Book Review: Aaron Benanav, “Automation and the Future of Work”

Does automation mean the end of work? In this relatively brief essayistic work, Aaron Benanav – fellow Berliner and sociologist – engages critically with one of the main topics that have found new and fruitful engagement in Marxist theory over the past two decades or so: the topic of automation and its relation to freedom. Advances in robotics, artificial intelligence, computing (or at least computational power), and algorithmic coordination have advanced so greatly in recent years that they have led to a reconsideration of the future of labor under conditions of strong labor-saving technological change. For many, not least in the Western labor movements, this has been a cause for concern and even fear. But others have seen in these developments the unfolding of the potential required to make a qualitative leap into a new kind of relation to work, and therefore a new relation to human freedom.

The tendency to see potential in these developments, rather than merely threat, is primarily found among the intellectual thought of the contemporary left. Certainly, for some reactionaries the thought of pesky labor organizations and demanding First World workers being shunted out by automation and being relegated to cowed serfs depending on a new corporate-feudal oligarchy for handouts is an attractive proposition. Among some liberal economists, there has been worry about the implications of shifts in the labor market for social stability and for the ability to contain ever-rising inequality. But among those engaging with real-world developments, it is primarily the left that has drawn attention to this process and its implications in our times. (Certainly the theme as such is not new, but this orientation was true in past iterations of automation discourse as well: examples are such figures as André Gorz, James Boggs, and Herbert Marcuse, as Benanav points out.)

Moreover, an important strand of Marxist thought, with its pedigree in writings of Marx and Engels themselves, points to automation and the labor-saving process as the foundation for the potential to reduce the total workload of society, and thereby to free up humanity for its true development. The development of the productive forces makes much human labor objectively unnecessary, even if under capitalism only labor has the ability to add value and is therefore essential to the system’s reproduction. This contradiction can be blown up with capitalism itself in order to free up humanity from its unneeded toil, leaving the machines, robots, and computers to do most of the work when humans no longer depend on such work for a living. In recent years, this has found a new impetus with among others the work of Srnicek and Williams in their Inventing the Future (Srnicek and Williams 2016) and sometimes ‘semi-ironically’ referred to by the loose slogan “fully automated luxury communism”, associated with the left media network Novara.

It is with this tendency that Benanav’s book critically engages, marshalling both economic-historical and theoretical arguments. His core thesis is that automation, which he defines as fully labor-saving technology as opposed to technology that merely increases productivity of existing labor, appears as a sense of crisis or a tipping point because of a persistently low demand for labor. This low demand is not just in evidence in the rich countries, but across the globe. However, against the automation utopians, Benanav argues contemporary automation is not the actual cause of that low demand for labor. Instead, he locates the causes of this low demand in the long term of secular capitalist stagnation that set in after the boom period of capitalism was over, i.e. in the 1970s. Since then, as Benanav shows, rates of output growth have slowly but systematically declined throughout the world.

The crux of his argument is not this familiar observation, however, but to point out that rates of productivity growth have also systematically declined since then, quite contrary to what the bullish talk of the automation theorists would suggest. During the boom period of capitalism (1950s-1970s) and as part of the process of economic development and reconstruction in the postcolonial era, manufacturing output enormously increased. Because the world market is a capitalist market, this tremendous increase in output caused significant and persistent overproduction at the global level (tempered only somewhat by population growth). This overproduction has not only hindered the possibilities of economic development through industrialization in much of the Third World, but it also caused rates of growth in manufacturing to plummet, causing deindustrialization. Indeed deindustrialization has been such a potent process that it has not only stripped much of the Western world of its industrial workforce, but even such ‘workhouses of the world’ as China have seen declining shares of manufacturing employment. Manufacturing output growth rates have over time become so low that productivity growth in this sector is consistently greater than the rate of output growth, implying the systematic shedding of employment.

This absolute surplus workforce has been shunted off into the service sector. But the service sector has the peculiarity that it is particularly unamenable to significant growth in sectoral productivity. (Indeed, as Benanav mentions, the primary way this tends to happen is by a process one could call ‘reindustrialization’: instead of launderettes you have washing machines, and fast food has subjected cooking to a factory mode of organization.) This leads to even lower rates of output growth and an even more sluggish accumulation rate overall, with declining rates of return inducing enormous financial bubbles for want of places to invest.

So as overproduction in global manufacturing has crashed output growth rates and taken productivity growth with it, it is not rising productivity rates but falling output rates that are the real drivers of global stagnation. The result is widespread underemployment and increasing inequality, as organized labor has been unable to adjust to the new feedback loop of stagnation and has found it harder to organize in the service sector. Meanwhile many governments have pursued neoliberal austerity policies aimed at reducing the costs of labor, both to remain competitive in the global market and to force workers into accepting an absolute increase in the rate of exploitation. Underinvestment resulting from the lack of returns for private capital – and almost as poor prospects of improvements in productivity across the economy as a whole – lead to technology hypes and huge speculative inflation of fixed assets like real estate as capital seeks a way out of the cycle of stagnation.

Seen from this vantage point, Benanav proceeds to critique some of the more utopian or at least optimistic solutions proposed by the automation theorists, in particular his comrades on the left (whom he repeatedly avows sympathy with, even as he disagrees on the analysis). Based on the preceding analysis, he critiques the proposals from the left (and some of the libertarian tradition) for a form of Universal Basic Income (UBI), as well as more traditional Keynesian and social-democratic amelioration strategies. What these have in common, Benanav argues, is that they essentially ignore that the power of capital over the production process and over investment and distribution of resources is the basis for the political power of capital, and that in order to counter the ‘rational’ decisions of capital in the pursuit of profit under conditions of stagnation an enormous amount of political force and power will be required.

Essentially Benanav’s critique of these approaches is in line with the traditional Marxist critique of social-democracy (in the post-1918 sense): if you have enough power to force capital to accept a UBI that supplements workers’ bargaining position rather than undermining it, or to accept something like the Meidner Plan, then you also have the power to overthrow capital as such, and should rather do so. In the conclusion, he points in that same direction: a regime of social relations based on maximizing free time, equal distribution of burdens, and rational collective control over economic decision-making is a matter of political will and strength, and does not depend on any particular level of automation or the coming of the Great AI.

Indeed the great strength of the book is precisely this: to remind the left automation theorists that while their Marxist analysis of the potential of liberation from work is quite correct, what they should not neglect is the other part of the Marxist theoretical legacy, namely the critique of capitalist production. Although Benanav mostly eschews using Marxist economic jargon and uses mainstream authors like Leontieff and Baumol wherever possible, in essence his argumentation is based on a demonstration of the operation of the falling rate of profit: capitalism’s own success undermines its conditions of accumulation, and this causes stagnation and bubble formation until sufficient value is destroyed to reboot the cycle. Benanav’s emphasis on the particular significance of manufacturing as the “engine of economic dynamism” is a reflection of the role of commodity production for the production, and therefore possible accumulation, of surplus value. Although large ‘service sectors’ are often portrayed as a sign of a modern and mature economy and therefore desirable, from a Marxist point of view it is ultimately the regime of surplus value generation and accumulation that determines the ‘health’ of the beast, not the distribution of this value across all other sectors. (Needless to say, the health of the beast is not the same as the health of the workers trapped in its belly, in either case.)

The point of all this is not to hammer one over the head with theory, but to point out that left engagement with automation and other forms of qualitatively striking technological change should not forget what Kliman (Kliman 2011) has aptly called the failure of capitalist production. Even for left theorists it can be tempting to take capital in a sense at its word, to be led to the mountaintop and shown the dazzling vistas that lie ahead and to think what we could do with all those tools and riches. But not all that glitters is gold. For all the hype generated by the tech billionaires and the slow but not negligible rate of improvement in global living standards from increased output (which the author acknowledges), capitalism is reminiscent of the man who sticks a bar between the spokes of his own bicycle: sooner or later it brings itself to a fall, and then looks around indignantly for who might have sabotaged its great promise. But it is its own mode of production and accumulation that does this, and it cannot help but do it, again and again. No amount of machine learning or robotics can alter this fate of its own power, whatever other potential the computers and robots may have.

That said, Benanav’s own suggestions for the norms of a noncapitalist world are in the foundations not very different from those discussed in such left automation theory works as Srnicek and Williams’ or Peter Frase’s Four Futures (Frase 2016), also premised on a fully automated future. Indeed, with its discussions of collective distribution of work and its slightly blithe argument that the denizens of the future socialist world should just agree with each other on how to solve economic challenges, it reminds one a bit of the slightly older tradition of such postcapitalist social design like Parecon (Albert 2004).

On this basis, I would suggest it is not so much to be read as a critique of left automation theory as a supplement to it, one that more systematically incorporates the Marxist critique of capitalist production as not just wasteful of its potential, but also as subject to (irrational) secular tendencies to stagnation and unemployment. Insofar as the book has weaknesses, it is mainly in this latter sphere, the discussion of the low global demand for labor. Benanav tends to treat the First World and the global South as fairly equivalent and subject to the same processes in this regard, but thereby tends to elide the structural differences between them. After all, the vast majority of global inequality is inequality between nations, not within them, and structurally huge unemployment and underemployment are characteristics of poorer nations more than of richer ones. Premature deindustrialization is certainly a matter of the world market, as Benanav rightly says, but that world market is not born equal, and likely his argumentation works all the more clearly and transparently in countries with an educated but wildly underemployed workforce like Syria or Egypt than in the Western metropole.

This also raises the question of what to do with the traditional Marxist argument of the need to develop the productive forces to achieve the necessary potential for a postcapitalist future – a cornerstone of both Marxist and non-Marxist state practice in the 20th century in the form of the developmental state, but an issue not much addressed by Benanav. This is slightly curious since precisely this insight is at the heart of the (so to speak) ‘left optimists’ in the automation discourse: it is precisely the appearance of a very advanced state of development of the productive forces that invites the suggestion that the potential for utopia is around the corner. This may not be justified, and others have historically hewed closer to Benanav’s apparent approach that the leap to freedom is a matter of political will more than of technological capacity. But a further discussion of this theme would, I think, add to the filling out of Marxist theory that the left automation theorists are so fruitfully and – dare I say – productively engaged in.

Overall, Automation and the Future of Work is a remarkably lucid and sharp argument made in hardly a hundred pages. It fits well together with the previously mentioned works of Srnicek and Williams as well as Frase as part of a larger (and very welcome) exchange of thoughts among left theorists on the implications of modern automation technology and the systematic restructuring of the global workforce since the 1980s. It adds to this discussion a firm rooting in the Marxist theory of capitalist production and accumulation as well as more neoclassical interpretations of the ‘secular stagnation’ thesis, which point in the same direction. Most importantly, by emphasizing declining output growth rather than increasing productivity growth, it provides an important empirical and theoretical check on overly bullish interpretations of capitalist Promethean power.

Cited:
Benanav, Aaron. Automation and the Future of Work. London 2020: Verso.
Frase, Peter. Four Futures: Life after Capitalism. London 2016: Verso.
Kliman, Andrew. The Failure of Capitalist Production: Underlying Causes of the Great Recession. London 2011: Pluto.
Srnicek, Nick and Alex Williams. Inventing the Future: Postcapitalism and a World Without Work. London 2015: Verso.