India as a Colonial Economy

The Debate on Deindustrialization

BY
ADITYA SARKAR

Delhi University

A major issue in academic debates about the nature of the Indian economy under
colonial rule has been the problem of de-industrialization, or the decline of indigenous
manufacturing industry. In the accounts of nationalist intellectuals, and in the
commonsense of much economic history-writing till the 1960s, it was more or less
assumed that the nineteenth century in Indian history saw a major decline of handicrafts
that was not compensated by the rise of modern industry, that this decline continued
into the twentieth century, and that large sections of the working population were
forced into agriculture. From the 1960s on, these positions were complicated
considerably by a number of revisionist critiques, and in one form or another the
literature bearing on the subject of de-industrialization has continued to grow right up
till the present day.

Briefly, this essay tries to explore some of the major issues and positions
involved in the debates around this problem. First, it examines Daniel Thorner’s
seminal critique of the census data that apparently demonstrated de-industrialization.
From this, it moves on to look at the controversy that exploded in the 1968 and 1969
editions of the IESHR around Morris D. Morris’ essay in 1968. While there were many
attacks on Morris’s argument, the present essay only discusses (apart from Morris’ own
paper) the critique offered by Toru Matsui. In itself, it can be argued that the debate
around Morris’ piece generated more heat than light, but it set the tone for much
subsequent research, and also put in place a certain crystallization of ideological fault
lines that has defined (some might argue bedevilled) the different positions on de-
industrialization. Ideology and analytical economic history were closely tied together in
the debates from this point on. This essay then goes on to study the debate between A.K
Bagchi and Marika Vicziany in the late 1970s. After this, it shifts to a brief examination
of some of the empirical studies made of handloom industry in the colonial period, as
contained in essays by Specker, Yanagisawa, and Sumit Guha. Finally, it attempts to
engage with the important alternative interpretation set forth by Tirthankar Roy,
arguably the most significant contribution to the debate in recent years, but also
extremely problematic and questionable in its assumptions and implications.

Before turning to a systematic survey of the literature, a couple of broad

qualifications need to be made. First, the hidden assumption made by both sides in the
de-industrialization debate has been that it is possible to quantitatively demonstrate the
crippling effects (or absence thereof) of colonial rule in the nineteenth century. While
 2 Aditya Sarkar| The Debate on Deindustrialization
Aditya 2

sustained quantitative analysis is indispensable, it can certainly be argued that the

strength of the case against colonialism need not depend on this sort of proof. In other
words, Morris’ bland assertion that the Raj’s main function was to secure ‘the welfare of
its subjects’ can be rebutted without necessarily building up a picture of devastating
economic decline: the exploitative and unjust nature of British rule is far too evident.
This point is significant because otherwise the polemical attributions of ‘left-
nationalism’ and ‘neo-colonialism’, which both sides in the debate have been guilty of
making (Roy’s dismissal of pre-existing literature on the colonial economy and Bagchi’s
intemperate reaction to Vicziany are instances of this) tend to stick. To argue, on the
basis of empirical evidence, that ‘de-industrialization’ is too simple a category to be able
to hold the experiences of nineteenth-century handicraft production, is not in itself a
symptom of an apologia for colonialism. The fact that, in the actual debate, the fault line
has often – though not always – been between radically different perspectives on the
nature of the colonial experience, should not obscure this, and lead one to simply read
ideological motivations off every position in the available literature.

Second, there is the additional qualification that while the different essays
assembled in this survey have a bearing on the de-industrialization debate, their value as
history, curiously enough, does not lie in the light they throw on de-industrialization as
a general problem. By and large, the bulk of the researches around this theme
concentrate on the history of textiles, and, even more specifically, on handloom
weaving. Roy is probably right in arguing that no comprehensive picture of de-
industrialization can come from a narrow sectorial focus. With the exception of the
essays of Thorner and Bagchi, the empirical content of the debate is less wide-ranging
than one might expect. Having said that, the importance of the historical insights and
evidence produced – from diverse positions – in the debate is unquestionable.

I
Daniel Thorner’s 1960 essay occupies an ambiguous place in the de-
industrialization debate, and this is an ambiguity that in a sense demonstrates the
artificiality of the ideological straitjackets into which historical research can be forced.
Thorner made the first important critique of the thesis that Indian handicrafts
experienced a measurable decline in the second half of the nineteenth century and the
first half of the twentieth. In doing so, he became a necessary point of reference in all-
subsequent debates on the matter. Roy’s brief essay on de-industrialization
(significantly, delivered as the ninth Thorner Memorial Lecture), for instance, tried to
assimilate Thorner’s critique of the census to a general argument about the
insufficiencies of ‘left-nationalist’ interpretations of Indian economic history. There is an
incongruity in this chosen alignment. Alice and Daniel Thorner were – and remained –
Communists, persecuted and driven abroad in the course of the McCarthy witch-hunts.
This fact is an important warning against the tendency to invest arguments with
political-ideological slants on the basis of simple, symptomatic readings of their
India as a Colonial Economy

position. It also shows up Roy’s tendency to greatly simplify and vulgarize a complex
and rich spectrum of left-wing economic thought, accommodative of greater differences,
and hence less rigid than he makes it appear. The fact that scholars as divergent as
Thorner and Bagchi can both lay claim to the Marxist tradition is a valuable reminder
of its internal diversity.

When he wrote, Thorner was reacting to the assumptions about de-

industrialization in India that currently constituted the orthodoxy in economic history.
Nationalists like R.C Dutt had argued that the policy of free-trade competition forced
upon India by Britain in the later nineteenth century, and the concomitant denial of
protective tariffs to indigenous industry had certain devastating consequences for the
Indian economy. The import of cotton goods trebled within 40 years (1870-1910), and
Lancashire flooded the Indian textile market, thus obliterating a vigorous tradition of
spinning and handloom weaving. The production of cotton goods within India fell, and
exports also contracted considerably. These arguments had relied heavily on import-
export statistics, and had assumed that imports showed a substitution of indigenous
production by exports. This, in turn, rested on the assumption that the size of the
market was static (Morris’ criticism of this is his one substantive contribution to a
debate that, ironically, his arguments seem to dominate). However, Thorner was also
tackling arguments about de-industrialization that came from other sources. Vera
Anstey, writing from a position diametrically opposed to that of Dutt and Shelvankar
described India’s economic progress as ‘arrested development’. Colin Clark used the
Census to demonstrate that the occupational structure of the Indian working population
between 1881 and 1911 had shifted decisively against industry and in favour of
agriculture.

It was the Census figures between 1881 and 1931 that provided the material that
Thorner handled in his critique of the ‘de-industrialization’ orthodoxy. On the surface,
the Census seems to bear out the orthodoxy quite unambiguously. In this period, the
Census indicates a growth of the total working population from 115.1 million to 140
million. Within this, the number of people involved in agricultural work seems to have
increased from 71.7 to 100.2 million, and the number of workers in manufacturing
activity seems to have fallen from 21.1 to 12.9 million. Thorner’s brilliant essay lays bare
the loopholes in this apparently explicit statistical statement, with the basic insight that
the categories used in the Census simplified forms and experiences of work that were
actually much more complex, and drew clear boundaries between activities that in
reality were much less differentiated. The clear separation made between people
involved in manufacture and trade is an important instance of this problem. Industry
and commerce are actually very difficult to disaggregate, since the household often
straddled both spheres of activity, and the makers of goods were also often sellers of
goods. Further, the census also separated agricultural labour from ‘general’ or unskilled
labour. In Thorner’s view, this is an arbitrary separation. It is possible to deduce from
 4 Aditya Sarkar| The Debate on Deindustrialization
Aditya 4

the Census figures an inverse relationship between agricultural and general labour: the
years where one grew were also, invariably, the years when the other declined. This
would suggest that the same people moved from agricultural labour to part-time non-
agricultural activities (construction work, transport services and so on) and back.
Finally, Thorner points out that the census figures about female labour were extremely
unsound, since they were based on faulty enumerative practices. Women often refused
to appear before the Census Commission, and the occupations of women were often
equated with those of their husbands. Thus, to arrive at a sounder picture of the
occupational structure of the Indian working force, argues Thorner, it is necessary to
combine manufacture and trade under a single head, combine the categories of
agricultural and general labour, and set aside the data for women (the figures for women
involved in manufacture seem to have been especially grossly overstated for 1881).
Having made these qualifications, Thorner arrives at the startling conclusion that there
was, over these fifty years, a 1% rise in the agricultural working population, and a 3%
decline in the numbers of people involved in manufacture and trade. In other words, the
occupational structure of India stood still.

Thorner’s analysis is not free of methodological difficulties. The complete

identity assumed between agricultural and general labour is problematic. The assumed
merging of manufacturing and trade can also be qualified, and it is possible to imagine a
slow movement towards more or less clearly separated spheres of activity over this
period. The exclusion of women from the analysis is questionable, since it leaves a large
proportion of the working population unaccounted for. Finally, there is Bagchi’s
interesting speculation (of which more later) that the absence of differentiation between
different forms of labour and livelihood might have been precisely a product of de-
industrialization: artisanal activity was probably far more specialized, for instance, in the
early nineteenth century. This is not fundamentally a critique of Thorner’s position, for
the latter accepts the possibility – and even seems to acknowledge the logical
probability – of a genuine de-industrialization prior to the 1880s. Further, Thorner
remarks upon the remarkable fact that in a period when population increased by over
100 million, there was no significant change in the occupational structure. If nothing
else, this might be a clear indication of stagnation. Thorner’s claim upon the issue of de-
industrialization is at the same time modest and inescapable: it is not possible, on the
basis of Census figures, to quantitatively prove an absolute decline in industry. Thus, while
his analysis did not claim to establish anything concrete, it served to destabilize old
certainties and open up a fruitful arena of debate. Most significantly, perhaps, Thorner,
writing in 1960, showed an astonishing sensitivity to the data that he sought to handle.
It is all the more important that this data happened to be the Census, and his particular
critique rested on the legitimacy of the categories of classification and enumeration that
it chose to employ. The critique of Census categories has been a major component of
recent historical research on a wide range of themes, and Thorner’s awareness of the
India as a Colonial Economy

problems with the way categories of knowledge are constructed is quite remarkable for
its time.

II
The hard empirical grounding of Thorner’s analysis is aeons removed from the
entirely speculative nature of Morris D. Morris’ 1968 article, ‘Towards A
Reinterpretation of 19th Century Indian Economic History’. Thorner’s essay was a
careful, meticulous and ultimately dazzling exercise in the interpretation of
demographic data. It would not be inaccurate to say that Morris’ piece is characterized
chiefly by the absence of any kind of data. To avoid being unjust to Morris, it is
necessary to accept that the absence of sustained analyses of the nineteenth century
Indian economy (barring the unacceptably underrated work of the economic
nationalists) was a genuine problem at the time that he wrote. The tentative tone of
Morris’ argument is fully understandable, and it is also true that he put forward his
hypotheses as a stimulus to further research, in which endeavour he succeeded, in
retrospect, quite dramatically.

That said, the content of Morris’ vast generalizations about the Indian economy
is often (in fact usually) quite untenable, and also quite inexplicable. Statements like
these are a case in point: ‘Certainly, the general object of the raj was the welfare of the
society.’ Or: ‘The British rule introduced the political framework of the…liberal nation
state…Public order was established on a scale never seen before….Taxation and
commercial regulations were largely eliminated.’ Or: ‘Indian society has been based
historically on a non-animal-powered agriculture’. (It is still unclear exactly where
Morris gets this particular insight into the Indian economy from). What is disturbing is
not these isolated statements, but the fact that the structure of his argument is largely
based on assumptions like these, and others of comparable looseness and unsoundness,
some questionable and others wrong. Essentially, Morris makes out a case that the
nineteenth century saw (as a result chiefly of the manifold benefits of British rule: public
order, efficient administration, the absence of war and conflict, improved
communications) significant positive economic changes: an expansion of agriculture
with no decline in marginal product because of the productivity of land, a shift to high-
value crops, slow population growth, and a consequent rise in per capita income. Toru
Matsui, Tapan Raycahudhuri, and others have questioned many of these confident
assertions. Raychaudhuri makes the point that the eighteenth century was by no means
as uniformly a period of decline as Morris’ essay assumes, and the early eighteenth
century saw much commercial growth. Further, the expansion of agriculture was not
uniquely a feature of British rule; previous dispensations (the Mughals for instance) had
followed a deliberate policy of reclamation of land for agricultural purposes. Matsui also
points out that the rise in per capita income in the nineteenth century probably did not
compensate for the many devastations caused by war and political conflict in the second
half of the eighteenth, and whatever the indicators of growth (meagre though these are)
 6 Aditya Sarkar| The Debate on Deindustrialization
Aditya 6

may suggest for the nineteenth century, the rise in per capita income would have been
too limited to provide any obvious succour.

Of immediate concern here, though, is Morris’ particular revision of the history

of handicraft production. Here, almost by accident, he seems to have put his finger on a
weak point in the de-industrialization thesis: the assumption that the market was static.
Thin though the argument was in its time, the researches of Roy and others in recent
years have, ironically, given it both substance and stature. Morris’ speculation is that
the import of cheap machine-made yarn helped indigenous weaving industry, cutting
down its costs of procurement. The general fall in cloth prices led to a ‘movement down
the demand curve’ (in other words, the price levels fell, the volume of marketed cloth
expanded, and demand was deepened). Further, there was ‘a shift to the right of the
demand curve’: the total quantity of cotton cloth available expanded. At worst, then, the
expansion of British textile imports ‘skimmed off the expanding demand’ in an
expanding market. The traditional sector might not have contracted absolutely, given
the expansion of demand. At one level, this argument is Morris at his most Panglossian.
However, it is also the most logically worked-out part of his essay, and has an empirical
basis that, ironically, he would have been unaware of at the time.

Toru Matsui’s critique of Morris is hard-hitting but sophisticated. While his

essay is wide-ranging and effectively attacks all the bases of Morris’ argument, it is only
necessary here to take note of the section relevant to the question of de-
industrialization. Matsui accepts the point that the competitive position of the handloom
sector may have been strengthened by the low price of imported yarn. However, he
argues, it is important to keep in mind the entirely destructive effect this would have
had – and did have – on the indigenous spinning industry. Further, he takes issue with
Morris’ general conclusion that the Indian weaver benefited from the import of cotton
from Britain. The general fall in prices owed as much to the cheapness of the cloth
imported from Britain as to that of the yarn. Only the import of yarn would have
benefited Indian weavers. Equally importantly, Matsui points out that Morris
recognizes, quite correctly, that it was the contraction of prices that caused the
expansion of demand, and not the reverse. In such a situation, if generalizations were to
be made, it would be more logical to conclude that the position of the Indian weaver
declined, especially in the absence of any evidence of qualitative technological changes
in the weaving industry. (This is one area where Matsui’s conclusions have become
outdated, given Tirthankar Roy’s studies of production processes in the traditional
sector). Matsui also argues strongly for the need for a more regionally disaggregated
and differentiated history of traditional industry. He points out that the demand for
cloth varied very widely, and that not all cloth was bought and sold in the open market.
In this context, it makes little sense to talk of a general ‘shift to the right of the demand
curve’. Regional experiences were, he argues, deeply inscribed into the picture of the
fortunes and misfortunes of traditional industry. The survival, prosperity or destruction
India as a Colonial Economy

of handicrafts under colonial rule in specific regional contexts demands study on its own
terms.

III
In 1976, the economic historian Amiya Kumar Bagchi, writing from a standpoint
that Tirthankar Roy would later castigate as ‘left-nationalist’, attempted an empirical
study of the process of de-industrialization in a specific geographical context: the
districts of Patna, Gaya, Purnea, Bhagalpur and Shahabad in Gangetic Bihar. Bagchi’s
investigations were a methodologically interesting exercise in comparative history.
Fully cognizant of the difficulties in formulating anything like a statistical outline of
employment, occupational structure, income or any of the other standard economic
variables in the context of the nineteenth century, he selected two very different kinds
of sources from two very different periods. Both these sources, however, he argued,
could claim to be the most reliable documents available in their time. The first was the
remarkable accounts of Dr Francis Buchanan (later Buchanan-Hamilton), based on his
survey of Bihar and Bengal in 1809-13. The second was the section of the Census of
1901 that touched on the same region. While Bagchi’s methods were to come in for
some sustained and cogent criticism from Marika Vicziany, his arguments did provide
an important empirical basis for the case that nineteenth-century India experienced de-
industrialization.

After a long defence of the merits of Buchanan-Hamilton’s investigations, Bagchi

analyses the relevant comparative sets of statistical data, and comes to certain
significant conclusions. In 1809-13, he estimates, the percentage of industrial to total
population in the relevant districts of Buchanan-Hamilton’s survey ranged from 14.2%
to 20.2%. The proportion of the industrial population dependent on cotton spinning and
weaving was extremely high, ranging from 58% to 69.9%. In 1901, on the other hand,
the population dependent on industry in the same region had shrunk to 5.5% - 11.6%,
and the numbers dependent on cotton spinning and weaving to the industrial population
was only 12.4% - 22.4%. Thus, the relative weight of the cotton sector in secondary
industry – which it had dominated in the early nineteenth century – had contracted
considerably by 1901. However, Bagchi argues that the cotton spinning and weaving
industries were not the only ones to experience decline. Hunter’s Statistical Accounts and
the District Gazetteer record a similar story of decline for silk manufacture, and for the
cloth-dyeing industry. The paper industry of Gaya and Shahabad contracted almost to
the point of extinction. Purnea, an important manufacturing centre (dealing in
commodities like lac ornaments, glassware, tooth-powder and blankets) witnessed an
almost complete deindustrialization. Further, the weaving of coarse cotton cloth had by
this time (the 1870s) almost entirely replaced the production of finer varieties. This
modest survival was due to the relatively steady demand for coarse cloth among the
poor, the depression of weavers’ wages to the levels of those of unskilled labourers, and
the taking up of weaving as a part-time occupation in the slack season, in order to
 8 Aditya Sarkar| The Debate on Deindustrialization
Aditya 8

supplement agricultural incomes. Thus, the mixing of occupations that Thorner

identifies as a structural component of the composition of Indian labour, may actually,
according to Bagchi, have been a consequence of the very process of de-industrialization.

In 1979, Marika Vicziany offered an important methodological critique of Bagchi

in the IESHR, and his reply appeared in the same issue of the journal. Important points
were raised in both essays, and despite the polemical edge of Bagchi’s riposte, there were
serious questions that were left unresolved. As debates go, the Bagchi-Vicziany
exchange was very unresolved. This flowed partly from significant methodological
problems within the critique itself, and its rather careless attribution of the wrong
sources to Bagchi’s account. Bagchi, in his turn, chose to be scathing rather than
substantive in his reply. While he effectively showed up the gaps in Vicziany’s critique,
he did not satisfactorily answer the larger methodological problems her essay had
raised.

Vicziany’s critique of Bagchi is founded on a basic dissatisfaction with Buchanan-

Hamilton as a source of information, and, more generally, with the notion of
comparability between his accounts and the 1901 census. The first part of her survey is
an examination of the reliability of Buchanan-Hamilton’s methods and conclusions. She
points out that his data collection was largely constrained by the kinds of information he
received from local people and rather unreliable guides, and his autonomy as an
investigator was greatly restricted. Often the guides seem to have been agents of local
zamindars and notables, and Buchanan himself seems to have felt that they were
instructed about what to tell him. Buchanan records several instances of tensions and
confrontations between himself and local people. Vicziany accepts that Buchanan’s
surveys may have had value as a general account of the Bihar districts. However, she
has legitimate doubts about the procedures of quantification he followed, since there is
very little about the statistical dimension of his investigations in the journals
themselves. The structure of local economy as recorded by Buchanan may have been
very different from the real, ‘objective’ structure, and the quantitative part of Buchanan’s
account seems to be suspect in terms of its reliability. His methods of enumeration seem
to have been necessarily rough and piecemeal – especially, according to Vicziany, when
it came to estimating the numbers of spinners.

Bagchi’s estimate of the number of persons dependent on spinning comes in for

some criticism for Vicziany. Bagchi had made two alternative assumptions, based on the
generally low earnings of spinners: a) that the average spinner supported one person
other than herself, and b) that she supported herself alone. Vicziany takes issue with
Bagchi over his refusal to consider a third possibility – that the spinner’s earnings could
not sustain her, and that spinning work, therefore, was part-time. Vicziany points out
that Buchanan himself had pointed to the part-time nature of spinning, and calculating
from his estimates, she argues that the average wage of a spinner in Bhagalpur and
India as a Colonial Economy

Shahabad ( as opposed to Purnea ) was substantially below subsistence level. Vicziany

also asserts that Montgomery Martin’s version of Buchanan’s statistical tables (which
she assumes are Bagchi’s prime source) is too vague to serve as a reliable source. Here,
though, she makes an astonishingly careless mistake which Bagchi is very severe on in
his reply: in reality, Bagchi had access to the original tables drawn up by Buchanan, and,
unlike Vicziany, did not have to rely on Montgomery Martin’s condensations. Vicziany
also reiterates Thorner’s point about the difficulty of separating manufacture,
agriculture and trade in the case of many people involved in artisanal activity: blanket-
weavers, for instance, were simultaneously farmers, as were iron smelters. The makers
of a particular kind of salt in Shahabad district, described by Buchanan, seem to have
been labourers working at different jobs through the year.

Vicziany also examines Bagchi’s attempt to adjust data so as to be able to

compare Buchanan’s surveys with the 1901 census, and reaches the conclusion that the
effort to establish comparability fails. Census supervisors, she argues, were very aware
of the crudeness of the categories they were forced to employ, and the divisions they
made were often for purely pragmatic purposes. They instructed enumerators to
distinguish clearly between the principal occupation and subsidiary occupations of
actual workers, and to separate workers from dependants clearly. The enumerators,
however, often failed to make this distinction. Vicziany suggests that Bagchi uses the
1901 figures for a much finer-grained analysis than they could actually yield, and that
the men who fashioned the census recognized its crudity. Finally, with respect to the
decline of the handloom industry, Vicziany points out that Bagchi’s use of qualitative
data relies exclusively on the District Gazetteers and the Statistical Account, ignoring
alternative sources like the Indian Industrial Commission, which focuses on the survival
of the indigenous cloth industry in the face of competition.

Bagchi’s reply to Vicziany, while devastating in its polemic, fails to address

seriously the more substantive points made by her. He more or less dismisses the point
about the comparability of data with a brief aside to the effect that all researchers suffer
from the same problems that Buchanan faced. This is a somewhat unsatisfactory reply,
since Vicziany had raised very concrete problems of analysis that flow from a study of
Buchanan’s methods of data collection. In any case, Vicziany’s central point was that
Buchanan’s quantitative analysis was of necessity founded on shaky data. The strength of
the qualitative data, the authenticity of description, and the perceptiveness of the
investigator do not negate the very real objective constraints he had to work with,
which were much greater than the problems a well-organized bureaucratic setup like the
Census Commission faced. Given the problems with the census data, the suggestion that
Buchanan’s quantitative data are genuinely unreliable is not unreasonable.

Another point made by Bagchi is an impatient reiteration of his argument that

the mixing of occupations was a product of deindustrialization. This is in itself an
 10 Aditya Sarkar| The Debate on Deindustrialization
10

interesting speculation, and certainly there was a decline in full-time craft specialization.
The problem is that it is difficult to locate the change in very precise historical time. It
may well be the case that Company rule, destroying traditional industrial occupations in
the late eighteenth and early nineteenth century, killed craft specialization for a
considerable length of time. However, it is unreasonable, in the absence of quantitative
data, to put this theory forward with the kind of finality that Bagchi does. Bagchi
argues that evidence from the seventeenth century on demonstrates considerable craft
specialization, that the late nineteenth century is marked by a disappearance of this, and
that colonialism is therefore clearly responsible for the intervening decline. This is a
plausible argument as long as one does not insist that it is final and susceptible to
quantification. Bagchi also tries to defend his method of deducing the number of
spinners from the number of dependants on a spinner. He rejects Vicziany’s suggestion
that the spinner’s income might not have been sufficient even to support herself,
pointing out that in Purnea the average spinner earned almost as much as a ploughman.
He is also on fairly strong ground when he defends the qualitative evidence for de-
industrialization, pointing to detailed descriptions and eyewitness accounts. The logical
slip he makes lies perhaps in the assumption that it is possible to generate quantitative
evidence that unambiguously confirms this. As he acknowledges, there was a great deal
of regional diversity in the experience of handicrafts. In this light, his implicit dismissal
of Vicziany’s critique as a neo-imperialist ‘smart ploy’ is somewhat unwarranted and
unnecessary. The Bagchi-Vicziany debate contains a great deal of empirical material and
a lot of sophisticated analysis and qualification of difficult data – on both sides. The
harshness of the tone of Bagchi’s reply, and his attribution of motives to Vicziany,
unfortunately, obscures much of this richness.

IV
The 1980s and 1990s saw a series of regional studies that touched directly or
indirectly on the subject of de-industrialization. The main focus of many of these studies
was the handloom sector in the late nineteenth and the early twentieth century, and a
great deal of empirical evidence, far richer than the de-industrialization ‘debate’ by itself
could contain, was generated. The relevant arguments, for the purpose of this essay, are
the ones put forward by Konrad Specker, Haruka Yanagisawa and Sumit Guha, whose
contributions are contained in an important volume, Cloth And Commerce: Textiles In
Colonial India, edited by Tirthankar Roy.

Specker’s essay is a brief survey of the development of the textile industry in

Madras in the nineteenth century. Unlike the literature on the question of de-
industrialization examined above, he does not seek to make any wide-ranging
generalizations on the development (or otherwise) of the colonial economy: his study is
focused and limited. Specker carefully traces the trajectory of the Madras handloom
sector in the nineteenth century, distinguishing between long-term trends and
responses to crises like famines. In general, he argues that the number of looms tended
India as a Colonial Economy

to increase between 1820 and 1870. This was not a smooth growth, though. It was
highly differentiated regionally, and the famines of 1824-26 and 1833 had a debilitating
impact upon production in Masulipatam and Guntur. A rise in the number of looms only
set in after the 1840s and 1850s in most areas. In Cuddapah, North and South Arcot,
Tanjore, Malabar and Coimbatore, there was a decline right up to 1870, partly as a
result of textile imports from England, and partly because of the cotton boom of the
1860s, which pushed up the price of raw cotton. Guledgud and Kornadu, on the other
hand, experienced considerable growth, since cloth produced in these areas faced no
competition. Quantitatively, Specker argues, there was no long-term ‘destruction’ of the
traditional textile industry. The real changes, though, were qualitative and product-
specific. Imports from England tended to force the traditional sector to shift towards the
production of coarse and low-grade cloth. However, product-specific advantages and
demand patterns kept the industry afloat in many old centres of production. At the same
time, contra Morris, the imports of machine-made yarn did not automatically strengthen
the competitive power of the Indian weaver. Under-employment became a marked
feature with the shift to coarse cloth production. Further, the greater the reliance of
weavers on the local market, the greater was the danger of susceptibility to rising
agricultural prices, especially in periods of crisis like the famine of 1877. In brief, the
textile sector did experience a certain falling-off and marginalization, but there was no
absolute decline, and the trajectories of growth and contraction were complex and
regionally differentiated.

These trajectories, however, changed qualitatively in the twentieth century, and

Yanagisawa’s essay (also incorporated in Roy’s edited volume) deals with these changes.
The process he points to is market segmentation, in the context of the Indian cotton
mills emerging as the major competitor to Indian handloom production as imports
levelled off in the first quarter of the twentieth century. Between 1901-02 and 1936-37,
the share of Indian mill production in the cloth market rose from 11.9% to 61.5%. The
handloom sector countered this competition by diversifying production (a process also
emphasized by Roy). Four types of alternative production emerged: fine-coloured cloths
with high-count yarns, silk and sometimes jari or gold thread; coloured cloths woven
with artificial silk yarn for lower-class consumption; coloured cloth exported abroad;
and rough cloth with low-count yarns. The last of these shifts, towards coarse cloth
production, was, according to Yanagisawa, largely a product of the late 1930s and the
1940s, and of war conditions. Yanagisawa also points to a shift in fashions, and
consequently in demand and consumption patterns, that partly determined the shifts in
handloom production. Production for female demand (especially among the upper
classes, where women’s clothes became costlier) was one of these shifts, as handlooms
began to cater to specific and new kinds of needs, supplying bodices and blouses.
Demand for handmade saris also probably increased. Yanagisawa argues that the
emancipation of the lower castes and the relaxation of social restrictions on their
clothing resulted in increased demand for new kinds of cloth, like hosiery, and also for
 12 Aditya Sarkar| The Debate on Deindustrialization
12

handloom-produced cloth for ceremonial purposes. Thus, the data generated by this
study demonstrates that there was no decline in the south Indian handloom industry in
terms of the number of looms despite the competition from indigenous mill-made cloth;
in fact, the market for handloom production in certain products expanded. However, as
Yanagisawa admits, the ‘economic condition of weavers may generally have worsened.’
The absence of any account of the changes in relations of power and production in
traditional industry seriously limits the scope of his essay, since the diversification and
sophistication of production bears no direct correlation with the living conditions and
social experiences of weavers.

Guha’s essay, on the handloom industry of central India from 1825 on,
intricately weaves together a number of economic processes into a complex and nuanced
historical narrative. As might be expected, the picture is a differentiated one, and the
trajectory of change is uneven, although patterns and correlations can be deduced. For
the 1820s, Richard Jenkins’ report on Nagpur indicates a situation of flourishing textile
manufacture and a large export trade. Jenkins’ census (which only took into account the
occupations of adult males) also demonstrates that a sizeable workforce was involved in
both spinning and weaving. From the 1820s to the 1860s, statistical material is much
more limited, but scattered information from official reports and the accounts of the
Maheji fair at Khandesh suggest that the market for Nagpur textiles was still immense.
Adverse changes, however, took place in the 1860s, as the prices of raw cotton soared.
Local weavers dealing in coarser kinds of manufacture seem to have experienced a
significant starvation of raw material, and a certain differentiation seems to have been
generated by the shift to the production of finer fabrics, which were less affected by the
boom in cotton prices. In general, though, weavers suffered, and this was exacerbated by
the high price of food grains (which would have constituted the bulk of their
expenditure), and the need to compete with cheaper imported textiles. The dependence
of C.P weavers on handspun yarn made this struggle all the more unequal. Guha finds a
marked decline in the weaving community of Nagpur city in 1866 as compared to 1825.
Cotton production seems to have picked up after 1868, but there is evidence of
considerable economic distress among weavers and spinners (especially the latter), as
real incomes seem to have contracted. Between 1870 and 1895, two processes seem to be
tied together. First, the weavers shifted to machine-spun yarn, and the spinning
industry underwent a great decline. Spinning seems to have been reduced to a
secondary or part-time occupation, often combined with agricultural labour.
(Interestingly, this lends weight to Bagchi’s general thesis). Second, there was a distinct
recovery in both the production and the consumption of handloom production, and the
demand for coarse cloth picked up. Employment in weaving picked up, although the
collapse of hand spinning meant a definite fall in cumulative employment between 1870
and the twentieth century.
India as a Colonial Economy

This trend was partly modified and partly consolidated after 1900, as a cycle of
scarcities and famines permanently destroyed hand spinning. Weaving employment also
contracted, though production didn’t fall substantially. Guha suggests that the poorest
and least credit-worthy weavers would have been most likely to perish in this period of
flux. There appears, in the statistical series of this period, a major discrepancy. While
the consumption of yarn in traditional industry remained substantially unmodified
between the 1880s and the 1940s, employment seems to have shrunk dramatically.
Between 1901 and 1941, numbers employed declined from 150,000 to 70,000. Once
again, distinct economic processes seem to have intersected here. Guha suggests that
much employment in weaving would have become part-time, and many people involved
in the production process would have not appeared in the census data, especially with
the disappearance of spinning. The increase in yarn consumption after 1930 was also
related to a technological change: the replacement of the throw-shuttle by the fly-
shuttle. Thus, employment declined, working conditions may have become worse and
more casualized for many, and productivity per loom probably increased – and all of these
were part of the same historical process.

The Central Provinces, despite the moderacy of Guha’s conclusions, seem to bear
out some key aspects of the de-industrialization argument quite clearly. The position of
weavers seems to have been fairly fluctuating and unstable throughout the period after
the 1820s. Employment seems to have witnessed a fairly steady contraction from the
1870s, though not as violent as that experienced in the previous decade. The spinning
industry was obviously decimated, and weaving labour seems to have in many cases
undergone a considerable deterioration in its terms of existence in the twentieth
century. The other side of the picture is output and productivity, which seems to have
undergone a certain qualitative expansion over time. In its basic outline, this does not
seem so very different from the south Indian case, where again the case against de-
industrialization seems to rest largely on the increase and sophistication of output.
However, the individual features of the de-industrialization argument need
disentangling and disaggregation, especially in the light of the complexity of the
empirical data produced by regional studies. One part of the case for a decline in
industry – the quantitative aspects of output and productivity – seems considerably
weakened, or at least complicated, as a basis for argument. It is necessary, though, to
examine the wider implications of the argument that India was ‘de-industrialized’.
Logically, this would surely indicate a falling off of living standards in substantial
sectors and a decline in employment. In other words, ‘de-industrialization’ also implies
the tying together of structural changes in industry with growing immiserisation. In
the case of textiles, such a historical correlation seems fairly plausible, though one needs
to keep in mind here the qualification that it is unsound to extrapolate too much from
such limited sectorial information.
 14 Aditya Sarkar| The Debate on Deindustrialization
14

V
Threading together the different critiques of the notion of de-industrialization,
Tirthankar Roy has sought to evolve a different model of understanding the colonial
economy (though he frequently evinces distrust for the use of colonialism as an
analytical category). His motives are fairly transparent, and he spells them out in no
uncertain terms: he seeks to develop a coherent non-Marxist account of the reasons for
India’s historical under-development, and to tie the changes in economic relations and
international trade in the nineteenth century to factors other than colonialism and
exploitation. In so doing, he touches directly on the issue of de-industrialization, and
offers to supply an alternative grid of understanding, one that often sounds like a pure
inversion of the ‘left-nationalist’ positions. It would not be unfair to the structure of Roy’s
argument to describe it as a ‘things-were-not-as-bad-as-you-claim’ statement. However,
it would be unfair to the real insights contained in his empirical investigations, and the
implications of these go well beyond any simple ‘neo-imperialist’ claim. Even when
disagreeing with Roy, it is possible to recognize that his work calls for a paradigm shift
in the study of traditional small-scale industry.

The most sophisticated statement of Roy’s position is contained in his 1999

book, Traditional Industry In The Economy Of Colonial India. However, his other essays,
his textbook on the colonial economy, and the 1993 monograph Artisans And
Deindustrialization also articulate much the same argument. In the text of his Thorner
Memorial Lecture in 2000, Roy has spelt out his disagreements with the proponents of
de-industrialization with lucidity. He cautions against extrapolation from textile data
alone, and he also suggests that the early nineteenth century provides a mixed picture,
one that it is unsafe to generalize on the basis of. He calls for attention to the
segmentation of the market, and argues that obsolescence (caused by relative
technological backwardness and cost-inefficiency in a context of international
competition) affected some sectors of the textile industry, while a wide range of crafts
suffered no obsolescence of any kind, but rather a ‘commercialization’. Obsolescence was
predominant in the early nineteenth century, and commercialization in the second half
and in the twentieth century. He takes the decline in handloom weaving through much
of the nineteenth century for granted, and argues that what needs to be explained is not
this natural reaction to the opening up of the economy, but rather the recovery and
survival of handlooms after 1880. This, he suggests, was a function of the comparative
advantages of hand and power weaving in different segments of the market, for different
kinds of cloth. Handlooms were better equipped to make certain kinds of traditional
garments for which there was considerable demand. In this process, the handloom
sector witnessed a growth of capital accumulation and intensity, the development of
qualitatively new relations of production, new investments and technological change.
Naturally, the less cost-efficient and productive weavers were pushed out of
employment’ others, however, managed to innovate successfully and meet the
challenges of the twentieth century. (One is tempted, on Roy’s behalf, to clarify his point
India as a Colonial Economy

with the observation, ‘Naturally, only the fittest survive!’, which is as good a summary
as any of his economic philosophy). Productivity increased, as did income, even as
employment fell. The data for the productivity and the export of handloom cotton and
silk goods, hides, skins, carpets and rugs from the early twentieth century all bear this
out.

Roy points to certain specific changes that defined commercialization in the

Indian economy: long-distance trade expanded, regional markets integrated on an
unprecedented scale, and contract law and private property rights came to characterize
all economic relations. The important point here is that these processes held good as
much for traditional industry as they did for modern industry. Production on contract
for distant markets replaced earlier, non-commercial forms of production. In a context
dominated by long-distance trade and the integration of markets, access to working
capital and reliable information were indispensable. However, only a small number of
entrepreneurs could command these resources, and they consequently expanded their
business and managed to exert a tighter control over manufacturing processes.
Capitalist relations thus evolved within the framework of traditional industry. Both
part-time employment and household involvement in small-scale production declined, as
specialization became the order of the day for labourers. Thus, argues Roy, there were
‘segments of decline and segments of growth’ within traditional industry in India, as
was the case with all successfully industrializing nations. There was a genuine impetus
and potential for growth from within the modernized and transformed ‘traditional’
sector. This was ultimately foiled not by colonialism, but by internal factors like the
rapid growth of population and the excessive protectionism of government policy after
independence.

Roy’s most significant arguments concern the changes in the organization of

production, and the redefinition of the form of the transaction between employer and
worker. He distinguishes between ‘spot-transactions’, where there is no prior agreement
on prices and quantities, and contracts, where there is, and where, consequently,
relations of production approximate much more closely to wage-labour’s negotiation
with capital. Long-distance trade is often a concomitant of the latter. The colonial
period, Roy points out, witnessed the growing contractualization of the employment
relation, and the marginalization of other, non-contractual relationships. The leather
industry saw one of the clearest breaks with the past, in that unskilled rural labour was
replaced by wage-labour (under equally hierarchical conditions) in urban factories; the
craft itself came to be relocated in towns. Other forms of organization, like karkhanadari
or teamwork, which characterized diverse kinds of production ranging from Benaras jari
to Moradabad brass, displayed more formal continuity. However, there were significant
structural changes here too under the impact of long-distance trade, as there came to be
a clearer division between capital and labour. Production processes in small-scale
traditional industry also changed, in that the division of labour was more clearly
 16 Aditya Sarkar| The Debate on Deindustrialization
16

articulated. Crafts – including coarse ones like leatherwork – came to be highly

specialized with the growing reach of the world market in leather. Separated from
agriculture, leather-workers also tended to bifurcate in terms of their tasks: flaying,
tanning and manufacture were separated from one another. Simultaneously, there were
important changes induced in technologies of production. Chemical processes in tanning
and brassware were changed, power plating replaced hand plating of wires in jari, and
the fly-shuttle and the warping mill were introduced.

The arguments made by Roy can be tackled at two levels. First, the question of
de-industrialization. Here, one comes to the question of periodization. The classic
arguments for deindustrialization were based largely on nineteenth-century evidence.
Roy adduces growth in the small-scale industry, on the contrary, almost exclusively
from the experience of the twentieth century. He tries to project this backward, but in
none too convincing a manner, since he has to admit the fall in employment in the
nineteenth century. Further, although Roy makes out a convincing case for the growth
of productivity in the traditional sector in the twentieth century (and to his credit
diversifies his enquiry beyond textiles), his blithe explanation of falling employment in
terms of a survival-of-the-fittest argument – whereby the inefficient were justly pushed
to the wall – resonates with all the platitudes of contemporary neo-liberalism, where
humane and rights-based notions of welfare are replaced by a fetishization of efficiency
and competition. In other words, Roy’s explanatory mechanism has no place for the
sufferings of those – in both the nineteenth and the twentieth century – who paid the
costs of India’s economic trajectory. To bring this back to the specifics of the debate,
Bagchi’s article (to take an instance) seeks to articulate two arguments for de-
industrialization, one based on quantitative data, and the other on qualitative data like
official descriptions and eyewitness accounts. Roy is able to problematize and cast real
doubts on the legitimacy – or, at least, the continuity – of a quantitative demonstration
of industrial decline, and the changes he brings to light were more ambiguous than the
de-industrialization model can admit. On the other hand, there is no harm in recalling
that the debate, in its roots, was also a debate about the misery of the Indian artisan and
labourer in traditional industry. The collapse of spinning, the contraction of
employment, and (ironically) Roy’s own superbly argued case for the entry of capitalist
relations into traditional industry, all illustrate the reality of the unfreedoms,
constraints and exploitation that working people had to suffer. Evidence of increased
productivity need not be an indication of increased prosperity or freedom for those who produced.
Even as far as the quantitative indicators go, the real qualifications to the thesis of de-
industrialization, in a sense, come from the detailed regional studies that demonstrate
the complexity of the trajectories that the handloom industry followed in the nineteenth
century. Roy’s commercialization hypothesis, while valid and insightful in its own right,
addresses a different period and a different set of data. It does not disprove the fact of
de-industrialization, though it does effectively disprove its permanence.
India as a Colonial Economy

The real strength of Roy’s work is its analysis of the changes in the organization
of production and the emergence of what might be described as quasi-capitalist relations
in the traditional sector of Indian industry. The significance of this is that it necessitates
fresh research along these lines, and thus perhaps heralds a shift away from a
controversy that has begun to look rather tired (Roy’s own Thorner Memorial Lecture
being an indication of its staleness and text bookishness). However, pioneering though
it is, Roy’s studies of production relations are also rather incongruous. A glance at his
bibliography reveals a reference to the English economic historian Maxine Berg, whose
brilliant studies of the transformations of traditional industry and labour in the
eighteenth century in England are quite possibly a major influence on his own work.
Berg had, in the 1980s, demonstrated the significance of changes in production
organization and the flexibility of management practices in the early Industrial
Revolution, and had brought into question the focus on the large-scale factory as the
prime (sometimes sole) digit of analysis in an organizational landscape that was
extremely diversified though quite uniformly exploitative. Roy’s data suggests the
validity of enquiries along similar lines in the Indian context, especially given the
present-day ubiquity (and, in the neo-liberal age, the sudden legitimacy) of brutally
exploitative working conditions and flexible employment practices in an industrial
world dominated by small production units. However, Roy’s own arguments are wedded
too closely to neo-liberal economic doctrine and an unproblematic belief in the free
market to allow much nuance after a point.

De-industrialization, to sum up, has been a controversial issue in Indian

economic history over almost forty years. In the different positions that have emerged
in the academic debate, there seems to be something to be said for – and against – both
sides. The overall picture seems more differentiated than any hard-and-fast position
would allow. It might tentatively be suggested that there was probably genuine de-
industrialization in a number of regions in the nineteenth century, though quantifying
this is difficult. Productivity appears to have risen afterwards, but the contraction of
employment countered this. The handloom sector was not destroyed; contrary to earlier
nationalist arguments, it witnessed a certain region and product-specific growth. This,
however, bears no correlation to the actual living conditions of weavers: in many cases,
the relation may actually have been inverse. Roy’s recent focus on the changes in
production relations lay out a potentially fruitful agenda for enquiry and speculation.
This agenda is hard to realize within the structure of Roy’s own theoretical framework
and certainties, though. The debates on small-scale traditional industry seem likely to
continue and grow. Whether these debates can operate within the context of the de-
industrialization controversy any more is quite another issue.
 18 Aditya Sarkar| The Debate on Deindustrialization
18

Bibliography

1. Daniel Thorner: ‘De-Industrialization in India, 1881 – 1931’ (from Alice and

Daniel Thorner, Land and Labour in India, Lucknow, 1962)
2. Morris D. Morris: ‘Towards A Reinterpretation of 19th Century Economic History’
(IESHR, March 1968)
3. Toru Matsui: ‘On the Nineteenth Century Indian Economic History: A Review of ‘A
Reinterpretation’ (IESHR, March 1968)
4. Tapan Raychaudhuri: ‘A Reinterpretation of Nineteenth Century Indian Economic
History?’ (IESHR, March 1968)
5. Amiya Kumar Bagchi: ‘Deindustrialization in Gangetic Bihar, 1809 – 1901’ (from
De et al, eds. Essays in Honour of Professor S.C. Sarkar, New Delhi 1976)
6. Marika Vicziany: ‘Deindustrialization of India in the 19th Century: A Methodological
Critique of Amiya Kumar Bagchi’ (IESHR, 1979)
7. Amiya Kumar Bagchi: ‘A Reply’ (IESHR, 1979)
8. Konrad Specker: ‘Madras Handlooms in the Nineteenth Century’ (from Tirthankar
Roy, ed., Cloth and Commerce: Textiles in Colonial India, Delhi 1996)
9. Haruka Yanagisawa: ‘The Handloom Industry and its Market Structure: The Case of
the Madras Presidency in the First Half of the Twentieth Century’ (from Roy, ed.
Cloth and Commerce)
10. Sumit Guha: ‘The Handloom Industry of Central India, 1825 – 1950’ (from Roy, ed.,
Cloth and Commerce)
11. Tirthankar Roy: ‘Traditional Industry in the Economy of Colonial India’
(Cambridge, 1999)
12. Tirthankar Roy: ‘De-Industrialization’ (from Alice Thorner, ed., Land, Labour and
Rights: 10th Daniel Thorner Memorial Lectures’, Delhi, 2001)