Informalisation and the social relations of production: Insights from urban Karnataka, 1990-2005

This article draws upon evidence from urban centres in the Indian state of Karnataka in order to explore the impact of informalisation on labour. It is argued that informalisation is best understood by outlining speci ﬁ c changes to the social relations of production. This issue is approached by comparing and contrasting Marxist claims about class formation with ﬁ ndings from an analysis of Economic Census data. The Economic Census has recorded a rise in waged labour in informal enterprises in urban Karnataka and a rise in sole trading in own-account establishments. It concludes by pointing towards the need for further research into the growing employment of ‘informal workers’ in the formal economy.


Introduction
It is widely argued that labour forces in developing countries have undergone a process of 'informalisation' (Breman, 1994;Chen, Vanek & Carr, 2004;Harriss-White & Sinha, 2007;Portes, Castells & Benton 1989).Yet there continue to be competing positions advanced about how to theorise the relationship between informality and labour.Th is article explores examples of the conceptual approaches on off er, critically evaluating claims about the informalisation of labour by focussing on a study of changes to urban employment in the south Indian state of Karnataka.Based on data derived from Economic Censuses conducted in 1998 and 2005, it suggests that formal employment relations have been eroded in the context of both large and small scale production processes, raising important implications for an understanding of informalisation.It is argued that informalisation is best understood as a historical process that aff ects the balance between specifi c social relations of production.
Th e fi rst section of this article explores some conceptual approaches to informality and informalisation.Th e meaning of 'informalisation' and the distinctiveness of the approach pursued in this article are outlined in this section.Th e relationship between the informal economy and the social relations of production is introduced in this section, including the relevance of Marxist analysis, leading to hypotheses about the impact of informalisation on work and labour.Why Karnataka's urban centres provide an instructive regional case study to test these hypotheses is justifi ed.Th e second section analyses state-level Economic Census data for the period 1990 to 2005, approximating the period of neoliberal policy adoption.Th e strengths and weaknesses of these data are outlined.Th is section examines evidence for changes to employment to address the hypotheses raised in the fi rst section.Th e concluding section contrasts this evidence with the conceptual claims underpinning the hypotheses, pointing to the need for further research.

Informalisation, labour and social relations
In this section, it is argued that informalisation is best understood by analysing changes in the balance between specifi c social relations of production.Th e approach off ered here builds upon the insights of important studies of the informal economy since the 1970s.An earlier view of the 'informal sector' as a marginal social and economic phenomenon caused by low economic growth per capita was criticised in the ensuing decades (ILO, 2002a;Chen, Vanek & Carr, 2004;Breman, 1994).Later studies emphasised the expansion of the informal sector and the economic interdependence of activities designated as formal or informal.A 'structuralist' approach claimed, furthermore, that enterprises in the informal sector and informal workers were subordinated to the development of large-scale capitalist fi rms (Portes, Castells & Benton, 1989).Studies began to acknowledge the expansion, permanence and complexity of the informal economy (Chen, Vanek & Carr, 2004).
Th e informal economy has also been characterised by an absence of 'decent work' or jobs recognised and protected by the legal system (ILO, 2002;2002a).Emphasis has been placed upon the need for policies that protect the rights and conditions of informal workers and policies that facilitate access to markets for the self-employed in informal enterprises (Chen, Vanek & Carr, 2004).Th is policy orientation has focused discussion upon those people that run informal enterprises or those who are hired as informal workers.Informal workers can refer either to those hired to work in informal enterprises or in enterprises with characteristics of the formal economy.Within the ILO framework, it is possible, empirically, to capture informal workers employed in both cases.For instance, one measurement used is to subtract the number of formal employees from the non-agricultural workforce, dividing the result between paid workers and the selfemployed.A second method, which includes an estimation of informal workers employed in formal enterprises, takes the fi gure for informal employment as a percentage of the non-agricultural workforce and divides it between employment in informal enterprises and informal employment outside informal enterprises (ILO, 2002a).
Furthermore, rather than seeing labour as being strictly divided between competing sectors, many of these studies have increasingly adopted the view that formal and informal enterprises and workers coexist along a continuum 'with decent work defi cits most serious at the bottom end, but also existing in some formal jobs as well, and with increasingly decent conditions of work moving up the formal end' (ILO, 2002: 4).Th e informal economy has increasingly been understood as a dynamic entity in which workers may move up and down the continuum and may even occupy more than one point in the continuum at a given time (Chen, Vanek & Carr, 2004).

Passing the buck: corporate restructuring and the casualisation of employment
Th is point is reminiscent of Breman's criticism of dualistic accounts of informalisation, originally published in 1976.Breman argued that social relations of production cut across the formal/informal divide (Breman, 1994).Th e industrial relations system in India is itself a case of a policy-based duality that may be misinterpreted as a structural division between sectors of the economy.Based partly on the belief that postcolonial India needed to embark on an ambitious industrialisation drive and partly on the need to protect livelihoods in small-scale, labour-intensive enterprises, this system was designed to regulate labour policies in an 'organised sector' of public sector establishments and non-agricultural private sector establishments employing 10 or more workers.Agriculture and the non-agricultural 'unorganised sector' were exempted from many of the planning and industrial relations laws established to regulate work and employment practices in the organised sector (Hill, 2009).
Furthermore, a distinction should be made between the 'formal economy' regulated by state planning regimes and 'formal regulatory environment' comprising government policies, laws and regulations, many of which are intended to apply to people employed in the informal economy (Chen, 2006: 83).Th is point suggests that there has been a problem regarding not only the absence or relaxation of labour protections in the informal economy, but also in the selective implementation and even systematic evasion of these policies (Hill, 2009).Th e relative decline of employment in the organised sector since the 1980s has highlighted problems with a system of governance that does not apply to about 93% of India's working population (ILO, 2002a;Hill, 2009).
Th is system has also been interpreted as an ongoing attempt to protect the status of workers who remain in the organised sector's 'islands of privilege ' (Rudolph & Rudolph, 1987: 267).It has been implied that the economic status of organised sector workers, particularly public sector employees, is partially the consequence of a redistribution of income and economic power from workers in the informal economy to those in the formal economy.Workers in the organised sector had accumulated privileges to which, it was implied, they were not entitled (Rudolph & Rudolph, 1987).Th e basis of this argument appears to be that the formal/informal division is, in part, a mechanism of social and economic redistribution as well as a form of stratifi cation.Th is has been challenged in other accounts (Breman, 1994;Holmström, 1984).Revisiting an argument from his earlier study of factory workers in Bangalore (Holmström, 1976), Holmström was clear-cut about this issue: 'Th ere is no dual economy.Th e relatively well-paid are not privileged at the expense of the unorganised sector, and there is no way that making factory workers poorer could help the others' (Holmström, 1984: 322).
Similarly, Breman argued that 'signifi cant diff erences' in income and status existed within the 'labour elite' of the formal economy (for example, between salaried employees in large enterprises, public sector workers and skilled blue-collar workers) which may be overlooked by scholars who 'exaggerate the uniformity of the working elite' (Breman, 1994: 35).In this regard, informality was expressed as a reifi ed concept, unable to capture changes in real social relations fully.Th e solution, he argued, was to distinguish 'diff erent articulated production relations' (Breman, 1994: 107).As well as the existence of a 'labour elite' , Breman pointed to a 'petty bourgeoisie' of self-employed and own-account operators, an unemployable 'lumpenproletariat' and, most importantly, a 'subproletariat' forming the largest section of the urban working population.Th e subproletariat was said to include casual workers in small workshops and the 'labour reserve of large enterprises' (Breman, 1994: 31).Th is was an important alternative to the idea that the formal/informal divide represented a structural divide.While there is little doubt that such a duality has been embedded in India's system of labour laws, this does not mean that it is refl ected in the actual social relations of Indian society.
A decade later, political economists Robert Cox and Jeff rey Harrod developed a typology of labour in developing countries based (using Harrod's term) on 'forms of social relations of production' .As well as an analytical framework, a theorisation of the distribution of power within economic relations was off ered.For Cox, types of production were ordered into a hierarchy, forming a structure of accumulation: 'Th e extraction of surplus fl ows from the subordinate and weaker levels of production to the dominant and stronger' (Cox, 1987: 5).Harrod (1987) argued that at any given moment some forms of social relations within a social formation were dominant and others subordinate.
For labour, this duality approximated the diff erence in power between unprotected workers and 'established' workers.For example, forms approximated by large-scale industry ('bipartite' or 'tripartite') were in a dominant position in relation to the social formation, while those characterised by small-scale petty production ('enterprise labour markets' and self-employed producers) were subordinated to it.Th e logic of this position was not only that workers employed in large corporations had better working conditions and living standards than those working in small scale enterprises: 'Th e lack of privilege of one is the result of the transfers made to the other' (Harrod, 1987: 31).Th is was a means of understanding social relations of production that adopted some concepts of Marxist political economy, while off ering a conceptual framework in which the central social and economic divide was between unprotected workers in small-scale enterprises and participants in large-scale industry rather than between waged labour and capital per se.
Some more recent accounts have also suggested that the inability of capitalist development to provide enough decent work for urban dwellers has changed the composition of urban classes in ways unforeseen in Marxist accounts.For example, urban historian Mike Davis has emphasised the expansion of an informal, slum-dwelling proletariat infl uenced by a range of reactionary political ideas (Davis, 2004: 28).For Indian political economists Sanyal and Bhattacharyya, there has been a 'process of dispossession without proletarianisation or exploitation' (Sanyal & Bhattacharyya, 2009: 35).Th e intersection of surplus labour with slum dwellings has reinforced a range of social and economic problems 'beyond' the workplace, from tenancy rights to health to the shaping of relations within the household.Sanyal and Bhattacharyya argue that such urban changes enhance the importance of the self-employed petty producer.Th ey argue that the locus of exploitation has shift ed from the capital waged labour relationship to a relationship between capital and the assets generated in a household-based 'need economy' in which the self-employed producer plays a central economic role.Making use of David Harvey's concept of accumulation by dispossession, they argue that capital accumulates by appropriating this value while struggles to resist it now have a hegemonic place in labour movements.
Sanyal and Bhattacharyya also hint at a framework similar to Hardt and Negri's concept of immaterial labour.Hardt and Negri too have suggested that the locus of exploitation has shift ed: '[It] is no longer primarily the expropriation of value measured by individual or collective labour time but rather the capture of value that is produced by cooperative labour and that becomes increasingly common through its circulation in social networks' (Hardt & Negri, 2004: 113).Th e implication of this position, like Sanyal and Bhattacharyya's point, is that the value expropriated from the capital-waged labour relationship is no longer central to the reproduction of capitalism; the expropriation of value has taken a form 'beyond' the archetypal workplace.Notably, Harvey too has emphasised the historical shift in capital's means of value appropriation.He has stressed the wide range of struggles against accumulation by dispossession that 'shift the terrain of political organisation away from traditional political parties and labour organising into what was bound to be in aggregate a less focused political dynamic of social action across the whole spectrum of civil society' (Harvey, 2003: 168).Th is argument incorporates a belief that the political and industrial organisation of waged labour in its place of primary employment can no longer be considered the focal point of social change from a Marxist perspective.
Th ese interpretations of the informalisation of labour problematise Marxist analyses of labour and class.As they suggest, Marxism is traditionally based upon ontological and epistemological assumptions in which industrialisation leads to proletarianisation and the concentration of waged workers in the factory system.Many Marxists have, in other words, assumed not only that this process of proletarianisation has continued to occur but that this occurrence is important as the central means of contesting and, ultimately, abolishing capitalist social relations.In part, classical Marxist claims have also assumed that the concentration of waged labour in the factory system would displace the employment of labour in small-scale industry as well as in other social relations of production.Yet, if understood in terms of changes to the social relations of production, the idea of informalisation suggests that a greater diversifi cation of social relations has occurred rather than the process of homogenisation posited in Marxist accounts.
Th is article explores this tension by looking at changes to urban social relations of production in the Indian state of Karnataka between 1990 and 2005.It focuses on the production of commodities in diff erent types of employment.It distinguishes two broad 'social relations of production': labour in which the worker is hired for a wage and labour in which workers are employed without payment of a wage.Each of these types of social relations is then divided into two sub-types of employment.Non-waged labour is divided into own-account operators who make use of non-waged labour in the production of commodities (e.g.family members) and those who do not (i.e.sole traders).Waged labour is divided into labour employed in informal enterprises and labour employed in formal enterprises.Based on the above Marxist narrative, one might expect waged labour to grow more quickly than, and eventually to displace, non-waged forms of labour.Secondly, one would also expect that the growth of waged labour would be increasingly concentrated in the factory system as an economy industrialised.Such a system might also tend to approximate the formal economy, encompassing large-scale industrial enterprises.Th e ensuing section of this article explores these hypotheses by analysing Economic Census data collected in Karnataka.
Karnataka provides a useful case study because it has been relatively fastgrowing, by Indian standards, during the period of neoliberal policy adoption since the late 1980s.It thus has the potential to illustrate the relationship between urban development and labour force transformation during a period of rapid economic change.Economic growth has had an impact upon urban development in the state.For example, Karnataka is India's leading state for software service exports.This is partly the consequence of a policy preference for IT, software services and ITenabled services (ITES).A series of measures have been introduced to encourage service exports to developed countries, to shift economic resources to private capital and to attract foreign investment.Subsidies and incentives for private capital have helped shape a strategy of export-driven growth and, more recently, a 'Special Economic Zone' (SEZ) policy.Forty per cent of all SEZs in Karnataka have been established for IT or ITES (Banerjee-Guha, 2008).The software services and, more recently, business process outsourcing (BPO) industries have clustered on Bangalore's suburban fringe.
Bangalore's large soft ware fi rms are concentrated in the east of the city and Electronic City in the south (Sastry, 2008).As well as transnational corporations such as Texas Instruments, HP and Motorola, large-scale employment has grown in some large indigenous fi rms that entered the soft ware services market as small or medium sized holdings in the 1980s, such as Wipro, Infosys and TCS.Th is growth has been followed more recently by manufacturing growth on the city's fringes, such as the Toyota factory in Bidadi, southwest of the city, and the Volvo truck factory at Hoskote in the northeast.Th ere has also been expansion in the north of the city surrounding the new international airport, which is also the base of Kingfi sher airlines.
By the 1980s, Bangalore was in a relatively strong position to take advantage of an export industry and foreign investment because of its aeronautics industry and other public sector projects established during the period of state-led expansion aft er 1947 (D'Costa, 2009: 633-635).Th e number of workers employed in industrial units grew from 46,000 to 213,000 over three decades until the early 1980s.Of particular note is Hindustan Aeronautics (HAL) which continues to provide industrial hardware for the Indian Air Force.Including HAL, the 'Big Four' (Indian Telephone Industries, Hindustan Machine Tools, Bharat Heavy Electricals) dominated employment in the organised sector in the city until the late 1990s.Th ese fi rms operate in the northern and eastern outskirts of the city.In addition to the public sector, Bangalore also has longstanding, large-scale employment in the garments sector.
Recent changes to the structure of capitalist production suggest a process of displacement and disruptive change.Refl ecting a national level shift , the share of public sector units in organised sector employment has fallen as the new soft ware services, BPO and foreign manufacturing fi rms have expanded.As Karnataka has continued to urbanise, meanwhile, the formal sector's share of employment has fallen (Nair, 2005: 81).Furthermore, despite the continued infl uence of large-scale factories and public sector units in the city, the lack of adequate housing means that perhaps a fi ft h of the city's population lives in slums (Nair, 2005: 119).Slums are an important base for production in the informal economy (Nair, 2005;Schenk & De Wit, 2000).
Finally, Karnataka provides an instructive case study because of the apparent ineff ectiveness of trade union responses to the process of informalisation.RoyChowdhury has outlined the pace of fragmentation among Bangalore's unions following serious industrial defeats in the early 1980s and the current problem of 'independent' unions 'concerned exclusively with fi rm-level issues, which typically arise within the factory or the shop fl oor, and not on broader issues related to industrial wages, conditions of work, and so on' (RoyChowdhury, 2008: 33).While RoyChowdhury situates the problem in the politics of the labour movement, she also questions whether it might be 'determined by a deeper structurality' since 'the objective basis of exploitation as also the face of class confl ict is more amorphous than it was during earlier phases of industrialisation where manufacturing capital met labour face to face on the shop fl oor' (RoyChowdhury, 2008: 36).An exploration of changes to social relations of production in urban Karnataka, and the relationship between these changes and the concept of informalisation, is a useful context in which to help address this question.

Employment trends in urban Karnataka, 1990-2005
Th is section uses Economic Census (EC) data to outline changes to the social relations of production and the sub-types of employment discussed in the previous section.EC data has been collected by state governments and collated nationally in 1977, 1980, 1990, 1998 and 2005 in order to measure changes in employment in the unorganised and organised sectors.Complete datasets for the sub-types of employment discussed here are available only for the 1998 and 2005 EC reports, allowing the researcher to capture some changes in employment for the time series, 1990-2005 and, more completely, for the period between 1998 and 2005.Unless otherwise stated, this section refers to non-agricultural economic activity in urban areas based upon EC data collected by the Government of Karnataka in 1998and 2005(Government of India, 1998;Government of Karnataka, 2000;Government of India, 2008;Government of Karnataka, 2008).
Importantly, the urban centres of Karnataka amount to more than the employment and economic activity of Bangalore, which accounts for about a third of the state's total urban population.1 Th e EC in Karnataka pertains to all urban centres in the state as well as its capital.It relates to urban centres across Karnataka, including 'class one' cities (i.e. with a population greater than 100,000 people) such as Mysore to the southwest of Bangalore, the municipality of Hubli-Dharwad and the city of Belgaum in northwest Karnataka, Bellary and Kolar in eastern Karnataka, Bijapur and Gulbarga in northern Karnataka, Mangalore in the southwest, Davangere in central Karnataka and Udupi on the western coast.It also refers to employment and economic activity in a large number of smaller 'class' towns across the state.However, much of the historical and geographical evidence of changes to the structure of labour forces analysed in this chapter come from studies of Bangalore.Most of the historical evidence regarding labour in Karnataka's urban areas is focused on the state capital.While the EC allows 1 According to a calculation of urban population data from the 2001 Census of India, Bangalore accounted for 31.74percent of Karnataka's urban population.Total urban population refers to the summation of the populations of all towns recorded in the Census (author's calculations based upon data from Government of India, 2001).us to disaggregate some district-wise data, this diffi culty can be partially addressed by looking for inconsistencies between fi ndings from the descriptive data in the EC and fi ndings from historical accounts.It will be demonstrated how certain information gleaned from the EC is corroborated by such analyses, i.e. the points where they are consistent and the points where the historical evidence deviates from the data.
EC data are based upon surveys of all households and fi rms that produce commodities ('establishments'), excluding all crop production and plantations.Th e census measures the number of establishments and people engaged in two important categories: 'own-account establishments' , which are 'normally run by members of the household' , and 'establishments with hired workers' that hire at least one paid worker 'on a fairly regular basis' .'Hired' is equated with waged or paid (Government of India, 2008).Th ese categories are used as proxies for the two types of social relations of production discussed in the previous section.Own-account establishments (OAEs) refer to nonwaged forms of labour (either sole traders or own-account operators who employ non-waged labour) while establishments with hired workers (EHWs) refer to waged forms of labour (workers employed in both the organised and unorganised sectors).Th e EC also makes a distinction between fi rms based on the number of people employed per establishment.For instance, it distinguishes directory establishments, which employ six or more workers, from non-directory establishments, which employ fi ve or fewer workers.Th e implications of changes to employment in fi rms by the number of people employed per establishment will be discussed below.
Indian statisticians have encountered some problems converting survey data into the reports of the EC.Observing and recording the nature of an 'establishment' may be diffi cult.For the purposes of the EC, an establishment is engaged in 'predominantly one economic activity… at a single physical location under single ownership control of a fi rm or enterprise' (Government of India, 2004: 9).However, there may be more than one activity undertaken within a 'physical location' .Th e activities of a single establishment may also be undertaken in diff erent locations.Data collectors have attempted to address this problem by hoping that each of the 'component technical units' is recorded as a separate establishment or, if not, that the activities of an enterprise can be reduced to the single activity that provides it with the highest revenue (Government of India, 2004: 9).In other words, the survey process must classify the economic activity of each establishment as a single endeavour in order for it to 'fi t' with national industry classifi cations.Th is means that the activities of establishments with diverse industrial interests may be distorted by the survey process.
Furthermore, the surveys rely on the self-reporting of establishment owners in conjunction with the observations of data collectors.Th ere are grounds for suspecting some distortion of the data because of false reporting.For instance, the unorganised sector in India is, by defi nition, not registered under legislation such as the 'Factories Act 1948' .While some labour protections are meant to apply to all establishments, there is evidence that these are systematically evaded (Harriss-White & Gooptu, 2000).Furthermore, there may be an incentive for large registered fi rms to under-report or reclassify their workforces.For example, large fi rms may distort employment musters in order to avoid anti-retrenchment rules under the Industrial Disputes Act, 1947(Hill, 2009).
In relation to the EC, these problems refl ect its reliance on an enterprise-based defi nition of informality.For instance, the EC data exclude some non-standard forms of labour including unpaid domestic labour.Unfortunately, they do not measure the use of casual, intermittently-employed or irregular labour.Th is means that the data do not allow us to count the number of informal waged workers employed in organised sector establishments (i.e. in establishments with 10 or more hired workers).Th is problem is reminiscent of criticisms of enterprise-based defi nitions in which informality is 'seen as a characteristic of enterprises, not of employment relations' (ILO, 2002a: 12;Hussmanns, 2008;Narayana, 2006).However, trends in the EC data can reveal underlying changes if they are treated with due caution.Th e strengths and weaknesses of the data are assessed in the fi nal section of this article.Th e data are able to capture some important changes in social relations of production.
Th is section proceeds as follows: fi rst, it looks at the trend in the employment of waged labour; second, it looks at changes to employment in OAEs (including the labour of sole traders and the participation of women); third, it looks at trends in small-scale production (the unorganised sector); and fourth, it analyses trends in large-scale production (the organised sector).

Prevalence of waged labour
Waged employment increased between 1998 and 2005, both in absolute terms and in relation to non-waged employment in OAEs.Firstly, most workers in Karnataka were employed in EHWs in 2005 (Figure 1).Sixty six per cent of all workers were hired.Seventy-nine percent of all workers in the state were employed in EHWs.In other words, 79% of all waged workers and self-employed workers worked in EHWs.In this regard, the concept of 'hired' refers to people employed for a wage.While there are people who work for others in OAEs (i.e. they are not self-employed), they are not regarded as hired workers in the EC, i.e. they are assumed not to be remunerated with a wage.Th is distinction refl ects the enterprise-based defi nition of informality used in the EC.According to the EC, hired employment was proportionally larger in Karnataka than the national average, in which 58% of all workers were hired and 69% of all workers employed in EHWs.Th ere was an absolute fall in employment in OAEs in Karnataka during the same period (Figure 2).Th is decline in employment was moderated only by a rise in sole trader activity (ownaccount establishments with one person 'usually working').In addition, the larger the number of people employed per OAE ('employees size category'), the greater the percentage decline (Figure 3).Th ese data are consistent with the proposition that people have been drawn out of non-waged forms of labour in OAEs into waged labour as the state's economy has grown.Th is result is diff erent from national-level data which indicate a 3.75% rise in employment in OAEs over the same period.However, there was an absolute decline in employment in OAEs in which six or more people were usually working, possibly indicating a similar trend (at a national level) in which non-waged workers were increasingly seeking waged employment over this period.Th e number of people employed in EHWs grew by 495,000 to approximately 2.3 million between 1998 and 2005 while the number of sole traders grew by 136,000 to 413,000.Overall, the number employed in OAEs fell by 35,000 to 623,000.Th is evidence suggests that waged employment is the most prevalent social relation of production in urban Karnataka.Th e number of sole traders grew faster between 1998 and 2005, but this group remains numerically less signifi cant than waged workers.

Changes to employment in own-account establishments
Falling employment in OAEs for the period 1998-2005 does not mean that home-based commodity production is no longer important.Economic restructuring appears to have compelled household members to seek waged employment as new job opportunities have become available and possibly also as household income has become more unpredictable in this period.Yet non-waged labour in OAEs, oft en based in households, continues to coexist with waged labour outside the household.Most production in OAEs in 2005 occurred in the retail trade, manufacturing, repairs and 'community, social and personal services' .Th e retail trade accounted for 44% of all employment in OAEs, representing about 272,000 workers.Women continued to play an important role in employment in OAEs.
For example, for manufacturing, women represented 39% of workers in OAEs against 23% in EHWs.Th is proportionally greater role of women in home-based manufacturing suggests that they continue to play an important part in household commodity production alongside a traditional role in childrearing and other forms of domestic labour.
In most industries, however, female participation is higher in hired employment than in household production.Nationally, this stood at 16% and 12%, respectively, in 2005.Female participation is generally higher than this in Karnataka and continued to rise between 1998 and 2005.Th e participation rate is similar, however, for EHWs and OAEs, at 20% and 19% respectively.Th is suggests that it is simplistic to talk of a linear process in which women are drawn out of household production into paid employment.It is more sensible to say that there is a rise in waged labour in which women play an important and growing part alongside a continued role in home-based production.

Waged labour in small-scale establishments
Although waged labour accounted for the largest share of the change to employment between 1998 and 2005, most of this change was captured in smaller-scale establishments employing less than 10 workers.Forty one per cent of workers in EHWs in 2005 were employed in establishments with between one and fi ve workers inclusive, while 15% were employed in establishments with between six and nine workers inclusive (Figure 4).Th ere was also signifi cant employment in establishments with 30 or more workers: 11% in the 30-99 size category, 4% in 100-199, 7% in 200-499 and 10% in the 500 plus size category.Th is is broadly similar to the national distribution.National level data also revealed a trend of falling employment per establishment between 1998 and 2005.Following this trend, 93% of the growth in the number of EHWs in Karnataka between 1998 and 2005 was represented by establishments with between one and fi ve workers inclusive.Establishments with two people usually working accounted for 86% of this growth.A further 12% represented establishments in the 6-9 size category.
Establishments with between one and fi ve workers inclusive accounted for 52% of the change in employment between 1998 and 2005 (Figure 5).Establishments with two employees accounted for 28% and the 6-9 category accounted for a further 20%.Th e larger share for waged employment in small-scale enterprises may be linked to the low rate of employment growth in the state.Employment growth has been much lower than output growth in Karnataka.While Net State Domestic Product (NSDP) grew by an annual average of 6.86 % in the 11 years to 2004-2005(RBI, 2008)), employment in all establishments grew by just 0.01 % for the period 1990-1998 and 2.44 % for 1998-2005.In other words, relatively capital-intensive economic growth may have played a role in the defi cit of permanent jobs in large-scale establishments, encouraging job seekers to settle for paid work in informal enterprises.

Waged labour in large-scale establishments
Although small scale establishments accounted for the biggest share of employment growth in Karnataka between 1998 and 2005, employment growth data by size category reveal two further important fi ndings.Firstly, there has been a decline in fi rms employing between 10 and 25 people.Secondly, there has been signifi cant growth for fi rms in the 200-plus employees size category (cf. Figure 5).Data showing the growth of large-scale employment is consistent with the growth of new large scale industry in the manufacturing, IT and ITES sectors from the 1990s onwards.With a fi ft h of workers employed in establishments with 100 or more employees in 2005, it is perhaps also consistent with the shift in the share of large-scale employment away from the still large public sector workforce in aeronautics and heavy electrical equipment towards these new private sector fi rms.Th e reason for the decline in employment in medium sized establishments (10-25 workers) is not immediately clear.It is possible that it represents producers driven into smaller scale production, fi rms driven out of business or even the expansion of more competitive or successful fi rms.Th e latter hypothesis is consistent with the experience of IT and ITES fi rms such as TCS, Wipro and Infosys that began as small 'start up' holdings in the 1980s and have successfully grown into large transnational corporations.Th ese data could also represent a combination of these processes.Growth trends are clearer if we look at percentage change data for the number of EHWs and employment therein for the period from 1998 to 2005.While growth among the number of establishments in the 1-5 employees size category is signifi cant, it is less than growth in the 30-99 range (243%), followed by the 200-499 range (89%).Furthermore, size category distribution is skewed towards the larger end for some major activities.
For fi rms engaged in 'fi nance, insurance, real estate and business services' , there was 314% growth in the 30-99 range, 263% in 200-499 and 167% for the 500 plus range, again partly refl ecting the growth of large IT and ITES fi rms (captured by the activity, 'business services').Data for the percentage growth of 'employment' in EHWs reveal a dualism (Figure 6).While employment growth for establishments with between one and fi ve workers inclusive was 38% and growth in the 6-9 category was 41% between 1998 and 2005, it was higher for establishments in the 200-499 (95%) and 500 plus categories (59%).Evidence for employment growth by district also reveals that the larger the size category, the greater the concentration of employment in Bangalore.Bangalore accounted for 47% of the total rise in employment, 74% of the growth in the 100-199 size category, 61% of that in the 200-499 size category and virtually all growth in fi rms with 500 or more employees.
Although there continues to be an interaction between non-waged forms of employment in OAEs and waged labour in EHWs, the analysis in this section shows that waged labour has accounted for the largest proportion of the recent change in employment in urban Karnataka.For the period between 1998 and 2005, the largest share of this change was represented by small-scale establishments (fi ve or fewer workers), although there was a smaller but signifi cant rise for employment in much larger fi rms (200 or more workers).Waged employment has grown in small-scale establishments as well as in large establishments concentrated along Bangalore's suburban fringes.

Conclusion, discussion and further research
Th is article has argued that the process of informalisation is best understood by outlining changes to specifi c social relations of production in developing countries.It has also problematised this argument in the context of the association between the concept of 'social relations of production' and Marxist analysis, and the latter's assumptions regarding the growth of waged labour in the factory system as capitalism develops.Based on these assumptions it has off ered two hypotheses: fi rstly that waged labour will grow more quickly than non-waged forms of labour as economies grow and, secondly, that waged labour will be increasingly concentrated in an industrial factory system that approximates employment in the formal economy.Th e EC evidence for urban Karnataka shows that waged labour became far more important as a mode of employment, particularly during the most recent period (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005).Nonwaged forms of employment in OAEs declined during the same period.Th is evidence suggests a polarisation in employment in which sole traders become more important in production within OAEs at the same time as signifi cantly greater numbers of people were drawn into waged employment.Th us, the evidence from Karnataka's urban centres is consistent with the fi rst hypothesis.
However, the EC data appears to contradict the second hypothesis.Th ere has certainly been signifi cant growth in large-scale establishments in the formal economy.Growth in establishments with 200 or more hired workers was higher between 1998 and 2005 than in establishments with less than 10 hired workers.However, the lion's share of the rise in waged employment was captured by informal enterprises, i.e. establishments employing fewer than 10 workers.Th e process of informalisation in Karnataka has thus taken the specifi c form of a rise in waged employment in informal enterprises alongside the growth of sole trading in OAEs.Th e low rate of employment growth in the state between 1990 and 2005 may have encouraged a greater number of people to seek paid work in informal enterprises due to the lack of opportunities for permanent employment in the formal economy.
However, although the EC data show that waged employment has grown more quickly in larger establishments, one cannot conclude that this represents commensurate growth in the formal economy.Th is is because these data tells us little about the conditions and rights of workers employed in these establishments.As outlined in the fi rst section of this article, a defi nition of informality based upon employment relations is superior to an enterprise-based defi nition, like the one institutionalised in the Indian system of national accounts, including the classifi cations used in the EC.It thus remains possible, perhaps even likely, that the growth of waged employment in Karnataka's large establishments is entirely consistent with broader claims about the informalisation of labour markets.Casual, subcontracted and irregularly employed workers in large establishments may represent a large reserve of labour employed 'off the books' and, consequently, invisible to, or unprotected by, state regulations.
Th e nature of the Indian industrial relations system suggests that there is an incentive for large-scale employers to restructure their employment records in order to avoid the legal obligations required in the organised sector.Employers may subcontract out work via intermediaries, force people to work without employment records or below the minimum wage, misallocate work titles, restructure their accounts, reorganise labour forces into smaller groups or pool small groups of permanent, casual and contracted workers together in order to enable easier restructuring or the provision of inferior wages and employment conditions.Government of India data for the period 1999-2005 show that there was a 42% increase in the number of unorganised workers employed in the organised sector (Hill, 2009).In this regard, Breman's original critique of the informal sector seems especially relevant.He argued that those workers denied permanent employment protected by legislation form 'a very substantial category of workers which overlaps the border separating the formal from the informal sector... [in which] the vast majority of the working population outside agriculture is concentrated' (Breman, 1994: 115).Exploring this hypothesis in the context of Karnataka requires greater research at a micro-level, such as surveys and observation of offi ce and factory workforces.
Th is hypothesis is complicated by the evidence that a signifi cant section of the growth in large-scale establishments has been accounted for by large soft ware services and ITES fi rms concentrated in Bangalore's urban fringes.Th ere is evidence that these fi rms tend to off er wellabove-average wages and tend to employ people from a relatively narrow social background.Based on surveys in Bangalore, for example, the labour force in IT and ITES is less heterogeneous in character than the wider labour market.Although there is a higher proportion of women employed in these sectors, workers tend to have a university or college education, come from an urban and upper-caste background and, unlike the wider workforce, are usually fl uent in English (Upadhya &Vasavi, 2006).Th is suggests that the sector does not absorb poorer groups into employment opportunities with higher pay.On the other hand, there is also evidence that workers employed in call centres (i.e.ITES) experience problems of long hours, job insecurity and excessive work intensity, presenting some openings for union organisation (Taylor, Scholarios, Noronha & d'Cruz, 2007).
More signifi cantly, there is also evidence that there has been growth in Karnataka in large-scale establishments in which the distinction between formal and informal employment may be less important.Th e garment manufacturing industry, for example, creates an overlap between characteristics of formal employment in large scale work units and informal employment relations (RoyChowdhury, 2008).Th us, fully exploring the implications of informalisation for labour in Karnataka requires further study of such grey areas between the informal and formal economies.Th is means that we must go beyond the limitations of the EC and other data sources that rely upon an enterprise-based defi nition of informality.
Th e evidence in this article suggests that informalisation in Karnataka's urban centres is represented by a rise in the employment of waged labour in informal enterprises and a rise in sole trader activity in own-account establishments, and points towards the need for further research into the growing employment of informal workers in the establishments of the formal economy.Arriving at this conclusion underscores the importance of conceptualising informalisation as a process that aff ects change in the balance between specifi c social relations of production.© Th omas Barnes, 2011

Figure 1 :
Figure 1: Employment in establishments by type, 2005

Figure 2 :
Figure 2: Employment in own-account establishment by employees size category, 1998-2005

Figure 4 :
Figure 4: Employment in establishments with hired workers by employees size category (distribution %), 2005