# copy the linklink copied!4. Labour market regulation 4.0: Protecting workers in a changing world of work

This chapter discusses the role of labour market regulations to adequately protect workers in a changing world of work. A key focus of the chapter is on employment status – a critical area because it acts as a gateway to various worker rights and protections. Ensuring the correct classification of workers is therefore a key first step to ensure access to labour and social protection, collective bargaining and lifelong learning. For some workers, however, there is genuine ambiguity about employment status as they find themselves somewhere in the “grey zone” between dependent and self-employment. While arguing that this grey zone should be kept as small as possible, the chapter examines the rationale and policy options for extending certain labour rights and protections to these workers. Finally, the chapter discusses the role of regulations in addressing abuses of monopsony power and rebalancing bargaining power between employers/clients and workers.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

In Brief
Key findings

Labour market regulation plays an important role in protecting workers, but a number of developments are challenging its role. First, the emergence of new forms of work poses a challenge to regulations largely designed for full-time, permanent employees working for a single employer. Second, there is increasing evidence of (and in some cases growth in) unbalanced power relationships between workers and their employers, which calls for a reassessment of how regulation can address both the consequences and sources of such power imbalances.

This chapter examines the role of labour market regulation in protecting workers in a changing world of work, and ensuring that firms that follow the rules are not put at a disadvantage. More specifically, it looks at the role of regulation in operationally defining employment status, extending protections beyond standard employees and rebalancing power asymmetries between employers/clients and workers. Its main findings are:

• Clearly defining the employment status of workers (e.g. the distinction between the self-employed and employees) in labour market regulation and enforcing it are crucial. Employment status acts as a gateway to various worker rights and protections – including employment and social protection, but also access to training and collective bargaining – see Chapters 5 to 7.

• Policy and law enforcement should minimise opportunities and incentives for the misclassification of workers. Certain employers may deliberately misclassify workers in an attempt to avoid employment regulation, tax obligations and workers’ representation, as well as to shift risks onto workers and/or gain a competitive advantage; and others may do so by mistake. Similarly, workers may choose among different employment forms in order to benefit from a better tax regime or simply avoid taxes – but at the risk of losing labour and social protections. Such misclassification therefore harms individuals, but also leaves firms that properly classify their workers at a competitive disadvantage, and damages public finances.

• Regulations or guidelines for determining employment status may need to be clarified, revised and/or harmonised. This would help reduce the size of the “grey zone” between self-employment and dependent employee status – i.e. workers who share certain characteristics of both forms of employment. Reducing the size of the grey zone would reduce uncertainty for both workers and employers, and reduce litigation.

• For workers who remain in the grey zone, where there is genuine ambiguity about their employment status, governments should seek to extend rights and protections. These workers have some of the characteristics of employees and, like them, they may find themselves in an unbalanced power relationship since employers often have a higher degree of control over the employment relationship than they do. At the same, they may be deprived of most of the standard rights and protections afforded to employees because they are usually classified as self-employed.

• Labour market regulations should be adapted so that workers in the grey zone can benefit, at least partially, from: fair pay, working time regulations, occupational health and safety, anti-discrimination legislation, as well as some form of employment protection. This will involve identifying those groups that need protection and clarifying who bears responsibility as an employer towards these workers. Not all protections can be extended to these workers in the “grey zone”. Chapters 5 to 7 discuss how this could be done for social protection, collective bargaining and training.

• Imbalances in power relationships between employers and workers (including many self-employed workers) can also emerge (or be worsened) when workers have few or no outside options and much lower bargaining power than employers (a situation usually called monopsony power in the labour market). There is growing empirical evidence that monopsony is important in many labour markets and that high market concentration of a few firms is associated with significantly lower pay and worse working conditions.

• Such situations can be made worse when workers are unable to organise and bargain collectively, which is usually the case for self-employed workers who tend to be banned from collective bargaining by antitrust regulation (see Chapter 5).

• The abuse of monopsony power in the labour market and its sources can be addressed by better regulation and more effective enforcement. This includes: i) extending coverage of labour market regulations to address the effects of monopsony on workers’ well-being; ii) more aggressively enforcing rules against employers colluding in the labour market e.g. through non-poaching agreements; iii) limiting the scope of non-compete agreements; and iv) using labour market regulation to redress information asymmetries between employers and workers. A comprehensive policy strategy to reduce labour market frictions and enhance job mobility would also help lessen the sources of monopsony power.

New, non-standard forms of work have received much attention in recent media, legal and policy debates – see for example ILO (2019[1]). Barely a day goes by without a newspaper article either praising or demonising work in the “platform” economy (in which workers provide services through online platforms), and many court cases are ongoing in which workers are challenging their employment status – giving the impression that employment regulation is increasingly out of date and policy makers unsure how to react. A recent OECD/European Commission survey confirms that addressing the issues raised by new forms of work is a major policy concern among member states (OECD, 2019[2]). However, these discussions are not limited to technology-induced forms of employment (i.e. platform work); they also cover other non-standard employment, such as on-call labour and own-account work more generally (see Chapter 2 for an overview of what is meant by new forms of work).

Underlying these debates is a fear that the standard, full-time and dependent employment relationship is under pressure and that, in the future, many individuals will be working in “flexible” work arrangements with little employment and social protection, few benefits and rights, and limited access to training. According to this view, the nature of firms is also changing and a growing number of them act as “intermediaries” in the production and delivery of products and services, rather than as producing firms in their own right. Consequently, firms will seek out new business models and engage in a “race to the bottom” in which working conditions become the basis for competition, rather than the quality or value of the product or service they provide. Such scenarios, if ever realised, would undermine many of the foundations on which welfare states were built last century, and would require a serious re-thinking of labour market, social and skills policies and institutions, as well as of traditional labour relations and social dialogue.

However, the debate has a tendency to get ahead of the facts, and policy makers should be careful to base any decisions they make on evidence rather than anecdotes, as well as on a balanced consideration of all the arguments.

First, it is important to remember that new, non-standard forms of work often emerge in response to the real needs of both employers and workers. For example, companies need to have sufficient flexibility to adjust workforces and working hours in response to fluctuating and unpredictable demand. Workers may be seeking greater flexibility to fit work around caring responsibilities and/or leisure in order to achieve a better work-life balance. Many workers also want more independence in the way they organise their work and hours. Diversity (and continuous innovation) in employment contracts allows both employers and workers to escape the constraints of a “one-size-fits-all” approach and find arrangements that are in the best interest of both. It is equally important to point out that “non-standard” employment (which is what many new forms of work are) does not necessarily equate to poor quality employment. Indeed, standard jobs can be low quality, while non-standard ones can be high quality. In fact, across the OECD, most individuals in temporary and part-time contracts now have access to similar rights and benefits as standard employees (although, this does not rule out that there might be practical barriers in exerting these rights and accessing benefits for temporary and part-time workers, and that their job might still be more precarious and of lower quality). Moreover, many high-tech professionals sell their services as independent contractors, and new digital intermediaries such as online platforms have made it possible for them to reach more rapidly a much larger, often worldwide, market.

Second, despite the fact there has been rapid growth in some “new”, forms of work like platform work and zero-hours contracts, they remain small as a share of total employment (see Chapter 2). Standard, full-time, permanent employment still remains the norm across the OECD (i.e. it accounts for the majority of all employment), even if non-standard employment as a whole may represent a relatively large share of all employment. There are several reasons for the continuing appeal of more stable, permanent employment arrangements. From the workers’ side, such contracts provide more certainty and allow them to plan ahead in both their private and professional lives. From the firms’ perspective, permanent contracts allow them to attract and retain talent (by increasing loyalty), which reduces hiring and training costs, and increases the pay-off from investing in staff (which raises productivity). Consequently, there is no logical reason to believe that standard work will completely disappear in the near future.

However, while the growth in non-standard employment may sometimes have been overstated and the concerns regarding their intrinsic job quality exaggerated, there are some important issues which policy makers need to address.

The first of these is the question of employment status. Many “new” forms of work have emerged in the grey zone between employee and self-employed status. This raises questions as to which labour rights and protections apply to these workers and whether or not such forms of work are being used merely to avoid costs and regulations at the expense of job quality. More broadly, there is an issue of worker vulnerability and power imbalance vis-à-vis employers, regardless of employment status – which is the second key issue that governments may need to address. The vulnerabilities inherent in the employment relationship can become more acute when workers have no means of organising and bargaining collectively (see Chapter 5), as well as in labour markets characterised by monopsony power. The third and final issue concerns the international aspects of some work in the platform economy, which offers great opportunities for some workers but also risks a race to the bottom in working conditions for others.

While many of these issues have come to the fore as a result of the rise in the platform economy, they are not necessarily new and apply to many pre-existing non-standard forms of work too. Policy makers should therefore ensure that any reforms also cover other forms of non-standard work.

This chapter will discuss each of these policy challenges in turn. Section 4.1 reviews issues related to worker classification and explores options for addressing false self-employment. It also considers how to identify workers in the grey zone between employee and self-employed status, where genuine ambiguity remains but there may be a need to extend certain labour rights and protections. Section 4.2 discusses policy options for extending protection beyond standard employees. Section 4.3 examines power imbalances and the role of labour market monopsony. Section 4.4 briefly discusses the role of international competition in the platform economy and Section 4.5 provides concluding remarks and sets out some policy directions.

## copy the linklink copied!4.1. Employment status: A gateway to workers’ rights and protections

One issue that has recently received a lot of attention in media, legal and policy circles is that of how to define a worker’s employment status. In particular, there have been a number of high-profile tribunal cases in which workers have challenged their status as “self-employed” and argued for more rights and benefits, in line with what employees would get.

The reason why employment status matters so much (and hence why there is so much at stake in these court cases) is that it determines access to worker rights, benefits and protections. As an employee, one will generally be entitled to the minimum wage (where it exists), working time regulations and overtime pay, holidays, sickness and accidents insurance, unemployment benefits, protection against unfair dismissal and discrimination.1 This will not usually be the case for self-employed workers (see Figure 4.1). Getting employment status right therefore matters from a worker’s perspective.

Ensuring workers are correctly classified also matters for employers and society more generally. Employers want to avoid situations where competitors gain an advantage simply because they avoid taxes and regulations by disguising employment relationships with their workers as self-employment. From a government perspective, traditional employment is an important source of public revenue, accounting for a greater share of taxes per capita than self-employment. Employment misclassification can result in a significant loss of revenues. For example, in the United States, the Department of Labor has estimated that between 10-30% of workers are misclassified and that this could have a significant impact on tax revenue (Brumm, 2016[3]) and the Government Accountability Office placed the tax gap due to worker misclassification at USD 44.3 billion for the tax years 2008-2010 (GAO, 2017[4]). In the United Kingdom, Her Majesty's Revenue and Customs (HMRC) has estimated that the self-employed account for GBP 5 billion of the GBP 7 billion uncollected tax gap for self-assessment tax income – but this includes also the effects of under-reporting of income and error (Adam, Miller and Pope, 2017[5]). Such erosion of the tax base will have an impact on everyone in society, including businesses attempting to reduce their tax obligations.

The purpose of this section is to bring some clarity to these debates, and to help policy makers think through the issues and the associated policy responses. A key conclusion is that it is important to distinguish between cases of false self-employment and those cases where genuine ambiguity might exist, because this has implications for how policy makers should address the situation.

### 4.1.1. Tackling false self-employment: Ensuring that individuals are correctly classified

False self-employment (also sometimes referred to as disguised, sham, bogus or pseudo self-employment) refers to cases where individuals are classified as self-employed but, to all intents and purposes, work as employees (see Figure 4.1 above). The deliberate misclassification by employers of workers in an attempt to avoid employment regulation, fiscal obligations, workers’ representation, as well as to shift risks onto workers and/or gain a competitive advantage, should be cracked down on. Similarly, workers’ misclassification of themselves as self-employed to avoid taxes should be addressed. As discussed above, misclassification harms individuals, but also may put firms that follow the rules at a disadvantage and risks creating a hole in public finances.

To successfully fight false self-employment, firms and workers first need to be clear about the rules and regulations. Most countries have in place tests and criteria for assessing employment relationships (Box 4.1) and government action should ensure that firms and workers are aware of, and understand these rules (as well as the spirit in which they were written). For example, in the United States, the Department of Labor published guidance in 2015 to help with the classification of employees and independent contractors.2 Australia has launched an “Independent Contractors Decision Tool” which helps firms, based on a series of questions they need to answer, identify whether their workers should be classified as a contractor or as an employee.

Governments should also consider making it easier/less costly for workers to challenge their employment status, for example by: placing the burden of proof on the employer (rather than on the employee)3, reducing court fees4, simplifying procedures5, reducing the risks to workers, and/or protecting workers against potential retaliation. Some platforms have tried to get workers to sign arbitration agreements under which they waive the right to sue them in court. In the United States, the National Labor Relations Board had initially argued that such waivers violated labour law (Waas et al., 2017[7]) – however, a recent decision of the Supreme Court ruled against this position, at least as regards class actions.6

It is important to allow labour authorities and/or unions to take cases to court. For an individual, filing a complaint with a court is typically expensive and the outcome is uncertain. In addition, workers may worry about retaliation from their employer and losing their job. In some countries, the labour authority has some power to enforce compliance with labour laws – although this is usually limited and does not include the possibility of ordering a civil remedy or taking a claim to court without the consent of the aggrieved worker(s). In Australia, Chile, Poland, Spain and the United States, however, the labour authority can seek for a civil remedy on behalf of the aggrieved workers even in the absence of consent, particularly in cases where an important public interest is concerned (OECD, 2014[8]). In Sweden, trade unions can take employers to court on behalf of the worker (Williams and Lapeyre, 2017[9]).

Box 4.1. Tests and criteria for determining employment status

In practice, the question of whether individuals should be treated as employees or self-employed is often determined by courts. In most countries, the term “employee” is not defined in legislation or, even where it is, leaves broad room for judicial discretion (Davidov, Freedland and Countouris, 2015[6]). In many cases, an employee is defined circularly as “someone who is employed by an employer”.

In general, employee status tends to be decided based on the principle of “primacy of facts” – i.e. regardless of what is stipulated in the contract of employment, it is the actual facts of the working relationship which matter to determine whether the person is genuinely self-employed or not (see Figure 4.1 above).7 One of the reasons for looking at the facts of the relationship rather than at what is stipulated in the contract is that there may be inequality of bargaining power between workers and employers – and the latter are more likely to be able to determine what is written in the employment contract (see Section 4.3).

To assess the employment relationship, a number of tests/criteria are used. In a few civil law countries, these criteria tend to be enshrined in law and there will be a presumption of an employment relationship if these criteria are met, with the burden of proof put on the employer to show that this is not the case. In most countries, however, (including common law countries), judges base their rulings on certain tests developed by case law.

There will be differences in the number of criteria used in these tests. For example, in Canada, courts sometimes use a fourfold test, while in the United States courts have tended to use a 13- or even 20-factor test (Davidov, Freedland and Countouris, 2015[6]). The ILO’s recommendation concerning the employment relationship lists 14 factors.

Notwithstanding these differences, the actual tests/criteria used are very similar across countries.8 Financial dependence of the individual on the client is one of the aspects considered. However, financial dependence is usually not sufficient for employee status to be established. In most countries, there also needs to be an element of subordination of the worker (and control from the client). This may be assessed on the basis of several criteria, including: the worker’s integration in the organisation; the extent to which the worker controls his/her conditions of work (including the place and time of work); who provides the tools, materials or machines; the regularity of payments; the extent to which the worker takes on financial/entrepreneurial risk; and whether or not the work must be carried out personally by the worker. There are differences across countries (but also within countries across judges and over time) on the relative emphasis placed on control versus dependency.

Strengthening the penalties for firms not complying with legislation could also help address false self-employment. Where consequences of abuse are minimal, firms may have little incentive to correctly classify workers. Policy options include: requalification of the employment relationship; imposing retroactive payment of taxes and social security contributions; imposing greater penalties if firms continue to breach the law in repeated comparable cases; and extending the application of tribunal judgments beyond the plaintiffs and to the entire workforce in a similar situation.9 In cases where responsibility can easily be passed on because there are multiple customers or contractors (e.g. in multi-party employment relationships), governments can hold the entire chain jointly and severally liable (“chain liability”) if labour regulations are broken. This was done in the Netherlands in 2016, when the Act Combating Bogus Self-Employed came into force.10

Actions to facilitate legal challenges and increase penalties should be combined with efforts to strengthen the labour inspectorate’s capacity to monitor and detect breaches, e.g. increased responsibilities and resources (including the number of inspectors), innovative methods to inspect those working from home/on platforms (e.g. new technological tools), and training. Some countries (e.g. Ireland, Spain and Greece) are also targeting inspection efforts on particular sectors or geographical areas known to have a greater prevalence of false self-employment (OECD, 2019[2]). The extra costs involved should be balanced against the potential revenue gains from clawing back taxes and social security contributions lost as a result of misclassification.

Because social security and tax authorities have strong incentives to ensure the right classification of workers, it might also make sense to ensure that they coordinate their efforts with other enforcement authorities. In the United States, for example, the Internal Revenue Service has made employment tax compliance, including repressing misclassification of workers as independent contractors, an enforcement priority and coordinates its actions with those of the Department of Labor (Internal Revenue Service, 2005[10]; Keneally, Saleski and Engell, 2015[11]). Similarly, in Sweden, the tax agency played a pivotal role in inducing many platforms to accept employer status (Söderqvist, 2018[12]).

In parallel to strengthening the penalties for misclassification, governments should seek to reduce any incentives for firms and workers to misclassify employment relationships as self-employment.11 In some countries, tax and/or employment regulation have created incentives for employers and/or individuals to shift away from standard employee relationships to self-employment, or other non-standard forms of employment. For example, in Italy, the legislation introduced in 1997 and 2003 to legalise temporary work agencies and collaboration agreements may have led to an increase in self-employed workers who are in fact in an employer-employee-like relationship, but are classified as own-account workers (OECD, 2015[13]). In Australia, employers have an incentive to use casual workers who fall back on tax-financed benefits rather than on employer-sponsored benefits (OECD, 2018[14]). Further examples for the Netherlands and the United Kingdom are discussed in Box 4.2. Some of these incentives were not deliberate, but many countries have in the past introduced specific advantages/benefits in order to encourage self-employment/entrepreneurship.

Some countries have tried to address these incentives in an attempt to make independent worker status more neutral compared with employee status. For example, a tax reform was introduced in the Czech Republic in 2004 to halt the spread of “false” self-employment, although it was overturned in 2007. While it remains difficult to isolate the effect of policy reforms from other factors, the incidence of own-account work increased less in the Czech Republic than in the Slovak Republic during this period (OECD, 2008[18]). In Austria, concerns that independent contractors (freie Dienstnehmer) would be used by employers to evade taxes and regulations led the government to gradually integrate them into the social protection system and, since 2008, they pay the same social security contributions as standard employees (OECD, 2018[14]). In Italy, the pension contribution rates for employers and workers in the case of dependent self-employment (collaboratori) have been gradually increased since 2012, with the aim of reducing incentives for misclassification (OECD, 2019[2]). In Latvia, the micro-enterprise tax has been increased from 9% to 15%, and the maximum turnover has been reduced in 2018, to try and crack down on bogus self-employment (Golubeva, 2018[19]).

Policy makers could also consider measures to encourage hiring on standard contracts by making them more attractive relative to non-standard employment relationships. Making standard employment more attractive could be achieved by easing the obligations or enhancing the flexibility associated with standard contracts for employers – while still ensuring an adequate level of protection for workers. For example, as part of its 2015 labour market reform, Italy provided a temporary amnesty from fines to employers who convert existing self-employed contracts into standard, open-ended employment contracts (Williams and Lapeyre, 2017[9]).12 In the Netherlands, the government aims to encourage small and medium-sized businesses to hire on open-ended contracts by reducing sick pay obligations (OECD, 2019[2]).

Box 4.2. Policy-induced rises in self-employment: The cases of the Netherlands and the United Kingdom

A recent OECD study analyses the tax treatment of the different employment forms in a set of eight countries, in order to assess to what extent the tax system contributed to the rising shares of non-standard work (Milanez and Bratta, 2019[15]). The authors find that tax treatment differentials cause total employment costs between employees and the self-employed to vary. As a result, non-wage costs are often higher for standard employees than for the self-employed and, at times, this differential may be large enough to shift employer-employee preferences toward self-employment.

The situation in the Netherlands and the United Kingdom provides two case studies of how policy choices have created strong incentives for choosing self-employment over other employment forms.

In the Netherlands, self-employed workers are not subject to most social security contributions and non-tax compulsory payments (e.g. pension payments). In addition, unincorporated self-employed workers are eligible for two reductions in the personal income tax (PIT) tax base: i) a self-employment deduction of EUR 7 280 for those who work more than 1 225 hours annually; and ii) an SME exemption equal to 14% of the gross wage, net of the aforementioned self-employment deduction. As a result, total employment costs are around 30% higher for traditional employees than for the self-employed at the average wage (Milanez and Bratta, 2019[15]).

This arrangement shifts pension and insurance costs to the individual. Although the self-employed may self-insure by purchasing private insurance or increasing pension contributions, many do not. For example, in the Netherlands, only one in three own-account workers takes out insurance against incapacity (Ministerie van Financiën, 2015[16]). In addition to this tax incentive, until recently, procedures for declaring employment status to the tax office facilitated growth in self-employment. Workers in a weak bargaining position were encouraged by employers to declare themselves as self-employed, and the employer risked little because the responsibility for declaring the correct employment status was on the worker. Under a new system introduced in 2016, the burden of misclassification is on the employer – who will be liable for all insurance and tax payments if the contractor is found to be an employee. However, enforcement of this new measure is pending following a very negative reaction from various stakeholders, including the self-employed themselves.

In the United Kingdom, the government funds an effective self-employment subsidy of GBP 5.1 billion, or GBP 1 240 per person per year (UK Parliament, 2017[17]). Initially, the tax differential aimed to promote entrepreneurship – but the policy fails to target this population effectively. Instead, it incentivises people to identify as self-employed so they can save money, without commensurate reduction in benefits: in the United Kingdom, the difference in benefits entitlement between employees and the self-employed is limited to contribution-based jobseeker’s allowance or statutory maternity/paternity/adoption/shared parental pay (see also Chapter 7). The implication is that regular workers subsidise the self-employed since they pay higher contributions without a commensurate increase in benefits (Adam, Miller and Pope, 2017[5]).

### 4.1.2. Workers in the grey zone between self- and dependent employment

In most cases where individuals are falsely self-employed, courts will be able to determine this relatively easily using the criteria and tests described in Box 4.1. However, there are also cases where the issue is less clear, and where a genuine ambiguity may remain (see Figure 4.1 above). Some workers share characteristics of the self-employed (for example, they can choose when and where to work; they use their own equipment); but they also share characteristics of employees (e.g. they cannot set their own rates of pay, they may have to wear a uniform, they cannot be replaced in executing their tasks by someone else).

This issue of the “grey zone” between employee and self-employed status has gained prominence in recent times as a result of the rise of the platform economy and the numerous court cases which have ensued from it.13 In one such case, the judge recognised the ambiguity and stated that “The jury […] will be handed a square peg and asked to choose between two round holes.”14 However, the issue of the “grey zone” is by no means a new challenge. As far back as 1944, Justice Wiley Blount Rutledge of the US Supreme Court stated that “Few problems in the law have given a greater variety of application and conflict than the cases arising in the borderland between what is clearly an employer-employee relationship and what is clearly one of independent entrepreneurial dealing”.15 Some researchers even trace the issue back to Roman law (Rubinstein, 2012[20]). In all likelihood, this grey zone will always exist – although its boundaries and scope may vary over time as new business models emerge, technology advances, court practices change and policy changes interact with all of these. Indeed, while there were hopes that many of the court cases against platforms would settle the issue once and for all, in practice the decisions of these cases have been inconsistent in both Europe and the United States (Cherry and De Stefano, 2018[21]).

From a policy perspective, this grey zone matters because workers who find themselves in it share some characteristics with dependent employees. Because of this, they will also share some of the vulnerabilities of employees. However, because these workers tend to be classified as self-employed, they will not benefit from most of the rights and protections given to employees. As a result, it can be argued that some of these rights and protections should be extended to workers in the grey zone. Indeed, workers in a situation of dependence and/or subordination are by definition in a position of unequal bargaining power,16 and one of the key objectives of labour law is to redress such inequality (and/or its consequences). The challenges for policy makers are to identify who are the workers in the grey zone and decide which labour laws and protections should be extended to them (and how).

In a first instance, the size of the grey zone should be managed and kept to a minimum. In some cases, regulations or guidelines for determining employment status may need to be clarified, revised and/or harmonised17. This would help reduce uncertainty for both workers and employers, and reduce litigation (Linder, 1999[22]). For example, in recent years, many countries have laid down criteria aimed at better circumscribing the status of a self-employed worker and distinguishing it from that of an employee – e.g. Belgium in 2006. In doing so, countries face a trade-off between simple rules and broad guidelines (Box 4.3). Simple rules provide clear and unambiguous decisions about employment status, but risk excluding certain workers who may also need labour protections. Broad guidelines leave considerable discretion to adjudicators (enforcement officers or judges) with the potential to extend protection to a much larger group of workers (but at the risk of introducing a greater element of uncertainty and arbitrariness).

Box 4.3. Simple versus complex rules for establishing employment status

In a few countries, the criteria used for determining employment status define very precise, simple and unambiguously applicable rules. For example, in Italy, there is a rebuttable presumption of an employment relationship in all cases of contract for services if at least two of the following conditions hold: i) the relationship lasts for more than eight months within the same year; ii) associated worker’s compensation represents more than 80% of the total compensation earned by the worker within the year; and iii) the worker has his/her work space at the employer’s offices.18 If a relationship meets at least two of these criteria, then the contract will be re-classified as an employment contract and the worker will be entitled to all the rights, benefits and obligations of a standard employee.19 Similarly, in Greece, an individual supplying work to primarily one employer for a period of nine consecutive months is assumed to be in dependent employment.

While simple rules like these make it relatively straightforward to determine employment status, they inevitably leave unprotected a number of other workers sharing some of the characteristics of dependent employees. Also, strict rules are easier to work around, and such an approach does not take into account new forms of employment that might emerge.

In most other countries, the tests developed through statutory law, guidelines for enforcement agencies or jurisprudence are more complex. The key difference with respect to simple, automatic rules is that all different factors must be jointly assessed in a holistic way. For example, in Canada, labour inspectors and health and safety officers are explicitly instructed to examine all aspects of the relationship and take into account the different factors, bearing in mind that they do not represent an exhaustive list and that their relative weight depends on the particular facts and circumstances of each case (Employment and Social Development Canada, 2006[23]).

The advantage of an ex post, holistic evaluation of the different factors characterising employee and self-employment status resides in its flexibility, which in practice allows for protection to be extended to a much larger group of workers who do not share all the characteristics of dependent employees but who are nonetheless quite similar to them. Precisely defining what it means to be an employee may, in fact, be impossible.20

However, by leaving large discretion to adjudicators to appreciate the specificity of each case and the weight to be given to different factors, this approach inevitably introduces a degree of uncertainty and arbitrariness.21 There is, in fact, a large empirical literature showing that different judges in the same jurisdiction tend to have persistent appraisal differences – see e.g. Waldfogel (1998[24]) and Aizer and Doyle (2015[25]). Thus, if the same case had been assigned to a different adjudicator, the probability of a different outcome is typically high (Fischman, 2014[26]). While this problem can be lessened by resorting to panels of adjudicators, the evidence suggests that it persists even in panels of randomly assigned judges, and cases decided by unanimous vote are no proof of the absence of inconsistency across panels (Fischman, 2011[27]).

Adjudicators in social and labour law also face these problems – see e.g. Autor et al. (2017[28]); Ichino and Pinotti (2012[29]); Breda et al. (2017[30]). As regards detecting employee misclassification, given the genuine ambiguity in the status of workers that has been uncovered in many of the court cases, it is no surprise that apparently similar situations have yielded different outcomes within the same jurisdictions (Davidov, Freedland and Countouris, 2015[6]). This setting may not be ideal for both businesses and workers. Employers would be confronted with a relatively high level of legal uncertainty and, possibly, unexpected changes of legal standards, thereby increasing their potential costs. Workers would find the enforcement of their rights being partially dependent on random events, such as the assignment of the court to a judge with favourable or unfavourable attitudes. This suggests that protections of workers should not be uniquely dependent on being granted employee status by adjudicators and should be, at least partially, extended to all situations where genuine ambiguity remains (see Figure 4.1 above).

An equally important task for policy makers is to clarify who, in triangular employment arrangements (such as those involving a user-firm, a subcontractor and a worker employed by the latter but providing services to the former within its premises) and those involving intermediaries (like many of those in the platform economy), is the employer and who, therefore, bears the responsibility for complying with labour market regulation. In such arrangements, there is a strong argument for clarifying obligations and, where necessary, spreading them across multiple legal entities – for example by holding the intermediaries and the clients jointly and severally liable, or by imposing liability on the intermediary and only subsidiary liability on the client (Box 4.4).

Box 4.4. Identifying who is an employer

Much of the discussion around new forms of work centres around the question of who is an employee, and who is not. A closely related (though less discussed) question is who is an employer, and who is not. This is particularly complicated in multi-party (or triangular) employment relationships, such as those observed in Temporary Agency Work (TWA) and subcontracting, but also in many of the new working arrangements that are emerging in the platform economy.

The question in such set-ups is who is responsible for worker rights and protections. In the case of TWA work – which was initially banned or heavily restricted in many countries (Countouris et al., 2016[31]) – the employment relationship is generally assumed to be between the worker and the agency, and the latter is therefore responsible for ensuring labour law is complied with. That being said, there might also be a legal requirement for the user company to guarantee certain worker rights and protections including, for example, health and safety at work, and to be jointly responsible for others together with the agency (OECD, 2014[8]).

The platform economy has complicated this landscape even further, and it is not clear to what extent the TWA experience might be a useful example for regulating platform work (Lenaerts et al., 2018[32]) – although the TWA model seems to have been accepted by many platforms in Sweden (Söderqvist, 2018[12]) and several platforms worldwide have taken the initiative to treat their workers as employees (Cherry and Aloisi, 2017[33]). Platforms usually argue that they are not employers but mere intermediaries providing the infrastructure for the worker to find clients. However, it is sometimes hard to argue that clients themselves should be considered to be the employer. Platform work typically involves a multiplicity of clients and tasks of a very short duration, even if these tasks are sometimes carried out on the premises of the client. At the same time, many platforms exert significant control over workers (through ratings systems, the management of payments, the withholding of information about clients, controlling the way work is carried out, the deactivation of accounts, etc.)

In some cases, where clear responsibilities cannot be assigned, there may therefore be an argument for platforms and users to bear joint and several liability for worker rights, so that a worker can claim against both or either. In other cases, the user might be argued to bear subsidiary liability – i.e. the worker can claim against the user in those cases where the platform does not comply with the regulation. Along similar lines, some authors have argued that the question of who is responsible for worker rights and protections should be analysed from the perspective of what are the key employer functions – from hiring workers to setting their rates of pay (Prassl and Risak, 2016[34]). The outcome of such an approach would equally be that employment law obligations are spread across multiple legal entities, rather than ascribed to a single employer in the classical sense of the term.

For those workers who are left in the grey zone (and therefore tend to be excluded from much of the scope of existing labour law), legislators should consider how (and which) labour rights and protections could be extended. Indeed, clarifying and stating as precisely as possible the scope of the law is unlikely to eliminate the grey zone entirely and hence the uncertainty for workers and employers. Many countries have identified specific groups of workers to which certain parts of the labour law are to be applied but with different approaches, each with their advantages and disadvantages. Some countries have identified very specific occupations; others have focused on the economically dependent self-employed; and still others have relied on vaguer (but broader) definitions. Each of these are discussed below.

Some countries have identified very specific occupations to which certain labour rights and protections have been extended. For example, in France, the following occupations are presumed to have an employment relationship under certain conditions: performing artists (also in Spain), models, professional journalists, and sales representatives and travelling salespersons (Pedersini, 2002[35]; ILO, 2005[36]). In Mexico, there are similar legal provisions for sales representatives, insurance salespersons, travelling salespersons, sales promoters and similar categories.

In several countries, policy attention has focused on the “dependent” self-employed (Box 4.5). A specific income threshold is usually set. In Spain, the economically dependent self-employed (trabajador autónomo económicamente dependiente – TRADE) need to rely on a single client for at least 75% of their income. In Portugal, an individual is considered dependent self-employed when at least 50% of their yearly income comes from one single client (regime dos trabalhadores independentes e que prestam serviços maioritariamente a uma entidade contratante). In Germany, workers are “employee-like” persons (arbeitnehmerähnliche Person) if they: i) work on the basis of a contract for service or a contract for work and services for other persons; ii) do so personally and essentially without collaboration with employees; and iii) receive 50% of their income from a single client (33% for artists, writers and journalists). In Canada, there is no fixed threshold, but “dependent contractors” cover “those non-employment work relationships that exhibit a certain minimum economic dependency, which may be demonstrated by complete or near-complete exclusivity”.22 Similarly, in Sweden, there is no precise definition, but dependent contractors (jämställda/beroende uppdragstagare) include “any person who performs work for another and is not thereby employed by that other person but who occupies a position of essentially the same nature as that of an employee” (Rönnmar, 2004[37]).

While the definition of dependent self-employed in all these countries is similar, the purpose of the category varies significantly. In Portugal, it opens up access to social protection (unemployment insurance, parental benefits, sickness and invalidity protection, as well as old age and survivors pensions). In Germany, employee-like persons have freedom of association, the right to collective bargaining, and the right to minimum leave of four weeks (Däubler, 2016[38]). In both Canada and Sweden, dependent contractors also have a right to collective bargaining and need to be given reasonable notice of the termination of the contractor relationship. In Spain, the TRADE category has access to a wide range of rights and protections, including: minimum wage, annual leave, entitlements in case of wrongful termination, leave for family or health reasons, and the right to collective bargaining (Cherry and Aloisi, 2017[33]).

Other countries have relied on a vaguer definition of an intermediate (or “third worker”) category to which some of the rights and protections of employees have been extended, though not all. These third worker categories often combine elements of financial dependency with elements of control/subordination. For example, in the United Kingdom, the statutory category of “worker” was introduced to broaden the reach of employment regulations and to include individuals who had been excluded from employee status by a judiciary taking a very narrow interpretation of the term. The worker category is not precisely defined but is intended to cover any individual who works under a contract to provide a personal service, independently of whether he/she has a contract of employment. These workers are entitled to protection against discrimination under the Equality Act 2010 – implying also a right to equal treatment with standard employees in basic working conditions. In addition, they are covered by selected labour regulations including those on working time, holiday pay and the minimum wage. They may also be entitled to sick pay and parental pay, though they do not benefit from minimum notice periods, health and safety provisions, protection against unfair dismissal and redundancy pay. In addition, they are responsible for their own social security contributions and tax (like the self-employed). While the original intention of the worker category was to broaden the reach of protective employment regulations to a greater number of workers, the creation of this separate category of workers may have shifted the objective of litigation down from obtaining employee status to merely obtaining worker status – as has been evidenced in recent court cases involving ride-sharing services.23

Box 4.5. The incidence of dependent self-employment

The “dependent” self-employed are usually defined as those own-account workers who rely for a large share of their income on a single client/employer. This can be a useful category to focus on in order to extend labour rights and protections to certain workers in the grey zone. But how important is this category of workers?

Measuring dependent self-employment is complicated given that: i) only a few countries have official definitions of dependent self-employment and, where these exist, they tend to differ; and ii) standard labour force and household surveys do not permit the identification of such workers.

Based on a special module of the European Union Labour Force Survey (EULFS), it is possible to get some estimates of the incidence of dependent self-employment, defined here as those own-account workers who generally have one dominant client. In many countries, dependent self-employment represents a non-negligible fraction of total self-employment (16%, on average) – Figure 4.2.24 Moreover, according to data from the European Working Conditions Survey, this share has risen between 2010 and 2015 across the countries covered by around 20%.

In Italy, the “semi-subordinate” worker category (lavoro parasubordinato) – covering co.co.co and co.co.pro contracts25 – was initially created to increase flexibility in the labour market and to reduce labour costs. These were self-employed workers characterised by: a certain continuity and length of the working arrangement; the individual nature of the service being provided; and some subordination to the client. These workers had access to employer-financed social protection (employers paid a reduced rate). In addition, the rules for termination of these contracts were less stringent than those of employees (OECD, 2014[8]). In the Italian case, the creation of the semi-subordinate worker category created clear incentives for employers to replace “standard” employees with workers in the less protected (and cheaper) intermediate category (Liebman, 2006[39]; Muehlberger, 2007[40]). In response to this situation, recent reforms in Italy have all but abolished these contracts (Cherry and Aloisi, 2017[33]). Both in terms of social protection and termination of contracts, the rules are gradually being aligned with those of employees on fixed-term contracts.

These two examples from Italy and the United Kingdom illustrate the dangers inherent in creating an intermediate category of worker where the boundaries of this category are vaguely defined or where it is created to introduce more flexibility in the labour market. The spirit of reforms to employment and social protection regulation should be to extend rights and broaden the reach of these regulations to vulnerable workers who were previously excluded, and not to create opportunities to take rights and protections away from workers who previously had them.

These potential pitfalls of creating a third worker category should be borne in mind in the context of the platform economy, which has changed the shape and size of the “grey zone” between employee and self-employed status, and which has spurred some to suggest a third worker category to cover those workers. In particular, creating a third worker category with vague boundaries would create a significant potential for downgrading. In the United States, Harris and Krueger (2015[41]) have argued for an “independent worker” category which would qualify for many, though not all, of the benefits and protections that employees receive, including the freedom to organise and collectively bargain, civil rights protections, tax withholdings, and employer contributions for payroll taxes. However, they would not qualify for hours-based benefits, including overtime or minimum wage requirements, nor for unemployment insurance benefits. In France, the government has introduced the idea of voluntary charters, which platforms would draw up and which would set out the rights and protections of platform workers. The French idea is an interesting one because it allows the government to buy time in terms of deciding on a certain policy action, while at the same time ensuring more rights and protections for these workers. The risk, however, is that these rights and protections will be set unilaterally by platforms (without consulting workers) at a low level of protection. Moreover, the lack of common protection standards across charters may reduce mobility and therefore heighten the monopsony power of platforms and exert downward pressure on pay (see Section 4.3 below). And a further concern is that, if they are not time-bound, they will end up forming a permanent third worker category (or, rather, multiple third worker categories) with diminished rights. Ideally, therefore, the charters should be seen as a temporary experiment from which governments can learn. In the longer-term, more sustainable solutions should be envisaged and the risk of creating a permanent third “platform” worker category should be avoided. New regulations should be flexible and broad enough to encompass various types of employment, regardless of the means through which work is obtained.

Alternative solutions could also be imagined to extend labour regulations to a greater number of workers in the grey zone. One approach would be to clarify as much as possible the definitions of “employee” and “self-employed”, and to treat everyone in the “grey zone” as an “employee” as far as labour law is concerned (until and unless it can be proved that these workers are properly self-employed).26 27 This was a route followed by the California Supreme Court in a recent case.28 Also, in Sweden, a Supreme Court decision dating back to 199629 concluded that, because the circumstances were ambiguous and it was difficult to make a clear judgement, the ruling went in favour of the plaintiff and an employment relationship was assumed (Åberg, 2016[42]). This could be the default position, while allowing specific parts of the law to create exceptions for those rights and protections which are more difficult (if not impossible) to extend to self-employed workers. For example, while it may be relatively straightforward to extend health and safety regulations as well as anti-discrimination laws to all own-account workers, it becomes much more difficult to do so for minimum wage legislation and working time regulations. These issues are discussed further in Section 4.2 below.

## copy the linklink copied!4.2. Extending rights and protections beyond standard employees

Self-employed workers are not usually covered by the same rights and protections as employees (see Figure 4.1 above). The key reason for this is that the self-employed are considered to be business undertakings, taking on their own risks in return for the prospect of earning profits. In principle, they do so with total autonomy, which is not the case for employees. In fact, the key reason why employment relationships are usually governed by labour law is to protect workers from the potential abuse of bargaining power and asymmetric control of information in the hands of the employer. The self-employed usually do not have access to the same social protection either (for example, they are often excluded from unemployment insurance) because the odds of losing one’s income depend on their sole efforts and decisions, and as such are considered to be part of the entrepreneurial risk they take on. Finally, because they are considered to be business undertakings, the self-employed are usually banned from collective bargaining as well, since this would be akin to price-fixing under competition law.

Yet, as argued in Section 4.1.2, there are a number of self-employed workers who share some (though not all) of the characteristics of dependent employees and for whom it might make sense to extend some of the rights and protections that employees benefit from (see Figure 4.1 above). Chapters 5 to 7 address the questions of how social protection and collective bargaining, but also training programmes, could be extended to these workers. The remainder of this section discusses some key aspects of labour market regulation and the extent to which (and how) they might be extended to workers in the “grey zone”. The section highlights that there are often significant practical difficulties in extending labour rights and protections to these workers – although these challenges are greater for some parts of the labour law than for others. In some cases, there might be a need to adapt existing regulations so that they can be applied to workers in the grey zone. More generally, the same policy issues may apply also to non-standard employees (such as those with an on-call or a zero-hour contract) who are in principle covered by labour law protections but are confronted to practical difficulties in exerting their rights (see Figure 4.1 above). Of course, a balanced approach is necessary to avoid that excessively burdensome regulations end up unduly curbing entrepreneurial activity and innovation.

### 4.2.1. Fair pay

For employees in standard employment arrangements, a legally binding minimum wage and collectively negotiated wage floors can help to prevent exploitation and address in-work poverty. Many studies of minimum wage impacts find that small increases in the minimum wage from a moderate level have no negative employment effects. This is contrary to standard theory but consistent with monopsony power and could suggest that there is some scope to set a minimum wage that exceeds the wage that would otherwise prevail in the labour market without harming employment levels (Card and Krueger, 2015[43]) – see also Annex 4.A.

Yet workers in the grey zone between dependent and self-employment (and, sometimes, other non-standard workers) are usually excluded from such arrangements. For example, neither minimum wages nor collectively negotiated wage floors tend to cover self-employed workers, who are considered “businesses” and are often paid per task that they complete (as opposed to the time-based compensation which most employees receive). While it may not be appropriate/necessary to introduce a minimum wage for many self-employed workers (e.g. entrepreneurs, highly paid freelancers or the “liberal professions”), there are many low-paid workers who will find themselves in the grey zone and/or who will face situations of monopsony power and bargaining asymmetry (see Sections 4.1.2 and 4.3). In those cases, it might be worth considering how mechanisms to achieve fair pay could be extended.

One way to extend minimum wage legislation to cover such workers would be to identify a group of workers in the grey zone (e.g. the dependent self-employed or certain occupations) and require employers to pay the equivalent of the minimum wage to individuals working on a piece rate basis. Piece rate legislation already exists in many countries as part of their minimum wage legislation and generally requires that employers pay individuals the equivalent of the minimum wage, regardless of the basis on which they are rewarded (i.e. by unit of time, or by unit of output).30 Employers are therefore expected to estimate the average productivity of piece rate workers and ensure that the average individual working at the average speed would earn the equivalent of the minimum wage. In most cases, such legislation is restricted to employees, and so the challenge lies in extending this to other vulnerable workers, including those in the grey zone between dependent and self-employment.

Proposals to introduce minimum rates for some groups of self-employed workers are being discussed in a number of countries. In the United Kingdom, it has been proposed by both the Taylor Review (BEIS, 2017[44]) and the Resolution Foundation (D’Arcy, 2017[45]). In the Netherlands, a legislative proposal has been put forward for the introduction of a minimum wage for the self-employed without employees and a revision of legislation to allow them to set minimum rates collectively (PvdA, 2017[46]). In Germany, a motion on the income of freelancers and solo self-employed, which also includes the demand for minimum fees, has been introduced in the federal parliament (Deutscher Bundestag, 2016[47]). In Poland, minimum wage legislation has recently (1 January 2017) been extended to work carried out under services agreements (umowy zlecenia, umowy o świadczenie usług). In particular, this covers individuals hired under services contracts and self-employed service providers who are not employers themselves or who do not contract work out to other parties. However, persons working under so-called specific task agreements (umowy o dzieło) are not covered, nor are providers of personal care services and service providers who determine the place and time of their services themselves and work on commission (OECD, 2019[2]).

While extending minimum wage legislation to some sub-groups of the self-employed may seem both desirable and (in theory) feasible, there are significant practical difficulties in doing so. These include: deciding what types of self-employed workers should be covered; defining who the employer is (which is particularly difficult in triangular working arrangements, see Box 4.4 above)31; measuring the average productivity of workers32 (especially for non-standard tasks); determining what counts as work33,34; and how to deal with work carried out across national borders (see Section 4.4 below). However, it is not impossible to overcome these difficulties. Since January 2018, for example, New York City has imposed a minimum wage for Uber and Lyft drivers.

Some voluntary initiatives have already been taken by a number of platforms to put a minimum floor under wages. For example, Favor, an on-demand delivery service in the United States, guarantees its drivers a minimum hourly wage. While its “runners” are paid by task, Favor will make up the difference if they do not meet the pay guarantee (Kessler, 2016[48]). Upwork has a global minimum wage of USD 3 per hour for tasks that pay by the hour. In the United Kingdom, Prolific also has a minimum wage per hour. In Australia, the online job-posting platform AirTasker has reached an agreement with unions to recommend minimum rates of pay to its users (Patty, 2017[49]) – although the platform does not guarantee this minimum pay (Kaine, 2017[50]). Some other websites, like the Czech Topdesigner.cz and the Spanish adtriboo.com, have set a minimum or even a fixed price for certain tasks, based on the average number of hours that workers spend on them (European Parliament, 2016).

However, because of the practical difficulties involved in extending minimum wage legislation to self-employed workers, alternative (or complementary) solutions could be sought for strengthening the pay of these workers. One of the most promising avenues might be to extend collective bargaining rights to certain groups of self-employed workers who find themselves in the grey zone between dependent and self-employment and, therefore, in situations of unequal power relationships. Canada, Germany and Sweden are examples of countries which have done this. One key challenge, however, is to ensure that labour market and competition policy are aligned on this issue – this is discussed more extensively in Chapter 5.

### 4.2.2. Regulating working time

Traditional concerns around working time have centred on the issues of excessive working hours. This is why labour legislation usually contains rules limiting working hours and requiring periods for rest and recuperation, including weekly rest and paid annual leave. It is difficult to see how some of these regulations might be extended to own-account workers who, in theory, choose their own working hours – even though, in practice, many (including those in the platform economy) often have limited ability to choose their own hours of work (Lehdonvirta, 2018[51]). Poor bargaining power combined with fierce competition between workers means that they may need to remain on-call if they do not want to lose out when new jobs/tasks are advertised. Moreover, in the case of certain micro-task platforms, workers spend as much time searching for tasks as they do in performing them (Kingsley, Gray and Suri, 2015[52]). Some platforms have introduced their own working hour limits (e.g. CloudFactory sets upper and lower limits on the amount of work each worker can complete in a week) and workers have adopted their own informal practices such as daily routines and quota setting to manage their time (Lehdonvirta, 2018[51]). Data collected through platforms can help in monitoring working time. However there are many complications with extending working time protections to such workers, including the fact that many of them have several clients/employers at any particular point in time, and therefore monitoring overall working time (and allocating responsibility) may be very difficult if not impossible. Again, certain exceptions could be envisaged for the economically dependent self-employed (e.g. minimum leave for employee-like persons in Germany) – but this may exclude many platform workers since they typically work for several clients at the same time (unless employer responsibilities are partially attributed to platforms – see Box 4.4).35 An alternative would be to focus on certain occupations.

For some workers, there are concerns about the unpredictability of working hours and/or the unwillingness of employers to guarantee a set number of working hours in the contract as they seek a maximum amount of flexibility to respond to fluctuating demand. This is an issue that goes beyond workers in the grey zone and relates to variable- and zero-hours contracts. Many countries have made reforms in this area in recent years (Box 4.6).

Box 4.6. Regulating variable- and zero-hours contracts

Many countries have special forms of atypical part-time employment contracts (or clauses within contracts) which either involve very short part-time hours or no established minimum hours at all. Some examples of these atypical part-time work arrangements include: casual workers in Australia, on-call contracts (lavoro a chiamata o intermittente) in Italy, min-max contracts and zero-hour contracts in the Netherlands, on-call contracts in New Zealand, zero-hour work in the United Kingdom, and if-and-when contracts and zero-hour contracts in Ireland. Many of these atypical part-time contracts have experienced rapid growth in recent years.

In response to concerns about the consequences of unpredictability in working hours on workers’ overall earnings, earnings volatility and their ability to plan ahead, there have been a number of reforms in recent years. This has particularly been the case for so-called “zero-hour” contracts. Reforms undertaken include: restricting the use of such contracts to situations where employers truly have a variable need for labour (Finland); requiring employers to provide information (such as the minimum number of hours) upfront or in the employment contract (Finland, Ireland and Norway); requiring advance notice of work schedules (Finland, Ireland, Netherlands, Norway and Oregon in the United States) or adjusting the contracted hours to reflect hours actually worked (Ireland); giving employees the right to request a more predictable contract after a certain period of time (Australia and the United Kingdom); compensation in case workers are called in but sent home again without work (Ireland) or in case they are expected to remain available outside guaranteed hours (New Zealand); and introducing provisions for sick pay and for compensation in cases of termination of employment (Finland). In addition, in the United Kingdom, there was a recent public consultation on introducing a higher minimum wage for such types of contracts to discourage their (over-) use (OECD, 2019[2]).

To a large extent, the concerns around variable-hours/on-call contracts derive from asymmetries in bargaining power between workers and employers. That is why some measures aim to strengthen the bargaining power of workers through, for example, the banning of exclusivity clauses (i.e. contract clauses which forbid workers from working for other employers – in the United Kingdom and a proposal in the Netherlands) or removal of obligations on the employee to accept work where the employer does not offer guaranteed hours (e.g. in New Zealand where firms must offer compensation for such availability). However, in some cases, countries have undermined the bargaining position of workers – for example, by expecting the unemployed to take up variable- (or zero-) hours contracts to be eligible for unemployment benefits (Kenner, 2017[53]).

### 4.2.3. Occupational safety and health

Some of the issues relating to working time also apply to occupational safety and health (OSH). Many new forms of work tend to transfer responsibilities for OSH from the employer to individual workers, who often lack the training or resources to take appropriate measures to ensure that their working conditions and the working environment are safe. Sometimes, strong competition between workers may result in corners being cut and unnecessary risks being taken. At the same time, labour inspectorates are often not adequately prepared to deal with these new forms of employment, which are often harder to reach, where it is not always clear that the existing legislation applies, and where the support from unions is often lower (Walters, 2017[54]). This makes it easier for such workers to avoid existing regulations like, for example, the need to obtain medical certificates in certain professions. Again, this is to some extent an issue of unbalanced bargaining power between employers and workers – although individuals themselves may also be short-sighted and/or want to avoid costs and regulations. In addition, there are the same difficulties in identifying the liable party due to the triangular nature of some relationships and the short nature of some of these tasks.

New forms of work also bring new or increased risks. For example, many platform activities are in the transport sector, where the risk of accidents is elevated. Recent evidence suggests that the arrival of ridesharing is associated with an increase of 2%-3% in the number of motor vehicle fatalities and fatal accidents as a result of increased congestion and road utilisation (Barrios, Hochberg and Yi, 2018[55]). There are also risks associated with online work – both physical and psychosocial, such as visual fatigue; musculoskeletal problems; work-related stress; chronic job and income insecurity; and isolation.

Again, the question of employment status is critical here as OSH regulation often only applies to employees. Special provisions for the dependent self-employed or certain occupations could be envisaged, while information campaigns and training could help strengthen OSH among own-account workers more generally. On the social protection side, some countries are beginning to think about how insurance for accidents at work might be extended to some previously excluded workers. For example, France has recently expected certain platforms to provide such insurance to some of their workers: if workers earning more than EUR 5 100 per year voluntarily insure themselves against the risk of occupational accidents or illness, the platform must provide reimbursement. The voluntary charters which France is considering might also lead to improved OSH for certain platform workers.

### 4.2.4. Anti-discrimination

Because of its potential negative impact on inclusiveness and basic human rights, but also on productivity, measures to tackle discrimination in the labour market on the basis of race, gender, religion, political opinion, socio-economic background, etc. have been put in place in all OECD countries (OECD, 2008[18]).

The emergence of the platform economy has an ambiguous effect on the ability to protect workers from discrimination. To the extent that platforms promote anonymity, they might help address discrimination. However, where such anonymity is not guaranteed, discrimination may be worse because of the lack of regulation and enforcement – see Edelman, Luca and Svirsky (2017[56]) for evidence of racial discrimination on Airbnb; Galperin and Greppi (2017[57]) for geographical discrimination on Nubelo (one of the largest Spanish-language online labour platforms); and Galperin, Cruces and Greppi (2017[58]) for evidence of gender discrimination, also on Nubelo. Moreover, while algorithms may offer the promise of removing human judgment and bias from decisions, there is some evidence that, in fact, they may reinforce human prejudice and embed biases of their own – see, for example, Sweeney (2013[59]) for such evidence on ads by Google AdSense.

This emerging evidence suggests that governments should think carefully about how non-discrimination laws might be extended to online platforms and independent workers more generally. Calls for labour platforms to collect (and publish) data on outcomes for various groups could be one step in the right direction. Focusing on certain occupations or the dependent self-employed could be another way of extending anti-discrimination legislation – see Section 4.1.2. In most European countries, anti-discrimination laws already cover the self-employed; exceptions include Lithuania and the United Kingdom, where they are not fully covered (European Commission, 2017[60]).

### 4.2.5. Employment protection

When workers have employee status, employment protection legislation usually protects them against unjustified breach of contract obligations on the part of employers, including remedies for unfair dismissal and wage theft. With rare exceptions, these rights are not contractible, in the sense that the employee cannot give them up upon signing the employment contract – see e.g. OECD (2013[61]).

However, these labour law protections do not extend to most contracts for services and/or intermediation contracts (such as terms of conditions for using platform apps), where the scope for contractible provisions is much larger. Yet, in a number of cases, wage and working conditions are set unilaterally by the platform (or the intermediary) or the requester (i.e. individual or company who posts tasks), with no scope for individual workers to negotiate any of the items in the terms of conditions that must be accepted in order to begin or continue working.36 For example, on certain micro-task platforms, requesters can refuse completed tasks without providing a reason, in which case the worker receives no pay – see e.g. Kingsley, Gray and Suri (2015[52]). Even with direct transactions, without platform intermediation, not getting paid for completed tasks is a challenge for many providers of work services – see e.g. Berg et al. (2018[62]). Similarly, terms of conditions of digital intermediation services often establish that the platform can deactivate a worker’s account without providing a justification, sometimes even without previous warning – see e.g. Ross (2015[63]), Kingsley, Gray and Suri (2015[52]), Kaltner (2018[64]), and Marcano (2018[65]).

The absence of adequate, simplified mechanisms of dispute resolution reinforces the asymmetry in the control of the relationship.37 Filing complaints with courts is expensive and long, and usually not attractive for workers in the case of small claims. In the case of micro-task platforms, these barriers are largely prohibitive, since the value of each task corresponds to very small amounts of money.

Building in some kind of simplified dispute resolution system for platform workers would therefore be desirable. For example, platforms could be required to provide a dispute resolution process that places the burden on the client to demonstrate that the work has not been completed (to the required standard) and allows workers a reasonable time to re-do rejected tasks – see e.g. Silberman (2018[66]). Similarly, platforms could be required to communicate swiftly the reason for account deactivation to the worker. The statement of reasons should identify the objective grounds for the deactivation decision, based on the grounds set out in advance in the platform’s terms and conditions, and reasonably refer to the relevant specific circumstances that led to that decision. That statement could also be set to define the limits of the possible legal dispute, in the sense that no additional griefs could be raised by the platform in the case of a lawsuit.38 A proposal for a regulation in this sense is under discussion at the European Parliament (European Commission, 2018[67]) and, in a recent court case involving a ride-hailing service39, the settlement included an agreement that, in the future, workers would receive a hearing before an arbitrator prior to any dismissal (Kovács, 2017[68]).

Simplified dispute resolution systems must be designed in such a way to guarantee impartiality. Standard private out-of-court arbitration may not be a solution, however, because in situations where the arbitrator has frequent interactions with one party, there is potential for abuse and the advantage of neutrality may be lost due to a conflict of interest (US House of Representatives, 2009[69]).40 Such a situation tends to be more frequent in cases opposing individuals to large corporations, not least because the latter are more accustomed with the arbitration process.41 This suggests that an alternative dispute resolution system could be more effective if provided or supervised by an administrative agency. In New York City, for example, the Freelance Isn’t Free Act (FIFA) is intended to protect freelancers from non-payment. Among FIFA’s protections is the requirement that hiring parties provide a written contract to freelancers for any work exceeding USD 800 over a 120-day period. Freelancers can bring a complaint to an administrative agency, which is empowered to issue payment orders as well as civil penalties. Alternatively, simplified dispute resolution systems could rely on social partners. For example, in the temporary work industry, there are a number Ombuds and arbitration schemes managed by social partners that could represent interesting models in the case of tripartite relationships (World Employment Confederation, 2018[70]). In the platform industry, indeed, three German platforms and the German Crowdsourcing Association drafted, in 2015, a Crowdsourcing Code of Conduct and established, in 2017, in conjunction with five other platforms and IG Metall, an “Ombuds office” to enforce the Code of Conduct and resolve disputes between workers and signatory platforms – see Berg et al. (2018[62]) and Chapter 5.42

## copy the linklink copied!4.3. Monopsony power, labour market efficiency and worker vulnerability

Many workers, including many self-employed workers, face an unbalanced power relationship vis-à-vis their employer/client, which makes them vulnerable and potentially in need of the protections that are normally granted to employees only (see the previous section). Unbalanced power relationships tend to emerge because employers/clients often have a higher degree of control over the relationship than workers (discussed in Section 4.1.2 above) and because the latter may have few or no outside options, giving rise to a degree of labour market monopsony (see Box 4.7). In many situations, employers’ power (often referred to as buyer power in the industrial organisation literature) is not compensated by sufficient bargaining power on the side of the workers and may, therefore, lead to lower employment and pay as well as poor working conditions. This is a particular challenge for self-employed workers who are often banned from collective bargaining by antitrust regulation (see Chapter 5).

Box 4.7. What labour market monopsony means in the labour literature

The labour economics literature often defines as labour market monopsony a situation in which employers’ power as buyer of labour services is not compensated by sufficient workers’ bargaining power, and workers have low or no outside options. Strictly speaking, the term “monopsony” refers to the extreme case in which one buyer dominates a specific upstream market and, in order to maximise its profits, can fix input purchases and prices below the level that maximises social welfare. The term “oligopsony” would more correctly refer to cases where few firms dominate buying in a market, and can affect input prices by reducing input purchases in that market. In the case of platforms, which face a multi-sided market, certain studies consider an operational concept of “intermediation power” (Bundesministerium für Wirtschaft und Energie, 2018[71]). Yet, the literature often refers to labour market monopsony to indicate any of these situations with respect to a specific segment of the labour market (Manning, 2003[72]; Bhaskar, Manning and To, 2002[73]). See also Annex 4.A for a more extensive discussion.

Unbalanced power relationships affect not just the self-employed, but workers more generally – including employees (but, in particular, workers who are in precarious employment). There are a number of policy options to tackle an imbalance of power in employment: i) collective voice can be strengthened or extended to workers previously excluded (see Chapter 5); ii) labour market regulation can be used to counteract the ill effects of power imbalances (as discussed in the previous section); and/or iii) these imbalances can be addressed directly, tackling the abuse of monopsony power and its sources. The three options are not mutually exclusive. As the first two options are treated elsewhere, this section briefly focuses only on the third option, with particular reference to non-standard forms of work – Annex 4.A provides a more extensive discussion of both existing evidence on labour market monopsony and related policies to address it.

There is a growing literature showing that, in some OECD countries, a significant fraction of employment is in highly concentrated labour markets – see e.g. Azar et al. (2018[74]), Abel, Tenreyro and Thwaites (2018[75]) and Martins (2018[76]). In specific labour markets of many OECD countries, a large number of studies have also estimated low residual labour supply elasticities – measuring how easily workers switch to a different employer in reaction to wage changes in a specific firm – which is typically considered as evidence of labour market monopsony – see e.g. Manning (2003[72]), Sokolova and Sorensen (2018[77]), and Naidu, Posner and Weyl (forthcoming[78]). While most of this evidence typically refers to employees, there are some studies quantifying the extent to which own-account workers, including platform workers, may be exposed to monopsony power. For example, Dube et al. (forthcoming[79]) provide evidence that workers on Amazon Mechanical Turk can have a residual labour supply elasticity as low as 0.1, while Chevalier et al. (forthcoming[80]) find values comprised between 1 and 2 for Uber drivers. Both estimates are clear evidence of strong buyer power in these labour markets. Although more evidence on the market for platform work is required, when compared with previous results in the literature for employees in many labour markets – see Annex 4.A, these figures suggest that certain platform workers are even more exposed to labour market monopsony than most ordinary employees.

Labour market monopsony (or, more generally, excess employers’ buyer power in the market for labour) can have a negative impact on prices – i.e. pay and benefits – and quantities – overall employment. Recent evidence suggests that high concentration in the labour market depresses wages and increases the wedge between pay and productivity. Labour market monopsony also appears to inefficiently reduce labour demand and employment, particularly at the bottom of the wage distribution, although the evidence is somewhat more indirect (see Annex 4.A).

Labour market monopsony is also a source of concern for business. On the one hand, a lack of competition in the labour market (for example through abuse of non-compete clauses by competitors – see the next section) may imply that innovative companies might be prevented from exploiting new opportunities. On the other hand, insufficient enforcement of competition law may disadvantage those firms that abide by the rules with respect to competitors engaging in illicit behaviours.

### 4.3.1. Enforcing regulation to address labour market monopsony

Excess monopsony power can be addressed by better regulations and more effective enforcement of these regulations. In fact, to date, enforcement agencies have paid somewhat less attention to labour market power and firms’ ability to pay workers less than their marginal productivity than to product market power.

Fighting against labour market collusion is one of the areas where further action is needed – see Annex 4.A for more extensive discussion of existing regulations. In most jurisdictions, competition law forbids collusion among buyers of intermediate goods or services, including labour services – see e.g. Blair and Wang (2017[81]). Illicit collusion occurs, for example, when companies competing for the same type of workers agree on refraining from hiring those employed by the others (so-called “no-poaching agreements”) or, except in the framework of collective bargaining (see Chapter 5), when firms competing in the same labour market agree to apply a common compensation policy to employees or independent contractors (wage collusion).

General statistics on collusion are difficult to collect, since figures on those illicit behaviours that escape investigation are typically not available. Anecdotal evidence suggests, however, that the phenomenon is far from negligible – see e.g. Krotoski et al. (2018[82]). Statistics on non-poaching covenants exist for franchising agreements, where these covenants are not necessarily unlawful (see Annex 4.A). Krueger and Ashenfelter (2018[83]) estimate that more than 50% of major franchise companies in the United States use no-poaching clauses in their franchising agreements.

Providing explicit guidance about labour market collusion is crucial to guide and set priorities for enforcement agencies. For example, US antitrust authorities have issued guidelines that explicitly refer to collusion in the labour market, present clear examples of illicit behaviours, and underline the importance of fighting them for their effects in the labour market (US Department of Justice and Federal Trade Commission, 2016[84]). Whistleblower protection is also needed for effective enforcement, since collusion is often discovered based on information provided by insiders (Dyck, Morse and Zingales, 2010[85]; Yeoh, 2014[86]). Specific protection is particularly essential in the case of independent contractors who usually are not covered by dismissal legislation provisions concerning whistleblowing – see e.g. OECD (2014[8]). Last, but not least, enforcement agencies should also be able to provide adequate sanctions for collusive behaviours.

Potentially colluding companies, however, could avoid unlawful labour market collusion by simply merging, thereby raising their buyer power in the labour market – see e.g. Marinescu and Hovenkamp (forthcoming[87]). More generally, if merging firms would together form the dominant buyer in the labour market, the merged entity would likely use its buyer power to reduce quantities and prices in that market, similarly to what non-merging colluding companies would do. Yet, one difficulty in assessing the impact of mergers on buyer power in the labour market has to do with the shortage of specific tools to analyse labour competition and, in particular, the difficulty of identifying the relevant market. Another difficulty is the evaluation of merger effects when merging firms are not direct competitors in downstream product markets.43 This is an area in which more research is needed and more investment in developing adequate tools by governments and enforcement authorities would be welcome.

Another area of possible policy action concerns non-compete covenants, which are contract clauses preventing workers from working for a competitor after they separate from the employer. In most countries, non-compete agreements are lawful and justified by the need to protect trade secrets and specific investment in the employment relationship by the employer (such as certain types of training and investment in knowledge). Yet, recent evidence suggests that employers may use these clauses in order to limit effectively the outside options of their workers. A number of cases, for example, have concerned low-skilled workers involved in production of standardised products, such as sandwich making, and with little access to company tacit knowledge. Starr, Prescott and Bishara (2018[88]) estimate that almost 30 million US workers are currently covered by non-compete agreements, and that many of these agreements are imposed broadly on workers with less than a bachelor’s degree, often with low income or no access to trade secrets.

A particular type of non-compete covenant is represented by the restrictions imposed on certain platform workers to continue a direct relationship with their client off the platform without paying a disproportionately high fee (Berg et al., 2018[62]). This constraint is particularly binding in the case of unilateral changes to the terms of service provided by the platform. In fact, the specific relationship that is sometimes created between a given worker and a given client on certain platforms implies a high opportunity cost for workers if they cannot quit the platform without losing their clients, which often forces them to accept unfavourable changes in the terms of conditions.44

In order to fight against non-compete agreements in situations where they are most likely used only to reduce workers’ outside options, governments could consider establishing a rebuttable presumption of abusive use (or even banning them), in the case of certain type of positions, pay levels or skill requirements, for which a clear justification, such as the protection of trade secrets, seems implausible. Where non-compete clauses are enforceable, governments could also consider banning “blue-pencilling” by courts – i.e. situations where courts can redraft unreasonable covenants in order to make them enforceable. In fact, this practice creates incentives to draft unenforceable, extensive clauses only to discourage uninformed workers from seeking alternative job offers. Last, but not least, since workers who are victim of abuses rarely initiate private damage actions, enforcement agencies, including labour inspectorates, should take a leading role and have the possibility of imposing sanctions or taking the case to courts for imposing them (see Annex 4.A for more discussion of current practices and issues in OECD countries).

Finally, one of the reasons why workers often have low bargaining power and few outside options is that they have much less information than employers. Workers often have only a vague idea of their rights, in particular if they are employed with a non-standard contract. For example, when signing a contract (or accepting terms and conditions) workers may not fully understand the obligations they are committing to (or the rights they are waiving). Regulation could therefore make sure that all contracting parties are fully aware of their rights and responsibilities. For example, certain scholars have suggested that governments should invest in a dedicated service for independent contractors, and in particular platform workers, offering general advice and counsel about their employment rights (Balaram, Warden and Wallace-Stephens, 2017[89]).

Lack of pay transparency may also increase monopsony power. Shortage of information about alternative opportunities reduces workers’ ability to change jobs or leverage outside opportunities to negotiate for higher pay (and better working conditions) (Harris, 2018[90]). Digital technologies have the potential to improve this type of information asymmetry, as workers can access a large number of job offers and compare them, thereby reducing search costs (see below). Yet, in many platforms, workers have few tools to search for available alternatives and have to spend much time searching for them (Kingsley, Gray and Suri, 2015[52]; Berg et al., 2018[62]). To improve pay transparency in the platform economy, employers and platforms could be required to publish information about the average pay per task, as well as on the average time taken to complete a task (which would help workers make more informed decisions about which tasks to accept).

One specific problem related to digital intermediation concerns the information which platforms collect on workers. For example, most platforms have a rating system for workers, which evaluates their performance history and which is supposed to improve service quality for the requester. However, the ratings algorithms often lack transparency (Rosenblat and Stark, 2016[91]). This informational asymmetry allows requesters or platforms to exclude certain workers from certain tasks, based on their ratings, while workers are unable to identify and refuse tasks proposed by bad requesters, who typically do not pay regularly or are inclined to provide negative assessments (Kingsley, Gray and Suri, 2015[52]). To address these issues, governments could consider regulating the ratings system by imposing rating symmetry (so called “five-for-five” policies) and algorithm transparency to platforms.

### 4.3.2. Other interventions to reduce frictions

Labour market monopsony tends to emerge in situations where there are few, large firms, and where frictions in the labour market, preventing workers from easily switching jobs in response to changes in wages or working conditions, are considerable – see e.g. Manning (2011[92]). All types of non-regulatory interventions to reduce frictions in the labour market are likely to contribute to lessen monopsony power in that market. A comprehensive strategy of simultaneously tackling all barriers to potential job mobility in the labour market should be developed (OECD, 2018[93]).

Interventions favouring geographical mobility play a crucial role. Existing evidence suggests that labour markets tend to be more concentrated in rural areas and/or areas where economic activity is more dispersed (Azar et al., 2018[74]; Rinz, 2018[94]). For example, housing policies could promote geographical mobility of workers to help people move to the regions where the best jobs are available. Similarly, occupational licensing should be used judiciously and standards should be harmonised across regions (and countries, where applicable) as much as possible, insofar as in some countries licensing has acted as a barrier to mobility, without clear benefits in terms of better service quality, consumer health or safety – see OECD (2018[93]) for further discussion. The same arguments apply to national (and international) recognition of competences and acquired skills – see Chapter 6.

Technological developments are improving the efficiency of the matching process. In many OECD countries, unemployment coexists with firms complaining about not being able to find suitably skilled workers to fill vacancies. Digital platforms can help employers find workers for tasks that their existing employees cannot perform, thereby enlarging the labour market by expanding opportunities in it – see Chapter 2. Similarly, digital intermediaries, such as commercial job boards, can provide access to a large number of vacancies and worker profiles, significantly reducing search costs for both workers and employers. The possibility of working from a distance made possible by digital technologies also allows workers in rural areas accessing jobs and tasks that before were precluded to them. However, in platform-mediated interactions, there is a potential for adverse selection and asymmetric concealment of information, which may require the development of institutions to certify the information provided by users (Autor, 2009[95]). Moreover, in certain cases, there is a risk of a race to the bottom because of competition from world regions with much lower labour standards and pay level in real terms (see Section 4.4 below)

Digital technology is also transforming the way the Public Employment Service (PES) and other placement providers operate, making it easier to exploit information about vacancies and jobseekers, thereby improving the matching process and reducing market frictions. By automating a number of tasks, digitalisation also allows the PES to concentrate resources on activities requiring personal interactions and more discretionary task actions. There are limits to what digitalisation can achieve in this field, however. While benefit applications and vacancy registration are easy to digitalise, it may be more difficult to do so for other services, such as personalised counselling and training. In addition, safeguards must be introduced in new, heavily digitalised systems to avoid creating a digital divide by handicapping more disadvantaged jobseekers (OECD, 2015[96]).

Promoting job mobility also requires making social protection more portable and less linked to a specific job or employer (see Chapter 7). In that respect, relying on voluntary provision of social protection entitlements by employers or intermediaries such as platforms (see e.g. Section 4.2 above) might not be a wise idea in the long-run. In fact, the lack of portability of employer-provided social protection plans may reduce mobility and therefore heighten the monopsony power of employers and intermediaries, exerting downward pressure on pay.

Specific features of certain platforms can also create obstacles to job mobility. For example, payment or remuneration in a moneyless form like bitcoins and vouchers are relatively common on certain platforms (Kingsley, Gray and Suri, 2015[52]) and tie the worker to the platform. Similarly, personal work histories, such as personal ratings, are usually lost upon changing platform (Berg et al., 2018[62]). Given that platforms de facto favour workers with good ratings, the loss of individual ratings represents a strong barrier to worker mobility, and may limit competition for workers across platforms. Governments could therefore consider further interventions to enhance worker mobility across platforms such as regulating moneyless payments and facilitating the portability of personal ratings.

## copy the linklink copied!4.4. International competition

Labour law tends to apply to the national or regional labour market, and it faces serious limitations when work is performed across national borders. Yet, with the rise of the platform economy, an increasing number of workers provide services internationally. As the ILO puts it, “digital labour platforms provide new sources of income to many workers in different parts of the world, yet the dispersed nature of the work across international jurisdictions makes it difficult to monitor compliance with applicable labour laws” (ILO, 2019, p. 44[1]). Clients and employers can be based in one country, the platform in another, and the workers in yet another. In those cases, it is not obvious what law should apply, if any, or how it should be enforced. In addition, when countries start regulating the platform economy, there is a risk that they will do so in very different ways which can not only cause difficulties for the platforms themselves, but could also result in a race to the bottom among countries trying to attract work through the platform economy (Cherry and De Stefano, 2018[21]). In such a context, it becomes very difficult to regulate working conditions and international cooperation becomes indispensable.

In the context of the European Union (EU), there is some precedent in terms of regulating cross-border work. While the EU legislation allows for choice, the default position is to apply the law of the country in which the employee is carrying out the work.45 Furthermore, recognising the imbalance in power relations between employers and employees, the choice of national law regime may not “have the result of depriving the employee of the protection afforded to him by provisions that cannot be derogated from by agreement under the law that, in absence of choice, would have been applicable” – Rome Convention of 1980, Rome I Article 6(1). In short, the choice of law allows some flexibility, but not at the expense of worker rights.

In most cases, however, platform workers are likely to have a contract for services (and not an employment contract) – in which case the above regulations do not apply. In the case of a contract for services, European law – Rome I Article 4(1) – again gives as the prioriy “the law of the country where the service provider has his habitual residence” or, failing that, “the law of the country where the party required to effect the characteristic performance of the contract has his habitual residence” – Article 4(2). However this will do little to protect the working conditions of platform workers (unless national regulations have specific provisions in place).

Given the limitations of the above regulations as far as platform workers are concerned, some have argued for further regulation at the EU level in the form of a “platform work directive” (Risak, 2018[97]). This could set out a list of rights for platform workers, regardless of their employment status. However, such a directive would still only apply to workers based in the EU and would not solve the problem of a possible race to the bottom in working conditions with countries outside the EU. It is also important to point out that for many genuinely self-employed and highly skilled platform workers such regulation would be less necessary, as they are typically less vulnerable and their activities could genuinely be seen as international trade in services between undertakings.

For the more vulnerable self-employed platform workers in the grey zone, voluntary initiatives and/or self-regulation might offer a solution (at least temporarily). Some initiatives have already sprung up. For example, the Dynamo Guidelines for Academic Requesters aim to encourage academics using Amazon Mechanical Turk to be good employers and pay fair wages – but there is currently no way of enforcing these guidelines (Salehi et al., 2015[98]). In Germany, the Crowdsourcing Code of Conduct has been signed by eight platforms, and there is now an Ombuds Office for workers to turn to in case they think they have been treated unfairly by one of the platforms that signed up to the Code of Conduct – see Silberman (2018[99]) and Section 4.2 above. For Mechanical Turk, there are also plans to set up a voluntary pledge for requesters to sign, and which would appear as a badge next to their name for workers to see. At the international level, the G20 countries have committed to “promoting decent work in the platform economy” (Labour 20, 2018[100]). While such efforts are at a very early stage, it is possible to envisage a set of guidelines for countries and platforms to sign up to and which might have a similar impact as the OECD Guidelines for Multinational Enterprises, which have promoted responsible business conduct in global supply chains.46

## copy the linklink copied!4.5. Concluding remarks

This chapter has examined how labour market regulation could be extended and adapted to adequately protect workers in a changing world of work, and to ensure that firms that respect these regulations are not at a disadvantage to their competitors. Employment status acts as a gateway to various rights and protections. Ensuring the correct classification of workers (and tackling misclassification) is therefore a key first step to ensure that those in new forms of work have access to labour and social protection, as well as to collective bargaining and lifelong learning. For some workers, however, there is genuine ambiguity about employment status as they find themselves somewhere in the “grey zone” between dependent and self-employment. In addition to urging policy makers to try and reduce the size of this grey zone as much as possible, the chapter discusses the rationale and policy options for extending certain labour rights and protections beyond standard employees. Tackling power imbalances between employers/clients and workers also involves addressing abuses of monopsony power, e.g. by fighting labour market collusion by employers, limiting the scope of non-compete covenants and redressing inequalities in the information available to employers and workers. Nevertheless, these issues could also be addressed, and the position of many workers improved, through social dialogue and collective bargaining (Chapter 5), skill policies (Chapter 6) and social protection measures (Chapter 7).

While this chapter has discussed policy options to extend labour rights and protections beyond standard employees, it is important to underline that a balanced approach is needed. New forms of work often emerge in response to the real needs of both employers and workers. Diversity (and continuous innovation) in employment contracts gives both employers and workers the necessary flexibility to find arrangements which are in the best interests of both. The objective of policy should therefore be to accommodate these changes without abandoning the aim of protecting the weakest and ensuring fundamental rights for all.

Box 4.8. Policy directions

Governments should ensure that all workers in the labour market have access to an adequate set of rights and protections, regardless of their employment status or contract type, and guarantee a level playing field among firms by preventing some from gaining a competitive advantage by avoiding their obligations and responsibilities.

In the area of labour market regulation, countries should:

• Tackle false self-employment by:

• Ensuring that employers and workers are aware of, and understand, existing regulations;

• Making it easier and less costly for workers to challenge their employment status;

• Strengthening the penalties for firms misclassifying workers;

• Strengthening the capacity of labour inspectorates to monitor and detect breaches;

• Reducing incentives for firms and workers to misclassify employment relationships as self-employment in order to avoid/reduce taxes and regulations.

• Reduce the size of the “grey zone” between self- and dependent employment by revising, updating and/or harmonising definitions of what it means to be an employee and/or a self-employed person in order to make these definitions as clear as possible and reduce uncertainty for both workers and employers.

• Extend rights and protections to those workers left in the “grey zone” (i.e. where genuine ambiguity in employment status exists) by a combination of:

• Identifying and targeting specific groups of workers to which certain labour rights and protections could be extended;

• Deciding which labour rights and protections to (at least partially) extend (e.g.: fair pay, working time protections, occupational health and safety, anti-discrimination and employment protection) and whether and how they should be adapted;

• Where necessary, clarifying and/or assigning employer duties and responsibilities in the case of triangular employment relationships (including platform work), which may require spreading such responsibilities across multiple legal entities.

• At the international level, build on the recent G20 commitment to promote decent work in the platform economy and consider ways of improving the working conditions of workers with little say over their remuneration and working conditions who provide services globally – including best practice principles or guidelines, which countries and/or platforms could sign up to.

Addressing power imbalances between employers/clients and workers also requires enhancing collective bargaining and social dialogue (see Chapter 5) and tackling labour market monopsony. Options that could be considered to fight against abuses of monopsony power include:

• Fighting labour market collusion, for example by providing explicit guidance on illicit behaviours, setting priorities for enforcement agencies and ensuring adequate whistleblower protection;

• Limiting the scope of non-compete covenants, including in contracts for services – particularly for certain types of jobs, pay levels or skill requirements, where they are most likely to be used to reduce competition in the labour market;

• Reducing the incentives for broad or unlawful non-compete agreements, by banning court redrafting of unreasonable covenants to make them enforceable, and by appropriately sanctioning the abuse of illicit clauses;

• Favouring the development of new tools and instruments to better analyse the effects of mergers and anti-competitive conduct in the labour market;

• Redressing inequalities in the information available to employers and workers by ensuring that workers are fully aware of their rights and responsibilities, improving pay transparency in the labour market and enhancing symmetry of treatment of workers and requesters on online platforms, including as regards mutual evaluations.

## References

[75] Abel, W., S. Tenreyro and G. Thwaites (2018), “Monopsony in the UK”, CEPR Discussion Paper, No. 13265, CEPR, London.

[42] Åberg, A. (2016), Working in a Cross-Border Situation - A Study on the Concepts of Employment and Self-Employment, Lund University, https://lup.lub.lu.se/student-papers/search/publication/8876224 (accessed on 13 December 2018).

[161] Adams, A. and J. Prassl (2018), “Zero-Hours Work in the United Kingdom”, Conditions of Work and Employment Series, No. 110, International Labour Office, Geneva, http://www.ilo.org/wcmsp5/groups/public/---ed_protect/---protrav/---travail/documents/publication/wcms_624965.pdf (accessed on 4 June 2018).

[5] Adam, S., H. Miller and T. Pope (2017), “Tax, legal form and the gig economy”, in The Institue for Fiscal Studies (IFS) Green Budget, The Institute for Fiscal Studies (IFS), https://www.ifs.org.uk/uploads/publications/comms/R124_Green%20Budget_7.%20Tax%2C%20legal%20form%20and%20gig%20economy.pdf (accessed on 8 June 2018).

[143] Adler, S. (2006), General Report on non-competition clauses (covenants not to compete) in labour contracts, https://www.ilo.org/wcmsp5/groups/public/---ed_dialogue/---dialogue/documents/meetingdocument/wcms_159933.pdf (accessed on 20 November 2018).

[25] Aizer, A. and J. Doyle (2015), “Juvenile Incarceration, Human Capital, and Future Crime: Evidence from Randomly Assigned Judges *”, The Quarterly Journal of Economics, Vol. 130/2, pp. 759-803, http://dx.doi.org/10.1093/qje/qjv003.

[108] Ashenfelter, O., H. Farber and M. Ransom (2010), “Modern Models of Monopsony in Labor Markets: A Brief Survey”, IZA Discussion Paper, No. 4915, IZA, Bonn, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1599013 (accessed on 18 November 2018).

[173] Association of Corporate Counsel (2018), Multi-Country Survey on Covenants Not to Compete, Association of Corporate Counsel, Washington, D.C., https://www.gtlaw.com/en/-/media/files/insights/alerts/2018/3/gtnoncompeteeuroinfopak.pdf.

[95] Autor, D. (ed.) (2009), Studies of Labor Market Intermediation, The University of Chicago Press, Chicago, Ill.

[115] Autor, D. et al. (2017), “The Fall of the Labor Share and the Rise of Superstar Firms”, NBER Working Papers, No. 23396, NBER, Cambridge, MA, http://dx.doi.org/10.3386/w23396.

[28] Autor, D. et al. (2017), “Disability Benefits, Consumption Insurance, and Household Labor Supply”, NBER Working Papers, No. 23466, NBER, Cambridge, MA, https://www.nber.org/papers/w23466.pdf (accessed on 22 October 2018).

[105] Azar, J., I. Marinescu and M. Steinbaum (2017), “Labor Market Concentration”, NBER Working Papers, No. 24147, NBER, Cambridge, MA, http://dx.doi.org/10.3386/w24147.

[74] Azar, J. et al. (2018), “Concentration in US Labor Markets: Evidence From Online Vacancy Data”, NBER Working Papers, No. 24395, NBER, Cambridge, MA, http://dx.doi.org/10.3386/w24395.

[89] Balaram, B., J. Warden and F. Wallace-Stephens (2017), Good Gigs - A fairer future for the UK’s gig economy, RSA, https://www.thersa.org/globalassets/pdfs/reports/rsa_good-gigs-fairer-gig-economy-report.pdf (accessed on 31 May 2018).

[114] Barkai, S. (2017), “Declining labor and capital shares | London Business School”, London Business School, https://www.london.edu/faculty-and-research/academic-research/d/declining-labor-and-capital-shares (accessed on 23 October 2018).

[55] Barrios, J., Y. Hochberg and H. Yi (2018), “The Cost of Convenience: Ridesharing and Traffic Fatalities”, http://dx.doi.org/10.2139/ssrn.3271975.

[44] BEIS (2017), Good work: the Taylor review of modern working practices, Department for Business, Energy & Industrial Strategy, https://www.gov.uk/government/publications/good-work-the-taylor-review-of-modern-working-practices (accessed on 4 June 2018).

[116] Bell, B., P. Bukowski and S. Machin (2018), “Rent Sharing and Inclusive Growth”, IZA Discussion Papers, No. 12060, IZA, http://www.iza.org.

[111] Benmelech, E., N. Bergman and H. Kim (2018), “Strong Employers and Weak Employees: How Does Employer Concentration Affect Wages?”, NBER Working Paper, No. 24307, NBER, Cambridge, MA, http://dx.doi.org/10.3386/w24307.

[62] Berg, J. et al. (2018), Digital labour platforms and the future of work: Towards decent work in the online world, International Labour Organization, https://www.ilo.org/global/publications/books/WCMS_645337/lang--en/index.htm (accessed on 6 November 2018).

[73] Bhaskar, V., A. Manning and T. To (2002), “Oligopsony and Monopsonistic Competition in Labor Markets”, The Journal of Economic Perspectives, Vol. 16/2, pp. 155-174, http://www.jstor.org/stable/2696501.

[109] Blair, R. and J. Harrison (2010), Monopsony in Law and Economics, Cambridge University Press, Cambridge, http://dx.doi.org/10.1017/cbo9780511778766.

[81] Blair, R. and W. Wang (2017), “Buyer cartels and private enforcement of antitrust policy”, Managerial and Decision Economics, Vol. 38/8, pp. 1185-1193, http://dx.doi.org/10.1002/mde.2857.

[149] Bork, R. (1993), The Antitrust Paradox (second edition), Free Press.

[30] Breda, T. et al. (2017), “Prud’hommes : peut-on expliquer la disparité des décisions ?”, Les notes de l’IPP 29, https://www.ipp.eu/wp-content/uploads/2017/11/n29-notesIPP-nov2017.pdf (accessed on 24 October 2018).

[3] Brumm, F. (2016), Making Gigs Work The new economy in context, University of Illinois - Urbana Champaign, https://blogs.illinois.edu/files/6232/430823/98170.pdf (accessed on 1 June 2018).

[71] Bundesministerium für Wirtschaft und Energie (2018), Modernisierung der Missbrauchsaufsicht für marktmächtige Unternehmen, Bundesministerium für Wirtschaft und Energie, Berlin, https://www.bmwi.de/Redaktion/DE/Publikationen/Wirtschaft/modernisierung-der-missbrauchsaufsicht-fuer-marktmaechtige-unternehmen.html (accessed on 14 December 2018).

[139] Cahuc, P., F. Postel-Vinay and J. Robin (2006), “Wage Bargaining with On-the-Job Search: Theory and Evidence”, Econometrica, Vol. 74/2, pp. 323-364, http://dx.doi.org/10.1111/j.1468-0262.2006.00665.x.

[140] Caldwell, S. and O. Danieli (2018), “Outside Options in the Labor Market”, Mimeo, Harvard University, https://scholar.harvard.edu/danieli/publications/outside-options-labor-market (accessed on 10 December 2018).

[170] Calligaris, S., C. Criscuolo and L. Marcolin (2018), “Mark-ups in the digital era”, OECD Science, Technology and Industry Working Papers, No. 2018/10, OECD Publishing, Paris, https://dx.doi.org/10.1787/4efe2d25-en.

[124] Capobianco, A. and P. Gonzaga (2017), “Algorithms and competition: Friends or foes?”, CPI Antitrust Chronicle 08/2017, https://www.competitionpolicyinternational.com/wp-content/uploads/2017/08/CPI-Capobianco-Gonzaga.pdf (accessed on 9 December 2018).

[43] Card, D. and A. Krueger (2015), Myth and measurement : the new economics of the minimum wage, Princeton University Press, https://press.princeton.edu/titles/10738.html (accessed on 5 November 2018).

[146] Carlton, D., M. Israel and M. Coleman (2014), “Buyer Power in Merger Review”, in Blair, R. and D. Sokol (eds.), The Oxford Handbook of International Antitrust Economics, Oxford University Press, Oxford.

[171] Carstensen, P. (2012), “Buyer Power and the Horizontal Merger Guidelines: Minor Progress on an Important Issue”, University of Pennsylvania Journal of Business Law, Vol. 14/3, https://scholarship.law.upenn.edu/jbl/vol14/iss3/4 (accessed on 16 November 2018).

[33] Cherry, M. and A. Aloisi (2017), “Dependent Contractors’ in the Gig Economy: A Comparative Approach”, American University Law Review, Vol. 66/3, http://digitalcommons.wcl.american.edu/aulr (accessed on 14 December 2018).

[21] Cherry, M. and V. De Stefano (2018), “A Labour Law Perspective on Conflict of Laws”, in Pretelli, I. (ed.), Conflict of Laws in the Maze of Digital Platforms, Schultess, Zurich.

[80] Chevalier, J. et al. (forthcoming), “The Value of Flexible Work: Evidence from Uber Drivers”, Journal of Political Economy, http://dx.doi.org/10.1086/702171.

[155] Chopra, R. (2018), “Keynote of Federal Trade Commissioner Rohit Chopra”, in Conference on Monopoly, Monopsony, and the Labor Market: Diminishing Worker Power in an Era of Market Concentration, The Economic Policy Institute, Washington, D.C., https://openmarketsinstitute.org/events/monopoly-monopsony-labor-market-diminishing-worker-power-era-market-concentration/ (accessed on 14 December 2018).

[31] Countouris, N. et al. (2016), Report on temporary employment agencies and temporary agency work, International Labour Office, Geneva, http://www.ilo.org/publns (accessed on 28 November 2018).

[45] D’Arcy, C. (2017), The minimum required? Minimum wages and the self-employed, Resolution Foundation, https://www.resolutionfoundation.org/app/uploads/2017/07/The-minimum-required.pdf (accessed on 5 November 2018).

[38] Däubler, W. (2016), “Challenges to Labour Law”, Pravo. Zhurnal Vysshey shkoly ekonomiki 1, pp. 189-203, http://dx.doi.org/10.17323/2072-8166.2016.1.201.215.

[163] Davidov, G. (2002), “The Three Axes of Employment Relationships: A Characterization of Workers in Need of Protection”, University of Toronto Law Journal, Vol. 52, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=555998 (accessed on 21 November 2018).

[6] Davidov, G., M. Freedland and N. Countouris (2015), “The Subjects of Labor Law: ’Employees’ and Other Workers”, in Finkin, M. and G. Mundlak (eds.), Research Handbook in Comparative Labor Law, Edward Elgar, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2561752 (accessed on 2 November 2018).

[169] Davies, A. (2009), “Sensible Thinking About Sham Transactions: Protectacoat Firthglow Ltd v Szilagyi * [2009] EWCA Civ 98; [2009] IRLR 365”, Industrial Law Journal, Vol. 38/3, pp. 318-328, http://dx.doi.org/10.1093/indlaw/dwp017.

[113] De Loecker, J. and J. Eeckhout (2018), “The Rise of Market Power and the Macroeconomic Implications”, NBER Working Paper, Revised 2018 version resubmitted to the Quarterly Journal of Economics, No. 23687, NBER, Cambridge, MA, http://www.janeeckhout.com/wp-content/uploads/RMP.pdf (accessed on 12 December 2018).

[47] Deutscher Bundestag (2016), Arbeit 4.0 – Arbeitswelt von morgen gestalten, http://dipbt.bundestag.de/dip21/btd/18/102/1810254.pdf (accessed on 5 November 2018).

[144] Dobson, P. and R. Inderst (2007), “Differential Buyer Power and the Waterbed Effect: Do Strong Buyers Benefit or Harm Consumers?”, European Competition Law Review, Vol. 28/7, pp. 393-400, https://ueaeprints.uea.ac.uk/26139/ (accessed on 21 November 2018).

[79] Dube, A. et al. (forthcoming), “Monopsony in Online Labor Markets”, American Economic Review: Insights, http://www.aeaweb.org/articles?id=10.1257/aeri.20180150&&from=f (accessed on 15 January 2019).

[85] Dyck, A., A. Morse and L. Zingales (2010), “Who Blows the Whistle on Corporate Fraud?”, The Journal of Finance, Vol. 65/6, pp. 2213-2253, http://dx.doi.org/10.1111/j.1540-6261.2010.01614.x.

[56] Edelman, B., M. Luca and D. Svirsky (2017), “Racial Discrimination in the Sharing Economy: Evidence from a Field Experiment”, American Economic Journal: Applied Economics, Vol. 9/2, pp. 1-22, http://dx.doi.org/10.1257/app.20160213.

[23] Employment and Social Development Canada (2006), “Determining the Employer/Employee Relationship - Canada Labour Code”, Interpretations, Policies and Guidelines, No. IPG 069, Government of Canada, Ottawa, https://www.canada.ca/content/dam/canada/employment-social-development/migration/documents/assets/portfolio/docs/fr/edsc/lois_reglements/travail/interpretations_politiques/index/069.pdf (accessed on 23 October 2018).

[67] European Commission (2018), Proposal for a regulation of the European Parliament and of the Council on promoting fairness and transparency for business users of online intermediation services, 2018/0112(COD), European Commission, Brussels, https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52018PC0238&from=EN (accessed on 5 December 2018).

[60] European Commission (2017), A comparative analysis of non-discrimination law in Europe 2017, European Commission, Brussels, http://dx.doi.org/10.2838/52129.

[182] European Commission (2011), Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements (2011/C 11/01), European Commission, Brussels.

[147] European Commission (2004), Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings (2004/C 31/03), European Commission, Brussels.

[26] Fischman, J. (2014), “Measuring Inconsistency, Indeterminacy, and Error in Adjudication”, American Law and Economics Review, Vol. 16/1, pp. 40-85, http://dx.doi.org/10.1093/aler/aht011.

[27] Fischman, J. (2011), “Estimating Preferences of Circuit Judges: A Model of Consensus Voting”, The Journal of Law and Economics, Vol. 54/4, pp. 781-809, http://dx.doi.org/10.1086/661512.

[164] Freedland, M. and N. Kountouris (2011), The Legal Construction of Personal Work Relations, Oxford University Press, http://dx.doi.org/10.1093/acprof:oso/9780199551750.001.0001.

[58] Galperin, H., G. Cruces and C. Greppi (2017), “Gender Interactions in Wage Bargaining: Evidence from an Online Field Experiment”, SSRN, http://dx.doi.org/10.2139/ssrn.3056508.

[57] Galperin, H. and C. Greppi (2017), “Geographical Discrimination in Digital Labor Platforms”, https://annenberg.usc.edu/sites/default/files/2017/11/27/Geographical%20Discrimination%20Galperin%20%26%20Greppi.pdf (accessed on 1 June 2018).

[4] GAO (2017), Employment Taxes - Timely Use of National Research Program Results Would Help IRS Improve Compliance and Tax Gap Estimates, United States Government Accountability Office, https://www.gao.gov/assets/690/684162.pdf (accessed on 7 January 2019).

[133] Garmaise, M. (2009), “Ties that Truly Bind: Noncompetition Agreements, Executive Compensation, and Firm Investment”, Journal of Law, Economics, and Organization, Vol. 27/2, pp. 376-425, http://dx.doi.org/10.1093/jleo/ewp033.

[19] Golubeva, M. (2018), Tax reforms not enough for changing the game?, http://ec.europa.eu/social/main.jsp?langId=en&catId=89&newsId=9040&furtherNews=yes (accessed on 2 November 2018).

[174] Gomulkiewicz, R. (2015), “Leaky Covenants-Not-to-Compete as the Legal Infrastructure for Innovation”, UC Davis Law Review, Vol. 49/1, pp. 251-304, https://lawreview.law.ucdavis.edu/issues/49/1/Articles/49-1_Gomulkiewicz.pdf (accessed on 4 December 2018).

[175] Green, T. (2015), “The shifting landscape of restrictive covenants in Oklahoma”, Oklahoma City University Law Review, Vol. 49/2, pp. 449-480, http://wallethub.com/edu/best-cities-to-start-a-business/2281/#complete-rankings (accessed on 5 December 2018).

[117] Gurkaynak, G., A. Guner and C. Ozkanli (2013), “Competition Law Issues in the Human Resources Field”, Journal of European Competition Law & Practice, Vol. 4/3, pp. 201-214, http://dx.doi.org/10.1093/jeclap/lpt017.

[121] Haar, B. (2018), “Regulation through litigation - Collective redress in need of a new balance between individual rights and regulatory objectives in Europe”, Theoretical Inquiries in Law, Vol. 19/2, http://dx.doi.org/10.1515/til-2018-0007.

[162] Hara, K. et al. (2017), “A Data-Driven Analysis of Workers’ Earnings on Amazon Mechanical Turk”, Cornell University, http://arxiv.org/abs/1712.05796 (accessed on 1 June 2018).

[90] Harris, B. (2018), Information Is Power Fostering Labor Market Competition through Transparent Wages, The Hamilton Project, https://www.brookings.edu/wp-content/uploads/2018/02/es_2272018_information_is_power_harris_pp.pdf (accessed on 6 November 2018).

[156] Harrison, J. (2012), “Complications in the Antitrust Response to Monopsony”, in Sokol, D. and I. Lianos (eds.), The Global Limits of Competition Law, Stanford University Press, http://dx.doi.org/10.11126/stanford/9780804774901.003.0005.

[41] Harris, S. and A. Krueger (2015), “A Proposal for Modernizing Labor Laws for Twenty-First-Century Work: The Independent Worker’”, Discussion Papers, No. 2015-10, The Hamilton Project, http://www.hamiltonproject.org/assets/files/modernizing_labor_laws_for_twenty_first_century_work_krueger_harris.pdf (accessed on 31 May 2018).

[160] Herrera Anchustegui, I. and J. Nowag (2017), “How the Uber & Lyft Case Provides an Impetus to Re-Examine Buyer Power in the World of Big Data and Algorithms”, Lund University Legal Research Paper Series Lund Comp Working Paper, No. 01/2017, Lund University, http://dx.doi.org/10.2139/ssrn.2998688.

[112] Hershbein, B. and C. Macaluso (2018), “Labor Market Concentration and the Demand for Skills”, IDSC of IZA Workshop: Matching Workers and Jobs Online, 21-22 September 2018, Bonn, http://conference.iza.org/conference_files/DATA_2018/macaluso_c26795.pdf (accessed on 8 January 2019).

[107] Hershbein, B., C. Macaluso and C. Yeh (2018), “Concentration in U.S. local labor markets: evidence from vacancy and employment data”, mimeo, University of Illinois, Urbana-Champaign, https://sites.google.com/site/clamacaluso/research.

[148] Hovenkamp, H. (forthcoming), “Whatever Did Happen to the Antitrust Movement?”, Notre Dame Law Review.

[29] Ichino, A. and P. Pinotti (2012), “La roulette russa dell’Articolo 18”, lavoce.info, https://www.lavoce.info/archives/27539/la-roulette-russa-dellarticolo-18/ (accessed on 24 October 2018).

[181] Illinois Attorney General (2018), Attorney General Madigan reaches settlement with WeWork to end use of overly broad non-competes, Press Releases of the Illinois Attorney General, http://illinoisattorneygeneral.gov/pressroom/2018_09/20180918.html (accessed on 18 September 2018).

[1] ILO (2019), Work for a brighter future – Global Commission on the Future of Work, International Labour Organization, Geneva, https://www.ilo.org/wcmsp5/groups/public/---dgreports/---cabinet/documents/publication/wcms_662410.pdf (accessed on 21 February 2019).

[36] ILO (2005), The employment relationship. International Labour Conference, 95th Session, 2006, Report V (1)., ILO, https://www.ilo.org/global/publications/ilo-bookstore/order-online/books/WCMS_PUBL_9221166112_EN/lang--en/index.htm (accessed on 7 December 2018).

[145] Inderst, R. and G. Shaffer (2008), “Buyer Power in Merger Control”, in American Bar Association (ed.), Issues in Competition Law, American Bar Association, Chicago, Ill.

[167] Internal Revenue Service (2012), “Voluntary Classification Settlement Program, Announcement 2012-45”, Internal Revenue Bulletin 2012-51, https://www.irs.gov/irb/2012-51_IRB#ANN-2012-45 (accessed on 29 November 2018).

[10] Internal Revenue Service (2005), Independent Contractor or Employee?, Internal Revenue Service, Washington, D.C., https://www.irs.gov/pub/irs-pdf/p1779.pdf (accessed on 29 November 2018).

[158] Irani, L. and S. Silberman (2013), Turkopticon: Interrupting Worker Invisibility in Amazon Mechanical Turk, ACM, https://cloudfront.escholarship.org/dist/prd/content/qt10c125z3/qt10c125z3.pdf (accessed on 6 November 2018).

[153] Jacobson, J. (2013), “Monopsony 2013: What’s the Latest on (Too) Low Prices?”, 61st Spring Meeting of the Section of Antitrust Law, American Bar Association, 12 April 2013, Washington, D.C.

[165] Johnstone, R. et al. (2012), Beyond employment : the legal regulation of work relationships, The Federation Press, https://www.federationpress.com.au/bookstore/book.asp?isbn=9781862878891 (accessed on 22 November 2018).

[50] Kaine, S. (2017), All care and no responsibility: Why Airtasker can’t guarantee a minimum wage, The Conversation, http://dx.doi.org/10.1111/rego.12122.

[64] Kaltner, J. (2018), “Employment Status of Uber and Lyft Drivers: Unsettlingly Settled”, Hastings Women’s Law Journal, Vol. 29/1, pp. 29-54, https://repository.uchastings.edu/cgi/viewcontent.cgi?article=1403&context=hwlj (accessed on 12 December 2018).

[11] Keneally, K., C. Saleski and C. Engell (2015), All Employers Face Independent Contractor, Employment Tax Scrutiny, https://www.forbes.com/sites/janetnovack/2015/06/24/all-employers-face-independent-contractor-employment-tax-scrutiny/#1116972c6ea8 (accessed on 29 November 2018).

[53] Kenner, J. (2017), “Inverting the flexicurity paradigm: the United Kingdom and zero hours contracts”, in Ales, E., O. Deinert and J. Kenner (eds.), Core and contingent work in the European Union, Hart Publishing, Oxford.

[48] Kessler, S. (2016), Could A Minimum Wage Work In The Gig Economy?, Fast Company, https://www.fastcompany.com/3058599/could-a-minimum-wage-work-in-the-gig-economy (accessed on 5 November 2018).

[52] Kingsley, S., M. Gray and S. Suri (2015), “Accounting for Market Frictions and Power Asymmetries in Online Labor Markets”, Policy & Internet, Vol. 7/4, pp. 383-400, http://dx.doi.org/10.1002/poi3.111.

[122] Knable Gotts, I. (ed.) (2018), The Private Competition Enforcement Review - Edition 11 - The Law Reviews, Law Business Research Ltd., London, https://thelawreviews.co.uk/edition/1001145/the-private-competition-enforcement-review-edition-11 (accessed on 19 November 2018).

[172] Korean Fair Trade Commission (2009), Guidelines for Concerted Practice Review, Korean Fair Trade Commission , Seoul.

[68] Kovács, E. (2017), “Regulatory Techniques for ’Virtual Workers’”, Hungarian Labour Law E-Journal, Vol. 2017/2, http://www.hllj.hu (accessed on 14 December 2018).

[82] Krotoski, M. et al. (2018), No-Poaching Agreements and Antitrust Laws: What Global Employers Should Know, Morgan Lewis, https://www.morganlewis.com/events/tech-mayrathon-no-poaching-agreements-and-antitrust-laws-what-global-employers-should-know (accessed on 13 December 2018).

[83] Krueger, A. and O. Ashenfelter (2018), “Theory and Evidence on Employer Collusion in the Franchise Sector”, NBER Working Paper, No. 24831, National Bureau of Economic Research, http://dx.doi.org/10.3386/w24831.

[123] Krueger, A. and E. Posner (2018), A Proposal for Protecting Low-Income Workers from Monopsony and Collusion, Brookings, Washington, D.C., http://www.hamiltonproject.org/assets/files/protecting_low_income_workers_from_monopsony_collusion_krueger_posner_pp.pdf (accessed on 25 October 2018).

[100] Labour 20 (2018), L20 Statement to the G20 Labour And Employment Ministers’ Meeting, G20, Argentina 2018, https://www.g20.org/sites/default/files/media/l20declarationen.pdf (accessed on 3 December 2018).

[51] Lehdonvirta, V. (2018), “Flexibility in the Gig Economy: Managing Time on Three Online Piecework Platforms”, New Technology, Work and Employment, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3099419 (accessed on 1 June 2018).

[32] Lenaerts, K. et al. (2018), Online Talent Platforms, Labour Market Intermediaries and the Changing World of Work, https://www.ceps.eu/system/files/CEPS_IZA_OnlineTalentPlatforms.pdf (accessed on 28 November 2018).

[39] Liebman, S. (2006), Employment situations and workers’ protection - Italy, http://ilo.org/wcmsp5/groups/public/---ed_dialogue/---dialogue/documents/genericdocument/wcms_205366.pdf (accessed on 5 June 2018).

[22] Linder, M. (1999), “Dependent and Independent Contractors in Recent U.S. Labor Law: An Ambiguous Dichotomy Rooted in Simulated Statutory Purposelessness”, Comparative Labor Law & Policy Journal, Vol. 187, https://ir.uiowa.edu/cgi/viewcontent.cgi?referer=https://www.google.fr/&httpsredir=1&article=1010&context=law_pubs (accessed on 4 December 2018).

[138] Lipsitz, M. and M. Johnson (2018), “Why Are Low-Wage Workers Signing Noncompete Agreements?”, American Economic Association Annual Meeting, 5-7 January 2018, Philadelphia, PA., https://www.aeaweb.org/conference/2018/preliminary/1515.

[120] Luz, R. and G. Spagnolo (2017), “Leniency, collusion, corruption, and whistleblowing”, Journal of Competition Law & Economics, Vol. 13/4, pp. 729-766, http://dx.doi.org/10.1093/joclec/nhx025.

[92] Manning, A. (2011), “Imperfect Competition in the Labor Market”, in Ashenfelter, O. and D. Card (eds.), Handbook of Labor Economics, Elsevier, http://dx.doi.org/10.1016/s0169-7218(11)02409-9.

[72] Manning, A. (2003), Monopsony in motion : imperfect competition in labor markets, Princeton University Press, https://press.princeton.edu/titles/7522.html (accessed on 25 October 2018).

[103] Manning, A. and B. Petrongolo (2017), “How Local Are Labor Markets? Evidence from a Spatial Job Search Model”, American Economic Review, Vol. 107/10, pp. 2877-2907, http://dx.doi.org/10.1257/aer.20131026.

[65] Marcano, I. (2018), “E-hailing and Employment Rights: The Case for an Employment Relationship Between Uber and its Drivers in South Africa”, Cornell International Law Journal, Vol. 51, pp. 273-295, https://www.lawschool.cornell.edu/research/ILJ/upload/Marcano-note-final.pdf (accessed on 5 December 2018).

[87] Marinescu, I. and H. Hovenkamp (forthcoming), “Anticompetitive Mergers in Labor Markets”, Indiana Law Journal, http://dx.doi.org/10.2139/ssrn.3124483.

[104] Marinescu, I. and R. Rathelot (2018), “Mismatch Unemployment and the Geography of Job Search”, American Economic Journal: Macroeconomics, Vol. 10/3, pp. 42-70, http://dx.doi.org/10.1257/mac.20160312.

[76] Martins, P. (2018), “Making their own weather? Estimating employer labour-market power and its wage effects”, QMUL Working Papers 95, Queen Mary University London, http://webspace.qmul.ac.uk/pmartins/CGRWP95.pdf (accessed on 9 November 2018).

[135] Marx, M. (2011), “The Firm Strikes Back”, American Sociological Review, Vol. 76/5, pp. 695-712, http://dx.doi.org/10.1177/0003122411414822.

[157] Marx, M. and L. Fleming (2012), “Non-compete Agreements: Barriers to Entry … and Exit?”, Innovation Policy and the Economy, Vol. 12, pp. 39-64, http://dx.doi.org/10.1086/663155.

[134] Marx, M., D. Strumsky and L. Fleming (2009), “Mobility, Skills, and the Michigan Non-Compete Experiment”, Management Science, Vol. 55/6, pp. 875-889, http://dx.doi.org/10.1287/mnsc.1080.0985.

[127] Meritas (2017), Employee Non-Compete Agreements in Europe, Middle East and Africa, Meritas, Minneapolis, Minn., http://www.meritas.org.

[15] Milanez, A. and B. Bratta (2019), “Taxation and the future of work: How tax systems influence choice of employment form”, OECD Taxation Working Papers, No. 41, OECD Publishing, Paris, https://dx.doi.org/10.1787/20f7164a-en.

[16] Ministerie van Financiën (2015), IBO Zelfstandigen zonder personeel, Rijksoverheid, https://www.rijksoverheid.nl/documenten/rapporten/2015/10/02/eindrapport-ibo-zelfstandigen-zonder-personeel (accessed on 18 March 2019).

[129] Motta, M. and T. Rønde (2002), “Trade Secret Laws, Labor Mobility, and Innovations”, CEPR Discussion Paper, No. 3615, CEPR, London, https://ideas.repec.org/p/cpr/ceprdp/3615.html.

[40] Muehlberger, U. (2007), “Work on the Border Between Employment and Self-Employment”, in Dependent Self-Employment, Palgrave Macmillan UK, London, http://dx.doi.org/10.1057/9780230288782_2.

[78] Naidu, S., E. Posner and E. Weyl (forthcoming), “Antitrust Remedies for Labor Market Power”, Harvard Law Review.

[176] New York State Attorney General (2018), A.G. Underwood Announces Settlement With WeWork To End Use Of Overly Broad Non-Competes That Restricted Workers’ Ability To Take New Jobs, Press Releases of the New York State Attorney General, https://ag.ny.gov/press-release/ag-underwood-announces-settlement-wework-end-use-overly-broad-non-competes-restricted (accessed on 18 September 2018).

[177] New York State Attorney General (2018), Non-Compete Agreements In New York State Frequently Asked Questions What is a non-compete agreement? Are non-competes legal?, New York State Attorney General, New York, N.Y., https://ag.ny.gov/sites/default/files/non-competes.pdf (accessed on 21 November 2018).

[126] Nicandri, A. (2011), “The Growing Disfavor of Non-Compete Agreements in the New Economy and Alternative Approaches for Protecting Empoyers’ Proprietary Information and Trade Secrets”, University of Pennsylvania Journal of Business Law, Vol. 13/4, https://scholarship.law.upenn.edu/jbl/vol13/iss4/6 (accessed on 11 December 2018).

[2] OECD (2019), Policy Responses to New Forms of Work, OECD Publishing, Paris, https://doi.org/10.1787/0763f1b7-en.

[93] OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264308817-en.

[101] OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-2018-en.

[159] OECD (2018), Taxi, ride-sourcing and ride-sharing services - Background Note by the Secretariat, DAF/COMP/WP2(2018)1, OECD, Paris, https://one.oecd.org/document/DAF/COMP/WP2(2018)1/en/pdf.

[14] OECD (2018), The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264306943-en.

[125] OECD (2017), Algorithms and Collusion: Competition Policy in the Digital Age, OECD, Paris, http://www.oecd.org/daf/competition/Algorithms-and-colllusion-competition-policy-in-the-digital-age.pdf.

[118] OECD (2017), OECD Employment Outlook 2017, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2017-en.

[13] OECD (2015), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264235120-en.

[96] OECD (2015), OECD Employment Outlook 2015, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-2015-en.

[8] OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2014-en.

[61] OECD (2013), OECD Employment Outlook 2013, OECD Publishing, https://doi.org/10.1787/empl_outlook-2013-en (accessed on 8 January 2018).

[18] OECD (2008), OECD Employment Outlook 2008, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2008-en.

[178] Office of the Attorney General of the State of Illinois (2018), Non-Compete Agreements: Frequently Asked Questions, http://www.illinoisattorneygeneral.gov/rights/Non_Compete_Agreements.pdf (accessed on 21 November 2018).

[179] Office of the Washington State Attorney General (2018), AG Ferguson’s initiative to end no-poach clauses nationwide continues with four additional chains, Press Release of the Office of the Washington State Attorney General, http://www.atg.wa.gov/news/news-releases/ag-ferguson-s-initiative-end-no-poach-clauses-nationwide-continues-four (accessed on 21 November 2018).

[49] Patty, A. (2017), Airtasker and unions make landmark agreement to improve pay rates and conditions, The Sunday Morning Herald, https://www.smh.com.au/business/workplace/airtasker-and-unions-make-landmark-agreement-to-improve-pay-rates-and-conditions-20170427-gvtvpo.html (accessed on 5 November 2018).

[35] Pedersini, R. (2002), ’Economically dependent workers’, employment law and industrial relations, Eurofound, https://www.eurofound.europa.eu/publications/report/2002/economically-dependent-workers-employment-law-and-industrial-relations (accessed on 7 December 2018).

[34] Prassl, J. and M. Risak (2016), “Uber, Taskrabbit, & Co: Platforms as Employers? Rethinking the Legal Analysis of Crowdwork”, Oxford Legal Studies Research Paper, No. 8/2016, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2733003 (accessed on 28 November 2018).

[141] Prescott, J., N. Bishara and E. Starr (2016), “Understanding noncompetition agreements: The 2014 noncompete survey project”, Michigan State Law Review, Vol. 2016/2, pp. 369-464, https://digitalcommons.law.msu.edu/cgi/viewcontent.cgi?article=1163&context=lr (accessed on 13 December 2018).

[46] PvdA (2017), Een Verbonden Samenleving - Verkiezingsprogramma 2017, Partij van de Arbeid, https://www.pvda.nl/wp-content/uploads/2017/02/PvdAVerkiezingsprogramma2017EenVerbondenSamenleving.pdf (accessed on 5 November 2018).

[180] Pynnonen, B. (1994), “Ohio and Michigan Law on Postemployment Covenants Not to Compete”, Ohio State Law Journal, Vol. 55, pp. 215-235, https://kb.osu.edu/bitstream/handle/1811/64679/OSLJ_V55N1_0215.pdf (accessed on 23 November 2018).

[110] Qiu, Y. and A. Sojourner (2019), “Labor-Market Concentration and Labor Compensation”, IZA Discussion Paper, No. 12089, IZA, Bonn.

[131] Rauch, J. (2016), “Dynastic entrepreneurship, entry, and non-compete enforcement”, European Economic Review, Vol. 86, pp. 188-201, http://dx.doi.org/10.1016/j.euroecorev.2015.12.004.

[94] Rinz, K. (2018), “Labor Market Concentration, Earnings Inequality, and Earnings Mobility”, CARRA Working Paper, No. 2018-10, US Census Bureau, https://www.census.gov/library/working-papers/2018/adrm/carra-wp-2018-10.html (accessed on 8 January 2019).

[97] Risak, M. (2018), Fair Working Conditions for Platform Workers Possible Regulatory Approaches at the EU Level, http://www.fes.de/ipaBestellungen/Kontakt:[email protected] (accessed on 30 November 2018).

[37] Rönnmar, M. (2004), The personal scope of labour law and the notion of employee in Sweden, https://lup.lub.lu.se/search/publication/617804 (accessed on 23 November 2018).

[91] Rosenblat, A. and L. Stark (2016), “Algorithmic Labor and Information Asymmetries: A Case Study of Uber’s Drivers”, International Journal of Communication, Vol. 10, pp. 3758-3784, https://ijoc.org/index.php/ijoc/article/view/4892/1739.

[142] Rosenthal, R. (2018), Employee non-competes: a state by state survey, Howard & Howard, Las Vegas, Nev., http://nevadalaw.info/wp-content/uploads/2018/01/Employee-Non-Competes-50-State-Survey.pdf (accessed on 13 December 2018).

[63] Ross, H. (2015), “Ridesharing’s House of Cards: O’Connor v. Uber Technologies, Inc. and the Viability of Uber’s Labor Model in Washington”, Washington Law Review, Vol. 90, pp. 1431-1469, https://digital.law.washington.edu/dspace-law/bitstream/handle/1773.1/1489/90WLR1431.pdf (accessed on 12 December 2018).

[128] Rubin, P. and P. Shedd (1981), “Human Capital and Covenants Not to Compete”, The Journal of Legal Studies, Vol. 10/1, pp. 93-110, http://www.jstor.org/stable/724227.

[20] Rubinstein, M. (2012), “Employees, Employers, and Quasi-Employers: An Analysis of Employees and Employers Who Operate in the Borderland between an Employer-and-Employee Relationship”, NYLS Legal Studies Research Paper, No. 12/13 40, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2047484 (accessed on 4 January 2019).

[98] Salehi, N. et al. (2015), We Are Dynamo: Overcoming Stalling and Friction in Collective Action for Crowd Workers, ACM Press, — April 18 - 23, 2015, http://dx.doi.org/10.1145/2702123.2702508.

[152] Salop, S. (2010), “Question: What is the Real and Proper Antitrust Welfare Standard? Answer: The True Consumer Welfare Standard”, Loyola Consumer Law Review, Vol. 22/3.

[150] Schwartz, M. (2004), Should Antitrust Assess Buyer Market Power Differently Than Seller Market Power?, US Department of Justice, Washington DC, https://www.justice.gov/atr/should-antitrust-assess-buyer-market-power-differently-seller-market-power (accessed on 22 November 2018).

[99] Silberman, S. (2018), Ombuds Office of the Crowdsourcing Code of Conduct, Presentation given to Sharers and Workers, Brussels, http://wtf.tw/etc/bru/slides.pdf (accessed on 4 December 2018).

[66] Silberman, S. (2018), “Rights for Platform Workers”, IG Metall Crowdsourcing Project Discussion Papers, IG Metall.

[154] Sims, R. (2018), Address to the 2018 Annual RBB Economics Conference, Speeches of the Chairman of the Australian Competition and Consumer Commission, https://www.accc.gov.au/speech/address-to-the-2018-annual-rbb-economics-conference#ref10 (accessed on 30 November 2018).

[12] Söderqvist, F. (2018), “Sweden: will history lead the way in the age of robots and platforms?”, Policy Network, https://policynetwork.org/opinions/essays/sweden-will-history-lead-way-age-robots-platforms/ (accessed on 29 November 2018).

[77] Sokolova, A. and T. Sorensen (2018), “Monopsony in Labor Markets: A Meta-Analysis”, IZA Discussion Paper, No. 11966, IZA, Bonn, http://ftp.iza.org/dp11966.pdf (accessed on 8 January 2019).

[137] Starr, E. (forthcoming), “Consider This: Training, Wages, and the Enforceability of Covenants Not to Compete”, Industrial and Labor Relations Review, http://dx.doi.org/10.1177/0019793919826060.

[88] Starr, E., J. Prescott and N. Bishara (2018), “Noncompetes in the U.S. Labor Force”, University of Michigan Law & Econ Research Paper, No. 18-013, http://dx.doi.org/10.2139/ssrn.2625714.

[136] Starr, E., J. Prescott and N. Bishara (2016), “Noncompetes and Employee Mobility”, University of Michigan Law & Econ Research Paper, No. 16-032, http://dx.doi.org/10.2139/ssrn.2858637.

[166] Stewart, A. and J. Stanford (2017), “Regulating work in the gig economy: What are the options?”, The Economic and Labour Relations Review, Vol. 28/3, pp. 420-437, http://dx.doi.org/10.1177/1035304617722461.

[59] Sweeney, L. (2013), “Discrimination in Online Ad Delivery”, http://dataprivacylab.org/projects/onlineads/1071-1.pdf (accessed on 1 June 2018).

[17] UK Parliament (2017), Self-employment and the gig economy, Committee on Work and Pensions, UK Parliament, https://publications.parliament.uk/pa/cm201617/cmselect/cmworpen/847/84705.htm (accessed on 8 June 2018).

[84] US Department of Justice and Federal Trade Commission (2016), Antitrust Guidance For Human Resource Professionals (October 2016), U.S. Department of Justice and the Federal Trade Commission, Washington, D.C., https://www.ftc.gov/public-statements/2016/10/antitrust-guidance-human-resource-professionals-department-justice (accessed on 15 November 2018).

[119] US Department of Justice and Federal Trade Commission (2016), Antitrust Red Flags for Employment Practices, US Department of Justice and Federal Trade Commission , Washington, D.C., http://go.usa.gov/xkdrq. (accessed on 16 November 2018).

[106] US Department of Justice and Federal Trade Commission (2010), Horizontal Merger Guidelines (08/19/2010), U.S. Department of Justice and the Federal Trade Commission, Washington, D.C., https://www.justice.gov/atr/horizontal-merger-guidelines-08192010 (accessed on 14 November 2018).

[69] US House of Representatives (2009), “Arbitration or arbitrary: the misuse of mandatory arbitration to collect consumer debts”, Hearing before the Subcommittee on Domestic Policy of the Committee on Oversight and Government Reform, July 22, 2009 Serial 111-125, https://www.gpo.gov/fdsys/pkg/CHRG-111hhrg64915/pdf/CHRG-111hhrg64915.pdf.

[7] Waas, B. et al. (2017), Crowdwork – A Comparative Law Perspective, http://www.hugo-sinzheimer-institut.de/fileadmin/user_data_hsi/Veroeffentlichungen/HSI_Schriftenreihe/Waas_Liebman_Lyubarsky_Kezuka_Crowdwork.pdf (accessed on 1 June 2018).

[24] Waldfogel, J. (1998), “Does inter-judge disparity justify empirically based sentencing guidelines?”, International Review of Law and Economics, Vol. 18/3, pp. 293-304, http://dx.doi.org/10.1016/S0144-8188(98)00013-1.

[54] Walters, D. (2017), An Inspector Calls? Achieving regulatory compliance on OHS in times of change, Tööinspektsioon, Tallinn, http://ti.ee/fileadmin/user_upload/failid/dokumendid/Meedia_ja_statistika/Teavitustegevus/SLIC/David_Walters.pdf (accessed on 2 December 2018).

[168] Wank, R. (1999), Workers’ Protection - National Study for Germany for the ILO, ILO, https://www.ilo.org/wcmsp5/groups/public/---ed_dialogue/---dialogue/documents/genericdocument/wcms_205364.pdf (accessed on 4 December 2018).

[102] Webber, D. (2015), “Firm market power and the earnings distribution”, Labour Economics, Vol. 35, pp. 123-134, http://dx.doi.org/10.1016/j.labeco.2015.05.003.

[151] Werden, G. (2007), “Monopsony and the Sherman act: consumer welfare in a new light”, Antitrust Law Journal, Vol. 74/3, pp. 707-737, http://www.jstor.org/stable/27897564.

[132] Wickelgren, A. (2018), “A novel justification for legal restrictions on non-compete clauses”, International Review of Law and Economics, Vol. 54, pp. 49-57, http://dx.doi.org/10.1016/j.irle.2017.10.004.

[9] Williams, C. and F. Lapeyre (2017), “Dependent self-employment: Trends, challenges and policy responses in the EU”, Employment Working Paper, No. 228, ILO, http://ilo.org/wcmsp5/groups/public/---ed_emp/documents/publication/wcms_614176.pdf.

[70] World Employment Confederation (2018), Compendium of voluntary initiatives promoting ethical recruitment practices by the national federations of the World Employment Confederation, World Employment Confederation, Brussels, http://www.wecglobal.org/uploads/media/WEC_Compendium_Final.pdf.

[86] Yeoh, P. (2014), “Whistleblowing: motivations, corporate self-regulation, and the law”, International Journal of Law and Management, Vol. 56/6, pp. 459-474, http://dx.doi.org/10.1108/ijlma-06-2013-0027.

[130] Younge, K., T. Tong and L. Fleming (2014), “How anticipated employee mobility affects acquisition likelihood: Evidence from a natural experiment”, Strategic Management Journal, Vol. 36/5, pp. 686-708, http://dx.doi.org/10.1002/smj.2237.

copy the linklink copied!Annex 4.A. Labour market monopsony: Evidence and regulation

Many workers face an unbalanced power relationship vis-à-vis their employer. This emerges because employers have a much higher degree of control over the relationship than workers (employers take hiring, firing and other organisational decisions and often may fix pay and working conditions) and because the latter may have few or no outside options. In many situations, employers’ buyer power is not compensated by sufficient bargaining power on the side of the workers – in particular in the absence of collective bargaining – and may, therefore, lead to lower employment and pay as well as worse working conditions.

The labour economics literature often defines as labour market monopsony a situation in which employers’ buyer power is not compensated by sufficient workers’ bargaining power and workers have low or no outside options. Strictly speaking, the term “monopsony” refers to the extreme case in which one buyer dominates a specific upstream market and, in order to maximise its profits, can fix input purchases and prices below the level that maximises social welfare. The term “oligopsony” would more correctly refer to cases where few firms dominate buying in a market. Yet, the literature often refers to labour market monopsony to indicate any of these situations with respect to a specific segment of the labour market (Manning, 2003[72]; Bhaskar, Manning and To, 2002[73]).

This annex briefly summarises existing evidence on labour market monopsony and discusses instruments to tackle the abuse of employers’ buyer power and its sources. Other policy instruments, which are not primarily intended to address labour market monopsony, could nonetheless affect its sources and/or effects. These are discussed in Section 4.3 in the main text and in Chapter 5.

## copy the linklink copied!Empirical evidence on labour market monopsony and its effects

There is a growing literature demonstrating that labour market competition is far from perfect. The early literature tended to attribute a key role to collective bargaining and labour market institutions in limiting competition in the labour market – e.g. OECD (2018[101]) and Chapter 5. Recent academic work, however, has begun to point out the key role played by labour market concentration and, more generally, buyer power in the labour market – in particular in the absence of collective bargaining.

A large number of studies for specific labour markets have estimated that workers cannot easily switch to a different employer in reaction to wage changes in a specific firm (i.e. residual labour supply elasticities are low) – see e.g. Manning (2003[72]), Webber (2015[102]), Naidu, Posner and Weyl (forthcoming[78]), and references cited therein. These studies typically find residual labour supply elasticities that are close to, or smaller than 5.47 They have usually been found to be smaller in America than in Europe, but even in the latter they tend to remain within single digits (Sokolova and Sorensen, 2018[77]). Single-digit elasticities are usually considered as evidence of labour market monopsony.48 While most of this evidence refers to employees, there is some evidence that residual labour supply elasticities are even lower for some independent contractors, including certain platform workers – see Section 4.3 in the main text.

Manning and Petrongolo (2017[103]) and Marinescu and Rathelot (2018[104]) have shown that both UK and US labour markets are very local, in the sense that job searches decline rapidly with distance. Using “scraped” data on online job postings from thousands of websites, Azar, Marinescu and Steinbaum (2017[105]) and Azar et al. (2018[74]) find that the average local US labour market – as defined at the level of the sixth digit of occupational classification within the same commuting zone – is highly concentrated, that is more concentrated than the threshold triggering concern of lack of competition according to the US horizontal merger guidelines (US Department of Justice and Federal Trade Commission, 2010[106]).49 Recent evidence for Portugal shows lower levels of concentration in this country (Martins, 2018[76]). Yet, the most reliable measures suggest that more than 15% of Portuguese employment is exposed to a higher level of labour market concentration than the above-mentioned threshold. Other studies for the United States and the United Kingdom have used a definition of the relevant labour market based on the industry and commuting zone. These studies find that about 25% of employment in both countries is in a local labour market with a level of concentration close or above the high-concentration threshold (Hershbein, Macaluso and Yeh, 2018[107]; Abel, Tenreyro and Thwaites, 2018[75]; Rinz, 2018[94]).

Market power in the markets for inputs (including labour) can have a negative impact on prices – i.e. wages and benefits – and quantities – overall employment, see e.g. Manning (2003[72]), Ashenfelter, Farber and Ransom (2010[108]) and Blair and Harrison (2010[109]). Azar, Marinescu and Steinbaum (2017[105]) and Azar et al. (2018[74]) find that employer concentration has a significant depressing effect on posted wages, while Hershbein, Macaluso and Yeh (2018[107]), Martins (2018[76]) and Qiu and Sojourner (2019[110]) find a negative relationship between concentration and pay using actual wages. Rinz (2018[94]) finds that this association is stronger at the lower tail of the earnings distribution, while Webber (2015[102]) estimates that residual labour supply elasticities are smaller and have a stronger effect on earnings at the bottom of the earnings distribution. Abel, Tenreyro and Thwaites (2018[75]) find that higher levels of labour market concentration are associated with lower pay especially amongst workers not covered by a collective bargaining agreement. Benmelech, Bergman and Kim (2018[111]) show that the negative relationship between concentration and actual wages becomes steeper over time, while the link between productivity growth and wage growth is stronger when labour markets are less concentrated. Hershbein and Macaluso (2018[112]) find that highly concentrated labour markets tend to have greater demand for skilled rather than unskilled labour. Qiu and Sojourner (2019[110]) estimate that workers in concentrated labour markets are less likely to be covered by employer-sponsored health insurance. Finally, evidence on lack or limited effects of the minimum wage on employment (when the former is not too high) is also considered as evidence of employment-depressing effects of monopsony, at least at the lower tail of the pay distribution – see e.g. Manning (2011[92]) and Card and Krueger (2015[43]) .

Recent research has also explored the link between market power in product markets and labour markets, finding that increasing price-cost margins have gone hand in hand with declining labour shares and decreasing workers’ bargaining power (De Loecker and Eeckhout, 2018[113]; Barkai, 2017[114]; Autor et al., 2017[115]; Bell, Bukowski and Machin, 2018[116]), suggesting that buyer power in the labour market may be associated with some degree of market power in the product market, with possible implications also for consumer welfare.50

## copy the linklink copied!Enforcing regulation to address labour market monopsony

To date, regulators have paid somewhat less attention to labour market power and firms’ ability to pay workers less than their marginal productivity than to product market power. There are four areas of action, in which legislators and enforcement authorities (including labour inspectorates and antitrust authorities) could consider taking a more active role in many countries: i) no-poaching agreements and wage-setting collusion; ii) non-compete clauses; iii) mergers; and iv) information asymmetries between employers and workers. These will be discussed one-by-one below.

### No-poaching agreements and labour market collusion

In most jurisdictions, competition law forbids collusion among buyers of intermediate goods or services – see e.g. Blair and Wang (2017[81]). The labour market is no exception in this respect. Firms can collude in the labour markets in several ways. For example, companies competing for the same type of professionals can agree on refraining from hiring or making job offers to workers employed by the others (so-called “no-poaching agreements”), thereby limiting the outside options of their employees or independent contractors. Alternatively, firms competing in the same labour market may agree to apply a common compensation policy to employees or independent contractors, thereby preventing market forces to raise wages as the result of supply and demand (wage collusion). Collusion, however, can occur in more elusive ways: for example, buyers can simply meet and discuss compensation and hiring policies, thereby reaching de facto an arrangement without formal agreement (US Department of Justice and Federal Trade Commission, 2016[84]). To be considered illicit, a collusive agreement does not need to concern direct competitors in the downstream output market51, nor does it need to have an adverse impact on final consumer prices.52

General statistics on collusion are difficult to collect, since figures on those illicit behaviours that escape investigation are typically not available. Anecdotal evidence suggests however that the phenomenon is far from negligible: US Deputy Assistant Attorney General for Civil Antitrust, Barry Nigro, said recently that he found “shocking how prevalent agreements have become between companies not to solicit or hire each others’ employees” – cited in Krotoski et al. (2018[82]).

Only a few jurisdictions have been significantly active in fighting labour market collusion. A number of cases of collusion in the labour market have been adjudicated by US courts and antitrust authorities in recent years – see e.g. Marinescu and Hovenkamp (forthcoming[87]) for a number of examples. In a few of these cases, illegal agreements have concerned collusion on compensation of independent contractors or employees on very atypical work arrangements.53 Similar cases have been adjudicated in some other jurisdictions, notably in France, the Netherlands, Spain, and the United Kingdom.54 However, law enforcement authorities and courts in most OECD countries have been somewhat less aggressive in fighting labour market collusion – see e.g. Gurkaynak, Guner and Ozkanli (2013[117]); and Blair and Wang (2017[81]). This heterogeneity may partially reflect the inherent difficulty of distinguishing tacit collusion agreements from lawful co-ordination among employers in countries with co-ordinated or branch-level collective bargaining (Gurkaynak, Guner and Ozkanli, 2013[117]; OECD, 2017[118]; 2018[101]).

Providing explicit guidance about labour market collusion is crucial to guide and set priorities for action of enforcement agencies. In fact, the greater number of cases in the United States is likely to be also the result of the explicit guidance provided by US antitrust authorities (US Department of Justice and Federal Trade Commission, 2016[84]; 2016[119]). The issued guidelines explicitly refer to collusion in the labour market, present clear examples of illicit behaviours, and underline the importance of fighting collusion in order to help “actual and potential employees through higher wages, better benefits, or other terms of employment” (US Department of Justice and Federal Trade Commission, 2016, p. 2[84]).55 By contrast, in other jurisdictions, guidelines typically refer to any type of joint-cooperation agreements, with examples from both cooperation in production and cooperation in input purchasing, at the risk of diluting the importance of fighting illicit agreements in the labour market.56

Providing adequate whistleblower protection is also crucial for effective action by enforcement agencies (Dyck, Morse and Zingales, 2010[85]; Yeoh, 2014[86]). As information which allows the identification of collusion agreements is often confidential and can be passed to enforcement agencies only by insiders, whistleblower protection is needed, in particular in the case of independent contractors who usually are not protected by dismissal legislation provisions covering whistleblowing – see e.g. OECD (2014[8]).

Enforcement agencies should also be able to provide adequate sanctions for illicit behaviours. Deterrence of possible future collusive behaviours from other companies would not be exerted if those caught up colluding are simply ordered to stop. Nevertheless, adequate leniency programmes, offering immunity to the first cartel member that blows the whistle, may play an important role – see e.g. Luz and Spagnolo (2017[120]).

Private actions seeking redress in courts should be seen as a complement rather than a substitute to public enforcement actions in deterring collusive behaviour. Beyond public enforcement by antitrust authorities, most OECD legislations allow private damage actions initiated by the victims of collusive behaviour (Blair and Wang, 2017[81]). However, individual employees often do not have the resources or incentives to sue employers for this type of antitrust violations because an antitrust suit is usually much more costly than the individual damage. By contrast, an antitrust claim concerning product markets is usually brought by a large firm claiming conspicuous compensation for damages resulting from the anticompetitive practices of a rival. Hence, private damage actions concerning labour market collusion would often require collective actions. In most countries, however, private collective actions are frequently subject to strict requirements as regards the parties that have the legal standing to initiate a claim – for example, associations and non-governmental organisations often do not have such standing, see e.g. Haar (2018[121]) and Knable Gotts (2018[122]). In turn, this may limit the capacity of filing such claims for independent contractors in the absence of representative trade unions (see also Chapter 5). Class actions may also impose considerable costs and risks on the initiating third parties or law firms, in the absence of previous infringement sanctions by antitrust authorities (Krueger and Posner, 2018[123]). As a result, collective claims usually follow the investigation of antitrust authorities.

The development of algorithms and artificial intelligence is opening up new opportunities for collusion, including in the labour market. Firms can indeed independently design algorithms to signal and coordinate a common compensation policy as well as to monitor and retaliate against deviators without any explicit communication (Capobianco and Gonzaga, 2017[124]). This is all the more likely in the labour markets for independent contractors and platform work, where in essentially all countries there is no legal constraint to adjust downward compensation for similar tasks over time. However, it is not obvious that the effect of algorithms is necessarily that of facilitating collusion. For example, if the use of different algorithms leads to cost asymmetry across companies, collusion might be difficult to sustain, due to the inherent difficulties of finding a focal point to coordinate and, as a consequence, of the low incentives for the most efficient firms to collude (OECD, 2017[125]). Moreover, it is not obvious how regulation can be adapted to allow enforcement agencies fighting effectively against this type of collusion without curbing innovation and depressing the growth of the digital economy (Capobianco and Gonzaga, 2017[124]). More research is still needed in this area.

A question that remains open is whether no-poaching clauses among franchisees should be considered unlawful. Krueger and Ashenfelter (2018[83]) estimate that more than 50% of major franchise companies in the United States use no-poaching clauses in their franchising agreements. Collusion agreements are typically defined as illicit pacts across different organisations, while franchisees are usually considered as part of the same organisation. Yet, when several franchisees are the dominant employer in a labour market, a no-poaching rule has clear anti-competitive effects in that market (Krueger and Posner, 2018[123]). Several cases involving franchisees are under examination by US courts, but typically concern situations in which franchisees are totally independent in terms of staffing decisions.57

### Non-compete covenants

Non-compete covenants are clauses in contracts that prevent workers from working for a competitor after they separate from the employer (and should be distinguished from moonlighting clauses that prevent employees from performing work for a competitor during the term of their employment).58 Sometimes the literature distinguishes between “non-compete” and “garden leave” clauses, the difference being that in the latter the worker is compensated after separating from the employer for the period of validity of the covenant while in the former she is not – see e.g. Nicandri (2011[126]). As in many countries and states a clause without worker compensation is not enforceable – see e.g. Meritas (2017[127]), for the purpose of this chapter, the term “non-compete covenant” or “non-compete agreement” refers to both type of clauses.

In most countries, non-compete agreements are lawful and justified by the need to protect trade secrets and specific investment in the employment relationship by the employer (such as certain types of training and investment in knowledge).59 These clauses are often considered to have a positive impact on innovation, in particular when companies cannot protect their investment in knowledge through patents or other types of contracts. Yet, the literature is far from conclusive on this matter, in particular due to the possible impact of restrictive covenants on knowledge spillovers – see e.g. Rubin and Shedd (1981[128]); Motta and Rønde (2002[129]); Younge, Tong and Fleming (2014[130]); Rauch (2016[131]) and Wickelgren (2018[132]).

Non-compete covenants, however, can also be used as a legal instrument to increase monopsony power in the labour market. The evidence on litigation suggests that employers may use these clauses in order to limit effectively the outside options of their workers, for example by preventing them from taking up similar jobs even in companies with which they do not compete in practice.60 Consequently, non-compete clauses typically result in lower wages and job mobility – see e.g. Garmaise (2009[133]), Marx, Strumsky and Fleming (2009[134]); Marx (2011[135]); Starr, Prescott and Bishara (2016[136]) and Starr (forthcoming[137]).

More striking, recent cases have concerned low-skilled workers involved in production of standardised products, such as sandwich making, and with little access to company tacit knowledge.61 More generally, available evidence suggests that non-compete covenants are pervasive in low-skilled jobs. For example, Lipsitz and Johnson (2018[138]) show that non-compete clauses are more common among minimum wage hairdressers than among those working for a higher wage. More striking, Starr, Prescott and Bishara (2018[88]) estimate that almost 30 million US workers are currently covered by non-compete agreements, and that many of these agreements are imposed broadly on workers with less than bachelor’s degree, often with low income or no access to trade secrets. This is even more challenging since low-skilled workers tend to be more vulnerable to monopsony than high-skilled workers: in particular, they often have fewer outside options and less bargaining power since they have less access to transportation, well-situated housing markets and job information – see e.g. Cahuc, Postel-Vinay and Robin (2006[139]) and Caldwell and Danieli (2018[140]).

In order to fight against non-compete agreements in situations where they are most likely used only to reduce competition in the labour market, governments could consider banning them, or establishing a rebuttable presumption of abusive use, in the case of certain type of positions, pay levels or skill requirements, for which a clear justification, such as the protection of trade secrets, seems implausible. A number of US states have introduced or are considering legislation limiting the use of non-compete clauses in particular for low-wage workers. For example, Massachusetts enacted a provision of this type for all covenants signed after October 2018 concerning either low-pay employees or low-skill jobs, with few exceptions.62 Similar provisions exist in three European legislations (Austria, Belgium and Luxembourg), where non-compete clauses are banned for workers earning less than a specified threshold that is close to, or even above, the median wage of full-time employees.63 Similarly, in the United Kingdom, the 2015 Small Business, Enterprise and Employment Act, by making exclusivity clauses for zero-hours contracts unlawful, has arguably resulted in a ban of restrictive covenants concerning future employment for these contracts.64 However, in most countries and states, while some restriction on the scope of restrictive covenants is often imposed, notably in terms of duration after separation and worker compensation during that period, a presumption (rebuttable or not) of abusive use is not made on the basis of the type of position, the level of pay or the skill requirements of the job. Instead, courts typically assess the reasonableness of non-compete agreements on a case-by-case basis (Meritas, 2017[127]). Moreover, regulations limiting the applicability of non-compete clauses often concern only employment relationships and do not extend to any type of contracts for services.65

Private actions initiated by workers themselves are rare and under-deterring. Even when restrictive covenants are unenforceable or unlawful according to statutory or case law, they may still be included in employment or service contracts as a way to put pressure on uninformed workers. In fact, costly private actions by workers would most often result in simply waving unlawful covenants, with no additional gain for plaintiff workers (Krueger and Posner, 2018[123]). As a result, private litigation cases are typically initiated by the employer aiming to enforce a restrictive covenant, rather than by employees hoping to waive them. Moreover, in most cases, workers are likely to honour these clauses by renouncing to search for certain jobs without consulting counsel or challenging the clause in court. For example, 19% of employees in California and North Dakota report having signed a non-compete clause, despite the fact that these clauses are legally not enforceable in these states (Prescott, Bishara and Starr, 2016[141]), suggesting that employers consider them having a mobility deterrence effect despite their formal lack of enforceability.66 Public guidance and counselling which clarify, in simple terms, under what conditions a non-compete clause is enforceable and lawful, can partially palliate this problem.67

Governments should also consider a ban on “blue-pencilling” – i.e. situations where courts redraft unreasonable or unlawful covenants in order to make them enforceable. When legislation allows “blue-pencilling”, workers are not guaranteed that restrictive covenants would be rendered void in the event of litigation, even when these clauses are clearly unenforceable as drafted. In such cases, employers could simply insert broad, unreasonable covenants not to compete in their employment contracts to put pressure on employees, to subsequently retreat to a narrower, more reasonable construction in the event of litigation. For example, a clause could state that the worker cannot accept alternative employment in an unreasonably long list of countries or regions, and the court could simply delete a few of them (or ask the company to delete a few of them) from the covenant to make the clause enforceable.68 For this reason, blue-pencilling has been rejected by courts in a few jurisdictions, such as Nebraska, Vermont, Virginia and Wisconsin (Rosenthal, 2018[142]) and, more recently, the United Kingdom.69 Yet, it is acceptable practice in many others, including a number of European countries – see e.g. Adler (2006[143]).

To the extent that private legal actions do not appear to exert an adequate deterring effect against the abuse of non-compete covenants with the intention of restricting mobility in the labour market, government and enforcement authorities should take a leading role in this area. For example, the attorneys general of the US states of New York and Illinois have recently been very active in investigating unreasonably broad or unlawful restrictive covenants, often reaching settlements involving significant sanctions with infringing companies.70 To be effective, however, enforcement agencies, including labour inspectorates, should have the possibility of imposing sanctions or taking the case to courts for imposing them. In particular, administrative sanctions could be considered in the case of inclusion of non-compete clauses in an employment contract for a type of job for which they are explicitly banned – such as for workers and independent contractors whose remuneration is below the enforceability threshold in those countries for which this threshold is defined in the law (see above).

### Mergers and anti-competitive conduct

Mergers can also have the effect of increasing employers’ buyer power in the labour market. In fact, if merging firms would together form the dominant buyer in an input market, the merged entity would likely use its buyer power to reduce quantities and prices in that input market, thereby increasing its profits. For this to occur, merging firms do not need to compete in the same downstream product market. For example, in the extreme case in which merging local input buyers sell their products in perfectly competitive markets, such as in markets for homogeneous tradable goods where prices are fixed by world demand and supply (such as markets for raw materials), the merger would not affect downstream product prices. Yet, in this case, the merger may generate significant welfare losses due to its impact in the input market – see e.g. Dobson and Inderst (2007[144]) and Marinescu and Hovenkamp (forthcoming[87]).

More generally, however, the merging firm could potentially pass part of the input price reduction onto downstream output prices, making final consumers gain from the merger, which implies that the welfare gain in the output market could offset the welfare loss in the input market. However, the actual pass-on of the reduction of input prices on downstream output prices will depend on a number of factors, including the degree of competition in the output markets, the degree of bargaining power of input suppliers and the type of outside options available to them – see e.g. Inderst and Shaffer (2008[145]) and Carlton, Israel and Coleman (2014[146]). In particular, worker bargaining power or good outside options for the workers are necessary conditions for a pass-on of any reduction in wages on downstream output prices. But, in the absence of collective bargaining, these conditions are unlikely to hold in many labour markets,71 where workers are capacity constrained – i.e. they cannot indefinitely extend their working time – and have a strong financial dependence on their employer – see e.g. Inderst and Shaffer (2008[145]) and Naidu, Posner and Weyl (forthcoming[78]).

In many jurisdictions, guidelines for antitrust authorities are cautious in considering the effects of mergers on input markets based on their possible effects on prices and quantities in those markets, even when likely effects on the downstream product markets are minor or absent. For example, EU guidelines suggest that the evaluation by antitrust authorities of a merger must be based on its possible effects in the final product market – see e.g. European Commission (2004[147]). In other jurisdictions, merger guidelines are more explicit on the fact mergers between input purchases should be mainly assessed by examining their impact in the relevant input market. For example, the US horizontal merger guidelines explicitly indicate that enforcing agencies should not “evaluate the competitive effects of mergers between competing buyers strictly, or even primarily, on the basis of effects in the downstream markets in which the merging firms sell” (US Department of Justice and Federal Trade Commission, 2010, p. 33[106]). US enforcement authorities and US courts have applied this principle in a few investigations of mergers of buyers of agricultural products,72 but cases of mergers with suspected monopsonistic effects in the labour market have rarely been investigated – see e.g. Hovenkamp (forthcoming[148]) and Naidu, Posner and Weyl (forthcoming[78]).73

One difficulty in assessing the impact of mergers on buyer power in the labour market has to do with the shortage of specific tools to analyse labour competition and, in particular, the difficulty of identifying the relevant market. This stands in stark contrast with the tools available to examine competition in specific output markets. While the recent literature has suggested a few practical methods – see e.g. Naidu, Posner and Weyl (forthcoming[78]), more research in this area and more investment in developing adequate tools by governments and enforcement authorities would be welcome.

Another difficulty in assessing the impact of mergers on buyer power in the labour market is the evaluation of their effects in the downstream markets when merging firms are not direct competitors in those downstream markets – see e.g. Marinescu and Hovenkamp (forthcoming[87]) for a few hypothetical examples. Certain scholars have suggested that a merger increasing concentration in the labour market may be harming the final consumer even if concentration in downstream markets does not pass the threshold level suggesting anti-competitive effects, although the size of the effect is hardly identifiable. For this reason, Krueger and Posner (2018[123]) suggest to introduce in the legislation a rebuttable presumption of anti-competitive effects on consumer welfare when a merger increases concentration in the labour market above a given threshold level. A connected debate – more philosophical but loaded with many implications in this area – turns around whether worker welfare should also be included in the definition of “consumer welfare” (which typically guides antitrust authorities) and, if not, to what extent enforcement authorities should also consider other welfare losses beyond those affecting the final consumer – see e.g. Bork (1993[149]); Schwartz (2004[150]); Werden (2007[151]); Salop (2010[152]); Jacobson (2013[153]); Sims (2018[154]); Chopra (2018[155]) and Hovenkamp (forthcoming[148]).74

Finally, effectively addressing other types of monopsonistic conduct in the labour market, such as predatory price-bidding,75 might be more difficult, in particular if courts and enforcement authorities adopt cautious behaviours to avoid over-deterrence – that is inefficient deterrence of conduct whose anti-competitive effects are offset by efficiency gains. In this case, antitrust authorities and courts may apply very strict tests that may fail in the presence of both monopsony in the upstream market and oligopoly in the downstream market. For example, if a predator in the input market has monopoly power in the output market, it may be able to sell its output at a price higher than its costs despite paying more than the marginal revenue for the input market in which it is pricing predatorily. In this case, the predator is making temporary smaller gains but no losses. Requiring temporary losses as evidence of predatory price-bidding76 may fall short of identifying the most important cases – see e.g. Harrison (2012[156]). More research in this area is needed.

### Informational asymmetries

One of the reasons why workers often have low bargaining power and few outside options is that they have much less information than employers. Workers often have only a vague idea of their rights and are uncertain about what they can legitimately ask for, in particular if they are employed with a non-standard contract on which information may be less abundant. For example, when signing a contract (or accepting terms and conditions) workers may not fully understand the obligations they are committing to (or the rights they are waiving). Even after signing the contract or accepting a job offer, and therefore, in some cases, having turned down alternative offers, workers may be pressed for signing covenants giving away specific rights without a clear understanding of their implications. For example, based on a survey of over 1 000 US engineers, Marx (2011[135]) finds that over two-thirds of those who signed a non-compete clause were asked to do so only after having accepted the job offer, and sometimes after having started working.

Regulation could therefore make sure that all contracting parties are fully aware of their rights and responsibilities. In the United Kingdom, the government’s response to the Taylor Review (“Good Work”) (BEIS, 2017[44]) includes plans to ensure that all workers receive information from day one on their working relationship, and what rights they are entitled to (OECD, 2019[2]). In the same line, a regulation proposal under discussion at the European parliament provides that terms and conditions for the use of intermediation services should be drafted in clear and unambiguous language which is easily understood by an average user (European Commission, 2018[67]). Similarly, in Oregon, regulations stipulate that non-compete clauses should be made clear before negotiating compensation, so that giving up future options could be traded off against pay (Marx and Fleming, 2012[157]). And the recently adopted Fair Work Week Act in the same state stipulates that large corporations in selected sectors (retail, food and hospitality) must provide a new employee with a written good faith estimate of the employee’s work schedule at the time of hire. Certain scholars have suggested, more generally, that governments should invest in a dedicated service for independent contractors, and in particular gig workers, offering general advice and counsel about their employment rights (Balaram, Warden and Wallace-Stephens, 2017[89]).

Lack of pay transparency may also increase monopsony power. Shortage of information about alternative opportunities reduces workers’ ability to change jobs or leverage outside opportunities to negotiate for higher pay (and better working conditions) – see e.g. Harris (2018[90]). Digital technologies have the potential to improve this type of information asymmetry, as workers can access a large number of job offers and compare them, thereby reducing search costs. Yet, in many platforms, workers have few tools to search for available alternatives and have to spend much time searching for them (Kingsley, Gray and Suri, 2015[52]; Berg et al., 2018[62]).

To improve pay transparency in the platform economy, employers and platforms could be required to publish information about the average pay per task, as well as on the average time taken to complete a task (which would help workers make more informed decisions about which tasks to accept).77 In addition, tools could be designed to help workers find requesters who pay well, and avoid those who pay poorly. However, the devil is in the details: publishing information on pay policies without making it easily accessible by workers could facilitate collusion among requesters, even in the absence of an explicit agreement among them (see above in this section). In practice, some websites and online communities have already sprung up which help workers find well-paying requesters. For example, the social networks Turkopticon and TurkerView allow Turkers (crowdworkers offering work on Amazon Mechanical Turk) to help one another with information and share experiences about their employers (Irani and Silberman, 2013[158]).

One specific problem related to digital intermediation concerns the information that platforms collect on workers. For example, most platforms have a rating system for workers, which evaluates their performance history and is supposed to improve service quality for the requester. However, the ratings algorithms, which are based on inputs from the requesters or on other productivity parameters (such as task acceptance rates and average delivery lags), often lack transparency (Rosenblat and Stark, 2016[91]) and workers sometimes have no clear information on how to improve their ratings (Berg et al., 2018[62]). In addition, certain platforms do not allow workers to rate requesters and/or to ask for the reason of a bad evaluation. This informational asymmetry allows requesters or platforms to exclude certain workers from certain tasks, based on their ratings, while workers are unable to identify and refuse tasks proposed by bad requesters, who typically do not pay regularly or are inclined to provide negative assessments, which results in lower average pay and heightened worker stress (Kingsley, Gray and Suri, 2015[52]). Even though some collective workers’ initiatives such as Turkopticon or FairCrowdWork.org have been recently developed to, inter alia, deal with this problem, there is probably scope for regulating the ratings system by imposing rating symmetry (the so-called “five-for-five” policies, in which workers and employers rate each other on the same scale) and transparency in algorithms on platforms.

The information accumulated by platforms could also give rise to exclusionary practices that are illicit under either anti-discrimination or competition law. For example, while independent contractors should be able to work for different platforms, the latter, by digitally tracking the former, can punish those who also work for the competitor, or alternatively engage in targeted predatory price-bidding by proposing different fees or compensation packages to these workers, thereby potentially excluding competitors from the labour market – see e.g. OECD (2018[159]). To the extent that pay, rating and task allocation algorithms are not transparent, these illicit behaviours might be extremely difficult to identify by enforcement authorities (Herrera Anchustegui and Nowag, 2017[160]).

## Notes

← 1. As well as access to training programmes, freedom of association and the right to bargain collectively.

← 2. A similar guide for tax purposes is available on the website of the US Internal Revenue Service (Internal Revenue Service, 2005[10]).

← 3. Some countries, like France (présomption de salariat) and the Netherlands, presume an employment contract when a number of criteria are met. In the Netherlands, for example, if a worker has worked on a regular basis for his or her employer for a period of three months (or at least 20 hours a month), then a contract of employment is automatically presumed (Davidov, Freedland and Countouris, 2015[6]). In Belgium, there is a presumption of an employment contract (subject to certain criteria) in certain “at-risk” sectors including caretaking/security, construction, transport, cleaning, agriculture and horticulture. In Ontario (Canada), the Employment Standards Act was amended in 2017 to put the burden of proof on the employer in cases where a contractor claims to be an employee. See also the online annex to OECD (2014[8]) for other examples of presumption of an employment relationship.

← 4. In the United Kingdom, for example, the introduction in July 2013 of launch and hearing fees of GBP 1 200 for an individual bringing a claim to the employment tribunal has led to a drop in claims of over 70%, which has affected disproportionately the bottom end of the claim distribution (Adams and Prassl, 2018[161]).

← 5. Portugal introduced a new, simplified judicial procedure to target the growth of bogus self-employment through changes in 2013 and 2017. It provides workers with a speedier court decision recognising the existence of an employment relationship.

← 6. Epic Systems Corp. v. Lewis, 138 S. Ct. 1612 (2018).

← 7. That being said, practice by judges will vary across countries and time. In some countries, for example, courts have tended to pay more attention to what is written in the employment contract (Davidov, Freedland and Countouris, 2015[6]).

← 8. Differences across countries tend to stem primarily from differences in practice rather than differences in theory (Davidov, Freedland and Countouris, 2015[6]).

← 9. In most cases, the decision of a tribunal will apply only to the plaintiff. To some extent, this will act as a deterrent to firms because it could create a precedent which would have an impact on similar cases in the future. However, the deterrent might remain limited if the outcome of the tribunal only applies to the few workers who brought the case in the first place (especially because many of these cases settle). The deterrent would be much greater if, once a case is decided, people in the same situation can make a liability claim that would automatically apply.

← 10. A complementary strategy consists in providing incentives to employers who have misclassified workers to seek settlement with the administration in exchange of a partial relief on penalties. In the United States for example, the Voluntary Classification Settlement Program (VCSP) provides substantial partial relief from federal employment taxes for eligible taxpayers who agree to treat workers prospectively as employees. In exchange, the employer pays 10% of the employment tax liability that would have been due if those workers were classified as employees in the most recent year. Moreover the employer is not liable for any interest and penalties on the liability and is not subject to an employment tax audit with respect to the worker classification of the class or classes of workers for prior years (Internal Revenue Service, 2012[167]).

← 11. There may be broader issues here than false self-employment. Where incentives are so strong as to lead to an “inefficiently high” level of self-employment, this could lead to a misallocation of labour resources as well as undermining social protection systems (particularly where lower-risk individuals choose to become self-employed and are allowed to opt out of parts of the social protection system, leaving only the “bad risks” behind and resulting in an increased reliance on social assistance, see also Chapter 7). In addition, to the extent that the self-employed participate less in training, very high levels of self-employment could act as a drag on productivity.

← 12. However, the problem with amnesties is that they create an expectation of further amnesties and therefore can encourage future deviations, except if they are introduced simultaneously with a real change in legislation that makes the reduction of sham arrangements more durable.

← 13. See e.g. Berwick v. Uber Technologies, Inc., CGC-15-546378, Cotter v. Lyft, Inc., Dkt. No. 94, 60 F.   Supp. 3d 1067 (N.D. Cal. 2015); Dynamex Operations West, Inc. v. Superior Court, 4 Cal.5th 903 (2018); O'Connor v. Uber Technologies Inc., Case No. 14-16078 (9th Cir. 2018); Uber B.V. & Ors v Aslam & Ors UKEAT/0056/17/DA; Kaseris v Rasier Pacific V.O.F [2017] FWC 6610; or Cass. soc., 28 novembre 2018, n° 17-20.079.

← 14. Cotter v. Lyft, Inc., Dkt. No. 94, 60 F. Supp. 3d 1067 (N.D. Cal. 2015).

← 15. NLRB v. Hearst Publ’ns, Inc., 322 U.S. 111, 121 (1944).

← 16. The employer generally drafts the contract and can use this to deny employment rights to individuals who would otherwise be employees (Davies, 2009[169]).

← 17. If countries have several employment tests and/or definitions, which vary across legal/policy areas (labour, tax, social protection), then there may be a case for harmonising them, as this could contribute to possible confusion. Even small differences across tests can result in very different results (Rubinstein, 2012[20]). For example, in the United States, the tests for defining employee status range from the broadest “suffer or permit” test, passing through the hybrid and the “ABC” tests, to the narrowest common law test (Waas et al., 2017[7]).

← 18. As introduced by Act 92/2012.

← 19. For a while, Germany also had a definition of employee for social security reasons which consisted of four criteria. If all four criteria were fulfilled, the person was classified as an employee; if none were fulfilled then the person was self-employed; and if two out of four criteria were met, the person would be presumed to be an employee for social protection purposes (Wank, 1999[168]).

← 20. Lord Wedderburn described the test used by courts to identify employees as an “elephant test” – “an animal too difficult to define but easy to recognise when you see it” (Davidov, 2002[163]).

← 21. Allowing for court discretion in the identification of employment relationships risks leading to results which the legislature may not appreciate. An interesting historical example is the Taft-Hartley Act of 1947, in which the US legislature reacted to a series of Supreme Court judgments from 1944 through 1947, which adopted an economic reality of dependence test instead of the common-law agency (control) test to ascertain employee status (Davidov, Freedland and Countouris, 2015[6]). This resulted in far more workers being classified as employees than congress liked, and the definition of employee in the National Labor Relations Act was subsequently amended to explicitly exclude independent contractors.

← 22. McKee v. Reid’s Heritage Homes Ltd., (2009) ONCA 916.

← 23. Judgement of 28 October 2010, Aslam v. Uber, London Employment Tribunal: 2202550/2015, paragraph 97.

← 24. Or 2.2% of total employment.

← 25. The co-ordinated and continuous collaboration contract (collaborazioni coordinate e continuative – co.co.co.); and the project-based collaboration contract (contratto di collaborazione a progetto – co.co.pro., abolished since January 2016).

← 26. In practice, this would be equivalent to defining the third worker category as a “residual category” capturing those cases where employment tests fail to come to clear conclusions about employment status.

← 27. An alternative approach which has been proposed is more radical, but would perhaps also encounter greater practical difficulties. The idea put forward is to no longer tie labour protections to employment status, but rather apply them to anyone performing work (Freedland and Kountouris, 2011[164]; Johnstone et al., 2012[165]; Stewart and Stanford, 2017[166]; Linder, 1999[22]).

← 28. See Dynamex Operations West, Inc. v. Superior Court, 4 Cal.5th 903 (2018), in which the court stated that delivery drivers should be considered as employees if at least one of the factors in the ABC test does not hold, thereby reversing the standard common law approach.

← 29. NJA 1996 s 311.

← 30. Two interesting variations exist in the United Kingdom and in the Netherlands. In the United Kingdom, there is a certain margin around the national minimum wage to allow even slower workers to earn a “fair rate”. Employers are required to monitor the average number of tasks/pieces completed per hour worked and divide this by 1.2. The hourly minimum wage rate should then be divided by that number to estimate the fair rate for each piece of work completed. In the Netherlands, in cases where work is difficult to supervise and the employees have a certain degree of freedom to organise their work, employers can ask for an exemption to the national piece rate legislation and, instead, pay a sectorally-determined piece rate.

← 31. One option is to make all parties jointly and severally liable for making sure that the worker is paid a fair wage. For example, in the Netherlands, the Act on Combatting Sham Arrangements came into force in January 2016. The law states that, in case of multiple customers/contractors, the entire chain is liable to make the correct payment of the agreed wage (supply chain liability).

← 32. This is an area where technological developments and the rise of the platform economy could potentially offer a solution since they present unprecedented opportunities for data collection which, in turn, could be used to calculate average productivity. This would help overcome some of the challenges traditionally encountered with piece rate pay.

← 33. For example, should platform workers be paid for the time that they have an app open and/or the time they spend waiting/searching for tasks? While the answer may depend on the particular type of work considered, the general response would probably be no – otherwise workers could simply log on to several platforms at the same time and be paid for waiting on each one of them. The hourly pay of platform workers also depends on whether the time working on tasks that are rejected is taken into account (Hara et al., 2017[162]). In a well-functioning market, such time would be accounted for in the rates that gig and self-employed workers set, so in their calculation of minimum piece rates employers should probably factor those in as well. Similarly, the expenses incurred by platform workers when performing gigs should be compensated for.

← 34. In Poland, the parties to an agreement covered by the new minimum wage legislation may agree on the method of determining the number of hours worked by the service provider. If they do not do so, it is for the service provider to indicate the number of hours worked. However, this rule does not apply to service providers who are temporary work agency workers, in which case, the number of hours worked is determined by the user undertaking.

← 35. In some cases, where there is a specific risk of injury to the worker or to others, working time regulations have already been extended to self-employed workers. For example, in accord with the EC Road Transport Directive, self-employed drivers are subject to the 48-hour average working week and other provisions of working time regulations.

← 36. These terms purport to govern issues such as how and when crowd workers will be paid, how work will be evaluated and what recourse workers have (or do not have) when things go wrong.

← 37. In turn, this asymmetry in the control of the relationship can enhance monopsony power by the platform or the requesters (see the next section).

← 38. This is standard practice in the case of the dismissal of an employee in many countries – see e.g. OECD (2013[61]).

← 39. O’Connor v. Uber Technologies, Inc.

← 40. Formal arbitration is also not necessarily shorter and less expensive than resolution by court.

← 41. See e.g. In re National Arbitration Forum Trade Practices Litig, 704 F. Supp. 2d 832 (D. Minn. 2010). See also Virtualpoint Inc. v. Poarch Band of Creek Indians, C.D. Cal., No. 8:15-cv-02025, (2016).

← 42. The Ombuds office consists of a board of five people – one worker, one trade union representative, one platform employee, one representative of the Crowdsourcing Association, and a neutral chair – and resolves disputes by consensus.

← 43. Furthermore, an open question remains on how to weigh the effects in labour and downstream product markets, in the cases where they are of opposite sign – see the discussion in Annex 4.A.

← 44. This is neither new nor platform-specific. In fact, it applies in principle to any triangular relationship between a client, an intermediary and an independent contractor, if the intermediary is not committed to maintain the same conditions of service but the client and the independent contractor commit on refraining from engaging in a direct relationship.

← 45. When there is an employment relationship, Article 6 of the Rome Convention of 1980 allows parties to choose the applicable law and, in the absence of such choice, “the contract shall be governed by the law of the place in which the employee habitually carries out his work in performance of the contract or, in the absence of such a place, the law of the place in which he was engaged”.

← 47. Most studies typically estimate residual labour supply elasticities for employees comprised between 0.5 and 5 – see e.g. Sokolova and Sorensen (2018[77]) and Naidu, Posner and Weyl (forthcoming[78]).

← 48. By contrast, a competitive market would require a residual labour supply elasticity close to infinity.

← 49. Azar et al. (2018[74]) find that the average level of concentration (the Herfindahl-Hirschman Index) in a market defined as a commuting zone by SOC-6 occupation by quarter is 3953. This figure is well above the 2500 threshold triggering concern of lack of competition that is mentioned in the horizontal merger guidelines of the US antitrust authorities – Department of Justice, Antitrust Division, and Federal Trade Commission (US Department of Justice and Federal Trade Commission, 2010[106]). Certain scholars even suggest that, due to search frictions, in the absence of worker bargaining power, negative welfare effects of buyer power arise at lower levels of concentration than in the case of product market power – see e.g. Carstensen (2012[171]), but this position is far from being consensual in the literature – see e.g. Schwartz (2004[150]).

← 50. On trends in price-cost margins, see also Calligaris, Criscuolo and Marcolin (2018[170]).

← 51. See e.g. California v. eBay, Inc., case No. 5:12cv05874, Document 85 (N.D. Cal. 2015) concerning an alleged handshake agreement entered into between senior executives of two companies specialised in different markets (accounting and tax preparation software and online auctions) but competing for a similar workforce.

← 52. For example, the European Court of Justice has condemned a wage-fixing agreement among Dutch telecommunications operators while underlining the lack of evidence of harm to the final consumer: “in order to find that a concerted practice has an anti-competitive object, there does not need to be a direct link between that practice and consumer prices.” Case C-8/08 T-Mobile Netherlands BV v Raad van Bestuur van de Nederlandse Mededingingsautoriteit [2009] ECR I-4529, [39].

← 53. See e.g. Your Therapy Source, LLC; Neeraj Jindal; and Sheri Yarbray , FTC File No. 171 – 0134, concerning therapist staffing companies colluding on rates paid to physical therapists in the Dallas/Fort Worth area (the case was concluded with a settlement in July 2018); and Beltran v. Noonan et  al, Civil Action No. 14-cv-03074, U.S. District Court D. Colo., concerning sponsors agencies conspiring to fix compensation for au pairs throughout the United States (a settlement on this case was agreed in January 2019).

← 54. See e.g. Cass. soc., 2 mars 2011, n° 09-40.547, concerning the no-poaching agreement between two information technology companies in France; Cass., com., 13 février 2001, 98-21.078, concerning a wage-fixing agreement among French temporary work agencies; JAR 2010/145, Gerechtshof's-Hertogenbosch 04-05-2010, BM3366, HD-200.056.33, concerning human resource agreements among a few Dutch hospitals with the purpose of restricting competition for practitioners, including independent contractors; Expediente S/0120/08, concerning a no-poach agreement among eight Spanish companies in the road freight transport industry; and Kores Manufacturing Co Ltd v Kolok Manufacturing Co Ltd [1959] Ch.108, concerning an agreement among two British competitors that required each other’s consent in case of one company recruiting the other’s employee. See also endnote 52.

← 55. Certain scholars argue that anti-collusion legislation may however be too rigid, thereby backfiring on both wages and employment, since companies may be reluctant to disclose their wage rules even to their own employees. The argument is that the asymmetry of information between employers and workers might increase monopsony power – see e.g. Harris (2018[90]) and the discussion of information asymmetries below in this section.

← 56. See for example Korean Fair Trade Commission (2009[172]) and European Commission (2011[182]).

← 57. See e.g. Yi v. SK Bakeries, LLC et al, U.S. District Court W.D. Wash., case number 3:18-cv-05627; Maurella v. H&R Block Inc. et  al., U.S. District Court N.D. Ill., case number 1:18-cv-07435; and The State of Washington v. Jersey Mike Franchise System Inc. et  al., KCSC, Wash., case number 18-2-25822-7). The attorney general of the US state of Washington is particularly aggressive in fighting no-poaching agreements among franchisees (Office of the Washington State Attorney General, 2018[179]).

← 58. Moonlighting covenants are usually enforced in most countries. Nonetheless, they can be particularly problematic in contracts that do not guarantee a minimum of hours of work, such as zero-hours contracts.

← 59. California, which bans non-compete clauses altogether, represents the best known exception (Section 16600 of the California Business and Professions Code). Similar provisions exist in few other US states, such as North Dakota, and Oklahoma (Gomulkiewicz, 2015[174]; Green, 2015[175]), and Mexico (Association of Corporate Counsel, 2018[173]). Michigan also used to have similar legislation but it was repealed in 1985 (Pynnonen, 1994[180]).

← 60. See for example Shores v. Global Experience Specialists, Inc., 134 Nev. Adv. Op. 61 (Aug. 2, 2018), concerning a sale agent at an event management company who signed non-compete covenant preventing him from taking a similar job in any US state, including those in which the company had no affiliates.

← 61. A famous case involved a fast-food company preventing its workers from taking up jobs in any capacity within two miles of any of its stores at any employer “who derived at least ten percent of revenues business from the sale of categories of products such as ‘deli-style’ sandwiches, for two years [and] anywhere in the United States.” (People v. Jimmy John’s Franchises LLC, Circuit Court of Cook County, 2016-CH-07746). The company agreed to a settlement and suppressed the clause (Office of the Attorney General of the State of Illinois, 2018[178]).

← 62. More precisely, the 2018 Massachusetts Noncompetition Agreement Act provides that non-compete covenants are enforceable only on “exempt” employees under the Fair Labor Standards Act, which means that they must earn more than a given threshold, have guaranteed minimum pay independently of their effectively-worked hours, and have executive, professional or high-level administrative duties, with few exceptions for specific jobs.

← 63. Art. 1 § 36 Angestelltengesetz (AngG – Employees Act), for Austria, art. 65 §2 of the Loi du 3 juillet 1978 for Belgium, and art. L-125-8 Code du Travail, for Luxembourg. In Luxembourg, the labour code in principle only regulates clauses that prevent employees from starting a new business competing with that of their previous employer. However, an appeal court ruling effectively extended this regulation to other types of non-compete clauses in 2014 – Cour d’Appel de Luxembourg, arrêt du 13 novembre 2014 (n°39706 du rôle).

← 64. The official explanatory notes accompanying the reform state that the Act “provides that a provision in a zero-hours contract which prohibits the worker from doing work under any other arrangement is unenforceable”, without explaining whether work is provided during or after the contract is terminated (http://www.legislation.gov.uk/ukpga/2015/26/notes/division/5/11 note 844). Yet, it remains to be seen how courts will interpret this provision in practice.

← 65. For example, the 2018 Massachusetts Noncompetition Agreement Act explicitly defines non-compete clauses as agreements between an employer and an employee, for the purpose of the act. Standard common law principle remain applicable to independent contractors.

← 66. See also the discussion in Kolani v. Gluska, 75 Cal. Rptr. 2d 257 – Cal: Court of Appeal, 2nd Appellate Dist., 7th Div. 1998.

← 67. As done, for example, by the offices of the attorney general of the US states of Illinois and New York (Office of the Attorney General of the State of Illinois, 2018[178]; New York State Attorney General, 2018[177]).

← 68. See e.g. Cour d’Appel de Luxembourg, 6 avril 2017, n°39706 du role.

← 69. See e.g. Egon Zehnder Ltd v Tillman [2017] EWHC 1278 (Ch).

← 70. See e.g. People v. Check into Cash of Illinois LLC, Circuit Court of Cook County, 2017-CH-14224; People v. Jimmy John’s Franchises LLC, Circuit Court of Cook County, 2016-CH-07746; as well as Illinois Attorney General (2018[181]), New York State Attorney General (2018[176]) and the cases cited there.

← 71. As shown by the low estimated values of the residual labour supply elasticity – see the previous section in this annex. When workers can bargain collectively, possible welfare effects will depend on the capacity of bargaining parties to achieve the outcome that would be pursued by a vertically-integrated firm in the downstream market – an outcome that may depend on a number of institutional settings – see e.g. Inderst and Shaffer (2008[145]) and OECD (2017[118]; 2018[101])

← 72. See e.g. United States v. Cargill Inc. and Continental Grain Co., U.S. District Court D.D.C., case number 99-1875 (GK), final judgement 29 July 2000. The merger of two major grain traders was refused despite the absence of allegations of harm to the final consumer, since grain prices to national processors are determined in world markets.

← 73. The labour market for health practitioners represents the most notable exception – see e.g. United States of America and the State of Texas v. Aetna, Inc. and The Prudential Insurance Company of America 1999 WL 1419046 (N.D. Texas); and United States v. Anthem, Inc., 855 F.3d 345, 356 (D.C. Cir. 2017).

← 74. The difference between the U.S. and EC horizontal merger guidelines (see above) could also result of different answers given by legislators to these questions (US Department of Justice and Federal Trade Commission, 2010[106]; European Commission, 2004[147]). See also the discussion above concerning standards for fighting collusion and the literature cited there.

← 75. Predatory bidding occurs when the dominant buyer temporarily raises the price of inputs to exclude weaker competitors and then recover its costs by lowering it.

← 76. See for example the doctrine imposed by the US Supreme Court in Weyerhaeuser Company v. Ross-Simmons Hardwood Lumber Company, 549 U.S. 312 (2007), as regards predatory price-bidding.

← 77. While not directly related to the platform economy, the existing piece rate legislation in the United Kingdom already requires that a notice be issued before the start of the pay reference period, with a clear reference to the expected mean hourly output rate of the task in question.

# Disclaimer

This work is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and arguments employed herein do not necessarily reflect the official views of OECD member countries.

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Photo credits: Cover © Jacob Lund, Dragon Images, PaO_STUDIO, Daisy Daisy /Shutterstock.com.

Corrigenda to publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.