development. In section 4, we construct a simple two-sector model of a dual economy to study the direction of movement in labour productivity in the non-agricultural sector of an economy. Empirical analysis of the model discussed in section 4 has been done in section 5. Conclusions based on these analyses have been summarized in section 6.

2. Trends of labour productivity ratio between the agriculture and the non-agricultural sectors

   In the last four decades, the labour productivity in the agricultural sector was significantly lower compared to the industrial and the service sectors in India, China, Indonesia, Malaysia, Philippines, Thailand, Japan and South Korea. Further, the agricultural sector witnessed a continuous decline in labour productivity and its share in national income. At the same time, throughout this period the agricultural sector has been much more important as a source of employment generation than as a source of income. It has acted as the primary residual sector for employment generation. Briones and Felipe (2003) noted that since the 1970s, while the output share in Asia declined at 2.5 percent per annum, the employment share declined at about 1.7 percent. Therefore, the role of the agricultural sector in the Asian countries has been very different compared to the industrial and the service sectors.

   In Figure 2.1, we present the labour productivity ratio between the agricultural and the non- agricultural sectors in India, China, Indonesia, Malaysia, Philippines, Thailand, Japan and South Korea (2).

   Looking at the trends, we observe that the labour productivity ratio is changing significantly over time. This changing pattern of the labour productivity ratio is different for different countries. For example, the ratio shows a declining trend in China, India and Malaysia. In case of Japan, Korea and Philippines we find an increasing trend. For Indonesia and Thailand the trend takes an "U" shape.

3. Historical evidence on the relation between changes in the employment structure and changes in production in the course of development

   The inter-sectoral movement of labour associated with the changing structure of production in an economy constitutes a different aspect of structural change. Based on the historical experiences of the developed countries, studies have noted that patterns of structural change in production and employment are quite similar in an economy. Changes in the employment shares of different sectors generally move in line with changes in the output shares of those sectors with a time lag.

   Data reported in the studies of Fuchs (1969) and Kuznets (1979) show that in the early phase of development, countries considered to be developed today witnessed a huge decline in the contribution of the agricultural sector both in national income and in total employment. The labour released from this sector was absorbed in the industrial and the service sectors. Finally, as the process of deindustrialization started in these developed countries, labour released from the industrial sector was absorbed in the service sector. Considering the non-agricultural sector we see that in the early phase of transition, the rise in output share of the non-agricultural sector was greater than the corresponding increase in employment share in these developed countries. Movement of labour from one sector to the other in accordance with changing output shares occurred with a time lag. Eventually, employment share of the non-agricultural sector converges to the output share of the non-agricultural sector. This is clearly evident from the experience of the 13 developed countries that Kuznets considered (depending on availability of data) over a period from the late seventeenth century to the middle of the 1960s (reported in the Appendix Table A.1).

   ...