I recently stumbled across an old Tyler Cowen post at Marginal Revolution — on “zero marginal product workers” — which perfectly illustrates the circularity of marginal productivity analysis, and how it hides power relations behind a facade of neutral economic laws. As Cowen initially states the argument,

we have had a recovery in output, but not in employment.  That means a smaller number of laborers are working, but we are producing as much as before….  If I ran a business, fired ten people, and output didn’t go down, might I start by asking whether those people produced anything useful?

But when we critically analyze the assumptions behind this argument, we find it to be as meaningless as other statements of marginal productivity theory.

First, let’s take a look at the metrics of worker productivity. The marginal productivity of any factor, including labor, is what its remuneration adds to the value of the finished good or service. But the payment of the different categories of labor within a firm — whether hourly workers engaged in the direct production process, or technical and managerial salaried employees — reflects their value to the people running that firm. And that value is determined by the interests — the pecuniary interests, and the individual and collective power interests — of those same people in control of the firm. So in fact the “productivity” of various forms of labor within the enterprise is assigned not according to some objective or immaculate standard, but according to how they serve the power interests of those who dominate the institutional structure.

And in the real world, the people engaged in occupations judged most productive by the senior management of a firm are engaged in what the late David Graeber called “bullshit jobs.” 

By way of analogy, let’s look at the American slave economy in the 19th century. A plantation overseer is the very model of someone in a bullshit job, because it’s a position that would be completely unnecessary and useless in any rational system. An enslaved person works less hard and produces less, either with or without an overseer, than she would if she were free and in control of her own work on her own land. But given the fact of her enslavement, she undoubtedly works harder with the overseer threatening her than she would without. To that extent, the overseer clearly has “marginal productivity” within the context of the slave system, because he adds more to the plantation’s revenue than the cost of his wage.

To continue the analogy, a major part of the white collar workforce in a corporation is engaged in the kind of bean-counting, gatekeeping, and supervisory activity which is necessary only for coping with the irrationalities and conflicts of interest that result from absentee ownership and hierarchy. In an enterprise owned and managed by workers for themselves, these activities would be entirely superfluous. But given all the irrationalities and conflicts of interest in a hierarchical capitalist firm, the people engaged in these bullshit jobs do produce some value for the people at the top of the pyramid.

Second, laid-off workers are judged to have zero marginal productivity because the revenue and profits of the firm have not declined despite their absence. But this assumes that revenue and profit are determined by factors in the outside market which are beyond the firm’s control. And it ignores the possibility that revenue and profit do not decline, despite reduced intrinsic quality or quantity of the product, because of the firm’s pricing power in an oligopoly market — the result of entry barriers, artificial scarcities, and artificial property rights like patents and trademarks.

The economic history of the past generation or two has been one of steady downsizing of labor, coupled with near-constant increases in corporate profit. But despite the ostensible lack of decline in “productivity” — as measured in corporate revenues — the universal perception from the inside, by those engaged in the production process, is of hollowing-out, asset-stripping, and loss of human capital. And the perception from the outside is of what Cory Doctorow calls “enshittification.” As Tom West observed in the comments:

My personal experience is that when workers are let go, often the observed productivity doesn’t drop significantly. However, many things that are not measured drop significantly over time.

The incidence of food-poisoning in the restaurant creeps up, the failure rate of the manufactured item slows rises, the washrooms in the office become more and more unpleasant until ‘good’ workers start leaving, etc.

All of these don’t immediately show up on the balance sheet, making the fired workers “zero marginal product”.

Or to put it another way, if you fired all the professors and replaced them with lecturers, I suspect that with respect to student graduation rates, the professors would have, by the numbers alone, “zero marginal product”.

However, the numbers alone don’t tell us the whole story.

I don’t doubt that, in many cases, technological advances have reduced the number of labor hours necessary to produce outputs of a given quantity and quality. But remember: in a competitive market, without artificial property rights of any kind, a reduction in the labor hours required in the manufacture of a product would result in a corresponding reduction in price of the final good — i.e. increased productivity in material terms would naturally result in reduced “productivity” in revenue terms. Abundance, absent monopoly power, is deflationary. So the very fact that corporate revenues hold steady, despite the need for less labor to produce a given output in material terms, is in itself cause for suspicion.

I doubt anyone who remembers what it was like talking to a live person in the days before automated telephone menus, or dealing with career retail workers before they were replaced by minimum-wage workers, considers the present state of affairs an improvement in terms of quality of service. But these changes were made as cost-cutting measures — and I’m sure all the cost savings went into increased profits or management compensation, and not to reduced prices.

So when we hear the assertion that some workers are unproductive, we should ask — unproductive for whom?