Skip to Content

Will Robots Cause Capitalism to Collapse?

In March the Socialist Party debated with Federico Pistono, the author of a book entitled Robots Will Steal Your Job, But That’s OK: How to Survive the Economic Collapse and Be Happy (reviewed in the March Socialist Standard). His argument is typical of many who think that the market-money-wages system we call capitalism is going to soon collapse as a result of the increasing pace of technological innovation leading to constantly growing mass unemployment. Peter Joseph, the founder of the Zeitgeist Movement, put it more dramatically in a TV interview the same month:

‘unemployment is a consequence of technology, entirely. The entire reason we have unemployment in America and across the world is explicitly based on the application of technology for cost efficiency. And this is not going to stop. And this will lead to what has been called by theorists the ‘contradiction of capitalism’, to the ultimate instability of our social system: the ability to produce more with less people and cheaper rates. It’s a complete clash of the system.’ (

A computer whizz-kid himself, Pistono describes various already-existing inventions that can displace humans in production, particularly due to advances in computer technology. Two he discusses in detail, because of their impact on relatively unskilled labour, are automatic vending machines (which would replace shop assistants) and driverless road vehicles (which would replace van and lorry drivers). He then asks why, if this is all possible, we are not seeing it:

‘Sounds futuristic? Every piece of technology needed for making this happen already exists, and has existed for many years. Then why is it not in place already?’

Good question. Why indeed?

Why machines are (and aren’t) introduced

When a machine is introduced in a particular production unit this reduces the number of workers (living labour) required there to produce the same goods or provide the same service. But, since the machine had to be produced by living labour, extra workers must have been taken on somewhere else to build it, so the question arises of whether the two effects on employment cancel each other out at the level of the economy as a whole.

At first, economists tried to argue that this was so but they soon recognised that they were mistaken and conceded that there would be a net reduction in the total level of employment, not as great as the number of workers displaced in the productive units affected but to a level less than previously. In other words, machines sack more workers than they take on.

Writing in 1821 not long after the Luddites had been smashing knitting machinery, David Ricardo concluded:

‘That the opinion entertained by the labouring class, that the employment of machinery is frequently detrimental to their interests, is not founded on prejudice and error, but is conformable to the correct principles of political economy’( Principles of Political Economy and Taxation, third edition, Ch. 31).

He added that this reduction in total employment could, and normally would, be offset if the economy expanded as a result of new capital investment in some other field of activity. Marx, writing nearly fifty years later, agreed. This – the expansion of capitalist production – is the reason why the introduction of machinery in the past has not resulted in steadily increasing mass unemployment.

Marx made a further point about the introduction of machinery: for a machine to be genuinely ‘labour saving’ in the sense of reducing the total labour-time required to produce something from start to finish, ‘less labour must be expended in producing the machinery than is displaced by the employment of that machinery.’ (Capital, Vol 1, Ch. 15, section 2). By ‘labour’ Marx meant not simply ‘living labour’ or its immediate product but also the ‘dead labour’, the product of previous living labour, incorporated in the raw materials, energy, buildings and machinery used in production.

In a rationally-organised society based on the common ownership of productive resources so that production can be carried on to produce directly to satisfy human needs instead of for profit whether or not a machine did this would be the main criterion for deciding whether to apply it to production. Not all inventions of machines do displace more labour than would be required to produce them but, in a rationally-organised society, even machines falling into this category could be introduced if it was considered that the specific labour that would be replaced was considered dangerous, unhealthy or boring.

But this is not what happens under capitalism. Built-in to the capitalist system is a drag on the use of machines. As Marx went on to explain:

‘For the capitalist, however, there is a further limit on its use. Instead of paying for the labour, he pays only the value of the labour-power employed; the limit to his using a machine is therefore fixed by the difference between the value of the machine and the value of the labour-power replaced by it.’

Under capitalism the immediate product of living labour is divided into a part that the capitalist firm has to pay for (wages) and a part that it doesn’t pay for (surplus value, the source of profit). This means that under capitalism a machine that would genuinely save labour – the time society has to spend to produce something – would only be introduced if it also reduced the total labour that the capitalist firm had to pay for, i.e. the dead labour incorporated in the machine and materials plus the living labour it employs. If this is not the case, then the labour-saving machine will not be introduced, as to do so would reduce the amount of unpaid labour that the firm extracts, i.e. the source of its profits. In fact, the lower wages are, the less the incentive to apply labour-saving inventions, and vice versa.

Marx gave some concrete examples to illustrate that under capitalism there is a difference between invention and application:

‘Hence, the invention nowadays in England of machines that are employed only in North America, just as in the sixteenth and seventeenth centuries machines were invented in Germany for use exclusively in Holland, and just as many French inventions of the eighteenth century were exploited only in England … The Yankees have invented a stone-breaking machine. The English do not make use of it because the ‘wretch’ who does this work gets paid such a small portion of his labour that machinery would increase the cost of production to the capitalist.’

This is the answer to Pistono’s question as to why the futuristic labour-saving inventions he describes have not been used on a wide scale: capitalist firms are only interested in using machinery if it will reduce their costs of production, i.e. the labour (living and dead) that they have to pay for. They will not and do not introduce machines that will increase their costs of production, even if their use would reduce the total amount of labour required to produce them.

This is why, as long as capitalism lasts, the rate of the actual application of machinery to production will always be less than the rate at which labour-saving machines are invented. In this sense speculations such as Pistono’s about the rate at which inventions will increase (he claims, somewhat hyperbolically, that this will become exponential in the coming decades) is irrelevant. It is not this that will determine the rate at which inventions will be applied to production as the rate at which ‘robots will steal our jobs.’ That will depend on the rate at which they reduce the labour that a capitalist firm has to pay for, which will be considerably slower than the rate at which labour-saving machines are invented.

Under capitalism invention is one thing, application another. The mere invention of some labour-saving machine does not destroy jobs; only its application does.

Will history repeat itself?

The trouble with many theories of economic collapse is that, if they were true, they need to explain why capitalism has not already collapsed a long time ago. Pistono’s argument is no exception. He is aware of this, as he quotes one critic as exclaiming:

‘Have you ever heard of this discipline called history? We’ve gone through the same crap 150 years ago, and none of what you say has happened!’

It’s actually a good question. Mechanisation has been going on since the Industrial Revolution started in the eighteenth century (in fact, that’s what this was) but it has not resulted in steadily increasing unemployment. Pistono’s reply is that it will be different this time as in the past the pace of labour-saving technological inventions has never been as fast as it is today.

We have just seen that Pistono commits the fallacy of confusing technological invention with the application of inventions to production. Even so, as in the past mechanisation has not resulted in growing technological unemployment, since the capitalist system expanded to absorb this, a weaker version of Pistono’s contention might still be valid: that the rate at which machines are introduced (despite the restrictions on this under capitalism) might still be faster than capitalism can expand. In this case unemployment would still grow.

While capitalism does expand in the long run it does not expand, as everyone is agreed nowadays, in a straight upward-sloping line. It goes in fits and starts, booms and slumps, with each succeeding boom reaching a higher level of production and employment than the previous one.

Because capitalism grows in this way, it needs a pool of unemployed workers, which Marx called ‘the industrial reserve army of labour’, that capitalist firms can draw on quickly in a period of boom and who become unemployed again when the slump comes. So, unemployment rises and falls with the capitalist business cycle.

Pistono does not go as far as Peter Joseph and claim that all unemployment today is technological, but he does advance the increase in unemployment since 2007 as proof of his contention that the increasing application of modern technology is causing unemployment to grow.

Some of today’s unemployment will indeed be technological in the sense of being living labour displaced by machines and unable to find new employment because capitalism is not expanding (Ricardo’s worst case scenario), but most is cyclical, the result of capitalism currently being in one of its periods of slump. It is the industrial reserve army of labour returning to its slump level. Also, capitalist economists talk cynically about a ‘natural’ level of unemployment (the rate below which they say a rise in price level would result). So, by no means all unemployment today is technological; in fact, only a relatively small proportion will be.

Predictions of a continual increase in mass unemployment will only turn out to be true if capitalism does not recover from the present slump, and even then wouldn’t increase at the rate this hypothesis suggests. If it does recover then unemployment will fall.

So the question can be reformulated as: Will capitalism recover from the slump or will unemployment go on increasing until the system collapses? Both past experience and the theory of how capitalism works based on this suggest that capitalism will eventually recover, however long it takes and however hard workers have to be squeezed. There is no way of knowing, though, exactly how long it will take.

In any event, capitalism will have to be ended by the conscious action of people who want to replace it by a system where the resources of the planet have become the common heritage of all. Then, there will no longer be any barrier to the robotisation of repetitive and boring jobs. Then, robots will ‘steal our jobs’ much more quickly than today and that will be OK, as there will be no harmful side-effects since access to what people need to live and enjoy life won’t be tied to working for a wage or salary. As Marx put it, ‘the field of application for machinery would therefore be entirely different in a communist society from what it is in bourgeois society.’