Why Has Capitalism Run Out of Steam?

By Dominique RouthierAugust 21, 2021

Why Has Capitalism Run Out of Steam?

Smart Machines and Service Work: Automation in an Age of Stagnation by Jason E. Smith

ENCOURAGED BY THE election of Joe Biden, the COVID-19 vaccine, and the so-called “major opportunity” of smart technology, commentators and investors now predict that the economy will rebound from the pandemic downturn, or even accelerate, once the “exogenous” shock caused by the coronavirus has been absorbed. But is this a plausible future scenario?

In his recent book, Smart Machines and Service Work: Automation in an Age of Stagnation, the Los Angeles–based Marxist critic Jason E. Smith makes the case that the dominant narrative of a tech-driven recovery is fundamentally out of tune with economic realities. When the economic data is properly disentangled from faulty definitions of productivity and untenable assumptions about perpetual economic growth, it becomes clear that we are not on the verge of an age of prodigious wealth creation spurred by smart machines. Rather, as the book title suggests, we are living through an age of stagnation with no reversal in sight.

Smith’s claim is clearly at odds with the past decade’s speculations about a fully automated society to come. For some — like New York Times best-selling author Martin Ford — the imminent Rise of the Robots poses a massive threat to our jobs and livelihoods. Others — like the leftist authors of the wildly popular “accelerationist manifesto” and subsequent book Inventing the Future — see full automation as key to unlocking the desired post-capitalist future. Whether shunned or celebrated, there is a widespread assumption that a fully automated society is right around the corner. Smith thinks otherwise. A more likely future scenario, according to Smith, is that despite advances in automation technology, the “ever-expanding ambit of the personal service sector” will continue its tearing apart of the wage contract and the social fabric while “communities dissipate into warring, atomized dysfunction.”

Smith’s critical analysis of the cyclically recurring automation hype aligns with Aaron Benanav’s equally crucial takedown of the “automation discourse” in Automation and the Future of Work. It also rhymes with the critique of tech-adoration in Gavin Mueller’s Breaking Things at Work. Like these and other Marxist critics, Smith responds to a salient intellectual trend throughout the 2010s: the discourse on futurity as seen through the conceptual lens of automation. There is indeed some critical work to be done. I, for one, welcome the recent “critical turn” in automation studies. And I think that a decade from now, a far-sighted book like Smart Machines and Service Work will prove to be the most lasting contribution to this important debate.

Most crucially, Smith shares with Benanav the idea that our economic wounds are not so recent as we might like to think. On the contrary, we learn that the present moment of crisis is systemically embedded in more than four decades of economic downturn. But, according to Smith, this long downturn signals something else than a temporary setback for capital. To Smith, and I tend to agree here, the current crisis exposes a more severe problem. What I gather from Smith is that our present historical moment of political unrest, riots, and uprisings might very well reflect capitalism’s protracted end crisis. A crisis that no machine, however “smart,” can forestall.

Worth noting, as Smith does, is the curious fact that if the release of the first iPhone in June 2007 marked the threshold to our era of “smart machines,” instant connectivity, and unhindered communication, it also marked the threshold to something else: the financial crisis of 2007–2008. While some economies have since quickly rebounded from the crisis — the most severe since the Great Depression — others had barely done so when the global pandemic of 2020 arrived to make matters worse. Much worse. Italy is the key example of a country that never even came close to a recovery. With a forecasted nine percent contraction in GDP for 2021, it faces the worst recession since World War II.

By all appearances, the US economy fared comparably well in the era of smart machines. Apple’s revenues quadrupled in the post-2008 decade, making it the first trillion-dollar company in market capitalization terms. The total annual revenue of Apple Inc. now easily trumps the annual GDP not only of Italy, but also of countries like Brazil, Canada, and Russia. For some observers, Apple’s success story relays the swift recovery of the US economy or even of capitalism at large. But did Big Tech set the pace for economic recovery after the last crisis? And will the smart machine weather the present crisis to serve as a global model for future prosperity and growth? Most likely not, according to Smith.

In a subtle and jargon-free application of what Marxists might call a “value-critical” approach, Smith strikes a deadly blow to the popular misconception that AI-powered smart machines are revolutionizing the core productive capacity of the capitalist world economy. Despite the past decade’s proliferation of smart machines and electronic gadgets in everyday life, the fact remains that productivity has been steadily falling for decades and seems to continue asymptotically toward zero. The “ubiquity of these devices,” Smith notes, “provides an important context for understanding the concurrent and urgent debates around automation.” But “[h]owever thoroughgoing the effects these machines and networks have had on the experiences of shopping, cultural consumption, navigating cities, or financial speculation, they have had negligible effects on one key economic variable: labor productivity in the workplace.”

That new technology, contrary to expectations, fails to boost productivity is a problem that has accompanied the computer age since its beginning. In the decade from when Intel released the first commercially produced microchip in 1971 until IBM introduced the personal computer in 1981, there were rapid advances in processing power, storage capacity, and hardware. But as the Nobel Prize–winning economist Robert Solow famously quipped in 1987 (it’s a line that Smith quotes): “You can see the computer age everywhere but in the productivity statistics.”

The problem that productivity does not necessarily keep pace with technological progress has become known, following Solow, as the “productivity paradox.” It’s a paradox economists have been trying to solve ever since. Smith discusses and critiques several more or less plausible economic hypotheses about why the so-called “third technological revolution” or “automation 2.0” fails to boost productivity. The main obstacle for economists trying to tackle Solow’s paradox, Smith points out, is that they “take for granted the resumption of a prior pattern of mid-century automation that may no longer apply.” One of Smith’s insights in the book is that automation then is necessarily both identical with and different from automation now: “If the first wave of automation took place in a postwar boom, current conversations about the coming disruptions to the labor market are occurring in the midst of severe and ongoing economic stagnation.”

Perhaps one reason that mainstream economists are struggling to understand why technological advances in computing and AI are not translating into productivity gains is that they assume that economic growth patterns are immutable natural laws that will repeat themselves in all eternity. But economic growth is contingent upon real-world circumstances — war, international trade relations, global levels of scientific and technological capacities, demographics, etc. — and hence subject to historical change. Today’s discourses on automation, Smith argues, rely on the unexamined assumption that current economic conditions are comparable to previous ones, and that technology, by implication, necessarily boosts labor productivity and stimulates growth. But this assumption rests on an extrapolation from two rather exceptional historical moments in the history of capitalism, the so-called “first” and “second” industrial revolutions.

In the wake of the “first” industrial revolution, in the late 19th and early 20th century, capital constantly introduced new forms of industrial extraction, processing, and distribution to supply what appeared to be an insatiable and constantly expanding world market. The invention of the steam engine and the introduction of fossil fuels, in combination with an abundant supply of material resources from the colonies and the hyper-exploitation of a newly “freed” and atomized labor force, boosted labor productivity and spurred a period of extraordinary economic growth that lasted until the Great Depression.

The “second” industrial revolution, in turn, took place after World War II and was characterized by yet another upsurge in labor productivity, due in part to new methods of “automatic” production and across the board application of time management techniques developed in the interwar years. A unique combination of circumstances fundamentally transformed both production and consumption, introducing an astonishing range of new commodities into the circuit of everyday life: TVs, cars, and refrigerators made up the peculiar decor of the Golden Age of Capitalism, which lasted roughly until the early 1970s. From that point onward, as the heterodox political economist Robert Brenner has convincingly shown, the global economy began a steady slope downward, putting an end to almost three decades of exceptional economic growth and ushering in a period of prolonged economic stagnation.

This is where Smith’s analysis picks up. The key to understanding our current conundrum, Smith argues, is the historical differences between industrial and service-driven economies. The ratio of investments in living labor relative to assets in fixed forms of capital (primarily machinery) has been decreasing throughout capitalist history. The nature of work has changed accordingly, shifting from manufacturing to services. The labor productivity gains of previous “technological revolutions” took place based on an industrial economy that extracted, processed, and manufactured goods for an expanding world market. In today’s gig economy of global platform services, the economic holding pattern of the industrial era has ceased to apply.


the share of workers now classified as working outside the manufacturing core of the U.S. economy […] has increased almost 50 percent. In absolute terms, the non-manufacturing sector tripled in size from 1953 to 2010, while the number of workers in manufacturing has actually declined from 20 to 15 million over the same span of time.

The transition from agriculture to industry ends in the service sector. Confronted with the proportional rise in service sector occupations relative to more traditional jobs in the manufacturing industry, it becomes increasingly difficult to measure and compare “labor productivity,” which has traditionally served as a yardstick for economic growth expectations. One of Smith’s major contributions to understanding this structural shift and the restructuring of the labor-capital relation lies in the fact that he disaggregates and reinterprets data to show that the term “services,” as traditionally applied, “amalgamates a number of activities that perform very different roles in the economy as a whole.”

Instead of lumping together under the term “services” any task performed outside the traditional manufacturing circuits, Smith suggests that we consider instead the economic function of a given task or occupation to determine whether or not it belongs to the service sector proper. One of the most convincing parts of Smith’s analysis, I find, is that in contrast to manufacturing, most services are best defined as acts of “unproductive labor,” in the sense that they produce little to no economic value. Pursuing this definition allows for a comparison between very different activities across the public and private sector: “In many ways, public services provided by the governments are similar to financial services and many retail activities in the private sector insofar as they do not directly produce value.” Essentially, the age of stagnation is premised on a zero-sum game of exchanging “worthless” services.

Interestingly, the economically “unproductive labor” at the base of contemporary service work also appears as a historically constitutive factor of the systematic underpayment and denigration of women since their en masse entrance to the workplace in the 1970s at the beginning of the transition into the service economy. The structural evolution in capitalism parallels widely held gendered assumptions about the inferior nature of so-called “low-skilled” service occupations. “[I]t is no accident,” Smith remarks, “that occupations classified as requiring few or no skills have traditionally been understood as ‘women’s work.’” The history of automation is also, then, a history of systemic discrimination. Today, most low-paid service sector jobs are performed by what Smith describes as a growing “servant class” consisting of people that, like the Black militant auto-worker James Boggs who is a central figure in Smith’s overall narrative, have been made structurally superfluous to the production of value. In this way, Smith’s critical analysis also touches — though one could have wished he would have expanded on this particular aspect — on deep-seated misogynist and racist cultures structural to the age of stagnation.

The term “stagnation,” or “secular stagnation” to be precise, is a favorite of American economist Larry Summers, a former economic advisor for the Clinton and the Obama administrations. Like many other economists, Summers sees Keynesian stimulants as a way to circumvent secular stagnation, reboot productivity, and kindle economic growth. But as we have already seen, to Smith the term “stagnation” signals more dire straits. Capitalism’s growth-engine — the performance of which is measured not in horsepower but in terms of labor productivity — has apparently run out of steam. If many services are low-productivity occupations or outright “unproductive labor,” as Smith powerfully suggests, it is highly unlikely that Big Tech — which consists mostly of “extremely refined advertisement delivery systems” and peer-to-peer service platforms — will kick-start the economy and set us on the path to full automation.

What I see as the distinguishing feature of Smith’s critique of “automation discourse” is the fact that he insists that the current crisis is intrinsic to the capitalist mode of production. The transition into a largely “unproductive” service sector was, so to speak, a necessary consequence of an economic system based on the value form of the commodity, private property, exchange, and fierce global competition. While the “exogenous” trigger for the present crisis was the outbreak of the coronavirus, the root of the problem reaches further into the economic texture than most mainstream economists like Summers would like to think. And the iPhone, or any other “smart machine” for that matter, isn’t anything close to a new steam engine. From a political economic standpoint, most of these smart machines are simply gadgets: “toys, not tools,” as Smith succinctly puts it.

In making the argument that capital accumulation has now come up against its limits, Smith is adding a chapter to a venerated history of Marxist crisis theory that includes figures like Rosa Luxemburg, Henryk Grossman, Robert Kurz, and Roswitha Scholz, whose particular brand of Marxist-feminist value theory still remains largely unknown in the Anglophone context. According to this tradition of Marxist thought, sometimes dismissively referred to as “collapse theory,” capitalism is fraught with internal contradictions and cannot maintain its trajectory of growth indefinitely.

Armed with insights from this tradition, Smith plows through the economic data sets of the past crisis decade and boldly revives Marxist crisis theory. Capitalism is an engine of technical innovation and economic growth, but it is also a self-defeating social system that tends toward crisis and eventual collapse. Barring a few gadget-producing tech companies, statistics indeed seem to confirm Marx’s controversial claim that there is a “tendency of the rate of profit to fall” at work in the history of capitalism. At least since the 1970s, global average profit margins have steadily dropped as productivity gains have flattened across sectors. This tendency highlights the need to tackle, head-on as Smith does, the question of why capitalism has run out of steam at the exact moment when it appears to be most innovative. Most likely, as Smith convincingly argues in Smart Machines and Service Work, Big Tech won’t save us this time around. Especially not if by “us” we include all those who do not belong to the affluent classes typically most invested, emotionally and economically, in the contemporary phantasm of a fully automated society.


Dominique Routhier is a writer and postdoctoral researcher based at the University of Southern Denmark. His work has appeared in numerous Scandinavian outlets and in international journals such as Rethinking Marxism and Nordic Journal of Aesthetics. Currently, he is finishing a book called With and Against: the Situationist International in the Age of Automation, 1956–1968.

LARB Contributor

Dominique Routhier is a writer and postdoctoral researcher based at the University of Southern Denmark. His work has appeared in numerous Scandinavian outlets and in international journals such as Rethinking Marxism and Nordic Journal of Aesthetics. Currently, he is finishing a book called With and Against: the Situationist International in the Age of Automation, 1956–1968.


LARB Staff Recommendations

Did you know LARB is a reader-supported nonprofit?

LARB publishes daily without a paywall as part of our mission to make rigorous, incisive, and engaging writing on every aspect of literature, culture, and the arts freely accessible to the public. Help us continue this work with your tax-deductible donation today!