What Explains the Job Creating Potential of Industrialisation in the Developing World? Kunal Sen Global Development Institute, University of Manchester www.kunalsen.org.uk
The False Promise of Industrialisation? In the early development thinking (a la Lewis), industrialisation was the route to both growth and employment generation. In the Lewis Model, as long as there was surplus labour, the capitalist sector (manufacturing) could expand, drawing labour from unproductive jobs in agriculture, retail trade and other sectors with disguised unemployment. However, successful industrialisation has remained a challenge for many developing countries. Neither has industrialisation, wherever it has occurred, been able to deliver the jobs that it had promised to do.
The Scarcity of Good Jobs
Manufacturing still remains the key sector for productivity enhancing employment growth in most developing economies
The Uneven Success with Industrialisation around the World
The South Asian Experience
Job Creation in Industry around the World 35 Employment Share of Industry in Total Employment 30 p e r c e n t 25 20 15 10 5 0 East Asia and Pacific Europe and Central Asia Latin America and Caribbean Middle East and North Africa South Asia
Job Creation in Industry in South Asia
Why do some countries do better in creating jobs through manufacturing than others? I will argue that industrialisation can affect job creation through three effects the scale effect, the compositional effect and the labour intensity effect. Using ISIC 4 digit UNIDO data for over 100 countries from 1990 to 2009, I will show that these three effects have worked in diverse ways in developing countries. I will then assess the theoretical literature on what drives these three effects. Finally, I will present empirical evidence on the determinants of these effects.
Revisiting Lewis Lewis provided a powerful framework of analysis on why manufacturing can be a key driver of growth and structural transformation in the developing world. In Lewis s model, manufacturing growth pulls workers from low wage/low productivity jobs in agriculture to high wage/high productivity jobs in manufacturing. But Lewis did not provide an answer on why some countries did better in creating manufacturing jobs than others.
The Lewis Model
Structural Transformation
What explains the rightward shift in the demand for labour curve in the capitalist sector? End of Surplus Labour
Q 1 Q 0 Scale Effect A 0 A 1 w D 0 D 1 A 0 A 1 Manufacturing L
Q 1 B C Q 0 A Labour Intensity Effect A 0 A 1 A 2 w D 0 D 1 D 2 A 0 A 1 A 2 L Manufacturing
Compositional Effect Capital Intensive Sector Labour Intensive Sector Q 1 1 B 1 Q 1 2 B 2 Q 0 1 A 1 Q 0 2 A 2 A 0 1 A 1 1 A 2 0 A 2 L 1 1 L 2
Q 1 1 B 1 A 1 Output increase in Capital Intensive Sector Q 0 1 A 0 1 A 1 1 L 1 w A B A 0 A 1 Manufacturing L
Q 1 2 B 2 A 2 Q 0 2 Output increase in Labour Intensive Sector A 0 A 1 L 2 w A B A 0 A 1 L Manufacturing
The Employment Creating Potential of Manufacturing How much employment is created by manufacturing consists of three effects: 1) Scale Effect: A rightward shift in the labour demand curve due to an increase in manufacturing output. 2) Labour Intensity Effect: A rightward shift in the labour demand curve due to an increase in the labour intensity of production, across all manufacturing sectors. 3) Compositional Effect: A rightward shift in the labour demand curve due to an increase in output in labour intensive sectors relative to capital intensive sectors.
A Simple Decomposition Denote L as employment, Q as real output. Let i be the industry, and t be the timeperiod. w is the weight of each industry in total manufacturing output. Denote labour intensity as L/Q = A. Start with the identity: L t = Q t [ i w it (L it /Q it )] (1)
Decomposing Total Employment Change Now consider change in L from one time-period (t 1 ) to the next (t 2 ). L t = Q t [( i w it A it ) + Q t [( i w it A it ) + Q t [( i w it A it ) (2) Here, Q t, i w it and A it are evaluated at t=t 1 The first term on the Right Hand Side of (2) is the scale effect, the second term the labour intensity effect, and the third term is the compositional effect.
Data Employment and nominal output from the UNIDO ISIC 4 digit industry data-base. From 1990 to 2010. To obtain real output, we used manufacturing price deflators. We took period averages for 1990-94, 1995-99, 2000-04 and 2005-10. We only retained countries where we had data on L, nominal Q and price deflators for all years in a given time-period.
Employment Changes in 1995-99, relative to 1990-94 600000 500000 400000 300000 200000 100000 0-100000 -200000
The Three Effects, 1995-99 800000 700000 600000 500000 400000 compositional effect labour intensity effect scale effect 300000 200000 100000 0-100000 -200000-300000
Poland Tonga Slovenia Mexico Malta Malawi Argentina Trinidad and Tobago Czech Republic Lebanon Peru Kyrgyz Republic Mauritius Nepal Hungary Ethiopia Singapore Cambodia Denmark Senegal Slovak Republic Cyprus Turkey Panama Bulgaria Tajikistan Vietnam Jordan Thailand Ecuador Indonesia Azerbaijan Bolivia Lithuania Georgia Eritrea South Africa Botswana Uruguay Philippines Estonia Austria Latvia India Brazil Mongolia Employment Changes in 2000-04, relative to 1995-99 2000000 1500000 THAILAND 1000000 MEXICO VIETNAM 500000 0 POLAND INDIA BRAZIL -500000
Poland Tonga Slovenia Mexico Malta Malawi Argentina Trinidad and Tobago Czech Republic Lebanon Peru Kyrgyz Republic Mauritius Nepal Hungary Ethiopia Singapore Cambodia Denmark Senegal Slovak Republic Cyprus Turkey Panama Bulgaria Tajikistan Vietnam Jordan Thailand Ecuador Indonesia Azerbaijan Bolivia Lithuania Georgia Eritrea South Africa Botswana Uruguay Philippines Estonia Austria Latvia India Brazil Mongolia The Three Effects, 2000-04 2500000 2000000 1500000 POLAND compositional effect labour intensity effect scale effect THAILAND INDIA 1000000 MEXICO BRAZIL 500000 0-500000 -1000000-1500000 -2000000
Cyprus Brazil Slovak Republic Madagascar Ukraine Ethiopia Singapore Uruguay Tanzania Malawi Malaysia Estonia Tajikistan Vietnam Trinidad and Tobago Slovenia India Eritrea Georgia Senegal Latvia Nepal Bulgaria Lithuania Poland Colombia Peru Czech Republic Jordan South Africa Turkey Azerbaijan Mongolia Panama Botswana Malta Fiji Ecuador Kyrgyz Republic Mauritius Austria Kazakhstan Hungary Denmark Indonesia Philippines Morocco Thailand Mexico Employment Changes in 2005-10, relative to 2000-04 4500000 4000000 3500000 3000000 VIETNAM INDIA 2500000 2000000 1500000 1000000 500000 0-500000
Cyprus Brazil Slovak Republic Madagascar Ukraine Ethiopia Singapore Uruguay Tanzania Malawi Malaysia Estonia Tajikistan Vietnam Trinidad and Tobago Slovenia India Eritrea Georgia Senegal Latvia Nepal Bulgaria Lithuania Poland Colombia Peru Czech Republic Jordan South Africa Turkey Azerbaijan Mongolia Panama Botswana Malta Fiji Ecuador Kyrgyz Republic Mauritius Austria Kazakhstan Hungary Denmark Indonesia Philippines Morocco Thailand Mexico The Three Effects, 2005-10 7000000 6000000 5000000 4000000 3000000 VIETNAM INDIA compositional effect labour intensity effect scale effect 2000000 1000000 0-1000000 -2000000-3000000
What are the determinants of the three effects? 1) Globalisation (trade, capital flows, technology) 2) Labour Regulations 3) Human Capital
Trade and the Scale Effect One determinant of the size of the manufacturing sector is a country s comparative advantage, which may in turn reflect factor endowments. In the model first proposed by Krueger (1977) and extended by Leamer (1987), the crucial variable determining trade and production structure is the land/labour ratio. Thus, land-abundant developing countries such as those in Africa and Latin America, would be more likely to specialise in primary commodities while developing countries in Asia would be more likely to specialise in (labour-intensive) manufactures.
Ricardian Theory and the Scale Effect An alternative view would explain a country s comparative (dis)advantage in manufacturing in Ricardian terms where differences in technology across sectors explain the effects of trade. In this case, the size of the manufacturing sector in a country is determined by its overall competitiveness which in turn is partly a result of technological capabilities in manufacturing.
Trade and the Labour Intensity Effect Industry level impacts on labour intensity may occur via induced productivity effects, as firms shed labour in response to external competitive pressures, due to either greater export orientation or increased import penetration (Levinsohn 1999) It could also be due to trade-induced technological transfers (for example, via an increase in the importation of capital goods) (Wood 1994). Or a fall in the relative price of capital goods (Das and Sen 2015).
Trade and the Compositional Effect A key prediction of the standard two factor Heckscher-Ohlin model is that with international trade, developing countries with plentiful supplies of labour will export labour-intensive commodities and import commodities with relatively higher capital requirements (Krueger 1978). As a labour surplus countries integrate, they will observe a change in the composition of its output towards more labour-intensive activities..
Labour Regulations Stricter labour laws push up manufacturing wages (relative to productivity) and make manufacturing uncompetitive (Lazear 1990, Besley and Burgess 2003, Botero et al. 2004) (scale effect). Stricter labour regulations can shift composition of manufacturing output to capital intensive sectors and/or increase the capital intensity of production (Gupta et al. 2009, Hasan et al. 2013) (compositional and labour intensity effects)
Human Capital Higher schooling attainment in the workforce can allow developing countries to compete in manufacturing (Mincer 1981) (scale effect). If skilled labour is complementary to capital, it can lead to a compositional effect towards capital intensive sectors.
Variables Globalisation: KOF Index of Globalisation, measures actual flows of trade and investment, and by restrictions on trade and investment, such as tariff rates. Labour Regulation: Workers rights, from Cingranelli-Richards (CIRI) Human Rights data-base, 202 countries, 1981-2011. Human Capital: Years of schooling, 25+ age group, Barro-Lee data.
Scale Effect and Globalisation -.04 -.02 0.02.04.06 20 40 60 80 100 (mean) dr_eg scale Fitted values
Compositional Effect and Globalisation 0.01.02 -.02 -.01 20 40 60 80 100 (mean) dr_eg comp Fitted values
Labour Intensity Effect and Globalisation 0.02.04 -.04 -.02 20 40 60 80 100 (mean) dr_eg labint Fitted values
Scale Effect and Labour Regulations -.04 -.02 0.02.04.06 0.5 1 1.5 2 (mean) ciri_worker scale Fitted values
Compositional Effect and Labour Regulations 0.01.02 -.02 -.01 0.5 1 1.5 2 (mean) ciri_worker comp Fitted values
Labour Intensity Effect and Labour Regulations 0.02.04 -.04 -.02 0.5 1 1.5 2 (mean) ciri_worker labint Fitted values
Scale Effect and Human Capital -.04 -.02 0.02.04.06 0.00 5.00 10.00 15.00 (mean) bl_asy25mf scale Fitted values
Compositional Effect and Human Capital 0.01.02 -.02 -.01 0.00 5.00 10.00 15.00 (mean) bl_asy25mf comp Fitted values
Labour Intensity Effect and Human Capital 0.02.04 -.04 -.02 0.00 5.00 10.00 15.00 (mean) bl_asy25mf labint Fitted values
The Correlates of the Three Effects Independent Variables Globalisation 0.0002* (1.93) Labour Regulations Overall Effect Scale Effect Labour Intensity Effect -0.004** (2.32) Human Capital 0.0001 (0.02) 0.0004*** (3.10) -0.004 (1.33) 0.001 (1.06) -0.0001 (1.52) -0.003 (0.15) -0.002*** (2.75) Compositiona l Effect -0.0001*** (2.36) -0.004 (0.42) 0.0006* (1.82) R square 0.11 0.25 0.31 0.10 No of observations 85 85 85 85
Conclusions Job Creation in Manufacturing is a combination of three different effects. These effects do not need to go in the same direction. In countries where manufacturing output growth has led to large employment increases, these effects have tended to go in the same direction. Globalisation affects these effects differently positive on scale, negative on composition. Human capital negative on labour intensity, positive on composition. Labour regulations important overall, but not for individual effects.
Policy Implications Challenge for policy makers is to harness the different ways that industrialisation affects job creation, so that the effects are complementary, not offsetting each other. To create synergies between policies for globalisation and human capital formation.