Chris Harman


The workers of the world

(Autumn 2002)

From International Socialism 2:96, Autumn 2002.
Later published as a pamphlet.
Copyright © International Socialism.
Copied with thanks from the International Socialism Archive at
Marked up by Einde O’Callaghan for the Marxists’ Internet Archive.

The eruption of the anti-capitalist movement worldwide over the last two and a half years has thrown up lots of old questions in new forms. The most central is the question of agency – of what forces exist that are capable of taking on the system and transforming the world.

For classical Marxism, the answer was simple. The growth of capitalism was necessarily accompanied by the growth of the class it exploited, the working class, and this would be at the centre of the revolt against the system. But today this view is being challenged from a number of directions, not merely from the ‘Third Way’ social democrat right, but also from some of the best known spokespeople of the anti-capitalist movement. In particular the notion of the ‘multitude’, developed by Michael Hardt and Antonio Negri [1], is widely seen as a more relevant category than that of the ‘working class’.

This is not the first time people have challenged the position of classical Marxism. It happened several times in the course of the 20th century. The spread of the revolutionary movement from Western Europe and North America to the rest of the world confronted people with the stark reality that the working class was not ‘the overwhelming majority’ of humanity, but still a small minority. This led one socialist trend in Russia, the Narodniks, to put their faith not in the workers but in the peasants. It led another, the Mensheviks, to declare that the Russian Revolution could not be a proletarian, let alone a socialist, revolution. Lenin, by contrast, insisted on the central, independent role to be played by the working class, even in the years before 1917 when he argued the revolution would not produce a workers’ state, but a ‘democratic dictatorship’. Trotsky went further and adopted a position which was effectively accepted by Lenin in the course of 1917: the workers had to take power, although their ultimate success in moving to socialism depended on the spread of the revolution to more advanced countries.

This did not end the argument. It arose again with the growth of revolutionary movements throughout what we now call the Third World in the wake of the Russian Revolution. The Stalinised Comintern from the mid-1920s put its faith in the ‘national bourgeoisie’ of colonial countries to be allies of the international revolution. Then in the 1950s and 1960s, after the victory of revolution in China and Cuba, the near-dominant view in the left internationally was that the peasants were the main hope for revolution. At this time fashionable sociologists declared that sections like car workers had become ‘bourgeoisified’ [2], and this was accepted by many on the left who saw them as ‘labour aristocrats’. [3] Only after the central role played by workers in the May events of France 1968 did this begin to change – and even then the examples of China, Cuba and Vietnam were seen as decisive everywhere outside Western Europe, the US, Canada, Australia, New Zealand and Japan.

Once the struggles of the late 1960s and early 1970s began to subside, the questioning of the role of the working class resumed. The French socialist André Gorz wrote a book whose title, Farewell to the Working Class, typified the attitude of a growing section of the left. In Italy ‘autonomist’ thinkers began to portray workers with secure jobs as a privileged group, cut off from the ‘real’ proletariat. Everywhere academics who had played with Marxism now began to insist that gender and ethnicity were as important, if not more so, than class – and these categories in turn were drowned by a deluge of competing ‘identities’.

The rise of the anti-capitalist movement has led people as varied as Susan George, James Petras, Naomi Klein, Michael Hardt and Toni Negri to challenge the enormous fragmentation associated with ‘identity politics’. But none of them has put the working class centre stage. Identification with the Zapatistas has led to a renewed emphasis on the role of peasants and indigenous peoples. The normal response to the fragmentation of identity politics has been to call for alliances between the fragments, with none of them playing any strategically central role. In Naomi Klein’s No Logo the working class is presented as decisively weakened by the spread of globalisation, ‘a system of footloose factories employing footloose workers’, with a ‘failure to live up to their traditional role as mass employers’. [4] Hardt and Negri’s Empire tried to theorise the notion of a new agency, or a ‘new social subject’, ‘the multitude’:

In a previous era the category of the proletariat centred on and at times was effectively subsumed under the industrial working class whose paradigmatic figure was the male mass factory worker ... Today that working class has all but disappeared from view. It has not ceased to exist, but it has been displaced from its privileged position in the capitalist economy. [5]

For them the agency of change becomes the ‘multitude’ – a sort of updated rainbow coalition of fragmented identities:

This is a new proletariat and not a new industrial working class ... It becomes ever more difficult to maintain distinctions among productive, reproductive and unproductive labour. As labour moves outside the factory walls it becomes increasingly difficult to maintain the fiction of any measure of the working day and thus to separate the time of production from the time of reproduction, or work time from leisure time ... The proletariat produces all in its generality everywhere all day long. [6]

The key demands organising this multitude no longer concern the length, intensity or payment of labour, but centre on ‘a social wage and a guaranteed income for all’, since ‘the social wage extends beyond the family to the entire multitude, even those who are unemployed, because the entire multitude produces, and its production is necessary from the point of view of social capital’. [7]

You find echoes of such notions in all sorts of writings emerging from the new movement. Thus the Argentinian philosopher and writer Leon Rozichtner has seen the neighbourhood asembleas of Buenos Aires as embodying the alternative to capitalism:

Previously, in the age of productive capital, one could think, with Marx, that the place for the radical confrontation by the exploited classes was the factory and the union. Now, when the transformations of finance capital have come to dominate nations and the whole productive apparatus and its services, the field of exploitation has spread to cover all aspects of everyday life: its power has penetrated and dissolved social relations, dispersing people, making personal interests antagonistic to the collective social power; the factory has ceased to be the only place where the social power of resistance is born. The field of expropriation has extended from the factory to the whole of society ... It is not only the industrial working class that can stop the functioning of this infernal social machine: it is the whole society which is building up inside it the power required to confront globalisation. [8]

Certain changes in capitalism in the last quarter of a century seem to give credence to such views. The restructuring of production internationally has led to the contraction of certain industries and the shift in the locus of others.

But the outcome is very different to that put forward by Hardt, Negri and the rest. Far from the working class internationally contracting, it has continued to grow. And the distinctions between this enlarged working class and other oppressed groups, far from becoming marginal, are as central as when Lenin and Trotsky polemicised against the Narodniks.

The worldwide picture

’The working class [exists] as never before as a class in itself ... with a core of perhaps 2 billion people’, around which there are another 2 billion or so people with lives which are ‘subject in important ways to the same logic as this core’. So I wrote three years ago. [9] A detailed study of the world’s workforce by Deon Filmer shows my figures to be roughly correct. [10] He calculated that 2,474 million people participated in the global non-domestic labour force in the mid-1990s. Of these around a fifth, 379 million people, worked in industry [11], 800 million in services,[12], and 1,074 million in agriculture. [13]

Each sector of the labour force includes people who employ others (big capitalists and the petty bourgeoisie), people who are self employed, and people who do waged (or salaried) labour for others.

In agriculture a very large proportion of people continue to work on their own account on plots of land they owned or rented. The proportion of cultivators dependent wholly on waged labour worldwide is still proportionately small – according to Filmer’s figures only about 8 percent, and only 3.6 percent in the ‘low income’ economies. However, he does not provide figures for those partially dependent on wage labour – and we know this figure to be very big in China (see later) and South Asia, which between them account for at least half the world’s peasantry.

The majority of people worldwide in the industrial and service sectors do get wages or salaries – 58 percent of those in the industrial workforce and 65 percent of the services workforce. But this still leaves a very big proportion who are self employed or involved in family labour.

Filmer concluded that the overall number of employed people worldwide was about 880 million, compared with around 1,000 million people working mainly for their own account on the land (overwhelmingly peasants), and 480 million working for their own account in industry and services.

The figure for ‘employed people’ includes some non-workers’ groups as well as workers. There is a section of the bourgeoisie in receipt of enormous corporate salaries, and below that the new middle class who get paid more value than they create in return for helping to control the mass of workers. These groups probably amount to 10 percent of the population. [14] That reduces the size of the employed world working class to around 700 million, with about a third in ‘industry’and the rest in ‘services’.

But the total size of the working class is considerably greater than this. The class also includes those who are dependent on income that comes from the waged labour of relatives or savings and pensions resulting from past wage labour – that is, non-employed spouses, children and retired elderly people. If these categories are added in, the worldwide total figure for the working class comes to between 1.5 and 2 billion. Anyone who believes we have said ‘farewell’ to this class is not living in the real world.

The dynamic of class

Filmer’s study provides no more than a still image of something which is undergoing continual change. The last half century has seen two interrelated processes sweeping the whole world. One has been the mass movement of hundreds of millions of people from the countryside to the towns.

Proportion of Population Living in Towns [15]




World as a whole



All developing countries



Least developed countries



Estimates suggest that by 2015, 49 percent of ‘developing country’ people and 55 percent of the world population will live in towns – and one in five in cities of larger than 750,000. [16]

Even in countries people often think of as rural, the urban population can be a majority – 78 percent in Brazil, 73 percent in Mexico, 59 percent in Ecuador and 56 percent in Algeria. Elsewhere it can be enormous – 45 percent in Egypt, 30 percent in China, 34 percent in Pakistan and 27 percent in India. [17]

The spread of urbanisation is associated, necessarily, with the increasing dependence of people on the market for a livelihood. A family of small peasants may be able to feed, clothe and shelter themselves as subsistence farmers, almost entirely out of the direct product of their own labour. City dwellers are unable to do so. They are likely to starve unless they can sell something – their own labour or products of their labour – however meagre. And even in the countryside recent decades have seen the growing importance of production for the market. Agriculture is often supplemented by forms of handicraft or primitive industry: ‘A survey has shown that for 15 developing countries where recent statistics are available, the percentage of the rural labour force engaged in non-farm work was 30 to 40 percent and was rising’. [18]

This trend is very marked in China, where more than 100 million people from peasant households seek at least temporary wage labour in the cities each year:

Since 1980, Chinese peasants ...have sought to improve their income by working in non-agricultural areas, such as local village-township enterprises and family businesses, or by migrating to cities to seek urban employment ... In the 1990s, not including migration within each county, 15 to 25 percent of rural labourers nationwide migrated to cities for six months or longer; 50 percent of them were people below the age of 23 ... At a national level, the proportion of non-agricultural income in the total income of rural households has increased from 10 percent in 1980 to 25 percent in 1985 and then to 35 percent in 1995. [19]

Many of those in a community of peasant households will work at wage labour jobs. Stepping on ‘two boats’ is a Chinese saying that describes people who commit to two jobs in order to secure their social and economic advancement ... for the younger generation, acquiring non-agricultural jobs has become crucial to avoiding the fate of peasant life and escaping rural poverty. [20]

In Egypt a sample of rural households has shown that 50 percent of income came from farming – and 25 percent came from ‘wages outside village’. [21]

Add these ‘semi-workers’ or ‘peasant workers’ to the numbers of people completely dependent on wage labour, and you get a figure which must be somewhere between 40 and 50 percent of the world’s population. In other words around the core of 1.5 to 2 million proletarians there are a similar number of semi-proletarians.

The myth of deindustrialisation

The argument that the working class has disappeared usually rests on superficial impressions about what is happening to the old industrial working class, at least in the advanced economies. So there is much talk about ‘deindustrialisation’, the ‘post-industrial society’, or the ‘weightless economy’.

Restructuring of industry through successive economic crises has certainly caused some formerly central features of the industrial scene in any locality to disappear. At the same time there has been an increased insecurity of employment and a rise in the proportion of jobs which are part time, temporary or on short contracts. But this does not justify the claim that the working class has disappeared.

Take, for instance, the number of industrial workers in the world’s biggest single economy, that of the US. At the end of the 1980s there was much panic in the US about ‘deindustrialisation’ in the face of challenges to US industrial pre-eminence in fields like auto production and computers. But in 1998 the number of workers in industry was nearly 20 percent higher than in 1971, roughly 50 percent higher than in 1950 and nearly three times the level of 1900:

Workers in Industry, US [22]









The number of manufacturing jobs in the US today is as high as ever before in history. ‘Old’ industries have by no means disappeared, or moved abroad. As Baldoz, Koeber and Kraft have noted, ‘More Americans are now employed in making cars, buses and parts of them than at any time since the Vietnam War’. [23]

This is a completely different picture to that painted by Hardt and Negri when they write of the trend towards ‘a service economy model ... led by the United States, the United Kingdom and Canada. This model involves a rapid decline in industrial jobs and a corresponding rise in service sector jobs’. [24]

The Japanese figures are even more astounding. The industrial workforce more than doubled between 1950 and 1971 and was another 13 percent higher in 1998.

Industrial employment has fallen sharply in a number of countries over the last three decades – in Britain and Belgium by a third, and in France by more than a quarter. But these do not represent a deindustrialisation of the whole of the advanced industrial world, but rather a restructuring of industry within it. The number of industrial jobs in the advanced industrial countries as a whole was 112 million in 1998 – 25 million more than in 1951 and only 7.4 million less than in 1971. There is a great danger of looking at the world through British or French glasses, and not seeing what is really happening on a global scale. So Toni Negri’s Italy may not be in the same league as the US or Japan, but the industrial workers have certainly not disappeared. There were 6.5 million four years ago, down only one sixth since 1971. [25]

Industry and services

These figures for industrial employment, it should be added, underestimate the economic importance of industry in general and manufacturing in particular. As Bob Rowthorn has rightly noted, ‘Almost every conceivable economic activity in modern society makes use of manufactured goods ... Many of the expanding service industries make use of large amounts of equipment’. [26]

The small decline in the total industrial workforce is not because industry has become less important, but because productivity per employee in industry has risen more quickly than in ‘services’. Slightly fewer manufacturing workers are producing many more goods than three decades ago. Their overall importance for the economy has not changed. Between 1973 and 1990, output in the advanced OECD countries grew by an average of 2.5 percent a year in industry, only a little less than the 3.1 percent growth of service output. But productivity growth in industry was 2.8 percent a year, in services only 0.8 percent. [27] The industrial workers are as important for the capitalist economy today as in the early 1970s. Glib statements like those of Hardt and Negri about their declining significance could not be more wrong.

But that is not all. The usual distinction between ‘industry’ and ‘services’ obscures more than it reveals.

The category ‘services’ includes things which are of no intrinsic importance to the capitalist production (for instance, the hordes of servants who provide individual capitalist parasites with their leisure). But it has always included things which are absolutely central to it (like the transportation of goods and the provision of computer software). What is more, some of the shift from ‘industry’ to the ‘service sector’ amounts to no more than a change in the name given to essentially similar jobs. Someone (usually a man) who worked a typesetting machine for a newspaper publisher 30 years ago would have been classified as a particular sort of industrial worker (a ‘print worker’); someone (usually a woman) working a word processing terminal for a newspaper publisher today will be classified as a ‘service worker’. But the work performed remains essentially the same, and the final product more or less identical. Someone who works in a factory putting food into a tin so that people can warm it up to eat at home is a ‘manufacturing worker’; someone who toils in a fast food shop to provide near-identical food to people who do not have time to warm it up at home is a ‘service worker’. Someone who processes bits of metal to make a computer is a ‘manufacturing worker’; someone who processes software for it on a keyboard is a ‘service worker’.

The trend in recent years has been towards firms ‘contracting out’ certain operations that used to be carried out ‘in house’ – for instance, catering and security. The result is that jobs once included in the ‘industry’ figure now appear under services. The Engineering Employers’ Federation in Britain has pointed out:

Manufacturing creates a large chunk of service industry through outsourcing of areas such as maintenance, catering and legal work ... Manufacturing could make up as much as 35 percent of the economy – rather than the generally accepted 20 percent – if it were measured using appropriate statistical definitions. [28]

Rowthorn has undertaken a statistical breakdown of the total ‘service’ category for the OECD as a whole. His figures show that goods-related services accounted for 25 percent of total employment in 1970 and 32 percent in 1990. There is a small fall in ‘total goods and goods-related services’ – from 76 percent of all employment to 69 percent. [29] But this is certainly not a revolutionary transformation in the world of work. He points out that in 1990, ‘free standing services’ only accounted for 31 percent of all employment [30], and concludes, ‘Goods-related production is still generating directly or indirectly about two thirds of all employment in the typical advanced economy, despite all the talk about a post-industrial economy’. [31]

The nature of the non-marketed services sector

But even Rowthorn’s figures considerably underestimate the size of the working class – that class whose labour is essential for the accumulation of capital. Many of Rowthorn’s ‘free standing services’ are essential to such accumulation in the modern world. Two in particular are absolutely indispensable for capitalist accumulation today – health provision and the education service.

The core of the health system of any modern capitalist country is concerned with ensuring that the labour force is fit and able to work. It is there to make sure the next generation of labour power is fit and well, and to patch up the members of the present generation if they suffer some ailment that removes them from the labour market temporarily. Even where this health provision takes place through the state, and so is not bought and sold, it is still an indispensable accompaniment to capitalist production.

This is, if anything, even truer of the education service. It grew up in the 19th century as capitalism found it had to train up its workforce to certain basic levels of literacy and numeracy (as well as discipline) if they were to be productive. It expanded through the 20th century to encompass longer and longer years of schooling, as the average levels of skills needed by the system rose. In nearly all countries the main sections of the education system remain in state hands. It does not sell commodities. Nevertheless, it too is indispensable for production. Those who are working within it are working for capital accumulation, even when they do not produce anything that is sold. [32]

The intimate connection between education and capital accumulation is emphasised every time ‘modernising’ governments draw up plans for educational ‘reform’. They make no secret of their view that education (or ‘training’) is an input into industry. It is about producing a mass ‘flexible’ workforce with the average level of numeracy, literacy and IT skills to adapt to the changing needs of capital for labour power as it restructures industry.

The expansion of education is connected to the rise in the productivity of labour in the ‘old’ industrial sector. This is, of course, in part a result of the greater workload forced on each worker. But it is also a product of the workforce having sufficient ‘training’ to adjust to changes. So a recent survey of the British workforce showed that 37 percent of men and 25 percent of women said they required ‘advanced or complex information technology skill levels in the work they perform’, while 51 percent of plant machine operators said they used IT in their jobs. [33] The capitalist class cannot guarantee the existence of a workforce with the requisite level of ‘versatile and adaptable’ skills without a growth in the educational sector. [34] Or, to put it another way, there is a rapid growth in the productivity of those workers who directly produce commodities – and this had led to a reduction in their numbers in countries like Britain and France (but not the US). But this productivity growth depends in part on an increase in the number of workers whose labour increases the productivity of these commodity-producing workers – ie an increase in the number of ‘indirectly productive’ workers.

The number of people employed in health and educational services has grown continuously over the past century, as part of the overall expansion of capitalism. Today there are over 10 million of them in the US (around one in 13 of the workforce) – US capitalism could not function without them. At the same time, however, the capitalist class resents paying more than is absolutely necessary to get their services. As a result, the long term trend is for them to be forced increasingly into conditions comparable to those of industrial and routine office workers. Various work-measurement systems once regarded as exclusive to the ‘old’ industries are now being forced on these sectors – payment by results, assessment and appraisal systems, increased concern with timekeeping, and enhanced discipline codes. Six percent more employees were subject to formal supervision or assessment in Britain in 2000 than eight years earlier. Around 5 percent more were subject to some sort of individual payment by results system. [35]

This does not mean health and education workers are all uniformly ‘proletarian’. Hospitals, schools and colleges are all organised along hierarchical lines. The higher grades are given salaries designed to cement them to the system so that they can be used to impose control on both the lower grades and the working class clientele of the institutions. They are part of ‘the new middle class’ (or even, in the case of those at the top of the elite universities, part of the ruling class). [36] But the bulk of the workforce is subject to continual pressure to work at a capitalist tempo for a level of remuneration determined by the labour market. They are for this reason part of the global working class, even though many continue to regard themselves as superior to the manual working class.

In fact, two related processes are taking place in all ‘advanced’ economies (and many ‘non-advanced’ ones). The traditional manual working class is put under more and more pressure as capital tries to squeeze its direct labour so as to get more profits from it. At the same time, the new ‘non goods-producing service’ working class is subject to proletarianisation as capital sets out to reduce the cost to it of a growing mass of ‘indirect’ labour.

The nature of the service workforce

There is a widespread myth that the ‘service’ workforce consists of well paid people with control over their own working situation who never need to get their hands dirty. So Guardian columnist (and former SDP member) Polly Toynbee writes: ‘We have seen the most rapid change in social class in recorded history: the 1977 mass working class, with two thirds of people in manual jobs, shrunk to one third, while the rest migrated upwards into a 70 percent home-owning, white collar middle class’. [37] So Hardt and Negri claim:

The jobs for the most part are highly mobile involving flexible skills. More important, they are characterised in general by the central role played by knowledge, information, affect, and communication. In this sense many call the post-industrial economy an informational economy ... Through the process of post-modernisation all production tends toward the production of services, toward becoming informationalised. [38]

In fact, however, any proper breakdown of the figures for ‘service’ employment provides a very different picture to this. Some of the most important ‘services industries’ employ overwhelmingly ‘manual workers’ of the ‘traditional’ sort. Refuse workers, hospital ancillary workers, dockers, lorry drivers, bus and train drivers, postal workers are all part of the ‘service’ workforce. And a very big part. In September 2001 ‘distribution, hotels, and restaurants’ accounted for 6.7 million jobs and ‘transport and communication’ for 1.79 million. [39]

In Britain the proportion of people in manual jobs is, in fact, much higher than a third. The Office for National Statistics’ Living in Britain 2000 shows 51 percent of men and 38 percent of women as being in its various ‘manual’ occupational categories in 1998. [40] The figures slightly exaggerate the number of manual workers since the ‘skilled manual’ category includes ‘foremen’ and the ‘non-professional own account’ group, which will knock a few percentage points off the male figure. But against this must be put the huge number of women – 50 percent – in the ‘intermediate and junior non-manual’ categories, where wages are typically lower than in most manual occupations and working conditions often at least as hard. A study shows how much things have changed from the time when clerical workers were regarded as a cut above manual workers of all sorts: ‘By 1978 the average earnings of semi-skilled men had overtaken their clerical counterparts for the first time. In addition, many of the traditional superior employment conditions of clerks have been gained by manual workers’. [41]

In the US in 2001, the total service-related occupations of 103 million people included 18 million in routine ‘service occupations’ with a decidedly manual cast to them (including nearly a million in ‘household services’, 2.4 million in ‘protective services’, 6 million in ‘food services’, 3 million in ‘cleaning and building services’, and 3 million in ‘personal services’). Then there were 18 million in routine clerical jobs, and 6.75 million sales assistants.

The great majority of white collar workers are in fact women from working class backgrounds. In Britain, a third of clerical workers come from manual working class backgrounds, a third from a clerical background and only a third from the so called ‘professional-managerial service class’. [42] Whereas their grandmothers most probably stayed at home after marriage, toiling to bring up the next working class generation, they expect to work all their adult life, combining the toil of paid employment with the added burden of childcare and housework. What is happening is a feminisation of a huge area of waged labour. This is not at all the same as its ‘bourgeoisification’ or even its ‘informationalisation’.

Altogether there are a minimum of 42 million ‘service sector workers’ in manual or routine white collar occupations in the US. These, it should be added, are the occupations that have been expanding most rapidly recently with the ‘creation’ of a mass of low wage jobs. On top of them, many workers in other occupational categories would have been doing work which was little different – for instance, many of the 3.2 million ‘sales representatives’ and the 4.3 million ‘technicians and related support’ workers. So would many of the ‘health assessment and treating occupations’ (83 percent female, unlike the ‘health diagnostic’ category above them which is 75 percent male), and many of the 5.3 million school teachers (75 percent female). [43]

Together these groups constitute well over half the ‘service sector’. Add to them the 33 million workers in traditional manual industries, and you have some three quarters of the US population made up of workers. If the ‘working class’ has ‘disappeared from view’ for people like Hardt and Negri, it is because they have been looking in the wrong direction.

By contrast, the numbers in occupations that are often seen as replacing traditional sorts of work are very small. There were just 2.1 million mathematics and computer scientists in the US in 2001. These figures hardly add up to an ‘informational’ economy, in which manual work is being marginalised. People in many old-style manual and routine white collar jobs may have to use some basic IT skills on top of their old jobs. But the number specialising in these skills is relatively small. In Britain in 2000, at the height of the internet boom:

In spring 2000, there were 855,000 people employed in IT-related occupations, an increase of 45 percent in only five years. The region with the highest proportion of employed people working in IT was London, at 4.8 percent, followed by the South East with 4.4 percent. Together, these regions accounted for 41 percent of all those working in IT ... Northern Ireland, Wales and the North East all had low proportions working in IT, at 1.3 percent, 1.6 percent and 1.9 percent respectively. [44]

Not all those who work for salaries are workers. There is a ‘new middle class’ of people who are salaried but who are rewarded by managements for helping them control the rest of the workforce and paid sums well beyond any value they might create. But this group is a relatively small proportion of the total workforce. I calculated 15 years ago:

The grouping contains only 9.7 percent of men between the ages of 31 and 75. This would seem to coincide with the sort of rough estimate for the size of the ‘new middle class’ one arrives at on the basis of figures for income, qualifications and managerial authority. It is a significant proportion, both of the total population and of the white collar work force. But it is a far cry from anything approaching the traditional manual working class in size or significance. [45]

The most up to date British and US statistics leave that judgement more or less intact.

Flexibility and the working class

A central theme of all those who see the working class as disappearing is that the jobs that remain are so precarious that little remains of the permanent working class organisations and communities that used to exist. The argument has been a continual feature of ‘post-Marxist’ arguments for the last 15 years both from ‘Third Way’ social democrats and those on the ‘autonomist’ left.

We are currently going through the world’s fourth great economic crisis in less than 30 years. Each crisis has involved sudden increases in unemployment – in some cases permanently – and the wiping out of old established centres of production (factories, docks, mines, etc). It has also speeded up enormously the restructuring of industry, not only on a national but also on a regional and global basis. [46] Capital and its apologists have then tried to take advantage of rising levels of unemployment and workers’ feelings of insecurity to remould the lives of the workforce around its own continually changing requirements. Its slogans have become ‘flexibility’ in labouring time, methods of work and labour markets. One of its rallying cries has been that ‘lifetime employment belongs to a past age’. Much academic research has taken such claims to be unquestionable truths. As Raymond-Pierre Bodin, Director of the European Foundation for the Improvement of Living and Working Conditions, has written:

It has become commonplace nowadays to speak of the development of atypical forms of work in the Western economies ... Works highlighting the erosion of the Fordist standard of employment embodied in the permanent full time contract as the main way of organising labour markets and integration into social life are now innumerable ...the term flexibility also seems to correspond to the changes taking place in this field. [47]

But this does not mean that in reality capital has been able to destroy workers’ resistance to such flexibility, or even that it itself can keep accumulating without continually reproducing relatively permanent labour forces within particular workplaces. One recent study for Britain shows:

Many of the commonly held assumptions about today’s world of work need to be seriously questioned. A wide gulf exists between the over-familiar rhetoric and hyperbole we hear daily about our flexible and dynamic labour market and the realities of workplace life. The evidence simply does not sustain the view that we are witnessing the emergence of a ‘new’ kind of employment relations, seen in the ‘end of the career’ and the ‘death of the permanent job for life’. [48]

People often do not see the limits to what capital can achieve in terms of ‘flexible labour markets’ because they lump together quite different forms of employment: part time employment, temporary employment, employment on short term contracts, and self employment on behalf of firms. But part time employment can also be permanent employment – as it usually is among women in Britain. Similarly, those working on short term contracts can find them renewed month after month, or year after year. They lack long term rights and are the first to go when crises hit, but they do not move into and out of jobs all the time in between. Finally, those in genuinely temporary employment may be indispensable to production and be provided on a long term but intermittent basis by agencies which are themselves large firms and dependent on maintaining a permanent pool of labour to supply to other firms.

These forms of employment have been growing in recent decades to different degrees within different industries and different countries – and often within particular regions of countries. So in Spain 35 percent of workers were in ‘precarious employment’ in 1992, in Britain and France 16 percent, and in Germany only 12 percent. In hotels and restaurants 72 percent of employment across Europe was precarious, but only 13 percent in transport and communications and 11 percent in ‘financial intermediation’. [49] And there was a concentration of precarious employment among young workers – a third of those in temporary or fixed term contract work are under the age of 25 and two thirds are under the age of 35. [50]

The overall pattern can be quite different to what is normally claimed. Thus, in Europe in the second half of the 1990s there was no increase in either the proportion of the self employed nor of ‘precarious employment’:

In the year 2000 the European population in work totalled 159 million people, 83 percent of whom were employees and 17 percent self employed. In 1995, the figure was 147 million with the same ratio between the self employed and employees. [51]

As regards ‘precarious employment’:

While this type of employment increased substantially during the first half of the 1990s, the relative proportions of permanent and non-permanent jobs remained almost unchanged between 1995 and the year 2000: permanent (82 percent), non-permanent (18 percent). [52]

Across Western Europe as a whole, ‘one out of five jobs has been precarious during the last five years’ [53] – but that still leaves four out of five jobs as ‘permanent’.

In Britain, the latest survey suggests that the economic recovery of the late 1990s was accompanied by a fall in precarious employment: ‘As many as 92 percent of workers held permanent employment contracts in 2000 as compared with 88 percent who did eight years earlier ... A mere 5.5 percent said they were working on a temporary work contract of less than 12 months in 2000, compared with 7.2 percent in 1992’. [54] The actual length of time which people spend in the same jobs has not changed very much. The proportion of employees who had been in jobs less than three months was 5 percent in 2000, the same as in 1986, and the proportion less than a year 20 percent, as against 18 percent in 1986. At the other extreme, the proportion in the same job for more than ten years was 31 percent as against 29 percent. The only really big change was a fall in the proportion of workers in the same job for more than two but less than five years between 1996 and 2000 from 21 percent to 15 percent. [55]

These figures do not prove, as is sometimes claimed by apologists for capitalism, that people’s feelings of insecurity are mistaken. Workers can change jobs frequently for two diametrically opposed reasons – either because there is a growing demand for their labour, allowing them to improve their situation by getting a different job, or because they are being forced out by employers who are shedding labour. So, for instance, few people were likely to give up their jobs voluntarily during the recession of 1990-1994 in Britain, precisely because employment generally was less secure than previously. And in the ‘more secure’ year 2000 some 27 percent of unskilled and semi-skilled manual workers had been in their jobs for less than a year, reflecting the recession that began in manufacturing industry while services continued to boom.

The figures do, however, show that ‘average job tenure has remained relatively stable since 1975’. [56] The idea that the working class has been ‘flexible-ised’ out of existence is completely mistaken. Most people continue to work in the same place, and to be subject to exploitation by the same employers for quite long periods of time. By the same token, they have the time and opportunity to connect up with the people around them to fight back against that exploitation.

The myth of instant mobility

The claim that the ‘permanent’ worker is a thing of the past is often connected with the claim that employers can move production – and jobs – at a moment’s notice.

So Hardt and Negri write:

The informatisation of production and the increasing importance of immaterial production have tended to free capital from the constraints of territory, and capital can withdraw from negotiation with a given local population by moving its site to another point in the global network ... Entire labouring populations, which had enjoyed a certain stability and contractual power, have thus found themselves in increasingly precarious employment situations. [57]

This vastly exaggerates the movement of capital, and the ease with which firms can move their operations from one place to another.

As I have explained elsewhere [58], capital as money (i.e. finance) can move at the touch of a computer key from one location to another (although determined governments can still impede its movement). But capital as means of production finds it much more difficult to do so. Physical equipment has to be uninstalled and reinstalled, transport has to be arranged for goods produced, a reliable workforce with the requisite skills found, and so on. It is a process that is usually expensive, taking years rather than seconds. What is more, physical production depends upon transporting goods to markets, and therefore closeness to markets is an advantage.

The result is that most of the restructuring of industry over the last three decades has usually been within the world’s existing industrial regions. As Rowthorn explains:

The developed world is now mostly divided into three blocs, comprising North America, Western Europe and Japan. These blocs are largely self contained in sophisticated manufactured goods. [59]

So in 1992, manufactured imports from Japan amounted to only 0.74 percent of Western Europe’s GDP and 1.5 percent of US GDP, while manufacturing imports from Western Europe only amounted to 1.2 percent of US GDP. And at that time, Japan’s total manufactured imports from the whole of Asia, including China and the Middle East, were less than 1 percent of Japanese GDP. [60]

There has, of course, been a shift in certain manufacturing industries to states which were not industrialised 40 years ago – otherwise the phenomenon of the Newly Industrialising Countries (NICs) and of certain expanding industries in ‘underdeveloped’ countries would be inexplicable. But there is little evidence to support the claim that ‘advanced countries are abandoning the production of manufactured goods. Many labour-intensive manufacturing activities in the advanced economies, such as clothing or routine assembly, have been put out of business by rising imports from developing countries’, but these imports have been financed ‘not by the export of services’ but ‘by the export of other manufactures, especially capital goods and intermediate products such as chemicals’. [61]

Overall imports to advanced OECD countries from non-OECD countries only grew from about 1 percent to about 2 percent of GDP between 1982 and 1992.

Rowthorn estimates that the total job loss from all the advanced countries through this shift has only been about 6 million jobs, or 2 percent of total employment (compared with total unemployment of around 35 million in these countries).

Baldoz, Koeber and Kraft point out that the restructuring of industry in the US has not involved a net flow of jobs abroad: ‘The US now has a larger percentage of the workforce working for wages than at any time since the 1950s – and for astonishingly long hours’. [62]

Restructuring means that much of this production does not take place in the old industrial centres, such as those in and around Detroit, but in the ‘sun belt’ states of the west and south. So most US auto workers no longer work directly for the ‘Big Three’ of Ford, General Motors and Chrysler, but for ‘trans-plant manufacturers’ like Honda, Toyota, Nissan, Mitsubishi and Daimler Benz, or for new parts manufacturers spun off by GM so as to weaken the union. [63] This is a far cry from the picture sometimes presented of all US auto jobs disappearing over the border into Mexico. [64]

Some industries find it easier to move than others. So, for instance, clothing is a particularly mobile line of production. The basic equipment – shears for cutting, sewing machines, presses – is light, cheap, and the products are relatively easy to fly from one part of the world to another. [65] Not surprisingly, many of the stories about firms shutting down and moving when faced with rising labour and other costs are about this industry. But even here there are limitations to mobility. Production of high quality goods can still be based in advanced countries. So there were 112,190 workers in the garment industry in New York city in 1990. And they certainly were not all ‘informational’ workers – 64,476 were production workers (mostly foreign born) and only 13,522 were ‘professionals and managers’. [66] At the time, the total number of garment workers in the US was around 300,000.

Another industry whose supposed mobility has been much emphasised by commentators from the right and left alike is software writing. It involves relatively little capital investment and transportation costs virtually disappear as processed data is sent almost instantaneously by ever-cheapening telecommunications links over thousands of miles. So software programs can be written in Bangalore in India for firms whose central headquarters are in Chicago, and airline tickets ordered in London can be typed into computer terminals in Delhi. And many Third World countries contain large numbers of highly educated fluent English speakers capable of doing such tasks: India alone has 4 million people with technical backgrounds and 55,000 graduates in engineering and sciences every year. [67] The industry does seem to fit Hardt and Negri’s description of ‘productive forces’ that are ‘completely delocalised’, in which ‘brains and bodies ... produce value’ but without necessarily needing ‘capital and its capacities to orchestrate production’. [68] And certainly, software production in India has been expanding at great speed – from employing 2,500 to 6,800 people in 1985 to 140,000 in 1996, of whom 27,500 were involved in the export sector. [69]

But a closer look at the industry in India shows that it is still dependent on fixed capital investment and cannot float off to another location at a moment’s notice. Bangalore has grown to be the country’s biggest software centre because it has facilities, many provided by the local state government, that other parts of the country do not. As a study of the industry explains:

Because of its dust-free environment, between 1956 and 1960 large public sector undertakings like Bharat Electronics Limited (BEL) and Hindustan Aeronautics Limited (HAL) were established by the government of India in Bangalore along with national defence research laboratories and the Indian Institute of Science ... In the 1970s the Indian Space Research Organisation and Bharat Heavy Electronics Limited (BITEL) were also located in Bangalore.

The results of this concentration of electronics and aeronautics industries in Bangalore were the backward and forward linkages that were created ... Bangalore was, therefore, a natural location for the computer and IT industry to locate ... Furthermore, Bangalore is the state’s capital, which has many advantages in terms of proximity to policy-makers and the government offices ... and to an airport. [70]

There was a concerted effort amongst policy makers to create an environment in Bangalore where high-tech industries would flourish ... The aim was to provide the facilities and infrastructure necessary to promote investment in the electronics industry, including a guaranteed supply of electricity, telecommunications facilities and a technical training centre. [71]

Without the assurance of guaranteed electricity and clean water supplies – neither of which is guaranteed in most Third World cities – the industry could not have taken off. In fact, with the expansion of the industry, firms often can no longer get these things without undertaking costly investments of their own. [72] So the firm Infosys ‘spends over $201,000 per new recruit in capital expenditure and training ... Employees have access to the latest in technology – a policy which is expected to lead to higher productivity.’ [73]

The firms which have found a suitable location for software production and which have then invested in training up a workforce are not likely simply to move somewhere else at the drop of a hat. For the same reason, firms with existing facilities in Europe, Japan or North America have not closed them down to move to Bangalore. Their Bangalore operations have been a response to difficulties in recruiting the skilled workforces they need at their home locations. This is shown by the relatively small size of the Bangalore software workforce. ‘It is not a big employer, especially by Indian standards. Approximations suggest that it currently [in 1996] employs between 7,000 and 15,000 people in Bangalore’. [74] These figures are very small compared with the global workforce in the industry. Meanwhile, labour shortages in India are now causing wages to rise by about 30 percent a year and some forecasters envisage ‘a time when ... India will no longer be considered a low wage location’ [75] since industry is beginning to suffer from environmental problems created by its own expansion (pollution, electricity and water shortages, and road congestion).

India in general and Bangalore in particular have found a niche in the world software market. But the niche is limited in extent, and is certainly not proof that software as an industry can simply wholesale from one part of the world to another.

Things are slightly different with the less skilled processes of inputting data into computer terminals. These require less expensive and sophisticated equipment and a lower level of training in the workforce. For this reason they are more widely dispersed in India than software production. As in the clothing industry, the less skilled operations are more mobile than the more skilled ones. But even in these industries there are constraints. The workforce has to have the right linguistic skills (fluent Urdu is no good for inputting US airline data) and has to be trained and committed to accuracy (mistakes on data inputting can lead to considerable expense), while the communications equipment (and therefore the electricity supplies) have to be reliable. When a company has all these things, it is not going to throw them away without a powerful motive.

Software production, like capitalist production in general, is subject to repeated rationalisation and restructuring through crisis. This leads to industries based in some localities contracting or disappearing, while others expand or emerge for the first time. But it is not ‘fluid’ in being able to move effortlessly from one location to another. The general trend for capitalism today is still for production to be concentrated in the advanced countries. Some sorts of production have shifted to a few favoured areas of the Third World – the NICs of east and south east Asia, and eastern China. But capital still finds it more profitable, in general, to locate itself in the regions which had industrialised by the mid 20th century. Workers may usually be better paid there, but a combination of established skills levels and existing investments in plant and infrastructure mean they are also more productive, providing much more surplus value for the system than most of their poorer brothers and sisters in the Third World. This explains why the picture for most of Latin America has been one of very slow average growth or stagnation, and of most of Africa of absolute decline.

Patterns of employment: the world working class and the ‘Third World’

Capitalism has created a world working class in the last century and a half. Industry and wage labour exist today in virtually every part of the globe. The industrial working class has a worldwide presence. But the combined and uneven development of the system means it is very unevenly distributed between regions. Rough calculations indicate that 40 percent of the world’s 270 million or so industrial workers are in the OECD countries, around 15 percent each in China, Latin America and the former USSR, around 10 percent in the rest of Asia, and around 5 percent in Africa. [76]

The unevenness does not just exist between the old industrial countries and the rest of the world. It also exists within the ‘Third World’.

Urbanisation and the spread of market relations are not necessarily the same as the growth of wage labour. This is especially true in countries where crisis means economic growth is slow or negative. Thus ‘wage employment is reported to have fallen in absolute terms in several African countries’ – by 33 percent in the Central African Republic, 27 percent in Gambia, 13.4 percent in Niger, 8.5 percent in Zaire [77] – and urban unemployment rates were in the region of 15 to 25 percent (up from about 10 percent in the mid-1970s). [78] For sub-Saharan Africa as a whole, ‘The chief employment in rural areas is self employment’. [79] Although manufacturing activity can account for up to 20 percent of employment in rural areas, it is usually of a very rudimentary sort – blacksmithing, brewing, tailoring or milling, usually in one person businesses.

Sub-Saharan Africa is the exception rather than the norm for the world system as a whole, or even for its vast, impoverished regions. In Asia and Latin America there has been a growth of wage labour. But it has often been outside what is usually called the ‘modern’ sector, and has often been accompanied by an equally rapid rise of self employment.

In Latin America the number of people in the employed non-agricultural labour force rose from 68 million in 1980 to 103 million between 1980 and 1992 (during what is often called ‘the lost decade’ of economic crisis and stagnation). But employees in ‘large business’ only grew from 30 to 32 million. By contrast the number in small businesses grew from 10 to 24 million, the number in the public sector from 11 to 16 million, the number in domestic services from 4 to 7 million, and the so called ‘informal’ sector mushroomed, from 13 million to 26 million. [80] The joint share of informal and small business activities of non-agricultural employment rose from 40 percent in 1980 to 53 percent in 1990. [81]

’Informal jobs per se are on their way to representing one third of the non-agricultural workers in the region ... Most of the increase in the informal sector is concentrated in own account workers’. [82] In Brazil in 1980, almost half the occupied urban population were not ‘formal employees’ [83], although more than half of these were wage workers without the protection of formal, legal benefit and 18.4 million workers declared in 1990 that they wanted to have a formal job. [84] This does not constitute ‘deindustrialisation’ and certainly not the disappearance of the working class in Latin America.

The Indian economy has been growing on average faster than most countries in Latin America for the last two decades, although starting at a much lower level than most of them. Output per head has been rising, while it has been falling in most of Latin America, and the share of industry in total output is now around 19 percent. But the growth in jobs, as in Latin America, has been overwhelmingly in the informal sectors.

In the 1980s, ‘despite a significant acceleration of the industrial growth rate ...the share of manufacturing in total employment ... declined’ with ‘negative’ employment growth in ‘private organised manufacturing [i.e. the formal sector]’. [85] Between 1977-1978 and 1993-1994 the proportion of the employed male urban workforce who were ‘regular employees’ fell from 46.4 percent to 42.1 percent (although the total numbers rose, since the urban population grew massively in these years), while the proportion who were ‘self employed’ rose slightly from 40.4 percent to 41.7 percent, and the number of casual employees rose from 13.2 percent to 16.2 percent. [86]

Most of the self employed are by no means privileged. A survey of Ahmedabad shows only one tenth of the male self employed as having a ‘separate business place’. A third worked on the streets, as vendors, rickshaw drivers, cart pullers and the like. There are 200,000 rickshaw men in Mumbai, 80,000 in Ahmedabad, and 30,000 in Bangalore, while Calcutta has around 250,000 street hawkers and Calcutta more than 100,000. [87]

The pattern that applies to India also applies, in varying degrees, to Pakistan and Bangladesh. [88] It is also to be found in other, more ‘advanced’, ‘developing’ countries. In Turkey employment in ‘large manufacturing’ establishments amounted to 979,839 in 1987, as against 550,670 in ‘small manufacturing’ establishments. And 44.2 percent of workers in all industry work in firms of more than 100 workers, as against only 24.2 percent in firms with less than ten workers. [89] Not surprisingly given this level of concentration of industry, 50 to 55 percent of workers were unionised. [90] But the number in the urban ‘informal’ sector was 1,854,000 in 1988 – and had grown to 2,152,000 by 1992. [91]

In addition to – and often merging into – those in the informal sector, there are everywhere those denied any opportunities for employment by modern capitalism: the unemployed. Their numbers vary considerably from region to region and country to country – depending, in part, on the ease of people making some sort of livelihood in the informal sector. So across the Middle East as a whole unemployment is reported to be 15 percent. [92] But ‘it reaches between 25 and 30 percent in Libya, Algeria, Iran and Yemen’. [93] In Greater São Paulo in October 1995 it was reported to account for 1,102,000 people out of a workforce of 8,221,000. [94]

In China the government is following a conscious policy of restructuring the state-run sector of the economy. The number of employers in selected enterprises fell from 45 million in 1993 to 27 million in 1998. [95] Some of these workers will have found other jobs, but by no means all – one source tells us that there were only 1.54 million jobs available in August 2001 for the 2.2 million workers registered at official labour agencies in 82 cities. [96] There were certainly far too few jobs to meet the aspirations of the millions of people coming to the cities from the countryside looking for employment. There are ‘150 million drifters from the countryside who travel from city to city in search of manual jobs’. [97]

Capitalist accumulation is causing the rapid growth of cities across wide swathes of the globe, and of occupations involving production for the market. In most regions (although not in most of Africa) there is also a growth of the number involved in wage labour of a relatively productive sort in medium to large workplaces. But even more rapid is the expansion of the vast mass of people precariously trying to make a livelihood through casual labour, selling things in the streets, trying to survive through working on their own account. At one extreme this mass merges into the petty bourgeoisie proper of small employers, at the other into the desperate poverty of those who can hardly get a livelihood at all – 48 percent of the urban population of Brazil live below the poverty line, and two out of five of these below the ‘indigence’ income needed to satisfy food needs but nothing else. [98]

The economics and politics of the informal sector

How does this mass of informal sector and own account workers relate to the workers with ‘formal’ employment?

There is one widespread, very simple, and very mistaken, answer. That is to see the workers with permanent jobs as ‘privileged’, as some sort of ‘labour aristocracy’. This is certainly how it can seem to those driven into the informal sector. In the formal sector there are usually considerably higher wage rates and often sickness benefits, paid holidays, and pensions of sorts as well. So in the cities of north east Brazil, it is claimed that to ‘be formally employed is almost a privilege, since less than half of those who want such a situation are in fact “enjoying” it’. [99] In India, across a wide range of occupations workers in the ‘organised sector’ tend to get paid a good deal more (30, 40 or even 100 percent) more than those in the ‘unorganised sector’. [100] In China, workers in large scale industry used to be promised the ‘iron rice bowl’ of a guaranteed income plus certain housing, sickness and pensions benefits – and people migrating from the countryside to seek jobs were excluded from these things by an internal passport system that denied them the right to reside in the cities.

Employers have not, however, provided such things out of the goodness of their heart. They need a certain stability to their labour force, particularly when it comes to skilled workers who they do not want to be poached by rivals during times of boom. States often want such stability as well, seeing welfare provision for a section of the urban workforce as a way of protecting themselves against sudden explosions of popular discontent.

So, for instance, in the decades after the revolution of 1910-1919 in Mexico, a political structure was created which attempted to integrate the unions into the ruling party, along with the employers and the peasant organisations. As one academic account explains:

The welfare policy promoted by the post-revolutionary regime was governed by a logic of sectoral benefit which favoured workers in the formal sector of the economy ... Social security mechanisms include access to a variety of benefits in addition to health care and incapacity allowances such as allowances in the event of disease, retirement pensions, compensation for accidents, compensation of families in the event of a death, payments of maternity allowances ... To the extent that the social security policy has handed out stratified benefits, it has acted in a discriminating manner. It has excluded on the one hand, all those who do not work in the formal economy and, on the other, those that, whilst working in the formal economy do not belong to a trade union. Preference is thus given to taking care of groups which are vertically organised in unions recognised by the state. [101]

The aim was not, however, to do a favour for the workers in the formal sector. It was rather to provide a mechanism to control them, allowing the state to take over ‘granting or refusing registration to trade unions, being able to arbitrate in labour disputes, legalising strikes, pronouncing their lawfulness or unlawfulness’ and tending to turn the trade unions ‘into quasi-governmental or quasi-corporative institutions ...where the union representatives have been raised above those they are representing’. [102] Whenever workers tried to act outside these structures, the state would repress them in the most violent fashion. The real beneficiary was the Mexican bourgeoisie which was able to emerge intact with an ultra-stable state after one of the most tumultuous revolutionary upheavals of the 20th century. In 1950 those in the richest 10 percent of the population were 18 times richer than those in the poorest 10 percent. By 1970 they were 27 times richer. [103] The rich were able to gain in this way because they had developed a mechanism for controlling those whose labour created their wealth. And the most important people for it to control were those who laboured in the most productive, that is, the most advanced, ‘formal’ sections of the economy.

The ‘iron rice bowl’ played a similar role for the rulers of China in the period from the 1950s to the 1990s. It guaranteed a stable workforce of experienced, productive workers in the core industries where most capital investment was tied up.

It often seems counterintuitive to argue that groups of workers who have better conditions than others do not benefit at their expense – whether the argument is used about Western workers and workers in the Third World, or formal sector Third World workers and informal sector workers. But in this case the ‘counterintuitive’ argument is correct. In many industries, the more stable and experienced a workforce, the more productive it is. Capital is prepared to concede higher wages to certain of the workers in those industries because by doing so it is able to make more profits out of them. Hence the apparent contradiction – some sections of the world’s workers are both better paid than others, and more exploited. It is this alone which explains why capitalists, motivated only by the drive for profit, do not usually invest on any great scale in regions like Africa, where wages are lowest.

That, of course, does not prevent capital from continually trying to hold down what it has to pay – and from seizing on new technologies and restructuring production to reduce its labour costs drastically. Hence the pattern in much of the world for the established ‘formal’ workforce to remain more or less intact, but for there to be some chipping away round the edges and for many new jobs to be in the ‘informal’ sector.

The great mass of the informal workforce in ‘developing’ countries today are people who are new to the urban workforce – either from the countryside (as with the more than 100 million peasants seeking employment in China’s cities) or women and young people seeking paid labour for the first time. But the pattern of capitalist accumulation over the last couple of decades means that the labour demands of modern, productive industry have not expanded on anything like the scale needed to absorb them into its workforce. Competition on a global scale has caused capitalists to turn to ‘capital intensive’ forms of production (with what Marx called a rising ‘organic composition of capital’) which do not require massive numbers of new workers. As a result, the only ways for most new entrants to the labour force to gain a livelihood are through the most meagre forms of self employment or through selling their labour power at such a low price and under such arduous conditions that small capitalists at the margins of the system can profit from exploiting it.

As a report on employment in Latin America points out:

Informal jobs per se are on their way to representing one third of the non-agricultural workers in region ... Most of the increase in the informal sector is concentrated in own account workers ... The result of this process has been a trend towards lower unemployment rates, but at the cost of a marked deterioration in average labour productivity. [104]

In general, the suffering of a very large chunk of the urban masses in such countries comes not from being super-exploited by large capital, but from the fact that large capital does not see a way of making sufficient profits out of exploiting them at all. This is even more clearly the case in much of sub-Saharan Africa. After squeezing wealth out of the continent during the period from the onset of the slave trade to the end of empire in the 1950s, those who run the world system (including local rulers who move their own money to Europe and North America) are now prepared to write off most of its people as ‘marginal’ to their requirements.

Marx described very well the process by which the informal sector grows, looking at British society 150 years ago:

The additional capitals formed in the normal course of accumulation serve particularly as vehicles for the exploitation of new inventions and discoveries, and industrial improvements in general. But in time the old capital also reaches the moment of renewal from top to toe, when it ... is reborn ... in a perfected technical form, in which a smaller quantity of labour will suffice to set in motion a larger quantity of machinery and raw materials.

... The additional capital formed in the course of accumulation attracts fewer and fewer labourers in proportion to its magnitude. The old capital ... repels more and more of the labourers formerly employed by it. [105]

The labouring population therefore produces, along with the accumulation of capital produced by it, the means by which it itself is made relatively superfluous, is turned into a relative surplus-population. [106]

As soon as capitalist production takes possession of agriculture ... the demand for an agricultural labouring population falls absolutely ... Part of the agricultural population is therefore constantly on the point of passing over into an urban ... proletariat. [107]

This dynamic produces a ‘stagnant’ component of ‘the active labour army’ with ‘extremely irregular employment’:

Its conditions of life sink below the average normal level of the working class; this makes it at once the basis of special branches of capitalist exploitation ... characterised by maximum of working time, and minimum of wages ... Its extent grows, as with the extent and energy of accumulation the creation of a surplus-population advances. [108]

The same causes which develop the expansive power of capital, develop also the labour-power at its disposal. The relative mass of the industrial reserve army increases therefore with the potential energy of wealth. But the greater this reserve army in proportion to the active labour-army, the greater is the mass of a consolidated surplus-population, whose misery is in inverse ratio to its torment of labour ... This is the absolute general law of capitalist accumulation. [109]

The interplay of the sectors

That is not, however, the end of the matter. Capitalism has one important use for those it refuses to allow to make a proper livelihood. It uses them to put increased pressure on those it does exploit in the most productive areas of the economy. Far from the growth of the informal workforce benefiting the workforce in the formal sector, it has been accompanied by an increased exploitation of workers in this sector – and in many cases by a deterioration.

The deterioration is most marked in Africa, where the scale of the decline in real wages for those with jobs is so great it beggars belief. A report in 1991 told of ‘a sharp fall in real wages ... an average 30 percent decline between 1980 and 1986 ... In several countries the average rate has dropped 10 percent every year since 1980 ... On average the minimum wage fell 20 percent over that period’. [110]

There was:

... a virtual collapse of real wages in Somalia, Sierra Leone and Tanzania ... The observed wage declines were much higher than the decline in per capita incomes in the region ... Wage earners have shouldered the brunt of the crisis ... Working people, especially in the cities, have been pauperised by inflation and devaluation. [111]

This collapse of wages goes a long way to explaining the ferocity of civil war in somewhere like Sierra Leone, where the starting civil service salary for the lowest grades was only one fifth of the 1978 figure by 1989 [112]: warfare seemed the only way to make a minimal livelihood to some of the young male population.

In Latin America the growth of the informal sector in the 1980s was accompanied by a fall of more than 10 percent in real industrial wages – although there was some recovery in the early 1990s before the Asian crisis hit the continent in the late 1990s. In India, real wages in the formal sector do not seem to have declined. But on average neither have they grown during nearly two decades of economic growth.

In both India and Latin America something else has been happening – the shifting of certain jobs in big industry from the formal to the informal sector. This allows management to cut some of their wage costs – and to put pressure on the remaining ‘formal’ sector of the workforce to accept worse conditions. So in one of the major areas of Indian industry, southern Gujarat:

The increased share of casual labour in the local industrial workforce has been a major change that has occurred during the last 30 years. I estimate that at least 50 percent of industrial workers are temporwalis [temporary workers] or contract workers ...

The distinction between permanent and temporary workers ... emanates from the owners’ permanent efforts to dodge labour acts and regulations in order to keep labour as cheap as possible and the carelessness of officials in charge of implementation and surveillance of the legislation ... Temporary workers generally perform unskilled jobs and do the hard work like loading and unloading, feeding machines, packing finished products and cleaning ... Their maximum pay does not exceed the daily official minimum. [They] do not get any fringe benefits and they are easy to dismiss ... They are beyond the reach of legislation, they therefore lack the attention of ... union leaders, for whom labour laws are the starting point for all their actions. Women have contributed substantially to the casualisation of the local labour force. So too has labour supplied by contractors from further afield, which managers regard as more ‘reliable’ and less prone to absenteeism than locally recruited labour. [113]

In parts of Latin America there has likewise been a tendency for some work which would have been done by the permanent ‘formal’ labour force in the past now to be done through the informal sector. Informal employment used only to be feasible:

for small firms and firms that hire only temporary manpower, such as farms for the harvest, and building firms ... This seemingly changed at the beginning of the 1990s ... This was a time of recession and more firms (supposedly small and middle sized) became informal to evade payment, not only of the indirect wage but of all taxes. Besides that, a new trend became significant, the so called tercuaruzação which meant the replacement of permanent employees (mostly formal) by self employed service suppliers ... in the most advanced sectors of the economy. [114]

This, it must be emphasised, does not mean the end of the formal section. It still has advantages for certain employers. In Gujarat the employers do not feel able to fulfil all their labour requirements with temporary or contract labour:

Permanent workers do receive benefits. Their employers contribute to the provident fund, pension, bonus, paid leave and they are more difficult to dismiss. Generally they are more skilled or experienced and considered indispensable and more loyal ... They are found in small and large factories with more than a rudimentary division of labour and comparatively more sophisticated technology, like the engineering industry. [115]

From the employers’ point of view, the use of casual and contract labour is a double edged weapon. It enables them to reduce labour costs and to force increased workloads on the workforce. But it also means they have a workforce which is more likely to be attracted to better jobs elsewhere if these become available. And it also undercuts attempts to portray themselves as the loyal ‘social partners’ of their workers.

So one Gujarat employer has explained:

We don’t go by contractors ... We employ only known persons. We need confidence and we get that through recommendation. They don’t dare disappoint. They have fear on their mind. [116]

And as Heen Steefkerk, who carried out this study, points out, the use of temporary and contract labour makes it less easy for the employer to pretend everyone is part of the same happy family:

More contract workers signify an altering course of local industrial transition. They indicate another work climate and changing labour relations. These trends imply a further objectification of labour relations, or, in other words, greater social alienation between workers and owners or managers. [117]

In Brazil’s most important industrial city, São Paulo, the formal sector shrunk slightly in the recession of the early 1990s, to expand again in the mid-1990s until it was bigger than ten years before, and although the informal workforce had grown by nearly 70 percent in the interim, the number of ‘formal’ workers employed in the private sector remained more than four times higher than the number of ‘informal’ workers. [118] Informal employment has been nibbling away significantly around the edges of the formal workforce. But it still leaves intact a massive force of workers who employers either cannot or do not want to ‘informalise’. It is wrong, as people like Paulo Singer do, to write of ‘deproletarianisation’. [119] Rather, what is happening is a restructuring of the workforce, with the hiving off by big firms of some tasks (usually relatively unskilled and therefore easily performed by a floating workforce) to small firms, labour-only contractors and the supposedly self employed.

It should be added that this phenomenon is by no means new in the history of capitalism. Casual employment has often played an important role in certain industries – for instance, in the docks in Britain until the late 1960s. And forms of contract labour are very old – it was common in the textile factories of the industrial revolution. In the mines in both the US and Britain in the 19th century, overseers or foremen (‘buttymen’) would recruit workers and pay them out of a sum given to them by the mine owners. These casual groups of workers may not always have felt themselves to be part of the working class. They were often detached from the struggles of other sections of that class for years, even decades, at a time. Yet the potential for struggling with those sections was always there, and when it turned into reality the struggle could be very bitter, with an almost insurrectionary tinge.

Frederick Engels observed precisely this development in 1889 when London’s dockers struck for the first time. He wrote:

Hitherto the East End had been in a state of poverty-stricken stagnation, its hallmark being the apathy of men whose spirit had been broken by hunger, and who had abandoned all hope. Anyone who found himself there was lost physically and morally. And now, this gigantic strike of the most demoralised elements of the lot, the dock labourers, not the regular strong, experienced, relatively well-paid men in steady employment, but those who have happened to land up in dockland, the Jonahs who have suffered shipwreck in all other spheres, starvelings by trade, a welter of broken lives heading straight for utter ruin ... And this dully despairing mass of humanity who, every morning when the dock gates are opened literally fight pitched battles to be first to reach the chap who signs them on, that mass haphazardly thrown together and changing every day, has successfully combined to form a band 40,000 strong, maintain discipline and inspire fear in the powerful dock companies ... Whatever the outcome of the strike may be, it means that the lowest stratum of East End working men has entered the movement and that the higher strata are bound to follow suit ...

And there’s more to it than that. Because of the lack of organisation and the passively vegetable existence of the real East End workers, the lumpen proletariat has hitherto had last say there, purporting ... to be the prototype and representative of the million starvelings in the East End. That will now cease. The hawker and his ilk will be pushed into the background, the East End working men will be able to develop a prototype of their own and so organise themselves. [120]

The point is very important. Internationally we are just emerging from more than two decades of defeat and demoralisation for workers right across the world. This bred a fatalism about the possibility of fighting, which was reflected in a mass of studies which depicted the suffering of the poor and the oppressed, showing them always as victims, rarely as fighters. Thus there are tons of materials sponsored by the International Labour Organisation on ‘social exclusion’ – a theme which suits the bureaucrats who run such bodies. In these studies themes like the ‘casualisation’ and ‘feminisation’ of the workforce become stereotyped, academic ways of dismissing possibilities of struggle – even if some of those carrying through the studies try to escape from the paradigm in which they are trapped. The stereotypes then provide trade union officialdom with excuses for avoiding struggle on the grounds that it cannot work. What begins as a mistaken assessment of the possibility of struggle becomes a real obstacle to unleashing such struggle.

The maquiladores

Often associated with talk of the supposedly ‘unorganisable’ informal sector is the enormous amount of literature on the maquiladores – the production units set up by multinationals in Third World countries which carry through certain routinised key tasks in a global production chain. The archetypal units are the clothing establishments run by firms such as Gap or Nike in ‘free trade zones’ or ‘EPZs’ in Third World locations such as Indonesia, Central America or the Philippines. Naomi Klein describes these graphically in No Logo:

There are at least 850 EPZs in the world ... spread through 70 countries and employing roughly 27 million workers ... Regardless of where the EPZs are located, the workers’ stories have a certain mesmerising sameness: the workday is long – 14 hours in Sri Lanka, 12 hours in Indonesia, 16 in Southern China, 12 in the Philippines. The vast majority of the workers are women, always young, always working for contractors or subcontractors ... filling orders for companies based in the US, Britain, Japan, Germany or Canada. The management is military-style, the supervisors often abusive, the wages below subsistence and the work low-skill and tedious.

As I walk along the blank streets of Cavite [in the Philippines], I can feel the threatening impermanence, the underlying instability of the zone. The shed-like factories are connected so tenuously to the surrounding country ... that it feels as if the jobs that flew here from the North could fly away again just as quickly ... Fear pervades the zones. The governments are afraid of losing their foreign factories; the factories are afraid of losing their brand-name buyers; and the workers are afraid of losing their unstable jobs. These are factories built not on land but on air. [121]

Such accounts provide a brilliant exposure of the greed and inhumanity of those who run the multinationals. But like many orthodox academic studies on the informal workforce (especially those sponsored by the International Labour Organisation) they are too pessimistic when it comes to the possibilities of fighting back.

First, the multinationals cannot afford simply to mistreat their workers. It is not as easy as the multinationals would like people to think for them to close down their facilities and move elsewhere if the workforce does explode in bitterness. Setting up the links in a global production chain takes a lot of effort by the multinational. It has to ensure mechanisms for achieving on the spot quality control of each stage of production, reliable transport links, a strong police structure to protect it from petty pilfering, clean water supplies [122], and, above all, it wants a labour force that is trained to the rhythms of repetitive work at regular, long hours. It cannot get this simply by taking people off the street and then dismissing them at a whim. This means that although it may recruit from members of the informal labour force, once they are employed it is likely to provide them with some of the conditions of a formal labour force – even if it does so in the midst of a police-labour regime. A study of maquiladores in Honduras, for instance, shows that wages are considerably higher than in the informal sector. The mean income of the workers is US $141 a month, as opposed to the $91 a month the workers earned in their previous employment and the less than a dollar a day that 47 percent of the country’s population have to live on. [123] The multinational needs a workforce it can get the maximum amount of surplus value out of, and that means one that will work efficiently and is able to sustain the production rhythms imposed on it without quality suffering.

When Henry Ford pioneered mass production, assembly line methods in the auto industry, he saw that the most effective form of exploitation lay in stabilising a handpicked workforce under tight managerial control. Thomas O’Brien has told how some of the first US multinationals to operate in Latin America took efforts to stabilise their workforces by providing minimal welfare facilities – providing accommodation in company towns, health clinics, schools, sports facilities, even paid holidays. The aim was to combine maintaining the workers at a minimal level of fitness with extending managerial discipline over workers to the home as well as the workplace. Even the notoriously exploitative United Fruit Company saw the benefit to profitability of such measures. [124]

The same logic applies to firms engaged in mass production through global assembly lines. Key functions affecting the quality of output (eg cutting clothes) will be performed by long term employees, even if casual workers are used for tasks which require brute force rather than minimal skill. This does not, of course, rule out the multinationals avoiding certain taxes by pretending to government officials supposedly responsible for overseeing labour legislation that their employees are temporary, informal or ‘unorganised’ workers.

This element of stability in the workforce is important because it means such workers can fight back, and win. Conditions in many South Korean clothing and footwear plants in the 1960s were exactly as Naomi Klein describes them. George E. Ogle has told of ‘the sweat, blood and tears of young women who worked in the export industries during the 1960s and 1970s – textiles, garments, electronics, chemicals’:

Eighty three percent of the employees in the textile industry are women. They are 16 to 25 years of age, and have come primarily from the countryside ... Skills used in the textile industry can be learned quickly. Once the machines are in place a steady supply of cheap, diligent labour is all that is needed. The country girls provided that labour. In 1970 there were already 600,000 female employees in manufacturing. That was about 30 percent of the entire labour force. Most of those were in textiles ... By 1980 there were about a million and a half female workers employed in mining and manufacturing ... Recruiters would go to the countryside to hire workers. The employment contract was understood to be as much with the family as with the person who was actually employed. The family, as it were, stood responsible for its daughter’s work and behaviour at the factory. Once employed the young women were housed in a company dormitory that was usually located within the company walls.

In the factory setting women are usually supervised by men. The men expect traditional-style obedience. They assume a traditional-style superiority. They speak in commands of traditional ‘low talk’ and when irritated, they may well respond with a traditional crack on the head or a slap to the face. [125]

Under the military government of Chun Doo Hwan in the 1970s and early 1980s attempts at resistance were viciously crushed. Typical was the attack on the union at the Wonpoong Textile Company in 1982:

The new union president ... Ms Kim Sung Koo, and a foreman on the floor (also union) were beaten and then fired for dereliction of duty. Two weeks later, a group of men, some of whom were from management and some others who were hired thugs, took over the union hall and kidnapped the newly elected replacement ... Chung Son Soon. Thugs kept her for 17 hours. They beat her, threatened her, humiliated her and then threw her out of a car somewhere in the outskirts of town. Barefooted and bleeding, she walked back to the factory. In the factory, her friends were conducting a sit-in ... Police, management men and thugs joined together to literally drag the workers out of the plant and throw them into the street. Union members were arrested. [126]

Such methods broke the back of most attempts by workers to organise themselves for two decades. But then in 1987 the military regime entered into a political crisis which saw the country swept by demonstrations involving sections of the middle classes. Against this background workers began to fight for their own interests. The revolt began in the big firms, the chaebol, first. More than two thirds of firms with over a thousand workers were hit by strikes. Their example soon inspired others:

The chaebol were organised first and then soon after businesses of all sizes were rapidly organised. Enthusiasm spread outside the boundaries of the ‘blue collar’ industrial workers, and into the ‘white collar’ service sectors of health care, research at government agencies, education institutions. And insurance companies ... [127]

About 33 percent of Korean workers are in small firms that employ between five and 100 people ... One clear pattern that emerged early in the 1987 labour eruptions was that the small and medium sized plants were being organised at the same rapid rate as the big chaebol ... Workers at these firms stretched out their courage and their consciousness and took the risk. Subsequently, a good portion of them have maintained themselves by organising regional associations for mutual protection. [128]

The maquiladores are not unorganisable. Nor are workers in the myriad small workplaces that make up much of the ‘informal’ sector in Latin America or South Asia. What happened in South Korea can happen elsewhere. But for it to happen requires more than the routinist approach to organisation that characterises trade union bureaucracies worldwide.

So, for instance, an account of the great Bombay textile strikes of 1982-1983 paints a different picture to that in Korea. The strike began as a semi-spontaneous upsurge from below (workers demonstrated outside the residence of Datta Samant, who was to become the strike figurehead, demanding that he ‘lead’ them) and developed into one of the biggest prolonged strikes in world history, lasting a year, involving hundreds of thousands of workers and dominating the political life of India’s commercial and industrial capital. But it never spread from the ‘organised’ sector of the larger workplaces to the small workplaces and the impoverished self employed weavers – indeed, many strikers began working in the informal sector without anyone regarding them as scabs. This enabled the employers to hold out for a year and defeat the workers, since they were never short of finished cloth.

The conditions in the informal sector were appalling:

A visit reminds one of scenes normally associated with the beginning of the industrial revolution: thousands of persons sleeping in or next to numberless ramshackle sheds in which the deafening sound of the power looms is heard 24 hours a day, no ventilation, no proper light, children doing tedious work for long hours, dust and dirt everywhere. [129]

But in Korea there was a network of activists prepared to put up with incredible hardship to organise such places and so to take advantage of the more general upsurge of struggle in 1987. Such a network was lacking in Bombay:

The trade unions active in the textile industry have never felt a strong urge to address themselves to the welfare of the power loom workers among whom there is hardly any degree of unionisation. With few exceptions the unions prefer the easy access to the workers offered by the mills to the tiresome and unrewarding labour of organising the workers in the power looms, although everyone agrees that the fate of the latter is far worse and that unionisation is badly needed. [130]

One union leader who agreed that organisation of such workers was in principal necessary complained of lack of people willing to undertake this arduous task: ‘You must stay with and live with them if you want to gain their confidence’. [131] Similar views are expressed in other quarters explaining the absence of trade union activity in this very important area. This neglect has serious consequences for the organisation of textile labour in the long run. It enables employers to use subcontracting as an integral part of the growth strategy of their enterprises. [132]

The victories in Korea show the possibility of organising informal and maquiladora workers, of pulling them behind struggles initiated by larger and more secure groups of workers. The defeat in Bombay showed the dangers for the more secure groups of not going out and bringing the informal workers into the struggle. The dangers are not simply a matter of wage cuts, job losses and deteriorating working conditions. Defeat can have a devastating impact on wider society. During the strike there was unity between the different religious and caste groups that make up the mass of Bombay’s lower classes. The aftermath of defeat saw the rise to a dominating position in wide areas of the city of the Shiv Sena, a political organisation based upon turning Hindus against Muslims, culminating in murderous riots against the Muslim population in 1992. Unity in struggle had created a sense of solidarity which then exerted a pull on the vast mass of the informal workers, self employed, the unemployed poor and the impoverished sections of the petty bourgeoisie. The defeat led to the sectional attitudes and communal conflicts of the petty bourgeoisie influencing the self employed, the unemployed and wide layers of workers.

It was a vivid example of how there are two different directions the despair and bitterness that exists among the ‘multitudes’ in the great cities of the Third World can go. One direction involved workers struggling collectively and pulling millions of other impoverished people behind them. The other involves demagogues exploiting the sense of hopelessness, demoralisation and fragmentation to direct the bitterness of one section of the impoverished mass against other sections.

That is why the working class cannot simply be seen as just another grouping within ‘the multitude’ or ‘the people’ of no intrinsic importance for the struggle against the system.


The overall picture is not one of a disintegrating or declining working class. It is one of a working class that on a world scale has grown bigger than ever, even if the rate of growth has slowed down with the successive crises in the world economy and the tendency everywhere to ‘capital intensive’ forms of production that do not employ massively new numbers of people.

Neither is the picture one in which working class employment is being transferred on a massive scale from the old industrial economies of the ‘North’ to the previously agrarian economies of the ‘South’. The new international division of labour is developing mainly within the ‘triad’ of North America, Europe and Japan – with a lesser part being played by the NICs of East Asia and eastern China. There is also an expansion of industrial employment within some of the burgeoning cities of the ‘South’ – but the expansion is uneven, barely touching whole regions, and is not mainly through transfer of jobs from the North.

In the North and the South alike there have occurred repeated crises with reorganisation of the structures of accumulation. This is producing a recomposition of the working class, similar in scale to the recompositions which occurred in the latter half of the 19th century when heavy industry began to overtake textiles as the centre of capital accumulation and in the interwar years when light industry and motors began to move to the centre. A twofold change is taking place now. There is the growing importance of the production of certain ‘immaterial’ commodities which are often classified as part of the service sector, but involve forms of work very similar to those in industry. And there is the growing importance of forms of labour which do not themselves produce commodities, but which serve to maintain and increase the productivity of the direct producers.

As these sectors become increasingly important for capital, it reacts by trying to cut their labour costs, producing an increasing proletarianisation of sectors that used to regard themselves as ‘middle class’. Meanwhile, there is increased pressure on the direct producers as well, with increasing intensity of work (disguised as ‘flexibility’) and, in some cases, an increase in working hours – the highest number of working hours per year is to be found in the US, with 1,991 for production workers in manufacturing, as against 1,945 in Japan, 1,902 in Britain, 1,672 in France and 1,517 in Germany. [133]

The working class is not disappearing. It is not becoming bourgoeisified. It is not turning into a privileged layer. It is not gaining somehow from the impoverishment of wide sections of the Third World, especially Africa. It is growing even while it is restructured globally.

The majority of the world’s population does still belong to other subordinate classes. In China, the Indian subcontinent and much of Africa the peasantry outnumber the workers. There are cases in Africa and parts of Latin America of those unable to find work in the cities attempting to re-establish themselves as small farmers. In some of the world’s biggest cities, the permanent workers are outnumbered by the floating population of the self employed, the unemployed and those with occasional casual jobs. In the advanced industrial countries there still exists the old petty bourgeoisie of small shop keepers, publicans, small businessmen and professionals, and alongside it is the new middle class of middle managers.

Workers often live, work and have family ties with members of these other classes. They can be influenced by the mood of these other classes – but they can also exert a decisive influence on the mood, as we saw with the case of the Bombay textile workers.

Certain issues encourage such different groupings to fight together. Community struggles erupt which unite all those who live in a certain lower class locality, regardless of the way in which they make their livelihood. They can share the experience of taking to the streets and of confronting those at the top of society together. It is in these struggles that notions of ‘the masses’, ‘the people’ the ‘multitude’ or the rainbow coalition seem to fit better than the notion of class. The most recent examples of such mass, multi-class upsurges were the wave of cacerolazo demonstrations from the inner city neighbourhoods of Buenos Aires that swept the De la Rua and Rodriguez Sáa governments from power in Argentina at the turn of the year – and the neighbourhood asemblea organisations that grew out of them. [134]

The anti-capitalist movement itself has some of the same characteristics. Its initial base, like that of the first movement of the late 1960s, has been among people not firmly rooted in the productive process – students, school students, young people not yet trapped into permanent jobs, workers who take part in its activities as individuals without any clear sense of class identity, lower professionals. As a descriptive term for such movements, ‘multitude’ is not completely misplaced. A disparate coalition of forces has come together to provide a new and massively important focus for the struggle against the system after two decades of defeat and demoralisation.

But the glorification of disparateness embodied in the term prevents people seeing what needs to be done next to build the movement. It does not recognise that what was so important about Genoa and Barcelona was the beginning of the involvement of organised workers in the protests. It fails to locate the most important deficiency of the movement in Argentina to date – the ability of trade union bureaucracies to build a wall between employed workers on the one hand and the neighbourhood and unemployed workers’ movements on the other.

The mistake is to see movements of disparate social groups as ‘social subjects’ capable of bringing about a transformation of society. They are not. Because their base is not centred in collective organisation rooted in production, they cannot challenge the control over that production which is central to ruling class power. They can create problems for particular governments. But they cannot begin the process of rebuilding society from the bottom up. And in practice, the workers who could begin to do this only play a marginal role in them. Talk about ‘rainbow coalitions’ or ‘multitudes’ conceals that relative lack of involvement in the movement of those working long hours at manual or routine white collar jobs – and with extra hours of unpaid labour bringing up children. It underplays the degree to which the movements remain dominated by those whose occupations leave them most time and energy to be active. Fashionable theories about ‘post-industrial society’ then become an excuse for a narrowness of vision and action that ignores the great majority of the working class.

What has been wonderful about the last two and a half years since Seattle is the way in which a new generation of activists has arisen to challenge the system. But what increasingly matters now is for this generation to find ways to connect with the great mass of ordinary workers who as well as suffering under the system have the collective strength to fight it. That is the lesson of Genoa. That is the lesson of Buenos Aires. That is the lesson ignored by those who provide a distorted account of the realities of production under present day capitalism, writing off the class whose exploitation keeps the system going.


1. M. Hardt and A. Negri, Empire (Harvard 2001).

2. For a survey of the literature taking this approach, see Introduction, in J.H. Goldthorpe, D. Lockwood et al., The Affluent Worker in the Class Structure (Cambridge 1971).

3. This was how many people saw Argentina’s car workers in the city of Cordoba – until they played a vanguard role in the Cordobazo uprising of 1969. So Aricó wrote in 1964 that ‘the industrial proletariat of the big enterprises ... constitutes in a certain sense a relatively privileged group, a labour aristocracy ... which enjoys high wages because their class brothers – unskilled workers, peones, rural proletarians, etc. – earn miserable wages’, while Carri saw their unions as ‘the principal means of imperialist penetration in the working class’. Both quoted in R Munck et al., Argentina From Anarchism to Peronism (London 1987).

4. N. Klein, No Logo (London 2000), p.223.

5. M. Hardt and A. Negri, op. cit., p.53.

6. Ibid., pp.402-403.

7. Ibid., p.403.

8. L. Rozichtner, El lugar de resistencia, Página 12 (Buenos Aires), 26 February 2002.

9. C. Harman, A People’s History of the World (London 1999), p.615.

10. D. Filmer, Estimating the World at Work, background report for World Bank, World Development Report 1995 (Washington DC 1995). Available on World Bank website,

11. That is, ‘Mining and quarrying, manufacturing, gas, electricity and water, and construction.’

12. That is, ‘Trade, transport, banking, commercial services, not adequately defined or described.’

13. There are another 1,200 million people of working age whose labour is for their own households, and so is not counted, even though many of them, especially in the countryside, would also have been involved in assisting in other sorts of work.

14. See, for example, my calculation for the size of the new middle class in Britain, in C. Harman, The Working Class After the Recession, International Socialism 33 (Autumn 1986).

15. UNDIP, Human Development Report 1998, Table 21, p.175.

16. Ibid.

17. Ibid.

18. S. Rodwan and F. Lee, Agrarian Change in Egypt (Beckenham 1986).

19. Danyu Wang, Stepping on Two Boats: Urban Strategies of Chinese Peasants and Their Children, in International Review of Social History 45 (2000), p.170.

20. Ibid.

21. S. Rodwan and F. Lee, op. cit.

22. Figures given by C.H. Feinstein, Structural Change in the Developed Countries in the 20th Century, Oxford Review of Economic Policy, vol.15, no.4 (Winter 1999), table A1.

23. Introduction, in R. Baldoz et al., The Critical Study of Work: Labor, Technology and Global Production (Philadelphia 2001), p.7.

24. M. Hardt and A. Negri, op. cit., p.286.

25. All the figures in this section are from C.H. Feinstein, op. cit.

26. R.E. Rowthorn, Where are the Advanced Economies Going?, in G.M. Hodgson et al. (eds.), Capitalism in Evolution (Cheltenham 2001), p.127.

27. Ibid.

28. Report in Financial Times, 12 February 2002.

29. R.E. Rowthorn, op. cit.

30. Ibid.

31. Ibid., p.131.

32. I spelt out some of the arguments here in greater depth in C. Harman, Explaining the Crisis: A Marxist Reassessment (London 1984), pp.105-108.

33. R. Taylor, Britain’s World of Work: Myths and Realities (ESRC Future of Work Programme Seminar Series, Swindon, May 2002).

34. The phrase used ibid. Much of the discussion on the left over skills over the last quarter of a century has been influenced by the ‘deskilling’ argument presented in H. Braverman, Labor and Monopoly Capital (New York 1974). But disappearance of particular skills learnt through long years of apprenticeship in industry has, in general, been accompanied by the rise in the average level of literacy and numeracy required to handle a range of jobs continually changing due to technological innovation. For a more sophisticated investigation of these issues than Braverman’s, see C. McGuffie, Working in Metal (London 1985).

35. R. Taylor, op. cit., p.18.

36. I discuss the notion of the ‘new middle class’ at greater length in C. Harman, The Working Class After the Recession, op. cit., pp.22-25.

37. The Guardian, 5 June 2002.

38. M. Hardt and A. Negri, op. cit., pp.285-286.

39. Office for National Statistics, Labour Force Survey (London 2001).

40. Office for National Statistics, Living in Britain 2000, table 3.14, available on

41. R. Crompton and G. Jones, White Collar Proletariat (London 1984), p.27.

42. Ibid., p.20.

43. All the figures are from Employed Persons by Occupation, Age and Sex, on pub.specia.requests/lf/aat9 [WebEd: I couldn’t find this reference but similar figures seem to exist at]

44. Office for National Statistics, Social Trends 2001 (London 2001), p.82.

45. C. Harman, The Working Class After the Recession, op. cit., pp.22-25.

46. As I have argued elsewhere, restructuring within the three great regions of the industrialised world – North America, Europe and East Asia – has been more important than restructuring across the whole globe. See C. Harman, Globalisation: A Critique of a New Orthodoxy, International Socialism 73 (Winter 1996).

47. R.-P. Bodin, Wide-Ranging Forms of Work and Employment in Europe, The Future of Work, Employment and Social Protection, International Labour Organisation,, p.1.

48. R. Taylor, op. cit., p.7.

49. R.-P. Bodin, op. cit., pp.3-4.

50. Ibid., p.5.

51. Ibid., p.2.

52. Ibid., p.2.

53. Ibid., pp.2-3.

54. R. Taylor, op. cit., p.12. The slight discrepancy between these figures and the European figures for Britain is not significant, since they are based on different surveys that come to very similar results.

55. These figures are from Office for National Statistics, Social Trends 2001, op. cit., table 4.6, p.88. The results are very similar in the survey in R. Taylor, op. cit., p.13.

56. Office for National Statistics, Social Trends 2001, op. cit., p.88.

57. M. Hardt and A. Negri, op. cit.

58. See C. Harman, The State and Capitalism Today, International Socialism 51 (Summer 1991), and C. Harman, Globalisation: A Critique of a New Orthodoxy, op. cit.

59. R.E. Rowthorn, op. cit., p.136.

60. Ibid., p.135.

61. Ibid., pp.131-132.

62. R. Baldoz et al., op. cit., p.9.

63. Ibid., p.7.

64. This, for instance, is the impression given by Naomi Klein in No Logo, when she writes about ‘General Motors ... moving production to the maquiladores and their clones around the globe’, N. Klein, op. cit., p.223.

65. Although, I personally was quite surprised at the sophistication of the equipment, with computer terminals attached to sewing machines, in the Bruckman factory in Buenos Aires, which had been occupied by its workforce.

66. F. Palpacuer, Development of Core-Periphery Forms of Organisation: Some Lessons from the New York Garment Industry, International Labour Organisation,

67. Figure given in A. Lateef, Linking Up with the Global Economy: A Case Study of the Bangalore Software Industry (International Labour Organisation, 1997),

68. M. Hardt and A. Negri, op. cit., p.294.

69. Figure given in A. Lateef, op. cit., ch 2, p.9.

70. Figure given ibid., ch 4, p.1.

71. Ibid., p.3.

72. Ibid., p.15.

73. Ibid., p.9.

74. Ibid., p.10.

75. Ibid., p.11.

76. Very rough calculations using and adjusting the figures contained in the tables in D. Filmer, op. cit.

77. International Labour Office, African Employment Report 1990 (Addis Ababa 1991), p.31.

78. Ibid., p.26.

79. Ibid., p.44 .

80. Figures from PRELAC Newsletter (Santiago, Chile), April 1992, diagram 3.

81. Ibid.

82. Ibid.

83. P. Singer, Social Exclusion in Brazil (International Labour Office, 1997), ch 2, table 7, available at

84. Ibid., p.17.

85. P. Nayak, Economic Development and Social Exclusion in India, ILO 1994, ch 2, p.1, available at Figures on employment from the 2001 census are not yet available.

86. Figures in J. Unni, Gender and Informality in Labour Markets in South Asia, Economic and Political Weekly (Bombay), 30 June 2001, p.2367.

87. The figures are provided, with sources, ibid., p.2369.

88. Ibid.

89. T. Bulutay, Employment, Unemployment and Wages in Turkey (Ankara 1997), p.196.

90. Ibid., p.193.

91. Ibid., p.200.

92. Economic Trends in the MENA Region, 2000, html
[WebEd: This is the only relevant link I can find:]

93. Ibid., ch 4.

94. P. Singer, op. cit., ch 2, table 10.

95. Yun-min Lin and Tian Zhy, Ownership Restructuring in Chinese State Industry, China Quarterly, June 2001, p.307.

96. China Labor Bulletin, Hong Kong, 2001.

97. According to Financial Times, 26 October 2001. Callum Henderson suggests a similar figure for the total number of unemployed in China. See C. Henderson, China on the Brink (New York, 1999), p.20.

98. P. Singer, op. cit., ch 2, p.3.

99. Ibid., ch 2, p.14.

100. See, for instance, the figures given in J. Unni, op. cit., tables 19, 20 and 22, pp.2375-2376. There are, of course, situations in which a sudden demand for labour can only be met from the informal sector, leading to wage rates temporarily above those in the formal sector. The same phenomenon occurs, for instance, with ‘lump’ labour in the building industry in Britain.

101. S. Gordon, Poverty and Social Exclusion in Mexico, ILO, 1997, p.10

102. Ibid.

103. Ibid.

104. PRELAC Newsletter, op. cit.

105. K. Marx, Capital, vol.1 (Moscow 1961), p.628.

106. Ibid., p.631.

107. Ibid., p.642.

108. Ibid., p.643.

109. Ibid., p.644. Engels provided a detailed empirical description of the ‘surplus’ pool of labour in the 1840s in his The Condition of the Working Class in England in K. Marx and F. Engels, Collected Works, vol.4 (London 1975).

110. ILO, African Employment Report 1990, op. cit., p.34.

111. Ibid., pp.37, 39.

112. Figure, ibid., p.40.

113. H. Steefkerk, Thirty Years of Industrial Labour in South Gujarat: Trends and Significance, Economic and Political Weekly (Bombay), 30 June 2001, pp.2399, 2402.

114. P. Singer, op. cit., ch 2, p.16.

115. H. Steefkerk, op. cit., pp.2399, 2401.

116. Quoted ibid., p.2402.

117. Ibid.

118. P. Singer, op. cit., ch 2, table 9.

119. Ibid., p.17.

120. F. Engels, Letter to Bernstein, 22 August 1889, in K. Marx and F. Engels, Collected Works, vol.48 (London 2001).

121. N. Klein, op. cit., pp.205-206.

122. For some of the efforts made by firms like Tesco and Coca-Cola to achieve such ‘governance’ over global chains, see R. Kaplincky, Globalisation and Unequalisation, and C. Dolan and J. Humphrey, Governance in Trade in Fresh Vegetables, both in Journal of Development Studies, vol.57, no.2 (December 2000).

123. K.A. Ver Beek, Maquiladoras: Exploitation or Emancipation, in World Development, vol.29, no.9 (September 2001). As the title suggests, this study is apologetic for the maquiladora firms, but there is no reason to believe the figures are phoney.

124. T. O’Brien, A Century of US Capitalism in Latin America (New Mexico 1999), p.52.

125. G.E. Ogle, South Korea: Dissent Within the Economic Miracle (London 1990), p.82.

126. Ibid., p.106.

127. Ibid., p.116.

128. Ibid., p.145.

129. H. van Wersch, The Bombay Textile Strike 1982-1983 (Bombay, 1992), pp.45-46.

130. Ibid., p.46.

131. Ibid.

132. Ibid.

133. Ministry of Health, Labour and Welfare, White Paper on the Labour Economy (Tokyo), 18 July 2001. The figures for France are for 1998 and for Germany for the former West Germany in 1997. The figures for Britain are higher than those normally quoted, which are lower because based on all workers, not just those in manufacturing (and exclude vast amount of non-recorded, unpaid overtime by white collar workers).

134. For a longer description of these, see C. Harman, Argentina: Rebellion at the Sharp End of the World Crisis, International Socialism 94 (Spring 2002).

Last updated on 5 May 2021