Abandoning individual enforcement? Interrogating the enforcement of age discrimination law

Abstract Discrimination law primarily relies on individual enforcement for addressing discrimination at work; yet those who are most impacted by discrimination are likely the least able to enforce their rights. The question then becomes: what role should individual enforcement play in discrimination law? Can we effectively abandon individual enforcement as part of the legislative model? Drawing on a mixed method, multi-year comparative study of the enforcement of age discrimination law in the UK, Australia and Sweden, this paper considers the gaps, limits and risks of the individual enforcement model in discrimination law. Integrating doctrinal analysis; statistical analysis of claims and cases, and data from the EU and OECD; qualitative expert interviews; and a survey of legal practitioners, this paper argues that while individual enforcement is inherently limited as a tool for achieving systemic change, it must remain part of any legislative model. Reflecting on the experience in Sweden, where individual enforcement of discrimination law is significantly curtailed, the paper posits that individual rights and individual enforcement remain important complements to other regulatory tools, particularly in jurisdictions with strong enduring age norms. Abandoning or severely restricting individual enforcement is unlikely to support either the macro or micro effectiveness of age discrimination law.


Introduction
Discrimination law in the UK primarily relies on individual enforcement for addressing discrimination at work. Scholars and practitioners have repeatedly emphasised the gaps and limits of relying on individual enforcement to address discrimination. 1 Those who are most impacted by discrimination are often the least likely to enforce their rights. 2 The question then becomes: what role should individual enforcement play in discrimination law? Can we effectively abandon individual enforcement as part of the legislative model?
Drawing on a mixed method, multi-year comparative study of the enforcement of age discrimination law in the UK, Australia and Sweden, in this paper I consider the gaps, limits and risks of the individual enforcement model in discrimination law. Integrating statistical analysis of claims and cases, including 1208 reported UK Employment Tribunal decisions, and 108 Australian cases; 72 qualitative expert interviews with 105 expert respondents in Australia, the UK and Sweden; 3 and a survey of 76 legal practitioners and advocates in the UK and Australia, 4 I argue that while individual enforcement is inherently limited as a tool for achieving systemic change, it must remain part of any legislative model for addressing discrimination.
In Part 1, I map problems with individual enforcement, drawing on my empirical study of enforcement in Australia and the UK. 5 In Part 2, I map the legislative framework in Sweden, where individual enforcement of discrimination law is significantly curtailed, and enforcement instead emphasises the role of trade unions, negotiation and statutory agencies. In Part 3, I compare the macro and micro success of the Swedish model of enforcement with that in the UK and Australia, drawing on OECD and EU statistics on experiences of older workers and discrimination. These statistics illustrate that neither system is ideal; there is scope for significant mutual learning across jurisdictions. Considering gaps in enforcement in Sweden, I argue that individual rights are an important complement to other regulatory tools, particularly in jurisdictions with enduring age norms. We cannot abandon individual enforcement. I therefore put forward reforms to strengthen individual enforcement; and consider the future of the Swedish regulatory model in a concluding section.
1. The problem with individual enforcement: the UK and Australia Scholars and practitioners are well versed in the difficulties of individual enforcement of discrimination law. 6 At a theoretical level, individual enforcement models reflect an 'individualistic vision' 7 of equality, protecting individuals from unfair treatment on the basis of protected characteristics. 8 This ignores or minimises the societal, institutional, communal and collective impacts of inequality. 9 It also prioritises a view of equality as a negative 'duty of restraint', 10 rather than something that should be proactively pursued to prevent harm occurring. Thus, individual approaches to addressing discrimination are unlikely to achieve meaningful systemic change. 11 In practice, for individuals to use legal mechanisms to address discrimination, they must undergo the process of 'naming, blaming and claiming' which is required for a dispute to arise. 12 Those who experience discrimination: may not recognise it as discrimination; may not be able to identify who is to 'blame' for the discrimination; or may be reluctant to make a claim using legal mechanisms. As a 2 Smith (2008), above n 1, at 132. See also S Wakefield and C Uggen 'The declining significance of race in federal civil rights law: the social structure of employment discrimination claims' (2004) 74 Sociological Inquiry 128 at 150. 3 Interviewees have been classified according to their jurisdiction: A for Australia, UK for the UK, Sw for Sweden, then allocated a number. 4 These methods are described in detail in A Blackham, Reforming Age Discrimination Law: Beyond Individual Enforcement (Oxford University Press, 2022) ch 1. 5 This section summarises the findings reported at length in Blackham,ibid. 6 See eg Gaze, above n 1, at 919. result, then, few individuals actually use available legal mechanisms: 13 age discrimination complaints in both Australia and the UK are rare, despite the reported prevalence of discriminatory treatment. 14 I have estimated that statutory agencies receive less than 0.09% (in Australia) or 0.08% (in the UK) of possible age discrimination complaints from older workers. 15 In the Australian Human Rights Commission's (AHRC) National Prevalence Survey, for example, of those who had experienced age discrimination in the last two years (n=309), 43% took no action; the least common way to take action was to discuss or raise the issue with an external organisation (5%)this was less common than leaving the job in question (6%). 16 There are many factors that might explain this reluctance to engage with legal complaints mechanisms. Those who are impacted by discrimination rarely fit the stereotype of the rational, utilitymaximising individual with the agency required to bring a claim; 17 instead, those most affected by discrimination are often the most vulnerable at work, lacking both rights awareness and the capacity to pursue their legal rights and entitlements. 18 As Smith concisely summarises: Anti-discrimination legislation is designed to protect disempowered groupsthose who traditionally experience marginalisation and exclusion. Expecting members of such groups to have the time, security and resources to pursue legal action in order to gain compensation and possibly bring about wider change represents a fundamental regulatory weakness. 19 For these individuals to undergo the process of 'naming, blaming and claiming' 20 would be unusual. The process of claiming is individually costly in terms of time, emotional energy, financial costs, social norms against claiming, and fear of retribution. 21 Further, claiming is made difficult by limited access to information and legal advice, legal complexity, the need for proof of discrimination, information asymmetries and power disparities between the parties, 22 restrictive time limits, 23 low perceived prospects of success, fear of adverse costs orders, and a legislative focus on individualised and retrospective remedies. 24 Those who experience discrimination are 'one-shotters', who rarely use legal mechanisms, whereas employers and respondents tend to be 'repeat players', who typically come 13 In the UK, in the 2008 Fair Treatment at Work survey, around 3% of respondents made an application to an Employment Tribunal (ET): R Fevre et al 'Fair treatment at work report: findings from the 2008 Survey' (2009) Employment Relations Research Series 103 at 136. In the 2004 Unrepresented Worker Survey, only 2.4% of respondents began ET proceedings: A Pollert and A Charlwood 'The vulnerable worker in Britain and problems at work' (2009) 23 Work, Employment and Society 343 at 351. These surveys were conducted before the introduction and subsequent abolition of ET fees. 14 In Australia see AHRC 'Willing to work: national inquiry into employment discrimination against older Australians and Australians with disability' (2016) p 321. In Great Britain see House of Commons 'Oral evidence: older people and employment' (2018) HC 359 Q43, available at https://web.archive.org/web/20230131003214/https://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/women-and-equalities-committee/older-people-and-employment/oral/ 76656.html (accessed 31 January 2023). 15 Blackham, above n 4, p 136. 16 AHRC 'National prevalence survey of age discrimination in the workplace: the prevalence, nature and impact of workplace age discrimination amongst the Australian population aged 50 years and older' (2015 Fredman, above n 10, at 370. out ahead. 25 These barriers are similar across both Australia and the UK. 26 Understandably, then, those impacted by discrimination may be more likely to 'lump it' or leave, rather than to make a formal claim or complaint, 27 especially if they wish to repair social relationships, or avoid future risk of hurt or harm. 28 Even if a claim is made, both Australia and the UK rely heavily on confidential conciliation to resolve individual discrimination complaints, especially as they relate to work. 29 This means most discrimination matters are resolved privately, behind closed doors, and often with a non-disclosure agreement. This has a number of systemic impacts, prompting a culture of secrecy and stigma around discrimination claims. 30 Further, it makes it difficult to identify 'lightning rod' cases that can change public opinion and effect broader change. 31 Instead, discrimination complaints tend to be resolved quietly, with small amounts of financial compensation, effecting limited change to individual or organisational behaviour.
Overall, then, the UK House of Commons Women and Equalities Committee has concluded that individual enforcement 'is not fit for purpose'. 32 Individual enforcement may undermine the transformative potential of equality law 33 and is ineffective for addressing age discrimination. 34 There is a compelling case, then, to think about how we might move beyond individual enforcement; or, indeed, abandon it entirely as a regulatory tool. It is here that the experience of Sweden is particularly illuminating.

The Swedish context
Like the UK, Swedish discrimination law has largely been shaped by EU Directives. However, unlike the UK, Sweden has an enduring tradition of collective bargaining and negotiation with the social partners. The Swedish model has been described as 'corporatist', that is, the 'concertation of economic and social policies amongst interest associations and state actors'. 35 The enforcement of discrimination law has been strongly shaped by this context; as a result, while both the UK and Sweden have adopted similar legislative frameworks to address age discrimination, individual enforcement has been effectively minimised in the Swedish context. For Carlson, then, The Swedish social partners, the employers, their organizations and the labor unions, keep a fiercely tenacious grip on controlling all labor market issues including discrimination protections, viewing employment discrimination claims as a collective labor, rather than a human rights, issue with no need for individual access to justice mechanisms. 36 This stands in stark contrast to the 'privatised' enforcement model in the UK. It is these national differences that make the comparison between the UK and Sweden particularly apt 37 and supportive of mutual learning. 38 (a) The legal framework Equality is embedded in Swedish constitutional law. Under the Instrument of Government Chapter 1, s 2, there is a non-binding policy provision that: 39 The public institutions shall promote the opportunity for all to attain participation and equality in society and for the rights of the child to be safeguarded. The public institutions shall combat discrimination of persons on grounds of … age.
Chapter 2, Article 12 also affords protection against discrimination: No act of law or other provision may imply the unfavourable treatment of anyone because they belong to a minority group by reason of ethnic origin, colour, or other similar circumstances or on account of their sexual orientation. 40 These constitutional protections are embedded in the Discrimination Act, 41 which is based, in part, on the Framework Directive. 42 The Act prohibits direct and indirect age discrimination in employment, 43 with exceptions for genuine and determining occupational requirements; and different treatment on the basis of age with a legitimate purpose and using appropriate and necessary means. 44 This closely reflects the Framework Directive.
Strong age norms remain in the Swedish labour market. 45 In relation to retirement, the so-called '67-year-rule' (as it then was) provides protection for employees to remain at work This analysis uses official and unofficial English translations of Swedish legislation. In the case of more recent amendments to the Discrimination Act (those after 2017), I used Google translate, as there was no official or unofficial English translation available (as at July 2022). While this is far from desirable, I compared this analysis with secondary materials to ensure its accuracy, confirmed the translation with native Swedish speakers, and complemented my doctrinal analysis with qualitative expert interviews to address any potential cultural misunderstandings. 40 See also Art 13 regarding gender. until the age of 67. 46 Thereafter, employers have the freedom to terminate an employee's employment with one month's notice, 47 or to deploy fixed-term contracts. Seniority rules are also often in place for redundancy (that is, the last-in-first-out principle) 48 and wage setting. That said, even these age norms are shifting: under 2019 amendments to the Employment Protection Act (1982: 80), protection from forced retirement increased to age 68 from 1 January 2020, and to 69 in 2023.

(b) Enforcement
The Equality Ombudsman (Diskrimineringsombudsmannen, DO) is responsible for supervising compliance with the Discrimination Act. 49 The DO may bring a court action on behalf of an individual with their consent, 50 as may a non-profit organisation. 51 The DO may only bring an action if an employees' organisation does not do so. 52 Disputes brought by an employers' or employees' organisation, or the DO, are heard in the Swedish Labour Court as the court of first instance. 53 Individuals may only bring a complaint to the District Court; 54 in these cases, the Labour Court is the superior appellate jurisdiction. 55 Cases brought by a trade union are subject to mandatory negotiation before proceeding to the Labour Court. 56 Someone who breaches the Discrimination Act is liable to pay compensation to the person affected. 57 This is not solely compensatory: according to Ch 5, s 1 of the Discrimination Act, when determining compensation, particular attention should be given to the aim of discouraging or counteracting infringements of the Act. In the context of work, an employer is also responsible for paying compensation for any loss, so long as that loss does not relate to recruitment or promotion. 58 If there are special grounds or reasons to do so, compensation can be reduced or set at zero. 59 Discriminatory provisions of individual contracts and collective agreements can be modified or declared invalid. 60 Termination of a contract can also be declared invalid, and workplace rules can be modified or declared to be of no effect. 61 Time limits for bringing a claim are set by reference to time limits in employment law. 62 Consistent with the Employment Protection Act (1982:80) ss 40-42, for example, actions relating to termination or summary dismissal have a time limit of two weeks for a declaration of invalidity, and four months if seeking damages. For other actions, time limits are consistent with the Employment (Co-determination in the Workplace) Act (1976:580). The DO may 'toll' or interrupt the statute of 46 Swedish Employment Protection Act, s 32a; see the CJEU decision in Case C-141/11 Hörnfeldt v Posten Meddelande AB, where the objectives behind the 67-year rule were held to be legitimate. See also A Numhauser-Henning and M Rönnmar 'Compulsory retirement and age discriminationthe Swedish Hörnfeldt case put in perspective' in P Lindskoug et al (eds) Essays in Honour of Michael Bogdan (Juristförlaget, 2013). 47 Swedish Employment Protection Act, s 33. 48 Though this has been 'watered down' by other developments: Votinius, above n 45, pp 293-294. 49 Diskrimineringslag (2008:567) ch 4, s 1. 50 Ibid, ch 4, s 2; ch 6, s 2. 51 Ibid, ch 6, s 2. 52 Ibid, ch 6, s 2. 53 Labour Disputes (Judicial Procedure) Act (1974:371), ch 2, s 1. 54 Ibid, ch 2, s 2. 55 Ibid, ch 2, s 3. 56 Co-Determination Act, s 11, Labour Procedure Act (1974:371), ch 4, s 7. 57 Diskrimineringslag (2008:567), ch 5, s 1 58 Ibid, ch 5, s 1. 59 Ibid, ch 5, s 1. 60 Ibid, ch 5, s 3. 61 Ibid, ch 5, s 3. 62 Ibid, ch 6, s 4. limitations, such that a new statute of limitations starts to run. 63 Time limit rules are complex; this may deter claiming for unrepresented workers. 64 In the Swedish system, costs typically follow the event; 65 however, parties can be ordered to pay their own costs in the Labour Court 'if the losing party had reasonable cause to have the dispute tried'. 66 This provision is rarely used in practice. 67 It typically takes around 12 months from the time of application until judgment. 68 The Labour Court generally receives few cases: according to the Court's website, each year, the Court receives somewhere between 400 and 450 cases, though many cases are subsequently withdrawn. 69 The court generally issues somewhere between 150 and 160 judgments per year; few of these relate to age discrimination. 70 In some years, though, the number of cases may be even lower: in the 2019 financial year, for example, the Labour Court received 272 cases, which resulted in 57 decisions. 71 Only 5% of all disputes related to the Discrimination Act. 72 The Labour Court issued no judgments concerning discrimination in 2019, and only six discrimination judgments during 2018. 73 As at 28 July 2022, the DO had recorded only four age discrimination cases decided by the Labour Court, 74 and four age discrimination settlements relating to work. 75 That said, as in other jurisdictions, few cases do not necessarily mean few disputes: most disputes will be resolved via negotiation before proceeding to the Labour Court. 76 Indeed, a matter that proceeds to the Labour Court is seen as a failure of collective negotiation; 77 matters are typically resolved well before a formal claim is required: if 'you come from a culture, where going to court is always a bad option, then someone has already lost if you go to court'. 78 Case and claim statistics therefore do not fully represent the successor otherwiseof the legal framework; this point is considered in more detail below.
(c) Transparency and equality law Sweden is notable for its approach to transparency and confidentiality in the context of discrimination law. The Swedish Constitution consists of four fundamental laws, which take precedence over all other laws. 79 One of these, the Instrument of Government, creates a right to freedom of information when dealing with government bodies. According to Chapter 2, Article 1: 63 Ibid, ch 6, s 5. Everyone shall be guaranteed the following rights and freedoms in his or her relations with the public institutions: … freedom of information: that is, the freedom to procure and receive information and otherwise acquaint oneself with the utterances of others. This, then, creates a general right of openness and transparency in the workings of government bodies, including the DO. This means that all documents sent to and created by the DO are public records, which can be requested by the public, and which must be released unless they contain sensitive personal data. 80 Thus, all settlement agreements held by the DO are also public records, and potentially subject to disclosure. Non-disclosure agreements are not possible in this context. 81 Further, if the DO conducts an investigation, the respondent is provided with full access to the investigation file. 82 This stands in stark contrast to the focus on confidentiality in other jurisdictions such as the UK and Australia: 83 in Sweden, confidentiality is contrary to the national requirement of public transparency.
That said, the potential for disclosure in Sweden can be bypassed if the parties resolve their complaint without the DO's assistance or intervention; in that case, the requirement of public transparency would not apply if both parties were private entities. Where settlements are publicised, though, 'that's an important aspect … [we] consider it is [the] most important aspect rather than just the compensation'. 84

(d) Beyond individual enforcement
The Swedish experience is notable as there is far less emphasis on individual enforcement than in the UK and Australia. This recognises the very real limits of the individual enforcement model: there are other tools to deal with discrimination in Sweden, and 'often the other tools could be viewed as more efficient' than discrimination law. 85 As one respondent in this study noted, the expectation of private enforcement, as a method, does not account for the great difficulty of actually doing that work … I think there is a flaw to think that this could be solved by private enforcement, unless you have a system with very heavy damages which we will not have in [our] legal culture and not in most European countries either. 86 The Swedish legal framework is therefore, by design, structured to move beyond individual enforcement. Indeed, it is difficult (if not impossible) for individuals to bring a case without the support of a trade union or the DO. 87 The weight of enforcement, then, is not on individuals, but on collective and agency processes. This is consistent with the broader view of the enforcement of rights in Sweden: as one respondent in this study noted, there are three parallel tracks of enforcing. One is litigation, going to court, with or without support, or trade union or worker representatives, or authorities; but individual cases in courts, one way or another. And the other is public enforcement, where you have some labour inspector or 80 See Diskrimineringsombudsmannen Handling of personal data and the principle of public access to official documents, https://web.archive.org/web/20230131004348/https://www.do.se/choose-language/english/handling-of-personal-data-andthe-principle-of-public-access-to-official-documents (accessed 31 January 2023). 81 Sw2. 82 Sw2. 83 Blackham and Allen, above n 29; Allen and Blackham, above n 30. See also Women and Equalities Committee 'The use of non-disclosure agreements in discrimination cases' (2019)  The issue in practice, though, is that it can be difficult for individuals to gain either DO or union support to challenge age discrimination. 89 It is difficult to find evidence or proof of discrimination, 90 making success in court proceedings unlikely. 91 Further, unions, the DO and individuals themselves tend not to prioritise or recognise age discrimination, limiting the success of enforcement. 92 There is a need, then, to scrutinise these non-individualised forms of enforcement more closely.
(i) The DO The DO is responsible for supervising compliance with the Discrimination Act, 93 working to ensure discrimination does not occur and promoting equal rights and opportunities, 94 including through informing and educating government agencies, enterprises, individuals and organisations. 95 The DO is to provide advice and support to help those who have experienced discrimination to assert their rights. 96 In the first instance, the DO is to attempt to achieve voluntary compliance with the Discrimination Act. 97 The DO may investigate whether an organisation is complying with the Act: 98 those who are subject to the Discrimination Act must, at the DO's request, provide information, give access to workplaces, and attend discussions with the DO. 99 The DO can order a financial penalty for non-compliance. 100 The DO may bring a court action on behalf of an individual with their consent 101 but only if an employees' organisation does not do so. 102 While the DO receives complaints of discrimination, not all complaints will be actioned or even acknowledged with more than a form letter. 103 Litigation is not always the most efficient way of achieving change: 104 the DO's focus has therefore shifted 'from the individual … to a structural focus, a move from dispute resolution to bringing cases of strategic value'. 105 Strategic litigation particularly focuses on cases that clarify the law or highlight and draw attention to problems that exist. 106 Overall, though, the DO is focusing less on litigation and more on its public enforcement function. 107 While the DO has extensive legal powers, it is costly to exercise those powers; 108 running a discrimination claim is costly and time-consuming, even for the DO, 109 and may not be the best use of finite resources.
Instead of seeking settlements and pursuing litigation, then, complaints are used to inform the DO's enforcement work, by highlighting barriers to equality of opportunity. 'knowledge base' for the DO's other work. 110 In some cases, though, a complaint may lead to the DO initiating a supervisory measure, or proceeding to court in cases of strategic importance. Other complaints might be passed to trade unions. 111 The challenge, though, is that this shift in the DO's approach was not effectively communicated to the public: one thing that the Ombudsman could have done better … is communication because this expectation about the reports and what the reports are, that the reports are not just a report of a general problem or a problem, there is also the expectation that the report will also lead to a Court action which is an expectation that cannot be fulfilled. … by not being sufficiently clear in that matter, [the DO has] created expectations or have contributed to maintaining an expectation that cannot be fulfilled. 112 This might lead individuals to become disillusioned with the DO and equality law more generally. The DO's website is now clear about there being no guarantee of action where discrimination is reported; reporting has also been expanded to encourage by stander reporting. 113 Perhaps consequently, the DO has been criticised for the low number of complaints it pursues. 114 The disparity between complaints received and complaints actioned is evident in the DO's statistics. Age discrimination complaints received by the DO have increased by 121% between 2015 and 2021. 115 Age complaints across all areas increased from 497 in 2020 to 624 in 2021, though this increase largely related to age discrimination in healthcare during the Covid-19 pandemic. 116 As Table 1 shows, however, few discrimination complaints received by the DO lead to supervisory investigations, court proceedings or judgments; 117 court proceedings initiated equated to 0.09% of all complaints received in 2021 (and court proceedings do not necessarily originate as complaints).
These criticisms and concerns over complaint handling led to a public inquiry into the DO and its handling of matters. The Inquiry report emphasised the DO's role in promoting settlements and agreements: The Equality Ombudsman's primary task, also in the future, should be to help the parties reach agreement. However, in our opinion the Equality Ombudsman should broaden its work involving consensual solutions and examine the possibility for parties to reach agreement in more cases. 118 The Inquiry report arguably reflects a fundamental disagreement about what the role of the statutory regulator should be: there was a general perception that there was a great problem that we did not bring individual claims to Court in a sufficient amount; [the] underlying presumption [is] that's what you should be doing. … They do still think we should work for settlements in some cases, which we don't. 119 The DO remains disinclined to promote settlementsunless a claimant wants itin part because it represents a poorer form of justice for those who experience inequality, compared to those who are wronged in other ways: I do think that there is a rule of law aspect here, which I think needs to be highlighted because somehow just because we are talking about this Commission, just because we are talking about these victims, somehow, the ordinary operations of how we normally ensure the law is complied with disappears. 120 The DO's focus on strategic litigation has left a gap in assistance for individual claimants. 121 To address this, local anti-discrimination bureaus have been established as non-governmental organisations designed to combat discrimination. Bureaus typically offer free legal advice to people who have experienced discrimination, and may take cases to court. 122 The DO is increasingly referring complaints to Bureaus. 123 However, Bureaus lack the capacity to address all instances of discrimination. 124 Instead of focusing on complaints and settlements, the DO emphasises its 'general supervision' or public enforcement function 125 and role in influencing public debate. 126 This might include, for example, conducting investigations into specific organisations or whole sectors. 127 This has distinct advantages over individual enforcement, as it does not rely on individuals making a complaint: you don't need necessarily an individual victim who is willing or capable to go to Court. You are not risking [that] the matters you handle are uniquely the ones that are reported. As we know, there are groups in a society that are not aware of their rights and are [un]likely to report. Some This approach offers significant advantages, then, over using complaints to determine a statutory agency's priorities, and recognises that gaps in complaining may not mean there is no need for agency intervention. 129 There is a risk, though, that organisations or entire sectors might delay or defer action to advance equality, on the basis that the DO is conducting an investigation: 'sectors [that] are under investigation state that: "we are waiting [on the] investigation to see." So, they are sort of saying an impartial body is looking into this, so we don't need to do anything'. 130 This might delay progress in addressing discrimination, putting further pressure on the DOnot organisationsto achieve change.
Focusing beyond individual claims also de-emphasises the individualised remedies that are pervasive in the UK and Australia. 131 As in those comparative jurisdictions, those who experience discrimination in Sweden tend to want something other than monetary compensation: I would say, [in] nine out of ten cases, the individuals are not in it for the money, you know.
[They say] 'I want this to stop or I want to prevent it from happening to someone else' or in the general interest of sanctions 'they should not get away with these kinds of matters', so very often the compensation aspect is secondary. 132 However, even where individual enforcement is de-emphasised, equality law is not well set up to move beyond individualised remedies; there are limited financial penalties or sanctions in equality law, for example: For some reason, when it comes to discrimination equality law, the sanction is supposed to be reparation for the individual that's affected and, preferably, that should be resolved by settlement whereby there is no admission of guilt or anything as part of it. We have tremendous difficulty seeing that as an attractive way of dealing with discrimination cases. 133 Thus, an emphasis on systemic, supervisory enforcement needs to be backed by more tailored remedies and sanctions: 'sanctions are the great deficiency'. 134 These sanctions do not need to be used to be effective; the threat of sanctions can be sufficient. 135 This is an important area of reform for equality lawin Sweden and elsewhere.
(ii) Collective mechanisms Unlike the UK, Swedish labour law is typified by substantial collective negotiation and bargaining between the social partners. In 2018, 88% of Swedish employees had the right to bargain, and 65.5% were union members (see Table 2 and Table 3). In 2013, the proportion of Swedish employees in companies that were members of an employers' organisation was around 81%. 136 Thus, both employers and employees have a strong collective presence in Sweden. This compares markedly with the other countries under study.
Understandably, then, trade unions and the social partners are seen as playing a substantial role in addressing discrimination at work in Sweden. As one respondent noted, 'the Swedish system works, or 128 Sw2. 129 cf the approach of the UK and Australian statutory agencies, which often hinges on complaints: Blackham, above n 4, ch 8. 130 Sw2. 131 Blackham, above n 4, ch 6.   Thus, discrimination disputes are likely being resolved at the local level, but this has not been measured or quantified. Trade unions estimate that around 1-4% of central grievance negotiations progress to the Labour Court. 139 In this context, progressing to court is seen as a failure of the system; ideally, the courts are not required, and disputes are resolved at the local level. For respondents, then, 'there's reasons we need to be very proud of this system, where it creates rather few court cases'. 140 However, it is questionable whether unions have engaged with age equality as a priority. For example, Tikkanen and others found that, while measures adopted by the Swedish social partners were often focused on combating discrimination in working life, this did not focus on any particular age group. 141 Collective agreements do not generally appear to refer to age discrimination, though they may cover age-specific issues such as notice periods, integration measures for younger workers, partial retirement and pensions. 142 There is limited evidence, too, that informal mechanisms are being used in relation to age discrimination specifically.
There is a question, then, as to whether trade unions are well adapted to dealing with discrimination issues in particular: 'does this really work in discrimination cases; are trade unions suited for taking into account this kind of pressures of interests, other interests [than] the interest between capital and work?' 143 Unions arguably have conflicting interests that make them inadequate enforcers of discrimination law; 144 there are problems with relying on established structures and players to challenge the status quo. 145 Indeed, claims of discrimination may (indirectly) implicate unions themselves: there are situations where trade unions … might very well have been involved in the hiring process. … And at least, we have been consulted, so that would be negotiated. So, in a sense, when that decision comes, we are a little part of it. Then that becomes a sort of outside challenge to the system when someone says, 'I have been discriminated against'. 146 It is unclear how unions will respond to discrimination claims that involve a conflict of interests between employees or groups of employees, as opposed to between the employer and employee. 147 Further, while discrimination claims may sometimes come from union members, 148 they are more likely to come from relative 'outsiders' 149those unprotected by collectively negotiated solutions. At the same time, the attention of unions tends to be on 'core' workers. 150 Tensions between individual and collective rights 151 may lead to discrimination law being seen as a threat by trade unions.
The very nature of discrimination lawor, at least, how discrimination law is perceivedmay also lead to hesitancy among trade unions: 'discrimination is so absolute, right? It makes it a bad fit for our system, which is very pragmatic'. 152 This partly reflects the different dynamic of discrimination law to other industrial laws: [trade unions] like to negotiate and in order to negotiate, you need something to negotiate about. Anti-discrimination law … first of all, the nature of anti-discrimination law is that there are hard borders; there are some things that you cannot negotiate about, right? Whereas, we are really comfortable negotiating with everything else, right? So that has made it also more difficult. 153 There is a strong perception that discrimination law is 'inflexible' and therefore not well aligned with established collective mechanisms. In practice, too, trade unions might want to avoid bringing discrimination claims and creating case law or precedent on specific issues, the better to facilitate negotiation in an ambiguous or flexible space: When we have a landmark case, that changes everything. … a landmark case, if it becomes well known, it can change the framework of those negotiations. … Sometimes, we negotiate in the shadow of the perception of the law … Or you negotiate in the shadow of not really wanting to know what the law says, because that's risky, because that might change the next negotiations. … We don't want to go to court on a thing, because we didn't want an answer. 154 Further, as in other jurisdictions, alleging discrimination has a strong moral weight or stigma attached: 'Discrimination is a very strong statement, in a sense. … If someone says, "You are discriminating," that's escalating a conflict'. 155 Thus, any reference to discrimination or discrimination law might be avoided to support collegiate negotiation. Trade unions may also avoid discrimination claims due to a lack of government funding to train and upskill union representatives in equality law: 'Maybe if we had seen that kind of massive training of trade unionists and employers, [discrimination law] would have had a bigger impact'. 156 Overall, then, discrimination law is seen as a limited tool for Swedish trade unions, as it does not always assist in negotiations; this compares, then, to the situation in other jurisdictions, where discrimination law might be useful: for colleagues in other countries, discrimination law is a fantastic tool, because it gives them somethingand also something to use in negotiations, clearly. Thus, in practice, trade unions appear to rarely seek to enforce discrimination law, either in negotiation or bargaining, or through supporting individual claims. While disputes might be resolved at the local level, it is difficult to assess how common this might be. 158

(iii) Positive duties
The Discrimination Act also makes provision for 'active measures' in working life, a form of positive duty. 159 Active measures are preventative, and seek to counteract discrimination and promote equal opportunities. 160 Until 2017, this did not extend to the ground of age: it was confined to sex, ethnicity, and religion or other belief. Active measures now apply to all grounds, including age.
Employers must undertake specified work in relation to active measures. 161 Work related to active measures must be continuous 162 and involves: • identifying risks of discrimination and obstacles to equal opportunities; • analysing the causes of these risks and obstacles; • taking preventative and promotional measures that may be required; and • following up and evaluating the work undertaken. 163 Employers need to conduct this work in relation to: • working conditions; • salary and terms of employment; • recruitment and promotion; • education and skills development; and • combining work and parenthood.
Employers and employees are required to cooperate in undertaking this work; 164 employers, then, need to provide relevant employee organisations with information to facilitate that cooperation. 165 Work on active measures needs to be documented each year by employers with more than 25 employees; 166 those with 10-24 employees only need to document their salary survey. 167 Failure to comply with the active measures requirements may lead to an order to fulfil the obligations and a fine. 168 Active measures have significant potential to put the obligation on employersnot employeesto identify and address equality risks. 169 They have the potential to reframe equality law to achieve systemic change. 170 172 This can be difficult to assess in more fluid areas, such as making a work environment supportive of people of all ages. 173 It is also arguably contrary to one of the key aims of positive duties, which is to promote organisational problem-solving: the whole point is that you're supposed to have a structural thinking about it and, with all due respect, it's very difficult to order someone to have a structural thinking about something. … The purpose of the legislation is to force employers to consider these matters and have an active work, and it's difficult to force people to have an active work without telling them what they should be doing, which we don't have the right to do. So, there's a little bit of a catch-22 there. 174 That said, enforcement of the active measures requirements can work hand-in-hand with supervision of the general prohibition of discrimination; if organisations lack the processes to undertake active measures, they are also at risk of infringing the stricter prohibitions against discrimination. 175 Further, trade unions may appreciate the fluidity of the active measures requirements, and the emphasis on cooperation in determining active measures work, which is more in keeping with the focus on negotiation in Swedish labour law: what we have been saying is that we want more of this sort of semi-mandatoriness [sic], which we have got now, more possibilitiesespecially this active measure, negotiate active measures, that's what we want and now we have that. 176 The challenge going forward, then, is to make active measures 'part of the ordinary systems of negotiation'. 177 Thus, while enforcement by the DO might be difficult, the breadth and flexibility of the active measures requirements may lend them to action and enforcement by trade unions, providing a way for trade unions to better engage with equality law.
(iv) Gaps in enforcement What this discussion shows, then, is that there are critical gaps in enforcement for age discrimination law in Sweden. While the Swedish 'corporatist' model has moved beyond individual enforcement, collective and agency enforcement mechanisms are rarely used. Very few cases are brought by the DO or by trade unions. 178 There is an 'ambiguity' as to who is responsible for the enforcement of discrimination law in the Swedish context, 179 and a general lack of interest in age equality. 180  If the DO, trade union or other non-profit organisation are unable or unwilling to bring a claim, individuals face substantial barriers in enforcing their rights, including the risks of adverse costs orders; 182 costs typically follow the event. As Carlson has mapped, costs can significantly outstrip the damages awarded. 183 Individuals left without the support of a trade union or the DO can take on significant financial risks in pursuing a claim: the employees represented by trade unions at the Labour Court, they are generally getting a very good legal representation, without huge economic risks. The ones not represented by trade unions, but [who] choose their own legal advice or lawyers, … we see a lot of cases where people are given bad legal advice, which costs them a hell of a lot of money. They lose their job, and they go to court to [challenge their] employers and they end up losing their job, and losing the case, and having to pay two, or three, or four hundred thousand Swedish kronor [in costs] 184 … I'm sitting there saying, 'You should have asked. I could have told you that it was impossible.' … [lawyers] are not risk-averse on behalf of them. 185 This makes it nearly impossible for individuals to pursue their rights without support: the system … is not conducive at all because of the costs and because of these things. … in reality what it means is that a great number of people will not get their matters to Court … these matters are extraordinarily difficult to bring to Court successfully. There is also a gap between what is perceived as discrimination, what is legally something you can determine as discrimination and what you can actually succeed with in Court [with] credibility and witnesses and all sorts of other stuff. 186 It is likely, then, that many instances of discrimination simply go unaddressed.

Macro and micro perspectives on the effectiveness of enforcement
For individuals, then, it can be nearly impossible to pursue a discrimination claim in Sweden without the support of an institutional actor. The courts appear effectively inaccessible for those without the backing of the DO, a trade union or a local anti-discrimination bureau.
However, the accessibility of individual enforcement is not the only way to evaluate the effectiveness of legal systems. Malmberg posits that there are two different ways of measuring whether enforcement is 'effective'. 187 The first, 'micro', view considers whether individuals (with or without the assistance of regulators or trade unions) can enforce their rights. 188 This micro view is focused on the accessibility of enforcement mechanisms.
The second, 'macro' view considers 'how rules manifest in society as a whole'; that is, is equality achieved in practice? 189 For Malmberg, enforcement through trade unions or industrial relations approaches will often lead to high levels of compliance or macro effectiveness. 190  the industrial relations approach … could be effective, from a macro perspective of enforcement, because … depending on how the industrial relations system works in different countries, there is better presence at the workplace to determine what's actually happening out there, making sure this is in line with the rules, and so on. 191 However, an industrial relations approach does not always offer sufficient guarantees of individual rights to achieve micro effectiveness. 192 On the surface, this appears to be an apt description of the current enforcement of age discrimination law in Sweden. 193 That said, it is unclear whether trade unions are adequately enforcing age discrimination law to achieve 'macro' effectiveness; indeed, the interviews conducted in this study seem to indicate that there is limited enforcement occurring at the local level. With strong and enduring age norms, a strong statutory agencyand strong agency actionis likely required to challenge the existing status quo. 194 Relying on the social partners to negotiate in a way that challenges age-normative standards will likely prove ineffective, particularly when those most affected by age discrimination are rarely 'core' workers. 195 Given these concerns, in this Part I askis there evidence of 'macro' effectiveness of age discrimination law in Sweden? Or, rather, what evidence, if any, can we identify of macro effectiveness? Answering these questions is difficult, given the lack of relevant data collected by government and international bodies. This is a challenge for empirical legal research more generally and reflects our (de)valuing of legal data. 196 However, there is some, limited, survey data that provides an indication of the macro effectiveness of age discrimination law and its enforcement in Sweden. In particular, if age discrimination law were being enforced effectively, we would expect to see: • high rates of workforce participation for older workers; 197 • high reported job quality for older workers; • low reported rates of age discrimination at work; and • high perceptions of equality law's effectiveness.
This raises a related question, though: who is our comparator, and what are 'high' and 'low' rates of these indicators? For this study, it makes sense to use the UK and Australia as the comparators (given they were the comparative jurisdictions under study); but, given this research has identified compelling problems with the enforcement of age discrimination law in those jurisdictions, achieving similar rates to jurisdictions with individual enforcement models may indicate enforcement problems. Age discrimination law in the UK and Australia has been generally ineffective at achieving workplace change; 198 thus, if the figures for Sweden are similar to those in the UK and Australia, enforcement mechanisms may be similarly ineffectual across the jurisdictions. 191 Sw1. 192 Malmberg, above n 189, at 223. 193 Another way of seeing these issues is to focus on the primary rightthat is, the right not to be discriminated against; and the secondary rightthat is, the right to reparations if the primary right is infringed: Sw2. The secondary right is only engaged if the primary right is not respected. In terms of the labour force participation rate for older workers (Table 4), Sweden significantly outperforms both Australia and the UK, for both men and women.
Perhaps, though, this simply reflects higher workforce participation rates in Sweden, across the life course? Table 5 therefore compares workforce participation rates across the life course (ages 15-64) with participation rates for those aged 55-64, at two times periods (prior to the Covid-19 pandemic, and during/after). While workforce participation rates are higher in Sweden for the general workforce, there is also significantly less disparity between the rates of participation for all ages, and those for older workers in Sweden. Sweden is significantly outperforming Australia and the UK in the employment of older workers.
Further, job strain among older workers in Sweden appears far less common than in Australia or the UK (Table 6). Older workers in Sweden are far more likely to experience workplace autonomy. Thus, it is not just that workforce participation rates are high: older workers in Sweden likely also experience better job quality than those in the UK and Australia.
Strong workforce participation rates and high job quality in Sweden may reflect stronger employment protection generally, not the effectiveness of age discrimination law. 199 Our attention must therefore turn to discrimination law specifically. EU surveys have found that the UK and Sweden perform similarly in relation to reports of age discrimination. In the 2019 Special Eurobarometer 493: Discrimination in the EU, for example, age discrimination was perceived as being less common in Sweden than in the UK (Table 7). 200 The issue, of course, is that strong age norms may mean that respondents do not perceive age-based treatment as being discriminatory.
Indeed, when respondents were asked if they personally had felt they had been discriminated against or harassed on the basis of age (being perceived as old, or young) in the last 12 months, 201 this was reported by double the number of respondents in Sweden (10%, or 96 respondents) as in the UK (5%, or 50 respondents) (Table 8). In both jurisdictions, age discrimination was one of the most common forms of discrimination reported (eclipsed only by gender discrimination, at 11% and 6% of respondents in Sweden and the UK respectively). That said, this may simply reflect the demographics of respondents in each jurisdiction: the Swedish sample was significantly older than respondents in the UK (Table 9). This may distort the reported responses, assuming older respondents are more likely to experience (or recognise) age discrimination.
Importantly, though, when respondents were asked whether efforts to fight discrimination in their jurisdiction were effective, 202 UK respondents were more likely to see efforts as effective than  QC7 Using a scale from 1 to 10, please tell me if you think that the efforts made in (OUR COUNTRY) to fight all forms of discrimination are effective. '1' means you consider that these efforts are "not at all effective", and '10' that these efforts are "very effective". respondents in Sweden (Table 10). This may indicate, then, that there is scope to do more to achieve both micro and macro effectiveness of age discrimination law in Sweden.
Conclusion: the future of the Swedish model Individual enforcement of discrimination law is inherently problematic. Few complaints are likely to be raised; even fewer will be successful in courts or tribunals. There is much to be learnt from the Swedish model for countries with 'privatised' enforcement models, like the UK. In particular, there are compelling lessons for the UK around transparency, moving beyond complaints to set statutory agency priorities, and expanding the role of trade unions and collective action to advance equality.
It appears, though, that the Swedish corporatist model, which severely curtails individual access to courts and tribunals, has fundamental limitations for achieving systemic change, particularly in an