Gig work in historical perspective

All digital platform businesses perform some kind of matching function: connecting participants who then engage in some form of exchange (directly or indirectly). Advances in the technology of networking and matching underpin the emergence of the far-reaching marketplaces developed by the most successful digital platforms. Once a particular platform attains a leading position in its market, strong economies of scale and scope in networking tend to reinforce its dominance.

Matching platforms can be divided into two broad categories: those which facilitate the exchange of assets and those which facilitate actual production (Reference Farrell and GreigFarrell and Grieg, 2016). Platforms which facilitate actual work and production have become common in several sectors, including transportation and delivery, odd jobs and miscellaneous tasks, and many forms of digital work (such as programming, writing, translating or design).^{Footnote 1} Productive labour performed through this class of digital platform typically incorporates the following five broad organisational features:

1. 1. Work is performed on an on-demand or as-needed basis. Producers only work when their services are immediately required, and there is no guarantee of ongoing engagement.

2. 2. Work is compensated on a piece-work basis. Producers are paid for each discrete task or unit of output, not for their time.

3. 3. Producers are required to supply their own capital equipment. This typically includes providing the place where work occurs (home, car, etc.), as well as any tools and equipment utilised directly in production. Because individual workers’ financial capacity is limited, the capital requirements of platform work (at least capital used directly by workers) are typically relatively small (although these assets can be significant in the lives of the workers who must purchase and maintain them).

4. 4. The entity organising the work is distinct from the end-user or final consumer of the output, implying a triangular relationship between the producer, the end-user and the intermediary.

5. 5. Some form of digital intermediation is utilised to commission the work, supervise it, deliver it to the final customer, and facilitate payment.

Only the last of these common features of platform work has any obvious connection to modern technology: obviously, the digital technologies which facilitate communication, management, supervision and payment have only recently been developed. Even this defining feature – digital intermediation – could apply to many different strategies of work organisation (including digitised rostering systems, web-based communication and monitoring systems), not just in digital platforms. And certainly none of the other items in the preceding list is new to 21st century digital capitalism. On the contrary, each has been utilised by employers across a wide range of industries, throughout the history of paid employment. On-demand work and piece-work compensation have been commonly applied in many industries, given their utility (in certain situations, not all) for ensuring that employers only pay for work they actually need and receive (Reference Grantham, Grantham and MacKinnonGrantham, 1994). Casual, seasonal and contract labour were the predominant forms of paid work as capitalism first emerged and consolidated (Reference Deakin, Clarke, De Gijsel and JanssenDeakin, 2000; Reference WoodWood, 2002). Reference QuinlanQuinlan (2012) shows that these practices were even described as ‘precarious work’ in 19th century policy discourse. Similarly, requiring producers to supply their own capital equipment is a long-standing feature of work in many industries, including transportation, resource harvesting, construction and personal services.

Consider, for example, the ‘putting-out’ system common in Europe in the early history of merchant capitalism.^{Footnote 2} It provides a good historical example of the long-standing application of several flexible and subcontracted work strategies that have again become common in modern digital businesses. In this system (also known as the ‘domestic’ or ‘cottage’ system), a merchant distributed production tasks to paid employees, supplying necessary raw materials and supplies. It was especially widespread in the manufacture of textile, clothing, footwear, cutlery, small furnishings and other simple consumer goods.^{Footnote 3} Producers performed work in their own homes, using simple capital equipment which they owned. But the output of their labour was owned by the merchant capitalist who supplied the initial materials; the producers were engaged merely to perform incremental value-adding labour on those materials. Their work was compensated on a piece basis, with payment occurring after the home worker returned the finished product back to the merchant. There was no promise of re-engagement to perform another batch of home production. The merchant took responsibility for selling the finished product to third-party consumers (and in some cases payment was deferred until that sale occurred). Indeed, it was the extending reach of these merchants, and their capacity to sell into increasingly far-flung markets (thanks to improved transportation, integration of markets and currencies, etc.), that facilitated the expansion of this flexible new form of work organisation. Except for the absence of digitised systems for coordinating, supervising and compensating work, this business model is quite comparable to those of modern digital platforms.

The triangular relationship between producer, end-consumer and intermediary typical of digital platform work (Reference Stewart and StanfordStewart and Stanford, 2017) also has many historical precedents. This triangulation obscures the relationship between the intermediary and the workers who perform the productive labour; in the modern context, it allows employers to inhabit a legal ‘grey zone’, where it is not clear whether producers are workers, contractors or self-employed (Reference Johnstone, McCrystal and NossarJohnstone et al., 2012). This ambiguity has so far allowed digital platforms to evade normal obligations imposed on traditional employers – although that immunity is being contested on many fronts. But this blurred intermediate position is hardly novel: it has been typical of many previous business models throughout the history of capitalism.

Indeed, businesses in many circumstances have long preferred to constitute their productive workers as ‘contractors’ or nominally independent producers, rather than strictly defined ‘employees’, for obvious economic and legal reasons: avoiding entitlements or benefits normally paid to employees, evading the impact of regulatory standards that apply to employment (such as minimum wages or limits on hours of work) and transferring risk for fluctuations in demand conditions to producers. In historical perspective, labour contracting and subcontracting practices were the predominant form of paid work in early capitalism until later in the 19th century (Reference Deakin, Clarke, De Gijsel and JanssenDeakin, 2000; Reference SteinfeldSteinfeld, 2001), even in heavy industries such as iron production (Reference ZmolekZmolek, 2013). Only with the advent of more regularised and centralised production technologies, along with social and legal reforms which required more reciprocity in the relationship between employers and workers,^{Footnote 4} did the practice of permanent waged employment extend its reach:

Labour hire and temporary staffing businesses, which also have a long (pre-digital) history, are another incarnation of employer efforts to subcontract labour and hence evade the risks and responsibilities associated with permanent paid employment. This model similarly exploits the intermediate and ambiguous legal space between producers and end-users. Waged work in many agricultural and industrial applications was commonly organised through nominally independent subcontractors or ‘gangmasters’ in the British and continental economies, through much of the 19th century (Reference BrassBrass, 2004).^{Footnote 5}

Strategies of outsourcing, contracting and subcontracting, therefore, have maintained a continuous presence throughout the history of capitalism. And this general practice remains important in many non-digital industries, including resource harvesting (like lumbering and fishing), hairdressing and other personal services, cleaning, maintenance and repair activities, and creative work like writing, arts and design (MBO Partners, 2016). In many of these schemes, producers are paid on ‘consignment’ from revenues generated when their output is finally sold by the intermediary – similar to the payment systems used by platform businesses. The triangular structure of subcontracting is also often associated with the requirement that workers provide their own tools and equipment (another feature of modern platform work).

Thus, apart from the specific nature of digital methods of communication, work allocation, supervision and payment, the work practices and relationships embodied in modern digital platform businesses do not seem ‘new’ at all. This historical perspective allows us to reconsider whether and how the business model and work organisation strategies utilised by modern digital platforms are indeed ‘innovative’. Consider the well-known case of the ride-hailing service Uber (and similar businesses like Lyft).^{Footnote 6} This business has successfully displaced traditional taxi work on the strength of an effective digital dispatch system – whereby clients can hail a ride (and pay for it) through an app on their smart phones, with useful features that include being able to track the location of their car online. Consistent with long-standing subcontracting strategies (Reference Johnstone, McCrystal and NossarJohnstone et al., 2012), Uber defines its drivers not as employees but rather as self-employed producers.^{Footnote 7} Yet, Uber sets the fare and route, collects payment from the customer through its online app (cash payments for Uber rides are not permitted in most jurisdictions), supervises and where necessary disciplines drivers and then pays drivers a portion of revenue based on pre-determined distance and time factors. The actual production process is no different from a traditional taxi: a worker collects a passenger and delivers them to a different destination. The online hailing app is more convenient, for many users, than manually hailing a taxi, or phoning a dispatch office. But it is certainly possible for traditional taxi services to utilise digital dispatch systems (including web-based and smart-phone systems) without adopting the same subcontracting labour strategies as Uber.^{Footnote 8}

What really distinguishes Uber from traditional taxi companies, therefore, is the organisation of work within its service, not the technology of production. Uber drivers provide their own vehicles, pay for all related expenses (including amortisation, fuel and maintenance) and are compensated by Uber on a per-fare basis, with no guarantee of hourly or daily income. Uber drivers incur the full costs of operating their vehicle (like taxi owner-operators), but also lose the fees deducted from their fare revenue by Uber (like waged taxi drivers). This model has allowed Uber to appropriate profits from provision of a taxi-like service, but without the capital outlays associated with either owning or operating vehicles, or purchasing licences/medallions.^{Footnote 9} Centralised control over its proprietary dispatch application, which drivers need to find customers, is the basis for its claim to this revenue – just as the merchant’s centralised capacity to connect home-made goods with final purchasers was the basis for its claim to a margin of total revenues under the putting-out system.

The analogy between modern digital platforms and the intermediated or subcontracted production practices of earlier centuries extends beyond the organisation of work. There are other parallels between the rapid expansion of digital businesses like Uber, and the rise of other subcontracted models (like the putting-out system) centuries earlier. Recall that the putting-out model itself supplanted an earlier, once dominant production system – small-scale workshop-based manufacturing under the guild system – just as Uber has displaced traditional taxi businesses. The motives for organisational innovation, and the means by which that transition was accomplished, are surprisingly similar (see Table 1).

Table 1. Ascendant regimes of work organisation.

Source: Author as described in text.

As with Uber, the putting-out system entailed no fundamental change in the actual process of production: the tools and techniques used in home work were no different than those used in the workshops of the guilds. The putting-out model was developed by merchant capitalists largely to subvert restrictions on entry imposed by the guilds, which allocated particular regional markets to authorised suppliers. In a similar frontal attack on regulation, Uber’s strategy has been premised on an effort to sidestep municipal rules limiting entry of taxi services;^{Footnote 10} the company has also seized the opportunity offered by its platform to avoid taxes and rules governing minimum wages or hours of work. In both the putting-out system and modern ride-hailing, workers provide the capital used directly in production. Of course, the intermediary business must also make a capital investment, but it is small relative to the total capital used in production, most of which is supplied by the producers. In the putting-out system, the merchant’s investment consisted of purchases of intermediate materials and developing the marketing infrastructure to ship finished goods to relatively far-off markets; in Uber’s case, it consists of software and computer capacity to run the dispatch and payment system, as well as marketing to promote brand awareness among consumers.

One novel feature of the business model of modern digital platforms is worthy of final note. In both putting-out and modern digital platforms, the intermediary business extracts a surplus over time from the ongoing production and sale of the product or service. But in the modern context, entrepreneurs are able to capitalise that surplus up front in the form of large stock market valuations – and then monetise those gains through public offerings, options, and other financialised strategies. In this regard, the practices and incentives of financialisation have both motivated and facilitated the rapid innovation and expansion of these businesses. There is reason to doubt the long-term sustainability of these sky-high platform valuations; equity markets often overshoot in their judgments of the profitability of novel businesses, carried along by manias and bubbles regarding new investing fads. But in the meantime, these financialised strategies certainly facilitate the accumulation of vast stockpiles of apparent wealth in the hands of businesses that, at their foundation, engage in fairly mundane, low-tech production (like rides, deliveries and odd jobs).

The rise and fall of the SER

Other than the use of digital technology to facilitate work allocation, coordination and compensation, there is thus little ‘new’ about the labour practices of digital platform businesses. The main aspects of their work organisation – on-demand work, piece-work compensation, home work and a triangulated relationship between producer, intermediary and end-user – are visible in long-standing practices of private businesses stretching back through the history of capitalism. The central labour relationships utilised by digital platforms can thus be better understood as a return to previous practices. This conclusion, however, begs another question. Why did those practices become less common for a period of time in the 20th century, despite their long historical pedigree? And why did employment, for a while, come to be associated with a different, more stable set of relationships and practices?

In contrast to the precarious, tumultuous world of work in earlier centuries, employment during much of the 20th century offered a more regularised, predictable arrangement – one more favourable for the economic security of those performing the work. This more stable system has come to be known as the ‘standard employment relationship’ – although it was only ‘standard’ for a circumscribed portion of the history of capitalism and can now be understood as an historical exception. Its origins are visible in the growth of centralised production in larger factories beginning in the 19th century, but it did not become a dominant template for work organisation until well into the 20th century. In the present era, the features of the SER are clearly receding – but that reversal is experienced more broadly than in just the limited world of digital platforms.

Reference Cranford, Vosko and ZukewichCranford et al. (2003) and Reference BoschBosch (2004) usefully catalogue the major defining features of the SER. Workers would typically work for just one employer, year-round, usually on a full-time basis, on the employer’s premises and utilising capital equipment supplied by the employer. The term of employment was indefinite: workers were rarely guaranteed ‘jobs for life’, but the mutual expectation was that employment would continue unless some intervening force (such as a downturn in the employer’s business or egregious misperformance by the worker) caused the relationship to be terminated.

Over time, labour market institutions evolved to reflect and reinforce the SER as the normative benchmark of employment practice. Labour laws defined rights and responsibilities associated with employment, on the assumption that a ‘job’ entailed certain reciprocal expectations of fairness and stability. Trade union and collective bargaining laws and practices were based on similar assumptions of a stable, cohesive workforce, congregated at a central location (the enterprise). Social programmes and employment benefits also came to be based on the assumption that paid work would take a form compatible with the SER through mechanisms like social security contributions collected from payrolls, and entitlements, for example, to pensions or unemployment insurance, contingent on periods of stable employment. In short, the SER reflected an understanding that extended well beyond the specific features of a given job, to incorporate a broader conception of social security:

Employers also experienced some benefits from the SER, including stability in work organisation, the ability to undertake more intensive production planning and institutional bounds placed around collective action by workers (Reference BoschBosch, 2004).

Even at its peak, however, and despite being buttressed by these complementary legal and social institutions, the SER was never universal. It is important, as Reference MillarMillar (2017) notes, that analysts of modern precarity do not uncritically elevate former practices as a universal, normative goal. Women, immigrant or racialised workers, and workers in numerous sectors of the economy (especially decentralised, highly competitive service sectors such as cleaning or hospitality) were far less likely to attain stability and permanency in their work (Reference Vosko, MacDonald, Campbell, Vosko, MacDonald and CampbellVosko et al., 2009). Indeed, the SER was always gendered, in the sense that the incomes earned by the core workforce in permanent, full-time positions were understood to constitute a ‘family wage’: sufficient to support the (assumed male) worker’s entire family, including his spouse (who, if engaged in paid employment at all, would not expect the same compensation or entitlements).

Economic and labour historians have catalogued factors which together help explain the rise of the SER as the dominant form of work organisation – beginning in the later decades of the 19th century and peaking during the postwar ‘Golden Age’ expansion in the industrialised countries.^{Footnote 11} Technology certainly counts among these key causal forces. The development of centralised technologies of mass production (especially in manufacturing), accelerating with the application of Fordist assembly line techniques, disrupted previous employment models. Capital requirements were beyond the reach of decentralised producers. And the operation of large facilities, with their intense internal division of labour, required the reliable presence of a consistent, disciplined workforce. The attendance and performance requirements of capital-intensive enterprises made it too risky to allow workers choice or discretion in working hours. Similarly, the job-specific skill requirements of mass production technologies enhanced the benefits to employers of a stable workforce, thus encouraging them to offer permanent jobs.

In addition to technology, however, broader macroeconomic and political-economic forces motivated the expansion of the SER, especially after the Second World War (Reference KallebergKalleberg, 2009). With the engine of capitalist accumulation firing on all cylinders, and governments reinforcing expansion through Keynesian full-employment macroeconomic policies, postwar unemployment was low. This created a stronger incentive for employers to recruit and retain workers with promises of stable employment – rather than assuming that contingent or subcontracted labour would be available whenever and wherever needed (Reference Grantham, Grantham and MacKinnonGrantham, 1994). The SER was also reinforced by political imperatives. For various domestic and global political-economic reasons, employers and governments in most OECD (Organisation for Economic Cooperation and Development) countries felt compelled to offer a more attractive compromise or social contract with workers. Norms about what constituted fair treatment on the part of employers changed: workers came to expect stable employment and associated entitlements and benefits as normal features of work. The rise to predominance of the SER paralleled the corresponding rise of a broader, redistributive understanding among employers, the state and workers, reflecting the unique conjuncture of economic, political and geopolitical circumstances of the postwar era.

Appreciating the historical and political-economic specificity of the SER provides a context for understanding the more recent unwinding of many of its typical features and practices. This unwinding is visible in many parts of the economy – not just among digital platform businesses. Indeed, evidence suggests that as little as half of existing paid work in developed Anglo-Saxon economies (such as the US, Australia, and Canada) still occurs within the confines of the SER model (Reference Lewchuk, Lafleche and DysonLewchuk at al., 2013; Independent Inquiry into Insecure Work, 2012; Reference StanfordStanford, 2016); precarious or contingent work in a multiplicity of forms (including part-time, temporary, casual, labour hire, independent contracting and marginal forms of self-employment) accounts for the rest.^{Footnote 12} The factors which supported the rise of the SER have largely reversed direction, helping to account for the generalised resurgence of precarious work. This certainly includes the direction of technological change – and not just the development of web-based platforms (Reference Howard and KingHoward and King, 2008). The growing relative importance of services industries (many of which, not all, are characterised by smaller scale production), and the decentralisation of other kinds of production (reflected in the shrinking average size of enterprises), have contributed to less capital-intensive, smaller scale production in which employers may worry less about recruiting and retaining a stable workforce.

More broadly, communications technology facilitates the vertical disintegration of production processes, with consequent outsourcing of multiple sub-functions to dispersed, smaller firms (Reference WeilWeil, 2014). Even the technology of surveillance and performance monitoring has likely contributed to the growing precarity of work: as it becomes cheaper and more effective for employers to monitor employee performance and obedience through increasingly intrusive forms of technology, they face less compulsion to provide positive inducements to workers to elicit performance (like permanent work or superior wages).

However, the erosion of the SER and the expansion of more precarious labour practices cannot solely, or even mostly, be seen as a technology-driven story. Broader macroeconomic and political-economic forces were important to the ascendance of the SER in the postwar era; those forces, too, have since reversed course dramatically, facilitating the return to more precarious work practices. With the advent of neoliberal macroeconomic management, full employment was abandoned as a guiding goal, replaced by an explicit commitment to restoring discipline to labour markets through the recreation of ‘equilibrium’ cushions of unemployment. Labour market slackness has become a more-or-less permanent feature in industrialised countries (particularly since the global financial crisis), and this contributes to the expansion of precarious work in at least two ways. Employers are less worried about being able to hire labour when necessary, and hence one key motive for offering permanent SER-type positions disappears. Workers are compelled by perpetual unemployment, combined with retrenchment of income protections for unemployed people, to accept precarious work. Regulatory structures, especially in the Anglo-Saxon economies,^{Footnote 13} have permitted and facilitated the expansion of precarious practices such as irregular shifts, zero-hours contracts, the elimination of rules regarding lay-offs and severance requirements, and freedoms for labour hire companies. Those regulatory structures, in turn, reflect an employer-friendly evolution in the broader political and cultural world, which also influences employment norms and expectations. In contrast to the peak of the postwar Golden Age, when workers’ demands for improved protection and compensation were broadly ratified, modern political culture promotes the idea of a ‘risk society’ (Reference Gottfried, Wilkinson, Wood and DeegGottfried, 2014), in which having a job is seen as a privilege.^{Footnote 14} Other features and practices of the neoliberal economy – including the vertical disintegration of supply chains, the intensification of franchising, the ubiquitous outsourcing of business functions, and the development of complex and far-flung global supply chains – have also contributed to the breakdown of the SER and its replacement with more precarious and contingent employment relationships.

In sum, both the rise of the SER as a benchmark for postwar employment relationships and its subsequent erosion under neoliberalism reflect the bigger shifts in the broad political-economic balance of power within the industrialised economies. Whether at the level of individual firms, industries or the macroeconomy, employers are less constrained in their ability to organise work to minimise their risks and responsibilities and optimise their profits. In this broader context, the use of precarious work practices within digital businesses can be understood as just one dimension of a broader shift in capitalist employment relations.

Gig work and the logic of labour extraction

Armed with this longer run perspective on the rise and fall of the SER, we now propose a more integrated analysis of the factors contributing to the resurgence of precarious work practices – and also the constraints that may limit their expansion, even in a digital world. This section will integrate preceding historical insights regarding the evolution of work practices under capitalism into a more holistic theoretical model focusing on the underlying labour extraction problem which confronts every employer (Reference GintisGintis, 1976). Paid employment is premised on an individual performing work for another entity, in return for compensation. Inherent in this relationship is the necessity for the employer to direct and supervise the work activity of hired help, to ensure that it is appropriately effective and productive. This challenge arises from the fact that the employee is not working ‘for themselves’, but rather to produce value-added which is owned by their employer, and ultimately sold for the employer’s benefit (not directly the workers’). The transition from the autonomous work of independent producers to the supervised work of paid employees required the development of systems of management, supervision, incentive and discipline so that employers could attain optimal effort and productivity from their waged employees (Reference BurawoyBurawoy, 1979; Reference ThompsonThompson, 1968).

This labour extraction challenge is complicated by the central fact that what employers typically pay for (the time of their employees) is distinct from what they want (exerted labour effort). Labour intensity can be defined as the amount of attention, energy and exertion that workers devote to their tasks, up to some limit imposed by their mental and physical stamina. From the advent of wage labour in the early years of industrialisation to the present, employers have addressed the challenge of labour extraction through varying management, technological, regulatory and cultural practices that maximise labour intensity, subject to the broader constraints and circumstances within which their businesses operate. The evolution of management labour extraction strategies helps to explain the changes in employment relationships that were surveyed above; it also helps in understanding the strengths and limitations of ‘gig’ work arrangements.

Maximising unit profit depends, with unit price and materials costs held constant, on minimising wage payments, maximising work intensity (i.e. how effectively paid work is translated into exerted labour effort) and enhancing the technical efficiency of the production process. But wages, labour intensity and the ultimate efficiency of production are not mutually independent, so the employer’s challenge becomes a complex juggling act – trading off the various levers which influence exerted effort, realised productivity and ultimate bottom-line labour costs. There can be no assumption that paying the lowest possible wages will maximise profits, if poor compensation negatively impacts labour intensity; this relationship provides a logical basis for employers to pay wages higher than market-clearing ‘competitive’ rates, as emphasised in the literature on efficiency wages.^{Footnote 15}

Theorists of the labour extraction problem emphasise a broad range of factors shaping employer strategies for minimising unit labour costs. There is always an embedded trade-off between ‘carrots’ and ‘sticks’ in management’s decisions (Reference GordonGordon, 1996): employers will generally choose some combination of positive rewards for performance (through gain-sharing or bonus-based compensation schemes, above-average wages to elicit loyalty and effort, implicit or explicit guarantees of stable employment, etc.) and negative punishments for disobedience or unacceptable performance.^{Footnote 16} The effectiveness of the latter option depends on several further factors: including the attitudes and expectations of workers, the cost of supervision, the legal ability to discharge workers and the ultimate loss that discharged workers experience as a result of being fired. That ‘cost of job loss’, in turn, is a composite outcome of several other factors, including how quickly a discharged worker can find alternative work, the extent to which their earnings in a new job match former wages and any income support payments they receive in the interim. Ultimately, the power of employers to control production is derived from the scarcity of good jobs so that workers always have something to lose if their relationship with their employer collapses (Reference Bowles, Edwards and RooseveltBowles et al., 2005).

Technology obviously shapes the labour extraction effort in many ways. But technology always interacts with the social organisation of work – not just how it is applied, but even how it is developed in the first place, with employers directing innovation to tasks most compatible with their interest in profitable labour extraction. In a strictly controlled Fordist production setting, where work is fragmented into specific tasks, output may be somewhat less dependent on the labour effort exerted by engaged workers (since output is determined largely by the pace of the machinery); this might reduce the premium that employers feel compelled to pay to elicit effort (Reference BravermanBraverman, 1974). At the same time, complex, capital-intensive production systems require considerable labour discipline (measured in attendance, attention, etc.), so the labour extraction problem is not ‘solved’ by the mere presence of an assembly line. In very complex worksites, the absence of a single worker might disrupt the work of hundreds of others; this enhances workers’ capacity to demand premium pay for requisite discipline and reliability. Centralised production technology can thus empower workers’ bargaining position or undermine it.

The decreasing average size of enterprises in the modern economy could be facilitating a shift towards fragmentation and hence more casualisation of work (Reference Sawyer and ToporowskiSawyer, 2000). Technical changes which allow for greater decentralisation of production with relatively limited amounts of direct capital (including some kinds of digital work) may also be facilitating the return to a more precarious model of work organisation (Reference Howard and KingHoward and King, 2008), including home-work systems (which require workers to supply their own capital). Employers often ‘sell’ home-work arrangements on the basis of supposed convenience or flexibility for workers, but in practice they are an effective mechanism for shifting capital costs to workers, and also for extending the reach of paid work time into greater portions of a worker’s day. On the contrary, technological advancement is normally associated with the accumulation of capital and a rising capital–labour ratio, and this should constrain the applicability of home-work systems. Many jobs clearly require more capital than individual workers could be expected to provide. Moreover, production may require proprietary technical knowledge which employers are unlikely to disburse, and most production still requires workers to gather in a specific workplace at a specific time. So while some work in the modern economy can be performed by workers operating with small amounts of invested direct capital equipment (and these are the jobs where the digital platform business model would seem most promising), this trend is not universal.

The technology of monitoring and supervision also affects employers’ choices regarding the trade-off between carrots and sticks in labour extraction (Reference GordonGordon, 1996; Reference GreenGreen, 2006). If it is expensive to monitor workers (e.g. by having to hire human supervisors, who in turn need to be supervised themselves), and difficult to punish or fire those who do not meet performance benchmarks, then employers will be relatively more disposed to use positive incentives to elicit effort, loyalty and retention. On the other hand, if supervision is inexpensive (e.g. thanks to automated monitoring technology), and workers can easily be disciplined (e.g. casual or temporary workers can simply be non-renewed, with no costs of dismissal or severance), then the ‘stick’ looks relatively more attractive. One oft-overlooked aspect of the technology of digital platforms is especially relevant in this regard. Many platforms (including Uber and Airtasker) utilise an online customer ‘ratings’ system to develop performance profiles of their associated producers. Producers receive evaluations from individual customers; the platform business retains the right to discharge producers whose ratings are considered inadequate. For the platform, this allows supervision and performance management to be outsourced, at low cost, to customers. For workers, of course, the system introduces enormous risks from unfair, arbitrary or inaccurate customer evaluations, and a compulsion to tolerate abusive or exploitive behaviour from customers for fear that their ratings may be adversely affected by complaining or resisting.^{Footnote 17} But so long as it is legally permissible to discharge workers on the basis of unverified consumer survey responses, this ratings system will be a powerful and inexpensive weapon in employers’ labour extraction arsenal.

The role of piece-work compensation in employers’ labour extraction efforts is worthy of additional discussion, given the importance of piece-work in digital platform businesses. On one hand, piece-work compensation seems like an obvious solution to the challenge of converting paid work time into exerted labour effort: the worker is only paid for production that actually occurs. This explains the consistent interest in these models by management theorists throughout the history of capitalism – from the dawn of Taylorism, through to modern digital applications. On the other hand, there are limitations to piece-work that constrain its usefulness for most jobs. For maximum impact, piece-work compensation must be calculated at the individual level, making the system unwieldy in jobs which require cooperation among teams of workers.^{Footnote 18} Moreover, in most jobs, labour output is difficult to measure – more complicated than counting widgets produced in an hour (or fast food deliveries in an evening). This measurement problem is especially acute where quality of output is important, not just quantity. Finally, most jobs involve a variety of tasks which must be performed under changing circumstances. In these conditions, both workers and managers need flexibility and the capacity to exercise judgement and problem-solving skills, rather than blind pursuit of a particular performance indicator. For all of these reasons, while piece-work has remained an important tool in the kit of employers, its applicability has been limited to a relatively small subset of the total work performed in a modern economy.

This gives cause to question the extent to which digital platforms, with their core reliance on piece-work compensation, could indeed spread throughout the economy. The limitations of piece-work compensation are indicated by odd-job digital platforms such as Taskrabbit or Airtasker, through which the end-user defines a job, and workers then bid for the work. But it is often difficult to fully and precisely describe the work that the end-user desires. The incompleteness of the subsequent (implicit) contract between user and producer gives rise to many disputes over whether a given job was completed adequately or completely.

The broader macroeconomic, political-economic and regulatory forces discussed above also enter the labour extraction calculation in several ways. The determination of wages clearly reflects the influence of regulatory institutions (such as minimum wage laws, the state of trade unionism and collective bargaining and other wage-regulating practices and institutions). In the context of modern digital platforms, the willingness of regulators to apply existing minimum standards (like minimum wages) to work performed through platforms has been spotty – partly because of the indeterminate status of platform-based producers in the eyes of traditional labour law and partly because many regulators welcome the ‘disruptive’ effect of digital platforms on regulatory levers to which they were not fully committed to in the first place. The regulatory environment also affects management efforts to boost labour intensity through threats of job loss, which depend on the legal powers of employers to monitor workers in increasingly intrusive ways, and discharge those with unacceptable performance. Social policy also impacts this process through the extent to which discharged workers receive income support while seeking another job.

The overall state of the labour market is another macro-level factor influencing employer labour extraction strategies. Wage pressures moderate in response to chronic excess labour supply – the sort that has been endemic in OECD countries over the past decade and more. But permanent unemployment and underemployment affect the labour extraction problem in other ways. When unemployment is more severe, the cost of job loss will be higher (since it will take workers longer to find a new job), and workers in general will feel greater compulsion to meet employers’ workplace demands.^{Footnote 19} Moreover, employers can be more confident of their ability to recruit additional labour when required to meet demand or operational conditions; this facilitates their willingness to utilise casual or temporary employment – including through platform-based models.

The general state of popular expectations regarding work and employment is also relevant, reflecting the big sweep of politics and culture. Effective work organisation requires consent as well as control (Reference BurawoyBurawoy, 1979; Reference EdwardsEdwards, 1990). Perhaps the greatest achievement of neoliberalism has been the construction of an attitude, common among young workers today, that they can expect nothing more from the labour market than an endless series of precarious ‘gigs’.