Over the last decade, the term ‘gig economy’ has risen to prominence in public discourse, but has failed to attract sustained attention from political philosophers. The gig economy is a subsection of the overall economy that predominantly relies on ‘on-demand work’:11 Prassl 2018, p. 11. workers22 Throughout this article, we deliberately speak of ‘workers’ in a broad sense: not only those permanently employed count as workers of a firm, but all those who actually work for it. receive short-term and freelance contracts rather than permanent jobs. Firms treat them as independent sellers of services and only hire them to perform a particular ‘gig’, that is, to complete a specified task or project. Gig work thereby increases the granularity of work contracts: work is sold in ever smaller quantities and, in extremis, firms only buy the exact amount of labour they need, at the particular moment they need it.

On a social level, the higher granularity of labour market transactions leads to an expansion of the reach of markets. Not only a one-off hiring decision, but the conditions of every single micro-labour contract become subject to the market forces of supply and demand. From an economic perspective, the ability to hire and dispense with workers at will is often seen as a measure for ‘removing friction’ in labour markets, and unlocking efficiency gains as ‘unproductive human capital’ is ‘set free’.33 On the flexibilization of labour markets in the 1990s, see Davies and Freedland 2007. Major expositions of this policy programme include European Commission 1994; OECD 1994. Whenever a firm can do without a worker’s additional unit of labour, she re-enters the labour market, which can then, in theory, allocate her work to the most productive use.

The business practices of ride-service providers like Uber and Lyft in the US best embody this trend, paying drivers by the ride and only if there are rides. But, beyond such extreme and well-publicized cases, many firms are gradually moving away from the paradigmatic form of employment in the industrial age—a job with a fixed number of hours, worked in set shifts, in one place, and for a predetermined salary (call this ‘standard employment’)—towards work that is contracted at short notice on the basis of current demand (call this ‘gig work’).44 Some other forms of work, like fixed-term employment with a longer duration or employment at temporary work agencies, lie somewhere in the middle of this spectrum. They often display many of the normatively problematic features of gig work, but to a lesser extent. While various statistics employ different criteria for determining whether or not someone counts as a contract worker, the number of people affected by this shift has dramatically increased over the past decades and is projected to rise further. According to a recent study, those in alternative work arrangements—like on-call workers, contract workers, and freelancers—made up 15.8 per cent of the US workforce in 2015, up from 10.7 per cent a decade before.55 Katz and Krueger 2019, p. 382. At Google, contract workers are now reported to outnumber permanent staff.66 Wakabayashi 2019.

The trend towards gig work poses a challenge to the paradigm of standard employment, which is modelled on the industrial worker and has dominated debates in political philosophy and labour economics.77 See also Vallas and Schor (2020), who focus on the specific context of the platform economy, where firms obtain a near-monopoly position in connecting buyers and sellers of labour. They contend that such gig-economy platforms require new regulatory approaches, as they reject responsibility for individual transactions, but wield considerable power over them. For one thing, many regulatory achievements of the past 150 years—social insurance coverage, dismissal protection, the five-day work week, minimum wages and paid sick leave—do not extend to gig work. Furthermore, such work gives rise to distinct concerns. In the context of standard employment, unemployment has been a major social concern. However, on-demand workers will typically not become unemployed, but underemployed: they will still have work, only less of it or at a lower rate. Existing social insurance, like unemployment coverage, offers little support, as it is not designed to cushion short-term income fluctuations.

This article addresses the recent, but neglected trend towards on-demand work and develops a new framework for its normative assessment that centres on the notion of risk. It proceeds in three steps. Section I presents our diagnosis. We draw on empirical evidence to detail the rise of flexible forms of employment and propose that the underlying shift in employment relations is best understood as a transformation of risk: owners of firms reduce their business risk by demanding more flexibility of their workers, thereby exposing them to greater personal risk. We identify five mechanisms that enable this risk transformation.

Section II argues that current trends and the resulting risk shift are normatively problematic. Forming the core of the article, this section has three parts. Section II.A evaluates the risk shift from the perspective of affected workers. We argue that it particularly affects low-skilled workers, diminishing their ability to form and enact long-term plans, thereby undermining their agency. Moreover, given their frequent lack of bargaining power (and compensatory benefits), their voluntary acceptance of gig work does not render it unproblematic. Indeed, we argue, the gig economy tends to entrench relationships of domination and, potentially, exploitation. Section II.B evaluates the risk shift from the perspective of society. We argue that firms in the gig economy operate in a socially unsustainable manner: they have a corrosive effect on the skill level of the workforce, on the provision of care work in affected families, and on social cohesion. Section II.C offers the caveat that, as it stands, our analysis reveals merely pro tanto reasons against the use of gig work; it examines the steps required to arrive at an all-things-considered judgement about the permissibility of gig work, and explains why, for the purposes of this article, no such judgement is needed.

Section III offers a more constructive, policy-oriented outlook. We introduce the idea of a Principle of Inverse Coverage (PIC), which stipulates that the shorter, more variable, and less predictable a form of employment, the higher should be the contributions to social insurance schemes made by the employer. This Pigouvian tax, which can be motivated along the lines of a carbon tax, has two goals: it aims to reduce the prevalence of gig work while further moderating the risk exposure of gig workers by funding a new insurance against income volatility.

在过去的十年里，“零工经济”这个词在公共话语中变得突出，但未能引起政治哲学家的持续关注。零工经济是整体经济的一部分，主要依赖“按需工作”：11普拉斯尔2018 年，第1页。11. 工人22在整篇文章中，我们特意在广义上谈论“工人”：不仅是那些长期受雇的人算作公司的工人，而且还包括所有实际为公司工作的人。获得短期和自由职业合同，而不是长期工作。公司将他们视为独立的服务销售商，只雇用他们执行特定的“零工”，即完成指定的任务或项目。因此，零工工作增加了工作合同的粒度：工作的销售量越来越小，而且在极端情况下，公司只在特定时刻购买他们需要的确切数量的劳动力。

在社会层面上，劳动力市场交易的更高粒度导致市场范围的扩大。不仅是一次性的招聘决定，而且每一份微型劳动合同的条件都受到市场供求力量的影响。从经济角度来看，随意雇用和裁员的能力通常被视为“消除劳动力市场摩擦”的一种措施，并在“非生产性人力资本”被“释放”时释放效率3。3关于 1990 年代劳动力市场的灵活性，参见 Davies 和 Freedland 2007。该政策计划的主要阐述包括欧盟委员会 1994；经合组织1994 年。 每当公司可以不用工人额外的劳动单位时，她就会重新进入劳动力市场，然后理论上可以将她的工作分配给最有生产力的用途。

美国 Uber 和 Lyft 等乘车服务提供商的商业实践最能体现这一趋势，即按乘车付费，并且仅在有乘车时才向司机付款。但是，除了这些极端和广为人知的案例之外，许多公司正在逐渐摆脱工业时代的典型就业形式——固定工时、轮班工作、在一个地方工作、按预定时间工作。工资（称为“标准就业”）——根据当前需求在短时间内签订合同的工作（称为“零工工作”）。44一些其他形式的工作，例如期限较长的固定期限工作或在临时工作机构工作，则处于这一范围的中间。它们通常表现出零工工作的许多规范上有问题的特征，但程度较轻。虽然各种统计数据采用不同的标准来确定某人是否算作合同工，但受这种转变影响的人数在过去几十年中急剧增加，预计还会进一步增加。根据最近的一项研究，2015 年，处于替代性工作安排中的人（如随叫随到的工人、合同工和自由职业者）占美国劳动力的 15.8%，而十年前这一比例为 10.7%。 55 Katz 和 Krueger 2019，第5页。382. 据报道，在 Google，合同工的数量现在超过了正式员工。 66若林2019 年。

零工工作的趋势对标准就业范式提出了挑战，该范式以产业工人为模型，并主导了政治哲学和劳动经济学的辩论。 77另见 Vallas 和 Schor (2020)，他们关注平台经济的特定背景，在这种情况下，公司在连接劳动力买卖双方方面获得了近乎垄断的地位。他们争辩说，这样的零工经济平台需要新的监管方法，因为他们拒绝对个别交易负责，但对它们拥有相当大的权力。一方面，过去 150 年的许多监管成就——社会保险覆盖、解雇保护、五天工作周、最低工资和带薪病假——并没有扩展到零工工作。此外，此类工作引起了不同的关注。在标准雇佣方面，未就业一直是备受社会各界关注。然而，按需工作者通常不会失业，而是就业不足：他们仍然有工作，只是减少了工作或以较低的速度工作。现有的社会保险，如失业保险，提供的支持很少，因为它不是为了缓解短期收入波动。

本文讨论了最近但被忽视的按需工作趋势，并为其规范性评估开发了一个新框架，该框架以风险概念为中心。它分三步进行。第一部分介绍了我们的诊断。我们利用经验证据来详细说明灵活就业形式的兴起，并提出最好将雇佣关系的潜在转变理解为风险的转变：公司所有者通过要求工人具有更大的灵活性来降低其业务风险，从而暴露他们到更大的个人风险。我们确定了实现这种风险转换的五种机制。

第二部分认为当前的趋势和由此产生的风险转移在规范上是有问题的。本节构成文章的核心，分为三个部分。第 II.A 部分从受影响工人的角度评估风险转移。我们认为，它特别影响低技能工人，削弱了他们制定和制定长期计划的能力，从而削弱了他们的能动性。此外，鉴于他们经常缺乏议价能力（和补偿性福利），他们自愿接受零工工作并没有使其成为问题。事实上，我们认为，零工经济往往会巩固统治关系和潜在的剥削关系。第 II.B 节从社会的角度评估风险转移。我们认为，零工经济中的公司以一种社会不可持续的方式运作：它们对劳动力的技能水平、受影响家庭的护理工作以及社会凝聚力产生腐蚀作用。第 II.C 节提供了一个警告，就目前而言，我们的分析仅揭示了反对使用零工工作的直接原因；它检查了达成目标所需的步骤所有的事情考虑的关于演出工作的允许判断，并解释了为什么，这篇文章的目的，是不需要这样的判断。

第三部分提供了更具建设性的、以政策为导向的前景。我们引入了反向覆盖原则(PIC)的概念，该原则规定，就业形式越短、可变性越大、可预测性越低，雇主对社会保险计划的缴款应越高。这种庇古税的动机类似于碳税，有两个目标：它旨在减少零工工作的流行，同时通过资助一项针对收入波动的新保险来进一步降低零工工人的风险敞口。