The world is on the cusp of the greatest transformation of the nature of work since the industrial revolution of the 19th century, if not, arguably, since the invention of the wheel. The confluence of developments in cybernetics, robotics, and artificial intelligence will, in all likelihood, result in transformations so profound as to rewrite the rules governing human societies, if it does not wipe them out entirely.
The logic of capital, imbricated in the rationality of the industrialized world and its less developed clients, is driving forward a polycephalus developmental complex, the minimal outcome of which will be the end of the mode of liberal capitalism dominant since the last decades of the 19th century. And what will its replacement be? To listen to the nostrums emitting from the various tech-nodes and research institutes throughout Europe, North America, and Asia, one would think that we are on the verge of a golden age, one of enhanced leisure where human creativity will have a scope previously unexampled in human history. Transfixed by a combination of self-interest and naiveté, the boosters of the imminent techno-order have either failed to recognize, or are too fatalistic to oppose, the prospect of a darker future for humanity, one in which extinction might not even be the worst outcome imaginable.
The topic of the effects of robotics technology on employment and economic development seems to have caught the attention of mainstream economists in the last couple of years. The economic downturn resulting from the collapse of the housing bubble and the concomitant disruptions in the repo market set off one of those waves of creative destruction of which the nattering nabobs of neoliberalism among the financial press are so fond. Yet this was a slightly different moment of creative destruction than that envisioned by Schumpeter when he (following a close reading of Marx) popularized the term in the early 1940s.
The innovation, which sparked the Schumpeterian gale, was not industrial but rather financial, based on novel ways of parsing and marketing risk. When the claims of the quants to have effectively understood and managed the underlying risks were shown to be vain, the global economy experienced a downturn comparable to (and in some respects worse than) that of the late 1920s. In the wake of the collapse, as companies struggled to regain the levels of production and profit of the first half of the decade, a clear trend emerged: a shift of capital outlay from human to technological resources. Writing in the New York Times in June 2011, Catherine Rampell noted:
“Two years into the recovery, hiring is still painfully slow. The economy is producing as much as it was before the downturn, but with seven million fewer jobs. Since the recovery began, businesses’ spending on employees has grown 2 percent as equipment and software spending has swelled 26 percent, according to the Commerce Department.”
This was not the first time that this had happened. Indeed, it is often case (as for instance in the wake of the recession of the early 1980s) that businesses use the necessity of rebuilding after a downturn to reconfigure production in ways that reduce the wage bill. But this is also part of a large pattern of displacing costs, of which the trend towards the automation of industrial production is the most profoundly consequential. As Paul Krugman mentioned in an article from a couple of years ago, economists often refer to this as “capital biased technological progress.” It might actually be more accurate to call it “technologically biased capital development,” as what’s really going on is a rejigging of the ratio of capital expenditure between technological and human inputs.
Krugman had earlier pointed out that the rise of mechanization in industry was progressively cutting the legs out of the proposition, so favored among right-wing commentators, that augmented skills and education were the path to resolving the problem of mass un- (and under-) employment. As Martin Ford has argued in a recent book (Rise of the Robots: Technology and the Threat of a Jobless Future), the increasingly versatile tasks for which robots can be used, along with the expanded availability of 3D printing, could mean the end of most forms of industrial employment.
But, as Ford comments, it is not only these jobs, the occupants of which are often sententiously informed that they must adapt (change locations, get more education, etc.) or lose out. The combination of advances in computing power and algorithmic understandings of human action means that the jobs of white-collar workers are also under threat. Holders of these positions, be they lawyers or doctors, software engineers or professors, are fond of creating narratives in which their place in the income distribution is the result of the possession of some sort of superior virtue (such as hard work in the acquisition of skills, or the ability to cope with extreme competitive stress.)
Putting aside, for the time being, the degree to which these narratives merely mask the underlying reasons for the levels of remuneration being what they are (and these tend to have much more to do with the vagaries of supply and demand than of any inherent personal virtue,) such claims are generally advanced in the confidence that the virtue requirements of such work preclude automation. But from law to healthcare to education, it is becoming increasingly evident that many tasks that previously were thought to require specifically human capacities for thought and discernment can actually be done either by a combination of computers and lower paid technical labor, or solely by machines.
This raises a serious problem for the functioning of market-based capitalism. As Marx argued in his discussion of the organic composition of capital in Volume 3 of Das Kapital, the tendency of the capitalist mode of production generally is to shift and ever larger proportion of total capital from variable (human) capital to fixed (technological) capital. On Marx’s account, this becomes increasingly problematic because, human labor being the only source of surplus value, the smaller the proportion of the total for which it accounts, the smaller will be the opportunity for profit and competitive advantage. So, for instance, if one is in the business of producing widgets and one discovers that the Widgetmatic 5000 will make ten times the number of widgets per day while requiring one tenth as many workers, pretty soon you (and everyone else in your line of production) will be using it.
But, unlike your human workers, the Widgetmatic 5000 cannot be compelled to work longer or faster. You can run it 24 hours per day, but the rate at which it runs is fixed by the specs. You might be able to achieve economies via sourcing raw materials or penetrating new markets, but ultimately all your competitors will do the same. You can switch to the Widgetmatic 6000 when that comes out, but your competitors will do likewise. More problematically, chances are that automation in your line of production will be paralleled by automation in other lines, thus reducing the proportion of capital in the hands of those with the highest marginal propensity to consume (i.e. workers and others in the lower reaches of the income distribution). Automation works great if it’s just you that’s doing it. The idea that it has salutary effects when everyone does is a fallacy of composition.
This is the basis for what Marx referred to as the Law of the Tendency of the Rate of Profit to Fall. It is important to note that this is a tendency rather than a law pure and simple. There can be countervailing factors. The story of modern capitalism is one of struggles against this tendency through the applications of different varieties of technology, the expansion into different lines of production, the reorganization of the productive process itself, and (at least in some cases) the institution of welfare state protections of basic income. As Ford notes in Rise of the Robots, it isn’t necessary for machines to completely replace humans for this to cause a major disruption in the way the market economy works. Even at current levels of automation, income inequality has reach levels not seen since the Gilded Age.
True, not everyone thinks that this is necessarily a problem in and of itself. In 2012, Paul Krugman and Joseph Stiglitz had an interesting back and forth about whether this income inequality was retarding economic growth. This is worth noting in the present context, because Krugman advanced the view that enough capital could be spread around the economy through production of big ticket luxury items that the danger of choking off spending on everything else could be avoided.
Ford acknowledges this position, but disputes whether a job market limited to high-end production could really allow the economy of modern mass societies to thrive. As problematic as it is on its own terms, this proposition becomes considerably more alarming when one realizes that robots could just as easily build (and design) the yachts, cars, and houses of those at the top of the income distribution, as well as writing and debugging the software to improve production in future iterations. It is this last prospect that is the germ of a much bigger problem.
Photographs courtesy of Joel Schalit.