A Decade of Bullshit Jobs

In a recent Jacobin essay, the sociologist Matteo Tiratelli took issue with David Graeber's celebrated idea of bullshit jobs. His purpose was not to say that they didn't exist, though he convincingly shows that the definition of the term is misleading, but that the heart of the matter is not a subjective sense of meaninglessness but a disconnect between work and the common good: "David Graeber’s Bullshit Jobs asked a vital question about contemporary capitalism. But its theoretical starting point led it down a blind alley. The central problem with work today is not that a growing proportion of people feel their own job to be pointless. It is that the economy as a whole is not oriented around satisfying true human needs. Instead, we find our working lives consumed by dysfunctional bureaucracies that constrain our capacity for play and creativity. We waste time trapped in processes that exist only to give our bosses more control over us. And we’re directed to jobs that wouldn’t exist if people had the power to choose where to invest our increasingly scarce resources. That is where the real bullshit lies."

Where Graeber and Tiratelli overlap is in the assumption that bullshit jobs play a disciplinary role, both by occupying us and so preventing dissent at the level of society and by keeping us under the thumb of  bureaucratic management within the confines of the firm. But for me this misses out a crucial question. If we accept that many jobs are pointless, in the sense that their disappearance would not be a negative and might even be a net positive for society, why do they exist? Put another way, why do capitalists spend money on tasks that don't need doing and which they appear not to value? In his How to be an Anti-capitalist in the Twentieth Century, Erik Olin Wright notes one of capitalism's self-destructive tendencies: "Each employer wants to pay employees as little as possible in order to maximize profits, but this then depresses the buying power of consumers in the market, which in turn makes it harder to sell the things capitalists produce". It doesn't make much sense to heighten the exploitation of labour and then fritter away the gains on unnecessary roles that produce no surplus value.

A key sociological point to make about bullshit jobs, at least in the examples favoured by Graeber, is that they are predominantly "white collar", as we used to say. We can all think of pointless jobs that are coded working class, such as security guards, but the dominant image that comes to mind is someone in an office laboriously updating a spreadsheet that no one else will look at. In other words, regardless of the physical and cognitive demands of the work, these jobs are typically occupied by people who would be considered middle class, whether under-employed professionals or harried clerks. Paying supernumerary workers to ensure a large enough middle class to both preserve political dominance and maintain a buying public for goods and services is rational in aggregate, but self-interest would rule it out at the level of the specific firm. Capitalists are themselves subject to capitalist ideology, hence they are capable of irrational actions, even actions contrary to their own immediate interests, if they serve to preserve the capitalist order. But even so, there is a clearly a gap in the theory when it comes to coordination. And as ever when we are faced with a collective action problem, it makes sense to look at the role of government.

Tiratelli sees an avenue of coordination in the welfare state. "The growth of education and health care as an employer of last resort is intimately bound up with the decline of manufacturing ... the social power that industrial workers once had allowed them to win substantial concessions from the state and from capital, concessions that often took the form of expanded welfare provision. While those manufacturing jobs have been automated and offshored, the welfare states they created have lingered on, often under attack but generally accounting for a growing share of GDP and of working-class employment." There is something to be said for this, not least that the conservative political desire to "shrink the state" never actually manages to reduce it in any meaningful way, even as it shifts resources from the needy to favoured groups (consider the cross-party consensus to preserve the two-child benefit cap versus the acceptance that public money should be spent on the private school fees of diplomats' children).

But Tiratelli's theory is also unsatisfactory because of his focus on the working class. What we actually see in the welfare state is increasing labour exploitation and proletarianisation. As he notes himself, "They are increasingly underfunded when compared with the demands placed on them. They are poorly paid, insecure, and casualized. Professional autonomy has been decimated. And the prestige attached to these vital roles has been systematically eroded". This description shares many of the characteristics of the "precariat", but that concept obviously extends beyond the working class to the middle class. Indeed, the emergence of the idea reflected the extension of traditional practices from working class jobs to middle class ones, such as casualisation and piece-work. What were once accepted as necessary features of the market became a social "problem" only after this shift. But precarity extends well beyond the public sector to encompass the private, and arguably is more rife in the latter because the former has retained greater labour rights. The problem of coordination in the private sector remains.

When I first addressed this issue a decade ago, in response to Graeber's original essay, I suggested that the solution to the conundrum lay in the work of Ronald Coase, specifically his theory of transaction costs for firms. In a nutshell, firms will keep certain skills inhouse to meet intermittent need because this is ultimately cheaper than outsourcing, however this leads to the creation of unnecessary administration tasks to fill the downtime. As the firm grows, this work is gradually consolidated into distinct roles: bullshit jobs. However, there remains the countervailing tendency for technology to reduce transaction costs, so making intermittent outsourcing more attractive (this is the essence of the "gig economy"). Combined with the suspicion that recovery from recession would see a higher turnover in firms and an end to what appeared to be job-hoarding, my conclusion was that "In the face of this pincer movement, the decline in bullshit jobs seems likely and thus Graeber's essay takes on the air of a lament for a soon-to-disappear world, not unlike the typing-pools and reprographic offices of yore."

That prediction (though only semi-serious) has proved wrong. Recovery has been weak and the long-tail of low productivity firms stubbornly persistent: the great shakeout hasn't happened. The statistical decline in bullshit jobs that Tiratelli finds depends on his assumption that such jobs are predominantly working class, but that in turn rests on a subjectivity about how society values work (e.g. cleaners, labourers, sales staff) and the consequent lack of self-worth this gives rise to. The subject is bedevilled by the methodological issue that worth is reported by the worker, not established by any external, objective means. Thus the bullshit jobs that infest the corporate world can be defended by their occupants simply adopting a belief that they "add value" to a business. The key insight implicit in Graeber's essay is that, however they are defined, there was a noticeable increase in bullshit jobs during the 1990s and 2000s, yet this occured at precisely the time that many firms were outsourcing peripheral functions and even offshoring core operations.

If this is not simple coincidence, it suggests that bullshit jobs are a characteristic of globalisation, but this would lead us to expect them to decline as globalisation has itself cooled after 2008 and even gone into reverse with onshoring and inhousing. This does not appear to have happened, at least not in the corporate sector. The pandemic and ensuing lockdown led many people to question the value of their work. This was both in the sense of putting their careers in perspective as people around them died and in discovering that they could accomplish a week's tasks in a couple of days if they worked from home and thus avoided the "frictions" inherent in an office job, though the greatest friction, poorly-managed meetings, got worse for many with the shift online. That some companies have now insisted that workers return to the office, ironically including Zoom, suggests that the disciplinary role posited by both Graeber and Tiratelli is real, though that clearly covers both useful and useless roles. What we still don't have is any greater clarity on why firms continue to indulge the latter. 

The simplest explanation is probably that firms do not know how to properly evaluate cognitive roles (consider the story of Elon Musk assessing programmers by the quantity of code they produced). They know there is waste and inefficiency, but it cannot be isolated as easily as on a production line. As no one firm has an advantage over any other in this respect, the overhead can be accepted simply as the cost of doing business, much as the inefficiency inherent in any bureacratic system has long been tolerated. The great promise of information technology was that it would improve bureaucracy by eliminating the human factor. The great promise of the Internet was that it would transfer the burden of administration onto customers. Both have only been partial successes. The next great hope is that AI will somehow do away with the need for many cognitive roles, though it should be obvious that, as with previous technologies, this is likely to simply generate novel bullshit jobs in its train. It appears that bullshit is simply a feature of any cognitive work, and as that type of work expands as a share of the labour force under the pressure of continuing automation, we can expect such jobs to proliferate.