On Jason E. Smith’s Smart Machines and Service Work

Jason E. Smith’s Smart Machines and Service Work argues that the rise of the service economy is the result of 1) the increasing productivity of manufacturing, and 2) the difficult, if not impossible, task of increasing productivity in services through automation.

These claims are, I think, correct, if also not earth-shattering. Of more interest in the specifics is Smith’s attempt to think through the relation of service work and the category of ”unproductive labor” from a Marxist perspective.  Smith rightly notes the incoherence of service as a category of economic analysis, yet retains it broadly by adumbrating  key categories of unproductive labor from Marx’s discussion in Capital volume three with all services. 

Part of the difficulty here is Smith’s reliance on definitions that Marx takes from Adam Smith, beginning with Smith’s definition of unproductive labor. For Jason E. Smith, the description of unproductive labor as labor that disappears in the moment of its production is not just what Smith means—and it is Smith’s idea—but also what Marx means. Yet at the same time, it is also fairly clear in Smart Machines that this definition cannot be what Marx means since Smith accepts the possibility that service work can be exploited. No surprise that Marx’s discussion of how teachers and other service workers can produce surplus value remains unquoted. 

And that shows, I think, my issue with the argument: its lack of engagement with value theory. Unproductive workers do not produce value for both Smith and Marx. The problem is that value doesn’t mean the same thing in either context. For Jason E. Smith, value seems easily quantified. It is increased labor productivity by sector. Rising productivity means an increase of value. 

Now, there’s a problem here in terms of Marxist theory: value is not the same thing as wealth. It is entirely possible in capitalist to have a situation in which more material wealth is produced but less value. This is a question of value composition, and it’s not dealt with at all by Smith. Falling capital investments and labor productivity rates give us unmediated access to value production. How? I don’t know. But it also explains Smith’s decision to treat all value production as a result of the interaction between capital and labor. Value composition is left out of the equation entirely. Now, why would that matter? Because labor isn’t the only source of value in Marx’s work. There’s also natural resources. There’s no sense in this discussion that the shift toward services occurs in the late Capitalocene, a period of intensified resource extraction and destructive climate change. 

You could argue that’s beyond the scope of this small book. Okay. Let’s look more closely at the categories of unproductive labor from Marx which matter most to Smith’s argument, the labor of supervision and labor of circulation. Both are unproductive, Smith argues, because they do not add value. Value is only created in production. The labor of supervision, which Marx explicitly compares to slave-driving (509), is correctly described as the labor of sweating workers for increased labor productivity. This is the labor of capitalists, to a degree, but it can be delegated in large firms to become wage labor. Marx also makes this point, but his reason for introducing this distinction is the opposite of Smith’s. For Marx, the separation of the labor of supervision from the capitalist means that the capitalist can be removed from the production process. The goal is to consider how one might separate capitalist control from its innovations in production (511). Organizing the labor process is a necessary aspect of production. Does that mean it produces value? It would seem possible, though Marx hates the idea because the only person who made it during his time was Nassau Senior, who reads less like a political economist than an apologist for capitalism. But Marx does treat this as labor that does not add value in production because it is outside the direct production process—and we will need to think about what that means in terms of the commodity fetish in just a moment.

The other labor in question is the labor of circulation. Here too the labor does not add value in the production process. It simply gets commodities from point A to point B. For Marx, it is of most importance for its ability to speed or slow the turnover of capital. Capital cannot be turned over into new investments without realizing its products as money. Or rather, it can, but this depends upon finance and debt strategies which will collapse if products are not sold. For Smith, circulation labor is just about getting commodities from point A to point B. It’s not just that they don’t add value but that they seem actively parasitic. 

And that’s what raises my ire in the book. The framing treats services, broadly construed, as parasitic on manufacturing productivity. This is a defeatist approach. The point, I suppose, is to show that service work’s lower productivity is part of a contradiction inherent to this formation of capitalism. At a certain point of labor productivity in manufacturing, poorly paid “low skill” service work begins to absorb some of the labor dumped from manufacturing, where the majority of value is produced. There are problems with this argument, though. In Smith’s telling, services are the last stop before a worker joins the reserve army of the unemployed. This may be true, but there is still a lot of other ”unskilled“ work in the first world in terms of agriculture, construction, meat packing, landscaping, and so on. Services continue to expand in the US, for instance, but much of that is going to health care. Other economic forces are driving that expansion. 

More than that, though, Smith’s claim that all non-commodity producing labor is parasitic on manufacturing productivity downplays that, for Marx at least, services can and are exploited by employers. If you hire a cleaning service, the cleaners are paid by the company to provide you, the client, a service. The company extracts value from their employees, meaning they produce value. Now, Smith would say, that’s true, but it can’t be intensified through technology. It can be sweated by managers with threats and so on, but it has an upper limit that can’t be increased. Fair enough. Smith’s concern, I suppose, is for capitalist accumulation. That’s stagnating. One might think this could be a possible locus for worker organizing and power. But for Smith is far more dour about this because of the dispersed nature of services, the lack of effect strikes in services have on society as a whole, and the difficulty of organizing across sectors. 

The problem, I think, all comes back to value. Smith insists value is created in production, not realized in exchange. With that perspective, the argument falls back on an old industrialist workerism: This leads to a variety of misapprehensions about the logical organization of capitalism. 

One of the worst, I think, is the blank reproduction of a claim from volume one of Capital that “labor-power in Marx’s formulation refers to the cost of the goods and services necessary to reproduce a given capacity to labor, day-in and day-out and over the course of generations” (123). At first glance, this seems fine and good. But it’s not so simple. First of all, this claim does appear in Marx but Marx has taken it, again, from Adam Smith. His use of it in Capital 1 appears early on during his discussion and thus operates at an extremely high level of abstraction. This is what we would think, in the abstract, that labor power is. But it is not, in practice, how capital functions. 

Smith’s book reflects one of the oddities of Marxism: a general sense that capitalism always works, even or especially when it doesn’t work for workers. There’s a sense that everything has been accounted for. Marx isn’t guilty of this but his use of means of subsistence to calculate surplus value can take him dangerously close, and for Smith, it will. Smith insists that the costs of reproducing “skilled” (aka highly educated or credentialed) laborers is the reason for differences in wages and salaries in different sectors. More on that in a moment. 

Marx here is working from Adam Smith’s claim that “a man must always live by his work, and his wages must at least be sufficient to maintain him” (book 1, chapter 8). For Smith, this sense of subsistence is bound up with a different mode of production—recall, Smith is writing not in the midst of developed capitalism but offering a dream of its possibility—and seems a version of “natural price” theory, which he uses elsewhere too (and in ways Marx will critique in subsequent volumes, if I recall correctly). We know that commodities don’t have natural prices. We also know, through historical work, that commodities in the eighteenth century didn’t have natural prices so much as moral prices, prices which were set by feudal traditions and enforced by social action, as EP Thompson explains in his essay on the moral economy. Labor, to the extent it has a price during Smith’s time, would have a moral price and be buoyed by other forms of production for its physical subsistence. Using it to analyze an economy premised on commodity exchange poses substantial difficulties.

Let’s bracket this historical problem, though, and approach the means of subsistence as Marx does, as a concept that enters into the production of value. As a concept, means of subsistence  binds the worker as a locus of physical and natural force to a larger natural metabolic process. In that sense, accounting for means of subsistence is necessary to imagine any economic organization.

The gist of Marx’s argument about means of subsistence, however, is that surplus value is value produced above value equal to the market costs of a worker’s socially determined means of subsistence. Marx’s presentation focuses on daily wages because employers purchase a day of labor-abstract power from workers and extract as much concrete labor—use-value producing labor—out of them in that period of time. What Marx wants to show is that surplus is not created through circulation, e.g., unequal exchanges between two merchants. If everyone buys low and sells high, no surpluses of production are possible. This is why Marx insists throughout the first part of Capital that a day of labor-power exchanges for its full value in terms of market costs for means of subsistence. He’s trying to track the source of surplus. Means of subsistence allow us to track what capitalism allocates for labor-power in production and how it can produce surplus value.

This is sometimes treated as a question of concrete labor time, as though people work for seven hours to reproduce their wages and then work another hour for their employers for free (another claim from Senior!). Marx’s account of the working day shows how intensely capitalists will try to extract more labor, either by extending the working day (though this is limited by the hours in a day and the worker’s ability to recuperate), or by intensifying work through co-operation and technological innovation. The problem is that capitalists can’t calculate value because they can’t be sure that their commodities will sell. They might approximate it by calculating their rates of profit, but even here capitalists anticipatewhat they may earn if their commodities sell. Now, they certainly try to extract as much work as they can from their workers. Of that, there is no doubt. But not in a way that they can track in terms of concrete labor time. They purchase a day of abstract labor-power and pay its market costs in means of reproduction. From the perspective of capital, this is a serious of commodity exchanges. 

All of which is to say, the purchase of labor-power and its payment in market costs of means of subsistence have nothing to do with the actual reproduction of labor. 

Why does this matter? We have to understand that, fundamentally, capitalism doesn’t care about labor’s reproduction. There is no rational process of labor reproduction. It is always strange to find claims about market rationality for labor reproduction built into the analysis of capitalism. Its defining feature, its production of commodities for exchange to accumulate value, is indifferent to workers’ needs. That’s an effect of commodity production, most especially the decentralization of market exchange that makes people into labor-power to be purchased. What capitalists purchase as a day or an hour or a piece is all that matters. That’s where ideas about “means of subsistence” actually matter. Marx’s two categories of capital account for all of this in the simple calculation of surplus value as the value produced in excess of the values consumed by production in terms of constant capital (e.g., means of production, raw materials, and so on) and variable capital (i.e., labor). 

Why does this matter? 

Means of subsistence suggests that capitalism cares about the ability of workers to reproduce their labor, that they must pay enough for workers to live. When one makes this central to the production of surplus value, it seems that capitalism understands what is in its interest or that it shares interests with workers. It doesn’t. 

What the reproduction of labor should direct us toward is the idea of appropriation. This is a concept utterly alien to Jason E. Smith, though central to work on social reproduction and care. Capitalism doesn’t simply exploit labor in exchange for wages. It also takes labor without paying for it. It’s not simply that women’s social reproduction work doesn’t cost capitalists much, so it’s easy to exploit and ignore (Smith 125). It’s that capitalism has been built on the backs of women’s reproductive labor and refused to pay for it. It’s that nature provides resources for production that also produce value for capitalism and that capitalism does not pay for those resources. And that’s why Smith’s lack of attention to the work of materialist feminists, social reproduction theorists, and to the environment matters. These are linked concerns, as Jason W. Moore’s theory of appropriation shows. For Moore, appropriation is as much a part of capitalist production as exploitation. The appropriation of cheap food, energy, labor, and raw materials is what drives capitalist accumulation. The means of subsistence problem is solved once we understand that capitalism relies on the appropriation of cheap food to lower its labor costs. When it can’t appropriate cheap food, it appropriates cheap labor through enslavement and unpaid social reproduction work. 

Now, this idea is fully present in Marx, not just in his discussion of original accumulation. It is  clearest in his discussion of workers and West Indian slavery in Capital 1. Capitalist production will use up its workers entirely, Marx tells us, as long as they have a continued steam of replacements for those killed in production. And Marx writes elsewhere, “Capital therefore takes no account of the health and the length of life of the worker, unless society forces it do so” (Capital 1, 381).