Rethinking Economic Change in India: Labour and livelihood 0203023102, 0415349893

Citation preview

Rethinking Economic Change in India South Asia, home to a quarter of the world’s workforce, has in recent years experienced economic growth and unprecedented levels of global integration. Despite this, there continues to be a large disparity in income. Rural poverty remains acute and extensive and manual labourers still constitute the majority of India’s poor. Rethinking Economic Change in India explores the historical roots of India’s high levels of poverty by placing labour history in the context of economic change within the region. Roy departs from the commonly used political-economic approach to the labour history of India, and considers the role of markets and resources in shaping the condition of rural labour, the transition in the economic position of women, the condition of informal labour in Indian industry and labour-intensive industrialization This highly original, thought-provoking book will prove invaluable to students studying labour economics, economic history and economic development in contemporary India as well as for academics interested in the field. Tirthankar Roy has established himself as the leading contemporary economic historian of India and his 2000 book, The Economic History of India, 1857–1947 has proved to be hugely influential.

Routledge explorations in economic history 1 Economic Ideas and Government Policy Contributions to contemporary economic history Sir Alec Cairncross 2 The Organization of Labour Markets Modernity, culture and governance in Germany, Sweden, Britain and Japan Bo Stråth 3 Currency Convertibility The gold standard and beyond Edited by Jorge Braga de Macedo, Barry Eichengreen and Jaime Reis 4 Britain’s Place in the World A historical enquiry into import controls 1945–1960 Alan S.Milward and George Brennan 5 France and the International Economy From Vichy to the Treaty of Rome Frances M.B.Lynch 6 Monetary Standards and Exchange Rates M.C.Marcuzzo, L.Officer and A.Rosselli 7 Production Efficiency in Domesday England, 1086 John McDonald 8 Free Trade and its Reception 1815–1960 Freedom and trade: volume I Edited by Andrew Marrison 9 Conceiving Companies Joint-stock politics in Victorian England Timothy L.Alborn 10 The British Industrial Decline Reconsidered Edited by Jean-Pierre Dormois and Michael Dintenfass 11 The Conservatives and Industrial Efficiency, 1951–1964 Thirteen wasted years? Nick Tiratsoo and Jim Tomlinson

12 Pacific Centuries Pacific and Pacific Rim economic history since the 16th century Edited by Dennis O.Flynn, Lionel Frost and A.J.H.Latham 13 The Premodern Chinese Economy Structural equilibrium and capitalist sterility Gang Deng 14 The Role of Banks in Monitoring Firms The case of the Credit Mobilier Elisabeth Paulet 15 Management of the National Debt in the United Kingdom, 1900–1932 Jeremy Wormell 16 An Economic History of Sweden Lars Magnusson 17 Freedom and Growth The rise of states and markets in Europe, 1300–1750 S.R.Epstein 18 The Mediterranean Response to Globalization Before 1950 Sevket Pamuk and Jeffrey G.Williamson 19 Production and Consumption in English Households 1600–1750 Mark Overton, Jane Whittle, Darron Dean and Andrew Hann 20 Governance, The State, Regulation and Industrial Relations Ian Clark 21 Early Modern Capitalism Economic and social change in Europe 1400–1800 Edited by Maarten Prak 22 An Economic History of London, 1800–1914 Michael Ball and David Sunderland 23 The Origins of National Financial Systems Alexander Gerschenkron reconsidered Edited by Douglas J.Forsyth and Daniel Verdier 24 The Russian Revolutionary Economy, 1890–1940 Ideas, debates and alternatives Vincent Barnett

25 Land Rights, Ethno Nationality and Sovereignty in History Edited by Stanley L.Engerman and Jacob Metzer 26 An Economic History of Film Edited by John Sedgwick and Mike Pokorny 27 The Foreign Exchange Market of London Development since 1900 John Atkin 28 Rethinking Economic Change in India Labour and livelihood Tirthankar Roy

Rethinking Economic Change in India Labour and livelihood

Tirthankar Roy

LONDON AND NEW YORK

First published 2005 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “ To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” © 2005 Tirthankar Roy All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN 0-203-02310-2 Master e-book ISBN

ISBN 0-415-34989-3 (Print Edition)

For Morris David Morris

Contents List of figures

x

List of tables

xi

Preface 1 Introduction

xiii 1

2 Economic history and modern India: redefining the link

23

3 Rural labour and colonialism

44

4 Agricultural labour: lessons from wage data

74

5 Was there an industrial decline in India in the early nineteenth century?

92

6 Labour-intensive industrialization

103

7 Women and industrialization

116

8 Women in the crafts

131

9 Labour and power: a critique of ‘subaltern studies’

143

10 Conclusion

154

Notes

158

References

165

Index

174

Figures 1.1 Wage labour in India, 1901–2001

8

2.1 Land and real agricultural income (1865=100)

41

2.2 Real yield per acre and person (1865=100)

42

3.1 The hypothesized relationship between density of population and labour-ratio

60

3.2 Density (acres) and labour-ratio (%), 1901–31

72

3.3 Density (acres) and labour-ratio (%), 1951

72

4.1 Average wage (Rs/month)

76

4.2 Wage rates, 1920–51

81

4.3 Agricultural labourers: number and earning, 1950–64 (1950=100) 82

Tables 2.1 Employment, 1901–91

28

2.2 Production and wages

31

2.3 Growth rates of Net Domestic Product (NDP) and Population, 1868–69 to 1946–47

37

2.4 Sources of growth in real GDP, 1900–47

42

3.1 Labour-ratio, 1881–1931

53

3.2 Labour-ratio of women (%), 1901–31

54

3.3 Labour-ratio by major region, 1901–51

55

3.4 Labourers and labour-ratio in the Indian Union, 1951–91

57

3.5 Land-population ratios, 1891–1931

63

3.6 Diverse transitions consistent with increase in labour-ratio

70

4.1 Money and real wage in agriculture, 1873–1951

76

4.2 Real wage in major provinces, 1873–1912

77

4.3 Major factors affecting demand for and supply of agricultural labour

83

4.4 Share of wages in agricultural income

85

4.5 Shares in employment and income: agricultural labour and the artisan sector

86

4.6 Weighted average money and real wage of agricultural labourers 91 in British India

6.1 Measures of industrialization in India, 1901–39

106

6.2 Contributions of large scale and small scale in Indian industrialization, 1901–39

107

7.1 Women workers, 1901–91

119

7.2 Gender and industrial organization, 1961

120

7.3 Women-worker ratio (%) in manufacturing, 1961

121

Preface South Asia, home to perhaps a quarter of the world’s workforce, has seen rapid economic growth in some spheres and persistence of poverty and underdevelopment in others. The poor, almost without exception, live on manual labour, which continues on a large scale in the region. What factors account for poverty in this sphere? What conditions enable attempts to escape poverty? This book tries to answer these two questions. It presents a brief general history of labour-intensive occupations in South Asia c. 1870–1970, with specific reference to rural labour, artisans and women workers. Any general narrative on labour needs to be a historical narrative, for occupational structure in South Asia has changed rather slowly in the long run. Economic historians in the region have often attributed the inertia in conditions of labour to various forms of inequality and repression, including, and especially, those sustained by colonialism. Such an approach, besides being overly political, underestimates the capacity of individuals to make a difference by utilizing opportunities offered by expanding market transactions. This book follows a different route. It explains the stable element in the labour experience with reference to resource endowments. The balance between natural resources and labour, and the prospect of diminishing returns to labour, can explain the low average earnings in which large numbers of the working poor remained trapped for a long time. The dynamic element is explained with reference to a process of marketization of labour. This latter process encouraged attempts to break out of poverty and raise returns to labour by reallocating work between places, occupations and institutions. The roots of these ideas go back to 1999, when I began writing a textbook on the economic history of India. I embarked on the task with a broadly sceptical attitude towards the then-dominant political narratives of Indian economic history. I made a somewhat clearer statement of the issues in an article published in the Journal of Economic Perspectives in 2002. The article forms the raw material for Chapter 2. The editors and readers of the journal who handled the article contributed greatly to sharpening its style and content. A more recent opportunity to polish these ideas presented itself when I taught, for the second time, a course in economic history at the Gokhale Institute of Politics and Economics. I was now armed with a typescript and a captive audience on which to practise. The graduate students who took the course probably regretted the decision; at any rate, the exercise produced a second attempt at summing-up that appeared in the Economic and Political Weekly. Chapters in this book, or fragments thereof, were presented in seminars at the University of Iowa, Gokhale Institute of Politics and Economics, Osaka University, the International Congress of Historical Sciences at Oslo, Labour History Association of India, Overseas Development Group of the University of East Anglia and the International Congress of Economic History at Buenos Aires. I wish to thank my hosts in these institutions, and colleagues and co-panelists who took part in the exchange of ideas on these occasions. In particular, I am grateful to Paul Greenough, Santhi Hejeebu, Jiürgen Kocka, Prabhu Mohapatra, Nitya Rao, Kunal Sen, Ajit Sinha and Kaoru

Sugihara. André Gunder Frank, Ramachandra Guha, Douglas Haynes, G.N.Rao, Suresh Tendulkar and Michael Twomey read specific chapters or the whole typescript and made extensive and painstaking comments. Their response to the work encouraged me, but also made its limits more apparent. Discussions with Alice Thorner on the history of women’s work were illuminating. Her gracious invitation to me to deliver the Daniel Thorner memorial lecture in 2000 was in a way the genesis of Chapter 5. Uma Kelekar provided valuable support as research assistant. Finally, I am grateful to Sudakshina Roy and Mrinmoyee Roy for forgiving me, most of the time, for returning late from the office on the excuse of writing this book. They make this labour worthwhile. Tirthankar Roy Pune January 2005

1 Introduction The aim of the book For some time past, the traditionally slow-moving South Asia has engaged the attention of development economists for having achieved appreciable economic growth. Along with respectable rates of growth, the region has seen unprecedented globalization with respect to consumption, capital, knowledge and information. The economist’s response to these events and the rapidity of the change is usually positive, even celebratory. In the process, those livelihoods are often overlooked in which the journey from poverty to well-being, from tradition to modernity, has been agonizingly slow. For example, manual labour, which provides living to millions of individuals in agriculture and manufacturing, has not seen significant growth in average earnings. Rural poverty remains acute and extensive. The work that women perform in the informal labour market is usually manual and low paying. Possibly as much as a quarter of the workforce, in this way, is engaged in poorly paid work. Individuals with poor endowments of capital or knowledge crowd these markets. Many amongst these individuals are women, and many are to be found in rural India. Not only has this sphere not shrunk, through much of the twentieth century it expanded. It expanded during a time-span that saw a veritable revolution in development strategy—from market-led in colonial India, to state-led during a long socialist interlude, and back to the market at the end of the twentieth century. So tenacious and so resistant to the devices of experts and ideologues, this sphere has not been passive after all. On the one hand, manual labour and poverty remained impervious to a whole basket of interventions because there were forces working towards intensification. These forces were too strong for the state or the civil society to master. But, on the other hand, the subjects themselves adapted to these forces by means of a variety of strategies, sometimes political, usually market-mediated. A coherent narrative of this complex process must be a historical narrative. If we cannot answer the question why poverty persists in South Asia without investigating conditions of labour, then we cannot investigate conditions of labour without going back to the past. Over much of the twentieth century, the more things changed in terms of new income opportunities, the more they remained the same in terms of the structure of employment. Manual, unskilled, unorganized, insecure, informal and earning a wage that changes slowly and is not indexed—these characteristics seem to define the rural labourer, the worker in small-scale industry, and the majority of women workers, in 2000 almost as well as in 1880. The factors that shaped India’s occupational structure were at work over that long a period. But even as the standards of living of ordinary workers changed little in the long run, labour institutions changed. If we take a sufficiently long view, they changed

Rethinking economic change in India

2

dramatically. The proportion of wage-labourers in the workforce increased through much of the twentieth century. Landlessness increased. The proportion of women among rural labourers increased, and that among manufacturing workers fell. The family as a workunit weakened both in agriculture and in industry. From the early twentieth century, a traditional bond between agricultural labour and a range of rural services began to break up. Likewise, the customary forms of attached labour and long-term contracts began to weaken. Occupational choice became more market-driven. Spot markets for labour came to rule the countryside by the third quarter of the twentieth century. In short, through the long run, even as manual, unskilled and poorly paid labour dominated the working situation in India, the work-sites, work arrangements and modes of labour allocation by families changed. These movements were not always a matter of choice. But sometimes they did have the character of deliberate attempts to resist marginalization. Why does manual labour persist in India? Why have wages, and therefore average standards of living, changed slowly? Why does poverty persist? What factors account for changes in employment conditions? What are these changes? This book contains a set of chapters that try to answer these questions and offer a coherent story on labour. The chapters are held together by the argument that whereas the balance between land and labour explains to a large extent the persistence of poverty, a growing market for wage labour enabled attempts to escape resource constraints by means of reallocation of work between places, occupations and institutions. Within the world of the rural labourer, there were cases of gainful shift even as the aggregate resource conditions turned adverse. Among other groups, the artisans of interwar India represent a significant example of gainful reallocation. Cultural factors shaped access to new opportunities, however. This is most clearly seen in the difference between the experiences of men and women. This book also argues that Indian historiography, either of economic change or of labour, does not answer these questions adequately. And it investigates the reasons behind this deficiency. I use the term ‘labour’ in this book to refer to the worker, to labour time and, perhaps most frequently, to occupations intensive in manual labour, according to context. Each of these senses is quite broad. However, the groups of people who figure in this story are more well-defined. Three overlapping groups dominate this narrative: rural wage labourers, artisans and women workers in agriculture or industry. When trying to describe the big picture, the reference period is approximately a century marked on one side by the 1870s when significant agrarian expansion began in the region, and on the other side the 1970s when a second agricultural revolution took shape. British Crown rule in India (1858–1947) overlapped with the period of study to a large extent, so that colonialism enters this narrative frequently. However, when dealing with parts of the detailed picture, the reference period varies and is specified according to the context. I use the terms ‘South Asia’ and ‘India’ inter-changeably if the reference is to a time prior to 1947. But these terms mean different regional units if the reference is post-1947. Few sources of occupational statistics now extensively used in the region have a long memory. The key stylized facts of the book are gathered from a source that has been neglected, even discredited, in recent economic and historical scholarship—the occupational statistics of the Indian censuses. Started in 1881, the decennial censuses yield data that a historian of occupations cannot afford to miss. But problems of definitional changes between censuses have led to their remaining underused, perhaps

Introduction

3

increasingly so. The problems are not trivial. But having had a random and transient character, these problems are not good arguments for dismissing observed patterns of change that lasted a century or nearly that long. Furthermore, in some cases, big stylized facts drawn from the censuses can be partially cross-checked with other datasets. The case for a return to the census, and making use of its long recall, is a sound one. The rest of this chapter is divided into three sections. The first takes up the principal stylized facts and explanations thereof, that is, outlines the ‘big picture’ of economic change and what role manual labourers played in it. The second section presents a selective overview of existing historiography. And the third contains a chapter summary. Since some of the chapters were originally written as independent essays, and yet all the chapters are driven by a single research agenda, by linking them, the third section performs more than a cosmetic role.

The narrative Why poverty continued From as early as the beginning of the nineteenth century, conditions of work in India began to change, possibly at a more rapid pace and on a larger scale than at any time before. Conditions of work changed under the influence of two largely exogenous forces: globalization and institutional change. How did these two forces reshape levels of living and modes of working? In reviewing the effects of globalization on labour, it is useful to distinguish between three periods of time: 1820–70, 1870–1930 and 1930–70. Integration of India in the expanding world economy had mixed effects on the labourintensive occupations: decline of traditional industry on the one hand and expansion of agriculture on the other. India experienced the Industrial Revolution at first in the form of a long decline in the average price of manufactures. In principle, this should depress demand for labour in industry, increase supply of labour to agriculture, and depress wages in both sectors.1 The few studies that exist on real wages of agricultural workers in the period 1820–70, the peak period of ‘de-industrialization’, have not found a discernible trend (Chapter 4). Nevertheless, relative returns may have changed. If real wages did remain roughly stable in the middle two quarters of the nineteenth century, the wage-rental ratio began to fall. For, real rates of land rent (and land prices which maintained a roughly stable relationship with rent) increased in most regions of India throughout the colonial period.2 Such a trend, if it occurred, would be consistent with the fact that India was a relatively land-rich and capital-scarce economy. On the other hand, the stable trend in real wages, despite growth in land-intensive exports, probably derived from the limited penetration of the labour market in rural areas in this period. From the middle of the nineteenth century, the Industrial Revolution began to exercise more positive effects. Demand for agricultural exports from India increased substantially, and new sources of demand for labour, such as mills, plantations, public works and emigration overseas, had emerged. According to the best estimates available, agricultural income grew at an average rate of 1.1 per cent per year between 1865 and 1910, which was close to the average that contemporary industrializing economies experienced.3 To meet the agrarian expansion, either more land was necessary or more investments were

Rethinking economic change in India

4

necessary to raise the yield of land. Yield per acre changed little in India as far as we can gather (Blyn, 1966, the dataset refers to 1891–1946). But idle land there was in the late nineteenth century. The reduction in demand for labour and capital in some segments of the economy, and the growth in demand in some others, meant that capital and labour needed to become more mobile, which they could in the late nineteenth century because of the growth of railways and communication network. The infrastructure revolution reduced the costs of reallocation of inputs in this period. Did the negative effects outweigh the positive ones and thus trigger impoverishment? Although some datasets suggest an absolute wage depression in rural India between 1870 and 1910, the evidence is unreliable (Chapter 4). Employment intensity of work surely increased in this time-span due to expansion in land area. More reliable data show that between 1900 and 1927–28, real wages of non-agricultural workers increased substantially (Sivasubramonian, 2000:264–78), and so did agricultural wages (Chapter 4). At least for a substantial part of the period 1870–1930, the net effect was positive for labour. Significantly, the major difference between the mid-nineteenth century and the late-1920s was in the extent of marketization of labour. The choices before the casual worker in the rural areas were considerably greater towards the end of the period than at its beginning, which may have rendered wages increasingly more flexible and less custom-bound. Wage-rental ratio does not appear to have changed much in the first quarter of the twentieth century. These trends would imply two things: (a) a substantial increase in demand for labour in agriculture, and (b) a slowdown in the decline in demand for labour in industry. It is even possible that the first caused the second, at least partly. Consolidation of property rights in land through a series of ‘settlements’ was the other major exogenous factor shaping rural economic change in colonial India. In pre-colonial India, rights in relation to land had at least four distinct layers: rights to own, rights to control or collect taxes, rights to cultivate and rights to the commons. Of these four layers, British reforms recognized and greatly privileged one layer, proprietary rights, while suppressing, or at best leaving poorly specified, the others. The reforms, in other words, potentially increased the insecurity of tenure for many while at the same time making some rights to assets secure, saleable and creditworthy. It encouraged participation of peasants in markets for land and credit, and increased the vulnerability of many to market risks. More exposure to the world market or the local (largely informal and imperfect) capital market increased aggregate risks. Did these risks intensify poverty in India? In perhaps the mainstream view of history, shared by Indian historians as well as some members of the colonial bureaucracy, the risks of credit default were indeed unsustainable and led to increasing land transfers from the poor peasants to the rich (see Chapter 3 for discussion and references). Increased asset inequality should lead to growing landlessness, which did occur in colonial India. But did increasing asset inequality also depress wages and thus lead to growing poverty? Although the new owners of land would still need the same quantity of labour, in reality, their monopolistic control of consumption credit might lead to more adverse terms for the new wage workers. Yet, the significance of institutional change and credit-based exploitation is perhaps overstated and misunderstood in historical scholarship. Land transfers occurred at both ends of the distribution, and not only from the poor to the rich. Legal reforms, for example, de-recognized rights to control land, which affected some

Introduction

5

members of the pre-colonial landed elites adversely. Further, while landlessness did increase, it increased long after commercial expansion began and, curiously, affected women more than men. The beginning of the rise in landlessness coincided with the exhaustion of land supply and demographic transition in India, approximately in the decade 1921–31. Population growth rates accelerated and continued to accelerate until 1971–81. From the 1920s until the 1970s, real wages of agricultural workers changed rather little, and wage-rental ratio for both urban unskilled and rural labourers declined.4 Late in the interwar period, however, formal and informal sector wages began to diverge somewhat because of greater organization of the urban mill workers and administrative workers. Their power increased substantially because of the convergence of the trade union movement with the nationalist movement. These trends are revealed in the significant increase in real wages of mill workers in these decades. The divergent trends continued after independence from colonial rule in 1947, when the formal sector workers were granted numerous legal privileges. Diminishing returns in agriculture was clearly the turning point in the genesis of Indian poverty. And yet, what is perhaps remarkable about Indian agriculture in the last subperiod, 1930–70, is not that it witnessed a fall in land-labour ratio, but that the change in wage-rental ratio did not induce technical change in agriculture, in the manner in which it did in parts of contemporary East Asia. Growth in yield per acre and yield per person in the interwar period was practically absent, even negative in regions like Bengal that experienced the most precipitous fall in land-labour ratio. The Malthusian scenario described above is perhaps interesting only so far as it enables asking why it proved so hard for Indian peasants to overcome it. Why Malthusian land-constrained growth was not replaced, if not by capital accumulation, by land-intensification, or by induced ‘landsaving’ innovation, is perhaps the key question in modern Indian economic history.5 In standard economic history scholarship on India, the question has been answered with reference to political economy (see ‘the retardation thesis’ below, and Chapter 2). Landowners and cultivators represented distinct classes, when lands were owned by rentearners or rural creditors. In Bengal, where the distinction between owners and peasants was institutionalized in the zamindari settlement, this might have been the reason why investments in land productivity did not occur. The risk-free income from rent and interest provided adequate living to the landowners, and made them lose interest in and, under certain arrangements like share-cropping, even hostile to land improvements. While there is an element of truth in this approach, it cannot evidently be generalized to all regions of India. Total output growth varied between regions, but the stagnation in yield affected nearly all regions more or less. The vast literature on induced innovation in agriculture suggests several hypotheses why technical change might be constrained in certain societies. Movements between one ‘innovation possibility curve’ and another, representing a more efficient method of production but one more biased towards the input that became more abundant in the meantime, can be inordinately slow, or never happen. For, such movements require institutional change, cultural change, conditions that facilitate collective action, research and development expenditure, public investment when switching to the new technique requires public goods such as common irrigation or technical education, and efficient capital markets (and perhaps intellectual property rights protection) when the new

Rethinking economic change in India

6

technique depends on private investment. We do not know which of these scenarios were particularly applicable to interwar India. It does appear, however, that the literature underrates the environmental barriers to technical change. Within this region, the incidence of poverty has had a strong apparent association with arid climates and the absence of man-made irrigation. Based on this association, my preferred version of barrier to technical change in Indian agriculture would give more-than-usual weight to climate and the physical resource endowment, though not in exclusion of the other potential barriers such as culture blocks or public spending. The greater part of the Indian subcontinent combines three months of monsoon rain with extreme aridity in the rest of the year, which dries up much of the surface water. The rains make growing one crop relatively easy, but growing another crop dependent on irrigation that requires, precisely because of the great scarcity of surface water in the dry seasons, expensive systems of harvesting, storage or relocation. These systems typically try to address two tasks simultaneously—seasonal and spatial transfer of water—because season affects space so much. In other words, the monsoon in a tropical region made earning subsistence rather easy, but making improvements in living standards exponentially difficult. Further, the dependence on a natural supply of water increased the risk of crop failure, and single-cropping reduced security against the failure of the monsoon crop. El Niño-type disturbances of the tropical oceanic atmospheric systems could aggravate matters. High risk of famine and crop failure was part of the cycle of agricultural production. This twofold scenario, relatively high capital cost of irrigation and uncertain returns, would have constrained investment in agriculture even in the absence of capital market imperfection, absence of collective action or dysfunctional cultural traits. Public investment in irrigation eased the barrier, but only marginally. The diminishing returns crisis in rural India broke out from the time of the Great Depression. It was too powerful a phenomenon to be offset by the better performance of industry in the interwar period. The preceding decade had seen an active land market, growing exports and a great deal of conspicuous consumption stimulated by easy credit. Facing a large deflation, these debts and rise in real wages squeezed the marginal peasantry hard. There was recovery after the mid-1930s, but the good times did not return. Economists and administrators now began to take serious note of the dynamic by which small farms were progressively subdivided because of population growth, had to support the same or a larger workforce and were losing productive power because of scarcity of water or overexploitation of land. Landlessness was rising far more rapidly than before. The downslide revealed itself in the stagnation of real wages that set in from early in the second quarter of the twentieth century. The ability of globalization to engender growth seemed to have come to a halt in rural India. And it remained impaired for nearly four whole decades through which agricultural real wages changed little. The inertia ended when a biological revolution reached Indian agriculture in the late-1960s, and was offered together with a whole package of input subsidies. Average real wages responded to productivity, albeit with a considerable lag.6 In the interwar period, the diminishing returns scenario affected only agriculture and was largely absent from manufacturing. This is perhaps not surprising for modern industry which could accumulate capital relatively easily. The small but robust mill industry in textiles could overcome its disadvantage in capital because it used natural

Introduction

7

resources abundantly available in India, used easily accessible technology and technical knowledge from Britain, and received capital and enterprise from groups prominent in foreign trade. More surprisingly, however, diminishing returns was absent even in traditional industry, which suffered a much greater disadvantage in accessing capital, and had earlier declined because of competition from imports. Chapter 6 argues that the productivity growth in traditional industry was made possible not by capital accumulation, but by a combination of a specific consumption pattern, and more efficient deployment of labour. Why modes of working changed If we go underneath these broad generalizations and look at work-sites and workarrangements, communities and families, the colonial period appears as an era of turmoil. Economic growth created numerous local shortages of labour. The very prospect of diminishing returns engendered ever more energetic attempts to look for new opportunities and new resources. Some of these movements had nineteenth-century antecedents and external causes, such as migration overseas. Others took shape later, such as internal migration of peasants and rural labourers, or a change in the organization of work in small-scale industry. These movements split up the history of male workers from that of female workers, which had been largely integrated earlier, for migration had a pronounced male bias. To a more limited extent, there was migration and reemployment of capital as well. Peasants sometimes moved to new territories with abundant land, as in the canal colonies of late-nineteenth century Punjab, and capitalist artisans relocated business near new industrial towns (Chapter 6). Regional agrarian studies have explored some of these movements, but usually piecemeal. Agrarian history has not yet produced a consistent story that can accommodate the central tendencies as well as the variations around these. The central piece in the story is ‘proletarianization’, or growth of wage labour in rural and urban India. We see this happening in the rising proportion of wage-labour in the rural workforce (Figure 1.1). We observe this in the steady dissolution of household industry (Figure 1.1), that is, in a weakening and transformation of the family as a workunit. In agriculture too, the average family weakened, as wage-labour was often accompanied by the migration of males. Wage-labour increased in extent because of a combination of factors. In agriculture, competition for land and titles to land intensified because of population pressure, leading to widespread abandonment of owner-cultivation and tenancy for labour, towards one end of the land distribution. In small-scale industry, a competition between the household and the wage-workshop led to the decline of the former. Commercialization increased the importance of capital, information, modern transport and communication systems and technology, and perhaps ‘networks’ of various kinds, as factors of production. The urban factory was better situated than the rural household in accessing these inputs. There was another level of proletarianization and another mechanism for the growth of wage labour. The average duration of employment contracts reduced in rural India, increasing labour turnover and casual hiring.

Rethinking economic change in India

8

Figure 1.1 Wage labour in India, 1901–2001 (sources for all figures are the Indian censuses, occupational statistics. In all cases, figures for 1901–31 refer to British India and the States, and figures for 1961–91 to the Indian Union. 1941 did not have a census. 1951 occupational data are not reliable. Both years are excluded from the graph.) ‘Farm servant’ was the census term for long-term contracts. The percentage of farm servants in agricultural labour households was in decline in the first half of the twentieth century. In Madras, where farm servant contracts were particularly prevalent, the fall in percentage was sharp. Migration played a role in reducing the employers’ need for ‘labour-hoarding’ in traditional agriculture (see Chapter 3 for more discussion). Since annual contracts involved predominantly males, its decline facilitated entry of women in agrarian labour markets. Mediated by a new type of market relationship based on credit, a different form of long-term contracts spread in rural areas at the same time. But debtbondage too, which was always spatially restricted, began to weaken in the last quarter of the twentieth century. Along with farm servants, another older form of ‘attached’ labour was the commitment to supply diverse services for the village, to which some rural labour castes were subject. The term ‘general labour’ was used in the early censuses (given up in the later ones) to refer to such people, who performed not just field labour, but across-theboard labour services for the village, usually on terms that were not negotiated often. There was a powerful hierarchical element in these services, in that the labourers had few choices regarding what tasks they would prefer to do or what payments they could ask for. This category of work tended to disappear from the late nineteenth century. Rural

Introduction

9

labourers asserted their freedom to choose the employer, and the right to receive a wage for any service supplied. There were many reports of conflict within the village over obligatory labour services and about choosing occupations (Chapter 3). These acts of resistance merged into the anti-caste and social emancipation movements unfolding at the same time. The break-up of these customary contracts had the effect of pushing individuals out of a portfolio of services and to either specialize or shift to a different portfolio. Earlier, these portfolios often included some handicrafts and some labour. In the case of industrial production, the market was usually a local one. As these portfolios broke up, artisan-cumlabourers tended to give up production and move towards labour. The market economy weakened the bond between local production and local consumption, encouraged longdistance trade, and made capital and information key to success in trade. These requirements weakened the rural artisan-cum-labourer and favoured the more resourceful artisan castes. For, the former had poorer access to capital markets relative to the more specialized producer groups. But even as the rural artisan-cum-labourer left production and moved into labour, the types of labour they supplied changed in composition. A traditional form of occupational mix dissolved into a more market-driven occupational one. These great transfers of population between occupations could not have taken place without equally large transfers between places. Large streams of migration began from the mid-nineteenth century, aided by the railways and the telegraph after 1850, to facilitate this huge reallocation of labour. Men increasingly had to leave home for careers elsewhere. And as competition for land intensified, those with inferior rights to land had to do so more often than before. They went to the mills, the mining towns, urban smallscale industries, overseas, public works, plantations, urban services or the railways. These shifts, which I call ‘reallocation’, were not just transfers of population between locales and jobs, but involved transfers between labour institutions. When rural artisan-cumlabourers left the village for work, they left almost always to take part in wage labour. Therefore, migration gave a significant push towards the break-up of customary terms of employment. Reallocation often required more than just a wage incentive. Transaction costs were important too. Much of the movement of labour into the modern sectors relied on ‘contractors’ and foremen (better-known as sardars in the context of recruitment into the cotton and jute mills) who could communicate with both the workers and the employers, and frequently took advantage of information asymmetry on both sides. Despite the unreliable intermediary, employers in the towns often found it hard to gather a large number of hands without some help from the former. Labour markets with sufficient depth to supply mill-hands on a large scale did not exist in the mid-nineteenth century, except briefly during famines.7 After 1900, voluntary internal migration increased. Eventually several types of transaction costs involved in hiring labour fell. A range of skills became available for hire in one place. Potential workers came to the millgates rather than wait for the sardar to take them there. The power of the intermediaries, consequently, declined. The contractor in later years was not so ubiquitous any more, and a great deal of hiring could take place via spot markets. Mill managers in the interwar years reported that they hired their daily hands from the ‘mill gates’ breaking with past

Rethinking economic change in India

10

practice, whereas it was only the new factories and new divisions that still relied on the agent.8 By contrast with the mills, how labour was reallocated in agriculture and traditional industry is much less known. There had been large shifts within and out of agriculture in the nineteenth century. Once again, these were often accompanied by a prominent role for contractors. However, little research exists on the intermediary outside the formal sector of the mines and the mills. Studies on the artisans suggest that migration was an important part of their history too in the early twentieth century. Skilled artisans, however, did not necessarily migrate to join labour markets. Unlike the rural artisan-cumlabourer, the skilled artisan often migrated to resettle elsewhere as an entrepreneur. These processes had a gender bias. In the ensuing market for manufacturing labour, it was often harder for women to take part than it was for men. However, in the spot market for agricultural labour, women did take part in an increasing extent. We cannot say whether or not these two tendencies always occurred together. To some extent they did. Earlier, some households had combined cultivation and industry. The men in these families took lands on rent or owned plots that they cultivated along with the women. Sometimes, these households also performed industrial work on the side. The nineteenthcentury censuses recorded women thus engaged as either ‘cultivator’ or ‘manufacturer’ depending on how the families saw themselves. But when the men left home in search of work elsewhere, the whole family’s engagement with industry on one side and with owncultivation on the other declined. The family as a work-unit invariably weakened when the men went from home. For, when the men left home, the women who remained behind rarely could continue or take up production work on their own, because women workers in industry or agriculture, who often commanded high degree of skills, had poor access to capital and marketing. Relative to men, women took part far less in contracting and in the wider and complex interactions that it required. Internal migration in South Asia involved a disproportionate number of men, a feature spanning colonial and post-colonial periods. This situation pushed a disproportionate number of rural women into agricultural wagelabour. In early twentieth-century rural India, that meant casual labour in agriculture. Also as mentioned before, the end of long-term contracts facilitated greater absorption of women in farm labour. Both cultivation and handicrafts experienced relocation of workers. But there was a difference in the result. Whereas average income in the crafts grew in the twentieth century, it stagnated in agriculture. New research has shown the conditions that made expansion in the crafts possible, despite often unbridgeable cost differences between the crafts and mechanized production (Chapter 6). This research has stressed the role of consumption and culture in sustaining traditional labour-intensive industry, the strengths of such industry deriving from scope for product-differentiation or design innovation and a variety of adjustments that made incremental productivity gains possible in this sector. The term ‘industriousness’, to mean a more efficient deployment of labour resources within labour-intensive industry, is a useful description of what happened in the crafts. Induced accumulation of human capital via education and training was conspicuously weak in this dynamics of intensified labour, both before and after 1947. Students enrolled in schools of all levels as the proportion of population of school-going age increased from 3–4 per cent in 1891 to 7–9 per cent in 1931. These conditions changed only slowly after 1947. A quarter century later, the proportion of scholars in the relevant age group was

Introduction

11

still low at 20–25 per cent. Only about a fifth of the students who started school in colonial India reached secondary levels. The British government in India had neither the money nor any serious intent to supply universal education and health care in effective quantity. The numerous Indian-ruled states, with only a few exceptions, echoed that indifference. Private funding was slow to shed the biases that had long confined education to only a few groups and to the men within these groups. These biases point to one final ingredient in the story, namely, cultural and social filters through which labour reallocation progressed. Culture enters this account via two principal routes. The first relates to an element of intellectual inertia that could sustain a very inadequate level of social effort in education and training for the poor. And the second relates to the conditions that made women vulnerable players in the labour market. Both these themes await further development than that contained in this book, though the second receives a somewhat detailed treatment in Chapter 7. Both illustrate the idea now all too familiar to social scientists that modernization processes have exclusions built into them. Chronology This study spans approximately a century between the 1870s and the 1970s. The key processes that form the subject-matter here took shape in this time-span. These are, increasing land-scarcity and diminishing returns, marketization of labour, and reallocation and industriousness that overcame resource-barriers, if temporarily. But these three processes did not overlap in time. The tendency of conversion of family or attached labour into casual wage labour was weak and localized until late in the interwar period, picked up pace from then on and quickened after independence. Roughly the same characterization of rate of change can apply to the exit of women from manufacturing, decline in household industry and casualization in the urban labour market. Indeed, a number of factors cascaded by midtwentieth century to create conditions for a decisive transition in modes of employment of manual labour. Examples are opportunities of migration, growth of local labour markets, and scarcity and competition for land. In the late nineteenth and early twentieth century, several of these trends had begun already. But for a long time afterward, they were only loosely connected, almost disparate. One connecting thread through them was commercialization of the product market (especially between 1870 and 1924) that created regions of rapid agricultural growth and spheres of excess demand for labour. The limited upheaval that this caused in the rural labour markets had at least the effect of loosening a variety of customary labourtying arrangements. Partial decline of cotton spinning and weaving between 1820 and 1880 in competition with British textiles also pushed groups of artisans into labour, or quasi-artisans to specialize in labour. Between 1870 and 1920, the railway network grew, and facilitated long-distance migration of labourers. There followed a many-sided process of inter-sectoral and spatial movements of manual labour. Until about the beginning of the 1920s these large movements were partly motivated by wage differentials between the traditional sector and the modern sectors such as the railways, plantations, mines, mills, public works, pockets of agrarian growth and overseas. Rural wages were consistently below, if only a little below, wages in modern sectors. However,

Rethinking economic change in India

12

from late in the interwar period, conditions began to change. Spheres of excess demand in agriculture were fewer. The modern sectors were not doing well, or not growing rapidly. And the rate of labour supply accelerated with population growth, the effect only slightly modified by a marginal fall in participation rates. The year 1931 marked the transition point. The good times for the peasantry were over, and a rural crisis broke out. The subsequent forty years saw a strengthening of the four conditions that made the 1930s such a setback—exhaustion of opportunities in agriculture, weakening of market incentives, relatively small growth of non-agriculture and demographic transition. All this entails a rather different narrative from what the mainstream historiography in South Asia has usually described in relation to either labour or economic history. This book, thus, is necessarily an argument in historiography.

Historiography A general economic history of manual labour should possess two attributes. First, since manual labour straddled different worlds, the history should be able to connect experiences across gender, sector, region and regime. Second, since conditions of labour and conditions of production were closely tied, it should be able to connect the labour experience with long-run economic change. In the three sections that follow, I offer an assessment of labour scholarship against these two benchmarks. The retardation thesis The bond between economic history and labour history was strong in India until about 1980. Until then, more or less the only general story available on economic change, a post-colonial blend of Marxism and economic nationalism, saw nineteenth-century globalization as a phenomenon that had induced distortions. Power—colonial power, class power or a compound—played an instrumental role in this process. Markets in the colonial condition were manipulated by those in control of political power. A specific configuration of power and the unequal exchange relations that these entailed pushed the peasants and artisans of India towards poverty during British colonial rule. Commercialization bred inequality rather than growth. The increasing extent of wagelabour and casual labour illustrated this general pattern of economic retardation (Patel, 1952; Patnaik, 1983). In the 1970s many economists and historians were interested in proletarianization as a phenomenon illustrative of a transition in the ‘mode of production’ in Indian agriculture. The historical understanding displayed in this neo-Marxist debate was broadly consistent with the thesis that colonialism created rural labour and poverty. I refer to this entire perspective as the ‘retardation thesis’. The principal sign of retardation was ruralization of economic activity. While Britain herself experienced industrialization and urbanization, potentials to industrialize were destroyed in India. Unemployment and poverty in rural areas deepened. And potentially beneficial economic changes such as commercialization strengthened inequality. The ‘unholy trinity’, to borrow a term from Surendra Patel, was a contributor to the retardation story and consisted of ‘subjugation’, ‘ruralization’ and ‘retardation’ (Patel,

Introduction

13

1992:57–8). India’s colonial status induced a ‘forced’ commercialization on the peasantry, and ensured a speedy destruction of the handicrafts. Indeed, according to Bipan Chandra (1966:61), Indian industry was ‘so readily, so easily destroyed because the British had used their political control’. In Jawaharlal Nehru’s words, ‘the burden on the land grew, with it, unemployment and poverty’. And India failed to industrialize, despite ‘all the ingredients and conditions of industrialization [being] present’ (cited by Joshi, 1967:453). In short, rural labour and rural poverty represented a process of politically induced economic retardation in a region otherwise ripe for industrialization and economic growth. These ideas originated in the writings of the Indian nationalists D. Naoroji, G.V.Joshi, R.C.Dutt, G.Subramaniya Aiyar, and others, c. 1900.9 They believed that India was becoming poorer in their time, and poverty was a creation of British colonial rule. In the interwar period, Nehru restated these ideas in his writings on Indian history, and R.P.Dutt, Secretary of the Communist Party of Great Britain, recast these ideas in a Marxist mould. From the 1970s, the retardation discourse also drew on a more global source: theories of underdevelopment. The argument was that economic policies and processes that created growth in one part of the world created underdevelopment in another (Frank, 1975; Bagchi, 1982). Nehru and his predecessors had argued the same point in almost the same language. Economic historians working in the USSR, who took a keen interest in India and were influential on a section of the Indian academia, took a very similar line, though employing rather different historiographic tools. In the 1980s, the paradigm of economic change in modern South Asia, now also inspired by theories of imperialism and underdevelopment, was essentially the nationalist story with minor changes.10 Apart from these intellectual roots, the retardation thesis drew strength from a deep symbiosis between history and policy making. The belief in politically induced market failure served as a strong link between the past and the present in the autarkic and dirigiste policy regime in force in India between 1950 and 1985. Economists and historians both invested ‘market failure’ with foundational status in their respective analytical systems. Practitioners in both fields believed that a broad identity of interest between the colonial and the local economic elite had been responsible for market failure in the past. Economists and historians agreed that markets and the open economy were instruments that needed to be restrained, if used at all. Economic history of modern India became a discourse of impairment, inflicted in the past by colonial policies of free trade and a minimalist state, and redressed in the present by a strategy that was emphatically isolationist and interventionist. Historians thereby gave meaning to a socialist policy regime that intervened heavily to restrain market forces and international relations. The retardation thesis was questioned from time to time. But before a serious reexamination could begin, an ideological upheaval made the model irrelevant. Economic nationalism began to lose out to the appeal of globalization in the ex-colonies. From the late 1980s, economists began to change their minds about the universality of market failure and the virtues of an interventionist state. With the obsolescence of political economy in the new regime, economic history, which remained committed to market failure, faced obsolescence. The premise that retardation and market failure were embedded in globalization could not inspire economists any more. Since the end of the socialist experiment and the return of economic liberalism in the academia, the pro-

Rethinking economic change in India

14

market sentiments of economics and the anti-market sentiments of orthodox economic history have made the two fields and their languages drift apart. To end this impasse and bring history back into the discourse on economics, to make the past regain substance and signiflcance for discussions on the present, we need a different historiography. To succeed, the project needs to unlink stylized facts in Indian economic history from specific configurations of power, and re-link these with themes on which economics and history can share a dialogue without necessarily sharing an ideology. Two such points of reference for this book are resource endowments and the agency of the market.11 Seen as a story of growth, my criticism of the retardation thesis is not only that it was irrelevant, but also that it was inadequate. The treatment of commercial expansion during globalization is a case in point. Indeed, many new types of market inside India linked to exports were imperfect. Nearly all were unregulated. In many instances the social ‘embeddedness’ of markets reinforced existing inequities rather than weakening these. And yet, stories of bounded, constricted, inequality-breeding markets cannot represent the average experience with market participation in colonial India. Agricultural growth rates recorded in the late nineteenth century and a limited industrialization bear witness to the positive potentials of the economic globalization process into which India was drawn, partly because of its status as a British colony. ‘Economic growth from below was possible’ (Tomlinson, 1995:173) and it was made possible by the fact that commercialization encouraged exploitation of unutilized resources such as waste land, search for new resources, better allocation of resources and industriousness. Seen as a labour history, while the retardation thesis gave rise to a poor-getting-poorer story, it was not efficient at fitting together the fragments that made up the overall labour experience. One of the prerequisites for a larger story on labour is equal attention to agriculture and industry, men and women, family work and wage work, for these worlds were necessarily connected in the decisions of many working poor. Such an integrated story remained missing. The heavy accent on politics made the economic history too onedimensional. Wage labour tended to be studied in isolation from forms of contract, occupational mix, earnings and gender. The role of the labour market in effecting reallocation was over-looked. Accent on particular modes of power such as colonialism did not live easily with the persistence of labour market trends into the post-colonial times. Subaltern studies and the new labour history The Marxist-nationalist narratives on labour went out of fashion in the early 1980s. Thereafter for some time the historiography of the poor in colonial South Asia was dominated by the group of historians known as ‘subaltern studies’. The subaltern studies were partly reacting to the Marxist economic history that had subjected the experience of the peasants and workers to the imperatives of class formation and partly trying to rewrite the history of the Indian nationalist movement (Chapter 9). Their message was that the poor could bring about change by resisting colonial power. That message turned out to be influential, but it made the discourse on agency overly political, insofar as it ignored market-mediated forms of agency.

Introduction

15

Subaltern studies transformed themselves in the 1990s to a form that was distant from both economics and the rural poor. Under the influence of French post-structuralism, ‘power’ detached itself from property relations and made knowledge its new habitat (see Chapter 9). An entire generation of historians repositioned their world-views from political to cultural domination, retaining through this change a belief in the centrality of power. The new faith was instinctively historical. But economics had no clearly defined role in it. Economic questions troubled it, and were shut out of it. Still, the early subalterns rekindled interest in labour and research flourished.12 And in a short time, labour history emerged as a well-organized and energetic enterprise, with a Labour History Association of India based in Amsterdam and Delhi. Partly in reaction to the submergence of economic and political processes into the cultural and partly in a need to incorporate cultural processes into mainstream labour research, the new labour history moved in a number of directions. Its interest in the constitution of the ‘working class’ connected it to Marxist historiography. Colonialism and popular resistance continued to be a key thematic, connecting it to the early subalterns. But workers’ consciousness was now seen in terms that were cultural, with explicit attention paid to ethnicity and everyday forms of resistance, connecting the new trend to post-Marxist labour history and the late subalterns. In the process, the new scholarship also contributed to the narrative history of labour against the backdrop of industrialization and nationalism, in a tradition represented by another, older line of scholarship (Morris, 1960, 1965; Newman, 1981). While thus reinventing itself, labour history also acquired weaknesses. The new scholarship neglected rural labour, as if at a loss for words after the subalterns had had their say on the subject. Deeply impressed by the sweep and radicalism of poststructuralism, it passively imbibed the latter’s distrust for ‘grand narratives’. There was, thus, a deliberate retreat from attempts to connect labour history with the economic history of colonial India. The interest in ‘class’ entailed a quest for group identity at some neglect of the larger economic context, indeed at a near-total neglect of the labour market. The emphasis on groups and group politics induced a fixation for work-sites where crowds could be found. Too much attention, thus, fell upon the mill and the mining town. To an equal extent, the new labour history overlooked workers such as the artisans who were neither tied nor concentrated; overlooked work-sites such as the households that did not involve an assemblage of wage-earners; and overlooked groups such as women that straddled several worlds. Sometimes, the invocation of ethnicity had an artificial quality. The ethnicized worker looked like a device to remedy the pitfalls of ‘class’. In some cases, the search for ethnicity introduced a bias for studying groups that were insignificant in terms of their economic presence. The new history also tended to emphasize regional particularities at some cost of regional similarities and connections, that is, at the cost of a pan-regional narrative on labour. These criticisms apply with particular force to the history of manual, semi-skilled, unregulated, informal labour in India, in which sphere barriers along gender, sector and regional lines dissolved away in a fluid, encompassing mix of decisions. These decisions involved choices between domestic and paid work, remaining sedentary and migrating, casual and contractual labour, agriculture and industry, industry and service, and production and labour. Most individuals and families divided their time between several uses of time. The changing mix of these uses merits serious consideration in any history

Rethinking economic change in India

16

of work in the region. Further, the task requires paying attention to the manner in which commercialization of labour encouraged market-participation in some spheres and discouraged participation in some others. Labour history scholarship by and large remained too focused, too politicized and too segmented to handle these connections. The weakest point of new labour history was perhaps its refusal to engage in general economic history. It was a labour history that was sensitive to differentiation within labour, but did not try to locate the average and did not try to explain the persistence of manual labour and poverty cutting across occupations. The retardation thesis did not have this flaw, for it had made colonialism the bridge between labour and economy. It could explain changes in the sphere of work and general economic history by a single analytical device—exploitation. But if the new labour history failed to situate labour in a context of overall economic change, the retardation thesis did so in a manner that failed to convince. Where does that leave us? From the early twentieth century, the economic history of India was the site of a contest between two rival approaches. The academic mainstream in post-independence India was represented by the retardation thesis, which believed in a broad identity of interest between colonialism and the local economic elite, and held that alliance responsible for retardation of the Indian economy in the colonial period. This approach held that the Indian economy on the eve of colonialism had the potentials to industrialize and experience rapid economic growth, but these potentials were destroyed by colonialism. Doing labour history in this perspective would mean charting out how repression engendered poverty. The other and the more marginal school believed that structural weaknesses constrained economic growth in India. These weaknesses of the economic system it traced to society, demography or environment. In part, these weaknesses showed up as resource imbalance and shortages. And, in part, the weaknesses showed up as persistently low levels of efficiency of labour. Doing labour history in this perspective would mean showing how these systemic forces and the labour market unfolded in mutual interaction. The former approach I have discussed earlier in this section. The time has come to take a serious note of the latter. Antecedents Persistence of poverty and unemployment in the South Asian context had long been bound up with the persistence of manual labour. The history of low wage and proletarianization, therefore, was also a history of technological stasis. It was a history of the difficulty of transforming traditional agriculture, traditional accumulation pattern and traditional ways of life. Until the 1940s, economists and administrators did sometimes explain Indian poverty and labour in these very terms, that is, in terms of inadequate capability to adapt rather than as an outcome of colonial politics. The accent fell sometimes on inadequacy of investment funds relative to labour, sometimes on incentives, sometimes on a low-level technological equilibrium. They observed that Indian levels of output-per-worker were far below the standard even in labour-intensive societies. In 1931, average rice yield per acre was 1,400lbs in India and 2,800lbs in Japan. A similar disparity occurred in rice yields between India and Egypt. International disparities in labour productivity in cotton mills were along the same

Introduction

17

direction, and often equally large. In 1930, an average worker in the spinning department of an Indian mill looked after 180 spindles. A similar worker in the USA looked after 1,140, and one in Japan 240 spindles. Not only were these gaps not closing in the interwar period, but possibly even widening. Such massive disparities called for investigation into what appeared to be a systemic inertia to improve efficiency. British administrators often held on, in good or bad faith, to a makeshift Malthusianism when making sense of economic change in India. However, as the economics profession grew, a parallel discussion on this theme emerged among university professors and academics not closely connected with, even mildly dismissive of, economic nationalism. The discourse was insightful, but not a very orderly one. Nor was there a well-developed public sphere in which economists could effectively communicate with the literate middle classes. The nationalists wrote in newspapers and pamphlets that had a wide reach. The administrators gave speeches attended by powerful persons. The professors wrote books in English that patiently waited for readers in the recesses of university libraries. Nevertheless, this discourse on efficiency revolved around a few positions that not only made sense, but also proved to have a long life. Gilbert Slater, professor at the University of Madras, asked what factors accounted for the poverty of the Indian peasants in comparison with their English counterparts. He dismissed the nationalist concern with excessive land tax, a tax that had been fixed in nominal terms and was no more than a small proportion of gross output value in 1918 when he was writing. His explanation revolved on the relatively ‘small number of working days in a year’ in India. The combination of tropical heat and monsoon rains imposed on the greater part of the region an extremely uneven natural supply of moisture and, consequently, enforced idleness on the peasants during one-third of the year or more. Idleness was pervasive not only among peasants and rural labourers, but also among artisans. Indeed, the average number of working days in a year was usually so limited in South Asia that no historical comparison of earnings seems to make sense without knowledge of employment intensity. Climate provided livelihood, but constrained utilization of labour, in this view (Slater, 1918–19:145). Slater did not ask, however, why employment intensity could not rise via more labour-using technical change in agriculture in the presence of enormous quantities of idle and low-wage labour time. Radhakamal Mukherjee, professor of economics at Lucknow University, wrote about land degradation as a result of population growth in some of the more advanced agrarian zones in interwar India. He had statistics on his side. Throughout the interwar period, the land-labour ratio fell and yield-per-acre was depressed. There was an institutional angle and an ecological angle to the story. Population pressure and inheritance practices led to steady subdivision of plots into ever-smaller units, as well as their scattering, so that economies of scale could not be utilized. Intensive cultivation on small and scattered plots required intensive harvesting of water. Pressure on groundwater increased, surface water was often wasted and water simply became costlier to mine. Extension of cultivation frequently took a toll on cultivable ‘wastes’, a part of which was pastures. The pastures were not always good quality land, and their destruction weakened the livestock. In turn, the slide in profitability pushed many smaller farmers to bankruptcy and, eventually, to labour. Mukherjee argued this story in a series of works on his home region, also the most populous region of north India, United Provinces (see essays in Mukherjee, ed., 1939; Chapter 3, this volume).

Rethinking economic change in India

18

The administrative discourse on the peasants often accused them of ‘improvidence’. In good times, the peasant spent lavishly on a daughter’s marriage and on gold jewellery. In bad times, the peasant had to mortgage land at high interest rates. Net gold and silver transactions took a toll of 2 per cent or more from the potential aggregate investment/income ratio. Since little of this gold came to the mortgage market, these savings did not convert into investment. The more astute observers understood that there was an element of insurance behind such ‘improvidence’. Climatic risks were excessive, and affected expected consumption. In the absence of insurance markets, precious metals were the best available stores of value. Because these metals were traded globally, their prices varied in a narrower band than the prices of nearly everything that sold in the villages. On the other hand, as the Secretary of Agriculture Edward Buck wrote in 1884, ‘a great deal of the improvidence with which the Indian cultivator is charged is simply due to the impossibility of being provident. How can he look forward when there is no certain prospect for him?’ (cited in Hall-Matthews, 2002:104). In other words, climatic risks depressed expected incomes from investment projects and, thus, reduced the demand for productive investment. In turn, the situation reduced the potential demand for local labour. The British economist Vera Anstey belonged to a small set of writers who believed that Indian poverty arose from, among other factors, poor education and the consequent poor capacity of the masses to absorb knowledge. Many nationalists too brought into focus the British colonial government’s appalling record on creating ‘human development’. But unlike them, Anstey believed that lack of education was to a significant degree a social failure. Education was not supplied because those having a say on the matter within local societies believed there was no demand for it: The danger spot appears to be women’s education, for which it is often said there is no effective demand’ (Anstey, 1929:471–2). In 1900, for every ten literate men, there was one literate woman. And education was not the only resource to which women had grossly disproportionate access. Such deprivation continued long after independence, which confirms Anstey’s belief that it was Indian society that ensured the exclusion of women from access to good health and education by sanctioning an appallingly low average age of marriage, high fertility and enforced domesticity.13 The theme of exclusion connects with a later discourse on the interaction between economic development and cultural change. There are perhaps two general ways that this interaction can be conceptualized. One of these involves observing how strongly market participants respond to market stimuli, and explains the quality of response in cultural terms. The other would involve observing differential capability of producers to participate in market exchange, and explains the ease of entry and barriers to entry in cultural terms. The former approach is associated with Max Weber, who proposed that the cultural disposition of merchants and artisans in India made them deficient as entrepreneurs. The idea was expounded in scholarship that developed around the ‘modernization theory’ of the 1950s and 1960s, and Indian cases played an important role in this enterprise. Critiques of ‘modernization’ in the 1970s and the 1980s, again using Indian evidence, pointed out significant levels of response of entrepreneurs to profits and risks, thus rejecting the conjecture. Marxist historiography too came down heavily on the notion that the poor were poor because of a deficient mindset.14 Neither criticism, however, was sufficient rejection of the second of the two positions. Culture can matter

Introduction

19

also by excluding groups from participation in the market. The spirit of enterprise is a characteristic of those groups already inside the market. It is market-exchange that reveals the spirit of enterprise. Both Weber and his South Asianist critiques studied groups such as the capitalists, who were insiders, and overlooked groups such as women or lower castes, who were often excluded from market transactions. These exclusions alert us to the prospect that culture is neither a passive nor a benign variable in explaining economic growth and stagnation. To cut the story short, nature on the one hand and society on the other constrained technical possibilities in India, according to these views. Nature was unreliable. And the society had a range of exclusions built within. The actions of the state were relevant, of course. And yet, state failure was evenly matched, even reinforced, by culture blocks. These concerns with structural constraints, though long overshadowed, have proven lasting. In current economic history thinking, the adverse role of population pressure, inheritance practices and diminishing returns have been emphasized again (Kaiwar, 1992, for example). Any discourse on caste in colonial India sensitive to caste as an economic category cannot fail to observe the powerful influence of caste on entry into and exit from occupations. The accent on ‘capability’ stands for a whole approach to development. And there is a powerful gender angle to capability failure and culturally induced exclusion (Nussbaum, 2000; Sen, 1999).15 The preoccupation in these early schools of economic thought with resource endowment and efficiency was reborn in a later academic tradition. Among early characterizations of peasant economies in the post-war development discourse, two became particularly influential. The first arose from the model of industrialization proposed by Arthur Lewis (see Lewis, 1954; Ranis and Fei, 1961; Jorgenson, 1961), and the second derived from the model of agricultural change proposed by T.W.Schultz (Schultz, 1964). Lewis described a prototype rural economy that was inefficient, in that it contained unemployed labour. Schultz’s prototype economy was efficient in using resources, but these resources had poor capability. Both prototypes were consistent with poverty and its persistence, but they gave rise to rather different prescriptions. Lewis proposed transfer of workers from rural to urban occupations. Schultz advocated raising agricultural productivity, instead of a patient wait for industrialization and migration. The first twenty years of post-independence Indian development strategy was strongly Lewisian in its single-minded reliance on industrialization and transfer of population from agriculture to industry. But the one condition for the model to work—rapid labourintensive industrialization—was missing from the scene. The stasis was broken by rekindling investment in agriculture, the way poinίed out by Schultz. And yet, writing at a time when parts of rural India were sliding towards violent peasant uprising, many Indian economists reacted to the idea of an efficient equilibrium with spontaneous scepticism. V.M.Dandekar (1966) gave that reaction the form of an argument.16 In Schultz’s ahistorical world, Dandekar inserted history via exogenous population growth and a fall in the land-labour ratio. Through demographic pressure, the peasant family farm entered a crisis of inadequate savings. Though Dandekar did not cast his arguments in terms of technical change, he was in effect asking why the small farm might fail to attain a land-saving investment as demography changed resource costs. Schultz’s book, Transforming Traditional Agriculture (1964), endorsed a contemporary rethinking of agrarian policy. Soon after the book appeared, parts of Indian

Rethinking economic change in India

20

agriculture did begin to transform themselves through a state-induced biological revolution. If the solution worked, it worked then in relatively small pockets such as Punjab that already had access to water for the dry seasons. A much bigger part of traditional agriculture displayed signs of gathering distress and political unrest and no green revolution in sight, such as Eastern India or the arid Deccan plateau. For some of these poorer regions of India, a remedy came in the 1980s. The resource that sustained this second biological revolution was groundwater, a common property the new regimes in rural states deliberately allowed to be harvested beyond sustainable levels. Beleaguered small-farm economy could finally lift itself from the spectre of diminishing returns, but now at a grave ecological cost. Both Transforming Traditional Agriculture and its principal critic in India, Dandekar, shared a diagnosis of Indian poverty in terms of structure rather than power. Neither author took colonialism or class seriously. Nor did the specialists on colonialism and class take much interest in these structural theories of stasis. The shared emphasis in the former body of thought on resources and efficiency provides the framework in which the present work can be located. Like these works, the labour history proposed here looks closely at traditional resources such as land and labour in accounting for poverty and the persistence of manual labour. That being said, the book is different from the older literature on several fronts. Neither model of ‘traditional agriculture’ was explicitly historical. Therefore, neither model allowed room for changes in labour institutions, markets and transfer of resources within agriculture or within industry. By contrast, this work is not only an attempt to explain poverty and labour, it is also an attempt to explain change in modes of working. The book emphasizes the capacity of individuals to break through poverty and stasis, and recognizes that it was the growing market for wage-labour that provided them with the principal means for doing so. It takes an essentially disaggregated view of the rural economy and the labour market. The numerous segments and fissures within the immense world of manual labour created opportunities and incentives for gainful transfer of population from surplus to shortage segments throughout modern history even as aggregate production conditions turned adverse to labour. It looks beyond rural labour to small-scale industry, which provides an important counterexample to diminishing returns. And lastly, this work is aware of the existence of culture blocks that decided who could effectively play the market game and who could not.

The chapters Chapter 2 presents an overview of the field. The reader wishing to get into the analysis directly can skip this scene-setting chapter. The narrative project behind this book is also restated here. The six chapters that follow implement the project with reference to agricultural labour, the artisans and women. The three specific transitions examined are, de-peasantization or growing landlessness (Chapters 3 and 4), de-industrialization or the artisans’ loss of livelihood (Chapters 5 and 6) and de-feminization of the industrial workforce (Chapters 7 and 8). As we have seen, in debates about colonialism, the experience of the rural labourer has been a central theme. But the empirical content of these debates is rather sketchy. Studies

Introduction

21

on regional agrarian history contain valuable information on labour, but do not link easily with a general account. A return to the rural labour theme, therefore, is useful. Chapter 3 re-examines a well-known thesis that peasants turned into wage-labourers on a large scale in colonial India. The tendency was attributed to colonial policies that led to loss of access to land and livelihood. The chapter explores the relevant data and scholarship, and finds that the incidence of labour did increase, but in a way that calls for a different explanation. Two processes were particularly active: fragmentation of land and changes in occupational mix that drove a large number of general rural labour into dependence on agricultural labour. Growing landlessness can lead to depression in real wages in the long run. Chapter 4 explores the data and concludes that the trends in wages were not uniformly downward. In fact, there is evidence of a rise in real wages during the first quarter of the twentieth century. There were periods of wage depression both before and after this phase. Between these periods, however, there was a significant difference. In the late nineteenth century, depressed wages derived from fixed money wage and rising prices. In the 1930s, depression derived from flexible money wages adjusting to prices with a lag. The fixed and flexible money wages represented different labour regimes. The latter was a market in which money wages were actively negotiated, the former one in which there was an element of custom in wage payments. If this reading is accepted, the long trends in real wages tell us a story of institutional change. Chapters 5 and 6 consider the experience of the artisans. In the retardation thesis, artisans lost livelihood and crowded into agriculture, adding to landlessness. That many artisans did lose their traditional living throughout the nineteenth and twentieth century is beyond question. What that process means for industrialization, however, is open to question. There can be two meanings attached to this trend: a quantitative one of a large decline, and a qualitative one of a destruction of the potentials to industrialize. The quantitative sense is not a serious matter, for a decline in the artisanate was a global phenomenon and was consistent with industrialization, not contradictory with it. The qualitative sense is more serious and worth a closer look. If ‘industrialization’ means rising share of industry in national income, a simple test of qualitative decline would be the share of artisans in national income. Between 1900 and 1935, the test rejects the qualitative decline thesis. Artisans broadly maintained their presence in the national economy in the early twentieth century. For the nineteenth century, we cannot conduct the test because we have no income data. We have some employment data, but employment is not a sufficient test. At best, some informed speculations are possible. Chapter 5 carries out such an exercise and concludes that, while a decline in employment did occur in the nineteenth century, it was more evenly distributed over time than has been believed, and was part of a process that entailed more income gain than loss. By the end of the nineteenth century, however, the forces of decline had spent themselves. If we take a long view from that point on, artisans not only survived but also contributed to industrialization in the twentieth century. In national income terms, they contributed to increasing average product in manufacturing in the same extent as capital-intensive industry. I call this process ‘labour-intensive industrialization’. Chapter 6 investigates the conditions that made such industrialization possible, and describes several features of an ‘industrious revolution’.

Rethinking economic change in India

22

Transformation of family and customary labour to casual wage labour was powerfully and persistently gender-biased. The scholarship on the history of women’s work in industry focuses on both barriers to entry into paid work, as well as barriers to exit from the family. Chapter 7 reviews the literature and develops the notion of a specific barrier to exit related to the quality of training received. Chapter 8 pursues these themes further through a descriptive survey of women artisans in colonial India. The last chapter addresses the progressive decline of economic questions within South Asian labour history. In the last twenty years, the study of rural labour in South Asia has taken a decidedly political turn under the influence of the subaltern studies collective. Partly in reaction to Marxist historiography in which ‘agency’ was submerged in ‘structure’, subaltern studies focused on more abstract and ubiquitous forms of power, as well as resistance to power. This accent on the ability of deprived peoples to make their own history endured, but in the process, subaltern studies itself imposed a structure on the way this ability was conceived and discussed. Chapter 9 offers a critical appraisal of the contribution of subaltern studies from the standpoint of economic history. It is possible to enter the theme of labour and livelihood from any of these points— decline of peasantry, change in women’s work, labour-intensive industry, or indeed, the polemic over historiography. But a general contour of the time and space seems appropriate before we get to the specifics.

2 Economic history and modern India Redefining the link British rule in India formally lasted between 1857 and 1947. How large, of what nature and how lasting was the impact? These questions have long guided the study of the economic history of India. The imperialist or ‘orientalist’ belief was that the empire heralded modernity in India. Karl Marx shared that belief with many of his contemporaries, although he also observed that modernity came with a cost. Twentiethcentury writers on imperialism and development emphasized these costs, and argued for an enduring link between colonialism and underdevelopment. The view that impediments to development were inherited from the damages of colonial rule, and not home-grown, was a key premise of Indian nationalist thought as well (see also Chapter 1). It was articulated, among others, in the writings of Jawaharlal Nehru himself about the origins of modern India. In 1947, this diagnosis of Indian poverty held that it was a product of ‘laissez-faire’ policies of the colonial regime, exploitation by foreign capital, and the non-interventionist stance of the Indian government under the British raj. In this view, the channels of influence in global, even local, economic exchanges were essentially political. In turn, these ideas supported the two key planks of India’s development strategy: strong sentiment against foreign trade and investment, and statism. Indian big business in 1947, the principal backers of the Indian National Congress, eagerly embraced the former and, somewhat uneasily, the latter. These policy stances now have few takers in the nations of South Asia. Since 1990, if not earlier, the world-view that habitually warns against globalization has been in decline. Faith in statism has diminished too. The study of India’s economic history has been affected by this shift. Scholarship continues along the imperialism-underdevelopment axis, albeit on a smaller scale than in earlier years. The belief that British rule created misery persists in amateur history. But this stance looks increasingly dated and disoriented, especially at a time when economic liberalization in India itself is drawing upon the tenets of classical political economy on which British policy in India was founded. Another reaction is to sidestep questions of India’s economic history and to focus instead on issues of recent decades. Indeed, the study of the economy history of India is at risk of losing wider relevance, audience and funding. This chapter argues that in order to restore the link between economic history and modern India, a different narrative of Indian economic history is needed. An exclusive focus on colonialism as the driver of India’s economic history misses those continuities that arise from economic structure or local conditions. The power of colonialism to impair productive capacity has been overstated in standard narratives on Indian economic history. In fact, market-oriented British imperial policies did initiate a process of

Rethinking economic change in India

24

economic growth based on the production of goods intensive in labour and natural resources. However, productive capacity per worker was constrained by low rates of private and public investment in infrastructure, excessively low rates of schooling, social inequalities based on caste and gender, and delayed demographic transition and the resultant demographic burden placed on physical capital and natural resources. The end of colonialism did not see a dramatic break in these conditions. In fact, the demographic burden became heavier in the post-independence period. Economic policy between 1950 and 1990 attempted harder than the raj to raise the quality of labour and rates of investment. A striking success of the new economic policy was the rise in agricultural yield per acre and per head. Still, economic growth continued to be intensive in semi-skilled labour. On the other hand, whereas British policy believed in exploitation of comparative advantage in trade, independent India turned firmly away from participation in the world economy, precisely at a time that the world economy experienced a boom as never before. Traditional resources like labour improved in quality to some extent but also lost a large potential market. When economic reforms in the 1990s re-integrated India in the world economy, the major beneficiaries were manufactures intensive in semi-skilled labour, in a late but welcome reversion to the colonial pattern of growth. This chapter begins with a descriptive tour of India’s economic history based on recent research. It returns later to the long-term continuities between colonial and post-colonial India, especially in resource endowments.

A descriptive tour: 1757–1947 It was a century from 1757, when the English East India Company established its supremacy in Bengal, to 1857, when the Crown took over administration of India. British Crown rule over India lasted ninety years, from 1857 to 1947. The period of British colonial rule was long enough to defy any simple summary. However, in discussing this period it is useful to focus on three features. First, structural features include the overwhelming importance of natural resources and labour to economic growth, fluctuations and welfare. Agriculture and labour-intensive industry and services were the main livelihoods throughout this period and beyond. Second, global features consider the fact that India’s economy was more open during this period compared to periods before and after colonialism. India participated in a global revolution in transport and communication, which for India includes especially the Suez Canal, the railways and the telegraph. The third set of features can be called colonial features. For example, that India was a colony imposed certain peculiarities on its balance of payments, like large remittances paid by the government to Britain. Being a colony, however, was not necessarily a fiscal burden. The ratio of investment to government expenditure was apparently much higher in British India than in Mughal India. The structural features of India’s economy changed rather slowly. For example, India’s economy was primarily agrarian before, during and since colonialization. However, the global and the colonial features shifted dramatically after 1947. Industry in colonial India had strong global ties, whereas after 1947 the policy of ‘self-reliance’

Economic history and modern India

25

involved a deliberate and drastic reduction in the influence of global factors on the domestic economy. Setting the stage: the century before British control An orientalist cliché, with adherents as great as Karl Marx and Max Weber, held precolonial South Asia to be stagnant and backward in political-economic terms. A corollary of this cliché was that economic modernity in South Asia began with European involvement in the region. Later research has proven this idea to be a myth. South Asia was already a major player in world commerce, and possessed a well-developed trading and financial world, when Europeans discovered it. Interestingly enough, the Indian trade at its peak involved an exchange of goods, not for goods, but for bullion. The sustainability of the trade depended on plentiful supply of precious metals from Spanish America. In this way, Indian trade helped to establish what Immanuel Wallerstein and other historians call a ‘world system’. Currently, the picture of the early-modern Indian economy derives significantly from research on the Indian Ocean trade. From the 1960s, the Indian Ocean trade in the seventeenth and the eighteenth century and, connectedly, the coastal economy, textile production and export trade, has been a major field of historical research. One clear thread has emerged in this scholarship that undermines any essential connection between the Empire and economic modernization. This is the centrality of India in the global network of trade in the seventeenth and eighteenth century. Indian merchants were not only long prominent in international trade, there was also compatibility rather than contradiction between the interests of the European companies and Indian traders until the mid-eighteenth century. The counterpart argument is that the Indian merchants retreated thereafter because of political intervention. Going further, one can argue that Empire, in fact, damaged prospects of an endogenous capitalist development. In different ways, such a sentiment shows up in comparative histories linking maritime trade with development.1 But radical claims in world history scholarship, such as the one made recently by André Gunder Frank that the centre of early modern world economy was Asia rather than Europe, are not reliable (Frank, 1998). For, such claims usually involve rather exaggerated assumptions about the share of regional commercial blocs in world trade, and the size of the export economy relative to the more subsistence-oriented economy within these regions. Notwithstanding the importance of export trade for some coastal areas of India, the significance of the Indian Ocean for the regional economy as whole was rather small, even negligible. At any rate, strong claims in this regard are often not strongly substantiated. Whatever data one can gather on the subject suggest the marginality of the trade, a point that was expressed somewhat brusquely by the ‘imperialist’ historian William Moreland (1920): The total trade of India going overseas in Akbar’s time could be accommodated in a couple of the cargo vessels leaving Calcutta port in the time of Minto’.2 Akbar’s time may not be a fair comparison. But even at its height (the late eighteenth century), export by the English and Dutch East India Companies was a minuscule percentage of probable regional, even sub-regional, incomes. The late eighteenth century political realignment in land and sea saw the retreat of the Indian merchants and ascent of the English East India Company as a territorial power.

Rethinking economic change in India

26

This transition remains critical to the significance of the eighteenth century for a historian of the imperial India. The standard view here is that the English used their political power to suppress Indian entrepreneurs and former collaborators. But there is a minority view that suggests that the Indian merchants declined because of the intrinsic weakness, even absence, of mercantile organization (Das Gupta, 2001). The century before 1857 was a time of territorial consolidation. In the early eighteenth century, Europeans dominated India’s maritime trade with Europe. By 1757, the English East India Company commanded political power in Bengal. The transition from trade to direct rule can be explained partly by the needs of trade itself. The mercantilists criticized the payment of bullion for Indian textiles. Local political circumstances that enabled the British to command land revenues of Bengal came as a less controversial means of payment. The local circumstances included the support of the elite disaffected by the local rulers. When the Company’s monopoly in trade ended in the early nineteenth century, it was seriously committed to building an empire. By 1857, the boundaries of colonial India, which were the basis on which nations were carved out in 1947, had been defined. A more or less uniform administrative system came in place in this time-span. In the economic sphere, there were several major changes. Agrarian ‘settlements’, which were contracts between the state and the cultivators on property rights and revenue commitments, were drawn. The British wanted to create a class of cultivators with secure property rights who would yield more revenue to them by pursuing profit-oriented cultivation. In effect, property rights often went to non-cultivating classes because of mistaken identity, imperfect information or political compulsions. The legal recognition of a property right, conditional on payment of land revenue, went along with the erosion of many customary rights over usage of land or what it produced. These rights were poorly understood, oversimplified or irreconcilable with private property right. Tenancy rights and rights to the use of common lands were victims of this confusion. Ultimately these settlements transformed rural institutions on resource use, restructured classes and had mixed effects on growth, the mix being variable over time and space. One universal effect of introducing secure property rights was the extension of markets in land. Since greater security for some rights inevitably increased insecurity for other rights-holders, the move also contributed, more silently, to the extension of a labour market. Another set of changes had their origin in foreign trade in an increasingly integrated world. Trade expanded quickly, and saw significant changes in pattern and composition. At the close of the Napoleonic wars, the market for Indian textiles had shrunk as a result of tariffs on Indian goods in England, which were steeply raised between 1797 and 1814. At the same time, technological change in weaving, most decisively Horrock’s powerloom, reduced the difference in cost of production, ending cotton textile’s pre-eminence as exportable from India though not ending the trade completely. The East India Company’s commercial monopoly ended in 1813. Already, many former employees and other individuals had established partnerships with Indians to carry on export of cotton, indigo, opium and sugar. Profits of these trades sustained new commercial-cum-port towns such as Calcutta, Bombay and Madras, helped some Indian groups accumulate capital that later found its way into industrial ventures and built close trading links between India and China, which was a market for Indian opium and a source for tea. These trades dwindled in the second half of the nineteenth century. Natural indigo was

Economic history and modern India

27

steadily displaced by synthetic indigo. Lancashire ‘discovered’ Indian cotton during the American Civil War, but subsequently more cotton was used by the newly established Indian mills themselves. The Sino-Indian opium trade began to dwindle from 1906, in the wake of worldwide anti-addiction movement. If bullion was India’s principal import in the eighteenth century, the composition of import began to change towards manufactures, notably British textiles from the second quarter of the nineteenth century. It is not easy to read in this period, roughly 1800–60, a general trend. We do not have the basic data to make an assessment of growth, stagnation or decline in the early nineteenth century. Nevertheless, there is a widely shared belief that the consolidation of British power in the economic sphere saw a violent and uncompensated disturbance. The fact of a decline, the period, the regions and the causes, remain imprecise. One thing we do know is that India’s traditional cotton textile industry declined between 1820 and 1880. At first an export market for Indian cloth disappeared. Later, hand-spun cotton yarn and hand-woven cloth suffered because of the import of yarn and cloth from the mills in England. The decline seems dramatic if seen against India’s earlier dominance in world textile trade. This single example of decay appears to have generated the ‘deindustrialization’ thesis, which at its narrowest holds that early British rule introduced a violent shock to India’s economy, and at its broadest holds that colonialism destroyed the potentials to industrialize. Both the narrow and broad inferences, however, are deeply controversial for three reasons. First, the decline was apparently restricted to cotton textiles. Second, the decline of the textile industry did not continue through the rest of the nineteenth century and on into the twentieth century as British colonial rule strengthened, which calls into question any connection drawn between politics and the potentials to industrialize. Third, a decline in cotton textiles was not capable of causing economy-wide distress. The pro portion of textile export in total textile production was very small, at its peak not more than 1–2 per cent. Further, losses for the Indian textile producers were largely balanced by large gains in consumers’ and producers’ surplus. To illustrate the gains, by 1880, prices of ordinary cloth and yarn were about 20 per cent of what they were by 1800. On the side of loss, many of the jobs lost because of competition with mechanized textiles consisted of poorly paid domestic workers with rather small implicit wages arising from small opportunity costs. A second and more plausible source of a regress was taxation. The new regional governments collected taxes more thoroughly than the regimes before them in areas where direct contract between the state and the cultivator (or ryotwari) was in existence, reportedly causing a decline in peasant production and demand in the 1830s and 1840s. This process was not present in all regions. It was absent in eastern and northern India where the state did not collect tax from the peasant directly. Further, the timing and intensity of the crisis remain imprecise. So far, more or less the only ‘evidence’ cited to show widespread distress is a rather dubious one, price depression (see also Chapter 5, note 3). The central role of agriculture in India’s economic history Agriculture has been the predominant sector for India’s workers for the last two centuries, right up to the present. As shown in Table 2.1, about 70

Rethinking economic change in India

28

Table 2.1 Employment, 1901–91 (millions, percentage share of main occupational classes in brackets) 1901a Population Workers c

Agriculture and allied occupations Modern industryd e

Traditional industry and construction f

Services

Others (mining and unspecified)

1931a

1991b

285.2

338.1

846.0

131.6 (100.0)

139.1 (100.0)

278.9 (100.0)

89.3 (67.8)

98.8 (71.0)

186.2 (66.7)

0.6 (0.4)

1.6 (1.2)

8.3 (3.0)

13.3 (10.1)

9.8 (7.0)

20.1 (7.2)

18.9 (14.4)

20.8 (15.0)

57.1 (20.5)

9.5 (7.3)

8.1 (5.8)

7.2 (2.6)

Sources: Heston (1983), Sivasubramonian (2000), and the Statistical Abstracts of India, various years. Notes a Undivided India. b Indian Union. c The main occupations were cultivation, livestock rearing, plantations, forestry and fishing. d Represented by officially regulated factories. e Represented by units outside officially regulated factories. f The main occupations were transport and communication, commerce, public administration, professions and liberal arts.

per cent of India’s employment was in the primary sector in the first few decades of the twentieth century. By the start of the twenty-first century, after fifty years of state-backed effort to industrialize, the share of the primary sector in GDP fell from over one-half at the time of independence to about one-quarter at present. Nonetheless, the majority of workers in post-colonial India continued to be engaged in the primary sector. Its share in employment fell only to about 60 per cent in the 1990s. Thus, conditions for agriculture have been a primary determinant of the well-being of most of India’s people. The typical weather pattern in most of India is nine months of dry weather and three months of monsoon season, which refers to the seasonal shift in wind direction between June and September that brings 90 per cent of total rainfall in the region. Rainfall during the monsoon season is usually adequate to raise one or two food crops in the months following the monsoon. But rainfall is rarely adequate for winter crops and marginally adequate in some of the drier regions, even for the main food crop. High risk, therefore, was a constant feature of economic life in most parts of India throughout history. If the monsoon rains failed even slightly, starvation was widespread and sudden. In the short run, famines affected all parts of the economy via violent shifts in consumption and labour force. For example, in Madras Presidency the great famine of 1876–78 took three and a half million lives, or 10 per cent of the population, and a larger percentage of workers.

Economic history and modern India

29

The work by Amartya Sen and others have changed historians’ approach to famines considerably. Famines are now seen more as processes rather than as accidents of nature. The great nineteenth-century famines, in 1876–78 and 1898–99, do illustrate such a view. The impact of famines depended on administrative failure, market failure and social failure. The state relied—in good or bad faith—too heavily on the market mechanism. But there was market-failure because transaction costs were high during famines. In turn, the state’s own refusal to compensate for missing markets added to the intensity of the famine. Social failure was revealed in the fact that certain castes, particularly the rural labourers, figured prominently among those who died. The roots of these failures were already present, perhaps even gathered strength, when a devastating weather shock occurred. That being said, nineteenth-century famines were also powerfully ecological phenomena. The regularity of the famines, the degree of crop failure, cannot be understood without reference to monsoon ecology, especially the conditions of the Deccan plateau. Some regions were far more prone than others because of their harsh agrarian environments. The probability of a weather shock was distributed very unevenly across space, and was correlated with the probability of a famine. The most vulnerable region was the vast arid Deccan tract. Weather-induced famines, remarkably, disappeared after 1900, to be replaced by more localized scarcities and ‘droughts’ What caused this change? The state had set up a relief system by then, but its reach was limited. And as the Bengal famine of 1943 displayed, the capacity of the state to manipulate grain markets had a menacing aspect. A more plausible explanation is the efficiency of the market mechanism itself especially after the full development of the railway network by 1920. Thereafter, droughts in India have had a tendency to localize in regions particularly ill-served by mass transportation. As a result of famine mortality, population growth in the first half of the nineteenth century was low (0.4–0.5 per cent) and subject to high fluctuations. Steady rise in population from 1921 was partly an outcome of the rarity of famines, along with better health care and possibly nutrition. Between 1914 and 1946, the rate of population growth was 1.2 per cent per year. In the long run, two observed tendencies seem attributable to pervasive and endemic risks. First, rates of private investment in India have generally been low. Instead, Indians who held assets displayed a marked preference for precious metals, which tended to be more stable in value, but generated smaller return than productive investment. Second, the high risk of famine mortality was possibly a reason why birth rates also tended to be high. During the interwar period, when mortality began to fall, the high birth rate led to a dramatic rise in population growth. In this primarily agricultural society, cultivation patterns and livelihood risks depended on the distribution of rainfall. Mean annual rainfall in India ranges from more than seventy inches on the western coast and Bengal delta to thirty inches or below in large parts of the interior. Areas with high rainfall tended to grow rice; those with low rainfall focused on coarser grains or millets. Rice and rainfall were generally associated with high population densities and low ratios of land to labour—because the combination of rice and rainfall normally meant lighter impact of famines and greater requirement for farm labour. The eastern coastal areas where British colonial rule first established itself had abundant water, fertile land, dense populations, well-developed foreign trade and

Rethinking economic change in India

30

relatively hierarchical societies. Land in these areas could sustain high rents, and thus a prosperous rent-earning class, who were rarely peasants themselves. From a mistaken identity, the British administrators conferred on them ownership rights under the Termanent Settlement’, overnight turning the direct producers into mere tenants. The interior regions conquered later were drier, more sparsely populated and cultivated coarser and sturdier grains. Peasantry here was less hierarchical with powerful kinship units, and these units tended to control land collectively. Farming here co-existed with extensive raising of livestock. In a large part of this region, property rights were granted to direct producers. From a mix of ecological and political reasons, the government invested heavily in extending canal irrigation in the drier interior regions. Coastal Madras, a rice region that saw canal construction on a large scale, was an exception. Between 1885 and 1938, cultivable area increased by sixty million acres, of which over half was irrigated. The latter half of the nineteenth century saw agrarian commercialization driven by trans-local markets. Between 1880 and 1925, the real volume of trade to and from India doubled. The value of exports quintupled between 1870 and 1914. More that half of Indian exports now consisted of agricultural goods such as grains, seeds, raw cotton and raw jute. The age of the artisan had ended, and the age of the peasant arrived. Close to half of Indian imports consisted of manufactured consumer goods, chiefly cotton textiles. The ratio of trade to national income increased from less than 10 per cent in the 1860s to about 20 per cent in 1914. British industrialization influenced the composition of trade, but the scale was also influenced by internal and external changes in the costs of trade. Ocean freight rates on bulk goods from India became one-third between 1873 and 1893. By the latter date, twothirds of exports from India passed Suez. Early in the nineteenth century, India’s product markets were constrained by a multiplicity of weights and measures, backward and risky transportation systems and extensive use of barter. Administrative reforms and political consolidation of British colonial rule weakened these constraints and enabled closer integration of local, regional and global markets. Between 1870 and 1914, agricultural prices consistently rose. Transactions costs fell. Land sales, land prices and rents increased. Credit transactions expanded. Labour became more mobile and more market-oriented. Millions went overseas. Profit opportunities led to changes in resource use. For example, in what had been the drier millet zones, after irrigation, a basket of ‘cash crops’ became common, examples are wheat, cotton, oilseeds, sugarcane and tobacco. The three conditions that had made agrarian expansion possible in the late nineteenth century—surplus land, investment in irrigation and growing exports—all weakened in the interwar period. As shown in Table 2.2, production of food crops was essentially unchanged from the early 1900s to the late 1940s, even though production of non-food crops and large-scale industrial production increased more rapidly. One interpretation of this slowdown in agriculture is that the resource-based trade-driven growth had reached its limits. Cultivable waste lands became scarce, investment in water slowed, and so did the world economy. Some growth continued in cotton and wheat, but increasingly dependent on yield-per-acre rather than acreage, in other words, dependent upon seeds developed or adapted in government laboratories rather than on wider access to water.

Economic history and modern India

31

That said, the principal source of agricultural stagnation was a crop and a region that had participated in a rather limited way in the whole transition—rice in Bengal. Thus, historians have also looked for other hypotheses for the slowdown with Bengal in mind. One theory focuses on class structure. In densely populated, rain-fed, rice-based areas like Bengal, the British had conferred property rights upon formerly rent-earning groups, perpetuating their power and blocking the way for basic restructuring in rural society. In the drier millet-based regions where ‘land was plentiful and hands few’, the state made revenue arrangements directly with the peasants, creating a positive incentive for private investment.3 Another explanation, complementary rather than competing with the former, can be resource-endowments. In the Bengal delta, income from rice had to be shared between too many people dependent on land. By early in the twentieth century, population growth in this region had led to the cultivation of inferior land. The rice areas that did well commercially, such as coastal Madras, had lower population densities and received canal irrigation that made it possible to combine rice with dry-season crops. Whatever factors were behind the stagnation of agricultural output, they were longlasting. The regional pattern of agricultural growth and stagnation since independence has been similar to the regional pattern of growth and stagnation before independence. Pockets of rural poverty today emerged as pockets of rural poverty in the latter half of colonial rule. Eastern India was prominent among these. Areas that experienced a ‘green revolution’ in the 1970s and 1980s were already advancing during the British rule. Northern and parts of Western India were examples. Land has been scarce in an absolute sense from about 1900. By and large, success in breaking the resource barrier after 1947 has depended on

Table 2.2 Production and wages Index of production Food crops

Nonfood crops

Modern industry

Index of real of income in traditional industry

Index of real wages Modern Industry

Non-agricultural, outside modern industry Skilled

Unskilled

1900–01 100 to 1904–05

100

100

100

100

100

100

1934–35 103 to 1939–40

146

293

127

185

160

151

Sources: Sivasubramonian (2000: tables 4.28, 4.40, 4.41, 6(g)); Blyn (1966: appendix 4~). Modern industry real wages are simple average of three indices, Bombay cotton mills, Ahmedabad cotton mills and Calcutta jute mills. The wage indices were estimated by Mukerji (1959,1960,1961).

irrigation, seeds, chemical fertilizers and, to some extent, exploitation of common property resources such as forests and pastures. How did the new regime of rights and markets contribute to standards of living? Curiously enough, most general estimates of inequality in land-holdings reveal an

Rethinking economic change in India

32

absence of change. But underneath this aggregate stability, there was quite massive reshuffling. There was decline of aristocracy, rise of rich and middle peasants, and gathering vulnerability at the lower end. Generally, the pre-colonial landed aristocracy declined, or left the village to pursue urban occupations. In Eastern India, where the precolonial elite were granted property rights, the decline was a more long-drawn affair, but happened all the same. An immediate cause of decline of the older elite was dispossession of sovereign powers. A more long-term factor was commercialization— possibilities of intensive cultivation and profit opportunities that gave rise to a new class of rich peasants and sub-landlords. Increasingly, they commanded more political as well as economic power than the older elite. Agricultural income indices are generally patchy. Between 1890 and 1950, no marked change in average real wages seems to have occurred. Real agricultural income per worker seems not to have changed much either. But inequality may have increased. At one end, non-wage incomes represented the earnings of the ‘small peasant’, who relied mainly on family labour and tilled land barely enough for subsistence, and usually had insecure property rights. At the other end were ‘rich peasants’ who had secure property rights, controlled enough land to generate a surplus, employed labourers, had better access to credit or were creditors themselves. As a rule, rich peasants gained from commercialization. That is, returns to capital increased. The evidence on small peasants is mixed. In some cases, they did well; in other cases, they gained in the nineteenth century but regressed in the twentieth. On a limited scale the small peasant turned into a labourer. Instances of the peasant losing land have received exaggerated importance in academic debate on the impact of colonialism. In one extreme view, such instances symbolized a general rural decline and dislocation caused by colonialism. In another view, stories of such reversal were neither very general, nor attributable to colonialism. After all, in the long run, the Indian small peasant faced steady fall in land-person ratio because of population pressure. Chapter 3 returns to this theme. Although there is no strong evidence to suggest the labourers became on average better off with the commercialization of agriculture, wages did rise in the major cash crop regions. Further, colonialism brought changes in the labourer’s social position. In precolonial India, labourers came from castes whose primary duty was to perform labour. Many were akin to serfs, and some at least were actually saleable. In the colonial period, this serfdom or slavery declined. The element of compulsion and force in employment weakened. Various forms of social oppression, such as enforced dress codes and codes of conduct with respect to upper castes, weakened too. The possibility of migrating to the cities and to other British colonies made occupational choice more diverse. The decline of attached labour was partly induced by the widespread exit of these castes from agricultural labour, and entry into plantations, mines, urban services, public works and government utilities. Industry India’s workforce is not significantly more industrial today than it was a century ago. In 1901, 13.9 million industrial workers formed 10.5 per cent of the workforce, as shown in Table 2.1. In 1991, 28.7 million workers made up 9.4 per cent of the workforce. The share of industry in national income has grown modestly from 11.1 per cent in 1900–10

Economic history and modern India

33

to 16.4 per cent in 1940–46, to 27 per cent in 2000. India’s independence in 1947 did not represent any marked break in the pace of industrialization as measured either by employment or share of national income. In describing industrialization in colonial India it is necessary to begin with a distinction between traditional and modern industry. Modern industry (or large-scale industry) had two characteristics: use of machinery, and factories subject to some form of regulation and modern managerial practices. By contrast, in ‘traditional’ industrial firms, machinery, size and hierarchical management played no significant role. Both traditional and modern industry shared one feature: intensive use of labour and/or locally available raw materials. The main examples of large-scale industry were cotton and jute mills. Examples of traditional industry include handloom textiles, leather manufactures, metal utensils, pottery, carpets and shawls, etc. Large-scale industry employed 2–3 per cent of India’s industrial workers at c. 1900 and a little over 10 per cent at 1947. Its share of the national income generated in manufacturing increased from less than 10 per cent to 40 per cent over this time. Factory employment in the colonial period was overwhelmingly dominated by the textile industry: mills for cotton and jute spinning and weaving; cotton ginning firms and jute presses; and a few isolated large firms in wool and silk spinning and weaving. The other mechanized industries were paper, sugar, matches, cement and steel. Technology and the capital goods were imported, but Indian mills used a far higher proportion of labour to capital than the comparable factor-proportions in the same industries in Britain. These modern factories were concentrated in two provinces: Bombay and Bengal. These provinces were not particularly industrial. Their attraction, especially the pull of the cities of Bombay and Calcutta, derived from their position as ports, railway termini, financial centres and commercial hubs. Modern industry was essentially a product of the times and a product of India’s contact with Britain. In cotton and jute mills, the idea of a mill, the technical knowledge, the equipment, a part of the capital and a section of the engineers at first came from Britain. The higher capital-intensity, new origin and dependence on British precedence led to ways of organizing work that did not exist before. It gave rise to cities such as Calcutta or Bombay, shaped urban labour markets, encouraged the growth of railways, ports, laws, banks and technical schools, and was a force behind the modernization of services. At the start of colonial rule in the 1850s, India’s capital market institutions were totally inadequate for channelling household savings to industrial investment. The real cost of capital was astronomical. It is not surprising that the pioneers in modern industry came almost entirely from communities that had specialized in trading and banking activities—that is, those who could raise money more easily than, say, an artisan or a farmer. By and large, fixed capital in modern industry came from its own sources of funds, or from borrowings from within a small set of people known to each other. They included both Europeans and Indians. Factory labour was a new form of work in India in the middle of the nineteenth century. Machinery, migration, urbanization and discipline were new ingredients in the workers’ lives. Did these changes improve income and welfare? From the early 1900s to the late 1930s, real wages of mill workers did increase quite substantially in the cotton mills of Western India, and marginally in the jute mills of the east (as shown in Table

Rethinking economic change in India

34

2.2). Still, most workers earned wages that were too little or too insecure to think of growing roots in the city and giving up connections with land and agricultural labour. However, the chances of occupational and income mobility were greater in the cities than in the villages. The city-dwellers never suffered the threat of famine in the way the rural population, and especially the lowly born within them, did. Historians have asked why modern industry remained limited in India. Two points of view exist on this question. Morris (1983) suggests that the scale of the home market was small for goods that used machinery. Bagchi (1970) suggests that the home market was shared by Indian and imported manufactures, and the colonial government did not protect Indian industry sufficiently from low-cost imports. For example, India never developed a capital goods sector, and did not see the kind of boost to machinery production associated with railway construction in mid-nineteenth-century United States. India’s railroads, the largest railway system in Asia, imported nearly all its equipment until the interwar period. Behind this policy, there was an explicit encouragement from the government to ‘buy British’, and possibly a disregard for experiments because the government guaranteed rates of return on private investment in the railways. The ‘buy British’ policy led to acute shortages during World War I, and changed in the interwar period, which saw some tariff protection for modern industry. By focusing on the extent of demand for products of modern manufacturing, both arguments sidestep the issues of resource-endowments and high cost of capital in India. Wider usage of machinery, whether for home or export markets, was not cost-effective because of the high cost of capital and the scarcity (and cost) of skilled labour. Machinery was used in those exceptional industries which processed raw materials abundantly available in India, and for which the machines and technicians could be easily imported. If we look away from this segment, the general situation was exactly as resource endowments would permit, that is, a vast world of traditional manufacturing, consisting of tool-based industrial production performed in homes or small workshops. Standard narratives of Indian industrialization have often neglected traditional industry from a mistaken belief that imports and modern industry killed it. Research on national income first challenged such a view, showing that income-per-worker increased quite significantly in this sector between 1900 and 1947, as shown in Table 2.2. Sivasubramonian finds that real product per worker in small-scale industry increased by about 60 per cent between 1900 and 1930, but fell in the next decade. Later work on specific industries showed evidence of a large rise in output per worker as well (Sivasubramonian, 2000: table 7.19). A different perspective that has taken shape more recently argues that the key dynamics in traditional industry was not that it became obsolescent, but rather that it was affected by commercialization in product and input markets (Roy, 1999). Commercialization involved a number of shifts: increasing integration of the market for the products of traditional manufacturing; and a shift away from production for own use or use as gifts and tributes to production for market; and a shift from local to longerdistance trade. These shifts led to changes in consumer and producer behaviour. As the market integrated, competition within the crafts intensified. There was decline of older institutional forms and the rise of new ones that used labour more efficiently. In particular, there was a decline in two types of non-specialized workers: women working in household industry and a group the early censuses called ‘general labour’, which

Economic history and modern India

35

performed a variety of labouring tasks in the villages and some manufacturing on the side. Leather manufacture gives an example of how commercialization affected traditional manufacturing. This was originally a rural craft performed mainly by rural serfs. In most places hides were subject to barter, but even where a market formally existed, servitude arose from both the caste-hierarchy and the interlocking of markets—the fact that the main customer of the leather artisan was also the peasant-employer. By the 1870s, an export market had arisen for leather, along with a need for different kind of processing. These changes weakened the serfdom of leather-makers, enabled massive migrations into the city, the rise of the merchant-owned urban factory and of wage labour. The case of handloom weaving is well-researched. Over a wide range of cloths, competition between traditional handloom manufacturing and the modern power-loom manufacturing was acute, and the share of traditional manufacturing eroded steadily throughout the nineteenth century. However, hand- and power-weaving also served segmented markets, and those segments of hand-weaving that did not compete with modern textile manufacturing saw a pattern of expansion in demand, commercialization, urbanization, technological and organizational change. A range of traditional manufacturing industries intensive in craftsmanship—carpets, shawls, engraved metals or silks—were always urban and always commercial. But the extent of urban concentration increased, and there was a qualitative change in clientele from powerful local patrons to exports. If Bombay and Calcutta with their large-scale factories represent one face of industrialization in India, numerous medium-sized towns such as Agra, Benares, Moradabad, Sholapur, Madurai or Jaipur, illustrate the continuing strength of traditional craft-based industry. Traditional manufacturing in these towns still displays artisanal roots. However, the numbers engaged in small-scale industry in these towns is now much larger, the markets have become more complex and more competitive, and freedom of entry and exit along with anonymity of transactors have become greater. Wages in real terms increased in traditional industry in the first half of the twentieth century, the rise being higher for the ‘skilled’ artisan, as shown in Table 2.2. Craftsmanship was a resource contributing to industrialization in India. In the largest industry handlooms, wages did not rise for the ordinary weaver. But earnings of the skilled weaver probably increased, and rates of profit were high, possibly rising, in the two decades before the Great Depression. We again have a scenario where returns to semi-skilled labour were uncertain, but returns to capital and craftsmanship increased. This whole process illustrates industrialization based on utilizing labour more productively, rather than on replacing labour by machinery. In that respect, colonial India was not fundamentally dissimilar to early industrialization elsewhere in the presence of surplus labour, such as ‘proto-industrialization’ in eighteenth-century Europe or industrialization in twentieth-century East Asia. Commercialization started the process. There was persistence, even strengthening, of traditional organization in the short run. But underneath that stability, a movement towards a labour market had slowly begun. In one respect, colonial India was different from these other cases. In Europe, modern industry had indirect roots in traditional industry. In India, the two developed side by side.

Rethinking economic change in India

36

Global flows of trade and capital India was a more open economy in the colonial period relative to both the eighteenth century and the first forty years of its independence. Openness can be seen in trade relations, income and capital flows, public finance and the monetary system. Before the nineteenth century, foreign trade was a negligible activity for India’s economy as a whole, though it was significant for certain regions. The ratio of trade to domestic product increased from 1–2 per cent at c.1800 to a little less than 10 per cent in the 1860s to 20 per cent by 1914. After 1947, the trade-GDP ratio in India steadily fell. It was 8 per cent in 1970, but has more recently risen to 13 per cent in 1985, 16 per cent in 1990 and 20 per cent by the mid-1990s. International flows of income and capital were also relatively larger in the colonial period than before or after. Net income from abroad formed 1–2 per cent of national income in India before World War I. Net income from abroad was well below 1 per cent of national income between 1950 and the mid-1980s. Until the Great Depression, India typically ran merchandise trade surpluses. In addition, India received financial investment from abroad in industry, commerce and infrastructure. In the international accounts, these two net receipts were balanced by three items of net payment: purchase of gold and silver, remittances made by the private sector and remittances made by the government. Government transactions were closely connected with the balance of payments. India’s government during the colonial period borrowed heavily abroad to finance its investment and other commitments. Repayment of these loans, along with regular remittance on account of charges made by Britain for costs of the administration of India, was a large net payment item in India’s foreign transactions. The money supply in colonial India was mainly influenced by the balance of payments. The primary objective of monetary policy was to stabilize the exchange rate. Stabilization of prices and outputs was meant to happen automatically. However when Indian interests and Britain’s interests came in conflict, stabilization in Britain’s external account was usually in the minds of those who decided Indian affairs. One of the most striking features of colonial India, and an enduring puzzle, is extremely low rates of investment. Net investment was 2–4 per cent of national income. Investment in machinery accounted for about half a per cent of national income. This low rate of investment has been attributed to colonialism. In an accounting sense, the relatively large remittance abroad on the government account led to lower government expenditure within India. For a given net receipt of sterling from abroad, a larger remittance implied a lower capacity to import. The period was one when a great deal of the machinery and raw materials needed by industry was imported. There was also a political problem on the side of financing government expenditure. The British government did not try too hard to tax either the local notables or the expatriates. Given limited taxable capacity, a larger remittance implied greater government borrowing abroad, which added debt service to the burden of remittances. Critiques of colonialism emphasized payments on government account, infamous as ‘drain’. These remittances held an element of transfer, in that some of the services for which these were due were overpriced. The British administrative elite, for example, were paid as grandly as its counterpart in Mughal India. However, a great deal of

Economic history and modern India

37

government expenditure was made for services that India needed but could not supply on its own, such as pensions to teachers and engineers, or payment of debts raised to finance railways and irrigation. After all, Britain and India were worlds apart in their technical, scientific and managerial capabilities. ‘Drain’, therefore, is extremely difficult to separate from legitimate factor payments. Even before separation, the scale of government remittances (0.5–1 per cent of national income) may not appear large enough to bear the ‘drain theory’. Explanation of low rates of investment has tended to focus on these colonial features. But the level of private investment was also low, which calls for more attention to economic structure. A climate of high uncertainty took a toll on the desire to invest. The hunger of Indians for gold and silver took a toll on productive investment and stability in the balance of payments. The slow pace of institutional development on the financial side was a negative factor for investment. For certain purposes such as remittance of money, the traditional money market was quite efficient. But that system normally did not deal in deposits, and was thus inadequate in channelling household savings into productive uses. Besides, the traditional money market was a closed system. It worked well when the transactors had a fair knowledge about each other, or could acquire such knowledge easily. Its doors were practically shut for all others. A more transparent and widely accessible financial system was a development that had to await the joint-stock banks. Modern banking and insurance technically began in the nineteenth century with the government-backed Presidency Banks. The majority of their clients were Europeans. The financial sector really expanded in reach from late in the interwar period with substantial infusion of small-scale Indian capital into these spheres. However, inadequate capital, asymmetric information and insider lending made the majority of the Indian-owned banks thoroughly unreliable institutions. India’s growth during the colonial period India’s pattern of economic growth during the colonial period is summed up by Table 2.3. The early colonial period between 1858 and 1914 saw positive economic growth for India. The rate of growth was small by modern standards, but not trivial by contemporary standards. India’s real national income grew at over 1 per cent between 1868 and 1914, per capita income at a little less than 1 per cent. These growth rates were rising late in the nineteenth and early in the twentieth century. One estimate shows that real national income grew at 1.8 per cent and per capita income at 1.2 per cent between 1865 and 1885, close to what Britain experienced in the last quarter of the century. By contrast, the interwar period of the 1920s and 1930s was a difficult

Table 2.3 Growth rates of Net Domestic Product (NDP) and Population, 1868–69 to 1946–47 (yearto-year growth rates, % annual) NDP 1868–98 1882–98

0.99 1.29

Population

Per capita NDP

0.40

a

0.59

0.51

b

0.78

Rethinking economic change in India

1900–14 1914–46

1.85

0.61c

0.61

d

1.18

38

1.24 −0.57

Sources: For 1868–98, Heston (1983: table 4.3A). For 1900–46, Sivasubramonian (2000: table 6.4). Notes a Growth rate for the period 1872–1901. b 1881–1901. c 1901–21. d 1901–46.

time, for the world, for India, for Britain and for India-Britain relations. On the positive side, the market for modern and traditional industry grew, in the case of the former because of tariff protection given to a set of industries. There was growth too in non-food crop production. On the negative side, there was acute stagnation in food production. Macroeconomic policy, the most infamous element in which was a slightly over-valued exchange rate which hindered India’s exports and paved the way for Japanese textiles to invade Indian market, caused political and economic stresses. World depression and excess capacity led to strains in busi-ness-government relations and in inter-firm relations. A combination of much slower growth in output and much more rapid growth in population meant that average levels of living stopped improving, or even declined, and the poorer sections of the agricultural population in particular faced harder conditions. India’s economic nationalism came on its own in a difficult time for India’s economy and the world economy. General assessments of the impact of British rule on the lives of ordinary people have been overly influenced by the economic experience of the interwar period. It is easily forgotten that the first sixty years or so of British colonialism delivered economic growth and a rising standard of living. At independence in 1947, the experience of stagnation, turmoil and a state-made famine of ferocious intensity loomed large in the minds of those entrusted with India’s economic future. In the economic philosophy that guided their vision, there was an instinctive reaction to the mercantilism of British rule in India. The next section begins with this turn.

Narratives of Indian economic history Until quite recently, most economic history research in India was conducted within a paradigm that saw development and underdevelopment, industrialization and deindustrialization, as two sides of the same coin. According to this informal consensus, markets and institutions built under the colonial situation retarded India and enriched Britain. Indian society and economy without colonialism, it was suggested, was capable of doing better than it actually did. Economic backwardness was identified with low levels of mechanized industry, a formulation that pervaded both Indian nationalist thought as well as leftist historiography. I called this paradigm the retardation thesis in Chapter 1.

Economic history and modern India

39

The paradigm tried to explain poor rate of growth of average income in colonial India. The retardation thesis suggested that the poor rate of growth was man-made. Colonial policies and the contours of the market economy that emerged, actually repressed the potential for economic growth. This thesis was based on five sub-theses.4 The first concerned external transactions. In the leftist-nationalist formulation, the ‘totality’ of colonialism was defined in terms of integration of India in world capitalism in a ‘subservient’ position, unequal exchange and ‘drain’. The second thesis offered an argument that the social surplus went to the wrong hands. Perhaps the best-known version of this argument was ‘forced commercialization’, which suggested that peasants, when shifting from subsistence to market, fell into domination by moneylenders who were averse to productive investments. The third thesis can be called ‘perilous commercialization’, the idea that the shift from food to non-food crops intensified famines. The fourth thesis was ‘de-industrialization’, the proposition that industrialization in Britain imposed large uncompensated costs upon India in the form of a decline in traditional industry. The fifth sub-thesis was a position on the public goods. The railways and the telegraph were seen as exploitative because they were introduced to aid imperial defence or foreign capital. The old school faced its first serious challenge when Morris D. Morris proposed a more positive view of nineteenth-century Indian development (Morris, 1968). The main message was that economic growth in nineteenth-century India was constrained by productive capacity rather than by politics, which played a benign or positive role. Indeed, the inconsistencies between facts and theory were manifold. Irrespective of the difference in their political status, British and Indian fortunes were complementary rather than contradictory. For example, GDP growth rates in India and Britain moved broadly in the same direction. Both were relatively high in the last quarter of the nineteenth century, and decelerated in the interwar period. Commercialization had a positive correlation with productivity growth. The vision of moneylender power in forced commercialization was a fiction. The thesis of perilous commercialization exaggerated the shift in cropping pattern, presumed knowledge about intensity of famines in the pre-colonial period, was inconsistent with data and under-played the fundamentally climatic origin of famines. De-industrialization was an inapt description for a region where millions of artisans not only continued in business, but also saw significant rise in output per worker. The uses and benefits of public goods such as the railways or the telegraph were not restricted to ethnic groups. The retardation thesis tended to see colonial market transactions as usually inequality-breeding, being usually a tool in the hands of the rich trying to gain a bigger share of a fixed amount of net output. Such outcomes could occur, but were not universal. Equally often, commercialization induced attempts to extend resources or enhance efficiency via better allocation of inputs. Can there be a narrative of economic history of India that explains the past, elucidates the present, is well-balanced, raises new research questions and does not become outlandish at the same time? A useful starting point for a rethinking would be to acknowledge that development and underdevelopment were not necessarily two sides of the same coin. Rather, Britain and India in the nineteenth century were two different coins. They were influenced by global factors and by mutual interaction, but also by their distinctiveness. To expect India to have followed the same path as Britain but for some external shock like colonization merely smuggles in a questionable teleology.

Rethinking economic change in India

40

The difference that seems to matter above all is resource endowments. India in 1800 or 1900, and South Asia as a region as late as in 2000, shared a common feature: an abundance of poorly equipped labour engaged in occupations subject to high climatic risks. Economic growth, in both colo nial and post-colonial India, has responded to this situation by focusing on labour-intensive methods of production, and by a preference for safe but low-return investment. Economic growth in the long run can be described as a two-stage process. In the first stage, unutilized traditional resources such as natural products and labour begin to be employed because of some kind of external stimulus, usually a market abroad. In the second stage, the stimulus continues, but some factors become scarcer such that diminishing returns to the abundant factor set in. In that stage, the economy either slows down, or sustains growth by more efficient use of the traditional resources. If land becomes progressively scarce, sustaining growth requires ‘land-saving’ technological change, chiefly biological inputs. If labour becomes progressively scarce, growth requires ‘labour-saving’ technical change.5 Both can raise the productivity of the input in scarcity. The former process of ‘extensive growth’ can give rise to an increase in total income or production, but no significant improvement in average income. The second stage can deliver rapid increase in average income. Nineteenth-century India experienced extensive growth. Integration in a rapidly growing world market supplied the external stimulus. Between the early nineteenth century and the onset of World War I, demand for food and raw material from industrializing Europe increased manifold. Peasants and merchants in India responded to these opportunities. Production for export increased, international capital movements grew rapidly and, within India, institutions and infrastructure necessary for a market economy to smoothly function were set up. These included uniform weights and measures, contract law, uniform currency system, the railways, the telegraph and a judiciary committed to defending private property rights. Until 1920 or so, land was not scarce. In fact, government investment expanded the supply of land to some extent. Consequently, the economy grew by mainly producing and selling agricultural products, and manufactures intensive in cheap natural resources. There were some costs to this growth. India’s traditional textiles declined because of British imports. Still, overall, the growth impulse proved stronger. Labour needed to move from sectors where it was unemployed to new pockets of demand. There was indeed a great deal of relocation of labour away from settled agriculture and handicrafts to new lands and new occupations such as plantations, mines, public works, new agricultural zones and migration overseas. In short, three factors—external and internal markets, extension of land and reallocation of labour—together produced a long unbroken phase of growth in production and income between 1865 and 1914. That whole process began to lose steam from early in the interwar period. The world economy was slowing, and land was running out of supply. The importance of land to economic growth can be seen from Figure 2.1, which shows the closely matching direction of two time-series: land and landed income. Given that landed income occupied about two-third or more of total income through the entire period, the figure tells us that an extensive growth occurred in the nineteenth century, which then lost momentum because the traditional resource on which it was based

Economic history and modern India

41

Figure 2.1 Land and real agricultural income (1865=100) (for source, see Figure1.1). was not growing any more. Population growth rate accelerated at the same time. Between 1891 and 1921, India’s population grew by about 15 per cent. In the next two decades, population grew by 50 per cent. Labour force grew at rates only slightly smaller. If we translate agricultural growth in per-capita terms, we observe signs of diminishing returns to labour, even land, in the interwar period (Figure 2.2). Further economic growth needed capital and/or efficiency gains, that is, steps that could lead to better use of scarce resources. Investment rates did rise in large-scale industry, which reflected not only the more advanced financial institutions this sector depended on, but also its access to technical knowledge and equipment arising from colonial contact. Investment rates could not so easily rise in small-scale industry which operated in highly imperfect capital markets. Strikingly, however, small-scale industry did experience significant efficiency gains, which probably derived from the transformation of less efficient family enterprises into capitalist workshops. Manufacturing incomes overall did not display the stagnation that characterized agriculture. By contrast with manufacturing, investment rates and efficiency-gains were weak in agriculture (Table 2.4). One possible clue to the problem was public investment, which slowed greatly in the interwar period. The nationalist story that India’s colonial status imposed certain constraints upon the government’s capacity to spend is cogent enough. It falls short though, for ignoring the issue of private investment and efficiency problems. Private investment was generally low, because of credit market imperfections and, less obviously, because of asset preference. The ‘hunger’ for gold repressed savings. Further, many contemporary

Rethinking economic change in India

42

Figure 2.2 Real yield per acre and person (1865=100) (for source, see Figure 1.1). Table 2.4 Sources of growth in real GDP, 1900–47 (% per annum) Agriculture

Non-agriculture

Growth

0.44

1.69

Contributions of labour

0.33

0.31

Contributions of capital (including land)

0.08

0.90

Contributions of total factor productivity

0.03

0.57

Productivity/growth (%)

7

34

Sources: Sivasubramonian (2000: table 7.21), Mukherjee, M. (1973). Contributions are products of base period factor shares and rates of growth of factor inputs.

observers also believed that the problem of poor economic growth was not that the production potential was not growing, but that production fell short of potentials. They attributed inefficient use of resources to such factors as poor education, caste hierarchy and diminishing return to land because of reckless expansion of cultivation.

Conclusion Since independence, the main factors in India’s economic growth remained those that are important in a labour-intensive society. A series of ‘green revolutions’ succeeded in breaking the barrier of no additional land, as the modern parallel to the irrigation projects of the early colonial period (see Figure 2.2). The relative abundance of labour, its

Economic history and modern India

43

allocation and mobility, continued to be of keen importance. After 1947, the marketization process for labour gathered strength in every sector, except in those under regulation. The persistence of surplus labour, increasingly arising from population growth, did not, however, permit a rapid rise in wages or earnings. In the long run, India has experienced a growth rate in average earnings consistent with its endowments, with or without colonialism. One key difference between colonial and post-colonial development patterns related to the strength of the external market. Post-1947 policy decisively weakened the stimulus of the world market. Colonial development was based on the classical principle that nations should grow by utilizing their comparative advantage, which often amounted to utilizing their endowments in a free world market. However, post-colonial development was based on an attempt to reshape the impact of resource-endowments with restrictions on trade. There were some gains and some costs involved in that shift. The biggest gain, perhaps, was a boost to local technological development. The cost was missed opportunities in trade. For India’s leaders at independence, looking back from the 1940s at the slow growth and collapsing international trade of the 1930s, the loss of international trade seemed of little consequence. However, the world economy greatly revived after 1950. Conceivably, if independent India had followed the colonial pattern of resource-led growth it could have achieved substantial growth in labour-intensive manufacturing, as it did so dramatically in the few years after trade liberalization in the 1990s. Conceivably, carrying over the colonial pattern of growth would have raised labour demand fast enough to generate strong upward pressure on real wages and to attract women to the labour market at a faster pace, eventually depressing population growth.

3 Rural labour and colonialism The proportion of wage-labourers in the rural workforce has more or less steadily increased in India in the last 100 years. This chapter investigates the reasons for the trend. Scholarship on the subject, extensive though it is, does not offer a clear explanation. Two general perspectives can be distinguished. One of these stresses increasing inequality, the other stresses increasing density of population and scarcity of land. In short, one stresses dispossession of land, the other stresses demographic pressure on land. In economic history, the accent has fallen heavily on the former approach. Increasing concentration of land-ownership, aided by a coalition of exploitative colonial and local power, it is argued, led to dispossession of the small peasants of land. Having put colonialism at the centre, the historical account lives uneasily with the fact that the trend continued, even strengthened, after colonialism. In studies dealing with more recent data, both density and polarization have been used, but in a casual way, and not always consistently. Some versions of the inequality argument tend to be invoked when explaining cross-section variation in the proportion of labour, but density is more usually cited in explaining changes over time. Can we discover an account of ‘proletarianization’ that is less offhand, and can bring the past and the present together into a consistent narrative? The present chapter seeks an answer to this question. The answer has two parts. First, proletarianization had many dimensions. And second, resource conditions, principally the balance between population and land, had a large effect on occupational shifts but not quite in a manner we might imagine. It concludes that inequality and colonialism fail to explain some of the key stylized facts, such as timing or gender-bias, in the trend. Indeed the history of rural labour cannot be reduced to a single dimension at all. There was indeed a kind of axis to the history of rural labour, namely, the growth of spot markets for labour and, connectedly, migration. That said, the transformation was a many-sided one, for different groups were affected differently by the growth of casual labour markets inside and outside agriculture. Peasants in some parts of India were pushed by scarcity and competition for land to depend on the labour market. Rural labourers who had earlier been part of the grainsharing arrangements became casual and wage-based labour. Rural labour that had previously engaged in diverse tasks tended to adopt agricultural labour as primary livelihood. Women experienced these changes, but in distinct ways. Women of peasant families who had looked after the family land tended to enter wage labour. Women of artisan families, when their husbands left home for wage-work in industries, joined the rural labour market. Scarcity of land, of course, plays an important role in this story, but not in a simple fashion. The relationship between land-labour ratio and participation in wage work was not a straightforward one because increasing density of population in land could have

Rural labour and colonialism

45

contradictory effects. Reduction in average size holdings could make the family just adequate, that is, reduce the need for wage labour. But further fall in holdings could make the family unable to sustain itself and increase the supply of wage labour. Descriptions of different regions confirm that in the early interwar period, both these movements were present, in different strengths in different regions. However, if one takes a longer view, the second effect probably emerged stronger irrespective of region. In short, the specific relationship hypothesized in this chapter can explain both observed cross-section differences in the proportion of wage labour in the rural workforce, as well as changes over time than either the inequality paradigm or a naive ‘over-population’ one. I begin with a broad-brush description of the ‘labourer’ as this figure appeared in reports from the early twentieth century. This section is followed by a commentary on measurement and explanation of proletarianization. The following sections review the evidence and offer a re-reading thereof. Profile of the labourer1 Conditions of labour in India varied according to region, ecology, crops grown, institutional or social backdrop. And yet, throughout the region, ‘labour’ conveyed a few invariable features. In simplest terms, the agricultural ‘labourer’ was a person who did not own or rent land. However, a great many among those who owned or rented land also performed wage-labour for others. This indeterminate margin between labour and cultivation posed a problem for occupational statistics. Marginal tenants and cultivators reportedly preferred to record themselves as ‘tenants’ or ‘peasants’ rather than as labourers. A further refinement in the above definition would involve recognizing that ownership of secure rights to use land was often a more important marker in rural India than clear rights of ownership. Thus, a peasant owning a small plot of land was often more vulnerable than a tenant with a secure right over a large plot of land. However, many of these shades of grey are not statistically tractable. The census divided agricultural labourers into a few broad categories, ploughmen, sowers and transplanters, weeders and reapers being the main ones. Usually, the same person might do all four tasks, except that sowing and reaping were seasonal tasks that required extra labour. Transplantation of paddy was generally women’s work, so was carrying the crop to the threshing floor. Other than that, sowing, weeding and harvesting were done by men and women, whereas ploughing was exclusively men’s work. Cotton and groundnut picking and tea and coffee plucking were done by men and women. Wages varied according to men and women, tasks, area and seasons. However, peakseason wages varied in a narrow margin. Money wages tended to be higher in irrigated land, in lands near cities, in the hills and in areas with high net emigration. To some extent, the money wage difference was neutralized by cost of living, as in the hills. Still, it seems that regions with higher money wages enjoyed higher real wages too. Women’s work had a few distinctive characteristics that changed rather slowly. In the early twentieth century, in nearly every context in which women worked, the work-site needed to be one where they could bring their children. Thus, ‘in the harvest season, the whole family set out together for the fields and the baby is strung from the branch of a tree watched by one of the elder children while the father and the mother are at work.

Rethinking economic change in India

46

Sometimes the mother arrives later than the rest of the family as she first sees to her household work. The mother will leave her work to give her child nourishment, and in the midst of the day they all gather under the shade of the tree to rest and have their meal’.2 For similar reasons, it was difficult for women to move away from the homestead for long periods of time during the day, or to be on the beck and call of a master. Therefore, relatively fewer women took part in tasks that required long hours in the fields, or in ‘farm-servant’ arrangements where the labourer agreed to a beck-and-call relationship. In short, women’s participation in wage labour, and the prevalence of casual labour, were correlated. A second prominent feature of women’s work is that women took part in long-distance migration, even short-distance internal ones, to a much smaller extent than men. Descriptions of mill workforce suggest that women sometimes stayed back to look after the family plot of land while the men left home. But such an explanation cannot be easily generalized to the labouring poor, nor is it all that well-substantiated for peasant groups. Other factors were present too—the high cost of living in the city, or just the fear of alien environments, for example (see, for example, Bhatty, 1929:90). A third feature of women’s work is the possible presence of a backward-bending supply. The evidence on this point is too fragmentary to assert how often, when and for whom such a mechanism really worked. But references do appear, usually for small peasant families that took part in wage-work on the side, that women working for a wage represented a mild distress situation, whereas whenever family income increased because of an exceptionally good season, there was some voluntary withdrawal of women from wage labour.3 Once again, it is hard to assert whether this was an exception or the norm. Migration, internal as well as to British colonies elsewhere in the world, was a powerful influence on the labour market in certain regions of colonial India. Several prominent streams of short-distance migration emerged from the late nineteenth century. Large numbers migrated every season from the uplands of the Godavari, Krishna, Guntur and Vizagapatam districts to the Krishna-Godavari deltas for farm work. Many from Chhattisgarh went every season to Berar cotton fields and gins. The wheat field of Narmada Valley, that is, Jubbulpore, Saugor and Damoh, received migrants from two directions, UP in the north and from Rewa, whence the Gonds descended the hills every season. Bihar workers migrated to Bengal at jute harvesting. Azamgarh workers were recruited for larger-scale earthwork in Bengal. From Ratnagiri, many went to work in the cotton fields of Broach. Santals from Damin-i-Koh ‘made their appearance’ in the plains of Western Bengal ‘wherever manual labour is wanted, [building] leaf huts in a few days, and before the end of the month [making themselves] as much at home as if [they] were still among the mountains’ (Hunter, 1872:205). And Punjab canal colonies received migrants from Rajputana. Agricultural labourers also went to the plantations, urban services, railways and other public works and, more rarely, the mills. Being seasonal and short-distance, few such movements were reflected in the census migration statistics, which was better able to capture net migration of more or less permanent kind. That the migrants enjoyed a better living standard than the norm in their places of origin is clear enough, despite the many known abuses that they sometimes had to go through at destination. However, just how much they were better off can be argued. Internal and seasonal migration did not necessarily mean access to higher wage rates. It did mean steadier employment or access to more days of work, and sometimes a better

Rural labour and colonialism

47

portfolio of occupations for the family. At least in the nineteenth century, a second factor associated with emigration was probably more powerful than wages: bad season and the prospect of a famine. Throughout the second half of the nineteenth century, as demand for migrants developed, and railways were ready to take them to the ports at a small cost, every bad season drove hordes of people out of their villages. A great deal of overseas migration, especially for South Indian labourers going to Ceylon or Burma, was of seasonal and short-term nature too. But, with longer duration and long-distance migration, the elevation of economic and social status was much more marked. Returnees were known to make a powerful impression at their places of origin. Thus, it is likely that over time the wage incentive became a more powerful inducement than the prospect of starvation in the village. Wage ‘contracts’ in the early twentieth century fell into three broad classes. There were casual wage earners usually paid by day in the peak season or paid by task irrespective of season. There were workers on long-term contract. Those bound by seasonal, annual or longer duration contracts were called ‘farm servants’ in the census, and those under the former type called ‘field labourers’. Those called ‘labourers’ in colonial sources sometimes included a third set of people—workers who were paid by a share of the crop. If such workers did all the work and brought their own equipment, this system was indistinguishable from a crop-sharing tenancy. Not surprisingly, therefore, labourers and tenants converged along a wide frontier, and there were frequent movements between the two categories in official statistics. The problem of equivalence of tenancy and labour probably compounded in older long-settled densely populated agrarian zones where tenancy was particularly prevalent. By contrast, in areas of new expansion, class structure tended to be more sharply polarized between landed owners and landless workers. Farm servants received a large part of their payment in grain, but this custom was probably weakening in the nineteenth century. Atkinson (1902), for example, believed that money payments were the rule in rural India at the time of his work on national income. Farm servants were internally quite mixed. There were some among them more or less like serfs in the mid-nineteenth century. That is, they were tied to particular families and plots of land for life, and supplied not just field labour but a variety of labour services for their masters at any time the latter needed these. In extreme cases, they were in effect sold when the land was sold. In some contemporary writings, these labourers were called ‘permanent’ farm servants. This form of serfdom was prevalent in Madras where certain castes were constrained by their position in rural society to work as servants. Elsewhere, the relative positions of the master and the servant were not very different. In Punjab, the farm servant earned larger incomes than many small cultivators, especially if the land bore prospects of profitable garden crops, which demanded experience and daily care. In Central Provinces, the farm servant contract was frequently renegotiated, and remained in force for various durations—the longer the duration, the greater the incentive. The Lamsena, who accepted a three-year contract, was promised the master’s daughter in marriage at the end of the term (India, 1931c:213). Despite the long duration of these relationships between individuals or families, wages were paid monthly and almost certainly not for the whole year. Why were farm servants needed? The usual answer is the need for ‘labour-hoarding’. In agrarian systems subjected to pronounced seasonality and negligible seasonal

Rethinking economic change in India

48

migration, a shortage in the peak season was always possible. To avert such prospects, employers were willing to offer annual contracts, the wage paid for doing nothing in the off-season imposed a smaller cost than the loss incurred in the peak season because of the absence of a worker. The view that permanent contracts reflected perceived labourscarcity, in which case tying labour to a plot of land made sense, has been advanced in the Indian case.4 What is curious though is that the seasonality motive for labourhoarding was present everywhere in India, but the incidence of farm servants sharply variable. Labour-hoarding as an explanation may work better for Punjab, in which region land-persons ratio was above average, increasing (1891–1931) and where many among the labourers were servants. But it is less convincing for Madras, where the proportion of servants was high but land-persons ratio was small. The caste-hierarchy aspect probably had a strong influence on the matter. While labour-hoarding captures the demand for farm servants well, social structure might account for the extent of supply, at least in some regions. In some cases, seasonality apart, steady demand for labour was high enough to necessitate labour-tying. This is reflected in the long average working hours of labourers in regions. Sometimes demand for labour derived from the quantity of livestock. Atchi Reddy (1991) finds that in the mid-twentieth century, annual farm servants in coastal villages of Andhra Pradesh slept in the cattle shed over night so as to feed the animals several times during the night. Multiple cropping in deltaic Andhra also raised demand for labour. In all such cases, both migration and labour-tying continued side by side as twin strategies to meet shortages. And yet, in the aggregate, farm servant arrangement was in general decline from the early twentieth century. So were wages in kind and general labour services. Early works on Indian agriculture uniformly reported serfdom and long-term contracts to be in decline, but the decline was in some cases located around the end of the nineteenth century, and in other cases early in the interwar period. In Madras, which had long been the principal concentration of long-term contracts, the proportion of farm servants among labourers fell from half to about a third. It declined further to just 1 per cent in 1951 (India, 1954). Until the 1950s, the distinction between ‘casual’ and ‘attached’ labourers was maintained in official statistics, but the term ‘attached’ was discarded thereafter on the ground that ‘attached labour is no longer attached to any particular household in the old sense. Such attachments are now conditioned more by economic considerations and may not extend beyond a season or a year at the most’ (India, 1973:65). In the aggregate, the percentage of farm servants in 1901 or 1921, and the percentage of attached labour at 1951 did not change very much. The proportion moved in the range of 12–15. And yet, unquestionably, long-term hierarchical or patron-client relationships of the older kind had declined, and a different form of attachment had risen in importance. The ‘attached’ labourers of the 1950s were a deceptive group. The abolition of zamindari tenures led many former landlords to take up cultivation, which meant a reclassification of their tenants as either casual labourers or attached labourers. That factor apart, many others among the attached labourers belonged in a different relationship from the farm servants. At c. 1900, it was often the case that the farm servants ‘get into debt to their masters, and as it is usually impossible for them to repay it, they become hereditary bond servants and their sons succeed them’ (India, 1931c:213– 14). In the mid-twentieth century sources, the so-called ‘attached’ labourers were usually

Rural labour and colonialism

49

debt-servants. The Dublas and Kolis of Bombay, the Halis of Gujarat, the Kamias of Bihar and the Goti of Orissa figured to some extent among the farm servants of 1901, but were bonded by debt at c.1960. Debt bondage was usually, though not always, a legal and documented contract. The farm servant relationship too was sometimes backed by paper though our knowledge about these documents or usage thereof is very inadequate.5 The transition from farm servant to debt-servant was not a smooth one. Both forms existed side by side for a long time, especially in the Tamil-speaking regions.6 If one takes a sufficiently long view, spot market for labour or casual labour increased, and long-term contracts decreased. At least in some cases there was probably a graduated transition, that is, the duration of the contract progressively shortened. As Atchi Reddy (1983a:67–79) has observed, some annual farm servants became transformed into seasonal farm servants. More generally perhaps the farm servant contract and the spot market co-existed and competed, and the latter won. Why did long-term relationships give way to casual ones? Later research which observed and studied this institutional change in rural India offered a range of explanations, including population growth, changes in crop mix, migration, attitudes, politics and legislation (Breman, 1974; Gough, 1981). Not all these variables can be generalized to all regions of India, or to the early twentieth century. I believe we can be more specific about the most important historical factors. Migration is likely to top this list. How did migration work to dissolve long-term relationships? In the economic history of labour, the transition from labour-tying to spot markets has often been explained with reference to technological change. In manufacturing, technological change enabled substitution of skilled artisans by relatively unskilled workers so that the ease with which one worker could be replaced by another increased (Grantham, 1994:8). In turn, workers could be hired from the casual labour market. Agriculture does not furnish an exact parallel. Still, some kind of a homogenization was indeed going on. For example, in the nineteenth century many farmers took up cotton, wheat or groundnut partially replacing the local consumables that they had produced before. Local knowledge having become less valuable now, these farmers could hire more outsiders than locals. At the same time, however, it was the decreasing cost of seasonal migration that enabled them to access such outsiders. Migration signified the opportunity of hiring from the market as opposed to reliance on customary relationship. It is easy to see that the very co-existence of migrant casual labour and permanent servants in the same area would bring customary ties to an end. Any wage difference between these two forms of labour will induce either the servant or the master to terminate a customary contract or attempt to enforce changes in it. I suggest in the next chapter that the gap between casual labour wages and farm servant wages widened in the late nineteenth century, encouraging attempts to leave the latter contract. The transition from farm servant or debt contract to casual labour had a gender implication. Farm servants were predominantly male. In 1900, 80 per cent of farm servants were male. Casual labourers, whom the census called field labourers, had equal proportion of males and females. In 1900, exactly 50 per cent of field labourers were women. Therefore, a change in composition alone contributed to a relatively greater presence of women among agricultural labourers. Given that women could work only part-time, casual labour was a more convenient contract for women than the farm servant contract.

Rethinking economic change in India

50

Which community did the labourers come from? Descriptions of labour communities can be found in reports on agriculture as well as in reports on industry. These reports suggest that in the peak season, wage labour could draw individuals from a wide range of occupational groups, castes and communities. However, two groups deserve particular mention. Both were prominent among agricultural labourers at the end of the nineteenth century, and both played central role in changes in labour institutions. In certain parts of South India, chiefly Malabar, South Canara and some areas of Tamil Nadu, were men (they were almost always men) of specialized labour castes tied to particular plots of land. They could be ‘sold’ when the land was sold. At the turn of the century that transaction sometimes meant a transfer to the new employer of the accumulated debts of the labourer to his old employer. Such ‘serfdom’ and specialization were present in other regions too, with difference in detail. The second type was mixed occupations. Many poor villagers pursued a portfolio of occupations, of which labour was just one. Specialization and attachment to plots were more often features of permanent contracts, whereas portfolio characterized the casual labourer. There was constant entry into the second kind of labour from some of the poorer artisan groups. The description of Chamars, whose principal occupation was tanning, is representative. The 1901 Census reported the Chamars of Bihar to have been engaged in, along with tanning, agricultural labour, ‘provision and care for animals, menial service, dealing in food and drink, weaving, working in metals, wood, glass, stone, canes, bamboos, leaves, etc.’ (Chandra, 1903:19). Again, on Khalpas (tanners) of south Gujarat: ‘Each village has its own Khalpas, who as village servants hold lands at one-half the ordinary rental. As a class they are poor, living from hand to mouth, and adding to their earnings from tanning by cultivating a little land or working in the field as labourers. They show few signs of improvement or progress’ (Martin, 1903:29). For such people, the early censuses used the term ‘general labour’. Both these types were transforming rapidly in the colonial period. Specialized and ‘tied’ farm servants, for example, left the village in large numbers to migrate overseas or move to other jobs. They came from the so-called ‘depressed’ classes, and played a leading role in the reform movements. Some of those engaged in what the census called ‘combined occupations’ were part of the village customary service arrangements wherein a share of the grain went to the artisans in exchange for year-round service. But these old arrangements were breaking down, making way for casual wage contracts for specific tasks. Also, among the rural artisans occupational diversity was declining, there was increasing specialization in labour. This is the reason why the census gave up the ‘general labour’ category after 1901 and the term ‘combined occupations’ sometime after that.

Measuring and explaining change In 1881, the workforce in India numbered 101 million, of which sixty-four million were engaged in agriculture (more precisely, cultivation). In 1931, the numbers were, respectively, 125 and eighty-two. There was, thus, an increase in the share of agriculture in the workforce. Simultaneously, there was an increase in the share of agricultural labourers in the workforce, in one estimate, from 15 to 30 per cent. In an earlier scholarship, these two trends were seen to be connected, and illustrative of the economic

Rural labour and colonialism

51

retardation unleashed by colonial policies. The first seemed to vindicate the nationalist thesis of ‘de-industrialization’, or a progressive shift of workers from declining industry to agriculture. The second suggested a progressive dispossession of the peasantry of land, or ‘de-peasantization’, a thesis embraced and refined in the later Marxist-Leninist perspectives on Indian agrarian history.7 Via these theories, a few simple percentages derived from the not-too-reliable British censuses became pivotal in interpretations of the link between colonialism and underdevelopment. Colonial India saw the beginnings of a modern economy with a larger role for capital, information and the market than in earlier periods. How did these changes alter the economic conditions of ordinary people? Answering this question led one closer to tackling the classic colonialism problematic: whether poverty was an outcome of colonialism or an outcome of structural conditions that the modernization process was too weak to change. In what became the widely accepted paradigm on rural labour, the effect of economic modernization on the peasants was seen to be an adverse one. In this view labourers, defined as workers in agriculture who neither owned nor rented land, represented a new form of relationship, and an expression of poverty and inequality introduced in colonial India via British trade and taxation policies. ‘Depeasantization’ and ‘de-industrialization’ were symptoms of this process of the ‘development of underdevelopment’. Like many other themes in Indian economic history, the one on agricultural labour withered away in the 1980s as the empirical project was captured by a rhetorical question: did markets in a politically unequal world give rise to exploitation? Thereafter, scholarship polarized into two disconnected domains: one holding on to a link between labour and distress, and the other taking a more empiricist approach pursued through regional histories. With ideology taking over, the factual content of the debate remained narrow. In the end, it was no more than a noisy argument over a few percentages culled from a few columns in census occupational statistics. Trends in wages were never seriously considered a relevant variable. There was a great deal of casualness in dealing with time and space. Gender aspects were ignored even though a gender angle was palpable in the most elementary reading of the census statistics. And, non-agricultural work was not investigated enough even though a significant part of agricultural labour was known to have combined occupations. As the macro-picture remained too simplistic, regional histories seemed increasingly at odds with the macro-picture. In the first significant attempt at formulating a general narrative on the Indian economy under British rule, S.J.Patel developed an interpretation of the conditions of rural labour between 1872 and 1931. This influential monograph proposed that agricultural labour was a result of disintegration of the ‘largely self-sufficient village communities in which the cultivators and the artisans had lived together for centuries on the basis of traditional arrangements regulating [the] exchange’ (Patel, 1992:47). The specific changes that caused disintegration in rural economic life were destruction of the artisanate because of cheap import of manufactures, cash rent and the need of the cultivators to borrow, world price fluctuations, domination of moneylenders and the creation of markets in land and land-mortgage. In short, the market-oriented economic policies and class-biased revenue policies of the colonial government ruined the intrinsic balance of the rural economy and led to large-scale pauperization. It is not clear whether

Rethinking economic change in India

52

this mechanism leads to immiseration or to polarization. The processes are compatible with either perspective.8 Patel’s thesis had two parts. The first part was factual, and consisted of the demonstration that the percentage of labourers in the agricultural workforce (that is, cultivators plus labourers) increased between 1881 and 1931. From now on, I call this percentage the ‘labour-ratio’. Trends in two labour-ratios—fall in industry and rise in agriculture—were seen as symmetric and causally related. Artisans were constrained by inadequate demand, leading to transfer of population from industry to agriculture and, in turn, agriculture was seen as receiving labour. The second part was the argument. In this part, the rise in labour-ratio was attributed to de-industrialization and the peasantry losing land. The critique of the immiseration-polarization thesis too had two parts. J.Krishnamurty, Daniel Thorner and Alice Thorner questioned the fact of a definite trend in labour-ratio.9 These authors subjected the occupational data of the first six British censuses to close scrutiny, and came to the conclusion that occupational classes amounted to an underestimation in rural labour force in the earlier censuses, and changes in classification introduced an upward bias in the labour-ratio. After their revisions, the extent of the rise in labour-ratio became smaller. But a rise remained. Dharma Kumar, on the other hand, questioned the argument (Kumar, 1965). The thrust of the latter critique was upon whether or not agricultural labour was a ‘creation’ of British rule, and the conclusion was that it was not. That is not to say that labourers were unaffected by economic changes in the nineteenth century. Indeed, the real message of this work was that the labour-ratio was not a sufficient index of changes in conditions of labour. More important, forms of labour changed, from varieties of bondage, attachment, serfdom and customary service towards greater formal contractual freedom. In later research, the key analytical building blocks of the immiseration thesis lost some of their force. In different spheres of scholarship, and some-what independently of one another, the relevance of de-industrialization, the idea of village community, marketless exchange, peasant indebtedness and moneylender power were rather strictly qualified. Indeed, too strictly qualified to be easily utilized in a narrative of crisis. This whole debate has the labourer somewhat behind. Gyan Prakash has recently argued that it allowed the narrative to be framed in colonialist terms without questioning those terms.10 That is, the debate allowed the study of the economic history of labour to be reduced to an appraisal of colonialism. Furthermore, neither critique touched upon or disputed the proposition that the ratio of labour in the agricultural workforce increased in colonial India and that this might validate Patel’s argument in a qualified form. A similar claim made by later proponents of the immiseration thesis seemed justified (Patnaik, 1983). To take a step forward, it is necessary to re-examine the factual content of the labour experience as well as reconsider the fit between labour-ratio on the one hand and the narrative that it is supposed to verify on the other. Surely, the attempt to discover a grand narrative on labour must go beyond the labour-ratio. It should at least look more closely at other unattended factual details relating to timing, gender, region and real wage trends. Further, the fit between the fact and the theory is not necessarily close. Labour-ratio can increase as a result of demographic pressure on land, rather than dispossession. Equally, an unchanging labour-ratio does not mean dispossession did not happen. A closer look at

Rural labour and colonialism

53

labour-ratio itself reveals some of these nuances and the difficulty, therefore, of a straight-forward reading. The evidence cited for the immiseration thesis was a steady rise in the percentage of labourers in the agricultural workforce. Patel compared labour-ratio for four censuses— 1901, 1911, 1921 and 1931—and showed that the labour-ratio increased, from 19.5 per cent in 1901 to 36.3 per cent in 1931. Patel also considered the late nineteenth century censuses, and the conclusion seems stronger. This conclusion was criticized on three main grounds. First, in the master table that Patel constructed, 1901–11 referred to ‘population supported’ and 1921–31 to ‘actual workers’. While the years within these two pairs were comparable, 1901 could not be compared with 1931. Second, Patel did not distinguish between male and female labourers. When disaggregated, ‘proletarianization’ appeared more pronounced for females and more subdued for males. Third, labour-ratio for the 1881 census was underestimated in Patel. If this is revised, the extent of rise in labour-ratio between 1881 and 1931 turns out to be much smaller. Krishnamurty (1972) used the ‘actual worker’ dataset, revised the 1881 figure, and distinguished male and female participation. The results do not show proletarizalization for males. A revised set of figures in Krishnamurty (1983a) again did not show an increasing trend in labour-ratio for males, though the 1931 ratio was higher than the 1901 figure (Table 3.1). This whole scholarship neglected a few dimensions of the occupational statistics. First, these processes had a gender bias. It was mainly women who experienced a large shift away from manufacturing (and services) towards agriculture between 1901 and 1931. The labour-peasant ratio too increased dramatically for women, more modestly for men. Second, 1931 appeared as a kind of watershed in the census dataset, suggesting a specific connection between the Depression and rural labour. Third, the trend had a regional bias in the colonial period, leading us to ask what factors made the regions different. Fourth, the rise in the proportion of labourers in the agricultural workforce continued into the post-colonial period, which at

Table 3.1 Labour-ratio, 1881–1931 Patel (persons)

Krishnamurty Men

Women

1881

15.0

27.1a

44.2a

1901

19.5

14.3

30.2

1911

20.5

15.4

32.5

1921

24.8

14.4

28.0

1931

36.3

19.5

43.8

Sources: Patel (1952), Krishnamurty (1972,1983a). Note a These figures come from an earlier work of Krishnamurty that was later revised, and may not be comparable with the figures for the later years.

Rethinking economic change in India

54

least questions, but does not necessarily dispute, a link between British colonial policy and rural distress. We begin with gender. Between 1901 and 1931 labour-ratio for both sexes increased. But it increased faster for women than for men. Overall, the ratio changed from 12 to 15 per cent of the agricultural workforce for men and from 23 to 36 per cent for women. For women, the ratio increased uniformly in all regions (Table 3.2). In UP, Bombay, Punjab or Central Provinces, male labour-ratio changed little, but women’s labour-ratio rose significantly. Analysts of occupational data treated women’s data with caution, even scepticism. Indeed, women tended to combine occupations, or were non-specialized, rendering the workforce data for women difficult to interpret. But is that sufficient argument for ignoring the data altogether? Suppose labour market institutions changed over time, and a rising proportion of labour among women workers captured that trend. If we ignore women’s data, we might then overlook a fundamental transition in rural labour markets. Why might 1931 appear as a sharp break from the past? Some caution is needed in interpreting the 1931 dataset. The 1931 data reported figures on ‘earners’ and ‘working dependents’, a subset of which was comparable to the ‘actual workers’ of the previous censuses. Tables in this chapter use only the figures for ‘earners’. Now, in that census year, there was conversion of earners into working dependents among the peasants, but not among the labourers. This conversion tends to inflate the labour-ratio if only earners are considered. That being said, this conversion in itself is a significant stylized fact. It illustrates well my point about increased vulnerability of the peasantry in that year. The 1931 census itself did observe the ‘somewhat remarkable’ increase in labourratio, but did not attempt a serious explanation. Other authors attributed it to the particular conditions created by the early years of the Great Depression (Joshi, 1958). The two noteworthy circumstances in 1930–31 were: (a) a crisis for the small farmers who were squeezed on two sides—high real value of their debts as soon as commodity prices crashed, and high real wages—and (b) gain for the labourers arising from high real

Table 3.2 Labour-ratio of women (%), 1901–31 1901

1931

Bengal

29.0

40.9

Punjab

5.1

19.8

UP

25.4

49.7

Bombay

30.4

79.8

Madras

33.0

69.3

Central Provinces and Berar

52.2

83.1

Source: Census of India.

wage. For, temporarily, money wages were slow to adjust to the massive fall in commodity prices. These gains for the wage-earner were wiped off in the drastic adjustment in money wages which began to occur from about 1932 and continued for several years. It is likely, however, that in the census year, there was a large-scale but

Rural labour and colonialism

55

temporary conversion of small peasants into labourers to supplement their reduced surplus. An indirect confirmation comes from acreage cropped data that show a fall in 1930–32 in almost all the major provinces. There was peasant distress involved here no doubt. But there was also greater attraction for labour, which received a windfall. That being said, a mystery remains about 1931. Comparing the 1901 and 1931 censuses by sex and broad occupational classes, no sharp decline emerges among male cultivators, in fact the number of cultivators and workers both increased among men. The peasant-to-worker ratio remained roughly constant at 80:20. The one sphere that suffered dramatic contraction was women cultivators. The absolute number of women peasants fell by as much as ten million, resulting in a sharp rise in peasant-worker ratio. If there is a proletarianization story here, that story may run as follows: the 1931 shock drove particularly vulnerable segments among cultivators into labour, women were predominant in these segments, and a vast number of those thus driven into labour were unable to return to their own cultivation. What was the mechanism by which this shift happened? The year 1931 was not a good time for land transfers. One plausible speculation is that a number of male migrants returned from the cities to the villages and took up cultivation that was until then being done by women. In turn, those same women became labourers. Let us now consider region. At the all-India level, labour-ratio for men increased by 3 per cent between 1901 and 1931. But in two regions, the ratio for men increased to a much larger extent in the early twentieth century. Of these two regions, in Bombay, there was only a spike at 1931 but otherwise no significant trend. But in Eastern India, there was a sustained rising trend (Table 3.3). To explain the regional differences, we need to look more closely at land scarcity or abundance across regions. Studies that claim a large increase in labour-ratio tend to stop at 1931. But 1931 figures were unusually high and may not be representative. What happened in the last twenty years of colonial rule? There was no census in 1941. The 1951 census returns a small labour-ratio. However, there is reason to believe that the 1951 census underestimated labour-ratio by not including all the landed labourers as ‘labour’ (Joshi, 1958), and by under-estimating the incidence of labour among the dependents of tenant and small-holder families (Thorner and Thorner, 1957). The Agricultural Labour Enquiry in the same year returned a higher percentage. This percentage refers to not workparticipation of individuals but the proportion of labour families in rural families. In principle, it is possible that this percentage overestimated the labour-ratio, for (a) labour families might

Table 3.3 Labour-ratio by major region, 1901–51 (percentages) 1901 India

a

1931

1951

23.8

28.8

30.4

Bombay

30.0

50.0

20.0

Madras

34.2

38.2

41.0

Central Provinces and Berar

47.9

55.3

40.0

Rethinking economic change in India

Punjab United Provinces/Uttar Pradesh

56

4.1

14.0

10.1

17.3

20.9

14.3

9.5

2.1

11.0

13.9 (10.5)

26.5 (23.5)

–

8.3 (7.3)

26.2 (26.0)

–

10.0 (9.0)

37.7 (36.7)

34.0

19.4 (14.6)

26.6 (20.8)

49.0

6.0 (5.0)

28.3 (22.6)

51.0

Eastern Region (labour-ratio for men shown in brackets): Assam Bengal, Bihar, Orissa Bengal West Bengal Bihar Orissa Sources: Census of India, various years, and India (1954). Note a For 1951, territorial units refer to the Indian Union and states within it.

include some non-labour earners, and (b) the size of labour families was smaller on average (that is per family, labour families had fewer workers). There is no easy way to test (a), but the prospect is unlikely because labour families had smaller alternative opportunities. The second possibility does not hold, for labour-families had much higher worker-dependent ratio than the other types of rural families. Correspondingly, labour families had higher rates of women’s work participation (India, 1954). Later surveys also confirm that proletarianization in fact continued in the 1950s and the 1960s (Bardhan, 1977). It is likely, therefore, that proletarianization was a persistent tendency between 1931 and 1951. What happened after colonial rule? Both the immiseration thesis and the Krishnamurty critique coexist uneasily with very clear indication of a significant and steady proletarianization in post-colonial India. The trend continued, even strengthened, between 1951 and 1991 (Table 3.4). The point that proletarianization was a fiction created by census definitions does not have validity in the long run. Nor does the persistence of the tendency allow us to connect colonial rule with proletarianization. The trend accelerated from around the Depression, or late in the colonial period. And throughout, the trend was strong among women. The literature on the post-1950 trend has had to grapple with its own database problems, in this case a persistent discrepancy between the two principal sources of employment data: the census and the National Sample Survey Organization. In turn, the discrepancy relates to a choice between stricter or broader definitions of ‘worker’. But these problems affect the level of labour-ratio, rather than the trend. There is no dispute over the

Rural labour and colonialism

57

Table 3.4 Labourers and labour-ratio in the Indian Union, 1951–91 Agricultural labourers (millions) Male

Female

Labour-ratio (%) Total

Male

Female

Total

1951

12.7

9.8

22.5

18.9

30.5

20.9

1961

17.3

14.2

31.5

20.7

30.0

24.0

1971

31.7

15.8

47.5

33.1

62.9

39.3

1981

34.7

20.8

55.5

30.9

58.2

37.5

1991

46.2

28.4

74.6

34.3

56.1

40.3

2001

57.4

50.1

107.5

39.9

54.8

45.7

Notes Sources for all figures are the Indian census, occupational statistics. The 1951 census workforce data are not reliable (see text), and should be treated as outliers. Scholars tend to agree that 1961 figures are comparable with the 1981 ones.

increasing trend in the ratio. The explanations, however, remain speculative. In the 1970s, it was suggested that tenancy legislation and technological change had induced land-owners to take a direct interest in land, so that leasing out declined, and the former lessors turned labourers. Along with increased supply for this reason, there was increased demand for hired farm labourers in the regions that experienced rapid agricultural growth (Bardhan, 1977). But these factors were circumstantial, and the trend of a rise has clearly proven long-standing. A few papers that were around at this time discussed the idea of a backward-bending supply curve for family labour, which might account for increased substitution of family for hired labour (Krishna, 1976). The idea did not persist, however. It implied a significant rise in income of the peasant family farm. Such a trend could be generalized to a small subset of the family farms. By and large the pure family farm tended to be marginal and continually in crisis. That leads us to the second major explanation. The number of uneconomical or barely economical farms was rising rapidly throughout the latter half of the twentieth century. Many scholars identified this tendency, and the pool of marginal farms, as the force behind the entry of peasants into labour. But the account for marginalization tended to be somewhat expansive. For example, ‘Along with…deterioration in land-man ratios …there have also been differentiation of the peasantry and polarization of the class forces’ (Jose, 1989:18). A recent work on the subject described marginalization in great detail, but stopped short of raising the politically incorrect theme of ‘overpopulation’ (Jha, 1997:19–28). Rather, the implied explanation for marginalization, based on a crosssection study, seems to be inequality in landholdings. If this sounds ambivalent, the answer to the question why the female labour-ratio rose faster is equally so. Density and polarization have again been cited. Bardhan (1977), for example, suggested increase in rural disparity, combined with the assumption that poorer households had higher female work-participation rates. And a more recent work (Jose, 1989:16) suggested that marginalization raised the female work-participation rates. It is

Rethinking economic change in India

58

not clear from this literature the precise ways in which the male and the female workparticipation rates might differ.

Land and labour: the logical relationship The persistent increase in the labour-ratio, and the regional variations, point to a structural rather than a political explanation for the trend. A balanced interpretation needs to consider demand for and supply of resources. Demand for labour cannot adequately explain rapid conversion of peasantry to labour, for real wages of labour were depressed, even fell, between those pairs of years when the largest increases occurred. The years 1951 and 1971, for example, saw a rise in labour-ratio but no significant increase in wages. I suggested above that conversion between 1921 and 1931 may have been partially driven by the attraction of higher real wage, but the effect was too short-lived. We are therefore driven to investigate the relationship between land and labour. The hypothesis is that increasing density of population leads to fragmentation as well as diminishing returns to land and weakens the marginal peasant, eventually turning him or her towards labour. This is a long-standing theme in Indian scholarship, and yet it has always remained in the wings rather than taking the centre stage. There is a reason for this. The correlation between land-persons ratio, or the inverse of population density, and labour-ratio has been found to be weak in cross-section data and regarded as unimportant, making way for class relations to take over the analysis.11 In fact, a weak correlation is not surprising. Why should we expect the relationship between density and labour-ratio to be linear in the first place? Consider the following. In 1950, Bengal had a high labour-ratio because the average size of a plot of land under occupancy by a peasant family was so small that the plot could neither engage the available family labour time nor provide enough subsistence. Nearly every family had excess labour time available for hire, and many reported themselves to the census as labourers despite owning some land. In 1900, the Narmada Valley, Vidarbha, Khandesh, or the ‘Agency tracts’ in coastal Andhra had high labour-ratio because there was unoccupied land available even as the wheat/cotton/sugarcane/tobacco boom had increased the return to land.12 The average farm size was large, and these lands needed migrant labour in large numbers. Between these two sets of cases, both had high labourratio but very different population density. The latter regions often reported an increase in labour-ratio as the agrarian expansion progressed. Given this scenario, any standard test of association between density and labour-ratio is bound to give a poor result. Yet, as these two examples suggest, a poor correlation does not mean an absence of a relationship. As the two examples suggest, the expected relationship should be U-shaped as shown in Figure 3.1. At relatively high land-persons ratio, the average farm requires to hire labour since family labour is relatively scarce. At some middle range of the ratio, family labour is adequate for the average farm size. At significantly low land-persons ratio, family labour is on offer as hired labour (or lives largely by labour). Further, at such range, the peasant family farm suffers from high insecurity. Its low asset base reduces creditworthiness. Small adversities, then, can lead to a transfer of the rights to cultivate.13

Rural labour and colonialism

59

How the surplus labour in a land-scarce regime finds employment is a question that cannot be easily answered. Several scenarios are possible according to local institutional structure. In paddy areas, for example, it is common to see the adult male member of a household cultivating the family plot throughout the year. But the female members who work in certain peak season operations that are also labour-intensive, work in the peak season both as hired labour and on their own farm. In such a scenario, the men are likely to record themselves as peasants, but the women as labourers. This U-shaped curve may not remain U-shaped if there is uneven adoption of new technology between regions. It is easy to imagine that a uniform increase in cropintensity arising from irrigation or new seeds, or what Hayami and Ruttan (1971) call ‘land-saving’ innovations, could shift the curve to the left. The smallest farms will then experience increased labour-absorption and reduce their supply in the wage-labour market, while the largest farms will increase demand for hired labour. A uniform adoption of agro-mechanical technology or labour-saving innovations could shift the curve to the right. Such technology worsens underemployment in the smallest farms but reduces labour-demand in the largest. A famous relationship hypothesized by Shigeru Ishikawa sequenced these two types of innovation—the labour-intensive ones preceding the labour-displacing ones. Across regions, however, the technological mix is likely to be partly driven by ecology and, therefore, to be uneven. The lesson is, technological change can confuse any pre-existing U-shaped relationship. Similarly, interstate migration can affect the relationship too. The basic relationship and these main possibilities are shown in Figure 3.1 below.

Land and labour: the empirical relationship Any reasonable scatter diagram trying to capture the association between land-persons ratio or the inverse of density of population and labour-ratio for the colonial period suggests that the relationship was a U-shaped or a V-shaped one. This can be seen in the set of figures presented in the appendix to this chapter (see pp. 83–5). The appendix figures that show the U-shape most clearly are the pooled cross-section and time-series dataset for 1901–31, the cross-section pictures for 1931 and 1951, and some of the major regions over time. The picture, however, begins to get confused after 1961. By the above reasoning, technological change was one of the factors that destroyed what once was a regular pattern

Rethinking economic change in India

60

Figure 3.1 The hypothesized relationship between density of population and labour-ratio (for source, see Figure 1.1). We need to proceed cautiously on reading this dataset, especially since the regional units are far too large. In the 1901 cross-section picture (omitted here), the U-shape shows up again, but the interpretation is not straightforward. At one end, indeed, was Central Provinces and Bombay with lower-than-average population density but higherthan-average labour-ratio. In the middle was Bengal, UP and Punjab, with density close to the all-India average and small labour-ratios. And Madras had relatively high average density of population, and high labour-ratio. So far so good. But Madras did not have high labour-ratio because of surplus labour arising from within the peasant family, but because of high caste-occupation correspondence. This instance suggests to us that the cross-section picture, whether or not it conforms to the expected relationship, is not easily readable because regions can be socially and culturally unique too. However, the over-time picture falls more readily into the rationale for a U-shape that I suggest above. The labour-ratio increased in Punjab as population density fell in the first quarter of the twentieth century. In Eastern and Northern India, on the other hand, the labour-ratio increased along with population density. By 1931, in most regions, the two ratios were moving in the same direction. Density was high enough in most parts of India to push a large number of peasants to the margin. In nearly all of India, including the prosperous districts of Punjab, the adversity factor was active in 1931 in pushing numerous small peasants, the group that Malcolm Darling called ‘one-plough proprietors’, towards labour. Elements in the above rationale for a U-shape appear in contemporary rural economics discourse. Radhakamal Mukherjee was one among several writers to explain the initially low labour-ratio in Bengal in terms of the profitability of small farms. At some point down the scale, hiring in labour became uneconomical for the small-holder.14 On the same ground a 1939 study of Hyderabad observed that ‘any decrease in the area of holdings is likely to decrease the demand for hired labour’ (Hussain, 1939–40:696).

Rural labour and colonialism

61

Harold Mann suggested a route by which a sliding down of the right-hand arm of the Ushape might happen. In labour-scarce large-farm scenarios, labourers could buy land more easily, and thus tend to leave the labour force and join small peasantry. This, he believed, explained a fall in the labour-ratio in certain districts of Bombay between 1911 and 1921 (Mann, 1924:51). He was not alone. The note submitted by the Government of United Provinces before the Royal Commission on Labour attributed a fall in labour-ratio between 1911 and 1921 in terms of active buying of land by labourers, encouraged by a relatively high-wage regime in that decade.15 ‘But he is mostly without staying power’, Radhakamal Mukherjee predicted with fair accuracy in 1926. A contemporary study of North Bihar observed that in a small family farm scenario, supply of wage-labour varied inversely with cropping intensity (Mukherjee, 1930–31:320). The same study also highlights the process Mann had described before. Subdivision of holding ffsdssffs at the top was accompanied everywhere by the labourers’ ‘extreme desire for the possession of cultivation of land’. Thus, holdings, no matter how small in size, always had ready buyers who hoped to declare themselves at the next census as ‘peasants’. Later research using the National Sample Survey and Farm Management Survey datasets in the Indian Union again discovered patterns in cross-section data that would be consistent with the U-shape: (a) the incidence of wage-work among farm families increased with a fall in average farm size (rising density), and (b) the incidence of unemployment or surplus labour-time among farm families decreased with size (or fall in density).16 In the background of the U-shaped curve, when we see the average labour-ratio rising, we could infer two things to be happening: (a) some regions with large weight (in terms of population of labourers) moved up the left-hand arm of the U-shaped curve raising the average labour-ratio, and/or (b) the whole curve shifted upward. I believe that from c. 1930, when land-persons ratio began to fall, more regions in India tended to experience the former type of movement. On the other hand, labour-ratio for women is an example of the latter process. The former process pre-occupied the ‘rural economics’ scholarship from shortly after the Great Depression. The argument about marginality was twofold: more and more people in potentially vulnerable segments were pushed below the margin and into real crisis, and the margin itself was changing because of degradation of land. Between 1920 and 1950 when ‘rural economics’ arrived in India through the writings of Malcolm Darling, H.Calvert, G.Keatinge, Harold Mann, Radhakamal Mukherjee, Gilbert Slater and his team, and others, the new but energetic scholarship concentrated on marginality. This discourse saw ‘uneconomic’ holdings as one of the main problems of contemporary agriculture. Important contributions saw this feature as an expression of pressure of population with a constant land size and quality, and stagnant non-agricultural sector, leading to faster subdivision of holdings. Keatinge, among the longest serving Directors of Agriculture, wrote: ‘the landless labourers [are] steadily recruited from the smaller land-holders as the pressure of population in the most densely peopled tracts squeezes them out’ (Keatinge, 1921:141). Mukherjee took this line of reasoning further by stressing the environmental crisis dimension. He argued that agrarian expansion in UP had reached its natural limits given land size and quality. In fact, population pressure had for some time been leading to intensification of cultivation involving exploitation of groundwater resources, which became progressively costlier to mine. The resultant

Rethinking economic change in India

62

increase in cost led to a lack of investment in land maintenance. The outcome was ‘a gradually diminishing returns to an increasing amount of labour and expense’.17 The argument could have been made of the late twentieth century green revolution in Eastern and Western India, which was intensive in groundwater, indeed recklessly so. Inevitably, it showed signs of running out of steam from the end of the 1990s. The marginalization theme persisted for a while after these pioneering contributions. For one notable example, Dantwala and Donde (1949:68) concluded, based on a detailed survey of ‘uneconomic holdings’ in Bombay, that ‘[T]he whole structure of the agricultural economy presents a gradual process in the down-slide of cultivators to the ranks of landless proletariat.’ The mechanism they described did not have a specific role for either credit-based exploitation or industrial decay. It was based instead on: (a) ecological problems resulting from population growth, and (b) fragmentation of holdings. In turn, these factors led to two problems: diminishing returns per person, and poor rates of investment arising from deteriorating creditworthiness of the average peasant. Many other writers wrote about the same kind of change in other regions of India. Marginality as such was ever present in Indian agriculture because of the exceptional risks of cultivation in semi-arid conditions and the poor rate of technological change. But the specific form of economic marginality that we see in the twentieth century involved a new type of risk: loss of land because of steadily falling average size and quality of plots. And this risk was beginning to affect an increasing number from about 1930. A hundred years before, the risk that the peasants suffered was usually a risk of losing life in a sudden famine. In that sense, some regions were more at risk than others. Ecology was the medium. In the twentieth century, economic risks became more acute—a tendency counteracted, if briefly in the middle of the colonial period, by expansion in land area and new profit opportunities. Population was the medium. When both the counteracting factors slowed from the interwar period because of a slowdown in the world market and in irrigation investment, descent into sub-marginality became acute.

Bengal: a study in land-shortage The long-term factors driving the marginal peasants to labour were working with particular force in Eastern India. This was a region where agriculture was relatively more productive (product-per-worker was among the highest), but insufficient for subsistence (product-per-person was among the smallest). While productivity did not change in the long run, product-per-person was declining, and that intensified a subsistence crisis among the rural population. At the same time, ‘the cumulative effects of population growth, partible inheritance and market forces produced an irresistible wave of land subdivision’ (Bose, 1986:164). Greater Bengal in the colonial period was the only region to witness a steady and precipitous fall in land-persons ratio, already low here (Table 3.5). Land-holding of agricultural labourers was also, on average, low in this region. In addition, the Bengal peasants’ rights to land were ill-defined because of the Permanent Settlement. Eviction was a constant theme in this region in reports and documents throughout the twentieth century. As late as in 1921, the Census explained the relatively low labour-ratio in Bengal in terms of the small average holding and adequacy of family labour. But by 1931, Eastern

Rural labour and colonialism

63

India was a large agrarian economy in which the distinction between labour and smallholding peasants, weak to begin with, was wearing off fast. The labourer in Bengal was devastated by the 1943 famine, and briefly emerged a major figure in post-famine diagnostic research. Karunamoy Mukherjee observed four years after the famine the telling signs of marginality: ‘over-crowding’ in share tenancy, insecurity of employment, fluctuating wages, and relentless displacement of the

Table 3.5 Land-population ratios, 1891–1931 (net sown area in acres per person) 1891

1931

Greater Bengal

0.76

0.56

Bombay

1.27

1.21

Madras

0.62

0.59

Punjab

0.82

1.14

United Provinces

0.61

0.80

Central Provinces and Berar

1.39

1.36

smallholder from occupancy rights (Mukherjee, 1947–48:170). The background was a ‘deepening agrarian crisis’ that Mukherjee did not specify, but whose symptoms indicated diminishing returns to labour.

Why was 1930 a turning point? Fragmentation of holdings had been going on from the time secure property rights in land were defined. Land-shortage took serious proportions from the time irrigation development had slowed or stopped. In different regions, these variables can be dated from different points. In Punjab, for example, the mortgage market was active from the 1870s, evident in statistics on land transfers. However, the rate of land transfers accelerated in the interwar period. In Bengal, absolute land-scarcity was in evidence even before 1900. And yet, we do not see a significant rise in labour-ratio in the major regions before the decade of the 1920s. The solution to this puzzle is that rural underemployment could remain in check because of two factors. First, the rural population had remained roughly stable. And second, non-agricultural demand could absorb surplus rural labour gainfully. From about the Depression, these counteracting factors weakened. The significance of 1930 becomes clearer if we see it against changes in the preceding 25–30 years. Between 1900 and 1920, not only was labour-ratio relatively unchanged, but also real wages (in nearly all regions of India) increased. There was no generalized rural distress at all. Neither employment nor wage data suggests a process of net accumulation of surplus labour until about 1920. On the other hand, there is evidence that demand for wage-labour was growing from at least three sources. First, the main form of agricultural investment, irrigation, was labour-intensive and so was some of the crops that irrigation enabled producing, such as sugarcane. Real wages rose more in areas where

Rethinking economic change in India

64

rates of investment were higher. Second, land-per-worker and yield-per-worker increased between 1900 and 1915 whereas population remained in check. Third, while labour may have been surplus in the sense of underemployment of family labour, wage-labour capable of migration and filling up pockets of excess demand was not plentiful. Add to this a fourth factor, there was some outflow of agricultural labour to small-scale industry, public works and urban services. What about reverse flows of population? As for de-industrialization, national income statistics has now established that there was considerable rise in industrial income, overall and per head, between 1900 and 1920 (Sivasubramonian, 2000). If there was a rise in the ratio of workers in agriculture, it cannot be attributed to a crisis in industry, that much we now know. From recent industrial history scholarship, we also know that there was increasing occupational and spatial inequalities emerging within small-scale and artisanal industries arising from internal competition. It is likely that there was some exit, possibly considerable exit, from industry as a result of this second process. Until the 1920s, the net balance of inflow-outflow was, in relative terms, not changing in any particular direction. And yet, land was running out of supply, competition for land was becoming keener, investment had slowed, commercialization of labour had progressed sufficiently and manufacturing was not doing well from the late 1920s. The sharp hike in labour-ratio at 1931, in a way, reveals the breakout of an incipient crisis in the rural economy. The Depression added a debt crisis in this scenario, introducing a sudden push towards de-peasantization even in the prosperous agricultural regions. There is now adequate evidence to suggest that the period of relative prosperity in rural India, 1900–20, led to a progressive and notable easing of the rural debt market. Rising prices in the 1920s led to peasants acquiring land at high prices, accumulating debts and ‘squandering his easily earned gains on extravagances…thinking the boom had come to stay’ (Burns, 1941:83). The village surveys initiated by Gilbert Slater during World War I reported time and again a tendency among small and middle peasants to buy land with borrowed money (Slater, 1918). In Punjab, average land price quadrupled between 1905 and 1916, a rise fuelled by, among other factors, ‘a corresponding rise in the average sum obtainable on a mortgage’ (Calvert, 1918–19:399). The time-range when all this was happening, 1900– 25, saw rising incomes, wages, output and credit supply, all round. Nearly every description of the more commercialized regions of India about this time suggests a picture of dynamism, peasant accumulation and a seller’s market in labour. Indeed, reports of this time suggest an unusually active rural credit market consistent with Malcolm Darling’s general view on Punjab that ‘prosperity and debt are evidently intimately connected’ (Darling, 1920–22:165). Statistics on the net purchase of gold and silver shows consistently high, indeed unprecedented, levels of consumption of precious metals in India in the first quarter of the twentieth century. Consumption of gold alone increased from five million fine ounces in 1900 to eleven million in 1910, twenty-two in 1915, and twenty-five in 1925 (Prakash, 1960). There was no question that the bulk of this gold was going into the peasant household. This unusual asset build-up was to be knocked down in the early 1930s. The Depression hit hard in this backdrop of creditfuelled agrarian expansion. It inflated debts, deflated assets, raised real wages and squeezed a great many small-holders in an unprecedented degree.

Rural labour and colonialism

65

The argument here can be given a general form. We know from research done in the last forty years that ‘de-peasantization’ cannot refer to the peasantry as a whole, sections of which gained from commercialization. And yet, it is plausible that the same general process drove sections among the peasantry to marginal or sub-marginal existence. After all, property rights and market-access were unequally distributed, and both became more valuable than before. In explaining proletarianization, we must concentrate on these marginal segments, and how they were doing in the 1920s and the 1930s.

Why women? Changing occupational mix Neither the thesis of rural distress, nor the alternative offered here based on landshortage, explains why women took up agricultural labour more readily than men. Why was de-peasantization gender-biased? The general position on women’s labour, accepted by default of a good alternative, is that the data are unreliable because of under—and over-enumeration. This position is oversimplified on three grounds. First, it is conjectural. Second, it ignores the fact that post-1950 data show continuous rise in women’s labour-ratio, and that the trend might have begun earlier. And third, it misses the significance of a striking symmetry in women’s work-participation between artisanal and agricultural labour, the two main occupations of the rural poor. Women’s presence in manufacturing as well as ownercultivation declined in the twentieth century, while participation in agricultural labour increased. On the strength of census statements, I believe that the transformation of women’s work witnessed the intersection of a number of separate tendencies. First of all, we need to understand why many women reported themselves as ‘peasants’ in 1881, and scarcely any reported themselves as peasants in 1931. Did women actually have cultivation rights that they lost the possession of? The 1881 census in Bombay Presidency hinted, rather intriguingly, at such a possibility: There are, of course, many cases of independent occupancy on the part of women’ (India, 1881a: 197). A study of land rights does not easily reveal rights that were gender specific. It is also clear that in many provinces women of peasant families were automatically classified as peasants. They were not wrongly classified. They did work on family farms or on the farms of relations. That said, ‘it is very hard to separate dependents from actual workers in this class’ (India, 1901b: 231). From these groups, an increasing number began to enter wage-labour in the early twentieth century. Why did they do so? It is not improbable that some kind of custom did disappear. Perhaps cultivating the fields of ‘relatives’ signified a form of tenancy, which disappeared as many other tenancies did. On the other hand, women of peasant families, once routinely classified as peasants because they worked only on the family plot or did not work, did begin to devote a lot of time to the labour market. Why did that happen? It was facilitated by an institutional change—the decline of the farm servants. This was the second large tendency with a significant bearing on women’s work. Before 1900, few women offered their labour in long-term arrangements, which was a male sphere. We can infer this from the low percentage of women (5–10) among the census category ‘farm servants’. Indeed, reference to women of farm servant families doing agricultural labour on a steady basis is scarce before 1900. Not surprisingly, Edgar Thurston’s long account

Rethinking economic change in India

66

of the Cherumans of Madras describe the whole range of labour tasks in the masculine gender, adding that the ‘wives sometimes accompany their husbands in their watches’ (Thurston, 1909, vol. 2:24). It is not difficult to explain why working women and what the Thorners called ‘beck-and-call’ system would be incompatible. It would interfere far too much with domestic duties, whereas casual or seasonal labour was more compatible with the domestic sphere. The second Agricultural Labour Enquiry (1955–56) produced evidence to suggest that the number of days supplied to cultivation work was the highest for attached labourers (326), in the middle range for male casual workers (200) and the smallest for female casual workers (120). If these figures are any indication of the potential supply of work-effort, it is easy to see why women would not enter attached labour arrangements. Around the turn of the century, labour market contracts were changing, with steady disappearance of long-term contracts and increase in casual labour or renewable/negotiable long-term contracts. Whereas among farm servants, women were rare, among casual labourers, men and women were present in equal strength. In fact, the censuses show that in the second half of the twentieth century, women outnumbered men in paddy-based labour markets. But women were subject to a third big process. If the development of spot markets in agriculture encouraged women’s participation in agricultural labour, the development of spot markets for industrial labour led to their withdrawal from industry: rural spot markets were local, but spot markets for industrial labour were far away from home. Traditionally, industry, cultivation and labour had coalesced for many poorer rural women in the nineteenth century. In part, this coalescence was effected via household industry, which allowed a considerable degree of non-specialization. This coalescence tended to give way, and household industry steadily weakened in the twentieth century. Given women’s partial access to markets for industrial goods and industrial labour, women tended to leave manufacturing and enter agricultural labour. Men, on the otherhand participated in manufacturing labour relatively more. Finally, there was migration, which had a strong male bias. Gathering land-scarcity and decline in household industry forced marginal families to look for wage-work far away from home. Rural women stayed behind more readily and far longer than men. In some of these cases at least, the exit of the men meant loss of access to the family plot. The women then entered the rural labour market. This whole thesis on why women tended to move into agricultural labour more readily than men is consistent with three key empirical observations about labour in twentieth century India: decline of household industry, migration of men and contractual change in agriculture. Women were not alone among rural classes that were changing their occupational mix in India during late nineteenth and early twentieth century.

Changing occupational mix generalized Broad terms like ‘de-peasantization’ and ‘de-industrialization’ oversimplify the figure of the rural worker, and the economic changes that they underwent in the early twentieth century. These terms presume too water-tight and exclusive distinctions between the

Rural labour and colonialism

67

‘peasant’, ‘artisan’ and ‘labourer’. In reality, these categories coalesced for a substantial section among the rural working population. In parts of rural India, small tenants, labourers and artisans tended to be overlapping categories. Labourers were a motley group consisting of landless workers aspiring to or trying to gain some cultivation rights, labourers with small quantities of land, peasants and tenants commanding insufficient land, families that combined industry and labour and families that combined ‘begar’ or the duties of a village servant with labour. They went too deep in any one direction according to economic conditions. These conditions could be short-term ones arising from seasonal or market fluctuations or could also be more long-term ones, such as land scarcity or regulations that discouraged tenancy. Recent research on the artisans suggests that a similar reconfiguration is necessary with the manufacturers. In rural India there were artisans specialized in industry and artisans who pursued a mix of occupations. The majority among the former groups was averse to agriculture and rarely owned land. But with the latter group, the distinction between ‘artisan’ and ‘agricultural labourer’ was thin or non-existent. Broad caste names that included millions of people—Chamars in the north, Pallans and Parayans in Tamilnad, Malas and Madigas in Andhra region, Mahars and Dheds in central India are just a few examples—were artisans or labourers according to the season and local resource conditions. So-called artisans of this kind were, according to region, included among the agricultural service castes or treated as industrial. Now, all these groups who depended heavily on agricultural labour, and yet might have been classified by the census as ‘general labour’, ‘manufacturer’, ‘tenant’ or even ‘owner-cultivator’, were going through two large transitions in the twentieth century. Some among them, especially the quasi-artisans, were previously part of arrangements where they served local needs in exchange for grain share. These arrangements were generally in decline, as we have seen, and being replaced by casual labour. At the same time, and overlapping with this change, some among those who earlier combined occupations, shifted relatively more towards agriculture and away from non-agricultural work. There was increasing concentration in agricultural labour, or manufacture, or in urban labour. These shifts can be seen as conscious voluntary moves into a new option in response to the prospect of higher earnings, or these shifts can be seen as absorption of workers unable to survive in traditional modes of work. We cannot completely settle whether these shifts were primarily demand-driven or supply-driven. We need to remember, furthermore, that completely specialized agricultural labour was not a sustainable option in most parts of rural India where agriculture provided no more than four to five months of employment. Rural labourers always needed to combine occupations, to pursue a portfolio. Along with some degree of voluntary or involuntary specialization, something else was happening. This portfolio of occupations was beginning to be market-driven. The decline of ‘begar’ or unpaid village services signified that element of freedom. This twofold process of change—increased freedom of choice over occupational mix and increased relative concentration—has not been well-researched, though some of its signs have been observed in agrarian history. Neil Charlesworth, for example, discusses ‘a more deliberate and flexible allocation of their [the Bombay peasants’] effort [after 1900], often involving a conscious choice towards wage labour’ (Charlesworth,

Rethinking economic change in India

68

1985:221). Breman (1974) and Yanagisawa (1996b) describe the slackening of nonmarket bondage on rural labour as outside opportunities open up. Roy (1999) has described the transformation of rural artisans into specialized labourers. The increasing facility of short distance migration played an important role in these stories. It is this class of evidence that I wish to dwell on in some detail. There is a great deal of qualitative evidence suggesting that rural labour castes performed a large variety of occupations before 1900, but were giving up some of these. Detailed ethnographic research on the so-called ‘untouchable’ castes at about the midtwentieth century captures both these processes well. Bhagat (1935) was one such study on a diverse population in Western India. Somewhat earlier, the famous caste anthologies prepared for different regions referred, if too briefly, to similar changes. The Holeyas of Mysore were both weavers and labourers, but their engagement in weaving was declining in the end of the nineteenth century. The traditional occupation of the itinerant Dommaras was diverse, but c.1900, they were ‘settling down as agricultural labourers’. The traditional occupation of the Madigas was described as leather, and yet by 1900 they were ‘mostly field labourers’. A number from the wine-making castes of both Travancore and southern Madras gave up wine-making and specialized in agricultural labour.18 The principal occupation of the central Indian Mahar was stated to be weaving. But, in the early decades of the twentieth century, they did little weaving, but were engaged in other types of rural work for payment (Russell and Hira Lal, 1916:144–5). Throughout India, the dissolution of customary ‘menial service’ contracts was often cited as the reason why these occupational shifts were happening. The custom required ‘begar’ or corveé for a variety of off-season work liabilities, including artisanal work. These ‘begar’ elements were no more accepted as the norm.19 Census occupational statistics on caste is not detailed enough to quantify these shifts. But when comparisons can be made, we observe the decline in the proportion of workers engaged in the traditional occupation of the caste among artisan-labourers. The north Indian Chamar, among the most numerous of northern groups, is an example. Between 1931 and 1961, work-participation rates among the Chamars increased, dramatically among women, but the percentage of men or women engaged in leather-related occupations fell by a large extent. Specialization in the other direction, from a portfolio of occupations to industry,is much rarer, but not absent. Within some of these artisan-cum-labour castes, a great many did move towards manufacturing and trade of the traditional occupation. Among some of the pastoralist groups, there was a distinct move away from sheep-rearing and towards manufacture, an example is the Punjab Gadaria (Ibbetson, 1916:303). There were possibly two general types of rural labour in India: those who were subjected to some form of proscription regarding ownership of land, and those who did not suffer in this way. The former proscriptive rights were more conspicuous in South India.20 The great South Indian castes known to have been thus ‘tied’ to land via a variety of customary field servant arrangements—Cherumans, Malas and Paraiyans—had, c.1900, relatively high proportions engaged in traditional occupation. However, there was a noticeable tendency among some of them towards owning or renting land. By 1961, there was greater occupational diversity, but towards other types of specialized labour and cultivation. There was no noteworthy shift towards industry or related work.

Rural labour and colonialism

69

One of the well-known problems with reading historical data on occupations is the existence of a large group called ‘general labour’, defined merely as persons who ‘take to miscellaneous tasks involving as little as possible of anything beyond bodily strength’ (India, 1881a: 196). Scholars such as the Thorners felt, on the basis of some conjectural statements by census officials, that this group consisted of agricultural labourers by and large. This is correct, but up to a point. It does not explain why the census needed to invent such a group in the first place. Nor does it explain why women crowded general labour relatively more than field labour. A more plausible position, one that is more consistent with modes of work in rural India, is that this group consisted of persons who combined a variety of services. Because women performed a range of tasks, they were naturally counted as ‘general’ labour in the nineteenth century, whereas women were rare among agricultural labourers. General labour was shrinking in size. If general labour is taken as a measure of occupational coalescence, that coalescence was breaking up. At the same time, the percentage of women among ‘pure’ labourers began to increase. When scholars discuss proletarianization, they too readily equate it with the peasants turning into labourers. The mechanism I just described and the one that the census suggests was a different one, diversified rural labourers turned into agricultural labourers.

Rural work: many transitions What I wish to stress is that the worker in colonial India experienced many processes of change. Peasants losing land and artisans losing livelihood were not the only important changes, nor can we see these changes as the axes of the narrative on labour. There were different types of worker, and they faced a range of options or were subjected to diverse forces over time. The only connecting thread between these stories is perhaps the extension of spot markets for labour in the village or outside and progressively easier access to these markets. Dispossession, therefore, did not necessarily play a large role in this diverse transition. Other factors at work included institutional change in the labour market, changing endowments of land and labour, differences between men and women in work-participation and breakdown of customary unfreedom. Perhaps the most dramatic transition involved those groups which earlier performed labour under conditions of extreme inequality in asset holdings. The most ‘depressed’ of the rural classes in South India belonged in this type. They saw a progressive loosening of caste-occupation correspondence in the twentieth century, as a result of changes within agriculture and opportunities outside agriculture in plantations, public works and overseas migration. Serfdom slowly ended and gave way to market-driven occupational choice. Invariably, this meant moving from hierarchical relationships to wage-based ones. However, to a limited extent, there was upward mobility within this group, that is, there was some extent of land-ownership and peasantization. A second large stem was, undoubtedly, de-peasantization. Among those who enjoyed some form of rights to ownership or cultivation, the peasants, there was increasing dependence on labour. It happened to a much greater extent in Eastern India, as a result of overcrowding on land and sometimes accompanied by degradation of land. The third and the fourth trajectories involved those groups which earlier did not depend too much on either labour or cultivation. These were groups that performed a

Rethinking economic change in India

70

variety of services. Women, for example, performed a range of services for the household. They cultivated land, participated in wage-labour and worked in the crafts. The third trajectory had women at the centre. Many women who earlier reported themselves to the census as owner-cultivator reported their occupation as wage-labour in later censuses. Was this just correction of an error? Not necessarily, in my view. An alternative hypothesis is that women did enjoy some form of customary rights to land, if mediated by the husbands or the collectives. With progressive scarcity of land, many old rights changed hands. It is plausible that some such reconfiguration took place across gender as well. Simultaneously, women withdrew from industry and entered spot markets for agricultural labour in an increasing extent. We cannot say whether these three subprocesses—withdrawal from ownership, from industry and entry in labour—were necessarily connected or not. To some extent perhaps they were. The fourth trajectory involves certain castes who performed a range of services for the community. Once again, coalesced occupations broke up or transformed into a new combination of occupations driven by compulsions and opportunities arising from the market. Clearly, the term ‘proletarianization’ cannot be used in a simple sense. It had a diverse meaning depending on which group we focus on. It could mean the reduction of peasants to labour, women’s entry into labour or disappearance of coalesced occupations. All these processes were co-existent, and the mix was regionally variable. That said, some groups and regions experienced mainly one or the other, whereas some groups such as women experienced all these tendencies. This many-sided change can be summarized in the form of a table (Table 3.6). The table distinguishes two broad types of rural workers: those relatively concentrated and

Table 3.6 Diverse transitions consistent with increase in labour-ratio Livelihood profile/ mobility

Remained rural/sedentary

Migrated

Relatively specialized

Peasants lost land, partially or completely, and joined rural 1 Peasants lost land labour and migrated 2 Labourers migrated in search of labour

Relatively diversified

l Women of artisan-cum-cultivator families joined Men of artisanagricultural labour because old livelihoods were in decline, cum-cultivator and women migrated less often families left home 2 Women of artisan families joined casual labour because the men migrated 3 Women joined casual labour because ‘farm servant’ contracts declined 4 Service castes and ‘general labour’ groups joined casual labour because obligatory livelihood combinations disappeared, farm servant contracts declined, and migration became easier

Rural labour and colonialism

71

those relatively diversified. It also distinguishes two general options for the rural poor: to stay on in agriculture, or to move to the cities. The processes discussed above are then classified. We need to remember that the categories that form the rows are not mutually independent. There was great deal of overlap between them. Also, several of the processes that appear in the column ‘remained rural’ can show up in the census in the form of increasing labour-ratio. But, of course, they all do not mean the same thing. A peasant selling the family plot and a woman artisan giving up manufacture for labour were different processes, even though both may be recorded in the successive censuses as a case of increase in labour-ratio.

Conclusion Proletarianization in colonial India has been explained in terms of a story in which class power and colonial power jointly made the rural poor progressively poorer. But such a narrative cannot be generalized to (a) all the poor, (b) all the peasantry, and (c) beyond colonialism. The facts do not support one single theory of policy-induced immiseration. There were other reasons at work. Two reasons seem paramount. First, competition for land was increasing because of decline in land-persons ratio. Among those who lost out early were women, who had inferior rights to land, and the small peasant in eastern India, where insecure rights were joined with acute shortage of land. The second mechanism was a market-driven specialization process. Many who had combined own-cultivation, labour and industrial work at the end of the nineteenth century, and had generally poor access to resources and markets, were driven by relative returns to specialization in labour rather than in production. Women figured in this group too, for women in rural India tended to be occupationally diversified. But women were not the only group so subject. The changing occupational mix also affected traditional rural castes that previously supplied a diverse range of services for the village, but were encouraged to or driven to specialize, and/or work for the market now. The macro-economic story that emerges from this review has at its centre two large tendencies that began well before 1947 but by no means ended at that date. First, there was diminishing returns in the presence of rising population and a constant supply of land. And second, there was institutional change that drove many producers with tenuous rights to assets to varying degrees of redundancy, specialization and labour.

Appendix: the relationship between density and labour-ratio In this appendix, the basic data on the U-shaped relationship between density and labourratio are presented. In Figure 3.2, data on six major

Rethinking economic change in India

72

Figure 3.2 Density (acres) and labourratio (%), 1901–31 (for source, see Figure 1.1). provinces for four censuses (1901–31) are shown. A third-order polynomial trend-line is fitted onto the scatter.21 The shape remains stable across major regions at any one point in time (1901 and 1931). It is not so clearly present over time for at least two of the six regions. In UP, Madras, Central Provinces and Berar and, to a smaller extent Punjab, labour-ratio progresses along a roughly U-shaped trajectory. Initially, each region

Figure 3.3 Density (acres) and labourratio (%), 1951 (for source, see Figure 1.1).

Rural labour and colonialism

73

contained pockets of labour-scarce agrarian zones. Agriculture in these pockets was growing rapidly around 1901. By 1931, the growth impetus was more or less exhausted and each region was facing pressure upon land. In Punjab and Central provinces, migrant labour formed a large part of the relatively high labour-ratio at the beginning of the period. In Madras, farm servants met the needs of employers to tie in labour. As population density increased, indicating a fall in average size of holdings, both migration and labour-tying weakened. Towards the end of the period, some degree of depeasantization began to add to the growing pool of casual labour. The cross-section relationship survives in the first Agricultural Labour Enquiry dataset, as Figure 3.3 shows us. A second-order polynomial is fitted onto the scatter.22 The 1951 enquiry was based on a completely different methodology and survey from those of the census. The survival of the U-shape, therefore, suggests the robustness of the relationship. It is noteworthy though that many more observations in 1951 cluster around the left-hand segment of the U-shape. When we come to 1961, nearly all observations cluster in the bottom portion or at the leftward segment, ending the U-shape more or less.

4 Agricultural labour Lessons from wage data If peasants became poorer and joined an already surplus worker pool, one would expect real wages to fall. Did real wages fall? This chapter argues that trends in wages in the later nineteenth century cannot be easily interpreted unless we have an explanation for an inconsistency between two sources. There is less of a contradiction between sources in the early twentieth century, and these show that the trend in real wages was usually increasing in the first quarter. Through both periods, sharp fluctuations in real wages could produce devastating short-term impact. The data also show that, if these fluctuations are ignored, then real wages began to fall or stagnate from shortly after the Great Depression, and that the stagnation lasted at least three decades. The chapter has five sections: database and findings on wage-trends, interpretation of trends, fluctuations, wage-share and interregional variation. An appendix describes details of construction of the dataset on average money and real wage between 1873 and 1968.

Database and major findings The historical data on wages are fragmentary, but not insubstantial when the fragments are put together. The data have not been exactly neglected in regionally oriented agrarian history, but that scholarship has rather shied away from large-scale generalizations using the data. By contrast, the post-colonial labour experience has been a field of extensive research, reviews, estimation and generalization.1 Work on national income could become a link between general economic growth and rural wage trends. But this link has not developed enough because of technical reasons. Research on national income could afford to bypass wages because agricultural income is conventionally estimated by the production method using crop output statistics.2 At the same time, national income research is yet to address functional distribution in a serious way. There is, it seems, a point in putting the historical wage statistics together. Official agricultural wage data come from two main sources. The first is a publication of the Department of Statistics, Prices and Wages in India. This source covers the period 1873–1923, then it was discontinued. The agricultural wage data in this source, however, become too thin and unusable after 1910/11, when the dataset was discontinued and replaced by quinquennial wage surveys conducted in the major provinces. The content and quality of the latter datasets vary. The second source is the sizable Report on an Enquiry into the Rise in Prices in India, better known as the K.L.Datta Committee (1914). This publication independently collected data on wages and published this data

Agricultural labour

75

for the period 1892–1912. Neither source gives sufficient details of collection procedure, or methods of dealing with the usual problems of reading rural wages. There are well-known difficulties in using any given set of wage data for rural labour to read trends in earning. Three of these are particularly important. First, individual workcapacities can vary a great deal, and vary especially with respect to wage at low levels of nutritional intake. Wage rate may include an implicit investment in work-capacity. In that case, one wage and another, and one worker and another, are not comparable. Second, wages vary according to casual or permanent status, because the employer may trade-off the advantages of a spot market for that of tying labour for peak season work. As in the former case, wage here may be adjusted by an implicit insurance premium. Again, there is a comparability problem. Third, wages vary a great deal according to season, especially in monsoon-dependent agriculture. Neither source of data is clear on capacity, casualpermanent distinction, or the seasonality issue. The only way we can use these data for comparisons over a long time is by making three assumptions: (a) while individual capacities may vary, the composition of the workforce does not change over time, (b) the data refer to casual workers (in which case the first problem is also taken care of, for wages in the spot market are unlikely to include investment components), and (c) the wages refer to the peak-season operation, which is a plausible case. For the time being we start with these assumptions, but I shall argue further on that the assumption (b) is weak, and may need to be modified. Specifically in the Indian case, wages in kind create a problem. Rural wages typically contained a cash and a kind component (later called ‘perquisites’). In principle, this problem does not affect the analysis here, because Prices and Wages supposedly collected its raw data in grain terms and converted it into cash, whereas the Datta Committee made adjustments for ‘perquisites’. In practice, neither procedure is easy to implement and leaves some doubt about the veracity of either source. Subject to these caveats, what do the data show? Prior to the start of Prices and Wages, wage data are fragmentary and unsystematic. Collating what exists for two major regions, Sumit Guha and Dharma Kumar, suggest that in Bombay and Madras respectively, the first half of the nineteenth century saw no clearly discernible tendency in real wage (Guha, 1985; Kumar, 1965). M.Atchi Reddy’s seminal work reiterates this conclusion for the southern coastal Andhra region, but does observe a rising tendency after 1844 (Atchi Reddy, 1986). The rise continued until the mid-1870s, when a devastating famine occurred. From whatever other scattered evidence that we have, a break in trend around the mid-century seems plausible. Data collected by the Wage Census of Bengal suggest broad stability in money and real wages in the first half of the nineteenth century, but a doubling of money wages (and a rise in real wage of about 50–60 per cent) between 1852 and 1911.3 Chaudhuri (1983) gathered a larger set of sources that suggest that cash wages increased significantly in Bengal after the midcentury. Data resources on wage changed significantly from 1873. The publication Prices and Wages allows us to create a time-series of real wages for 1873–1909. Figure 4.1 and Tables 4.1–4.2 show that, in the 1870s, there was some gain for labour, though Madras and Bombay—both particularly affected by the 1876 famine—saw a much smaller rise in the 1870s. Consistent with the findings of Atchi Reddy, who uses a different source, the Prices and Wages seems to capture the end of a gradually rising tendency in real wages

Rethinking economic change in India

76

in the 1870s. If this inference of a rising trend is correct, and if this rise was marketdriven, it can be explained in terms of expansion in land area, the still quite restricted labour mobility and the exit of labour from agriculture to the public works and plantations. These hypotheses, however, are based on big ‘ifs’ and might need to be modified, as I argue further on. There occurred a remarkable change in trend after this period. The 1880s and the 1890s saw a steady fall in real wage. The fall derives from a

Figure 4.1 Average wage (Rs/month) (for source, see Figure 1.1). Table 4.1 Money and real wage in agriculture, 1873–1951 Money wage Rs/month

1891=100

Real wage 1873 prices

1891=100

Source: Prices and Wages 1873

4.6

77 4.6

92

1891

6.0

100 5.0

100

1901

5.2

87 3.6

72

1891

5.8

100 4.9

100

1901

7.5

129 5.2

106

1912

11.1

191 6.4

130

32.8

565 5.6

116

Source: Datta Committee

Source: India (1954) 1951

Agricultural labour

77

Note For sources and methods used in the construction of the tables and figures in this chapter, see appendix at end of this chapter.

near-total stagnation in money wages even as prices were rising. This result on real wages in the late nineteenth century has made Prices and Wages famous, or infamous, depending on how credible it is seen to be. The conventional practice among historians has been to take the wage stagnation to be authentic, and explain it either as a result of distress or structural change. There was sustained upward pressure in food prices in the 1880s because of exports as well as depreciation of the currency. The nationalists argued that these circumstances intensified a subsistence crisis. In recent scholarship, Bagchi (2003) cites Prices and Wages data as evidence of immiseration caused by the currency problems of the end of the century. There is some truth in the inflation argument. But there is also a lack of caution. Both exports and depreciation exerted an upward pressure on demand. Even if prices increased, one would expect the demand pull to

Table 4.2 Real wage in major provinces, 1873– 1912 (1891=100) Major provinces

Prices and Wages 1873

Greater Bengal

1883

1891

Datta Committee 1901

1891

1901

1912

100

119

100

85

100

110

124

Punjab

98

125

100

106

100

117

169

Madras

78

80

100

60

100

100

131

Bombay

104

107

100

69

100

107

122

Central Provinces and Berar

93

107

100

67

100

105

131

United Provinces

97

105

100

76

100

101

130

lead, via a multiplier effect, to a general expansion, a rise in labour demand and a rise in money wages. Now, all the signs of a rise in labour demand were present, why should wages not respond? The last quarter of the century saw continued expansion in net cropped areas, no change in population of labourers, and considerable exit of labour from agriculture. The stability in money wages in the last quarter of the nineteenth century is remarkable precisely because it does not respond to acreage expansion. This is counterintuitive, unless of course we are prepared to accept that rural wages were decided by non-market factors. If we apply ‘de-peasantization’ to explain the money wage stability in the last quarter, we need to remember that rising prices were almost universally seen in British India, with good reason, as beneficial to the peasantry. Dharma Kumar too accepts the decline in real wage for Madras to be authentic, and attributes it to land-person ratio (Kumar, 1983b). But like currency-based explanations, resource-based ones also cannot be taken very far for the late nineteenth century when the land-frontier was still quite some distance away.

Rethinking economic change in India

78

If neither of these hypotheses for stagnation in real wage seems completely credible, both are firmly contradicted by the Datta Committee dataset. For the overlapping years (1890s), the Datta Committee dataset shows that money wages rose sufficiently to register a healthy increase in real wage, at the peak of a monetary inflation. The Datta Committee criticized the Prices and Wages for not being scrupulous enough about recording the actual market wage every season. Seemingly confirming this suspicion, long series of years in Berar, Orissa or Assam return identical money wage rate. On the other hand, the Datta Committee dataset for the earlier years was apparently based on recall, whereas Prices and Wages at least had a system of collecting wages every year. Both datasets suggest a relative slowdown in the late 1880s and the 1890s, compared to the decades before and after these years. To that extent there is agreement. But the levels of flexibility in money wages differ markedly. What is also striking about the Prices and Wages is that the stability in money wages seems to disappear after 1900. I take this inexplicable and counterintuitive change of dynamics as a clue to the problem of reading the Prices and Wages dataset. Unless we take the Datta Committee’s criticism of the dataset at face value, a plausible solution to the muddle about the period 1880–1900 is that the Prices and Wages data did bear signs of a transition from non-market to market-driven wages. In other words, perhaps the two sources were looking at different types of labourers. K.L.Datta, whose compilation of a wage-price series became more famous than his analysis, was explicitly looking for casual labour wages. He drew a vast number of primary and secondary sources together, often travelling through the countryside for the purpose. He did not explain all the sources in adequate detail, but was clear on the point that the only rate of wage worth comparing across regions was peak season casual wage. The Datta dataset, therefore, referred to tasks. By implication, the wage data in the rival source that Datta disfavoured, Prices and Wages, were time wages, collected by villagelevel agents, and almost certainly related to the more regular component in the labourforce, namely, those under long-term contracts. Only such people could be unambiguously called ‘labourers’ at any time of the year, whereas many among the peakseason workers might not always represent themselves as ‘labourers’ at other times of the year. If that is the case, one would expect two things to happen. The wages of those under permanent arrangement being generally lower than casual wages (but their employment more secure), one would expect Prices and Wages figures to be lower. And second, because it was a long-term and caste-bound relationship, these wages contained a customary element, and were likely to have remained relatively unchanged for long periods of time. A comparison of the two sets shows that both predictions are borne out. Customary wages consisted of both cash and product components. The existence of a fixed cash payment would impart a stability in money wage rates. Customary wages were also sometimes paid as a share of grain output. If the share was fixed, fluctuations in output would make real wages unstable, though money wages may remain stable. For a fall in grain output would cause prices to rise. A second and more plausible explanation is that the price series used in the official sources was wrong and not revised very often. Reliable cost-of-living indices are unavailable. The prices used are output prices. It was the case that grains that dominated the consumption of the rural poor were not important in output data.

Agricultural labour

79

Whether or not the decline in real wage was spurious or real, in periods when prices were rising as steadily as prices did in the late nineteenth century, a change in customary labour services had to happen. From the employer’s side, rising prices would make fixed grain payments costlier, and might well encourage a shift to cash wages or hiring in from the spot market, as Radhakamal Mukherjee suggested. From the worker’s side, rising prices would make fixed cash payments costlier and, therefore, encourage more frequent bargaining. Indeed, we do have a massive quantity of evidence suggesting that the customary arrangements began to dissolve from about this time.4 What I am suggesting is that trends in money wages in the turn-of-the-century Prices and Wages figures were either wholly unreliable, or implied a non-market payment situation. And if it was a non-market situation, then a sustained inflation was enough to weaken non-market arrangements in rural areas. By this reasoning the quicker adjustments in money wages revealed in the Datta Committee dataset or datasets available about twenty years later stood for wage movements outside or after the nonmarket regime, that is, in the casual or peak-season market. We cannot verify directly how sound this hypothesis is, but at least it is a plausible resolution of the contradictory and counter-intuitive picture we get from the Prices and Wages data. Despite two famines, production conditions in the last quarter of the nineteenth century were generally good. Monetary factors might cause an increase in prices. But labour demand increased too. If workers did supply the extra work, they must do so on one of three conditions. First, population growth or migration eased labour supply so that real wages did fall. Second, they asked for higher cash wages. And third, they found locked into customary arrangements. Labour supply overall was not rising in the last quarter of the nineteenth century, even if spatial movements were beginning to increase. We must, therefore, settle on the second and the third hypotheses. The group that could ask for a higher wage every season was still a small one. The group that found itself locked into customary wage arrangements was a large one. The inflation, together with growth of labour markets and opportunities of migration, contributed to the shrinking of the latter group and expansion of the former. There is a related issue of convergence in wages. If we assume that all wages were equilibrium wages and isolation was the chief characteristic of the labour market before the railways made migration easier, then the process of integration of labour markets should lead to a convergence in wages. Migration implies that some low-wage areas would experience exit and some high-wage areas experience entry. If we assume, on the other hand, that some wages were customary before and not decided by demand and supply, then the same process can lead to divergence in wages. In that case, the railways would enable not only spatial mobility but also contractual mobility. That is, some workers and employers would be able to break out of custom and demand or offer equilibrium wages. If customary wages were generally less than equilibrium wages before, then the new wages offered would be higher than the earlier ones. In that case, the cluster of wages observed after mobility, a mix of customary and market wages, may well show divergence rather than convergence. A recent paper on Indian wages (Collins, 1999) finds no evidence of either in the Prices and Wages dataset, but finds evidence of divergence in the Datta Committee dataset. The paper concludes from this evidence that there was insufficient mobility of labour. Implicitly, it is assumed that all wages before and after migration were equilibrium wages. If we modify that assumption somewhat,

Rethinking economic change in India

80

and allow for the existence of customary wages, the interpretation of these findings would be that there was in fact accelerated mobility. At about the first decade of the twentieth century, nearly all available sources show that there was rise in money and real wages. In some cases a mild rising trend continued till 1925.5 The fastest growth in the last phase was recorded in Punjab, Madras and UP— the main regions to experience continued agrarian growth based on cash crops in the early twentieth century (Table 4.2). In parts of Madras, not only had wages of casual labourers ‘more than doubled in the last twenty years’ from 1916, but these wages were also higher than the customary payment made to the permanent servants.6 Further, while women rarely worked as permanent servants, in the casual labour market both men and women participated. The family income from labour, together with a shift from nonmarket to market contract, increased rapidly in Madras between 1896 and 1916. Indeed, in South India, this phase of rising wages was inseparably linked with the breakup of permanent relationships all round. Wherever agriculture was doing relatively well, the opportunity to migrate to Burma or Ceylon was present and there was some influx of immigrant labourers, the wage-gap between the servant and the casual worker became stark, leading to emigration of the former. These circumstances caused what one contributor to Gilbert Slater’s enterprise called ‘a dislocation in the [local] labour market’. Interestingly, this whole episode of dislocation saw ever more insistent pronouncements on the part of the employers, whenever they had a chance to express their view, about the sanctity and moral authority of the old patronage relationship.7 For the interwar period the collection of regional datasets are only approximately consistent. The most complete of these series refers to Bombay and was compiled by the Labour Gazette of the provincial government, later used by Rath and Joshi (1966) and Sivasubramonian (2000). We can add to these dataset, a fairly comprehensive set of observations on real wages for field labourers in five other provinces, based on the quinquennial wage censuses that replaced the Prices and Wages (see appendix for details). The Bengal figures were cross-checked by those reported by other contemporary sources. They tally closely. UP and Punjab data were likewise cross-checked with two other sources. The resultant levels can be questioned. But, as the variation between wage figures across regions was small and narrow over time, the trends are representative. The trends (see Figure 4.2) can be summarized as follows: 1 The few years from the end of World War I to c. 1920/21 saw a rise in money and real wage. In Madras, spurred by a groundnut boom, this was the case. The first half of the 1920s saw no significant change in money and real wage. 2 From the second half of the 1920s through the 1930s, money wages were falling. Food prices were falling too. Real wage decline started within a few years of this trend. After the adjustment was complete, real wages c. 1935 were usually slightly below those c. 1925. 3 Halfway into the massive inflation generated by World War II, real wages were depressed, but thereafter, some adjustments in money wages to inflation did occur. However, the adjustment did not result in a steady upward rise in real wage. Rather real wage fluctuated

Agricultural labour

81

Figure 4.2 Wage rates, 1920–51 (Rs/day, real wage in 1873 prices) (for source, see Appendix). sharply. At the end of the decade, real wages were again below or about comparable with levels that prevailed shortly before World War I. Grafting 1951 money wage on this chart suggests that the stagnation that set in after and in the course of the post-Depression adjustments in money wage was long lasting. Later research has shown that the stagnation in real wage lasted until the late 1960s, in some regions until the mid-1970s, with only a short episode of rise in the early 1960s.8 Figure 4.3 shows the stagnation using Mukherjee’s series for the purpose (see appendix for sources and the data on wages). For a remarkably long period of time, 1930–70, labourers on average either did not gain monetarily or could not retain the fruits of short-term prosperity. If we compare the Prices and Wages figures with those that later wage surveys produced, we cannot escape the astonishing result that the average worker in rural India earned an amount in 1965 that was not higher, possibly even smaller, than that earned in 1875. More precisely, the labourer gained in the early twentieth century, to regress in the 1950s and 1960s. Not surprisingly, the 1960s ended in fairly widespread outbreak of revolutionary sentiment in rural India.

Rethinking economic change in India

82

Figure 4.3 Agricultural labourers: number and earning, 1950–64 (1950=100) (for source, see Appendix). Interpretation of trends: demand for and supply of labour In a canonical model of economic development that had enormous influence on policy discourses in South Asia, W.Arthur Lewis proposed that the aggregate supply curve of rural labour for both agricultural and non-agricultural work was very elastic because of the presence of surplus labour. The implications of the model are that employment growth is largely driven by demand for labour, and real wages are likely to remain constant in the course of demand expansion. In later studies on rural labour, the existence of surplus labour found strong support, but the prediction of constant real wages was often violated, at least in the short run in local markets. The lesson was that, because of segmentation, the local labour markets did not always move together. That being said, historically and in the aggregate, real wages of rural labour did remain stable for long periods at a time, which is at least consistent with, if not a conclusive proof of, the existence of surplus labour and an elastic supply curve. At least in the first quarter of the twentieth century, there was definite evidence of a rise in real wages, as we have seen. And one of the two major sources on the nineteenth century suggested that the trend had started from 1891. Plausibly, the trend of a mild rise was indeed more general and enduring than historians have so far believed. Forces of expansion in demand for labour, in other words, outweighed the forces that shifted the supply curve outward. It is a more established fact that real wages were stagnant, if not falling, between 1930 and 1970. At the height of the Great Depression, real wages rose. But all regions experienced adverse adjustments from the late-1930s, leading to a long spell of stagnant wages. The spell was finally broken in 1970, when productivity became the chief source of growth in agriculture.

Agricultural labour

83

To understand these trends, we need to separate the major factors influencing labour demand and supply. The mainstream historiography focused on an induced increase in supply. In this view, between the census years 1881 and 1931, the supply curve of rural labour shifted out as a result of de-peasantization and de-industrialization. Irrespective of the shape of the supply curve, this should imply a rising proportion of labourers in the workforce. It should also imply a fall in real wages, the evidence of which is conflicting as we have seen. This story, furthermore, was unduly pessimistic on demand for labour, which should expand with cropped area and cropping intensity. Table 4.3 shows separately the rates of change in some of the major variables that might have influenced the agricultural labour market. There are four variables in the table. Two of these, population growth and net job-loss in non-agriculture, should depress wage-rate or indicate an outward shift in the supply curve. Cropped area and cropping intensity should indicate an outward shift in the demand curve for labour. The first point we need to notice is that the variables did not all change in the same direction in every period. Rather, their effects were usually contradictory, suggesting a picture of long-term stagnation rather than a significant push either way. A second point is the contrast between the late nineteenth century and the mid-twentieth century. The effect of land-extension was relatively large in the former and the effect of population

Table 4.3 Major factors affecting demand for and supply of agricultural labour Average annual change Population (%)

Cropped area (%)

Percentage cropped area irrigated (%)

Net job-loss in nonagriculture (million)a

0.6

2.7

0.2

0.2

1911–31 0.6

0.1

0.9

0.1

1951–71 2.7

0.9

1.2

−0.8

1881– 1911

Note a The difference between actual number engaged in small-scale industry in the end-year and the product of start-year percentage of workforce in small-scale industry multiplied by end-year workforce is calculated. The latter figure measures what should have been the number working in small-scale industry if industrial structure had not changed. This difference measures the number that notionally exited small-scale industry. The procedure is repeated for large-scale industry. The difference between the two figures is the extent of net job-loss. Minus sign indicates job-gain. The period 1931–51 is excluded because of poor quality of data and territorial changes in 1947.

growth large in the latter. Through the ninety years shown in the table, the importance of cropping intensity increased, and the negative effect of job-loss declined, both gradually. In the colonial period, net job-loss was positive, as the de-industrialization thesis suggests. However, extension of land and, later, cropping intensity, sustained labour demand. The period between 1865 and 1915 saw a steady rise in the net sown area and the emergence of significant pockets of demand for rural labour from within agriculture. Table 4.3 confirms the well-known stylized fact that improvements in efficiency or yield-

Rethinking economic change in India

84

per-acre played a rather small role in the agricultural growth in colonial India. Demand for labour was dependent on quantity of land, whereas quality of land was at best stagnant, at worst degrading. In turn, this is consistent with the negligible rates of investment in land per acre. Aggregate private investment (in rural and urban sectors) was 1 per cent of national income in the 1930s. The ratio of public investment was small too. Investment in agriculture was a minuscule share of national income.9 The dynamics changed somewhat in the first twenty years after independence. Both expansion in non-agriculture and land yield played a positive role, but population played a strongly negative role. Of course, it is hard to be precise on the magnitude of these effects on the labour market, unless we know the employment elasticities and, finally, the slopes of the demand and supply curves. The leap from micro-level data to macroeconomic functions is not an easy task, given data quality, and is best deferred. Wherever land quality and quantity remained unchanged, the long-term tendencies in real wage can be adequately explained with reference to the land-persons ratio. In Bengal, where land-persons ratio fell the most, real wage fell by more than 60 per cent between 1911 and 1936. In Madras and Bombay, there was long-term stagnation. In Punjab and UP, especially western UP, where land-persons ratio rose, real wages rose between 1900 and the middle of the 1930s. All regions, however, experienced adverse adjustments from the late-1930s. The spell was finally broken in 1970, when intensification became the chief source of growth in agriculture.

Fluctuations One of the most obvious features of real wage data is sharp fluctuations, a feature that persisted well into the post-independence period, and one that has been commented on by other authors (Rath and Joshi, 1966; Guha, 1985:141). Over 2–5 years, sometimes longer, real wage fluctuations were much more strongly influenced by the quality of the season than by money wages. The most disastrous fall in real wages occurred during famines. The Prices and Wages data record starkly the impact of the 1876 and 1896 famines. As Table 4.2 showed, the biggest falls in real wages occurred in regions worst affected by the famines. The first step to explaining real wage fluctuations, therefore, is to tackle this classic problematic: why did money wage change slowly over a five to seven-year horizon? A simple correlation exercise shows that money wage varied positively with the previous year’s price (r2, both variables in levels, is 0.37), and with a measure of expansion in cultivation, that is, land-population ratio (r2 0.49, both variables in first difference).10 Famines saw conversion of savings into consumption leading to a fall in the scale of cultivation. If acreage fell in one year, money wage fell too. At the same time, money wage was unresponsive to current price, which tended to shoot up in a harvest shock. These two circumstances produced the perverse response of real wage to seasonal conditions.11 If we assume that violent seasonal fluctuations were a constant feature of the region’s economy from ‘time immemoriar, it is easy to speculate that the pattern of short rise in real wages broken by sharp reversals was in fact an enduring condition. Over an indeterminate long run, this dynamic may well have played the role of a ‘built-in

Agricultural labour

85

depressor’ on real wages, borrowing a term from Daniel Thorner. It was briefly broken about 1900–20 as a result of the rarity of violent famines and expansion in land and irrigation. But by then another depressor was already beginning to act—demographic pressure.

Wage-share The phases in real wage trends generated similar movements in wage-share in rural income (Table 4.4). Not surprisingly, from the second decade onward, as net agricultural product was not increasing rapidly any more, the share of labour was expectedly more contested. Artisans

Table 4.4 Share of wages in agricultural income Share (%)a 1875

9.9b

1901

16.7

1911

14.1

1921

7.6

1931

14.4

1951

9.1

1961

8.4

1971

10.2

Sources: Colonial period: wage bill estimated as explained in the appendix. Domestic Products come from Heston (1983), Sivasubramonian (2000). For the period 1950–65, estimates by Mukherjee (1995). Notes a British India until 1938, and Indian Union between 1950 and 1965. b The 1875 figure is an estimate based on data supplied in Atkinson (1902). The results are not exactly comparable.

and rural labour constituted the largest of the poorer working masses in India. An opposition emerges in the experience of the artisan and rural labour in the early twentieth century. Artisans experienced a fall in labour-ratio, but a rise in income-ratio (if we exclude World War I when there was a setback). Rural labour experienced a rise in labour-ratio but fall in income-share (Table 4.5). What made these two types of poorer earners different, and yet made their experiences remarkably comparable? In the received narrative, small-scale industry was constrained by inadequate demand, leading to transfer of population from industry to agriculture and, in turn, agriculture is seen as burdened by an increased supply of labour. Now this story gives rise to two testable predictions: 1 There was transfer of population between the sectors.

Rethinking economic change in India

86

2 Industrial wage or productivity was relatively more depressed than agricultural while the transfer lasted. Neither proposition finds strong support. There is surprisingly little evidence of a conversion of artisans into cultivators until very late. And trends in relative products were exactly in the opposite direction to that predicted by received theory. Contrary to the prediction of the de-industrialization thesis that industry was doing badly and agriculture doing better, in reality, industry was doing better, whereas agriculture was beginning to be depressed. A more plausible story is that both sectors contained some surplus labour, and both persisted with extremely labour-intensive processes of production. The key difference between these was that agriculture suffered from a scarcity of capital, limits to organizational changes and diminishing returns to labour. Industry could stave off diminishing returns because of increasing industriousness and some capital accumulation (see Chapter 6).

Table 4.5 Shares in employment and income: agricultural labour and the artisan sector Sector Share in workforce (%)

Share in national income (%)

1910–20

1930–40

Small-scale industry

8.4 (1919)

8.2 (1938)

Agricultural labour

15.0 (1911)

19.0 (1931)

Small-scale industry

5.9 (1919)

9.1 (1938)

Agricultural labour

7.5 (1911)

5.0 (1931)

Levels and variations What did the level of wage mean in terms of quality of life? Using the principle that ‘in any country, the jail should form a standard of living below which the conditions of the free man should not fall’, Lyons (1920–22) estimated that a subsistence wage in 1919 prices, taking into account coarse grains, would amount to Rs 1 per day. The 1921 average money wage was about one-third this norm. The average family income of a mill worker in Bombay in 1921 was Rs 52 per month.12 With the average number of workers in the family at 1.2–1.5, the wage rate is just about equal to the minimum nutritional norm. The cotton mill worker and the field labourer were at two ends of the labour hierarchy. The distance was not easily bridgeable. Few rural unskilled labourers tried for jobs in the cotton mills. More usually they went to other unskilled semi-rural labour markets. These included seasonal factories like gins and presses, mines and public works. Seasonal factories in Madras paid an average daily wage of about Rs 0.40 in 1928, when the average agricultural wage was Rs 0.36. The unskilled labour wages in the coal mines of Bengal were marginally above the average rural wage—Rs 0.6 daily—in 1929. There are other regional datasets available that point to the same conclusions, that (a) the rural daily wage rate was about 70–80 per cent of the nearest comparable labour markets—

Agricultural labour

87

unskilled labour in industry and mines—and (b) the entire range of unskilled labour wages fell well below nutritional norms. There were differences in agricultural wages between regions, and these differentials persisted. Over the long period, rankings of the major provinces changed, but not dramatically. The relatively faster growth of wages in Punjab and UP continued in the interwar years, with the result that their rankings changed over the colonial period. The most dramatic fall in rankings occurred in Bombay. In principle, we would expect labour mobility to reduce variations over time because of easier supply conditions. On the other hand, the uneven pace of agricultural growth and urbanization can create unequal ‘demand-pulls’ upon regional wages. The net effect may be unpredictable in the medium term, but should be convergent in the long run, as the land-intensive agrarian growth process slows because land runs out, and population growth and migration further ease labour supply. The evidence on dispersion, however, is ambiguous. If we use the Prices and Wages figures as a benchmark, then statistical measures of dispersion between 1901 and 1951 show a rise. If we use the Datta Committee as the benchmark, these measures show a significant fall. The much narrower dispersion of wage figures in the former source suggests the hypothesis that the source tracked customary contracts rather more. Why did wages vary between regions? There can be several reasons. The first hypothesis deals with consumption pattern differences. In a labour-surplus economy, wages should follow subsistence closely. The hypothesis is that there was no variation in real wages, there was substantial real wage equality, but nominal wages varied because the grains consumed in each region had different prices. This is at first sight plausible, if we take rice as the staple of high-wage areas such as Bengal and Assam, millets the staple of Bombay and Madras, and wheat that of Punjab and Western UP. Rice regions had higher money wages, and rice was about 50 per cent costlier than millets. While this adjustment brings East, West and South closer to each other (but does not erase the variations completely), it raises the Northern wages, for wheat was cheaper than rice. Overall, it marginally smooths the distribution, but does not alter its basic character. There are, furthermore, two objections to this adjustment. First, in reality, the labourers in rice regions did not necessarily live on a diet of the average quality of rice. Cheaper grains were available. And second, the Agricultural Labour Enquiry dataset shows that average nutritional status varied in the same pattern as nominal wages, implying that consumption levels did vary. As expected, in nutritional terms the rice regions go down in ranking substantially. The second hypothesis is that, if the proportion of attached labour is high in a region, average casual labour wages may be relatively low in that region because the demand for such labour will be relatively low. However, this correlation is not strong for 1951 (r2 0.09, negative sign). Wages of casual labour, it would seem, were not driven by institutional factors but by peak-season labour demand. The third hypothesis centres on bargaining power. There is a significant correlation between inequality in land-holdings, a proxy for bargaining power and money wage (r2 0.47 and coefficient negative for 1951, Agricultural Labour Enquiry dataset). However, it is not easy to interpret this relationship. The critical problem is that we cannot take it for granted that in the wage-inequality relationship, the latter is necessarily the independent variable. Think of a region where wages are low because land-persons ratio is low. Here,

Rethinking economic change in India

88

the poorer peasants, with small plots of land, are likely to suffer from weak capacity to retain land because of poor creditworthiness. We might expect a relatively higher rate of land transfers in such a region and, therefore, greater inequality in holdings. But in this case, both low wage and high inequality are results of a third factor—the low asset base of the poorer peasant. This is not a hypothetical example. It has characterized Bihar and Eastern UP for a long time, and it characterized, in my reading, the Eastern Bengal peasantry in the late interwar period, a well-known case in the economic and political history of Bengal.13 The fourth hypothesis centres on productivity of land. Demand for labour per acre can be approximated by output value per acre. Supply of labour, on the other hand, depends on land-persons ratio. The correlation—real wage upon yield and land-persons ratio carried out for one year with the Agricultural Labour Enquiry (1954) dataset—is positive and significant (r2 0.39). Later research also finds a modified form of the association to be significant (Jose, 1988). A simple way to understand the relationship is that some crops—cotton, sugarcane, rice—were more labour-intensive as well as more profitable than others. In regions that had a significant acreage occupied by these crops the demand for person-days per acre was higher leading to a tendency for wages to be higher. Bombay in the late nineteenth century had relatively higher wages for the same reason. Parts of Bengal and UP had higher wages also for the same reason. Punjab, on the other hand, had higher wages because it was also labour-scarce relative to demand for labour.

Conclusion What does wage statistics tell us about narratives on agrarian labour in colonial India? The standard paradigm has postulated a double process consisting of ‘de-peasantization’ and ‘de-industrialization’, both creating an excess supply of labourers in rural India. One implication of such a process, if we see it as the sign of gathering distress, is a decline or depression in real wage in the long run. We do observe depression and decline, but not for all the colonial period. A ‘poor-getting-poorer’ story for colonial rural labour is again untenable as a general model. Official statistics suggest that there were two distinct phases in real wage depression, 1880–1900 and 1930–70, separated by an intermediate period when there was generally wage gain. Both phases, however, were broken internally by short fluctuations in wages. Between them, the two phases were very different. The first phase was characterized by money wage stability together with price rise. The second phase was characterized by money wage flexibility. Money wages usually, but not always or completely, followed prices. In other words, the second phase displays signs of a market for casual labourers at work, whereas the first phase does not. Herein lies the clue to an explanation of real wage depression in the first phase. Until 1900, a large proportion of rural workers was still on annual contracts. I believe that this fact imparted on the data a substantial stability in money wages, in one of the two contemporary official sources. The nature of the contract was changing thereafter. On the other hand, the most plausible explanation for wage depression in the second phase seems to be a mounting excess supply of labourers given a decisive downturn in land-persons

Agricultural labour

89

ratio from about 1920. In a general way, post-Depression money wage adjustments and trends in land scarcity were correlated across major regions. In the long run, real wages were influenced by demographic transition, land scarcity and stagnant land yields. Over shorter periods, real wage statistics also display effects of institutional change from patronage and caste-based relationships towards spot market and terminable contract. In the very short run, real wages depended on how quickly money wages reacted to prices, the response being variable over time. The focus on land-labour balance leads us also to a remarkable contrast in the experience of the two largest occupations of the poor: rural labour and artisanal work. Agriculture from about 1920 showed signs of suffering from diminishing returns to labour. Industry could stave off diminishing returns because of increasing industriousness and some capital accumulation. The next two chapters investigate these dynamics of the artisan industries.

Appendix Sources for wage data 1873–1912: Government of India, Director-General of Commercial Intelligence and Statistics, Prices and Wages in India, Calcutta, various years. 1891–1912: Report on an Enquiry into the Causes of Rise in Prices in India. 1920–51: India (1954); and quinquennial wage surveys in major provinces. 1951–64: Mukherjee (1995). Some problems and adjustments relating to wage statistics The Prices and Wages report monthly wage, later sources report daily wage. I assume the monthly wage refers to the busy season when for all thirty days in a month the worker might find work. Tables 4.1–4.3 use monthly wages, but the converted daily wages can be found in Table 4.6 below. When using the Datta Committee figures, we concentrate on the three basic operations: ploughing, sowing and, with rice, transplanting. Where applicable, I employ intensity of employment data from the Agricultural Labour Enquiry, assuming intensity to be a function of the relatively invariant parameter, crop mix. The data on female wage rate is generally less detailed than male wage rate. For the earlier years, Prices and Wages do not mention which sex the wage refers to. Where not mentioned explicitly, I take wage to refer to male wages. Elsewhere, I use the femalemale wage proportion averaged over four provinces where detailed data on female wage is available, upon male wage rate, to derive the corresponding female wage rate. The average wage for all India is a weighted average over all states/provinces for which we have data. For colonial period, only one set of weights are used, 1901 agricultural labour population, to avoid biases implicit in the census occupational statistics.

Rethinking economic change in India

90

Wage estimates: 1873–1912 The 1873–1901 figures based on Prices and Wages are averages of six major regions (Greater Bengal, Greater Punjab, Bombay, Madras, Central Provinces and Berar, United Provinces). The 1891–1912 figures based on the Datta Committee are averages of six major regions. For 1891, the two datasets produce remarkably similar figures, for 1901 these produce very different figures. The index numbers are calculated taking the Datta Committee figure for 1891. The price index used is the weighted agricultural price index in M.McAlpin, ‘Price Movements and Fluctuations in Economic Activity’ in Kumar (1983b). Wage-estimates: 1916–46 The average wages shown in Figure 4.2 are based on forty-two data points in the six major provinces. The sources for these figures, and Table 4.6 below, are: a Bombay. b Ghosh (1969) reports real wages for field labourers in the five other provinces, based on the quinquennial wage censuses that replaced the Prices and Wages. The Bengal figures were cross-checked by those reported by Radhakamal Mukherjee at the Royal Commission on Agriculture Evidences. They tally closely. UP and Punjab data were cross-checked with two other sources: Chaturvedi (1947), Bhattacharya (1981). The provincial average figures are averaged weighted by respective labour populations. The figures that appear in the money wage column in Table 4.6 are these weighted averages. Wage estimates: 1951–68 The 1951 figure is from Agricultural Labour Enquiry. Raw data are daily wage, converted into monthly. Standard price indices covering pre- and post-1947 years are not available. The extent of price increase between 1946, the last year of the previous series and 1951, part of a new series, may not be precise. The estimates for 1951–68 extrapolate the 1951 real wage to the series on average real income taken from ‘Real income of agricultural labourer’ in Mukherjee (1995). GDP in agriculture and wage-ratio The figures for the colonial period available in the standard sources are for all India, whereas the labour force figures refer to British India. I scale down GDP agriculture by the ratio of population in British India vis-à-vis the states. The wage bill is estimated by multiplying wage rate with number of workers and average work intensity (for 1951 in all cases). This is done for male and female workers separately.

Agricultural labour

91

Table 4.6 Weighted average money and real wage of agricultural labourers in British India (Rs/day, real wage in 1873 prices) Year Monday wage

Real wage

Year Money wage

Real wage

Year

Monday wage

Real wage

Source: Prices and Wages Statistics

Source: Datta Committee

Source: Quinquennial wage surveys, as reported in Ghosh and other works

1873 0.157

0.157

1892 0.173

0.130

1916

0.306

0.147

1874 0.155

0.130

1893 0.173

0.134

1921

0.333

0.114

1875 0.161

0.148

1894 0.176

0.144

1926

0.460

0.162

1876 0.171

0.171

1895 0.181

0.151

1931

0.270

0.171

1877 0.168

0.127

1896 0.190

0.145

1936

0.243

0.125

1878 0.171

0.117

1897 0.197

0.127

1941

0.187

0.161

1879 0.172

0.117

1898 0.200

0.159

1946

0.931

0.264

1880 0.166

0.138

1899 0.206

0.168

1881 0.167

0.168

1900 0.216

0.150

1882 0.173

0.182

1901 0.223

0.163

1883 0.171

0.180

1902 0.230

0.175

1951

1884 0.177

0.162

1903 0.233

0.190

1955

0.195

1885 0.176

0.164

1904 0.244

0.194

1960

0.193

1886 0.176

0.167

1905 0.254

0.186

1965

0.171

1887 0.175

0.173

1906 0.263

0.164

1968

0.189

1888 0.177

0.164

1907 0.273

0.161

1889 0.184

0.148

1908 0.283

0.160

1890 0.180

0.141

1909 0.290

0.160

1891 0.184

0.155

1910 0.294

0.194

Source: ALE and mukherjee’s series on real income’s 1.094

0.219

5 Was there an industrial decline in India in the early nineteenth century? Many, perhaps most, historians believe in a significant industrial decline in colonial India as a result of the imports of British manufactures. Some place such a phenomenon in the early nineteenth century, others believe it continued in the late nineteenth and early twentieth century. An industrial decline can lead to a large fall in aggregate demand and, thus, cause an economic regression. An industrial decline can push industrial workers into agriculture, reduce agricultural wages, reduce demand for industrial goods and intensify rural poverty. On the other hand, an industrial decline that signifies the availability of a cheaper imported alternative might improve real wages and purchasing power in the nineteenth century. If the adverse effects were stronger, an industrial decline was not just an industrial decline, but also stood for a destruction of the potentials to industrialize. In this sense, it is sometimes said that India dries whereas Britain ‘industrialized’ in the nineteenth century. The idea of an industrial decline accompanied by some or all of these bad effects— aggregate demand crisis, rural unemployment, rise in poverty—has travelled in the last fifty years from nationalist axioms to conjectural historiography to canonical development models.1 In the course of this journey, the idea acquired a life of its own, and was therefore never tested rigorously. Did an industrial decline occur? Did it have these effects? Neither question is an easy one to answer. The usual measure of decline is employment. In employment terms, the scale of Indian industry did shrink between 1800 and 1950. But a fall in employment does not necessarily mean a fall in demand. It can mean an increase in efficiency. There is evidence that in the last fifty years of colonial rule, the fall in employment did derive to a significant extent from efficiency. For the period before, we cannot conduct a test as to why the fall occurred. Did the decline have those bad effects? In development models, it is usually presumed it did. But the facts of economic history are ambivalent. The propensity of an industrial decline to cause rural unemployment and poverty was limited in the nineteenth century when agriculture had the capacity to absorb people because land was available. The propensity of an industrial decline to cause rural distress was greater from the inter-war period when agriculture was saturated and the rate of population growth began to increase. But this was a time when the very fact of an industrial decline was open to question. The propensity of an industrial decline to cause distress can also be mediated by institutional structure and gains for consumers—issues that remain speculative if touched at all. The timing, as the above discussion suggests, is an important issue. When the decline happened has an influence on how severe its effect might be. There is confusion in the relevant scholarship over timing. Some scholars suggested that a decline occurred in the

Was there an industrial decline in India

93

late nineteenth and early twentieth century. This view has been revised, and the current consensus is that in the later part of the colonial period, the most vulnerable section of manufacturing—small-scale industry—was actually doing well. Another group of scholars argued recently that a significant decline occurred in the early nineteenth century, unwittingly lending support to the ‘imperialist’ claim that colonialism made domestic enterprise progressively better-off. Was there an industrial decline in the early nineteenth century? This chapter answers that a fall in employment did occur in the nineteenth century, but it was more evenly distributed over time than has been believed, and was part of a process that entailed demand gain rather than loss. The chapter is divided into four sections. The first delves into the origin of the idea of an industrial decline. The second deals with the problem of timing, the third with welfare loss and the fourth section interprets the results and summarizes the main arguments.

The origin and evolution of an idea In 1750, India was supplying about a quarter of the world’s industrial output, and possibly a larger percentage of world textile exports. By 1900, India’s share in world industrial production had dropped below 5 per cent. Historians have seen in this retreat a counterpart of what Britain experienced in the same time-span. While the West ‘industrialized’, India experienced the inverse of industrialization, or ‘deindustrialization’. Indian traditional textiles seem to illustrate this downturn the best. The idea of a decline has been reinforced by studies on early modern coastal trade that project India to be a proto-industrializing region before technology imposed a rearrangement of comparative advantage (for example, Perlin, 1983). At a broad level, de-industrialization hypothesizes a fall in industrial employment, output and income, and a shift of resources from industry to agriculture. This broad empirical generalization, however, accommodates at least three sets of arguments that lend de-industrialization the significance that it carries in the historiography of colonial India. The most general of these is a thesis about comparative development. By defining ‘industrialization’ to mean substitution of labour-intensive by capital-intensive products, it can be argued that industrialization in Britain impeded industrialization in India. This sense of de-industrialization as the distorted mirror image of industrialization has been articulated in Indian economic history (for example, Bagchi, 1978; Chandra, 1968). The idea that the third world saw an antithesis of industrialization was also taken up in a scholarship on the origins of underdevelopment.2 Some recent works on Indian industrial history call into question this usage of ‘de-industrialization’. National income estimates reveal growth in real income per worker in artisanal industry between 1900 and 1946 (Sivasubramonian, 2000). These works suggest that a certain kind of industrialization took shape in India from at least the mid-nineteenth century that relied on cheap skilled labour, natural resources and a particular consumption regime. It did not take place by getting rid of artisans like the hand-weaver. Rather artisans made a positive contribution in it (Roy, 1999).

Rethinking economic change in India

94

The second argument is that there was, as a result of de-industrialization, a decline in overall economic activity. A special crisis led to a general one. This effect occurred in two ways, an aggregate demand crisis (Bagchi, 1978), and a peasantization of the artisanate that ‘led to a structure of society which has often been characterized as “semifeudal”’ (Bagchi, 1982:82). Morris questions the point about aggregate demand by pointing out that the big fall in world textile prices in the nineteenth century plausibly raised total demand for cloth in India (Morris, 1968). Yarn prices more than halved between 1819 and 1880, whereas cloth prices fell by a factor of three to four. Morris implicitly argues that the positive forces outweighed the negative ones of competitive destruction of income and jobs. The question of net gain or net loss, however, has not been settled in a satisfactory way because, among other reasons of unavailability of income data on the early nineteenth century. The third set of arguments concerns timing. Precisely when an industrial decline occurred is of some importance to the question about impact. The most famous of the nationalist writers who first raised the industrial decline issue were writing near the turn of the century and it appears that they were talking about a process that had already finished happening. Whether or not the process really happened in the first half of the nineteenth century, the period implicit in the nationalist discourse, or in the second half has a bearing on how acute its effect might have been. The first half is said to have seen a number of adverse changes—agrarian depression, suppression of local political elite, loss of the world market for Indian textiles and competition for Indian textiles in the domestic market.3 All these can in principle intensify a demand crisis. After 1850, on the other hand, some positive factors began to play an ameliorating role. These factors would include a fall in material cost for handloom weavers, rise in rural demand as a result of growth in agricultural markets at home and abroad, and rise of the mill industry. If industrial decline did concentrate in the first half, it is plausible that it aggravated a depression. If the decline was more evenly spread over time, it is equally plausible that its effects were partly and increasing outweighed by the positive factors. Indeed, national income estimates, which began to be available from the 1870s, show no sign of a crisis either in industry or the economy, so that the latter view is valid at least for the post1870s. Timing is an issue on which there is considerable confusion. In historical scholarship, there is some recognition that timing matters (for example, Little, 1982), and that the early and the later nineteenth century are qualitatively different periods. However, a large part of empirical testing has used the censuses that cover the period 1881–1931. By using this dataset, Patel (1952) and Chattopadhyay (1975) generalize de-industrialization to the later colonial period, and Thorner (1962) and Krishnamurty (1972) dispute the reliability of that dataset. Other more innovative tests of the hypothesis cover the early nineteenth century but do not necessarily enable us to narrow down the period of occurrence further.4 Responding to Daniel Thorner, who found no evidence of an industrial decline in census data (1881–1931), Chandra (1968) argued for setting the clock back to the early nineteenth century. Responding to Roy (1999) more recently, who suggested that artisans contributed to industrialization in the later colonial period, three discussants demanded that the ‘temporal location of the debate’ be shifted much further back (Ray, 2000; Markovits, 2000; Sarkar, 2003). The time question, therefore, remains open and, perhaps, strategic too.

Was there an industrial decline in India

95

In short, to analyse the nineteenth-century experience we need to answer two questions: when did an industrial decline begin and when did it end? While textile producers lost jobs, did they also lose an income large enough to outweigh the gains from the cheapening of manufactures? This chapter is an attempt to answer these two questions. In investigating the time question, it relies on recent developments in textile history scholarship as yet unutilized in this context. And in examining the second question, this chapter conducts an experiment.

The timetable The problem with settling these issues with direct evidence is that the period 1800–50 is rather ill-served with relevant primary material, leaving the statistically oriented historian ‘very much in the dark’, as Colin Simmons concluded (Simmons, 1985:606). The principal source of industrial statistics until the nineteenth century is the records of the Indian Ocean trade in the major European archives. By 1800, the trade was changing in composition and direction, and textiles were no longer an important item of export. On the other hand, the East India Company was yet to settle down as a territorial power, having to fight small and large battles on almost all fronts except the eastern one. Governance records at this stage have a bias for East Indian Company finances rather than the local economic situation. Travelogues, an important source for the early nineteenth century, were impressionistic on industry. Survey reports such as Francis Buchanan-Hamilton’s famous journals contain important information, but their reliability is disputed and impossible to cross-check. In recent years, a few detailed regional histories and two general estimates have redressed the gap to some extent. But these materials do not suggest a simple story when put together. The two attempts at estimating employment using pre-census data cited above both confirm decline (Bagchi, 1978; Twomey, 1983). Bagchi compares two separate estimates for 1820 and 1901, to conclude a large fall in employment in the Bihar region of eastern India. The authenticity of the earlier estimate has been rightly questioned (Orr, 1980; Vicziany, 1979). Twomey conducts a more persuasive test, and confirms a large decline in textile employment, so large in fact that no conceivable counter-trend in other industries could possibly have compensated for the fall in employment. From these estimates, the question as to whether or not there was an industrial decline is easily answered, but not the question when. We cannot say anything about the corresponding income loss either. The evidence on industrial decline in prior works on the economic history of India and its major regions is surprisingly thin. To take the canonical work of Romesh Dutt, the chapter on manufactures does no more than reproduce sections from the evidence before a Select Committee of the East India Company by several top-ranking British officials and a few British manufacturers/importers (Dutt, 1906). The relevant parts of this citation are generally statements based on hearsay and prognosis. One product, Dacca muslins, flgures too conspicuously in this evidence. Choksey (1945) on Bombay-Deccan follows Dutt in his method, and is no more convincing. The economic histories of Madras Presidency and Bengal by A.S.Raju and H.R.Ghosal respectively discuss decline almost entirely relying on reports of the dwindling export trade. Comments on Lancashire

Rethinking economic change in India

96

competition in the home market are again impressionistic (Raju, 1941; Ghosal, 1966). The factual basis for ‘industrial decay’ in Raman Rao’s (1958) economic history of Andhra Pradesh is even less substantial. It is noteworthy that comments on industrial trends in the regional economy surveys in the Cambridge Economic History of India have largely recycled the partial and unreliable research contained in these secondary sources. It is not improbable that a crisis did happen and went unnoticed by the East India Company officials too busy trying to consolidate their command over an expanding territory. But there can be other, more plausible explanations why the phenomenon did not become an issue until later. My hypothesis is that the decline was milder if measured as net rather than gross decline and if measured as income rather than employment. In fact, even by 1900 or 1920, decline in hand-spinning was not seen either in administrative reports or in the media as the kind of climacteric that M.K.Gandhi made it out to be when using the hand-spinner as a powerful nationalist symbol. A fresh crop of research on regional textile histories appeared between 1985 and 1995. These works confirm that neither the decline in hand-spinning nor the decline in handloom weaving happened in a skewed manner. The only development that can be unambiguously periodized in the first quarter of the century is the fall of Indian textile exports to Europe. But, notwithstanding India’s impressive presence in world trade, it is not likely that foreign trade had occupied a major part of the textile economy within India. At the end of the eighteenth century, cloth export from India amounted to about fifty million yards (Twomey, 1983), whereas total production within India could not have been smaller than 1,800–2,000 million yards. Parthasarathi (2001) estimates textile exports from South India in the late eighteenth century at about 1,800 tons, whereas the production of cloth in South India could not have been smaller than 50,000 tons. Foreign trade, in other words, was marginal to the textile economy overall. The impact of Lancashire in the home market is a different story. English yarn first reached Indian ports in the second decade of the nineteenth century. Quantity and value of imports rose rapidly from 1820. Exports to Calcutta were much larger in volume than exports to Bombay or Madras until late in the nineteenth century. Studies of southern and central India suggest that the decline in hand-spinning was rather mild in 1820–50, and accelerated only after 1865, about which time the major railway routes opened.5 Bengal, being more accessible by sea than central India or the Deccan, may well show a different pattern, one which conforms more closely to classical de-industrialization timetable. But Bengal was probably an exception. As late as 1870, the share of imports in total yarn used in Madras Presidency was 37 per cent. In 1890 the percentage was fifty-five (Specker, 1989). A similar point can be made about weaving. Weaving, indeed, never really suffered to the same extent as hand-spinning. There was almost certainly some decline, but of an uncertain magnitude, since total consumption might indeed have increased in the nineteenth century. The dissimilarity in the experiences of spinning and weaving has been discussed in textile history scholarship. If there was a decline, how large was it, and when did it peak? Dharma Kumar cites J.G.Borpujari’s research that suggests the decline was not of serious order before 1865 (Kumar, 1972). Data gathered by Specker (1989) shows that, but for famines and epidemics, nearly every attempt at a loom count in early nineteenth-century Madras Presidency found the number of looms on the increase. His description of South Indian

Was there an industrial decline in India

97

weaving suggests that in the period 1800–50 imported cloth was only a minor factor in the experience of weaving. Famines, loss of the export market and decline of the elites were more important factors. However, contrary to what Specker seems to suggest, the latter two factors could not be quantitatively very significant. The experience of central India in this regard has been studied by Harnetty (1991) and Guha (1989), both suggesting that it was only after the major railway routes opened c. 1870 that the impact on cloth production became serious. The Maratha regimes immediately preceding the British in central India (c. 1775–1815) patronized urban handicrafts on a large scale, leading to a number of weaver settlements springing up in small and large towns in the region. Some of these languished after the change of regime, but some of the others more fortunately served by mass markets and transportation routes survived and expanded. Harnetty cites Robert Jenkins, the British Resident at the Nagpur court (1807–27), to describe the Koshtis of the Berar towns in this period. They enjoyed a ‘high standard of living’ and were ‘an important class in the towns’. They produced high-valued goldthread-bordered garments of very fine texture. Their leading men were employers of labour, who had to go through a long and arduous apprenticeship in the craft. Change, however, was beginning to come from the end of the second quarter of the century. The extinction of the native courts reduced the demand for such articles, changes in fashion led to a preference for lighter and cheaper goods, and competition from British cloth in the plainer varieties began to be felt. And yet, not until after the Cotton Famine of the 1860s did the third factor take serious proportions. Quite likely, the total handloom production in the region was growing between 1825 and 1865, thanks to an agrarian boom in several zones of central India that stimulated demand for coarse cotton cloth. The major dynamics in central India was on the demand side, a decline of elite consumption, while the mass market was growing for cheaper varieties. But the same weavers who supplied one market were not necessarily adept at serving the other. As a result, central India in this period saw migration of Kori and Momin weavers from further north. Trade statistics suggest that it was only after 1867, when the Bombay-Nagpur railway was completed that a surge of import of cloth into the region began. Interestingly, however, the scale of imports rose rapidly between 1867 and 1885, and stabilized thereafter. In a rather short span of time, substitution of handloom weaving by English cloth had reached its limits. Based on these reports, we are driven to two conclusions. First, Bengal and the rest of India might not show similar chronology of decline and survival. In Bengal, decline in spinning and weaving probably happened at a sharper pace and earlier than elsewhere. In southern and central India, both spinning and weaving felt the pressure of import from the time of the Cotton Famine. From that decade, the positive forces working on handloom cloth demand had begun to consolidate. Agricultural exports and agrarian commercialization strengthened local demand, and export of textiles from South India to the new settlements of expatriate Indians began to increase. Specker (1989), for example, shows a remarkable 40 per cent increase in the total number of looms in south India between the 1850s and the 1870s. National income statistics start from this decade, and show no sign of an overall industrial crisis. By every indirect index that we can use, the period 1870–1914 saw net expansion rather than contraction in weaving. If we postulate that the major part of the decline in spinning and weaving occurred after 1865 in most

Rethinking economic change in India

98

parts of India except Bengal, we must accept the possibility that the positive forces substantially offset the negative ones. These four observations—large employment loss, little visible effect on aggregate demand, differential impact of imports on spinning and weaving and an expansion in cloth consumption—suggest that whatever employment loss there was did not generate a corresponding income loss. This conjecture cannot be directly tested, but it can be defended with a notional accounting.

An aggregate demand crisis? A notional accounting of the gains and losses from textile imports would have to distinguish four types of gains and losses: 1 Hand-spinners’ loss of income as the scale of handloom cloth production out of handspun yarn falls. 2 Hand-weavers’ loss of income as the scale of handloom cloth production falls; handweavers’ loss of income as the income per yard from cloth with handspun yarn falls. 3 Hand-weavers’ gain in income as the cost of raw material used in handloom cloth production out of machine-spun yarn falls. 4 Consumers’ gain in income as the price of cloth falls. The first three constitute, assuming the aggregate supply curve is horizontal and shifts as cloth prices fall, the net change in producers’ surplus. The fourth measures the net change in consumers’ surplus. There was also a flfth kind of effect. The early nineteenth century also saw assumption of British power in several regional states that had once been major buyers of luxury articles. Generally urban enterprise suffered all round because of the enervation of the older elite, a process studied in the context of the early nineteenthcentury Maratha principalities of central India by Harnetty. We assume that competition between hand-loom and power-loom cloth ensured that the same price ruled at all sectors, and competition within the hand-loom industry ensured that the same profit rate ruled for all handlooms, irrespective of what yarn was used. The notion of weavers’ gaining from fall in raw material cost is problematical. Weavers gain only to the extent they do not pass on the full effect of the fall to the consumers. We assume, however, that competition ensured that they did pass on the full extent of the cost reduction to the consumers, in which case, weavers gained nothing. We also assume that the total market size did not change and consumption remained at the level of the initial production. In that case, the consumer gained by production times change in price of cloth. We use the following symbols: k=share of hand-loom production made out of handspun yarn Y=Hand-loom production (=consumption) m=spinners’ share in value-added 1−m=weavers’ share in value-added s=proportion of value-added in price of cloth per yard P=price of cloth per yard Between 1800 and 1900, the spinners and the weavers suffered losses to the extent of

Was there an industrial decline in India

99

ms(P1800Y1800k1800−P1900Y1900k1900)+(1−m)s(P1800Y1800−P1900Y1900) For spinners’ and weavers’ combined loss of income to exceed what the consumers gained out of the same quantity of cloth requires ms(P1800Y1800k1800−P1900Y1900k1900)+(1−m)s(P1800Y1800−P1900Y1900)> –Y1800∆P Define a maximum-stress-scenario in which k1800=1, k1900=0, Y1900=0. In that case, the condition simplifies to −s>gP, where gP is ∆P/P In the more general case, the condition is ms∆(PYk)+(1−m)s[∆(PY)]>Y∆P Writing out fully, dividing through by PY, and writing gY and gP for ∆Y/Y and ∆P/P respectively, the expression becomes ms[(1+∆k)(gY+gP+gYgP)+∆k]+(1−m)s(gY+gP+gYgP)>gP In the maximum-stress-scenario, that is, gY= −1 and ∆k= −1, the expression simplifies to −s>gP How much did cloth prices change? The estimates here vary, and the early nineteenth century averages remain obscure. The average price used here —Rs 0.5 per yard c. 1820—is no more than illustrative. How is this figure arrived at? More reliable figures for 1880 suggest that average cotton cloth prices per yard ranged between Rs 0.13 (Indian mill cloth export) and Rs 0.20 (British export). Hand-loom average was likely to have been between these points. Taking these figures as benchmark, and assuming that average prices in 1880 were about one-third of 1820 levels, we settle on Rs 0.50 as an acceptable indicative price for 1820. With gP= −2/3, the condition −s>gP requires income/production ratio to be impossibly high. Realistically, the percentage would rarely exceed 30, and ordinarily be much smaller. With gY= −1/2, which was closer to reality (Twomey, 1983), the condition for losses to exceed gains becomes −s>2gP/(2m−(1−m)(gP−1)) With m set at 0.5 and gP at −2/3, the condition becomes s>0.7 Let us assume now that a certain part, b, of Y was produced as ‘luxury’ at a price aP, where a is greater than one, and b a positive fraction. This aristocratic demand, we assume, became zero in the terminal year. Reworking the expression above, the condition for there to be a net loss of income under maximum-stress becomes:

Rethinking economic change in India

100

−s>gP/(1+ab) With a=2, b=0.2, gP=−2/3, s>0.48 With gY=−1/2, the condition for losses to exceed gains becomes −s>2gP/(2m−(1−m+ab)(gP=−1)) With a=2, b=0.2, gP= −2/3, s>0.53 The right-hand side reduces to a realistic 0.3 at a=3 and b=0.4, that is, with luxury prices at three times the ordinary goods prices on average, and about 600 million yards of luxury cloth production for every 1,400 million yards of ordinary cloth production. These ratios are on the very high side, but cannot be ruled out offhand. The assumption, however, that luxury cloth disappeared totally is exaggerated and, on that ground, we can continue to assert that the net income loss was not positive even after accounting for a large luxury cloth production. In other words, the condition for net loss to be positive remains too stringent no matter how much we relax the assumptions. Leaving aside the complication created by luxury production, let us assume the following values: P1800=Rs 0.5/yard, P1880=0.2, m=0.5, s=0.4; and take Twomey’s estimates of production and the proportion of handspun yarn, Y1800=2,000 yards, Y1880=1,000 yards, k1800=1, and k1880=0.1. The gains for the consumer is approximately Rs 600 million, and the loss for the producers approximately Rs 316 million, or a realistic ratio of 2:1. Luxury production might bring the ratio down to 1.5:1. Basically these results are driven by an imbalance between two sets of ratio. The spinners’ income was a small part of the value-added/price per yard, whereas the drop in material cost and cloth prices were much larger proportions of price. The spinners’ income could not be more than 10–20 per cent of the price, but the drop in material cost was at least 50–67 per cent. The former was the basis for the income lost. The latter was the basis for the income gained, and the latter tended to outweigh the former.

Conclusion Was there an industrial decline in nineteenth-century India? Indeed there was. As the calculation above shows, at the assumed average price of cloth and other ratios, approximately Rs 300 million of income was lost over 70–80 years (c. 1830–1900), an annual average decline of Rs 4–5 million. Assuming a population of 150 million and an average income of Rs 20, this works out to about 0.1–0.2 per cent of annual income, and possibly 1–2 per cent of industrial income every year, sustained over a long period of time. Undoubtedly, there was an income effect of fall in price of manufactures on the demand for other industrial goods. But a part of this income effect benefited imports, and for the balance to fully compensate for the loss of demand requires making rather strong assumptions about income elasticity. The net loss of demand was surely smaller than what these magnitudes imply, but possibly still positive. Combine this conclusion with

Was there an industrial decline in India

101

estimates of employment loss, and we get a reasonable picture of what happened in the early nineteenth-century manufacturing. In this purely empirical sense, there was industrial decline in nineteenth-century India. We stumble on the qualitative questions of time and impact. Regional histories suggest that in many parts of India, weaving, and even hand-spinning, began to decline only after expansion of the railway system brought down transport costs significantly. Visible decline of weaving was concentrated in a much later period, 1860–90. Yet national income statistics show growth in income in this period. Clearly, the compensating effects of the rise of new enterprise and even the income effect of a fall in cloth prices on handloom weaving itself seem to have outweighed any substitution effect in the late nineteenth century at least. Did the decline entail a general economic crisis and a large welfare loss? The answer is almost certainly ‘no’. We can argue this on two grounds, absorptive capacity of the non-industrial economy, and the disparity between employment loss and income loss. At 1850, the economy had considerable absorptive capacity. There was plenty of land available. Many artisans joined the flows of overseas labour migration. That being said, a shift of workers from industry to agriculture remains no more than a plausible conjecture. It is typical of the period that we know very little directly on ‘peasantization’ of the artisan. The process could not have been either automatic or easy. Many artisans did not have the financial or social capital needed to occupy and develop new land frontiers. Caste might well have been a barrier to occupational diversification, if not an impenetrable one. Many among those who lost employment were domestic labour, for whom opportunities were limited. Productive labour absorption in agriculture, therefore, cannot be simply assumed to have happened. The stronger ground to dispute a catastrophic impact is that though the decline was very large in number of workers, it was rather small in cost of labour-time and, therefore, in terms of income. A full-time job loss of 4–5 million in 60–70 years works out to at least 30 per cent of the industrial population; whereas the income loss works out to about 1–2 per cent of annual industrial income. How did such a large employment loss as this one cause such a small crisis in income? The answer is that spinning had always engaged labour time that was already in surplus and, therefore, poorly paid. The small payment per yard to the spinner was distributed thinly over an extraordinarily large population. Spinning was labour-intensive. Approximately 2–3 spinners’ full-time labour was needed to keep one weaver fully engaged. Using this ratio, Twomey estimates the full-time employment requirement in hand-spinning at 4–5 million in 1800. Assuming weavers and spinners received the same amount of money per yard, clearly, whatever amount weavers earned per month or year would have to support 2–3 spinners. The realistic income of a weaver was about Rs 5 per month. At that standard, a spinner’s income could not exceed Rs 2 per month, which was well below subsistence. Realistically, the income was much lower still, for spinning almost certainly received smaller payment per yard than weaving. Given the excessive labour-intensity of spinning, the only way India’s textile economy could work was by utilizing an equally large pool of slack labour. The rhythm of monsoon agriculture in tropical zones created a long off-season in agriculture, which was the main factor that sustained the spinning economy. Not surprisingly, descriptions of the spinner community generally show two types to be predominant: women of peasant

Rethinking economic change in India

102

households (not weaver households, who ordinarily found work within the weaving process), and rural labour castes who engaged in some manufacturing only part-time. It is possible, indeed likely, that if we had a good occupational census for 1750 or 1800, the percentage of industrial workers in the total workforce would be as high as 25– 30, which collapsed to 8–9 at 1900. But it would amount to an exaggeration of reality to deduce from this evidence that an ‘industrial’ economy ‘deindustrialized’ in the process of this change—the large proportion of industrial workers at 1800 did not signify industrial potentials, enough incomes, or promising rates of saving and investment. All it signified was an economy with an extraordinary burden of surplus manpower. The early colonial economy was one in which industrial tasks were performed as a form of service by the most primitive methods utilizing abundant supply of slack labour time. The income loss from the end of this system could not be very large. And the employment loss that resulted from the decline of hand-spinning can be called ‘de-industrialization’ only if we are prepared to call the world where domestic workers sweat for less than subsistence wage in a whole range of semi-skilled tasks an ‘industrial’ society.

6 Labour-intensive industrialization From the last quarter of the nineteenth century, small-scale labour-intensive industry experienced an increase in total and average income. What conditions made this growth possible? The present chapter shows that the growth arose from market segmentation and innovations in technology, organization and products. By 1900, the overwhelming majority of India’s manufacturing workers were employed in industries that did not use either machinery or large factories. A hundred years later, over two-third of manufacturing employment remained intensive in manual labour and craftsmanship. A subset of this large segment had roots in traditional industry, that is, employed or adapted skills, training systems, recruitment systems and production relations that originated in pre-colonial artisan traditions. That labour-intensive industry should survive in a region with a large population and low wages is not surprising. However, standard works on Indian industrialization have rather overlooked labourintensive industry and focused on the experience of a small set of relatively mechanized mills.1 Such oversight often derives from two premises. First, industrialization means mechanization, or a significant rise in capital-intensity. And second, such a process puts an end to labour-intensive industry. While Britain experienced the dual process, India, it is said, experienced mainly the second. These two beliefs played a large role in policy formulation in post-colonial India. Industrialization became a matter of raising the rate of investment, and designing large factory yards with gigantic equipment—a vision Jawaharlal Nehru called ‘temples of modern India’. This vision did not go unchallenged. Followers of M.K.Gandhi were concerned that such pursuits would destroy the handicrafts, which for them represented, not exactly commercially viable enterprises, but a harmonious rural utopia threatened by modernism. Independent India, therefore, adopted a paradoxical industrial policy. On one side, the competitive strengths of labourintensive industry were underrated and not utilized fully.2 On the other side, there arose capital-intensive units that could survive only by dint of progressively higher tariff walls. When trade and industrial policy liberalized in the 1980s, these symbols of socialist industrialization began to crumble, or had to embark on a tortuous and long-drawn adaptation process. Labour-intensive industry too included segments that suffered decline in the presence of competition. However, another segment surged dramatically by being able to access world markets more easily than before. Behind the old regime, historiography played a powerful role. The key tenets of policy derived from an understanding of what had happened in history. If the policy turned out to be unsound, so was the premise on which it was based. Contradicting the view that labour-intensive industry was destined to fade out, not only did such industry survive in colonial India, but it also generated healthy growth in income per worker for as long as we can measure it. How did the artisan qua artisan experience income growth? Driven by

Rethinking economic change in India

104

a question such as this, historians have recently begun to formulate a different narrative of industrialization, including rather than excluding the artisan.3 It is now accepted that there was never an outright displacement of labour-intensive industry, rather there was some obsolescence and some survival. These survival stories suggest three important lessons. First, there is perhaps no one universal model of industrialization. Rather, industrialization can be shaped by local factors, for example, a consumption pattern that sustains craftsmanship. Survival stories do not represent anomalies in a grand narrative of industrialization, but motivate us to redefine the term. Second, artisans were not a passive target of forces of competition imposed on them from the outside, but they responded creatively to these forces, and shaped these very forces to some extent. They could produce different goods of different quality, access new markets, cut costs, make changes in tools and processes, shift location, and share knowledge, information and risks. Third, even as many artisans lost livelihood because of competition with machinery, industrialization potentials based on craft skills were not distorted thereby. The new scholarship, thus, supports the work of economic historians who recognize that a straightforward substitution of labour-intensive industry did not occur globally either, because such industries possess unique strengths (Piore and Sabel, 1990; Sabel and Zeitlin, 1985). They can continue even in a scenario of relatively high wages arising from these strengths. In this chapter, I shall discuss the change in perspective. The rest of the chapter has five sections, the first four dealing with statistical overview, demand, supply and weaving respectively. The last draws some general lessons.

Measuring lahour-intensive industrialization Declining employment-shares of industry in the workforce have figured rather too prominently in discussions on Indian industrialization. But employment-share is a narrow measure. An increasing importance of industry in the domestic economy can be measured by at least three indices: a rise in the share of industry in workforce, a rise in the share of industry in income and a relatively large contribution of industry to productivity growth. The first ratio measures industrialization in terms of number of livelihoods, the second is a composite of livelihoods and their earning capacity and the third is purely a measure of earning capacity. The three indices should be considered together, because it is possible that they do not move in the same direction over a relatively short period. This was especially so, as we shall see, with early twentieth century India, and perhaps was a general case with industrialization in labour-surplus societies. These three measures are, of course, related by definition. The relationship can be expressed by the following identities: ∆p=∆(pana)+∆(PiAi)+∆(psns) and ∆(y/p)=∆(ya/pa)+∆(yi/pi)+∆(ys/ps)

Labour-intensive industrialization

105

Where p stands for average productivity, or income-per-worker, n for employment-share, y for income-share, and the subscripts refer to agriculture, industry and services respectively.4 How did India do in terms of each of these measures of sectoral shift? Table 6.1 presents the results. The period of the table is restricted mainly to 1901–31. The year 1901 was the first of a reliable income series. The 1930s results are unreliable because the decade saw unusual movements in the labour market. In a decade of large fall in prices, there were massive wage-cuts in small-scale industry, but money-wage stability in government administration, which made the 1930s atypical. By contrast, the peak years of the Great Depression, 1930 and 1931, saw far less drastic adjustments in the labour market. The 1930s experience is shown in the last column of the table, but will not be discussed beyond these general remarks. Both 1901 and 1931, incidentally, were census years. We can draw four broad conclusions from the table: 1 Employment-share of industry declined. 2 Income-share of industry increased. 3 The contribution of industry to average income-per-worker was larger than the contribution of agriculture. 4 By all three indices, the services sector was ahead of both agriculture and industry. What these data together suggest is a modest beginning of industrialization, measured in terms of income-share and earning-capacity. Industry shed some jobs, but raised its contribution to overall earning capacity. The fall in the number of jobs, therefore, cannot be taken to mean a fall in demand or earning capacity. My interpretation is that industry required fewer people for the same work because of an organizational change. Work shifted from household industry to factories and wage workshops, implying an increase in average hours worked. If we assume that the rise in real income per worker in smallscale industry derived only from a rise in hours-per-worker, then hours-per-worker increased by about 34 per cent between 1900 and 1947. In reality, the extent of the rise was some-what smaller than this, but large and positive nevertheless. An interesting, and probably somewhat misleading, feature of the table is the trend in the services sector. A large part of the services sector incomes were directly tied with private enterprise in industry, including commercial, financial and transportation sectors. In that respect, the trends display industrialization itself. However, the high and steadier average incomes in this sector to some extent reflect the presence of public administration in it. Since administration data are of better quality than those of the private sector, the former seems to dominate the averages in the services dataset. Industry can be further segregated into two parts: the relatively capital-intensive largescale industry and the relatively labour-intensive small-scale industry. Examples of largescale industry were cotton spinning-weaving mills. Examples of small-scale industry included hand-loom textiles, leather manufactures, metal utensils, pottery, food processing, woodwork, carpets and shawls. A large part of labour-intensive industry, thus, consisted of handicrafts. In a loose way, three sets of features marked these two segments as separate: the use of machinery and electric power in the former, whereas the latter was largely tool-based; the relatively larger scale of factories in the former, whereas the latter was predominantly located in small wage-workshops and family enterprise;

Rethinking economic change in India

106

Table 6.1 Measures of industrialization in India, 1901–39 Share in employment (%) 1901

1931

Share in income (%) 1901

1931

Contribution to average productivity growth (%) 1901–31

1931–39

Agriculture

75.0

75.9

45.9

38.0

14.0

−97.6

Industry

10.5

8.9

11.8

14.3

22.0

0.4

Services

14.5

15.2

42.3

47.7

64.0

197.2

100.0

100.0

100.0

100.0

100.0

100.0

Total

Source: Sivasubramonian (2000), based on data contained in Tables 4.2, 4.41, 6.1, 6.4. Note The figure for each year stands for average of three years, including the year preceding and following the one displayed.

and the application of Factory Act on the former, whereas the latter operated mainly in the unregulated labour market. Table 6.2 shows that, relatively speaking, small-scale industry receded both in employment-share and in income-share. The fall in income-share is to be expected even if both sectors expanded employment at the same rate, since large-scale industry had higher income-per-worker or earning capacity. What is significant is that the earning capacity of small-scale industry was growing somewhat faster than that of large-scale industry between 1901 and 1931. In this limited sense, colonial India experienced a process that can be termed ‘labour-intensive’ industrialization. This rise in earning capacity was not a smooth one. Excluding World War I, labour-intensive industry’s share in income was indeed rising throughout. But the war was a big setback. It caused massive shortage of imported inputs and an equally large rise in markets abroad. In 1913, smallscale industry was substantially dependent on imported inputs, whereas their markets were local. By contrast, large-scale industry had a more limited dependence on imported inputs, and a part of their markets was abroad. The impact of the war, therefore, was uneven. And, as the last column in the table shows, so was the impact of money-wage adjustments after the Depression. Expectedly, the segment recruiting from the unorganized labour market saw a bigger fall in labour’s earning capacity after the wage cuts. The results of Tables 6.1 and 6.2 are sensitive to the national income estimates for the period. How reliable are these estimates? The estimates of domestic product in agriculture and large-scale industry are derived from physical production statistics, whereas those in the services and small-scale industry derived by multiplying average incomes with employment. Average incomes are calculated over a non-random collection of

Labour-intensive industrialization

107

Table 6.2 Contributions of large scale and small scale in Indian industrialization, 1901–39 Share in industrial employment (%) 1901

1931

Share in industrial Contribution to average income (%) productivity growth in industry (%) 1901

1931

1901–31

1931–39

Large-scale industry

4.3

9.6

22.2

32.9

44.6

138.4

Small-scale industry

95.7

90.4

77.8

67.1

55.4

-38.4

100.0

100.0

100.0

100.0

100.0

100.0

Industry

Source: See under Table 6.1. Note See under Table 6.1.

incomes in specific occupations. Within a reasonably well-defined occupational cluster, income-variations are usually small. Also, if the number of observations is large enough, the non-random character of the dataset matters less. That being said, the income dataset is not as strong as one would wish it to be. That fact, together with the apparent contradiction between employment-share and income-share of industry, has led to some scepticism about the income data. Careful recalculation of these datasets recently by s. Sivasubramonian has confirmed the trends. Furthermore, recent scholarship on labourintensive industry also suggests that the rise in earning capacity in small-scale industry is not a statistical artifact. Direct measures of physical output exist for the major branches of the textile industry. These measures show rising output, sometimes in quite dramatic extent. By far the largest industry in colonial India, cotton handloom weaving, more than doubled its output in the interwar period. The evidence suggests that small-scale labour-intensive industry in the period considered did not suffer an overall loss of demand for their goods and, in turn, a fall in earning capacity. How do we reconcile this finding with the fall in employment? The answer proposed above was that a change in the conditions of work raised the work capacity of the average worker and, therefore, reduced the required number of workers for the same scale of production. Specialization was one aspect of the transition. In the mid-nineteenth century, many ‘industrial’ workers were not specialized industrial workers, but belonged in rural families that pursued industrial work part-time along with agricultural labour, cultivation and domestic work. A hundred years later, industrial workers were to a much greater extent specialized industrial workers. Today, specialization is the norm and the combination of manufacturing with cultivation is exceptional. When the products of traditional industry became more commercialized in the nineteenth century, two things happened. First, some low-productivity segments in handicraft textiles were destroyed by foreign competition. Second, household industry decayed because of increasing competition within the handicrafts. The rural family as a

Rethinking economic change in India

108

unit of work was disadvantaged compared to the urban factory in a number of ways. The former could not always install new technologies, implement division of labour or economies of scale, and in the putting-out system, often involved higher transaction and agency costs for the merchant. This whole process raises questions about the demand for and the supply of these goods.

Demand The conclusion of the previous section was that labour-intensive industry expanded income-per-worker at a faster rate than the rate of growth in national income. In most cases, such as hand-loom weaving, there was also a significant rise in the scale of total income. Where was the increased production of labour-intensive manufactures being sold? Recent research on labour-intensive industry has shown that competition between machinery and crafts was more exceptional than the rule. Competition was intense in cotton textiles, but outside that sphere and, in the usual case, mills and artisans did not make similar goods at all. And a part of what the artisans made was both highly skilled and met specific wants. Traditional preferences for consumer goods of a certain kind or quality enabled labour-intensive systems to survive, especially where the specific quality derived from manually imparted designs or craftsmanship. Indeed, this story applies even to that large product segment in which competition from machinery was most acute— cotton and silk textiles. In the case of hand looms, the sari allowed for particular types of design that only the hand loom could effect. The hand loom survived partly because that type of sari continued to be in demand. Outside woven cotton textiles, competition between machinery and crafts was rare. Where the design component was of overwhelming significance in the quality of the product, such as carpets, competition from machinery did not count in a serious way. In short, the market for labour-intensive and capital-intensive industries were segmented to a large extent, and some of the specific markets the former served did expand. Who was buying these goods? Although agriculture was generally stagnant in the period covered by this database, some forms of income were growing in the early twentieth century. Real wage in largescale industry increased by 85 per cent between 1900–04 and 1935–39. Real wages of skilled and unskilled workers outside agriculture and manufacturing increased by 50–60 per cent. Within agriculture, food-grains production was stagnant, but non-food-grains, which included major industrial raw materials, expanded. Thus, demand for such mass consumables as handloom cloth was undoubtedly growing among a section of the peasants and workers. Some of the highly skilled crafts in shawls, leather and carpets emerged as successful exports (see Chapter 2). A great deal of labour-intensive industry did suffer obsolescence against mechanized industry products. And quite a range of goods lost their traditional markets because of change in tastes. The cultural context that created a demand for certain forms of craftsmanship and consumption was changing. Over a long period of time, the sari has not only made way for a variety of other forms of garments, but sari designs have also tended to standardize, enabling mass production systems to compete more keenly with

Labour-intensive industrialization

109

hand looms. Given that competition and segmented markets co-existed, what we observe as growth in real income or output is the net effect, which was positive but not very large. The growth in demand for small-scale industry’s output had two opposing effects on demand for labour. The first was a scale effect—a rise in the demand for labour. But the second was a substitution effect—the effect of commercialization on efficiency, which actually reduced the need for the number of persons. In the Indian case, the net effect was marginally negative, so that employment in labour-intensive industry fell even as output and income increased. It is to these efficiency issues that I now turn.

Supply We should perhaps begin by asking the question: in the course of growth in output, how did labour-intensive industry avert diminishing returns to labour? Capital accumulation, or adding capital to the production process, usually expands productive capacity over a long enough time-range. What about ‘labour accumulation’? It is reasonable to expect that it would run out of steam relatively quickly, once production capacity is fully utilized, unless something is done to avert diminishing returns. What steps were needed to make a labour-intensive industrialization sustainable? Diminishing returns were indeed avoided in the early part of Indian industrialization, as is suggested from the second last column in Table 6.2. The average returns to labour were increasing. Case-studies of specific industries confirm that conclusion. In cotton hand looms, for example, physical output per loom or worker more than doubled in the early twentieth century. How was this achieved? Three types of subsidiary processes, I think, were very important. The first relates to complementary inputs. Even if labour was in surplus overall, complementary inputs were not always easily available. There is evidence that an informal capital market was taking shape in the crafts of this period, especially in urban small-scale industries. What is conventionally called the informal money market in India supplied very little capital for small-scale industry. The relatively new formal financial sector consisting of joint-stock banks dealt marginally with them as well. By insisting on English literacy and paper-work, they believed they solved persistent information problems, but in effect greatly curtailed credit supply at the same time. Fixed capital needs were largely being met by mechanisms created within production systems. We observe glimpses of these systems in recent scholarship on the artisan industries, especially those located in major agglomerations. Another example of the same process was growth in local markets in equipment and tools. Both capital and tools commercialized much faster after 1947. The second process can be called labour-augmenting technological change, that is, changes in techniques that did not displace labour, but enabled labour to produce more. The third efficiency-building process was changes in industrial organization. New organizations have been described in recent research as an outcome of commercialization. Segments of the Indian economy were commercialized, and increasingly commercialized, in the eighteenth century. But the nineteenth century added some factors that generalized and extended this tendency, in particular, to peasant production and small-scale industry. These factors included a dramatic fall in external and internal transportation costs, a fall in transactions costs with uniformity imposed in

Rethinking economic change in India

110

currency and weights and measures, and political unity within a large region making inter-regional trade easier than before. In effect, labour, capital and information could become more mobile and more marketized. Within small-scale and labour-intensive industry, there was increasing integration of the market for its products; there was also a shift away from production for own use, local transactions, or use as gifts and tributes, to production for long-distance trade. Competition within these industries increased, and caused the decay of many manufacturing towns or villages that earlier used to serve a local demand, and were swamped by cheaper or better quality products from outside. At the same time, a few agglomerations emerged as concentrations of production, trade, capital and wage-labour. A brief illustration may be in order.

Hand-loom and power-loom weaving At the close of the twentieth century, several hundred thousand hand-loom weavers survived in South Asia, despite the fact that the average labour productivity in the hand loom was not only much smaller than the power-driven loom, but the gap was widening. Historically, the survival of hand-looms arose from the persistence of demand for traditional garments. Since 1947, a second factor was added—a government policy sympathetic to hand looms. Both of these factors have weakened in the long run. Traditional apparel preferences have weakened, if gradually. And protective policy began to be dismantled from the 1990s. The number of hand looms, therefore, fell steadily through much of the twentieth century. Still, in those segments in which preference for hand-loom cloth remained relatively stable, opportunities of earning a reasonable wage and accumulating capital were present. And yet, in the first half of the twentieth century, particularly the period covered by Tables 6.1 and 6.2, demand was stable and competitive forces induced a rise in productivity. Hand looms competed between themselves for markets. Sometimes weavers migrated away from regions that were disadvantaged in this competition, and resettled as capitalists or workers in the more dynamic textile towns. While they did so, the older organizations and ways of working tended to break up. Migration, being usually male-biased, always weakened the family as a work-unit. Structurally the family has certain advantages in artisan production. Hand-loom weaving, for example, often involves a division of labour based on gender and age. The family enabled such a division of labour at little cost. Further, in the absence of formal training and protection of property-rights in skills or designs, the family remains a means to serve as an instructional system and as a means to restrict access to skills. However, these advantages are strong in some types of cloth and not in some others. For example, design-intensive cloths that require a great deal of product differentiation and flexible specialization sustain the household better, whereas standardization weakens these advantages. The household tended to decay because of incompatibility between household production and requirements of efficiency in commercial production, such as a more extended division of labour or utilization of economies of scale. Simultaneously, household production imposed new types of costs as long-distance trade and putting-out arrangement expanded, for example, problems of monitoring quality. Changing opportunity costs of family labour, especially in an urban setting, could also break up the family as a unit of work. At the same time, the new locations emerged as fields for

Labour-intensive industrialization

111

experiment in new organizations such as the factory and in the new technology that required a factory environment to succeed. These two apparently contradictory processes—accumulation in niche markets and the gradual fading out of niches—ensured that capital and skills acquired in hand looms tended to be re-invested outside hand looms. There was constant ‘creative destruction’ inside the handloom industry. In its ashes, there emerged a new technologicalorganizational regime. One of the main fields of investment of money made in hand-loom production or trade was the power loom workshop. The ‘power loom’ was a weaving factory consisting of a small number of power-driven looms. The story of their phenomenal growth in the last quarter of the twentieth century suggests several lessons about the changing balance of advantages between large-scale and small-scale industry in India. Descriptions of hand-loom weaving in western and southern India illustrate best the process of commercialization, capital accumulation and organizational change. The changes were the most visible, even dramatic, in the case of medium-sized textile towns. The industry in western India evolved into an almost entirely urban and wage-labourbased industry. Two of the four largest hand-loom towns in 1940, each with 20,000 or more looms, were located in these two states. These included Sholapur, where the industry was mainly in factories each containing from ten to forty looms. The workers and owners were generally migrant weavers from arid regions east of the town. They migrated in increasing numbers as the town emerged as a major market for cloth, yarn and cotton in the interwar period. Other examples, different only in points of detail, were Malegaon, Surat, Burhanpur, Belgaum and Bhiwandi. All of them accumulated capital by making and selling hand-loom cloth in the interwar period. And all of them invested that capital in power looms after 1950 (Haynes, 2001). Today, western India is a marginal region in hand looms, but contains perhaps 800,000 power looms. In western India, the underlying change in labour market gathered speed after 1947. A somewhat contrasting profile comes from Tamil Nadu in South India. Among major Indian textile-producing regions, Tamil Nadu had the highest density of hand looms, approximately ten looms per thousand persons in 1950 and again in 1982. The average for India in 1982 was four. Here too, weavers and merchants steadily switched to power looms, but such shifts did not make the hand-loom industry marginal to the extent these did in western India. The South Indian scenario has seen greater continuity and coexistence between traditional industry and the intermediate regime, possibly because of slower changes in consumption. The hand-loom specialty, bordered sari, has had a longer and richer life in this region than nearly anywhere else in India. And yet, the strong handloom heritage ensured a steady supply of well-trained capital and labour into the powerloom industry. Today, Tamil Nadu also has 400,000–500,000 power looms. Its output consists of cotton sheets, grey or checked, that are taken up in large quantities by the apparel industry. South India also illustrates a more gradual decline of the household than was the case in the west. The advantages of the conventional division of labour between weaving and processing were still strong in the bordered hand-loom sari. These advantages were smaller in the kind of cloth the power loom took up, the nature of weaving being simpler, and the separation of weaving and processing more decisive. At the end of colonial rule, Sholapur, Salem and Madurai were among the five largest weaving towns in India. Sholapur had 10,000 hand looms in work, of which a fraction

Rethinking economic change in India

112

survived in 1997. Salem in the early 1950s had more than 20,000 of which 8,000–10,000 were active at the end of the century. Madurai had another 10,000, the number in fact grew fivefold in the next forty years. In the mid-nineteenth century, neither Sholapur nor Salem was a major textile town. Salem did have an ‘investment’ of the East India Company, but a rather small one, and one which did not survive beyond the 1820s. Their growth into major textile towns arose from nineteenth-century developments like the railways, the cotton export trade and spinning mills. Towards the end of the century, and especially after the railways connected cloth, yarn and cotton trade, Sholapur and Salem developed into entrepôt in hand-loom cloth. Madurai was a trading point too, but dealt mainly in its own output, and had a large local market among both residents and pilgrims into this temple town. As trade grew, capitalists emerged too. Many of them rose from the ranks of the artisans. They included traders in cotton, yarn, dyes, cloth and gold-thread. These materials began to be imported or exported from the colonial period. Pioneers in these trades later became not only rich, but often also leaders of the local weaving communities. At the end of the twentieth century, a random sample of leading industrial families and social leaders in these towns would be dominated by families that were artisanal two generations previously, with ancestral business in long-distance trade in cloth, dyes, yarn or gold-thread, and with a history of rapid accumulation in the interwar period or shortly thereafter. There are few exceptions to this profile of success in the major hand-loom or power-loom towns. As the towns grew in scale of production, capital was invested in factories. Sholapur illustrates this process best and Salem to a smaller extent. In Madurai, where the handloom tradition remained strong, factories were non-existent in weaving, but did emerge in dyeing of yarn, which was a major source of the reputation of Madurai’s hand-loom industry. In each town there was a general tendency towards increased division of labour. One universal direction of change was separation of weaving, processing and dyeing. At the same time, there were sustained improvements in quality and speed of each process. Originally, the three tasks tended to coalesce either in the family firm, or under various forms of communal pooling of labour. Over time, separate specialized groups began to perform these tasks. And within each segment there was a drive to make experiments with newer tools and processes. The major examples of technological change included the fly-shuttle, ‘semiautomatic’ looms, new hand-driven machines for warping and sizing, and synthetic dyes. The favourable conditions for large-scale adoption of these tools started with a ‘commercialization’ of the major products of hand-loom weaving, which created a space for capitalists to operate, and generally benefited capital more than the wage-earning classes. At the same time, conditions adverse for technological change weakened. The family firm located in the rural areas had limited means to employ new tools or labour, limited means to experiment with new tools and supplied markets dependent on seasonally variable rural demand. The diffusion of new tools needed different organizations, that is, needed new firms, collectives, different markets and products. Conditions for new types of units that employed hired labour and could utilize new economies of scale arose in the cluster of towns in western India, via urbanization of a variety of textile-related services, and migration of labour and capital (see Haynes and Roy (1999) on migration; and Roy (2002b) on technical change).

Labour-intensive industrialization

113

From the semi-automatic handloom that was usually installed in a larger workshop, the idea of a power-driven loom was a small step. The traditional hand-loom set up in a pit in the hand-loom weaver’s living room was a far cry from the power-driven loom. But the automatic hand-loom mounted on a frame inside the factory and the power-driven loom looked similar and worked on the same principle, but for the source of energy. In the interwar period, many power-driven looms were discarded as scrap by the mills. Buying such a loom and reconditioning it to fit the weaver’s factory shed was not expensive. Relatively well-off weavers started to replace hand looms by power-driven looms in products where such a switch was possible. The first such looms in India appeared in hand-loom towns about 1900 and were run with fuel oil. At the same time, and probably unaware of this small beginning, A.C.Chatterton, the Director of Industries, Madras, predicted that ‘small power loom factories might be worked with great success in this country’.5 From the 1930s, the power-looms spread much faster when many interior towns with hand loom industries received electricity. The ground was prepared for what was to become India’s largest industry at the end of the century, and the most fertile ‘training ground for the development of indigenous manufacturing genius’, to quote the prescient Chatterton again. From 15,000 looms in 1940, the power looms grew to engage at least 1.5 million in 1995, and possibly more in 1997. If these magnitudes of the growth of the industry are staggering, so is the lack of attention it has received from historians and economists.6 From the mid-1980s, India’s rather insular trade regime began to be liberalized. Since then, the most important export commodity to benefit from the opportunity to export and access technologies has been textiles and clothing. In turn, the boom in apparel export saw explosive growth in the power-loom industry, which supplied nearly all the demand for fabrics from apparel producers. Between 1985 and 1997, looms expanded by an estimated 700,000–800,000, which might amount to an additional employment of about two million persons. In 1991, the power looms employed about four million persons. The figure stood for 20 per cent of all wage-labour in industry, and about a third of wagelabour in informal sector industry.7 Innocuous in itself, the power loom represented a new system of mass production of cotton textiles in India. The cotton mills of Bombay and Ahmedabad were vertically integrated systems. They integrated spinning, weaving and processing in one place. The power looms represented a system of production where each of these stages of textile production was separate from each other, even though usually located in the same agglomeration. The power loom represented, in other words, a production system based on market transactions, as opposed to vertical integration. Midway between the hand loom and the integrated cotton mill, the power loom proved far more flexible than both. The later growth of power looms and decline of the mills, though partly policy-driven, can also be seen as an outcome of an obsolescence of the vertical integration model that worked well in the cotton textile industry of the late nineteenth and early twentiethcentury India. It can be seen, in the language of industrial organization, as a transition from hierarchy to market. Why did such a change occur? Vertical integration and large scale made sense in the mid-nineteenth century when the first cotton mills were started, in the context of absence of well-developed markets in labour, knowledge, machinery and capital. By 1880, a weaving factory required, for its success, a captive spinning mill and processing workshop, recruitment systems that

Rethinking economic change in India

114

involved large numbers, access to foreign machinery and technicians, and the involvement of merchant firms who alone could raise money. All of this entailed economies of scale as well as vertical integration. Therefore, the large-scale integrated mill was advantageous in the nineteenth century. However, 100 years later, the missing markets problem had disappeared. Migrant labour was available easily even in small towns. Capital was available from informal and local sources such as rich artisans. Machinery could be purchased and repaired easily. In other words, setting up a weaving factory no longer entailed setting up spinning and processing shops as well. The power loom represented the new market-resource regime in India that naturally strengthened the advantages of small-scale labour-intensive industry.

Conclusion I described above is a process in which industrialization occurred via more efficient use of labour, rather than replacement of labour with physical capital. Demand for such goods was specific and growing. While labour-intensive and capital-intensive sectors did compete in some spheres, competition was not the norm. In recent research, the segmented nature of their markets has been stressed rather more. In the aggregate, there was a rise in demand for the goods of labour-intensive industry. Without attempts to raise the quality of labour and quantity of complementary inputs, pure labour-intensity was always at risk of running out of steam. In early twentieth-century India, we do see several types of changes that might lead to efficiency—growth of informal input markets, labouraugmenting technological change and increasing employment of wage-labour. The story has lessons for interpretation of Indian economic history. Its key elements seem capable of explaining the pattern of industrial growth after 1947 to a large extent. Despite efforts to build a capital-intensive sector, a substantial part of Indian industry continued to be labour-intensive. Despite low, even stagnant real wages, such industries did generate growth in income-per-worker. The historical account of how firms and workers responded to the risk of diminishing returns to labour illustrates some ways that efficiency gains might have been ensured. A second set of lessons concerns histories of labour and capital particularly. Labour history in India has tended to focus on the formal sector, mainly the mills. The main thematic here is a transplantation of workers from traditional to modern occupations, turning peasants-labourers-artisans into an urban working class. The question often asked is, how closely this working class approached ‘class’ as a concept. The paradigm of transplantation, however, is not an appropriate one for labour-intensive industry, where industrialization did not necessarily involve transfer of population from traditional to modern sectors, but a gradual change in the traditional unit itself. A labour history in this sphere needs to be sensitive to enduring links between the past and the present. The gradual transformation of the family as a work-unit, for example, should be a core theme in such a narrative. Similarly, business history in India as a rule dealt with mercantile-

Labour-intensive industrialization

115

financial communities that financed the mills. Labour-intensive industrialization reminds us that a great deal of capital for industrialization also came from such people as rich artisans, and such institutions as the informal capital market. One important effect of the decline of the family as a work unit was a steady withdrawal of women from manufacturing work, to which I turn in the next chapter.

7 Women and industrialization Through much of the twentieth century, work-participation of women, which on average declined a little, increased in agriculture in relative terms and fell in manufacturing. One possible factor behind this change was the increasing importance of casual labour in agriculture. Women entered casual labour markets more easily than they did long-term contracts. Another factor that played a role was the partial withdrawal of women from the industrial workforce. The labour-intensive industrialization in the artisan sector described in the previous chapter involved, on the one side, increasing use of wage-labour and, on the other side, a steady and pronounced de-feminization of the industrial workforce. For a full account of changes in women’s work, we need to answer the question why women were prone to leaving industry. The present chapter discusses the difficulties that women workers faced in continuing work when work itself shifted from the home to the factory. Historical scholarship focuses on both barriers to entry into the factory as well as barriers to exit from home to join work-sites far away from home. This chapter reviews the literature and focuses on a particular barrier to exit related to the age at marriage. The barrier in question concerns the quality of training. This chapter begins with a tour of the basic stylized facts in a comparative historical setting, followed by a discussion on the Indian dataset and, finally, a review of models of transition in women’s work.

Comparative history: the U-and the M-graphs That men and women face different considerations when thinking about a move into the labour market is well-known. The difference arises because of the need, in the case of women, to accommodate wage-work into the domestic routine, in which the care of small children is the most crucial task. Before modern industrialization began, industrial work was performed largely by households, the family itself was the unit of work, and the balancing of domestic with commercial work did not pose a problem. But, it did become a problem as manufacturing work began to shift, to borrow an expression from Boserup (1970), ‘from the hut to the factory’. To continue working, women needed to leave home and children for long stretches of time during the day. The difficulty that this situation posed shows up in occupational statistics in the form of two quite universal stylized facts. First, women’s participation in urban industrial labour markets tended to at first fall and then rise in the course of industrialization. From the nineteenth century, the mill existed side by side with the family in the same industry. In some cases, mills out-competed the household. Women’s participation in industry in this relatively new scenario of competing industrial organizations required institutions enabling an easier balance between paid work and domestic work. If these factors did not

Women and industrialization

117

exist, or were slow to emerge, for a while, participation rates in industry could fall. But even though initially combining domestic and wage-work was difficult and there was some exit from the workforce, in the long run these problems weakened because of commercialization of domestic work and technological change that saved time in domestic work, so that the participation rate began to rise again. In the course of industrialization, therefore, women’s presence in the urban industrial labour market is expected to follow a U-shape. The second stylized fact is that women’s participation rate in relation to age group shows that the rate peaks for two groups and is low otherwise. The first peak occurs for a group in which adult but unmarried girls form a majority, usually 15–24 years, the rate then falls in a group in which the majority are married with small children, and the rate rises again when many among the married women return to the workforce. If the overtime rate follows a U-shape, the across-age rate is expected to follow an M-shape. How universal are these patterns? Generally speaking, the evidence on the U-shape is mixed, but partly for the reason that economic demographic data for the nineteenth century remain of uneven quality.1 Britain in the late nineteenth century probably saw a decline in women’s participation in industry. The reasons remain controversial. In one view, the withdrawal of women was a voluntary response to an increase in average family earnings, giving rise to the ‘male breadwinner’ household and economic culture. The accent falls on supply of labour. In another view, the withdrawal was an outcome of competition at the workplace. The accent falls on demand for labour (Rose, 1992; Seccomb, 1986; Humphries, 1995). Boserup (1970) observed that in developing countries industrialization generally involved a decline in women’s participation in industry. She observed that, at a point in time (in the 1950s), the percentage of women in the manufacturing workforce was significantly lower than the overall participation rate of women (the only exception being Hong Kong). India was part of this set. However, in cases of relatively rapid industrialization from Asia, the U-shape was either absent or traversed quickly. In late nineteenth-century Japan, the proportion of women in the manufacturing workforce was steady in the interwar period, at 30–35 per cent, even as work shifted from rural farming families to urban factories (Taeuber, 1958:87). Among the Asian ‘tigers’ in the 1960s and 1970s, industrialization did not see a fall in women’s participation rate, whether overall or in industry. If there was a U-shape, the transition from its one arm to the other was effected pretty quickly.2 The percentage of women in the manufacturing workforce rose more or less continuously in the post-war period. In developing Southeast Asia generally the percentage was rising in the last quarter. Historical data from Taiwan suggest the presence of a very slight decline in the proportion in the early twentieth century, but a significant reversal in the post-war period. By contrast, South Asian data do show, not only a U-shape, but also that the transition from one arm of the ‘U’ to the other was exceptionally slow. For an unusually long period of time spanning much of the twentieth century, this region crawled along the left arm of the ‘U’. Further, the Indian percentage c. 1970 (twelve) was one of the smallest in the world. The norm in post-war developing Asia and Latin America was well above 30. The M-graph appears in many industrial countries, and is found to be correlated with demographic transition, namely, decline in total fertility rate such that married women could enter the labour force quicker (Durand, 1975). Post-war data for East Asia

Labour-intensive industrialization

118

generally support the M-shaped participation rate. However, in South Asia, the ‘M’ was largely absent (Horton, 1996; Visaria, 1997). Until recently, age-wise participation rate in Indian data followed a smooth inverted U-shape, with neither a significant decline nor a significant upsurge in the 25–34 age-group. Marriage indicated a higher probability of exit than of return to the workforce in the Indian data. Within the Indian academia, opinion has been divided on the reliability of the Ushaped graph. Economic historians have tended to give more credence to male workforce data than to female workforce data, implying that the fall is not a wholly reliable trend.3 In this respect they followed J.A.Baines, the 1891 Census Commissioner of India, who was so dissatisfied with the 1881 data on women that he declined to report any sex-wise workers’ data in 1891. The early censuses probably overestimated as ‘workers’ many women who contributed to commercial work rather marginally. They were, it was suspected, routinely included in the workforce and classified under their husbands’ occupation. Later censuses presumably corrected that error. Interestingly, in the 1990s, another body of experts argued that the later censuses underestimated women workers.4 The hypothesis that the fall was primarily a reporting error is not very satisfactory, however. That it took the census a century to discover and correct an ‘error’ is not an attractive hypothesis. Moreover, demographers, who have been prone to treating the ‘U’ as a global tendency, have taken the fall to be authentic, and representative of a structural change. Generally, demographers reject the conceptual possibility, embraced from time to time by Indian economists, that the fall was a matter of voluntary choice. D.R.Gadgil, for example, had stressed what I call ‘barriers to exit’ from the home environment in search of paid work outside. Demographers stressed rather more fall in demand for women-intensive jobs, a perspective I call below ‘de-industrialization’.5 What factors account for the international disparities? Why did some societies find a way to move quickly over the ‘U’, if they did face a ‘U’ at all, whereas South Asia did not? What accounts for the relative scarcity of married women in the workforce in the region? Why were the levels of participation rates so low in the region? To answer these questions, we need to consider some models of women’s participation in manufacturing. But before that, let us examine the Indian data somewhat more closely.

The Indian data The principal difficulty in classification of women workers was that women often took part both in the occupation of the adult men of their families and in the wage labour unconnected with what the men did. Where the former work occupied them to a greater extent, the difficulty arose as to whether they should be treated as primarily dependents or primarily workers. While there was diversification at this level, in wage work, there was a great deal of gender segmentation. Some work only women did, some work only men. And women entered mainly those jobs that hired only women. There was hardly a hint of competition in historical sources. To the contrary, there were many references to a great deal of compatibility between men and women in respect of their mode of entry into the same work-site. In the cotton mills, women worked usually in the mills where their husbands worked. In the mines, ‘they of course work with or near the male members’

Women and industrialization

119

(India, 1921a: 170–1). We can infer from these reports one constant feature of women wage workers, they were all married. Hand-spinning of cotton, basket weaving, grain processing by hand and stone cutting were some of the traditional manual processes in which women far outnumbered men. Of the more modern sectors, metal processing, chemicals or machinery manufacture and printing presses rarely hired women. Mines hired many women, not in extraction but as headload carriers. Cotton mills hired women as reelers and winders in the spinning department. In cotton gins, women fed cotton into the gin. As that task required a large number of workers, nearly half the workers in a gin were often women. In hand-loom factories, women performed yarn processing tasks again as they did at home. In bidirolling, women alone were employed. There was also near-monopoly of women in certain activities allied to the processing or use of minerals. Quarries of hard rock, for example, employed women as small stone breakers. Pottery, brick and tile factories, and cement factories hired women too. If there were changes and shifts within these spheres, we cannot capture them for want of sufficient research. However, we do know that in the spinning department of the cotton and jute mills women tended to be replaced by men after the 1922 Factories Act. At the aggregate level, Indian data shows two patterns quite clearly. First, there was indeed hint of a long ‘U’ in the percentage of women in industry. And second, as far as we can measure it, the trend was correlated with a shift of industrial work from the family to the factory. Table 7.1 shows the percentage of women in the workforce, in industry, in agriculture, and the work-participation of women over the period for which a reasonable comparison can be made. The table excludes 1951 data, which were uniformly ‘outliers’ and probably underestimated industrial workforce. In the period covered by the table, women’s presence in industry declined and their presence in agriculture increased. The work-participation rate fell slightly in this period, for both men and women, but did not lead to an absolute fall in the workforce. The break-up of women workers into family and non-family is available only from 1961. As Table 7.2 shows, there was an apparent association between women working and household industry. For men we cannot spot a similar bias. The 1960s saw the sharpest fall in women’s presence in manufacturing. Between 1961 and 1971, as household-workers ratio fell, the women-worker ratio fell too. Change in the former accounted for as much as 66 per cent of the fall in the latter. Although we cannot apply this test before 1961, it seems plausible that industrial organization and the shift

Table 7.1 Women workers, 1901–91 workers as % of work force

Total workforce

Workparticipation rates (%)

Agriculture

Industry

Men

Women

Percentage of female workers Agriculture

Industry

1901

31.0

31.1

34.3

63.8

29.8

62.6

17.3

1911

31.8

32.5

34.3

62.9

30.8

68.8

12.7

1921

31.3

32.1

32.1

61.3

29.6

70.3

11.0

Labour-intensive industrialization

120

1931

22.2

21.2

24.1

53.6

16.3

62.6

11.7

1961

31.5

36.1

25.8

57.1

27.9

79.6

9.6

1971

17.4

20.8

12.4

52.5

11.9

80.1

7.7

1981

20.2

24.1

14.1

51.6

14.0

79.4

9.1

1991

22.5

27.3

15.0

50.8

15.8

78.8

8.0

2001

31.6

38.9

71.9

Notes Source of all figures is the Indian census, occupational statistics. The 1951 census workforce data are not reliable, and not reported in the table. In all cases, figures for 1901–31 refer to British India and States, and figures for 1961–91 to the Indian Union.

Table 7.2 Gender and industrial organization, 1961 (workers in millions) Household

Non-household

Total

Men

5.9

7.1

13.0

Women

4.1

0.8

4.9

10.0

7.9

17.9

Total

from family to wage-labour played a singular role in influencing women-worker ratios in manufacturing in the long run. Women’s presence varied across region and industry (Table 7.3). However, the association between family and women workers remained quite stable across industry and region. In both cases, the presence of household industry explains a significant part of the variation, as shown: 1 Industry: 54 per cent (r2 of regression of women-worker ratio on household-worker ratio). 2 Region: 39 per cent (r2 of regression of women-worker ratio on household-worker ratio). Interestingly, the relationship was relatively weak across space, suggesting some cultural influence on the women-worker ratio. We see this clearly in the case of much of northern India (UP, Rajasthan, MP and Punjab), where household industry accounted for more than 55 per cent of the industrial workforce, and yet the women-worker ratio was smaller than the average. Similarly, industrial composition too plays a somewhat independent role, suggesting a technological influence on the women-worker ratio. Women’s presence was small (less than 10 per cent) in metals, machinery, paper and printing; moderate in textile products, wood and minerals (10–30); and high in basic textiles (above 30).

Women and industrialization

121

Table 7.3 Women-worker ratio (%) in manufacturing, 1961 By industry

By region

Manufacturing

27

Southern India

28–43

Food-tobacco

38–42

Western India

14–18

Cotton and silk textiles

41

Northern and Central India:

Jute textiles

24

Bihar

34

Textile products

21

Jammu and Kashmir

35

Wood

27

Madhya Pradesh

29

Minerals

27

Uttar Pradesh and Punjab

14–18

Leather

12

Rajasthan

27

Chemicals, machinery

0

Eastern India:

Metals

9

Assam

69

Orissa

44

West Bengal

14

In addition to these aggregated profiles of women’s participation, we can reconstruct some caste-wise participation rates. These data permit us to talk about shifts out of ‘hereditary’ occupations. Generally, women’s work-participation tended to be higher for those castes who were already engaged in agricultural labour. The 1961 census observed that the women-worker ratio among the ‘scheduled castes’ was considerably above the average for the general population. Also, among castes already engaged in labour, the tendency to diversify out of labour to some form of production was generally weak, and particularly weak with women. A representative exercise can be done for the caste group Paraiyans of Madras, whose traditional occupation was agricultural labour. Between 1901 and 1961, the overall work-participation rate increased significantly (even if we take, as there seems to be some reason to do so, the 1901 ratio to be an underestimate). Between 1901 and 1961, the Paryaians’s occupational profile shifted away from agricultural labour. But the shift was much weaker for women who remained predominantly oriented to agricultural labour. Thus, the Indian census data suggest that the explanation of women’s participation in manufacturing workforce hinges on conceptualizing the relationship between gender and industrial organization.

Labour-intensive industrialization

122

Five models of transition Overall participation rate: the household model In several cases from comparative history, the percentage of women in manufacturing work and the female labour-force participation rates seem to be correlated. Both are relatively low in India, both high in East Asia. One hypothesis could be that the overall participation rate is the independent variable, and drives the participation rate in industry. In the Chicago paradigm of allocation of time developed through the work of Gary Becker, Jacob Mincer and others, women’s work-participation depends on a balancing of time and money.6 The key insight of the paradigm is that a variety of domestic goods have economic value. Domestic goods include non-marketed services that the family consumes, such as child-care, and are supplied largely by women. The allocation of time at the disposal of the household to the production of domestic goods or to wage-work depends on preferences and relative prices. Market wage is a cost of producing domestic goods, and a means with which marketed goods can be purchased. Now, as wage increases, an income effect induces the family to consume more of both domestic and market goods, whereas a substitution effect induces conversion of family-time into wagetime. To transform this model into an explanation for the U-shaped graph, we need the assumption that, as industrialization progressed, men’s wages at first rose faster and rise in women’s wages followed. The former would, via both effects, lead to a net conversion of women’s time towards the produc-tion of domestic goods. However, a rise in specifically women’s wages would strengthen the substitution effect and lead to a move into the market. Now, this assumption is not satisfactory. If we have to assume that men’s wage rose first and women’s wages followed, that amounts to assuming that men were preferred during early industrialization. But is that not what we set out to explain in the first place? Further, the implication that withdrawal and return were driven by rises in wages can be questioned, for there is no evidence of a significant rise in wages. In industries that were particularly intensive in women workers saw a decline in demand. De-industrialization The ‘de-industrialization’ literature in India noted a symmetry in employment shift— decline in the share of industry and rise in the share of agriculture in the workforce. This was attributed to a decline in the demand for traditional industry. It was suggested that decline in demand for traditional industry pushed unemployed workers out of industry and into agriculture. The literature did not distinguish men and women, and thus failed to notice that the symmetric employment shift characterized women workers rather more strongly than men. The literature also assumed that the two sides of the symmetric relationship were causally related, whereas they need not actually be so. Later, demographers offered a modified and more acceptable version of the thesis. It was suggested that some work which was done by women earlier declined for lack of demand, as this work became semi-mechanized. Of the more commercially oriented

Women and industrialization

123

industries in India that employed a large number of women, cotton hand-spinning was an example. A range of occupations allied to the rural household’s own consumption of goods and services, chiefly grain processing, was routinely done by women, and progressively declined. There was a reason behind women-intensive work disappearing earlier and faster. These works involved relatively little training, skill and capital, and thus, (a) tended to be pushed to that segment of rural labour which commanded little market value even in the busy seasons, and (b) were especially vulnerable to mechanization. There are two problems with this demand-based thesis, however. The first problem is conceptual. It assumes the existence of a certain kind of gender-specific segregation of tasks, wherein women crowded those tasks that were more vulnerable to mechanization. Why, one might ask, did such segregation exist in the first place? What constrained opportunities of occupational diversification within industrial work for women, but not for men? The second problem with generalizing from these examples is an inconsistency with data. If falling demand was such a powerful influence, we would expect a positive correlation to emerge between women’s participation and income-growth in traditional industry overall. For, if exit of women signifies disappearance of whole industries, there should be a fall in income. In fact, the correlation is a negative one. Women’s participation fell even as income growth was positive in the handicrafts in the early twentieth century. This is not a conclusive critique, but it should alert us to the prospect that within traditional industry, segments that saw rise in demand outweighed segments that saw fall in demand. Why could labour not shift from one segment to another? What we really need to explain is why unemployed women spinners had to leave industrial work altogether, and not seek employment in spinning mills. To some extent, women did move from home to factory when the industry itself relocated—grain-milling is a partial example.7 But the general rule was attrition, as opposed to relocation, of women’s work when industries mechanized and/or shifted from the home to the factory. This scenario leads us to the third model of transition—barriers to entry into the factory. Barriers to entry We have seen an aspect of supply of labour, demand for labour, and now we consider competition at the workplace. I call the cluster of explanations that end up arguing some form of deliberate substitution of women for men by their employers ‘barriers to entry’. Within this cluster, however, the point of emphasis can vary a great deal. The only recent work on the subject of women workers in the factories of colonial India has focused squarely on barriers to entry, but of a political kind: The colonial State…helped to achieve and foster a particular accommodation between capitalism and patriarchal forces which allowed the retention of women’s labour in the non-capitalist and household sector or at the lower end of the wage labour market. (Sen, 1999a: 239)

Labour-intensive industrialization

124

Reference to the ‘colonial state’ raises two problems. The state could play a role, if the British Indian state took sides in the gender race, only in the large regulated mills, which accounted for not more than 5 per cent of manufacturing workers in the early twentieth century. The real site of the decline of women’s presence was not here, but in the unregulated small-scale manufacturing. Second, by setting sight on colonial state and patriarchy, we risk losing sight of another variable that capitalists ought to be much more concerned about. Was it not profitable for employers to replace women by men? If yes, why do we need patriarchy? If not, why did capitalists succumb to irrational sentiments? We need to look more closely at barriers to entry that were less political and more economic. Boserup (1970) judged correctly that the exit of women from wage-work during early industrialization was connected with a shift from ‘the hut to the factory’. As industrial work shifted from households to factories, women tended to leave the workforce. Boserup implicitly equated ‘industrialization’ with capital-intensive industry, an equation usual in the 1960s development scholarship. Her question was why the largest of the factories did not want women, or replaced women by men at the first chance they got. The explanation focused on various pieces of government regulation, which gave women workers the institutional means to handle their domestic duties better, but were seen by employers as reducing their flexibility. Maternity provisions are well-known examples. Another relevant piece of legislation was the prohibition of night-shift work for women, which became effective in India in the 1890s. This particular clause of the Factory Act, and its possible implications for women’s participation in the textile mills of Bombay, were debated extensively in India in the early twentieth century. Later research on the cotton and jute mills exposed other aspects of perceived flexibility that often led to the substitution of women for migrant male workers (Chandavarkar, 1994; Sen, 1999b). Despite its long antecedent, the adverse legislation argument is a limited one for two reasons. First, the example of Japan in the nineteenth century shows that these barriers could be overcome. The second objection to barriers to entry is that government regulation did not touch more than a small fraction of the workforce in the developing countries. The Japanese mills drove out the cotton mills of Bombay from the China market within a few years in the 1890s. They had an advantage, as the Bombay mill-owners noted, with a sense of despair because they could not hope to match that particular advantage. This was low-wage female labour. Young unmarried women from farming families moved to cotton-spinning mills in Japan. Cotton mills elsewhere in the world employed farm girls too.8 Their wages were very low, though in order to employ them the mills needed to create special systems such as all-women dormitories. They were employed only for a few years, and the pay covered little more than subsistence. And yet, through this transfer of population, ‘the industries in Japan secured cheap labor, the farmers secured some cash income, the girls secured dowries—and the contracts were completed by the age then regarded as appropriate for marriage’ (Taeuber, 1958:115–16). While the Japanese employer created conditions enabling women’s entry into the factory, the Indian mill-owner was stubbornly indifferent to the most elementary enabling conditions. Most women workers were married with small children. Until the late-1920s most mills did not have child-care facilities, and yet had begun to prohibit entry of small children into the space where their mothers worked. The women were left at the mercy of

Women and industrialization

125

the sardars (the contractors-cum-foremen) for breaks from the shop-floor, or dependent on older children, and were compelled to use what tree-shade they could find outside the mill to nurse small children.9 There were many other, less glaring, cases of neglect and indifference. Not surprisingly, the presence of women in the mills began to decline quickly after the 1922 Factories Act. No less powerful barrier than employer indifference were the social and cultural attitudes of the workers themselves. A group of women complained before the Royal Commission on Labour that ‘the chastity of the female employee working in the factory is always in danger’ (India, 1928–31:65). On the point of patriarchal prejudices, Indian trade unions were partners and collaborators of the millowners. One mill union, for example, stoutly resisted ‘the scandalous system of men and women working jointly on the same machine’, which an innovative South Indian mill tried to implement in the 1920s. Many other unions shared the sentiment.10 This difference between Japan and India had probably more to do with economics than prejudice. The critical question is: why did the Indian employer not want to spend more on welfare or accept the cost of special legislation, considering that women usually received smaller wages than men for the same work? Underlying the difference between them in the status of women inside the factory, there was a calculation of how valuable or indispensable the women workers were. Both societies were patriarchal and neither allowed women freedom to work or live as they wished. The crucial difference, I believe, was in the fact that the women workers in the Japanese mills were unmarried. Being unmarried and young, they were more suited to training than married women might have been. The Indian women workers were married and thus seen as not reliable material to be trained and put to the skilled tasks. Women were part of the core workforce in Japan, whereas the Indian women workers were usually a marginal part of the factory. The Japanese women worked the ring spindle while the Indian ones were set to unskilled tasks. The Japanese women were part of the steady workforce, at least for a few years until they were married, the Indian women were casual workers. If culture does enter this story of marginalization of women, it enters via marriage norms, and not via employers’ biases. The second objection to ‘barriers to entry’ is that government regulation did not touch more than a small fraction of the workforce in the developing countries. Using South Asia as an illustration, between 1911 and 1961, more than one-and-a-half million women notionally exited manufacturing in this region.11 In the mills covered by the Factory Act, Disputes Act and other regulations, the extent of the notional decline was only a few thousands. The real sites of the decline were small factories and workshops that employed wage-labour, showed a strong preference for male labour, but were not subjected to any regulation worth the name. The gender-bias of mill-owners or the legal safeguards do not really matter, in a quantitative sense, to explaining why women’s presence in manufacturing fell in India. To explain why we find few women working in these units, we need to consider the hypothesis that perhaps women could not seek these jobs. We consider supply of labour again, in the form of what I call barriers to exit from the family.

Labour-intensive industrialization

126

Barriers to exit: marriage Numerous reports point to a deep-rooted and universal feeling of uneasiness and unwillingness on the part of the Indian woman to work in a factory alongside men. How do we account for such a feeling? Married women faced a singular barrier to exit the household: children, especially young children and infants. Even in the presence of adequate institutional safeguards at the factory site, it was not easy to go round this barrier, for the traditional society would not easily accept these institutions to be adequate substitutes for maternal care. A consideration such as this would tend to bias married women’s work towards household industry, part-time or casual work rather than permanent contractual work, and bias it towards agriculture and services, rather than manufacturing. For, manufacturing by contrast with farming, would require steadier commitment and possibly staying away from children for longer periods during the day. There were situations when work and home could meet without major lifestyle adjustment. Sometimes, home could be taken to the work-site, and sometimes work could be brought home. As we have seen in Chapter 3, rural labourers quite often took children to the field. The next chapter describes a hand-loom factory in South India where women took children inside. Universally in the cotton and jute mills in the nineteenth century, especially the jute mills, women took infants into the factory. Inside such a jute mill near Calcutta, infants can be found lying on sacking, in bobbin boxes and other unsuitable places, exposed to the noise and danger of moving machinery and a dustladen atmosphere, and no year passes without a certain number of serious and minor accidents, and sometimes even deaths, among such children. (India, 1928–31:65) On the other hand, household industry was a situation where work was performed at the very site where the children lived. But these were more the exceptions than the norm in industry. There were many situations in which the connection between work and home was severed. As the above citation reveals, taking home to the work-site could be traumatic, whereas bringing work to the home as in household industry was simply uneconomical. Generally speaking rural women found it difficult to relocate to urban areas when work urbanized. The men simply did not earn enough to take their family along. Migration, in other words, always had a strong gender bias in India, a point Gadgil (1965) stressed in his study on women workers in India in the early twentieth century. Over a sufficiently long run, marriage does tend to lose significance. Why does such a powerful barrier as marriage and the care of infants eventually begin to weaken? Technological and institutional changes were mentioned before. Another factor is that the fertility rate begins to decline as rise in wages outside increases the opportunity cost of producing children. Another possibility is that as the fertility rate declines independently, women are released from domestic work to a greater extent and can return to the labour market. Between the two hypotheses, the direction of causality is reversed. And neither

Women and industrialization

127

view answers the question, if there were barriers to entry earlier, how did those barriers disappear?. Still, in a traditional society, marriage was undoubtedly a watershed, in that it was perhaps easier for unmarried women to take part in an unfolding industrialization process, provided of course the average age at marriage was not too low. Barriers to exit: average age at marriage Even in a traditional society where marriages at a certain age were inevitable, unmarried women had some flexibility. In Japan, they could negotiate with parents a small postponement of age at marriage and the consent to work in a factory. Through a small window formed by the difference between the age of attaining adulthood and the age at marriage, women stepped out of tradition into the new world of an unfolding industrialization, even though ‘once marriage occurred, the major regulates of conduct were the rules of the traditional society rather than the forces of industrial society’ (Taeuber, 1958:214). This observation reminds us of the momentous significance of marriage as a marker between two mentalities in a transitional society. For unmarried women, young children were not the barrier to exit. But age at marriage was. The lower the age, the more likely it was that unmarried girls would not be considered old enough for migration, travel or simply leaving school. The lower the age at marriage, the earlier the post-marriage barriers would come into force. They would, then, work from home, if they worked at all. The later the age of marriage, the greater should be the proportion of unmarried women in the workforce and the proportion of workers among unmarried women. In short, the gap between attaining adulthood and age at marriage was the window of opportunity for women in a traditional society that saw industrialization on its margins. This window of opportunity was not a passive range. Three things—schooling, working and marriage—laid simultaneous claims on the 15–19 age-group, which was perhaps the most relevant one for my hypothesis. Historically, we observe two long-term tendencies in work-participation in ages below the mean age at marriage. To cut the story short, in the long run, women started working later, but in larger numbers. In other words, the Mshaped age-wise participation graph shifted further out, while the peaks became sharper than before. On balance, the evidence suggests a net addition of unmarried women to the workforce. How wide or narrow was the window of opportunity during early stages of industrialization? It was relatively wide in Japan, about 4–5 years—the maximum length an unmarried girl would spend in a spinning mill at the turn of the century. On the other side, the window was probably a negative range in South Asia. The mean age at marriage in South Asia has been historically the lowest in the world and rose rather slowly in the twentieth century. The mean age at marriage for women in India was thirteen in the decade 1891–1901, and sixteen in the decade 1951–61. The age was lowest by far by any contemporary standard. In 1920, the mean age for women in Japan was twenty-three. In Europe, the age was twenty-five or more even at the beginning of industrialization. In neighbouring Burma and Ceylon, the average age for women in the first decade of the twentieth century was 18–20. The mean age did not necessarily increase in the long run in regions where it was high to begin with. However, if it fell, it fell marginally.12

Labour-intensive industrialization

128

With the average age of marriage as low as it was in India, the big barrier to leave family in search of work was age itself. This is how culture could hurt. Very rarely indeed were adult women workers inside a mill still unmarried. If in Japan unmarried girls populated the spinning mill shop-floor, in India, the women workers inside the mills were usually married. Not only would the married women not work with men, but they tended to go where their own husbands went. Married women, thus, had limited options about where they would go for work, and employers did not want to spend on welfare or training of such a casual component of the workforce. In turn, they were vulnerable to potential displacement. As far as I am aware, in recent scholarship, the significance of the age of marriage for industrialization and the labour market has been underestimated. But the census commissioners in colonial India, with their noses firmly to the ground, saw the significance. The 1931 report of India wrote apropos women’s presence in industry: The vast majority [of women] are married and married young. They have their domestic duties to perform and other work has to be done at home or nearby. Their general lack of education excludes them from engaging in any but the simplest forms of labour. (India, 1931a: 414, emphasis added) I present below four hypotheses about the nature of women’s work in industry, inspired if not driven by this sharp but terse statement. The first point concerns international differences in levels of women’s participation when the labour market expands. With the average age being as low as it was in India, young unmarried girls had no hope of joining the labour market. That would explain the difference in levels of participation rates between societies. From an identical participation rate when all work took place ‘inside the hut’, some societies could see a peak forming at 15–19 age-group, as work-opportunities expanded from the hut to the factories. But, since Indian women were nearly all married by thirteen, such a peak was missing in India, and the average participation rate fell. This gave rise to a pronounced Ushape. The second point is about the length of the U-shaped graph. The average age also explains the slowness of the transition, namely, the reason why the declining phase or the left arm of the ‘U’ occupied the greater part of the twentieth century. The average age at marriage was a cultural variable, and therefore took a very long time to respond in India. Specific policy intervention might make a difference to redressing the balance, but such intervention needed at least a diagnosis of the problem. Such diagnosis never came. The third point concerns the M-shaped graph. The age at marriage explains why we do not see the M shape or the second peak in age-wise participation rates. Instead of an increased participation (before marriage), a decline (post-marriage) and a rise again (reentry), in India there was only one rather low and rather flat peak. Being already married, women faced barrier to exit at an age when the first peak normally occurs. There being no entry before marriage, there was no re-entry either. The M-shape disappeared. On the other hand, those women who did make the transition from the hut to the factory were too vulnerable. Their age and marital status made them face a barrier to entry, or at least a high risk of displacement.

Women and industrialization

129

The fourth point is about the particular weakness of married women in the labour market, and the reason why they were pushed so often to unskilled tasks. The point is about intra-household distribution of industrial training. Many early mill workers came from an artisan background. Many wage-workers in small-scale industry too were largely artisanal in background. To understand why married women would not be preferred as the core workforce in such settings, we need to understand how informal training systems worked. In an artisan family, industrial training occurred at home or in master-apprentice teams outside home. The longer the children stayed with their parents, the longer was their training. Now, oddly enough, the girls in artisan families—who worked in industry all the time—never figured in apprenticeship relations, were never trained in commercial skills, or trained in handling capital. Surely, when the time came to seek outside work, this differential endowment made men more ready to go outside than women. For, the opportunity cost of women staying at home was smaller. This feature can be explained by age of marriage. When women married young and left home, the parental home was able to recover only a small part of the returns to the skills it imparted. There would then be a bias to impart only generic training to girls. Specific training creates a specific asset. It creates an individually valuable worker or capitalist. For those marrying young, the parental home would neither have the incentive nor the time to create such specific asset. The priority of the marital home, on the other side, was the training of the children rather than that of the mother. Many earlier studies found that modernization and technological change were genderbiased. The reason why women were often more vulnerable than men in the manufacturing process was that they were put to tasks that were so routine and so labourintensive that they could be replaced easily by a machine. Inside the cotton spinning mills in Bombay, for example, women in the nineteenth century performed winding and sizing operations manually. Eventually when that stage was mechanized, women’s presence in the factory floor fell. In the 1970s, some authors used the term ‘inadequacy’ to characterize women’s tasks inside the mill, and were criticized for using culturalist language.13 But there was indeed a real problem here. The issue of barriers to entry in a factory is closely tied with division of labour, in which women tended to perform the more vulnerable roles. I propose that we can understand what we often observe as a given division of labour inside a factory as a sign of the quality of training that girls received in their parental homes. The more generic the training inside the home the more generic one’s work would become outside the home. And the lower the average age of marriage, the more generic is the training at home. There is perhaps a more general moral of this story. I am arguing here that the demographic transition and capability of women were correlated, via the average age at marriage. Girls marrying young are too under-trained and under-schooled to compete effectively with men. I wish to show in the next chapter that the historical material on women’s work in the crafts reinforces the hypothesis.

Conclusion In a traditional society which was beginning to see an urban industrial labour market emerge on its fringes and which did not permit women a lot of independent decision-

Labour-intensive industrialization

130

making about where to work, the gap between attaining adulthood and marriage was the window through which women could step out of tradition, enter the urban labour market before retreating into tradition again. If this is accepted, we must consider the following hypotheses. First, the earlier the age at marriage, the narrower was this window. This factor can explain differences in participation rate, and a faster exit from work when work required going out of the home, that is, the protracted U-shape. Second, the earlier the age at marriage, the weaker was the incentive of parents to impart specific skills to their daughters. Therefore, the earlier the age of marriage, the more vulnerable were married women at their workplace. This factor explains the absence of a second peak in women’s participation rate in South Asia. There is a lesson here about the nature of interdependence between economics and culture. Critiques of modernity remind us that exclusion is built into modernization processes. Indeed, exclusion is part of the process by which workers move from a lower segment of the job market to a higher segment as industrialization progresses. By ‘exclusion’ I mean a systematic and predictable difference among groups in the probability of making the transition successfully. There are built-in tendencies that make some groups less able to make the transition. In a passage in A Theory of Economic History, John Hicks (1969) hinted at a hypothesis about exclusion: The passage from one grade to another is largely a matter of training; and training is a process with which the labour market does not readily cope… Training “on the job”…shares the same defect’. In other words, education, being an impure public good, involves market failure. But informal education, which is a socially culturally mediated process, can involve exclusion or a cultural failure. When these two join hands, we can have a drawn out effect such as that which the South Asian women faced in the industrial labour market. The hypothesis here was suggested in the course of compiling descriptive data on women’s work in artisan industries. The next chapter collects and classifies this historical material. The material goes well beyond the specific concerns of the present chapter. Still, one of the driving themes of the next chapter remains the inequality of endowments between men and women.

8 Women in the crafts Chapter 7 suggested that women worked, but hesitantly outside the family or the community. Perhaps their mode of working inside the traditional milieu would answer why. With that aim, this chapter gathers information on women at work in small-scale industry. I begin with a note of caution. The raw material is fragmentary. Women were generally invisible or barely visible in colonial period descriptions of work. Somewhat detailed information became available in the course of the Royal Commission on Labour in India (India, 1928–31) on women in large-scale industry. Perhaps for the first time, many women workers were interviewed by government officers. There was also a discourse on women’s work in the social service literature of the early twentieth century, though when it came down to details, these details went no further than the mill premises (YMCA, 1926). By contrast, in agriculture and small-scale industry we get no more than passing glimpses of women at work. These fragments do not add up to systematic chronological accounts. What we get instead are some significant hints about the workenvironment. Since the work-environment mattered so crucially to women workers, there is perhaps a point in putting the material together. Despite the difficulties of a chronological reading of the raw material, a chronological perspective does underlie my treatment of this material. It derives from the long, nearly unbroken, decline in womenworker ratio in manufacturing, with which the previous chapter was concerned. In organizing the material, I use a few simple tools. Women’s work was sometimes closely tied with men’s work, and sometimes not. I call the former scenario ‘genderintegration’ and the latter ‘gender-independence’. Whether or not men’s and women’s works were connected, usually men controlled capital and marketing. Correspondingly, women worked the routine labour-intensive processes, and rarely had contact with the buyers. These features-connectedness, access to capital and marketing, and labourintensity—are of some importance to understanding occupational choices and changes. Gender-integration is important because in the presence of integration, decline in women’s work almost always implied a weakening of the whole household as an industrial organization. In the gender-independent case, decline in women’s work was not necessarily linked with the fortunes of the household in the market. The relatively greater access of men to capital suggests that men and women might face different opportunity costs of leaving industrial work altogether. Lastly, the high labour-intensity and the routine character of women’s tasks meant that these tasks were particularly susceptible to mechanization and substitution of women for men. The simpler the task, the less indispensable was the individual worker.

Rethinking economic change in India

132

Gender integration: textiles In cotton hand-loom weaving, the household was universal c.1900. Women and girls performed warping of the yarn, winding the thread on to bobbins for use as weft yarn, and shared with men the task of sizing, that is, applying starch paste on the warp thread. Weavers as a rule were men, sometimes assisted by young adults, almost invariably boys. About one in ten women in weaver families knew how to weave, they learnt it usually half-heartedly and from observation rather than from systematic apprenticeship such as the boys underwent.1 The situation changed in the late twentieth century. Already in interwar South India, in towns that were growing rapidly as weaving clusters, more women had started weaving for wages. In Salem c.1950, a significant percentage of women weavers could be found in small hand-loom factories. Today, women are common in hand-loom factories and, in Salem itself, export-oriented garment units employ a large number of young women workers.2 There is scattered evidence from the early twentieth century that the relatively high work-participation rates in weaving families were connected with the demographic characteristics of the population. Weaving families were on average slightly larger than the average family, had larger number of workers on average and a higher ratio of female-to-male workers (Bengal, 1937:37). There are many descriptions of women’s work in cotton hand looms. The ‘independent weaver’ in this industry was defined as one who owned a loom and worked in one’s own house. The members of his family assist him in winding, warping, sizing and occasionally in weaving too. The women throw the warp and wind the yarn for pirns. They sometimes assist him in sizing. The children help him in bobbin winding or in the manipulation of the harness if it is the drawloom. The weaver sizes and weaves. When the cloth is ready he takes it to the nearest bazaar every evening or to shandies every week for sale. In some places he himself hawks it in streets. (Venkatraman, 1935–36:62) Warping, winding and sizing were not always done inside a home, but sometimes hired out to neighbourhood women. And frequently these were done by teams of adult and young women in a commonly owned premise. Such pooling of labour of the community was standard in rural Bengal and Madras c. 1900. Thurston (1899a:9) wrote in respect of South India, As you enter a weaver’s grove, it appears at first as if those occupied in this industry were engaged in a pretty game. Rows of women walk up and down the shady aisles, each holding aloft in the left hand a spindle, and in the right a bamboo wand, through a hook at the end of which a thread is passed.

Women in the crafts

133

At the centre of the weavers’ quarters, in an open space shaded by trees, a few sticks were fixed on the ground. The warper would take a hank of yarn, fasten a thread to the first stick, and then go on unreeling it round the sticks until the last one was reached, when the course would be repeated backward. Only one thread (or one end) at a time could be reeled this way, and before the threads required for a given width of warp were reeled, the warper would have walked several miles. A survey team found in 1941 a twelve-year-old Bengali girl who walked seventeen miles to prepare the warp of a ten yard sari of about 3,000 ends, for which she earned Rs 0.125 or a little over two pence.3 When this work was finished, the thread was sized, that is smeared with starch paste, either with a brush or by immersing the whole warp in the compound. Even such intensive labour was not enough to engage a weaver for more than a few days at a time. And for several months of the monsoon, the weaver and his family would be idle for no other reason than the rains made warping impossible. In the dry months, the work was usually done in the afternoon, as in mornings the women had to do domestic work. The formation of warp in silk was almost an identical process, but for sizing. Thurston, again, can be cited: Bamboo sticks, 120 in number, are fixed upright in the ground, generally in the shade of a tope or grove, at a distance of a cubit from one another, and ten women or children, carrying ratnams [hand-spindles] in their hands, walk up and down this line, one behind the other, inter-twining the thread between the bamboos, until 1,920 threads of various colours, according to pattern desired, are thus arranged. For this work each gets half an anna—a small remuneration for walking four miles. To form a warp sufficient for eight women’s cloths, 40 miles have thus to be traversed. (Thurston, 1899b: 10) Warping systems of this kind combined three features: communal, manual and outdoor. Such systems, however, were decaying throughout the region in the twentieth century, though even today they have not disappeared completely from rural South India. The drum warping system, installed in a factory set up, was more capital-intensive than its predecessor. But it popularized the use of a warp beam, thus enabling a longer length of warp to be prepared at a time, and allowed for sizing and warping to be done as a joint operation. In the weaving towns of Tamil Nadu and the Deccan, it was a wellknown implement and had displaced stick-warping extensively in the mid-twentieth century. But it was not yet popular in Bengal, or in the north. That was partly because the warping mill was yet to find a satisfactory way of combining warping with hand-sizing, which gave hand-loom cloth distinctiveness. In places where it was adopted, weavers had to find ways of bypassing the problem. These warping mills continued, with only a slight change in technology, into the power-loom weaving of the present day. But today, the warping factories allied to power looms are run almost entirely by male labour. In handloom weaving, warping is still done partly by women, whether inside a mill or outside. While warping came into larger-scale semi-mechanized work-sites, cotton hand-loom weaving too was becoming concentrated in factories, in western India particularly. These factories, called ‘karkhanas’, were common in silk weaving also (Joshi, 1936:120).

Rethinking economic change in India

134

Both these shifts in weaving and in processing had a gender bias, even though the extent of the bias was variable over time. The urban hand-loom factories employed migrants. The single male migrants dominated the working class. And yet, as the factory movement strengthened, many of these migrants brought their families into the textile towns often braving a much harsher lifestyle in the process. For example, until the 1930s, many among these weavers lived with their families in single-room tenements not larger than eight feet by ten feet. Living conditions improved thereafter. Between 1883 and 1910, weaving towns in Bombay-Deccan experienced a noticeable change in…the methods of production. That change was the gradual specialization of the preliminary processes of sizing and warping, by a different group of workers consisting of women and inefficient weavers. (Joshi, 1936:72) The presence of women was rising again as migration became steadier and less malebiased. In a somewhat different context of the hand-loom factories of the Malabar Coast run by the Basel Mission, ‘the fear, that the organization of the factory will eliminate women labour or that women would not come to the factory for work, is unfounded’ (Venkatraman, 1935–36:72). And yet, on average, the shift from rural to urban, family/community to factory, settled to migrant labour and, above all, manual to semi-mechanized technical options, tended to displace women. K.S.Venkatraman, a contemporary authority on the handloom industry, generalized this point based on his field-work in the mid-1930s: ‘a careful study of the practice of the industry will show that men have taken the place of women sizers who have gradually disappeared as a result of the introduction of long warps’ (Venkatraman, 1935–36:54). The process was mediated by an interconnected change in technology, organization, market and location. The textile towns needed mass production systems in weaving, in turn requiring mass production on the processing side. The household could not process long warps. The factory could. And men were either preferred in factories, or more easily available. Venkatraman studied Tamil Nadu and Bombay-Deccan, whereas N.G. Ranga, another specialist of the subject, had studied Madras-Deccan and coastal Madras a few years before. Generally, in the area toured by Ranga, the majority of the weavers ‘get their preparatory work done by their own women, and so a weaver’s wife does not get any outside earnings’ (Ranga, 1930:68). However, women supplying household labour for free did merge sometimes with a system of limited local hiring of women. A family might hire the services of neighbourhood women on a casual basis to tide over a temporary rise in demand. Because such hiring existed, we can calculate women’s earnings separately from the household income. The earning of a small family of MadrasDeccan in 1925 was Rs 8. There were rare instances when a woman would work full time (from six in the morning to six in the evening, with an hour’s break) in warping and earn Rs 6 per month. Clearly, the shift from family to wage labour, unless accompanied by a male bias, was a distinct economic improvement for the family. That said, the magnitude of the improvement was not always as large as this example would suggest. In the South Indian silk-cotton weaving town Rayadurg, there were a few weaving factories working

Women in the crafts

135

in the mid-1920s. ‘ln most cases, a weaver and his wife go to work in the employer’s house’ (Ranga, 1930:76). Men engaged in rather plain cotton sari weaving earned Rs 0.63 per day, and women, employed by a warping factory on preparatory processing, earned Rs 0.18 on average. It is easy to see why low wages might persist. The shift from the family to the factory required in theory a smaller number of workers for the same job, for factory workers worked longer hours. Redundancy, thus, was built into the shift. For women, the effect of redundancy would be more intense, as they had fewer occupational choices than men. Factories were rarer in South India compared to West India, but not unknown. There is an interesting description of the Rayadurg factory in Ranga (1930:66): The houses of employers in which these workers labour…have tiled roofs and are too long and too narrow and are crowded with half a dozen looms, many charkas, a dozen women and a score of men in addition to an unruly crowd of children, so that in them it is very hard to hear one another speaking. Note the presence of children on the factory floor, a point I called taking home to the work-site in Chapter 7.1 shall return to this point later. Ranga’s informative survey of South Indian weaving developed contrasts between weaving in the arid Madras-Deccan, which was in a depressed state in the 1920s, and weaving in the coastal Deccan, which was in a state of expansion and relative prosperity. It is not always clear whether he deployed this contrast as a descriptive tool, or there was a real difference here. There was probably some truth behind this broad generalization, and it can be understood with respect to their contrasting agrarian experiences. The very dry and famine-prone Madras-Deccan saw large-scale emigration of weavers towards the textile towns of the west, overseas and the plantations in the north-east. Coastal weaving, at the same time, saw some kind of rejuvenation as an agrarian expansion and commercialization improved levels of living in the countryside. Ranga saw the contrast nowhere more starkly than in the ‘level of contentment’ of the weaver womenfolk. In the Deccan, A woman works from the morning till the next morning, with a short break of five to six hours. Her children, her charka, the hanks she must prepare; her possibly drunken husband, or husband with many concubines, and her unkind neighbours; all cause her immense anxiety; in addition to the tiresome work of cooking, feeding, turning the winding wheel and preparing the warp she has to do… She goes on cursing, grumbling, murmuring and muttering abuse and receives reproaches and blows from her husband many a time during the day. In many a case she is as addicted to drink and drugs as her husband. (Ranga, 1930:114–15). But on the coast, the women were ‘less hardworked and better treated… as the conditions of the weavers are better there’. In either case, what she earned in money, she had neither banks nor insurance to deposit it in. ‘Every penny she earns is given to her husband, and

Rethinking economic change in India

136

though she has some voice in deciding the family budget, the husband has the final say’ (Ranga, 1950:114–15). Edwardes (1900) on the Patwegar silk weavers of Bombay-Deccan does mention tersely that ‘their women also lend a helping hand’. However, what makes the accompanying text interesting is the description of the long apprenticeship that the boys of weaving families went through in weaving and dyeing.4 It is interesting because we have no counterpart on the training of girls in whatever they did as adults. This is the norm in contemporary descriptions of craft households, with one major exception: embroidery and carpet in Muslim households (Edwardes, 1900:51). The gender-integration scenario reappears in many other contexts within the textile industry. Although the (wild silk) tasar-weavers of Chotanagpur are uniformly referred in the masculine gender in a turn-of-the-century source, several photographs in the monograph show women both as spinners and as weavers. Indeed, ‘the great merit of the industry [tasar near Gaya] is that it gives employment to men, women and children; the first weave, the second spin, and the third set the warp. Labourers are not employed by the patwas at all’ (Mukherji, 1903:111, 115, 122). Among the silk weaving Momins of Deccan, usually urban communities, ‘their women help largely in the weaving and embroidery of wearing apparel, and add as much as the husband to the monthly income of the family’ (Edwardes, 1900:55). Among transhumant shepherd-weavers of Bombay, ‘one of the weaver’s family, generally a woman or a boy, [spins] the wool into yarn’ (Brendon, 1899:3). Similarly, Gareri women in south Bihar were spinners of wool, while their men reared sheep and wove coarse blankets (Banerjei, 1899:32). A rather different form of gender-integration is observed in sericulture. These were cases in which men and women members of the family were engaged in the same industry, but the men were usually recorded in official sources such as the census as cultivators and the women as industrial workers. Thus, among cocoon rearing families of Murshidabad in central Bengal, many had mulberry land. Mulberry growers in this region were men, whereas cocoon-rearing itself was a female occupation. One often sees an old and experienced woman being called in at [the] critical periods to judge whether feeding should be stopped or feeding should be recommended. Women do most of the work in connection with the rearing house, while men look after the mulberry, cut and bring it home for the silk-worms, and assist the women in feeding and cleaning. (Mukherji, 1903:11) Many of the mulberry growers and cocoon rearers ‘would call themselves cultivators, and they would thus be excluded from the [census] tables on the silk industry’ (ibid.: 11). Mukherji’s report was written at a time (c. 1900) when the silk weaving industry was flourishing in Bengal, while cocoon-rearing industry was languishing because of a series of epidemics among silk-worms. The decline in cocoon-rearing in that case would not lead to a change of occupation for men, but may lead to a change of occupation for women. This kind of division of labour was common in sericulture industry throughout Bengal. In the semi-wild silk (endi) industry of Mymensingh, cocoons as usual were reared in the houses of Muslim cultivators by women throughout the year. They carry on

Women in the crafts

137

the work of rearing the worms at home in the midst of their household duties’ (ibid.: 146). What is noteworthy in these contexts is that the men almost invariably worked the more valuable resource or controlled the source of earnings. The men among cocoonrearers in Bengal did more work with mulberry cultivation, and were thus in control of land. In the case of endi production, ‘the threads, when ready, are sold by the male members of the family in the local markets’. Further, Mukherji’s report on Bengal sericulture was written at a time when government-sponsored efforts were under way to replace the household industry in silk-worm rearing with seed-nurseries or grainages, and to introduce sericulture as a subject in industrial schools. The workers in the grainages as much as the pupils in these schools were to be men.

Gender-integration outside textiles Among the semi-nomadic iron-smelters of Mahabaleshwar hills, the Dhavad Muslims, women ‘do as much work as the men, bringing head-loads of fuel and grass from the forest’. Because of cheaper foreign ironware, the indigenous industry was nearly dead c.1900. The Dhavads lived by cutting grass, gathering honey and making and selling iron nails, tongs and frying pans (Scudamore, 1907:8). The village smith, Lohar, did not die out and practised a great deal of job-work for the cultivators in the twentieth century. They were also hired in significant numbers in railway workshops, foundries and other newer forms of metal-working factories. These communities followed quite a different practice. For blacksmiths generally, ‘the womenfolk never appear to render any assistance’, the truth of which can be easily verified from census data (Watson, 1907:35). The 1881 census of the Baroda state singled out several professions of western India, the important ones being tailoring, leather-work and pottery, where ‘females of many of the artisan…classes work with their husbands and relatives’ (India, 1881b: 207). As in hand-loom weaving of western India, some of the other examples of the genderintegration model saw an extensive shift from the home to the factory in the interwar period. Coarse woollen blankets are one example. A more important example is tanning. There was a large-scale shift of tanning from the rural to the urban and from the home to the factory. Among rural tanners and leather-workers, women ‘helped’ men universally. But in the factory the workforce in tanning was emphatically male. In leathermanufacturing, households and women workers withstood migration somewhat better. Despite the growing popularity of foreign shoes, the village shoemaker in parts of western India retained his hold on the local market, Each tanner or shoemaker works for himself, aided by brothers or other relations, and usually in the lighter parts of his work by the women of the family…he women [of Dhors] look after the house and assist the men in the lighter parts of tanning work, such as breaking barks. (Martin, 1903:24, 28) In the Mochi workshops of Bombay city

Rethinking economic change in India

138

The women assist by doing the silk embroidery which edges the mouths of the shoes, or other ornamental work of this nature. In this way they earn about annas 4 per day. They live in overcrowded dwellings, as many as three families sharing one room. (ibid.: 26) Most were migrants. Somewhat later, migrant ‘pardesi’ Mochis in the Khandesh industrial town of Dhulia ‘work in their own houses with the help of outside labour and of the members of their family, the latter generally helping in the final processes’ (Joshi, 1936:126). Even though the extent of survival of the family in leather was relatively greater, the household did not withstand the rural-urban shift everywhere equally well. The interwar Calcutta’s leather industry, for example, was being peopled by migrant Punjabi Muslims, who constituted a pre-dominantly male workforce.

Gender-independence In households not necessarily connected with industry, some types of industrial work were often exclusively done by women. There were perhaps two broad classes within this group: the first was rural, and the second urban. Examples of the former describe very part-time, very low-paying, often waste-recycling, kind of labour, which engaged the otherwise unemployable sections of the labour force. The latter were often better organized and more commercialized. It is in this latter sphere that we encounter allfemale apprenticeship systems. The spinning of matka silk (pierced cocoon, cocoons from which the moth has escaped) of central Bengal is an example of the former type, and perhaps representative. Matka-spinning and matka-weaving give occupation to the poorest of women and the least artistic among the weavers… Women doing the work during their leisure time at home, do the spinning of about 400 cocoons per day, not at one sitting but as they get time. It takes six or seven hours’ continuous work spinning 400 cocoons in this manner… A woman spins only about an ounce of thread every day A…seer of matka thread sells for Rs 2 to Rs 6… The profits of matka-spinning industry are extremely low; and it is only old and feeble women, as a rule, who carry on the industry. The only thing to recommend it to them is its ease and inexpensiveness. The woman can do the work at her leisure, in the midst of her ordinary domestic duties, and the appliances required are only a takur [spindle] and a latai [wooden roller to wind the thread on] and a mud vessel to hold the cocoons… After a month’s toil she may make a rupee or two. (Mukherji, 1903:22–3, emphasis added) Scattered examples of specific women’s work performed by vulnerable sections of the female workforce can be found elsewhere. For another example, among the Surat-

Women in the crafts

139

Ahmedbad silk weavers, generally Muslims, the warps were usually arranged by poor women of the Khoja community (Edwardes, 1900:41). North Indian embroidery, on the other hand, is an example of the second type of enterprise. In northern India, embroidery was more than a commercial skill, it was part of the education of girls in Muslim families, rich or not-so-rich.5 The highest stages of its development, as a commercial craft and as a style, can be seen in the towns of the United Provinces, of which Agra and Lucknow were probably the most important. When under economic pressure, adult women formed a team and carried out job-work for urban merchants. This kind of domestic team-work steadily expanded in the interwar period. Women also sometimes dyed cloth, usually for their own use in their houses. If embroidery was mainly a female work, professional dyeing was mainly a male work. But in Chittagong in south-eastern Bengal, the rule was broken. Local ‘Mugh and Mussulman women are reported to carry on dyeing’ commercially. Men and women alike took part in the dyeing operations allied to pottery and the manufacture of wooden toys inside household industry (Banerjei, 1899:9–11). One large example that falls somewhat between gender-integration and genderindependence is carpet weaving among semi-nomadic tribes of Sind. These communities raised sheep and handled a lot of wool. Some of them were also, at least in the twentieth century, cultivators. However, even when the connection with wool had weakened, their women maintained a tradition of carpet weaving for non-commercial use. The women of Chandias and Brahuis of Nasirabad in Sind were examples. ‘One old woman of seventy asserted that when a girl is born the mother commences making a carpet which becomes the mother’s wedding gift when the girl is married’ (Twigg, 1907:15). These people were not exceptional at least in Sind and Baluchistan. But compare them with the labour situation inside a carpet factory of South Asia. On the eastern coastal town of Eluru, factory ‘weavers are Muhammadans, and, as the women are gosha, only the men are wage-earners’ (Harris, 1908:13). The characterization would apply to the carpet factories of Agra, Amritsar and Srinagar in the interwar period, almost wholly occupied by male master-apprentice teams. The great women-artisanate tradition of Assam and northeastern India belongs in the same class as the carpet weavers of Sind-Baluchistan. They were women of peasant households who had developed home weaving as a domestic cultural tradition. The tradition commercialized only late in the twentieth century. Another large example in the gender-independence scenario is hand-spinning. This example has occupied, not unjustly, an overwhelming part of the discourse on women’s work in industry. However, a great deal of this discourse is impressionistic. The first thing we need to note about hand-spinning is that it was quintessentially women’s work: very small pay, very labour-intensive and usage of the most unemployable of the rural workforce. Such labour was easily a prey to mechanization because it was much too labour-intensive to be economical for the weaver, and because the income-loss was negligible for the community as soon as an alternative became available. The peak period of the decline of hand-spinning was not well-served by documentation, official or other. However, spinning was faintly remembered at the end of the century. Based on these reports, we know that spinning labour came from across the social spectrum. Women of Brahman households took part as much as poor widows such as the ones described in the matka example above. As one respondent remembered the routine: ‘Females in every household would get up early in the morning and sit in small parties of five or six and go

Rethinking economic change in India

140

on spinning’ (Slater, 1918:66). The mention of caste is significant, for team-work was not likely to form in too mixed a company. Another work that was exclusively performed with women wage labour from the beginning of a mass market was bidi (country cigarettes) rolling. The census proposed that the popularity of foreign cigarettes had increased the smoking habit among both urban and rural Indians (India, 1931b: 203–4). However, after being initiated, these new converts did not necessarily persist with foreign cigarettes. The demand for hand-rolled cigarettes, therefore, increased greatly. Bidi quickly emerged as a lucrative business. Bidi was never a household industry on a significant scale. Large factory yards employed mainly women wage workers. The work was simple and monotonous, and the pay good enough only if large volumes were produced. This market never seriously attracted men. A different type of gender-independence is encountered among those rural communities primarily engaged in labour. Large numbers of people in central, southern, western and northern India were agricultural labourers in the peak seasons, and took part in a whole variety of industries and services at other times. It was to emphasize the contrast with such diversified labour communities that the term ‘caste weaver’ was sometimes used in contemporary sources to refer to specialized artisans who considered themselves above labouring and even cultivation. Among the former people, it was often the case that women took part in agricultural work to a greater extent than men. As Gilbert Slater’s team found in rural Madras during World War I, men were frequently engaged in weaving at the same time as their women took part in casual labour in agriculture (Slater, 1918:66). This kind of combination—men in weaving and women in labour—became steadily rarer in the twentieth century. It is possible to advance a hypothesis why it became rarer. The kind of occupational ‘portfolio’ that such poor rural households displayed was not a strategy that could give them rapid income growth, for individually these occupations were very low-paying indeed. It was more likely a strategy for survival, a strategy for tiding over crisis such as famines, or tiding over periods when agricultural work was difficult to get. Steadier and less risky prospects of wage-labour in agriculture could break up such mixed occupations easily. After 1900, when famines and scarcities became rarer, such specialization away from industry did become more viable for rural households.

Continue or abandon? Numerous examples suggest that when the household faced a crisis of survival, men tended to persist with manufacture much more than did women. Venkatraman (1935– 36:55, 58) was particularly insightful when he generalized the gender difference in this way: As weaving is fast ceasing to be the partial occupation of agriculturists and the ideal domestic industry of women and is tending to become more and more the specialized calling of a distinct group of workers as a result of the advances made in the technique of the industry, there is naturally a fall in the total number of cottage workers…. The weavers …are slowly concentrating in towns and colonies where there are certain facilities viz.,

Women in the crafts

141

the preparatory processes being done cheaply and quickly by special warpers and sizers. This means that women who are generally employed as winder, warpers and sizers in a self-sufficient cottage system of production are now thrown out. Men replaced women via a change jointly in technology and organization. The women of artisanal families sometimes continued into the factory system as we have seen. But more often they stayed back in the village. And when they did, they withdrew from industry both because the gender-integration and the household as a unit had broken up, and because women had too little access to capital and consumers to carry on in the industry on their own. The other side of the specialization that Venkatraman described was not only withdrawal of women from the workforce but also their increasing participation in agricultural labour. On the other hand, the prejudice against field-labour among weavers, necessarily men, tended to be strong. Men persisted with their hereditary calling. A common example of this persistence comes from silk weaving faced with adversity. Silk weavers shifted to cotton weaving rather than giving up industry altogether. Even matka-weavers, the poorest among silk weavers, ‘being weavers…are ashamed to…work as field-labourers’. Examples of similar shifts within textiles for men weavers come from coastal Orissa, Dharmavaram and among the dyers of Bombay-Deccan (Mukherji, 1903:43, 121; Ranga, 1930:78; Fawcett, 1896:2–3).

Conclusion I advance three hypotheses about why women might leave industry when continuing in industry required taking part in wage-labour. These hypotheses illustrate the conclusions of Chapter 7 somewhat more fully. As Venkatraman reminds us, technological change required certain processes to be done in a different location, at a faster pace, with full-time or specialized labour. Temporarily at least, it was difficult for families to migrate outright, given the costs of such a move. Migrants, therefore, were always predominantly male.6 In a number of instances in which we see women working long hours outside the home, the children necessarily accompanied them. This was almost always true of peak-season field labour, and was true of the Rayadurg factory described above. In the cotton mills, women’s work inside the factory depended on a support system outside to look after young children and bring them to their mothers at certain times and places. Not all mothers could access or create such a system. And those that could needed to have large families. For, older siblings played a crucial role in this scheme. My second hypothesis is that men, being in control of capital and marketing and having undergone more rigorous training, faced greater opportunity cost of outright exit from manufacturing than did women. They persisted with industry much more, if necessary, leaving the family and resettling in a different town. Women as merchants were probably not unknown, but the rare instances in which we see them involved odd personalities and odd domestic circumstances.7 This is a point about uneven levels of investment in human capital, which can be made the other way round. Since women from

Rethinking economic change in India

142

a very young age were constrained to perform domestic work and child rearing, the family might see it as too costly to train them in those tasks that required sustained interaction with the world outside. My third hypothesis is that it is precisely these constraints that made it more difficult for certain products to survive commercialization. The outcome of distinctively women’s labour such as waste silk yarn, carpets, embroidery, or a sari meant for display of its maker might be distinctive and might have potential market, but these products were not backed by a strong commercial organization. When a market did grow for such products, only in rare instances did women artisans supply that market. In carpets, they were almost wholly displaced, in sari and embroidery partially displaced. Commercial organization evolved in some of these cases to utilize the particular craft skills that only women could supply, but usually, involved men in trade and women in production. The distinction between gender-integration and gender-independence matters in this context. When a product moved from the household to the factory, we are talking about a gender-integrated scenario weakening. The decline in women’s work in such a case implicated (a) technological change that made the factory a superior organization, and (b) differential gender endowments in production skills that made men better able to join the factory. When a product earlier made by women, but not by the whole household, began to be made in a factory or by men, we are talking about a gender-independent scenario breaking up. That kind of change implicated not differential skill-endowment, but differential endowment of capital. Women might be able to make that item better than men, but could they market it? Not usually, unless there were certain institutional changes. Such changes occurred particularly in Lucknow chikan embroidery or Assam handlooms in the late twentieth century. But these were rare instances.

9 Labour and power A critique of ‘subaltern studies’ In the 1970s, the orthodoxy in economic history proposed that peasants were a passive victim of adverse power relations. Thereafter, the weight of economic history in regional scholarship weakened, and the study of peasants and labourers in rural India took a political turn under the influence of the ‘subaltern studies’. In the 1980s, this collective of historians argued, partly in reaction to mainstream economic history, that the poor could attempt to shape their own destiny by resisting exploitative power. The message endured into the 1990s. But in the process, the historiography of ‘agency’ tended to impose a structure of its own on the way this ability of the poor was conceived and discussed. It made the discourse too political. Repression, resistance and collective protest seemed to exhaust the landscape of agency. Throughout, this book brings in other dimensions of agency. I have argued here that ‘power’ is not necessarily, not usually, a helpful construct in studying the economic history of the working poor. Resources and markets are better constructs. ‘Agency’, in that case, connotes much more than angry mobs using ethnic symbols to express solidarity, the level of empiricism at which the early subalterns confined themselves. Narratives of making-a-difference, can also involve migration, innovation, occupational shifts, allocation of family time, or organizational change. These methods, needless to add, were mediated by the market. Furthermore, whereas a repression-resistance paradigm implicitly commits itself to studying groups and group action, the marketmediated approach can potentially bring in the issue of micro-level decisions and, via these, bring the relationship between the collective and the individual to the fore-ground. Previous chapters illustrated some of these varied ways that the poor could shape their destiny. This chapter takes up the subaltern studies paradigm for a critique from the perspective of a non-Marxist historiography. To my knowledge, in the highly fertile field of critiquing the subaltern, that particular plant has not been seen yet.

Introduction Subaltern studies, which completed twenty years of existence in 2001, has generated as much disquiet and confusion as creative scholarship in the last decade.1 The concept was first applied in a specific context—peasant insurgency. The intention was to locate an autonomous field of peasant consciousness that expressed itself through resisting repression by the ‘elites’. Quickly thereafter, ‘subaltern’ occupied a place at the centre of colonial Indian historiography. In the context of an unsettled debate between ‘class’ and ‘caste’, ‘subaltern’ seemed less problematical than both, and became a benchmark for

Rethinking economic change in India

144

writing the history of the underprivileged people in South Asia. The notion of an autonomous mindset gave a clue to the question animating a generation of political historians, why the subalterns did not come together as a class with revolutionary potentials. The project has faced criticisms. Generally, the criticisms have come from insiders, in that the critics share the belief that the concept ‘subaltern’ inaugurated a significant chapter in Indian historiography. The criticisms stem from a sense that the idea behind the word is restrictive, perhaps increasingly so as subaltern studies stepped out of the history of Indian nationalism and moved into the cultural history of colonialism. But few historians would go so far as to argue for a rejection of the term in writing the history of the poor in colonial India or contemporary South Asia. The methodological success of the project can be seen in the way scholars of South Asia now employ the term ‘subaltern’ casually as if their references are known and acceptable to all readers.2 The most lasting impact of this enterprise, thus, is in the language employed in histories of the poor and the ordinary people. But this ascendancy came at a cost. Even as history and anthropology came closer through a shared interest in repression and resistance, other ways of thinking about the experience of the poor and the underprivileged—economic history foremost—became submerged in the political mode of reflection. More or less the only dialogue between economic and political historians that the subaltern studies ever could generate was possible with the Marxists for whom economic relations were naturally coercive and confrontational. Ironically, outside that establishment, economics and economic history were steadily losing interest in coercive modes of conduct at the same time that these dialogues were going on. At any rate, subaltern studies never seriously tried to clarify the relationship between the political and the economic. The disciplinary breach widened with the late subaltern studies (voluems VII to X). These essays are nearly indefinable as a set. However, two things that do hold several of them together are (a) an interest in the subjugation and appropriation implicit in Western knowledge about colonial societies, and (b) a profound disinterest in economics. In the section that follows, I present a longer outline of this intellectual journey and the questions that it faced from sympathetic critics. That done, I go on to voice a critique of my own that is much less sympathetic than the ‘insider’ criticisms that I refer to above. In the next section, I wish to elaborate on the causes and consequences of the disciplinary breach and closure of dialogue that the subaltern studies brought about in the study of the working poor of colonial India. The last section summarizes the conclusions of the essay. It is indisputable that in some themes of political history, the ‘subaltern’ made a theoretical and empirical contribution. But as a historiographical tool applied to the experience of the poor, the term ‘subaltern’ is by definition limiting rather than helpful. This argument, which applies more to the programmatic claims and pronouncements rather than to any specific article in the subaltern studies volumes, consists of four points. First, the term ‘subaltern’ is limiting because of the compulsion built-in within the very word to contemplate only one axis of human interaction over all others—the repressive— and one form of agency over others—protesting repression. The idea that the poor could make their own history, a key message of the subaltern studies, is there to stay. However, the poor could do so along several fronts—domination-submission-resistance being one of these, and market participation, for example, was another. Second, market and

Labour and power

145

resource conditions could weaken or strengthen relationships of subordination. Third, the actions of the colonial state, via markets or political means, could similarly weaken or strengthen subordination. And fourth, the colonial state and the colonized subject relationship cannot be seen as fundamentally repressive. It was an ambiguous and a changing relationship. For the same reason, colonial knowledge too had a fundamentally ambiguous relationship with power and repression. No total critique of colonial knowledge about subject societies can become convincing.

Outline Subaltern studies began in the early 1980s as a critique of the history of anti-colonial movements as written in the then dominant schools of political history. It was a critique of the history of anti-colonialism as an eliteled top-down mobilization effort. The focus on elite interests and actions was common to the so-called Cambridge School and the version of nationalism written by the more vocal participants in anti-colonial movements. Peasant movements figured as subplots in this story of freedom struggle. Subaltern studies brought the peasant to the centre. These studies also emerged from a dissident left background, related with contemporary attempts to write history from below, but discontented with the orthodox Marxist way of reducing peasant movements to a subplot in the story of struggle against capitalist exploitation. In neither did the peasant appear as the conscious subject of his own history (Guha, 1982). These ‘totalizing’ narratives subordinated peasant consciousness to structure or elite consciousness, both supplying models of external determination of the history of the poor, the oppressed and the peasants in particular. Subaltern studies recovered peasant consciousness as something distinct and independent of elite actions and elite consciousness. It did so by arguing that subaltern subjectivity is constructed through resistance. The very word ‘subaltern’, bringing in Antonio Gramsci into the theorization, implies a battle, in this case a battle to regain rights or resist processes of repression, subordination and marginalization. Subalternity is here defined along only one axis—that of repression. But it is evident that there is another axis along which the subalterns chosen for doing subaltern studies can be located—their economic situation. There is no subaltern without limited access to assets and incomes. There is no subaltern without deprivation. Now deprivation is not static and, if not, it can independently influence subalternity. The poor try all the time to become non-poor, and sometimes they might succeed. From the beginning the subaltern studies were inattentive to the relationship between these two axes, or even about admitting that there is an independent variable that might matter to subalternity. Can economic change strengthen or weaken political situation? There was no answer. Implicitly, the result of not thinking out this relationship fully was a submergence of the economic in the political. In the practice of South Asian history in the following decade, the notion that repression and resistance should naturally be the focus when one is studying the history of the poor has not been subjected to a serious questioning. Once repression was left to be all that mattered to the historian, the journey of subaltern studies from nationalism to the cultural history of colonialism was an easy step. The gurus of French post-structuralist thought asserted that there was not one history of

Rethinking economic change in India

146

human progress, but repression and resistance were written into histories of progress, especially the liberal humanist narrative of history as a progress of elite consciousness suggesting at the same time that those that failed to keep up did so because of their own shortcomings. A critique of orientalist knowledge emerged from the same roots via the writings of Edward Said. From about subaltern studies VI, the subaltern was less of a political rebel and more of a voice silenced or transformed in categories of Western knowledge about colonial societies. Domination was still the centre, but domination was seen less often in political-materialist terms and more often in cultural-discursive terms. With this turn, three things happened. First, ‘early’ and ‘late’ subaltern studies became distinct. Second, the subaltern studies school entered the world stage using the postmodern bridge. Third, the breach with economic history became total. The subject of subaltern studies was colonial societies and not the poor any more. Every colonial is, by a wide definition, a member of the subaltern citizenry. Economics could be shut out from history with a clear conscience. The late subaltern is hard to characterize. But if there is any unifying theme that runs through a fair number of the late 1990s essays, it is ‘insurrection of subjugated knowledges’. The phrase comes from Michel Foucault, whose radical proposal on historiography has two basic premises: (1) theory serves political goal and (2) that theory rules, or claims validity, which has the maximum power-effect. His proposal was an alliance of scholarship with disqualified knowledge. Since theory exists to serve power, the rehabilitation project had nothing to do with empirical verification of the claims of different theories. ‘Genealogies’, in this sense, are ‘precisely antisciences’ (see Foucault, 1994, on genealogical history). The project, however, was not easy to implement without falling into contradiction. In practice, it went through on the basis of a polarization of included and excluded discourses. But Foucault also argued that domination is everywhere, it is the stuff that human interaction is made of. If domination is everywhere and every discourse is simultaneously dominating and dominated, the project is unsustainable. And if discourses are indeed polarized into ‘unitary’ and ‘subjugated’, the view of power as having no centre is unsustainable. If historians must work at all, they must work by assuming that in any hierarchy of discourses there might be another element involved—their claim to scientific validity. This is not to deny that domination can work through knowledge. But historians must address the coincidence or contradiction between power-hierarchy and truth-hierarchy. The late subaltern studies dealing with colonialism have not been discriminating about this problem. There tends to be too much focus on one overarching form of unfreedom— colonial power—and neglect of those regimes of unfreedom that ‘disqualified knowledge’ of various types served. Colonialist sources are discounted as expressions of power even as dependence on such sources continues. Verification is of low priority, leading to essays that are too bound to the words in texts. Narrative history is distrusted, so that histories that cannot but be done in the narrative mode—economic history in particular—have to be excluded. Early subaltern studies were the subject of a lively debate. The late subaltern studies, while opening up the field to incorporate a variety of experiences, also generated as much alarm as dialogue. Critiques of both revolved on a few fixed points, with what might appear as tiresome consistency. Four of these were especially important. First, making subjectivity the centre of history, displacing and at times de-emphasizing structure,

Labour and power

147

exposed the programme to the criticism of neglect of the material context and modes of exploitation in which elite-subaltern relationships were embedded. Since the history of structural change is by nature more tractable than the history of subjectivity, there was an in-built immobile element in subaltern studies. Second, dividing the world into poles made it difficult to handle situations whenever the elites and the subalterns were not battling one another. Such situations tended to be handled with the notion of either some kind of strategic collaboration or ‘submissiveness’. The word ‘cooperation’ was not used for that would overturn the whole programme. Third, a history that has domination at its centre cannot stand without a definition of what the subaltern figure really is like untouched by power. From early on, this need led to the formulation of ideas of homogenized peasant communal consciousness that became a target of repeated criticism. The formulation of the ‘subaltern’ as a universal category did not go well with a critique of the idea of the universal human subject. Fourth, subaltern studies from the beginning suffered from a need to define and locate axes of domination, with inevitable exaggerations and oversights. In the late subaltern studies, for example, overemphasis on the colonial-indigenous binarism led to understatements of inequities and oppressions within the local society. From its start, subaltern studies regressed into binarism just as surely as critics questioned binarism. Other important points have been made from time to time; but the four major ones were: subjectivity versus structure or agency versus determinacy; conflict versus collaboration; subaltern as universal category versus the critique of universal categories; and polarity versus plurality.

Critique The repression fetish ‘Subaltern’ is defined in dictionaries as a ‘commissioned army officer of lower rank than a captain’. The subaltern is a soldier. The very word invokes a battle. Its use in history begs in a historiographical project. The project takes it for granted that to study the history of the poor is to study resistance. One starts using it with a statement of faith: ‘what is fundamental to relationships in South Asian society is not negotiation…but domination’ (Rosalind O’Hanlon in Ludden, 2001:149). By means of this axiom, subaltern studies privileges one type of relation over all others. It equates history with repression. It recovers subaltern subjectivity necessarily via negation of colonial or elite repression. All this is crippling as soon as we step out of that chapter in the history of nationalism where it all began. Surely agency can be pursued in, say, the economic sphere, via a group’s ability to change its conditions through economic adversities or uncertainties by means of entrepreneurship, collective effort, or migration? Surely cooperation can be, in principle, just as significant a form of agency as protest? These possibilities do not mix with the word ‘subaltern’ and, therefore, are excluded from the subaltern studies narrative. In turn, that has made subaltern studies, and a large part of colonial Indian historiography, insular. The retreat from dialogue is the most acute where a dialogue is

Rethinking economic change in India

148

needed the most—economic history. Complaints about the insularity of the language employed by the later subaltern studies authors have been made before. The early subalterns continued a dialogue with the Marxists, for both shared a vision of the world polarized along repression lines. A much more decisive barrier came in place in respect of non-Marxist perspective in economic history and South Asian historiography in general. Why this closure? K.Sivaramakrishnan answers, because ‘political-economic approaches divested the subaltern of agency’ (Ludden, 2001). Behind the dismissal there lurks a trained incapacity to incorporate economics in stories of agency. Economics looks at agency of the economic subject in a rather more open and contingent fashion than the primitive tools of exploitation and repression in which colonial Indian historiography is still trapped. Subaltern studies, as Sivaramakrishnan reminds us, broke disciplinary boundaries between history and anthropology. But it is really the historians’ refusal to come to grips with the essential openness in economic interactions that drove postcolonial history into the warm embrace of post-modern anthropology. Admittedly, the end of a dialogue was not confined to the subaltern studies. In a way it was part of a general crisis in the social sciences. Until the 1980s, the belief that economic change and cultural change were interdependent was widely held in both economics and history and was the ground for a dialogue. Marxism and modernization theory supplied models of interaction between the economic and the cultural. With their decline, the belief in interdependence itself receded. Economists grappled with culture quite independently of any help from historians, who generally found economics too tied to a modernization language and therefore politically incorrect. Market-participation and agency The identity between non-elite and subordination fails in situations where both the elites and the non-elites stand to gain from transactions. In those cases, cooperation, co-option, communication and co-sharing the rules of the game are just as important axes of relationship as are repression and resistance. A clear message from economic history is that economic changes under colonialism enabled such transactions on a wide front. Take the example of the rural tanners-cum-labourers who migrated out of extremely one-sided rural relationships to work in tanneries that came up in Dharavi or Calcutta in the interwar period. For large numbers of such people the colonial economy provided escape routes from servitude and famine. Notwithstanding the fact that these moves could lead them to new inequalities, these moves represented ways in which the subaltern could attempt to change their own status by co-sharing the rules of the game with the elite. Subalternity, then, becomes a special case, a contingent situation. If there was no battle, there was no subaltern. Unless we have good grounds to believe that mutually profitable transactions were impossible, conditions of battle, whether these arose from peasant consciousness or economic structure, have to be seen in a more general context that also includes conditions of compliance, cooperation and participation. Incorporating economic history creates a fundamental discord within the paradigm of repression and resistance. Economics indeed did remain in the backdrop of a great deal of the early subaltern studies, concerned more specifically with the peasants. In these narratives, more or less all that economics did was to account for the initial conditions of

Labour and power

149

a polarity. Subaltern studies conceded to the ‘economistic’ Marxists a possibility that power could and usually did move in several fields simultaneously—economic, social, and political—and drew upon a variety of resources. But material conditions and distribution of resources played almost no role in the dynamics of elite-subaltern relationship. This single-minded and static concern with power would seem, from a modern economic historical point of view, archaic. A non-Marxian historiography is likely to suggest two characteristic dynamics in colonial Indian economy: one strengthening and the other weakening subordination of the non-elite by the elite. The former was growing imbalance between labour and capital as a result of population growth, poor effort at human capital formation and the exhaustion of land surplus. In the long run, real wages in agriculture did not increase much even as return to capital increased. There is a simple explanation to this in terms of demand and supply of resources. A sustained excess supply of unskilled labour relative to capital tended to depress the worker’s bargaining power. In a society already divided, unequal economic power could easily join with social inequities to turn into an all-encompassing subordination. The second process was market formation, which could empower and liberate some subaltern groups and individuals from non-market forms of subordination. It could do so by breaking the obligation upon the subalterns to supply product and labour to the local elite, thereby weakening some elite-subaltern relationships. It could happen via a simple exit from one set of relationships, such as migration from rural to urban areas, a theme surprisingly undeveloped in subaltern studies. It could happen by steps that did not reduce dependence in labour-capital relationships, and yet increased the non-elite’s choices by increasing their access to subsistence and occupational mobility. Most ruralurban and agriculture-industry shifts had such implications. Both these processes suggest that the possibility of a change in elite-subaltern relationship—either way—can derive from redistribution of economic opportunities, rather than via political repression or resistance to such repression. This is the lesson of economic history. The subaltern studies, however, are constrained to ignore it. The suggestion that markets can in principle be seen as liberating is blasphemy in the context of the reading of colonial modernity that the subaltern studies have taken from the poststructuralist master narrative. Not surprisingly, the only type of economic history or vision of long-term economic change that the subaltern studies can be somewhat comfortable with is the neo-dependency one that necessarily sees markets in the colonial setting as repressive tools. An example of this compatibility is David Ludden’s introduction to a recent collection of critical essays on subaltern studies (Ludden, 2001). Ludden’s own references to economic globalization predictably stop at the extreme form of opposition to globalization represented by Arturo Escobar. The critique of modernity that the subaltern studies bring into Indian historiography leaves no room for economics to play any role in the position of the non-elites. There are three conceptual difficulties. First, economics allows openness in the outcome of transactions, whereas subaltern studies imply that all roads of change are closed except active or passive protest. Second, economics cannot give up the language of modernity. The notion of economic modernization is deeply entrenched in economics, and derives fundamentally from the measurability of an improvement in well-being. Third, and perhaps most important, the theories of economic modernization increasingly look like a

Rethinking economic change in India

150

more open and less ethnocentric version of modernization theory, in so far as these believe in the necessity of cultural changes for economic growth to happen. Growthinducing changes do not necessarily involve harnessing power and domination, but sometimes turning hierarchies to cooperation while running organizations or managing resources. In that sense, economic modernization can involve greater freedom. To yield some ground to economic history amounts to admitting that there are indeed trajectories of progress and that societies are unequally placed along these trajectories in their social characteristics. Repression does not have the universality that it does in a Foucaultian world. Instead, it is one kind of condition, sometimes conducive to growth sometimes adverse. The colonial state and subalternity Subalternity sinks into fuzziness unless one can effectively polarize the elite and the nonelite, the rulers and the ruled. To deal with this problem some polarities are exaggerated in subaltern studies. Right from the beginning, the point was made that there were different types of repressive relationships in colonial India. One of them was along the state-subject axis. Outside that, there was a dense network of subordination among and between the so-called subalterns. Along the colonial-state-colonial-subject axis, subalternity is easier to define because these were theoretically distinct poles in the exercise of power. Inside the indigenous world of repression, subalternity cannot be either defined or located precisely, for every repressed person or group was also a repressor in turn. This endless spiral of repression shows up best inside the family. The mill worker, protagonist of many subaltern tracts, with a new-found economic security behind him, often insisted on early marriage of his daughter or on asserting tradition in other ways. Hospitals for childbirth, for example, were rarely resorted to in the jute mill labour colonies even as late as the interwar period. Infant and maternal mortality often touched far higher levels in the mill towns than in the rural areas where the workers came from. A great deal of the emphasis of subaltern studies, and especially that of Ranajit Guha’s own writings, falls upon the colonial state. Subaltern studies from its political history origin through its cultural studies reincarnation is primarily a critique of the colonial state. There are several problems with this emphasis on state-subject relation. First, it is arbitrary. Second, colonial and indigenous networks of politics were mutually interacting, and one cannot be described without the other. Third, it greatly narrows the scope of historiography. A great many poor people in South Asia were in remote contact with the state. Artisan communities, for example, were groups that the colonial state was never seriously interested in. Such people remain outside history whenever historiography gets too bogged down in what the British and their collaborators did or thought. Fourth, it smuggles in a bad faith. The notion that here was one big stem of oppression and all the others were its branches ignores the fact that oppression had indigenous and diverse roots, and colonialism suppressed many of these roots because it needed to establish its own authority. Many subaltern studies authors and their critics recognize the existence of indigenous repression. But nowhere is there a serious admission of the possibility that the effects of state power could be reducing the effects of other types of power. And yet, one dimension of the much-maligned ‘colonial modernity’ was the

Labour and power

151

attempt to empower some groups repressed in the indigenous networks of power. Tenancy and moneylender legislation are examples. Another dimension of colonial modernity was elite actions directed at empowerment of the non-elites, such as social reforms or women’s education. Subaltern studies are constrained to emphasize instead complicity between indigenous and colonial elites. A less loaded reading would see that colonial power stood in a relation of ambiguity with indigenous centres of power. For the same reason, colonial knowledge too had an ambiguous relationship with power and repression. The ambiguity of colonial knowledge The attempt to read ‘coloniality’ in colonial knowledge in the late subaltern studies is sometimes compromised by the need to rely almost wholly on colonial documents for the purpose. There is simply no other information structure of comparable detail, volume and observational intent that this vast body of knowledge can be weighed against. Of course colonial knowledge was biased. But how biased it was there is no way of measuring, because there is no standard available outside this body of knowledge. In some exceptional cases, subaltern studies contributors have made innovations in this regard. But, generally, beliefs about the degree of bias in colonial sources remain statements of faith. Before the subalterns, the American anthropologist Bernard Cohn was the most influential author to have written about the power-knowledge nexus in colonial India. ‘The knowledge’, Cohn wrote in a celebrated essay, ‘which this small group of British officials sought to control was to be the instrumentality through which they were to issue commands and collect ever-increasing amounts of information’ (Cohn, 1996:21). That information, in turn, reinforced the structure of colonial knowledge. He was referring particularly to the development of colonial knowledge about India in the eighteenth and nineteenth century, through which Indian knowledge and the Indians themselves ‘were to be converted into instruments of colonial rule’. The much larger scale administrative projects taken up later, such as the census and the ethnographic surveys containing perhaps a richer information resource on the life and work of ordinary Indians than had been seen until then, most subaltern historians would classify, following Cohn, to belong in the same programme. The remarkable spiral between ‘information’ and ‘knowledge’ Cohn depicted in the first citation above could leave the information theorist at a loss. Modern technical usage of both these terms keeps these two concepts strictly distinct, defines these differently, and in such a way as to be able to understand their interconnection. The classical communication theory, for example, defines information not with respect to its functions but to a state of uncertainty, information is reduction in uncertainty. The theory also implies that the one who possesses or controls information does not have full control over its communication to others because of the ‘noise’ that necessarily enters all communication channels. Knowledge differs from information in a number of ways, by an element of intentionality for example, or by the fact that knowledge is constructed through a process of recognition of information to be correct or useful (Lehrer, 2000). The upshot is that the two tasks—gathering information and creating knowledge—are far from identical, and the conversion of information into knowledge is a far more open-

Rethinking economic change in India

152

ended and contingent process than might appear from reading Cohn. How might that insight change our reading of colonial knowledge? Colonialism had to create an information base on the people of India before it could filter it into knowledge about colonial subjects. The two tasks were distinct. This was so precisely because pre-colonial and indigenous structures of knowledge supplied no worthwhile description of the work that engaged the ordinary people of India. Information was needed to reduce the uncertainty left behind by pre-colonial regimes. There was nothing there that was comparable with what colonial sources, including the fashionably discredited censuses or caste surveys, supplied. Colonial knowledge, before it could become in the language of Cohn fully ‘categorized, classified and bounded’ to function as a tool for control, had to come into existence as a raw purposeless mass of information. Now, this is a far more indeterminate project than the critiques of colonial knowledge make it out to be. First, systems of governance operate at different levels, principally, administration and ideology, and the fit between the two is not necessarily very close, as one might imagine from reading Cohn or the subaltern studies treatment of colonial sources. Administration in a colonial society took on a life and dynamics of its own. Indeed, some of the most valuable descriptions of life and livelihood of the poor in colonial India arose out of the autonomous uncoordinated research by officials whose connection with governance went no further than using the government press to publish their findings. A great deal of the work done for the census, the craft monographs, the caste surveys and the village studies belonged in such research efforts without plausible political roots. Second, the very process of creating information about the poor, if it represented a consolidation of power and hierarchy, was also equally a radical inversion of the profound indifference from which the pre-colonial elite forgot to document the experience of the underprivileged classes. Colonial knowledge, in its re-hierarchization aspect, was an expression of power. But colonial knowledge, in counteracting the silences left over from hierarchies that ruled in the past, was empowerment. Third, precisely because the project was both of these things, it never could serve up a finished ideological product, and never had the chance to become an ‘orientalist’ output. A part of colonial knowledge, especially that part created independent of explicit political command, consisted of unclassified fragments that had been created to meet the silences before any meaning could be invested in them. Fourth, these fragments of information without an orientalist content was ‘public good’ in the sense economists use the term. These were non-rival in the extent of usage, and non-excludable in terms of access. Such information could be used equally for producing an image of Western rule, for demanding reforms that empowered the poor, or for writing a critique of colonialism. All three uses merit equal weight. I am arguing in effect that the communication of colonial knowledge/information was susceptible to ‘noises’ of a political kind. Information could never have the desired power effect because of this factor. An example is the famine reports that documented, for the first time in India’s famine-ridden history with adequate detail, how the other half died during a severe scarcity. In turn, these produced a relief policy that is still in force, and continues to represent India’s modern egalitarian credentials. This knowledge was empowerment for the famine-prone who included some of the most oppressed peoples of India.

Labour and power

153

Conclusion I now sum up the main arguments. Generally, criticisms of the subaltern studies project have been written by insiders, that is, by historians who harbour no serious misgivings about the term ‘subaltern’. The critical argument developed is rather that there is an intellectual history of subalternity outside the subaltern studies and that the relationship between these two traditions tends to be undeveloped and overlooked within the latter. In this chapter, I develop an outsider’s point of view. Conceding that in the specific context of revolt and resistance, the term has made a contribution, I dispute its usefulness in a general history of the rural poor. I propose that the very term ‘subaltern’, by begging in a research programme bound up with polarity, coercion and resistance, has been limiting in the study of the underprivileged peoples in South Asia. This critique has two basic points. First, the term ‘subaltern’ refers to a vague set of people in a vague sort of way. If they are the poor, then subaltern studies fails in meeting the basic criterion for an adequate historiography of the poor, a good economic history. A large part of this scholarship is supposedly about conditions of labour, and yet it pays surprisingly little attention to wages, migration, contracts, negotiation, the household, weather, techniques, livelihood and a host of other elementary aspects constituting manual labour. If they are the victims of domination, then they are indeterminate, for every victim was in principle also an oppressor. A search for ways out of that problem has almost always led subaltern studies to select and privilege some axes of domination over others, and to overlook the ambiguities characterizing most relationships, between principals and agents, or between the state and the subjects. Second, ‘subalternity’ begs in a world-view centred on the idea that the non-elite must be studied via a repression-resistance narrative. This grand narrative, the indoctrinating power of which is evident in the trusting way the term is employed by an entire generation of scholars to refer to whoever it refers to, is limiting in two ways: it oversimplifies the experience of the poor in South Asian societies; and it deepens the rift with economic history, which admits of a greater openness in interactions. The way forward for history writing is to retain the original intent, the search for agency, while interpreting agency in a much broader way. That is, by limiting the use of the word subaltern and the repression fetish hidden inside it.

10 Conclusion The message of this book can be restated quickly. Over a century marked on one side by the 1870s when significant agrarian expansion began and on the other side the 1970s when a second green revolution emerged, economic change in India can be seen as a process wherein acute resource-constraints were temporarily overcome by reallocation and increasingly industrious labour. In this economy with persistent labour surplus, occupational structure changed rather slowly. Poverty continued because manual labour continued, and because the prospect of diminishing returns to manual labour was always present. That is not to say that conditions of labour did not change. A process of marketization of labour encouraged attempts to raise returns to labour through reallocation between places, occupations and organizations. Via these shifts, labour institutions changed, the two key examples of which were decline of household industry and rise of casual wage labour in agriculture. The whole transition was influenced by culturally specific selection of market participants and even market success. A historiographic preference lurks behind the story just outlined. Economic history of India was for a long time the site of a contest between two rival approaches. One of these, the academic mainstream in post-independence India, believed in a broad identity of interest between colonialism and the local elite, and held that alliance responsible for retardation of the economy. Persistence of manual labour and poverty illustrated the very retardation process. The other school believed that the production possibilities in colonial India were subject to weaknesses that it traced to society, demography or ecology. Manual labour and poverty were an outcome of this structure. This book, evidently, belongs in the second approach. Colonialism here plays no essential role as a driver of economic history. Neither the stable element nor the dynamic element in the labour experience responded significantly to 1947 when India attained freedom from colonial rule. This account of persistence of poverty amidst turbulence in the labour market is not a complete account of either labour history or economic change. There are many points of detail that remain undeveloped. Those points apart, this study underplays or ignores the post-colonial state when dealing with a time when state intervention was pervasive. A state-less narrative is certainly limited, it is at best a device to focus on what I find missing or underrated in the conventional historiography of labour. That said, one can go too far in bringing the state back in. A great deal of the recent attempts at appraisal of poverty and poor growth in post-colonial India, attempts made in the wake of an ongoing drive to liberalize the economy from government control, tend to be absorbed in ‘what the government did wrong’ and ‘what the government could do and didn’t’ types of discourse. I wanted to shift the ground somewhat. One other important theme that receives less than adequate attention in this work is collective action and collective bargaining. This issue does not figure adequately, partly

Conclusion

155

because there is very little material available directly on it. What I might say on either subject would be inevitably speculative. Besides, I do have an excuse for placing this issue outside the scope of the study. Collective bargaining has not played a sustained and significant role in shaping employment relations or labour markets in the informal sector in India. The history of these formal institutional actions is a history of largely unsuccessful attempts that petered out from the mid-1980s. On the other hand, it will be hard to sustain that local and informal attempts to reshape employer-employee relationships at the workplace sometimes using the agencies of the state did not occur in systematic ways. In some regions, such attempts even played prominent political and ideological roles in post-independence India. While not disputing their significance, I would defend the limits imposed on the study by stating the belief that reincorporating politics into this largely economic narrative, while a potentially fruitful project, is unlikely to change the main conclusions of the book. Essentially, this book underscores systemic features relatively more in explaining poverty. These features, in turn, tie colonial and post-colonial India into a single narrative of continuity and change. The story seems to offer some lessons for economic history of India in comparative setting. These lessons can be divided into two sets: one illustrates points of continuity between colonial and post-colonial India, and the other stresses the difference.

Continuity First, there is a lesson here about redrawing the boundary between the present and history. The border, implicitly in historical scholarship, is usually set at 1947 when India attained independence from colonial rule. Rarely do economic historians step beyond this date and into the recent past. Exceptions to this practice are beginning to appear in some fields of research, but the norm has not yet been threatened. This book, by undermining the significance of colonialism, makes a plea for including the post-colonial into the field of historical vision. Post-colonial India did not see a cessation of that growth process of which commercialization, relative resource abundance or scarcity, scramble for nonlabour resources, reallocation of labour, cultural preferences and exclusions, formed the core elements. These elements defined an underlying continuity between colonial and post-colonial India, transcending politics and regime change, even though the manner in which they affected individual lives and whom they affected changed constantly according to local situation, the resource in question and specific interventions. An illustration shows how the model here can dissolve the conventional dividing line between history and the present. Chapters 3 and 4 suggested the idea of a long interregnum in conditions of rural labour. Between 1930 and 1970, real wages in agriculture remained depressed, agricultural productivity remained a forgotten policy agenda and the land-labour imbalance turned steadily worse. This interregnum ended with political turmoil of varying intensity, in Eastern India particularly. In the nonagricultural sectors, a similar regime of sluggishness lasted somewhat longer, again threatening political repercussions at the end of the twentieth century. In both cases, the preconditions for a slackening of growth were building up from before 1947.

Rethinking economic change in India

156

Culture responded slowly to government policy, which was one dimension of continuity in the growth process. Culture enters this labour-centred history via several routes. First, persistence of traditional consumption helped labour-intensive industry survive competition from mass production. Second, culture influenced participation of women and the mode of participation in the labour market via the average age at marriage. At work-sites, women faced barriers to entry or congregated in tasks that were more likely to be displaced. In the view of this book, such vulnerability derived from differential endowments of training and knowledge, which took shape inside the family. Third, a relatively high population growth rate maintained for over eighty years is crucial to any narrative that takes land-labour balance seriously. Fertility response to changing mortality risks was exceptionally slow in South Asia. Persistence of norms about family and gender roles might explain why. Fourth, the pace of formal schooling in South Asia before and after 1947 was also exceptionally slow. For British India, the inertia can be blamed on a lack of government resources and the convenient policy of non-interference in matters of social development. For independent India, the inertia is not easily attributable to fiscal pressures. A residue of cultural agency may have remained. It is plausible that India’s deeply hierarchical society discouraged private spending on public welfare. Attitudes, preferences, mentalities, thus influenced what goods, services and labour were likely to form the economic mainstream, and what were likely to suffer exclusion. The journey from ‘tradition’ to ‘modernity’ in South Asia, as indeed anywhere else, did not see a uniform displacement of the old by the new, nor did it necessarily mean an increase in welfare. Rather, it saw some form of tradition (handicrafts, for example) continue, and some others (women’s industrial work) abandoned.

Difference A clearer sense of continuity brings difference into sharper focus. Colonial and postcolonial India differed crucially in respect of the influence of market forces, especially the world market, on the use of traditional resources. By contrast with colonial India, the post-colonial regime embarked on a strategy that curbed competition and introduced isolation from the world market. It was during the 1960s and the 1970s that an economy with inefficiencies and unutilized labour slid further back in comparison with other successful developing economies. Incentive structures were seriously distorted by means of socialistic policies, and efficiency-enhancing institutions were poorly provided for. In respect of a deliberate distortion of market-based instruments for promoting efficiency and incentives, the post-colonial regime was distinctly regressive by comparison with the colonial. In popular discourse, South Asia’s rather dismal record on poverty is intuitively attributed to the legacy of colonialism. A significant, perhaps larger, contribution was made by the post-colonial-induced economic retardation. A simple proof of this statement can be found in the difference in per capita industrial growth rates between South Asia and East Asia for thirty normal years before and thirty normal years after World War II. The differential was larger in the second period. Over the twentieth century, as population and labour force have grown, competition for resources has extended beyond agricultural land and into other types of natural

Conclusion

157

resources and common property. Environmental pressures, in other words, bring in a specifically contemporary dimension to the story of competition for resources. Already in the colonial period, controversies about overexploitation of pasture and cultivable wastes had emerged. Water too was always potentially controversial. Over-harvesting of water was alleged in some older writings. Reckless water harvesting has been the foundation to the second-generation green revolution in India that occurred in the final decade of the twentieth century. With most parts of India now on the brink of a serious crisis of groundwater, this centrality of a traditional resource is increasingly in evidence. The question of common property resources has been addressed in a certain fashion in the environmental history scholarship on South Asia. In line with old economic history, it has tended to place somewhat greater emphasis on the role of the colonial state in causing disequilibrium in resource use, at some neglect of market-driven changes in land-use pattern. The question of how markets, civil society institutions and the state interacted in the clamour for common property resources has also remained under-researched. With resources that tended to fall in absolute scarcity in post-colonial India, such as pasture or water, endogenous response to scarcity by capturing, rent-seeking, free-riding or conservation seem to have been joined by state-mediated actions working in the same direction or the opposite direction. The nature of the state made a difference to the outcome of such games. In matters of conservation at least, arguably the move from colonialism to a democratic framework did not necessarily produce socially more desirable outcomes. Post-colonial India represented a veritable leap forward when compared with the colonial period in respect of rates of investment. In respect of foreign savings, government savings and private savings, India was a remarkable success story. Other regions of South Asia were less remarkable in this respect, but a distinct improvement from the colonial record nonetheless. A capital-intensive induced industrialization was enabled by this success in raising resources. On the other hand, with market-based incentives and efficiency weaker than before, the conversion of investment into output involved serious inefficiency, wastage and loss of potential output. The government in post-colonial India tried, much harder than its predecessor, to expand the base for training in vocational skills. That effort paid off in broadening the social base of technical education and was a positive factor in the re-integration of India in the world market for services at the beginning of the twenty-first century. From 1971, women’s presence in the industrial workforce began to increase. Behind the reversal lie hidden remarkable changes in the social norm, painfully slow but powerful in total impact. Consistent with my hypothesis about the earlier phase of a decline, the average age at marriage of women had crossed fifteen in the previous decade, significantly raising the work-participation in the 15–19 age-group. Culture does change. And it is more than likely that the residue of attitudes that made for exclusions of certain groups from the market-place will dissolve away more rapidly in the near future as India reintegrates with the world economy.

Notes 1 Introduction 1 These effects follow from the assumption made in a ‘specific factors’ model that labour is mobile between agriculture and industry, but land is not. The most well-known recent application of this model in economic history scholarship is O’Rourke and Williamson (1999). 2 On rent rates in the nineteenth century, see Kumar (1983a), pp. 57–8, 67, 133–4, 237. 3 The range in the latter was 1.5–2.5, see Hayami and Ruttan (1971:27). 4 On real wages see Sivasubramonian (2000: ch. 4) for 1930–46. On rents, see Guha (1993). 5 On the significance of different types of innovation in agriculture, see Hayami and Ruttan (1971: ch. 3). Application of the induced innovation model to agrarian change largely owes to this book. For historically sensitive surveys and restatements, see Hayami (1997: ch. 1) and Ruttan (2001). See also Boserup (1965, 1981), and Hansen and Prescott (1998). 6 The countries of South Asia had somewhat divergent experiences in this regard since the end of colonialism. Land scarcity was acute in the former East Pakistan or present-day Bangladesh, and in large parts of India too. But in the former West Pakistan, or present Pakistan, land-using growth continued in the 1960s and the 1980s. To some extent at least, the significantly higher growth rates in Pakistan relative to the rest of South Asia in these decades derived from the existence of unutilized land and water. 7 Morris (1960) explained the strength of this institution from the labourers’ point of view. The workforce was linguistically diverse and, therefore, communication with the managers, right from the recruitment stage, involved an intermediary. This factor was clearly present, but does not explain why the institution weakened from late in the interwar period. 8 India (1928–31), evidence vol. III, part I, Central Provinces and United Provinces, pp. 8,144. 9 For a review, see Chandra (1966: ch. 1). 10 For general statements, see Bagchi (1982), Chandra (1992), Habib (1985), Sarkar (1983). 11 Colonial power does not play an essential role in this story because there was no fundamental link between politics on the one hand, and resources or the market for manual labour on the other. This is also the reason why the archives containing official transactions remain stubbornly uncommunicative on conditions of work involving manual labour, unskilled labour, artisans and women. 12 Major works which draw inspiration from subaltern studies include Chakrabarty (1989) and, in a more critical spirit, Chandavarkar (1994). A representative but not exhaustive list of recent publications in new labour history might include Nair (1998), Sen (1999), Simeon (1995), Joshi (2003), Basu (2004) and several contributions in Parry et al. (1999). See Mohapatra (2000) for a useful survey of the ‘renewal’ of the field. 13 Shortly before Anstey, the American journalist Katherine Mayo wrote a book, Mother India, on the treatment of women in India (Mayo, 1927). This largely anecdotal work dealt with issues such as low age at marriage, illiteracy and discrimination of the girl-child by parents, all real problems that characterized the majority of Indian families. Mayo, however, spoilt the potential of her own book by what she called her ‘argument’, that moral debility was a trait ‘that truly characterized the Indian of long-past history’. The implicit imperialist claim that Indians were unfit to govern themselves immediately provoked a large number of books

Notes

159

with titles such as Father India, Happy India and Unhappy India. M.K.Gandhi called the book ‘the drain-inspector’s report’, a label that became more famous than the book itself. Gandhi’s choice of words was unfortunate. Post-colonial India treated the issue raised by Mayo, neglect of the girl-child, with about the same concern that it showed to draininspectors’ reports. 14 There are many contributions around the Weberian conjecture. See, particularly, Morris (1967). See Rutten (2002) for a recent survey of the field. 15 Another old idea, and a rather curious one, enjoyed a remarkably long life. Radhakamal Mukherjee had attributed Indian poverty, inter alia, to an excess of cattle in the villages, which were not allowed to be slaughtered on religious ground, but were too weak to work, and yet competed with the humans for food. In the 1970s, the ‘sacred cow’ controversy was revived. Participants included leading figures in contemporary agricultural economics scholarship such as K.N.Raj, V.M.Dandekar, A.M.Khusro and C.H.Hanumantha Rao. 16 A useful review of Dandekar’s work is Deshpande (1996). Dandekar and Rath (1971) outlined the policy framework that followed from the general approach.

2 Economic history and modern India: redefining the link 1 Among recent works, Parthasarathi (2001) generalizes the point to artisans. 2 See discussion of Moreland in Das Gupta (2001:26). The same estimates of trade volume appear in Moreland (1920:221). 3 The citation is from Stokes (1994). 4 For general statements of the orthodox position covering some or all of the five theses, see Chandra (1968, 1992), Habib (1975, 1985), Sarkar (1983), Bagchi (1982), Patnaik (1983). 5 See Hayami and Ruttan (1971: ch. 3).

3 Rural labour and colonialism 1 Studies, old or new, directly bearing on the labourer have been scarce even within the discipline ‘agricultural economics’ that became an established field in Indian academia in about 1930. However, some scholars were more interested than others about the labourers and the labour questions. The following description draws on three such roughly contemporaneous observations: written evidence of Radhakamal Mukherjee in India (1926– 28), vol. VII (evidence taken in the United Provinces), written evidence of Sir George Paddison, Commissioner of Labour, Madras, in India (1926–28), vol. III (evidence taken in Madras), and Darling (1934: ch. XIV). In post-independence scholarship, research on rural labour was stimulated particularly by the writings of Daniel and Alice Thorner—see, for example, Thorner and Thorner (1957). In recent work in agrarian history, M.Atchi Reddy, Neeladri Bhattacharya and Haruka Yanagisawa have focused particularly on the labourer. 2 Paddison before India (1926–28:315). 3 For example, India (1926–28), vol. VI:5, evidence of FJ.Plymen, Director of Agriculture, Central Provinces and Berar. 4 For example, Sanghvi (1969:51). 5 A notable exception is Atchi Reddy (1983b) who used a number of these documents collected from coastal Andhra Pradesh.

Notes

160

6 Any worker under contract is attached to an employer in the sense that leaving the contract while it is in force involves penalty. However, Indian scholars have frequently used the term ‘attachment’ to refer to an extra degree of unfreedom. In this implicitly coercive sense of the term, a worker is attached in not being able to choose an employer at entry, or not being able to leave the old employer at end of contract. See Basant (1983) for a discussion of the term ‘attached’ labour from these perspectives. 7 The original statement of the thesis can be found in Patel (1952). See also Patel’s essay reprinted in Prakash (1992b). The Marxist ‘mode of production’ debate in the 1970s, carried out largely in the pages of the Economic and Political Weekly, more or less absorbed the immiseration logic, and saw ‘‘free’ labour, and some forms of ‘unfree’ labour as well, as new relationships in colonial India. But it stressed rather more forms of labour and polarization within the peasantry, and related all of this to the development of capitalistic relations. See contributions on ‘colonial mode of production’ in Patnaik (1990). By contrast, a few regional agrarian histories and some of the field-surveys by agricultural economists in the interwar period stressed relative resource positions more than production relations in explaining changing conditions of labour. See Kumar (1983b), Atchi Reddy (1996). The Marxist debate was superseded by the ‘subaltern studies’ in the 1980s (see Chapter 9, this volume), in which the labourer figured predominantly in a political role. 8 For general discussions on the thesis, its implications, and its tests, see ‘Introduction’ in Kumar (1992), ‘Introduction’ in Prakash (1992b), Breman (1974), ‘Introduction’ in Yanagisawa (1996a). 9 Krishnamurty (1972,1983a), Thorner, Alice (1962), Thorner, Daniel (1962). 10 ‘Introduction’ in Prakash (1992a). 11 Jha (1997:26) cites C.R.Reddy’s test of the relationship. In other works, density has been mentioned as a factor, but rather casually, and the relationship not been explored fully. 12 Dealing with large regional units is always hazardous, however. For example, ‘coastal Andhra’ included land-abundant agrarian zones as well as district, especially in the north, that had faced population pressure and outmigration since the nineteenth century. On the land-labour situation in this region, see Rao and Rajasekhar (1991). 13 It may be for this reason that low land-persons ratio regions often turn out to have relatively high inequality of occupational holdings. See India (1954). 14 Mukherjee at India (1926–28:392). 15 India (1928–31), vol. III, part I, p. 138. 16 Rath (1983), Krishnamurty (1983b). 17 Mukherjee before India (1926–28:380–1, 390). 18 These instances come from Thurston (1909), vol. 2, pp. 185, 332, 398, vol. 4, p. 325. 19 Apropos menial service and the Chamars, see Ibbetson (1916:300); and on Madigas in a similar context, Thurston (1909), vol. III, p. 325. See also, Darling (1934:272) on the decline of ‘begar’, which he attributed to four causes: rise in wages and prices after the war, urbanization (both factors raised the opportunity cost of begar), increase in population (which presumably increased the landlord’s choice of labour, and spread of education and social emancipation. On the decline of ‘begar’ again, see Mukherjee (1930–31:325). 20 See, for example, Kumar (1992). This further division between ‘serfs’ and ‘free’ labourer has been noted in the literature. Patel (1952) and Joshi (1958) examine whether or not labourratio was higher for provinces that had a larger percentage of ‘depressed classes’ among the population. They test the hypothesis that labour-ratio was higher in regions that had a relatively large number of persons whose traditional socially sanctioned occupation was labour alone. The official term ‘depressed’ classes is not really adequate for the test. There was great variation within this term according to occupation and the strength of the casteoccupation correspondence. A more precise test can use degrees of territorial bondage or the link between caste and rights, as in Kumar (1965). A similar exercise has not been done for other areas of India, but agrestic servitude was by all accounts much weaker in northern

Notes

161

India and the Bengal delta. There is, in other words, qualitative support for the thesis that the labour-ratio and the occurrence of castes that were specialized as labourers were, in fact, correlated. 21 The estimated equation (where y=labour-ratio and x=1/density) is

y=151.33−344.39x+279.50x2−63.63x3

with r2=0.30, and F=3.51, which is significant at the 5 per cent level. 22 The estimated equation (where y=labour-ratio and x=1/density) is

y=137.47−208.17x+89.57x2

with r2=0.37, and F=2.03, which is not significant.

4 Agricultural labour: lessons from wage data 1 See, for a review, Bardhan (1977). Other important contributions on post-independence trends in real wages include Mukherjee (1974) and Jose (1988). 2 Thus agricultural wages are not a central dataset in Sivasubramonian’s national income estimates, Sivasubramonian (2000). 3 Cited by Radhakamal Mukherjee before India (1926–28:392), Evidence taken in the United Provinces. 4 See, for a similar argument relating wage trends with decline of farm servant arrangements, Atchi Reddy (1986, 1991). On decline of product wages and long-term arrangements generally, see Radhakamal Mukherjee in India (1926–28:390, 397), evidence taken in the United Provinces. Keatinge (1921:141) placed the decline of long-term contracts called saldari in west-central India well before 1921: ‘The tendency nowadays is for the labourers to prefer to work for daily wages’. Kessinger (1974:123–4) placed the decline of sepidari in Punjab from about the early 1920s, but this was based on recall. G.C.Mukhtyar (1930) wrote of the decline of the hali system of Gujarat. And Darling (1934:272) mentioned a ‘a new wind’ in the 1920s Punjab whereby customary payment obligations tended to be discarded and questioned even in the most traditionalist settings. 5 Real wages in Bengal increased in the pre-1914 decade, but stagnated thereafter until 1922 (Mukherjee in India, 1926–28:393). Although official real wages in Bombay remained steady for the first twenty years of the twentieth century, ‘the improvement in the position of the labourer is greater than the figures appear to indicate’ (Keatinge, 1921:143). G.Findlay Shirras (1924) reported a significant rise in real wage in Bombay between 1900 and 1922. H.W.Lyons reported a rise in wages in the Narmada Valley in 1902–20, (Lyons, 1920– 22:450). On rise in wages in the early twentieth century see also Charlesworth (1985:218– 19); M.McAlpin quoted in Charlesworth (1985:224); Bhattacharya (1981); Fukazawa (1983:207); Kumar (1983b:239); Atchi Reddy (1986). 6 Slater, ed. (1918:33). The wages of ‘padials’ or debt servants were also uniformly and significantly lower than casual labour wages. 7 For some examples from Malabar, see Slater (1918:145,193). 8 Mukherjee (1995), India (1954), Bardhan (1977), Jose (1988). 9 See also Islam (1978) for a similar point on Bengal. 10 For the interwar period, the only complete series we have is the Bombay one. For Bombay, the correlation between money wage and acreage does not work, at least in the short run. The absence of a relationship might mean that, while money wage increased when there was

Notes

162

increase in acreage per person (India in pre-war period), money wage did not fall when there was decline in acreage per person (Bombay in the interwar period). The rate of change in money wages might still be influenced by acreage in the long run. 11 Why money wages were stagnant in the short run is a promising research question, which calls for more attention to information problems characterizing rural labour markets. 12 India (1928–31), Chapter XII for these and the following figures. 13 There are other problems with the wage-inequality relationship. In the above correlation, if two extreme cases of low inequality are removed from the dataset, Punjab and Assam, the correlation disappears. Excluding these two outliers, concentration varies in a much narrower range than do wage rates. Further, inequality might favour workers. More inequality in land-holdings could raise demand for labour because the larger farms cannot be worked with family labour any more.

5 Was there an industrial decline in India in the early nineteenth century? 1 Hymer and Resnick (1969). Eswaran and Kotwal (1994), explaining poverty in India, uses industrial decline as the critical marker between colonial and post-colonial India. 2 Influential works in development that use the Indian example to illustrate such a point include Robinson (1979:104), Frank (1975:22), Baran (1978:277–83). 3 The idea of an agrarian depression sometime in the second quarter of the nineteenth century has been articulated mainly in regional histories on the Bombay-Deccan and Madras. See Fukazawa (1983) and Thomas and Natarajan (1936). For a recent discussion of the theme, see Washbrook (1999). There are qualitative grounds to expect depression conditions to exist in some parts of India because of the regime’s tax policy. However, more or less the only hard evidence cited in support is the agricultural price index. Such evidence can be misleading, for agricultural prices could be adjusting for fall in manufactured goods prices in the same period. 4 Bagchi (1978) and Twomey (1983). See also, for critiques of Bagchi, Orr (1980), Vicziany (1979). 5 This is evident from Harnetty (1991), Guha (1989), Specker, (1989).

6 Labour-intensive industrialization 1 The representative works are Bagchi (1970), and Morris (1983). 2 A section of the colonial bureaucracy in the first half of the twentieth century held a different view of Indian industrialization, one that was strikingly different from both the Nehruvian heavy industry model and the Gandhian corrective imposed upon it. This view held that traditional industry on a large scale continued in India because of specific features of consumption and unique strengths of labour-intensive technologies. The handlooms, for example, survived largely because of continued demand for the sari, which required productdifferentiation and certain types of design that handlooms were better able to supply. This semi-official discourse proposed steps to strengthen the productive capacity of labourintensive industry. Alfred Chatterton, an influential writer and bureaucrat who worked in South India, was a key figure in this movement (see Roy, 2002a). In the clamour for rapid industrialization after 1947, this gradualist model was submerged.

Notes

163

3 A selection of recent works that represent the change of perspective would include Roy (1999), Haynes (2001). 4 If Y stands for income and N for employment, then, p=Y/N, pa=Ya/Na, pi=Yi/Ni, ps=Ys/Ns, na=Na/N, ni=Ni/N, ns=Ns/N, ya=Ya/Y, yi=Yi/Y, ys=Ys/Y. By definition, pana+pini+psns=p. The terms (pana/p) etc., are the contributions of each sector to increase in average productivity in the economy. A similar relationship can be derived between these measures and incomeshares. 5 Cited by Chatterjee (1907). 6 On the history of the powerloom industry, see Haynes (2001), Roy (1999), Roy (1998). 7 A caution, the denominator is generally believed to understate participation rates.

7 Women and industrialization 1 ‘The cross-sectional mean levels of female activity rates form a U-shaped pattern in relation to levels of economic development’, see Durand (1975:131). The U-shape was first observed by J.N.Sinha in a 1965 paper presented at the United Nations Population Conference. For time-series data capturing more of the rising phase, see Leser (1958). Among recent research on the U-graph, Mathur (1994) needs to be mentioned. 2 Horton (1996). The participation rates in this dataset use different age cut-offs, India’s being the most liberal (5+). See also Liu (1984). 3 Krishnamurty (1983a), Thorner (1962). Thorner (1984), in a later piece, admitted that the practice of ‘sweeping the dubious female figures under the rug’ was too drastic a solution to the problem. 4 Visaria (1997). 5 The important empirical works in this class are Ambannavar (1975), Gadgil (1965), Thorner (1962). See also the survey article by Krishnaraj (1985). 6 See, for example, Mincer (1962) and Goldin (1995). 7 ‘Wherever rice-mills are being established, some of the women of lower castes who used to get a living from rice-husking find employment’, Banerjea (1933). 8 One example being the ‘boarding-house’ mills of New England, see Morris (1960). 9 Based on interviews of numerous women workers before the Royal Commission (India, 1928– 31). A few were dismissed from their job at their sardars’ instigation after they appeared before the Commission. 10 Evidence of Labour Union, Tuticorin Mill, India (1928–31:89). 11 This figure is the difference between manufacturing workers in 1961 multiplied by the 1961 proportion of women in manufacturing, and manufacturing workers in 1961 multiplied by the 1911 proportion of women in manufacturing. These many persons ought to have been in the workforce but for the decline in women-worker ratio in industry 12 Dandekar (1974), Goyal (1975), Agarwala (1962); Japanese National Commission for UNESCO (1962), Taeuber (1958:214), Hajnal (1965), Crafts (1978), Dixon (1971). 13 For references and discussion, see Krishnaraj (1985).

Notes

164

8 Women in the crafts 1 Circa 1925, Ranga (1930:69) described what was perhaps the general situation among ‘caste weavers’ until then: ‘About 10 per cent of the women of Karnool are able to weave, but it is considered derogatory for a Padma Sale woman to weave.’ 2 A recent informative survey of women weavers in South India is Parikh et al. (1991). 3 India (1942:295). A 1937 report on Bengal estimated that in this system of warping, an average worker walked 2,400 miles per year (Bengal, 1937:39). 4 For another description of the apprenticeship of boys, see Ranga (1930:113). Boys of ten were sent (near Dharmavaram) to other weavers, they were paid Rs 1/month. They handled the dobby for the first year, prepare warp in the second, and were taught weaving from the third, when their wage increased to Rs 2.5/month. At the age of seventeen or eighteen, they tried to set up their own shops. In the Salem area, the period ranged from twelve to twenty. They worked for ten hours a day, and ‘even the six to eight hours sleep that [they] get is disturbed by the tip tap noise of the blessed loom’. 5 The craft monographs prepared around 1900 supply some instances of homemade articles. One example comes from Fanny Parks’ accounts of the customs of the North Indian home, such as the visit c. 1840 to Hyatulnisa Begum, a princess of Delhi: ‘Four trays, filled with fruit and sweetmeats were presented to me; two necklaces of jasmine flowers, fresh gathered, and strung with tinsel, were put round my neck; and the princess gave me a little embroidered bag filled with spices. It is one of the amusements of the young girls in a zenana to embroider little bags, which they do very beautifully’ (Parks, 1975:215). 6 One observer of textile industry in Eastern UP wrote: ‘Only the male members go to the industrial centres… The females remain in the village, The ties of home…make them unwilling to migrate permanently. Besides, the cost of living in industrial towns is very high’ (Chaturvedi, 1947). 7 Such as the two women friends of Appanna, a Devanga master-weaver from Vizianagaram. The women resided in Ichhapuram and Rajahmundry and were known for being ‘able to force up the sales far better than his male agents’ (Ranga, 1930:53).

9 Labour and power: a critique of ‘subaltern studies’ 1 For two recent collections, each with a comprehensive introduction and an exhaustive bibliography, see Chaturvedi (2000) and Ludden (2001). 2 A sufficient example is Amartya Sen, see Sen (1997).

References Abbreviations used: IESHR:

Indian Economic and Social History Review

MAS:

Modern Asian Studies

EPW:

Economic and Political Weekly

IJE:

Indian Journal of Economics

IJLE:

Indian Journal of Labour Economics

Agarwala, S.N. (1962) Age at Marriage in India, Allahabad: Kitabistan. Ambannavar, J. (1975) ‘Changes in economic activity of males and females in India’, Demography India, 4:345–64. Anstey, V. (1929), Economic Development of lndia, London and New York: Longmans, Green. Atchi Reddy, M. (1983a) ‘Female agricultural labourers of Nellore, 1881–1981’, IESHR, 20:67–79. ——(1983b) ‘Labour relations in Andhra Pradesh agriculture 1881–1981’, IJLE, 3: 160–87. ——(1986) ‘Trends in agricultural wages in some south Indian districts: 1800–1980’, IJLE, 28:307–49. ——(1991) ‘Work and leisure: daily working hours of agricultural labourers, Nellore District, 1860–1989’, IESHR, 28:73–96. ——(1996) Lands and Tenants in South India: A Study of Nellore District, 1850–1990, Delhi: Oxford University Press. Atkinson, F.J. (1902) ‘A statistical review of income and wealth in British India’, Journal of the Royal Statistical Society, 65:209–83. Bagchi, A.K. (1970) Private Investment in India, 1900–1939, Cambridge: Cambridge University Press. ——(1978) ‘Deindustrialization in India in the nineteenth century: some theoretical implications’, Journal of Development Studies, 22:135–64. ——(1982) The Political Economy of Underdevelopment, Cambridge: Cambridge University Press. ——(2003) ‘The great depression (1873–96) and the third world: with special reference to India’, in G.Balachandran (ed.) India and the World Economy 1850–1950, Delhi: Oxford University Press. 19:61–78. Banerjea, D. (1933) ‘Decaying industries of Bengal’, Mysore Economic Journal, 19:61–78. Banerjei, N.N. (1899) Monograph on the Woollen Fabrics of Bengal, Calcutta: Government Press. Baran, P.A. (1978) The Political Economy of Growth, Harmondsworth: Penguin. Bardhan, K. (1977) ‘Rural employment, wages and labour markets in India: a survey of research’, EPW, 12:A34–A48, 1062–74, 1101–8. Basant, R. (1983) ‘Attached labour and the classification of agricultural labour: some issues’, IJLE, 26:129–45. Basu, S. (2004) Does Class Matter? Colonial Capital and Workers’ Resistance in Bengal, 1890– 1937, Delhi: Oxford University Press. Bengal (1937) The Report on the Survey of Handloom Weaving Industry in Bengal, Calcutta: Government Press. Bhagat, M.G. (1935) ‘The untouchable classes of Maharashtra’, Journal of the University of Bombay, 4:130–74 .

References

166

Bhattacharya, N. (1981) ‘Agricultural labour and production: central and south-east Punjab, 1870– 1940’, in K.N.Raj et al. (eds) Commercialization of Indian Agriculture, Delhi: Oxford University Press. Bhatty, E.C. (1929) ‘The economic survey of an Indian village’, IJE, 10:81–99. Blyn, G. (1966) Agricultural Trends in India, 1891–1947: Output, Availability, and Productivity, Philadelphia: University of Pennsylvania Press. Bose, S. (1986) Agrarian Bengal, Cambridge: Cambridge University Press. Boserup, E. (1965) The Conditions of Agricultural Growth, London: Allen and Unwin. ——(1970) Woman’s Role in Economic Development, London: Allen and Unwin. ——(1981) Population and Technological Change: A Study in Long-term Trends, Chicago: University of Chicago Press. Breman, J. (1974) Patronage and Exploitation: Changing Agrarian Relations in South Gujarat, India, Berkeley and Los Angeles: University of California Press. Brendon, B.A. (1899) A Monograph on the Woollen Fabrics of the Bombay Presidency, Bombay: Government Press. Burns, W. (ed.) (1941) Sons of the Soil: Studies of the Indian Cultivator, Delhi: Government Press. Calvert, H. (1918–19) ‘The price of land in the Punjab’, IJE, 3:389–99. Chakrabarty, D. (1989) Rethinking Working-Class History: Bengal 1890–1940, Princeton: Princeton University Press. Chandavarkar, R. (1994) The Origins of Industrial Capitalism in India: Business Strategies and the Working Classes in Bombay, 1900–1940, Cambridge: Cambridge University Press. Chandra, B. (1966) The Rise and Growth of Economic Nationalism in India, Delhi: People’s Publishing House. ——(1968) ‘Reinterpretation of nineteenth century Indian economic history’, IESHR, 5:35–76 . ——(1992) ‘The colonial legacy’, in Bimal Jalan (ed.) The Indian Economy, Delhi: Viking. Chandra, R.N.L. (1903) A Monograph on Leather Manufactures in Bengal, Calcutta: Government Press. Charlesworth, N. (1985) Peasants and Imperial Rule: Agriculture and Agrarian Society in the Bombay Presidency, 1850–1935, Cambridge: Cambridge University Press. Chatterjee, A.C. (1907) Notes on Industries of the United Provinces. Allahabad: Government Press. Chattopadhyay, R. (1975) ‘De-industrialisation in India reconsidered’, EPW, 10: 523–31. Chaturvedi, S.C. (1947) Rural Wages in the United Provinces, Allahabad: Government Press. Chaturvedi, V. (ed.) (2000) Mapping Subaltern Studies and the Postcolonial, London: Verso. Chaudhuri, B. (1983) ‘Agrarian relations: eastern India’, in D.Kumar (ed.) The Cambridge Economic History of India, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. Choksey, R.D. (1945) Economic History of Bombay-Deccan and Karnatak (1818–1868), Poona: Gokhale Institute of Politics and Economics. Cohn, B. (1999) Colonialism and its Forms of Knowledge, Delhi: Oxford University Press. Collins, W.J. (1999) ‘Labor mobility, market integration, and wage convergence in late 19th century India’, Explorations in Economic History, 36:246–77. Crafts, N.F.R. (1978) ‘Average age at first marriage for women in mid-nineteenth century England and Wales’, Population Studies, 32:21–6. Dandekar, Kumudini (1974) ‘Age at marriage of women’, EPW, 9:867–74. Dandekar, V.M. (1966) ‘Transforming traditional agriculture: a critique of Professor Schultz’, EPW, 1:25–36. Dandekar, V.M. and N.Rath (1971) Poverty in India, Poona: Gokhale Institute of Politics and Economics. Dantwala, M.L. and Donde, W.B. (1949) ‘The uneconomic cultivator’, Indian Journal of Agricultural Economics, 4:9–47. Darling, M.L. (1920–22) ‘Prosperity and debt in the Punjab’, IJE, 3:145–66. ——(1934) Wisdom and Waste in the Punjab Village, London: Oxford University Press.

References

167

Das Gupta, A. (2001) The World of the Indian Ocean Merchant 1500–1800. Delhi: Oxford University Press. Deshpande, S.H. (1996) ‘Professor Dandekar on Indian agriculture’, Journal of Indian School of Political Economy, 8:577–91. Dixon, R.B. (1971) ‘Explaining cross-cultural variations in age at marriage and proportion never marrying’, Population Studies, 25:215–34. Durand, J. (1975) The Labor Force in Economic Development, Princeton: Princeton University Press. Dutt, R. (1906) The Economic History of India in the Victorian Age, London: Kegan Paul. Edwardes, S.M. (1900) A Monograph upon the Silk Fabrics of the Bombay Presidency, Bombay: Government Press. Eswaran, M. and Kotwal, A. (1994) Why Poverty Persists in India, Delhi: Oxford University Press. Fawcett, C.G.H. (1896) A Monograph on Dyes and Dyeing in the Bombay Presidency, Bombay: Government Press. Findlay Shirras, G. (1924) Report on an Enquiry into the Agricultural Wages in the Bombay Presidency, Bombay: Government Press. Foucault, M. (1994) The Archaeology of Knowledge, transl. A.M.Sheridan Smith, London: Routledge. Frank, A.G. (1975) On Capitalist Underdevelopment, Bombay: Oxford University Press. ——(1998) ReOrient: Global Economy in the Asian Age, Berkeley and Los Angeles: University of California Press. Fukazawa, H. (1983) ‘Agrarian relations: western India’, in D.Kumar (ed.) The Cambridge Economic History of India, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. Gadgil, D.R. (1965) Women in the Working Force in India, Bombay: Asia. Ghosal, H.R. (1966) Economic Transition in the Bengal Presidency (1793–1833), Calcutta: Firma K.L.Mukhopadhyay. Ghosh, K.K. (1969) Agricultural Labourers in India, Calcutta: Indian Publications. Goldin, C. (1995) ‘The U-shaped female labor force function in economic development and economic history’, in T.P.Schultz (ed.) Investment in Women’s Human Capital and Economic Development, Chicago: University of Chicago Press. Gough, K. (1981) Rural Society in Southeast India, Cambridge: Cambridge University Press. Goyal, R.P. (1975) ‘Shifts in age at marriage between 1961 and 1971’, Demography India, 4:336– 44. Grantham, G. (1994) ‘Economic history and the history of the labour markets’, in G.Grantham and M.MacKinnon (eds) Labour Market Evolution: The Economic History of Market Integration, Wage Flexibility and the Employment Relation, London and New York: Routledge. Guha, R. (1982) ‘On some aspects of the historiography of colonial India’, in R.Guha (ed.), Subaltern Studies I. Writings on South Asian History and Society, Delhi: Oxford University Press. Guha, S. (1985) The Agrarian Economy of the Bombay-Deccan 1818–1941, Delhi: Oxford University Press. ——(1989) ‘The handloom industry of central India: 1825–1950’’, IESHR, 26: 297–318. ——(1993) ‘Agricultural rents in India c. 1900–1960’, in M.Hasan and N.Gupta (eds) India’s Colonial Encounter: Essays in Memory of Eric Stokes, Delhi: Manohar. Habib, I. (1975) ‘Colonialization of the Indian economy, 1757–1900’, Social Scientist, 3:20–53. ——(1985) ‘Studying a colonial economy without perceiving colonialism’, MAS, 19:355–81. Hall-Matthews, D. (2002) ‘Famine process and famine policy: a case study of Ahmednagar District, Bombay Presidency, India, 1870–84’, unpublished thesis, Oxford University. Hajnal, J. (1965) ‘European marriage patterns in perspective’ in D.V.Glass and D.E.C.Eversley (eds) Population in History, Chicago: Aldine. Hansen, G.D. and E.C.Prescott (1998) ‘Malthus to Solow’, NBER Working Paper 6858, Cambridge MA.

References

168

Harnetty, P. (1991) ‘Deindustrialization revisited: the handloom weavers of the Central Provinces of India’ MAS, 25:455–510. Harris, H.T. (1908) A Monograph on the Carpet Weaving Industry of Southern India, Madras: Government Press. Hayami, Y. (1997) Development Economics, Oxford: Clarendon Press. Hayami, Y. and Ruttan, V.W. (1971) Agriculture in Economic Development, Baltimore: University of Maryland Press. Haynes, D. (2001) ‘Artisan cloth-producers and the emergence of powerloom manufacture in Western India 1920–1950’, Past and Present, 172:170–98. Haynes, D. and Roy, T. (1999) ‘Conceiving mobility: migration of handloom weavers in precolonial and colonial India’, IESHR, 36:35–68. Heston, A. (1983) ‘National income’, in D.Kumar (ed.) The Cambridge Economic History of India, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. Hicks, J. (1969) A Theory of Economic History, Oxford: Clarendon Press. Horton, S. (1996) ‘Women and industrialization in Asia: an overview’’, in S.Horton (ed.) Women and Industrialization in Asia, New York: Routledge. Humphries, J. (1995) ‘Women and paid work’, in J.Purvis (ed.) Women’s History: Britain 1850– 1945, Berkeley and Los Angeles: University of California Press. Hunter, W.W. (1872) Annals of Rural Bengal, London: Smith, Elder. Hussain, A. (1939–40) ‘Some aspects of agricultural labour in Hyderabad’, IJE, 20: 695–707. Hymer, S. and Resnick, S.A. (1969) ‘A model of an agrarian economy with nonagricultural activities’, American Economic Review, 59:493–506. Ibbetson, D. (1916) Panjab Castes, Lahore: Government Press. India (1881a) Census of India, Bombay Presidency, vol. I, Bombay: Government Press. ——(1881b) Census of lndia, Baroda, vol. I, Baroda: Government Press. ——(1901a) Census of India, Central Provinces and Berar, vol. XIII (Part I), Nagpur: Government Press. ——(1901b) Census of India, Central India, vol. XIX (Part I), Nagpur: Government Press. ——(1912) Report on an Enquiry into the Causes of Rise in Prices in India, Calcutta: Government Press. ——(1921a) Census of India, Central Provinces and Berar, vol. XI (Part I), Nagpur: Government Press. ——(1926–28) Royal Commission on Agriculture in India, vol. I–VI (evidence taken in various provinces), Calcutta: Government Press. ——(1928–31) Royal Commission on Labour in India, Report, and vols I–IXX (evidence taken in various provinces), London: HMSO. ——(1931a) Census of India, Bombay Presidency, vol. XVIII (Part I), Bombay: Government Press. ——(1931b) Census of India, Madras Presidency, vol. XIV (Part I), Madras: Government Press. ——(1931c) Census of India, Central Provinces and Berar, vol. XIII (Part I), Nagpur: Government Press. ——(1942) Report of the Fact-Finding Committee (Handloom and Mills), Commerce Department, Calcutta: Government Press. ——(1954) Agricultural Labour Enquiry: Report on Intensive Survey of Agricultural Labour, vol. I (All India), New Delhi: Ministry of Labour. ——(1973) ‘National commission on agriculture’ (extracts from the National Commission on Rural Labour), IJLE, 16:60–93. Islam, M.M. (1978) Bengal Agriculture 1920–1946; A Quantitative Study, Cambridge: Cambridge University Press. Japanese National Commission for UNESCO (1962) An Outlook of Studies on the Population Problem in Japan. Jha, P.K. (1997), Agricultural Labour in India, New Delhi: Vikas. Jorgenson, D.W. (1961) ‘The development of a dual economy’, Economic Journal, 71:309–34.

References

169

Jose, A.V. (1988) ‘Agricultural wages in India’, EPW, 23:A46–A58. ——(ed.) (1989) Limited Options: Women Workers in Rural India, New Delhi: ILO-ARTEP. Joshi, C. (2003) Lost Worlds: Indian Labour and Its Forgotten Histories, New Delhi: Permanent Black. Joshi, N.M. (1936) Urban Handicrafts of Bombay-Deccan, Poona: Gokhale Institute of Politics and Economics. Joshi, P.C. (1967) ‘Pre-independence thinking on agrarian policy’, EPW, 2:447–56. Joshi, V.R. (1958) ‘Growth of agricultural labour with special reference to UP’, IJLE, 1:175–95. Kaiwar, V. (1992) ‘Property structures, demography and the crisis of the agrarian economy of colonial Bombay Presidency’, Journal of Peasant Studies, 19: 255–300. Keatinge, G. (1921) Agricultural Progress in Western India, London: Longmans. Kessinger, T. (1974) Vilayatpur 1848–1968, Berkeley and Los Angeles: University of California Press. Krishna, R. (1976) ‘nemployment in India: a survey of concepts and definitions’, Washington DC: The World Bank. Krishnamurty, J. (1972) ‘The growth of agricultural labour in India’, IESHR, 9: 327–32. ——(1983a) ‘Occupational structure’, in D.Kumar (ed.) The Cambridge Economic History of lndia, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. ——(1983b) ‘Some aspects of the distribution of surplus labour time in rural India’ in A.Robinson, P.R.Brahmananda and L.K.Deshpande (eds) Employment Policy in a Developing Country: A Case Study of lndia, vol. 1, London and Basingstoke: Macmillan. Krishnaraj, M. (1985) ‘Research on women’s work in the seventies’, in K.Saradamoni (ed.) Women, Work and Society, Calcutta: Indian Statistical Institute. Kumar, D. (1965) Land and Caste in South India, Cambridge: Cambridge University Press. ——(1972), ‘Economic history of modern India’, IESHR, 9:63–90. ——(ed.) (1983a) The Cambridge Economic History of India, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. ——(1983b) ‘South India’, in D.Kumar (ed.) The Cambridge Economic History of India, vol. 2. c.1757–c.1970, Cambridge: Cambridge University Press. ——(1992) ‘Introduction’ to Land and Caste in South India, reprint by Delhi: Manohar. Lehrer, K. (2000) Theory of Knowledge (Dimensions of Philosophy), Boulder: Westview. Leser, C.E.V. (1958) ‘Trends in women’s work participation’, Population Studies, 12:100–20. Lewis, W.A. (1954) ‘Economic development with unlimited supplies of labour’, Manchester School of Economics and Social Studies, 22:139–91. Little, I.M.D. (1982) ‘Indian industrialization before 1945’, in M.Gersovitz et al. (eds) The Theory and Experience of Economic Development: Essays in Honour of Sir W.A.Lewis, London: George Allen and Unwin. Liu, P.K.C. (1984) ‘Trends in female labour force participation in Taiwan: the transition towards higher technology activities’, in G.Jones (ed.) Women in the Urban and Industrial Workforce: Southeast and Eawst Asia, Canberra: ANU Press. Ludden, D. (ed.) (2001) Reading Subaltern Studies, New Delhi: Permanent Black. Lyons, H.W. (1920–22) ‘A subsistence wage’, IJE, 3:448–59. Mann, H.H. (1924) Economic Progress of the Areas of the Bombay Presidency 1911–1922, Bombay: Government Press. Markovits, C. (2000) ‘Colonialism and traditional crafts’, EPW, 35:3939–40. Martin, J.R. (1903) Monograph on Tanning and Working in Leather in the Bombay Presidency, Bombay: Government Press. Mathur, A. (1994) ‘Work participation, gender and economic development: a quantitative anatomy of the Indian scenario’, Journal of Development Studies, 30:466–504. Mayo, K. (1927) Mother India, London: Jonathan Cape. Mincer, J. (1962) ‘The labor force participation of married women: a study of labour supply’, in H.G.Lewis (ed.) Aspects of Labor Economics, Princeton: Princeton University Press.

References

170

Mohapatra, P. (2000) ‘Situating the renewal: reflections on labour studies in India’, New Delhi: V.V.Giri National Labour Institute. Moreland, W.H. (1920) India at the Death of Akbar. An Economic Study, London: Macmillan. Morris, M.D. (1960) ‘The recruitment of an industrial labor force in India, with British and American comparisons’, Comparative Studies in Society and History, 2:305–28. ——(1965) The Emergence of an Industrial Labor Force in India: A Study of the Bombay Cotton Mills 1853–1947, Berkeley and Los Angeles: University of California Press. ——(1967) ‘Values as obstacles to economic growth in south Asia: an historical survey’, Journal of Economic History, 27:588–607. ——(1968) ‘Towards a reinterpretation of nineteenth century Indian economic history’, reprinted from Journal of Economic History, in D.Morris et al. (eds) Indian Economy in the Nineteenth Century: A Symposium, Delhi: Indian Economic and Social History Association. ——(1983), ‘The growth of large-scale industry to 1947’, in D.Kumar (ed.) The Cambridge Economic History of India, vol. 2, Cambridge: Cambridge University Press. Mukherjee, B.B. (1930–31) ‘Agricultural labour in north Bihar’, IJE, 9:320–31. Mukherjee, K. (1947–48) ‘Problem of agricultural labour in India’, IJE, 28: 163–77. Mukherjee, M. (1973) ‘Sources of growth of the Indian economy’, Sankhya, Series B, 35:133–40. ——(1974) ‘Share of Agricultural Labour in National Income’, Indian Journal of Industrial Relations, 9:563–85 ——(1995), Selected Papers on National Income, Calcutta: K.P.Bagchi. Mukherjee, R. (ed.) (1939) Economic Problems of Modern India, vols. I–II, London: Macmillan. Mukerji.K. (1959) ‘Trends in real wages in cotton textile mills in Bombay City and Island from 1900–1951’, Artha Vijnana, 1:82–95. ——(1960) ‘Trends in real wages in jute textile industry from 1900–1951’, Artha Vijnana, 2:57– 69. ——(1961) ‘Trends in real wages in cotton textile industry in Ahmedabad from 1900–1951’, Artha Vijnana, 3:124–34. Mukherji, N.G. (1903) A Monograph on the Silk Fabrics of Bengal, Calcutta: Government Press. Mukhtyar, G.C. (1930) Life and Labour in a South Gujarat Village, Calcutta: Longmans. Nair, J. (1998) Miners and Millhands: Work, Culture and Politics in Princely Mysore, New Delhi: Sage. Newman, R. (1981) Workers and Unions in Bombay, 1918–29: A Study of Organisation in the Cotton Mills, Canberra: ANU Press. Nussbaum, M. (2000) Women and Human Development: The Capabilities Approach, New York: Cambridge University Press. O’Rourke, K.H. and J.G.Williamson (1999) Globalization and History: The Evolution of a Nineteenth-century Atlantic Economy, Cambridge, MA: MIT Press. Orr, A. (1980) ‘De-industrialisation in India: a note’, Journal of Development Studies, 16:96–7. Parikh, I.J., Garg, P.K. and Menon, I. (1991) Women Weavers, New Delhi: Oxford and IBH. Parks, F. (1975) Wanderings of a Pilgrim in Search of the Picturesque, London, 1850, reprinted Karachi: Oxford University Press. Parry, J., Breman, J. and Kapadia, K. (eds) (1999) The Worlds of Indian Industrial Labour, New Delhi: Sage. Parthasarathi, P. (2001) The Transition to a Colonial Economy: Weavers, Merchants and Kings in South India, 1720–1800, Cambridge: Cambridge University Press. Patel, S.J. (1952) Agricultural Labourers in Modern India and Pakistan, Bombay: Asia. ——(1992) ‘Agricultural labourers in modern India and Pakistan’, in Prakash (ed.) World of the Rural Labourer in Colonial India, Delhi: Oxford University Press. Patnaik, U. (1983) ‘On the evolution of the class of agricultural labourers in India’, Social Scientist, 11:3–24 ——(ed.) (1990), Agrarian Relations and Accumulation: Mode of Production in Indian Agriculture, Delhi: Oxford University Press.

References

171

Perlin, F. (1983) ‘Proto-industrialisation in precolonial south Asia’, Past and Present, 98:30–95. Piore, MJ. and Sabel.C.F. (1990) The Second Industrial Divide, New York: Basic Books. Prakash, G. (1992a) ‘Introduction’ in Prakash (ed.) World of the Rural Labourer in Colonial India, Delhi: Oxford University Press. ——(ed.) (1992b) World of the Rural Labourer in Colonial India, Delhi: Oxford University Press. Prakash, V. (1960) ‘An estimate of stock of precious metals in India’, in V.K.R.V.Rao et al (eds), Papers on National Income and Allied Topics, vol. I, Bombay: Asia Publishing. Raju, A.S. (1941) Economic Conditions of the Madras Presidency 1800–1850, Madras: University of Madras. Raman Rao, A.V. (1958) The Economic Development of Andhra Pradesh 1766–1957, Bombay: Popular. Ranga, N.G. (1930) The Economics of Handlooms, Bombay: Taraporevala. Ranis, G. and Fei, J.C.H. (1961) ‘A theory of economic development’, American Economic Review, 51:533–65. Rao, G.N. and Rajasekhar, D. (1991) ‘Commodity production and the changing agrarian scenario in Andhra: a study in interregional variations, c.1910–c.1947’, in Bhattacharya, S. et al. (eds) The South Indian Economy: Agrarian Change, Industrial Structure and State Policy, 1914–1947, Delhi: Oxford University Press. Rath, N. (1983) ‘Measuring rural unemployment in India’, in A.Robinson, P.R.Brahmananda and L.K.Deshpande (eds) Employment Policy in a Developing Country: A Case Study of India, vol. 1, London and Basingstoke: Macmillan. Rath, N. and Joshi, R.V. (1966) ‘Relative movements of agricultural wage-rates and cereal prices: some Indian evidence’, Artha Vijnana, 8:115–32. Ray, R.K. (2000) ‘The spinning wheel is my husband…’ The Book Review, 24. Robinson, J. (1979) Aspects of Development and Underdevelopment, Cambridge: Cambridge University Press. Rose, S.O. (1992) Limited Livelihoods: gender and class in 19th century England, London: Routledge. Roy, T. (1998) ‘Development or distortion? The powerlooms in India, 1950–97’, EPW, 33:897– 911. ——(1999) ‘Growth and recession in small-scale industry: a study of Tamil Nadu powerlooms’, EPW, 34:3137–45. ——(1999) Traditional Industry in the Economy of Colonial India, Cambridge: Cambridge University Press. ——(2000) The Economic History of India 1857–1947, Delhi: Oxford University Press. ——(2002a) ‘Economic history and modern India: redefining the link’, Journal of Economic Perspectives, 16:109–30. ——(2002b) ‘Acceptance of innovations in early twentieth century Indian weaving’, Economic History Review, 55:507–32. Russell, R.V. and Hira Lal (1916) The Tribes and Castes of the Central Provinces and Berar, vols I–IV, London: Macmillan. Ruttan, V.W. (2001) Technology, Growth, and Development: An Induced Innovation Perspective, New York: Oxford University Press. Rutten, M. (2002) ‘A historical and comparative view of Indian entrepreneurship’, Economic Sociology (European Electronic Newsletter), vol. 3. Sabel, C.F. and Zeitlin, J. (1985) ‘Historical alternatives to mass production: politics, markets and technology in nineteenth century industrialization’, Past and Present, 108:133–76. Sanghvi, P. (1969) Surplus Manpower in Agriculture and Economic Development with Special Reference to India, London: Asia Publishing. Sarkar, S. (1983) ‘The colonial economy’, in Modern India: 1885–1947, Delhi: Macmillan. ——(2003) ‘The limits of nationalism’, Seminar, 522:19–24. Schultz, T.W. (1964) Transforming Traditional Agriculture, New Haven: Yale University Press.

References

172

Scudamore, W.V. (1907), A Monograph on Iron and Steel Work in the Bombay Presidency, Bombay: Government Press. Seccombe, W. (1986) ‘Patriarchy stabilized: the construction of the male breadwinner wage norm in 19th-century Britain’, Social History, 11:53–76. Sen, A. (1997) ‘On interpreting India’s past’, in S.Bose and A.Jalal (eds) Nationalism, Democracy and Development: State and Politics in India, Delhi: Oxford University Press. ——(1999) Development as Freedom, Delhi: Oxford University Press. Sen, S. (1999a) ‘At the margins: women workers in the Bengal jute industry’, in J.Parry, J.Breman and K.Kapadia (eds) The Worlds of Indian Industrial Labour, New Delhi: Sage. ——(1999b) Women and Labour in Late Colonial India: the Bengal Jute Industry, Cambridge: Cambridge University Press. Simeon, D (1995) The Politics of Labour under Late Colonialism: Workers Unions and the State in Chotanagpur, New Delhi: Manohar. Simmons, C. (1985) ‘“De-industrialization”, industrialization and the Indian economy, c.1850– 1947’, MAS, 19:593–622. Sivasubramonian, S. (2000) The National Income of India in the Twentieth Century, Delhi: Oxford University Press. Slater, G. (ed.) (1918) Some South Indian Villages, London: Humphrey Milford. ——(1918–19) ‘South Indian economics’, IJE, 2:145–57. Specker, K. (1989) ‘Madras handlooms in the nineteenth century’, IESHR, 26: 131–66. Stokes, E. (1994) ‘Dynamism and enervation in North Indian agriculture: the historical dimension’, in D.Ludden (ed.), Agricultural Production and Indian History, Delhi: Oxford University. Taeuber, I.B. (1958), The Population of Japan, Princeton: Princeton University Press. Thomas, P.J. and Natarajan, B. (1936) ‘Economic depression in the Madras Presidency (1825–54)’, Economic History Review, 7:67–75. Thorner, A. (1962) ‘The secular trend in the Indian economy, 1881–1951’, Economic Weekly, 14:1156–65. ——(1984) ‘Women’s work in colonial India, 1881–1931’, London: SOAS. Thorner, D. (1962) ‘“Deindustrialization” in India, 1881–1931’, in D. and A. Thorner, Land and Labour in India, New York: Asia. Thorner, D. and Thorner, A. (1957) ‘Types of employer-employee relationship in Indian agriculture’, Indian Journal of Agricultural Economics, 12:84–96. Thurston, E. (1899a) Monograph on the Silk Fabric Industry of the Madras Presidency, Madras: Government Press. ——(1899b) Monograph on the Cotton Fabric Industry of the Madras Presidency, Madras: Government Press. ——(1909) Tribes and Castes of Southern India, Madras: Government Press. Tomlinson, B.R. (1995) ‘Productivity and power: institutional structures and agricultural performance in India and China, 1900–50’, in J.Harriss, J.Hunter and C.M.Lewis (eds) The New Institutional Economics and Third World Development, London and New York: Routledge. Twigg, H.J.R. (1907) A Monograph on the Art and Practice of Carpet-making in the Bombay Presidency, Bombay: Government Press. Twomey, M.J. (1983) ‘Employment in nineteenth century Indian textiles’, Explorations in Economic History, 20:37–57. Venkatraman, K.S. (1935–36) ‘The handloom industry in south India’, supplements to The Journal of the Madras University, 7 and 8. Vicziany, M. (1979) ‘The de-industrialisation of India in the nineteenth century’: a methodological critique of Amiya Kumar Bagchi’, IESHR, 16:105–46. Visaria, P. (1997) ‘Women in the Indian working force: trends and differentials’, Artha Vijnana, 39:1–136.

References

173

Washbrook, D. (1999) ‘India 1818–1860: The two faces of colonialism’, in A.Porter (ed.), The Oxford History of the British Empire, III, The Nineteenth Century, Oxford: Oxford University Press. Watson, E.R. (1907) A Monograph on Iron and Steel Work in the Province of Bengal, Calcutta: Government Press. Yanagisawa, H. (1996a) A Century of Change: Caste and Irrigated Lands in Tamilnadu 1860s1970s, Delhi: Manohar. ——(1996b) ‘Elements of upward mobility for agricultural labourers in Tamil districts, 1865– 1925’, in P.Robb, K.Sugihara and H.Yanagisawa (eds) Local Agrarian Societies in Colonial India: Japanese Perspectives, Richmond: Curzon. YMCA (1926) Women in Industry in the Orient, New York: The Womans Press.

Index agriculture: early-nineteenth century 187; growth 4, 16, 36, 49; markets 35; productivity 4, 6, 19, 20, 23, 49 apprenticeship 189 attached labour 10, 57 Bagchi, A.K. 40, 89, 108, 110 Bengal: famine 34; labour-ratio 64, 68, 70, 73–4; migration 55; property rights 6, 35–6; textiles 112, 157, 159; wages 97, 101 Bombay Presidency: artisans 128, 157–8; labour-ratio 64, 70; labourers 58–9; peasants 76, 79; wages 87–8, 97, 101 caste 59 casual labour see labour Central Provinces and Berar: labourers 68, 70; textiles 112 census 3, 62–5 class 18 class structure 6 climate 7, 20–1, 33–5 colonialism 24, 60 consumption 12, 124–5, 189 culture 12–13, 21, 180 Dandekar, V.M. 21, 184 Darling, M.L. 72, 75, 184–5 de-industrialization 4, 15, 32, 46, 60, 107–8, 110, 118, 141–2 de-peasantization 60, 75 diminishing returns 6–8, 22, 47, 72 Drain theory 46

Index Dutt, R.C. 15 East India Company 29–31, 109–10 economic growth: colonial India 44, 46–7, 49 economic policy: colonial India 15, 27, 45, 61; post-1947 15–16, 27, 50–1, 120 education 12–13, 21 factory regulation 143, 144 famine 7, 33, 55 farm servants 10, 54, 56–60, 77 general labour see census; specialization globalization: late-twentieth century 16, 28; nineteenth century 4 gold 21, 75; see also investment and saving Great Depression 7, 64–5, 74–6 green revolution 23, 36 hand loom 111–12, 125–30, 152–8; see also textiles historiography 5, 14–19, 22, 27–8, 45–8 household; artisan 9, 13; farm 9, 13; general 18 Hyderabad 71 induced innovation 6–7, 183 Industrial Revolution 4 industrialization 16, 39–42, 107–8, 119–20 industry: growth and productivity 121–6; organization 126–8; technological change 130 inequality 5, 38, 101 investment and saving 6–7, 21, 43–4, 49 Ishikawa, S. 69 Kumar, D. 61, 87, 90, 111, 186 labour: contracts and institutions 9–10, 56, 58; factory 40;

175

Index ‘hoarding’ 57; market 9–11, 24; resource 8, 23; technological change 58, 67 labour-ratio 61, 68–72 labourer: definition and type 53–4 land: market 7; resource 6, 16–17, 36, 48, 67 land tax 20, 32 landlessness 5, 7, 9 leather industry 41–2 Lewis, W.A. 23, 93 Madras Presidency: agricultural growth 36; famine 33; farm servants 10, 57, 77; labour-ratio 64, 70; labourers 79–80; migration 55; property rights 35; textiles 129, 152–6, 161, 163; wages 87–8, 93, 97, 100–1; Malthus, T.R. 6, 20 market failure 15–16 marriage norm 146–9 Marx, K. 27 Marxism 14–15, 18, 22 migration: internal 8, 10–13, 54, 55, 92, 128, 159, 189; overseas 55 mode of production 185 modernization theory 22 Morris, M.D. 18, 40, 46, 108, 183–4, 187–8 Mughal India 29, 44 Mukherjee, R. 20–1, 71–2, 91, 184–6 Naoroji, D. 15 nationalism 14–16, 45 Nehru, J. 15 Patel, S.J. 15, 61, 63, 185–6 population growth 6, 23 post-structuralism 18, 169 poverty 3–8, 181 ‘power-loom’ weaving 127–32 prices 100–1 proletarianization see de-peasantization property rights 5, 31, 38;

176

Index see also settlements proto-industrialization 42, 107 Punjab: canal colonies 8, 55; farm servants 57; labour-ratio 64; wages 97, 100–1 rent 4–5, 183 resources 6–8, 16–17, 21, 23, 35, 36, 38, 47 risks 6, 7, 21, 33, 34 Royal Commission on Labour in India 151 Schultz, T.W. 23 Sen, A. 33, 189 settlements, agrarian 5, 6, 31, 35 Sind: artisans 160 Slater, G. 20, 72, 93 Soviet school 15 specialization 78–81 standards of living 100 Subaltern studies 17, 166–70; economic history 171–3; the State 173–6 tenancy 67 textiles 32, 39, 110–11, 152–61; see also de-industrialization Thorner, A. 62, 65, 80, 109, 184, 188 Thorner, D. 62, 65, 80, 184, 188 trade; early-modern 29–30, 111; nineteenth century 31, 35, 42 traditional industry 41–2, 123–4 unemployment 23 UP (United Provinces/Uttar Pradesh): artisans 160; labour-ratio 64; labourers 72; wages 100–1 village community 61 wage: agricultural 5, 8, 38, 54, 95–7; artisan 42, 155; convergence 92; fluctuations 97–8;

177

Index sources 86–7, 103–4; trends 88–95, 186–7 wage differential 14 wage-rental ratio 4–5 wage-share 98–9 water 7, 21 Weber, M. 22 women: education 21–2; work-participation 140–9 women in agriculture 11, 54, 65, 76–8 women in industry: general 134–7; India 137–40; textiles 152–61 workforce 39, 60

178