Nordic Economic Policy Review 2025

Changes in union density in the Nordic countries

Anders Kjellberg

Abstract

Over the last 25 to 30 years, union density has experienced a significant decline in Denmark, Finland, and Sweden, while remaining relatively stable in Norway and completely unchanged in Iceland. In the first three countries, the weakening of Ghent systems (state-subsidised union unemployment funds) stands out as the primary factor contributing to union decline. In Sweden, this trend is further exacerbated by widening disparities between blue-collar and white-collar union density. This study analyses both the historically high union density in the Nordic region and the notable decline in certain countries, focusing on five key features of Nordic industrial relations and their evolution. Additionally, the impact of situational factors and social customs is explored. Particular emphasis is placed on newcomers to the labour market, specifically immigrants and young people. In Norway and Sweden, the substantial influx of recently arrived immigrants poses challenges for union recruitment efforts. To effectively engage with immigrants and young people, maintaining a strong union presence in the workplace is crucial.

Keywords: Unions, union density, Ghent system, situational factors, social customs

By international standards, union density is very high in the Nordic countries. About 50–70% of the employees are members of trade unions, and the figure is even higher in Iceland (Table 1).

Union density has declined considerably in Denmark, Finland and Sweden regardless of whether 1990, 1993 or 2000 is chosen as the starting point. In Finland and Sweden, it peaked at about 85% in the mid-1990s. In Norway, the level of unionisation has remained almost the same since 2000, and in Iceland, it has risen slightly.

Table 1. Union density in the Nordic countries since 1990

	1990	1993	2000	2005	2009	2010	2013	2015	2017	2019	2020	2021	2022	2023
Denmark (1)	76	77	/74	72	71	70	71	70	68	68	70	69	68	68
Denmark (2)	74	75	/72	68	65	63	61	59	57	57	57	56	55	54
Finland (1)	73	80	74	73	73	71	68	68	63	59
Finland (2)	72	79	71	69	67/68		65		60			55
Iceland		87	89				89	90		92
Norway	57	57	53/51	51	50	51	50	50	50	50	51	50	50	50
Sweden	81	85	81	78	71	71	70	69	69	68	69	70	69	68
- blue-collar	82	86	83	77	70	69	66	64	61	60	61	62	59	58
- white-collar	81	83	79	78	72	73	73	74	73	72	73	74	73	73

Note: Denmark: Including unemployed people. Employed and unemployed people 1990–1993 register-based data November aged 16–66; 2000–2023 AKU Quarter 4 aged 15–64. Denmark (2) excluding ‘ideologically alternative’ or ‘yellow’ unions.
Finland (1): OECD-AIAS-ICTWSS. Finland (2): 1990 refers to 1989, 1993 refers to 1994, 2000 refers to 2001, 2005 refers to 2004. Ahtiainen 2001, 2011, 2023.
Iceland 2019 refers to 2018.
Norway 1990-1993: Stokke (2000); 2000: 53% from Nergaard & Stokke (2010); 51% from Nergaard 2024:12; 2001-2022: Nergaard 2024:12; 2023: preliminary data obtained from Kristine Nergaard, Fafo.
Sweden: LFS/AKU annual averages employees aged 16–64, excluding full-time students with jobs. Kjellberg 2024, Kjellberg 2019/2024: Appendix 2.

One of the aims of this article is to explain why union density is much higher in the Nordic countries than elsewhere. Iceland is a special case as all employees are required to pay union dues, even non-members. Another is to explain why union density has declined considerably in Denmark, Finland and Sweden over the last 25–35 years but only modestly in Norway.

In the first section, after looking at some main similarities and differences between the Nordic countries, we will examine five Nordic features to explain the high union density:

The combined centralisation and decentralisation of industrial relations and trade unions. Centralisation refers to centralised employers’ associations, national unions and union confederations with far-reaching powers to conclude collective agreements about wages, working-hours, occupational pensions and other issues. Decentralisation refers to the widespread union presence at workplace level, where union representatives together make up ‘trade union clubs’, negotiate about the local implementation of national collective agreements and recruit new members. Decisions on strikes and other conflict measures, however, are centralised at the national level.
The preference for self-regulation over state regulation
The Ghent systems in Denmark, Finland and Sweden
The socio-economic divisions in union movements
Large proportions of the employees in the public sector.

In the second section, we examine each of these features to see if changes to them explain the large falls in union density in Denmark, Finland and Sweden and why density has not fallen in Norway. Two other aspects are also considered:

Situational factors and social customs affecting the decision to join a union or not
Newcomers to the labour market: immigrants and young people.

Section two is followed by some concluding remarks.

1 Why is the rate of unionisation so high in the Nordic countries?

1.1 The combined centralisation and decentralisation of industrial relations and trade unions

Research shows that the combination of centralisation and decentralisation of industrial relations and trade unions in the Nordic countries prevents fragmentary unionisation and facilitates recruitment via extensive networks of ‘trade union clubs’ and workplace union representatives (Andersen et al. 2014). The high density of Nordic employers’ associations (the proportion of the employees in companies affiliated to employers’ associations) that negotiate collective bargaining agreements at sectoral/branch level (Table 2) means there is no employer hostility to trade unions and union membership in large parts of the labour market. This is promoted by the long tradition of co-operation between the labour market parties institutionalised in basic agreements in Denmark (1899), Norway (1935) and Sweden (1938). Finland was a latecomer in this respect, with national collective agreements only making their breakthrough after World War II.

Table 2. Main characteristics of Nordic countries

	Denmark	Finland	Iceland	Norway	Sweden
Population, 1 January 2024 (millions)	6.0	5.6	0.4	5.6	10.6
Foreign-born population, 1 January 2024	13.5%	8.2%	20.6%	17.4%	20.3%
Employees with fixed-term jobs in 2023 as % of total number employed aged 20-64	8.2%	12.2%	8.9%	6.4%	10.8%
Public-sector employment as a % of total employment, 2021 (Iceland 2019)	28.0%	25.4%	25.0%	30.9%	29.3%
Union density	68% 2023	55% 2021	92% 2019	50% 2022	68% 2023
- in private sector	60% 2015	46% 2021		38% 2022	64% 2023
- in public sector	82% 2015	77% 2021		79% 2022	78% 2023
Density of employers’ associations	68% 2018	64% 2022	78% 2018	81% 2022	87% 2021
- in private sector	52% 2018		70% 2018	72% 2022	83% 2021
Coverage by collective agreements	82% 2018	89% 2022	90% 2018	64% 2022**	88% 2023
- in private sector	73% 2018	84% 2022*		47% 2022**	83% 2023
Extension of collective agreements***	-	X	X	X	-
Required minimum union density in workplace for collective agreements	50% HK	-	-	10% blue-collar	-
Bargaining levels (wages	Two tiers	Two tiers + one tier	Two tiers	Two tiers	Two tiers
Dominant bargaining level	Industry	Industry	Industry	Industry	Industry
Statutory minimum wage	-	-	-	-	-
Balloting on bargaining and mediation proposals	X	-	X	X	-
Linked balloting results ****	X	-	-	-	-
First private-sector basic agreement	1899	1944	-	1935	1938
Strike days: annual average, 2014–2023	38,000	203,000		91,000	3,300
Ghent system (state-subsidised union unemployment funds)	X	X	-	-	X
Competing unemployment funds	Alternative	YTK	-	-	(Alfa)
‘Ideologically alternative’ (’yellow’) unions	X	-	-	-	-
Supplementary union income insurance schemes	X	-		-	X
Income ceiling for unemployment insurance	X	-		X	X
Tax deduction for union dues	X	X	(X)	X	-

Note: * Excluding collective agreements by the state extended to cover all employees in an industry, coverage in the private sector was 64% in 2021/2022.
** Excluding extended agreements. Including these, coverage in 2022 was 58% in the private sector and 72% in private + public sector (Nergaard 2024: 30).
*** Collective agreements that the state extends to cover all employees in an industry.
**** The Danish state can bring balloting results from different industries and bargaining areas together into one single unit.

Union density, density of employers’ associations and the coverage rate of collective agreements refer to the share of employees who a) are union members, b) work in a company or public agency (local/central government) which is affiliated to an employers’ association, and c) are covered by a collective agreement.
Source: Kjellberg 2024.

By contrast, trade unions in the UK (union density 22% in 2023), USA (10% in 2023) and Japan (16% in 2023) fight company-by-company for recognition and bargaining rights. Both one-sided decentralisation, which is a characteristic of those three countries, and one-sided centralisation (the Netherlands) appear to push down union density. When explaining the high rate of unionisation in Belgium (49% in 2019) compared to the low figure in the Netherlands (15%), Ebbinghaus & Visser (1999:152) point to the strong union presence in Belgian workplaces and the weak presence in Dutch ones as a difference ‘with big consequences’ for the large unionisation gap between the two countries.

Union representation in the workplace is a decisive advantage when it comes to protecting and supporting workers, achieving improvements in the workplace and, as a consequence, demonstrating that unions matter. Face-to-face contact with union representatives and other members maintains membership as a social norm, a social custom.

For employees in workplaces without union representation, and consequently less social pressure to join, selective incentives, like union income insurance schemes, can be expected to have a relatively greater impact. This is in line with Ebbinghaus et al. (2011: 120–121), who show that the effect of workplace representation on union density is smaller in countries with a Ghent system than elsewhere. As a result, workplace representation will have relatively greater importance in Norway than in Sweden for maintaining high union density. This is reinforced by the Norwegian practice that centrally negotiated collective agreements for blue-collar workers in the private sector are only implemented at workplace level if the workplace union demands it, and at least 10% of the employees in the bargaining area covered by the agreement in the workplace are union members (Kjellberg & Nergaard 2022: 61).

A Norwegian survey from 2019 confirms the importance of workplace recruitment. As many as 40% of union members were recruited by a union representative or colleague in the workplace, while 30% were recruited via student membership (Nergaard 2020a-b). Only 25% said that they joined on their own initiative. The most common reason for joining a union was also workplace-related, i.e. to receive help in the event of problems in the workplace, which is very similar to the top reason in a Swedish study by Calmfors et al. (2021a-b).

The practice of the government extending the coverage of collective agreements to all companies in an industry is widespread in Finland, indicating that in many workplaces the unions have great difficulties setting up branches (workplace union clubs) and enforcing collective agreements.

Besides showing that workplace unions are relatively less important in Ghent countries, Fazekas (2011) demonstrates that extending collective agreements encourages freeriding because it negates the need for local union presence. This has consequences for membership numbers, as a union in the workplace almost doubles the probability of an employee being a member compared to workplaces with no union activity (Fazekas 2011:160).

Workplace union organisations are, however, very important in Ghent countries, too. Union workplace presence is strengthened by the positive Ghent influence on union density as higher membership facilitates the recruitment of union representatives, who in turn recruit more members. In that way, workplace unions reinforce the Ghent effect. Accordingly, developments that undermine the membership recruitment capacity of Ghent systems can be expected to have negative consequences on the presence of ‘trade union clubs’ in the workplace.

1.2 The preference for self-regulation over state regulation

Another feature of the Nordic model of industrial relations, or more correctly, the Nordic models, is the dominance of self-regulation over state regulation, i.e. the preference for collective bargaining agreements rather than labour law and other types of state intervention (Kjellberg 2017). Self-regulation presupposes a high density of both unions and employers’ associations. In contrast to most EU countries, none of the Nordic countries has a statutory minimum wage. The Swedish model of industrial relations is the closest to a Nordic ideal type in terms of the degree of self-regulation. The government in Sweden is much less involved in wage formation than in Denmark (where mediation proposals are often imposed by law), Finland (which used to have a tradition of tripartite bargaining) and Norway (compulsory arbitration).

Denmark and Sweden are the only Nordic countries with neither statutory minimum wages nor mechanisms for extending the coverage of collective agreements (Table 2). By contrast, France has both. When the French government raises the minimum wage by a fixed per cent, it serves as a ‘mark’ for the whole labour market corresponding to the Nordic industry norms. The French state also extends all collective agreements, resulting in 98% of the employees being covered. With such extensive government regulation of wage formation, the unions risk appearing redundant. Not surprisingly, under 10% of the French employees are in unions.

Finnish collective agreements are extended to cover whole industries provided at least 50% of the employees in the bargaining area are covered by the central agreement. No such rule exists in Norway, where the extension mechanism is only applied in industries where workers risk being exposed to poor conditions – in practice, only if foreign workers are paid less and have worse working conditions than the norm (Kauhanen 2025).

In Denmark and Sweden, the unions’ right to take industrial action against companies that are not members of employers’ associations is the closest equivalent to extension mechanisms and is, therefore, a key means of maintaining the self-regulation model. In Sweden, although the number of industrial disputes per annum to force employers to conclude collective agreements is small, it is a powerful tool to uphold the high coverage of collective agreements. In this context, the right to take sympathy action (strikes, blockades, etc.) is of central importance. The Nordic countries are distinguished by the large funds available to unions taking industrial action and their extensive right to do so. In countries with no extension mechanism, unions have a particular interest in recruiting members in workplaces where there are no collective agreements so they can push them through having achieved sufficient local strength.

1.3 Ghent systems in Denmark, Finland and Sweden

The Ghent variant of unemployment insurance is often considered a selective incentive to join unions. Strictly speaking, this is not true. In all Nordic Ghent countries – Denmark, Finland and Sweden – employees can choose to join a union’s unemployment fund without joining the union that runs it. In Sweden, the proportion of employees who are in union unemployment funds but not the union has increased considerably since the late 1980s. In 2023, no fewer than 43% of the members of the unemployment fund linked to the LO-affiliated Commercial Employees Union were not members of the union itself and three-quarters of the members of the Hotel and Restaurant Workers’ fund. This hollows out the Ghent effect in these industries. It is estimated that an average of one in four members of Swedish union unemployment funds is not a member of the union (Kjellberg 2024, Table 61).

Despite this, new evidence suggests the Ghent system still serves as an important recruitment tool for trade unions. According to a study by Calmfors et al. (2021a), the top four reasons for joining are:

Assistance in the event of a dispute with the employer
Access to supplementary income insurance
Access to unemployment funds
Better prospect of keeping the job in the event of redundancies.

Of the eighteen reasons listed, no fewer than three of the top four are about the risk of losing the job. Two of them concern access to unemployment benefit and one employment protection. Considering that membership of a union is not a prerequisite for joining its unemployment fund, it is remarkable that access to a union unemployment fund is ranked as high as number three. Calmfors et al. note that membership of a union and its unemployment fund is often perceived as a ‘union package’ whether both are selected or not.

It is worth noting that the reasons listed above are given by union members. The increasing number of people who are only members of unemployment funds shows that many non-members correctly do not consider union and fund membership to be a ‘union package’.

Although this traditional package is gradually losing ground, the three-tier combination, i.e. union – union unemployment fund – union income insurance, is a new union package that makes freeriding impossible. To benefit from the income insurance, you must be a member of both the union and its unemployment fund.

Ghent systems have a positive impact on union density in two ways, one of which involves the union in the workplace:

By facilitating membership recruitment.
In turn, the increased number of union members expands the base for setting up ‘trade union clubs’, further improving the prospects for recruiting and retaining members, which creates or reinforces a social custom of unionisation.

It might be expected that this double Ghent effect would be particularly important in industries like retail and restaurants, where it is difficult to recruit members due to a high share of fixed-termed and part-time jobs and high labour turnover. It is hardly a coincidence that union density in Norway, where there is no Ghent system, is very low in industries like trade (26% in 2022 compared to 59% in Sweden) and hotels & restaurants (16% in Norway, 38% in Sweden). However, the extremely high proportion of members of the Swedish hotel and restaurant unemployment fund not affiliated to the union itself has reduced the Ghent effect in this industry. Not surprisingly, union density has declined sharply among hotel and restaurant workers.

‘Low-wage’ unions like the Swedish Commercial Employees Union and the Hotel and Restaurant Workers’ Union have income insurance schemes, but their recruitment capacity is limited because the wages of their members seldom reach the ceiling for the ordinary (state-subsidised) unemployment insurance. Income insurance schemes only provide unemployment benefit above this ceiling. As a result, most members of low-wage unions do not qualify for supplementary unemployment benefit.

In Finland, there are no supplementary income insurance schemes because there is no income ceiling for unemployment insurance. Some Danish unions have income insurance, but some of them offer only voluntary individual income insurance resulting in a high extra union dues because of their non-collective character. Other Danish unions have cheaper obligatory collective income insurance schemes.

As we have seen, unions in Norway and other non-Ghent countries have a relatively greater need for workplace presence than unions in the Nordic Ghent countries. At the same time, the absence of a Ghent system may make it more difficult to form local ‘union clubs’ in Norway.

1.4 The socio-economic divisions in union movements in the Nordic countries

The Nordic union movements are distinguished by a far-reaching socio-economic division between blue-collar, white-collar and ‘academic’ unions, all of which have their own confederations. This is most evident in Sweden. The total dominance of blue-collar unions in LO-Sweden is related to the broad Swedish definition of blue-collar workers or arbetare. For instance, assistant nurses and health care assistants organised in the LO-affiliated union Municipal Workers’ Union, and most retail employees are categorised as arbetare in official statistics.

The self-organisation of white-collar workers into separate unions and confederations is considered to have facilitated their unionisation as it makes it easier to identify with the union. The confederations of professional associations (Akava, Akademikerne, Saco, etc.) have no equivalents outside the Nordic countries.

1.5 Large proportions of the employees in the public sector

The high proportion of public sector employees, who usually have a higher rate of unionisation than private sector workers, is also conducive to a high union density in the Nordic countries. The long expansion of Nordic welfare states resulted in large public sectors, which promoted the growth of professional unions and other unions dominated by public-sector employees.

Norway and Sweden are the Nordic countries with the highest proportions of public-sector employees (Table 2). Finland and Iceland have the lowest, while Denmark is in the middle.

As Boeri et al. (2001:24) noted, the private/public sector unionisation gap in the late 1990s was particularly large in Norway and other non-Ghent countries. In fact, the density gap between Norway and Sweden, which is the largest between the Nordic countries, is entirely concentrated in the private sector. In both countries, public-sector density in 2022/2023 was almost 80%, while only 38% of Norwegian private-sector employees were union members compared to 64% in Sweden (Table 2). This indicates that the Ghent effect is concentrated in the private sector.

Having examined the influence of five conspicuous features of Nordic industrial relations on the high union density in these countries, we will now look at them again – as well as two additional factors –this time to explain the declining union density in Denmark, Finland and Sweden in contrast to stability in Norway.

2 Why declining union density in Nordic Ghent countries but not Norway?

2.1 Combined centralisation and decentralisation of industrial relations and trade unions

Apart from the Finnish forest industry, the sector/industry level is still the dominant one for collective bargaining in the Nordic Region. The growing importance of local negotiations to implement national agreements increases the importance of workplace unions, but the proportion of employees covered by ‘union clubs’ and local union representatives has fallen, at least in Sweden (Kjellberg 2024).

Another challenge is the emergence of new globalised companies (Tesla, Spotify, Google, etc), which have negative attitudes toward unions and oppose collective bargaining. A growing number of construction subcontractors employing mainly foreign-born, non-union workers ignore the Nordic model of industrial relations and operate ‘outside’ the model (Kjellberg 2023a).

2.2 The preference for self-regulation over state regulation

The Swedish government’s radical remodelling of the Ghent system in 2007 and 2008 led to a massive loss of union members. No other state intervention in Swedish history has affected union density so negatively.

The Finnish government’s active role in reforming the wage formation system contrasts sharply with the Swedish process up to the 1997 Industry Agreement. Most controversial is the plan for local agreements with non-union representatives in companies covered by extended collective agreements and not affiliated to employers’ associations.

2.3 Erosion of the Ghent systems in Denmark, Finland and Sweden

The trends for union density in the three Nordic Ghent countries illustrate that these systems may also have negative consequences, which Norway has avoided. Firstly, the cost of union membership may appear more reasonable when it does not involve a comprehensive ‘union package’, which includes membership of an unemployment fund. When the costs seem too high for a growing number of people in the Ghent countries, fund membership combined with non-union membership (above all in Sweden), alternative unions (Denmark) and unemployment funds with no links to traditional trade unions may seem like an attractive low-cost option (Denmark and Finland).

Norwegian unions have not been subjected to institutional changes like those in Denmark, Finland and Sweden, where remodelled Ghent systems have eroded the capacity of unions to recruit. In Sweden, the 2007–2008 unemployment insurance reforms (mainly higher membership fees and lower benefits) resulted in a considerably higher price for the ‘union package’.

Following an initiative by Finnish employers who believed it was important to have a non-union unemployment fund, the independent cross-occupational YTK was founded in 1992 (Shin & Böckerman 2019: 3). Since then, it has expanded considerably at the expense of trade union unemployment funds. With about 530,000 members, it is by far Finland’s largest unemployment fund and covers one in five of the employees. A few other independent funds have also been established. The competitiveness of YTK is strengthened by an association connected to it (‘YTK Worklife’, founded in 2005) that provides insurance and individual services to its members. Due to the emergence of the non-union YTK fund, the proportion of workers in union unemployment funds but not unions grew more slowly in Finland than in Sweden.

In Denmark, cross-occupational unemployment funds were introduced in 2002 when the centre-right government changed the law. This promoted the growth of ‘ideologically alternative’ or ‘yellow’ trade unions, weakening the LO unions in particular (Kjellberg & Ibsen 2016). The large drop in membership prompted LO-Denmark to merge with the white-collar confederation FTF in 2019 to form FH.

At the end of 2023, the ‘yellow unions’ had almost 387,000 members, corresponding to one in five Danish union members. The cross-occupational unemployment funds are linked to yellow unions, which in general neither sign collective agreements at industry level nor implement them at workplace level, which is why they can offer low-cost memberships (Ilsøe 2013:85-86). One of these unions is even called Bedst og Billigst (‘Best and Cheapest’). The ‘yellow’ unions do not have workplace representatives either. They offer individual services in the event of disputes with employers, membership of their unemployment fund and other insurance services, including voluntary income insurance schemes. The members are found, above all, among younger people at workplaces in the private sector without collective agreements and without representatives of traditional unions (Ibsen et al. 2013).

In Sweden, the centre-right government raised the fees for unemployment funds considerably in 2007, making union membership quite expensive as union dues generally included the fee for the unemployment fund as well, although many unions subsequently separated the two. The government also established a link between the fees paid to funds and the unemployment rate for the members of each fund. In July 2008, it reinforced this link. The higher the unemployment, the higher the fee to pay to the unemployment fund. As unemployment is usually much higher among blue-collar workers, they had to pay considerably higher fees than white-collar workers.

The price of ‘the union package’ was further raised in 2007 by another government reform enacted simultaneously with that mentioned above. From 2007 the government abolished the tax reduction corresponding to 25% of union dues and 40% of the fee paid to an unemployment fund. All other Nordic countries have tax deductions for union dues. A Norwegian study based on data for 2001–2012 calculated that the private sector union density in 2012 would have been five percentage points lower without the increased tax deduction. Furthermore, tax deductions have the strongest relative impact on “newcomers to the labour market, such as younger workers and immigrants, or workers with a more marginal attachment to the labour market, such as workers with part-time or temporary jobs” (Barth et al. 2025:15).

Recent research demonstrates that the introduction of Ghent systems did not result in increased union density per se (Rasmussen & Pontusson 2018). What matters is the level of state subsidies. The situation in Sweden confirms this. Due to reduced state subsidies, membership of an unemployment fund became much more expensive while the benefits deteriorated, and requirements were tightened (Lindellee & Berglund 2022). The result was a massive flight of members. In 2007 and 2008, Swedish unions lost 245,000 members and the unemployment funds more than 460,000 members, of which the union unemployment funds lost roughly 400,000 members (Kjellberg 2024:125). Union density fell dramatically, from 77% in 2006 to 71% in 2008. A drop of six percentage points in two years is also remarkable from a global perspective.

Diverging white-collar/blue-collar union density in Sweden

Until the Swedish unemployment fund fees from 2014 were restored to about the same level as before 2007, union density declined far more among blue-collar workers than white-collar ones (Table 1). After that, white-collar density has remained almost unchanged while blue-collar density has continued to fall. Apart from the lower white-collar fund fees in 2007–2013, the more frequent and attractive union income insurance schemes for white-collar workers help explain the growing gap between white-collar and blue-collar union density. Another contributory factor is that it is more difficult to organise blue-collar workers due to the higher proportions of young people, immigrants and employees on fixed-term and part-time contracts, structural characteristics that often overlap.

In addition, white-collar workers are overrepresented in the public sector. However, it is remarkable that in Sweden white-collar union density has declined considerably in precisely the public sector – and only in this sector. One possible explanation is that union income insurance schemes might not be considered very important by public sector white-collar workers due to their lower risk of unemployment (Calmfors et al. 2021b:51).

However, before 2007, blue-collar union density was already declining more quickly than white-collar density. The average annual fall among employees in general varied in 1999–2006 from zero to just over one percentage point and was significantly greater among blue-collar workers (on average almost one percentage point) than among white-collar workers (0.5 percentage points). As a result, blue-collar union density (84% in 1999) and white-collar density (80%) converged to 77% in 2006.

Since the mid-2010s, average Swedish union density has remained relatively stable at 68–69%, with a temporary peak at 70% during the second pandemic year (2021). However, under the surface, blue-collar and white-collar union densities have continued to diverge. From both groups being equal at the same level, 77% in 2006, blue-collar density in 2023 has dropped by 19 percentage points (to 58%) and white-collar density only by four points (to 73%).

Blue/white collar gap in Norway

No similar blue-collar/white-collar divergence has occurred in Norway. Between 2008 and 2017, union density among both categories of workers declined by only three percentage points (Kjellberg & Nergaard 2022:62). Nevertheless, Norway also has a substantial gap between blue-collar and white-collar density: 43/57% in 2017. Part of the explanation is the extremely low union density among some groups of blue-collar workers, such as restaurant workers and cleaners. Regarding the whole period 2001–2017, blue-collar density declined by seven percentage points compared to two for white-collar workers. In 2001, the gap was already nine percentage points.

The growing proportion of white-collar workers combined with their higher union density has promoted the stability of average Norwegian union density. A similar effect can be discerned in Sweden since the mid-2010s.

2.4 The socio-economic divisions in union movements in the Nordic countries

The strong growth of ideologically alternative or ’yellow’ unions in Denmark has fundamentally changed the Danish union landscape, among other things by accelerating the merger of LO-Denmark and the white-collar confederation FTF into Fagbevægelsens Hovedorganisation (FH). Excepting the ‘yellow’ unions, Danish union density fell dramatically, from 72% in 2000 to 54% in 2023.

2.5 Large proportions of employees in the public sector

Public sector cuts, privatisation and outsourcing via public procurement have reduced the share of employees in the public sector. In Sweden, it fell from 43% in 1993 to 36% in 2000 (Kjellberg 2022:26). Many jobs have moved from the public sector to the private service sector, that is, from the sector with the highest rate of unionisation to the one with the lowest (Table 3).

Table 3. Union density by sector and industry in the Nordic countries since 1990

	1990	1993	2000	2005	2009	2010	2013	2015	2017	2019	2020	2021	2022
*Denmark*	74	76	75	72	69	68	69	67	67	67
Private sector			64	59			62	60
Public sector			96	97			81	82
*Finland*	72	79	71	69	67/68		65		60			55
Industry & construction	80	82	84	86	84		81		72			63
Private services	49	65	55	50	50		52		48			42
Public sector	85	85	88	88	82		76		73			77
*Norway*	57	57	53 /51	51	50	51	50	50	50	50	51	51	50
Private sector		44	40				37			36		38	38
- Industry & construction		57	54	/48		45		44		43		44
- Manufacturing				/56		53		52		52		53	53
- Construction				/35		33		30		30		31	30
- Private services		36	33	/32		32		32		33		35
Public sector		80	81	/76		77		79		77		79	79
*Sweden*	81	85	81	78	71	71	70	69	69	68	69	70	69
Private sector	75	78	74	72	65	65	65	64	64	63	64	65	64
- Manufacturing	87	89	86	84	79	78	78	77	76	75	76	75	76
- Construction	86	85	85	80	71	70	67	65	64	61	60	61	59
- Private services	66	71	67	67	60	60	61	60	60	60	62	63	62
Public sector	91	94	92	89	84	85	83	81	79	79	79	80	79

Note: Denmark 2005 refers to 2004, 2015 refers to 2016. OECD-AIAS-ICTWSS.
Finland 1990 refers to 1989, 1993 refers to 1994, 2000 refers to 2001, 2005 refers to 2004. Ahtiainen 2001:33, Ahtiainen 2011:35, Ahtiainen 2023:40.
Norway 1993 refers to 1995, 2000 refers to 2001. 1993-2000: labour force surveys in Nergaard 2024:13; 2005-2022: register-based data Nergaard & Ødegård 2022:12, Nergaard & Ødegård 2024:16-18, 29-31 and Nergaard 2024:15 and supplementary data obtained from Kristine Nergaard, Fafo. For total union density, see Table 1.
Sweden LFS/AKU: annual averages employees aged 16-64, excluding full-time students with jobs. Kjellberg 2024. Industry, construction and private services Q1 1990 and 1993. Kjellberg 2019/2024.

As in Sweden, private services in Finland increased its share of the employees at the expense of manufacturing industry (Böckerman & Uusitalo 2006). These structural shifts have a long-term negative effect on average union density.

As the lowest price often wins in public procurement, the space has increased for unfair competition, unfair working conditions and companies with negative attitudes toward unions (Kjellberg, 2023a).

2.6 Situational factors and social customs affect the decision to join a union or not

In Ghent countries, unionisation usually varies with the business cycle. This was particularly evident in Finland and Sweden in the 1990s. Between 1990 and 1993, the Finnish unemployment rate rose from 3.1% to 16.5% and union density by seven percentage points (Kjellberg 2022). Similarly, in Sweden, unemployment rose from 2.4% to 10.2% and union density increased by four points in the same period.

Before that, in the late 1980s, the Swedish economy had been ‘overheated’, and union density dropped to 81% after a peak at about 85% in the mid-1980s. With the good times, some felt that they could do without union membership, but for peace of mind many thought it best to be in an unemployment fund. It was at this point that the number of private-sector white-collar workers in union unemployment funds but not unions started to take off (Kjellberg 2011:83).

The decreasing unemployment from the mid-1990s further broadened the scope for more individualistic behaviour. In Sweden, membership of union unemployment funds but not unions now spread to blue-collar and public-sector workers. By 2000, union density had fallen by 6-8 percentage points in Finland and by four in Sweden. An interview study in 1993 of union members, members of unemployment funds and other workers in the Stockholm region showed that young workers, in particular, weighed the costs of membership against the benefits (Kjellberg 2001/2017, Kjellberg 2024: 316). Other Swedish research in the 1990s and later confirms the growth of a more instrumental approach to unions, according to which the current situation of the individual plays a more important role.

In a broader sense, the benefits and costs of being (or not being) a union member also include the reactions from colleagues, union representatives, family and friends. In workplaces with a strong union presence, non-members may pay a price in the form of disapproval or even ostracism, although by the early 1990s the latter was a thing of the past. This is in line with social customs theory. As we have seen, the presence and strength of unions in the workplace play key roles in recruiting and retaining members. With no Ghent system or an eroding one, this becomes even more important unless union income insurance schemes or something else serves as a substitute.

According to the Norwegian survey by Nergaard (2020a-b), a majority of the non-unionised workers will consider joining due to situational factors:

“The majority of the non-unionised workers will consider joining if they can find a suitable union, if they should come to a workplace where this is common, or if the workplace proves to be insecure. Only a minority rejects the possibility of joining outright and irrespective of the situation. Younger workers state more frequently than others that they will consider joining, given certain preconditions.” (Nergaard 2020b:3).

Fixed-term employment is one situational factor that reduces the propensity to join a union. The Nordic countries with the highest shares of fixed-term jobs are Finland and Sweden, in sharp contrast with Norway (Table 2). Contributing to the growing proportion of employees with fixed-term contracts in Sweden were the successive amendments to the 1974 Employment Protection Act, which made Swedish legislation among the most liberal in the EU in the 1990s (Svalund & Berglund 2018: 265). The most insecure form of employment, ‘general fixed-term employment’, introduced in 2007, expanded fastest. The over-representation of temporary blue-collar jobs, particularly among young people and among the growing number of foreign-born workers, is one of the circumstances that makes it harder to organise blue-collar workers than white-collar ones.

Contrary to what might have been expected, union density did not increase in the Nordic Ghent countries during the financial crisis. As mentioned, a contributing factor in Sweden was that rising unemployment resulted in higher fees for membership of unemployment funds, particularly among blue-collar workers (Kjellberg 2011). The weakening of the link between membership of a union and membership of its unemployment fund caused by the erosion of the Ghent systems has also dampened the impact of economic downturns on union density.

By contrast, during the COVID-19 pandemic starting in 2020 and the economic uncertainty that followed, both union and unemployment fund density increased (Kjellberg 2024). Unlike during the financial crisis, fund fees did not soar this time as differentiated fees were abolished in 2014. On top of that, unemployment benefit was made more generous. Not surprisingly, both blue-collar and white-collar density increased by two percentage points from 2019 to 2021.

In Denmark, union density increased from 68% to 70% during the first year of the pandemic. During the inflation years 2022 and 2023, it fell back down to 68%. In Sweden, union density also decreased during the inflation years. The rapidly increasing inflation and the accompanying reduction in real wages in 2022 meant some people did not feel they could afford to join unions as the price of food, electricity, petrol, etc., rose. Immigrants came under particular financial pressure due to their over-representation in low-wage jobs. After blue-collar union density rose from 60 to 62% during the pandemic years, it fell by twice as much – from 62% to 58% – during the inflation years.

2.7 Newcomers to the labour market: immigrants and young people

When studying declining union density, newcomers to the labour market are of special interest. These are, first and foremost, young people and newly arrived immigrants.

In Norway, which is a member of the EEA, EU enlargement was followed by the arrival of large numbers of labour migrants from the new Eastern and Central European member states. Labour migrants have a much lower union density (32% in 2021) than other immigrants (43%) or native Norwegians (54%) (Nergaard & Ødegård 2024:28-29). Immigrants who arrived before 1990 have a higher union density (56%). The proportion of workers in workplaces without collective agreements is particularly high among labour migrants, who are also overrepresented in private services and live only temporarily in Norway. In addition, many of them are commuting or do not intend to stay in Norway. When comparing all foreign-born workers with native-born Norwegians, the union density gap in 2022 was almost the same as in 2010 or 2000: about 13-15 percentage points.

In the 1960s and 1970s, tens of thousands of Finnish labour migrants found work in the Swedish manufacturing industry, where they were strongly encouraged to join unions. In the last few decades, many of the migrants have been asylum seekers or labour migrants from non-EU countries with limited knowledge of Nordic trade unions and labour-market models. In blue-collar occupations, most of them end up in low-wage jobs in the private service sector, which is characterised by small workplaces without union representatives, high labour turnover and temporary contracts. Such structural shifts in the labour market combined with new migration patterns are not conducive to high levels of unionisation.

Regarding labour migrants from third countries (non-EU/EEA countries), the strong dependency on the employer during the first four years in Sweden (after which they have the right to permanent residency) has a deterrent effect on joining unions (Frödin & Kjellberg 2018). The high number of refugees who arrived from non-EU/EEA countries in 2015 (almost 160,000) added to the proportion of workers in vulnerable positions. In contrast to almost all other countries, asylum seekers in Sweden have the right to work pending an asylum ruling.

In Sweden, a unionisation gap between foreign-born and native blue-collar workers arose after 2006, when 77% of each category consisted of union members. In 2023, only 50% of foreign-born blue-collar workers were union members, compared to 62% of those born in Sweden. During the same period, the share of foreign-born blue-collar workers more than doubled (from 16% to 34%) and was even higher in the public sector (from 16% to 41%), where many immigrants have low-paid care jobs. In white-collar jobs, the proportion of foreign-born workers increased from 10% in 2006 to 20% in 2023.

It takes some time for migrants and young people to establish themselves in the labour market and join unions in greater numbers. Norwegian research shows that union density among both labour migrants and other immigrants increases considerably the longer they live in Norway (Nergaard & Ødegård 2024: 20-23, 33). Therefore, the rate of unionisation among migrants is expected to decline during periods when many newcomers arrive. Cools et al. (2021: 24) also found that union density among immigrants to Norway increased strongly over time after their arrival, but not so much that it reached the level of native Norwegians. More than half of the gap was explained by the labour-market characteristics of the migrants, most importantly that they tend to work in firms and industries with low union density.

Union density, both among immigrants and non-immigrants in Norway, is much higher in firms with collective agreements compared to those without agreements (Nergaard & Ødegård 2024:24-26). As labour migrants are less frequently covered by collective agreements and more often employed at workplaces without a union presence, this has a negative influence on their average rate of unionisation.

The high proportion of temporary contracts among blue-collar migrants and young people has a similar effect. In all Nordic countries, this is associated with the shift from manufacturing to low-paid jobs in private and public services.

There are several explanations for a union density gap between young and older workers, spanning from more individualistic attitudes and job characteristics of the young people to life course explanations (that the propensity to join unions varies with age). According to a Danish study, more young workers in 2014 than in 2002 thought that: “Trade unions are necessary for securing the interests of the workers” (Høgedahl & Møberg 2022:12). How can this be reconciled with the low union density? The same study shows that “many young workers are found in parts of the labour market with low trade union density and collective agreement coverage” (ibid:16). Generational differences almost disappear when variables related to the job characteristics of young people are taken into account.

Nergaard & Svarstad (2021) found that young workers in Norway are more likely to join unions if they find jobs at workplaces where it is common to be organised, in other words, workplaces where it is a social custom to be a union member. In this respect, there is a clear parallel with immigrants. In the private sector, only about one in four young workers were in workplaces with high union density in the late 2010s. The significance of social customs and union presence in the workplace is supported by the Swedish study Calmfors et al. (2021a-b), which found that ‘not being asked to join’ and ‘lack of information’ are more important reasons for not joining a union among young people and immigrants than among older workers and native Swedes.

Similarly, Ibsen et al. (2017) found found that the higher the union density, the higher the likelihood of new employees in Denmark joining a trade union. They identified a tipping point (somewhere between 45% and 65% union density) at which the social custom of union membership was self-sustaining. However, they also found that the inclination to join a union, particularly among young workers, also increased gradually at a lower workplace union density (ibid: 512-513). Toubøl & Jensen (2014) found that workplace union density – in practice correlating with union presence at workplace level – is the most important predictor of whether or not an employee is going to join a union.

A Swedish study shows similar results to the Danish one by Høgedahl & Møberg. The union density decline in later age cohorts is not associated with values changing over time (Vestin & Vulkan 2022: 24-25). Within all cohorts, union density increases up to 30 years of age, but in the cohorts born after 1970 the rise is not as steep and does not reach the same levels as in previous ones. There are some signs of more individualistic values, but they are not followed by a decline of trust in unions. On the contrary, the later, more individualistic cohorts have more trust in unions. Despite this, they are less inclined to join. The authors “suggest the need for a greater focus on structural and institutional factors” like the 2007 reform of the Swedish Ghent system (ibid:25). To reverse the trend in recent cohorts, they argue that “significant changes of structural and institutional incentives” are required.

If union density does not increase after the age of 30 and it is lower than before among the young, density will gradually decline as the younger cohorts grow older. There are indications of such cohort effects in Sweden. During the first two years of the remodelled Ghent system, union density declined almost twice as much among workers aged 16–29 years as among those aged 30 or older (Kjellberg 2019/2024, Table 8). More than a decade and a half later, the decline for the whole period from 2006 to 2023 was about the same in all age groups. Leaving aside the initial two years (from 2006 to 2008), the decline was largest among workers aged 30–44 and 45–64. There is a risk that union density will continue to decrease as the younger, less unionised cohorts gradually replace the older ones in the labour market.

3 Concluding remarks

As we have seen, the recruitment of newcomers to the labour market – young people and immigrants – is considerably easier when there is a union presence in the workplace. Consequently, the relatively low number of union representatives in industries where young people and immigrants are overrepresented makes the recruitment of new members more difficult. The erosion of the Ghent systems in Denmark (‘yellow’ unions and cross-occupational unemployment funds), Finland (the growth of the independent fund YTK) and Sweden (the growing membership of unemployment funds without joining a union) give workplace unions a key role in the recruitment of members. This is the case, in particular, for blue-collar unions with many low-paid members for whom the union ‘package’ is perceived as too expensive and supplementary income insurance schemes run by unions are of limited value.

By contrast, income insurance schemes that require union membership are an important recruitment tool for many white-collar unions. In Sweden, this is mainly the case in the private sector where the risk of unemployment is higher and employment protection weaker than in the public sector. It is no coincidence that Swedish white-collar union density 2006–2023 declined by eight percentage points in the public sector while it remained unchanged in the private sector.

A challenge for Swedish unions is the declining number of ‘union clubs’ and union workplace representatives. In 2019, the number of elected union representatives (all levels) was 254,000 compared to 360,000 in 1995. This decline is even more striking considering that the workforce grew considerably in the same period, and the proportion of small workplaces increased due to the growth of the private service industries.

In recent decades, the Nordic model of industrial relations has been challenged by the emergence of new industries with companies like Google, Spotify, Tesla and game developers with negative attitudes to unions and collective bargaining. Another tendency is the growth of long chains of subcontractors in construction and, to some extent, cleaning, in which there is exploitation of workers posted to Sweden (for example, Polish building workers), labour migrants (for example, cleaners from Mongolia) and other foreign-born workers (for example, asylum seekers with jobs). Berry picking is another example. Some companies use collective agreements as a façade – without following them in practice – and with no union members or local union representatives to supervise compliance with the agreements. In connection with sub-contracting and public procurement, the companies offering the lowest price usually win the contracts in industries with a high risk of unfair working conditions and unfair competition.

The first category of companies (ICT companies, etc.) is dominated by white-collar workers with relatively favourable wages and terms of employment, while the second one (cleaning, etc.) is dominated by blue-collar workers with low wages and less favourable working and employment conditions, in some cases connected to work-related crime.

The Nordic collective bargaining models are based on a high density of both unions and employers’ associations. There is a broad consensus in all countries that self-regulation, which requires strong labour market parties, is preferable to state regulation, among other things because it provides greater flexibility and adaptation to different industries and local conditions. Although the three Nordic Ghent countries still have a higher – or much higher – union density than Norway – the model is challenged in the long run by declining union density, above all in industries with a high share of low-wage – often foreign-born – workers with a weak individual bargaining strength. It is also in these industries that unfair competition, exploitation of vulnerable workers and work-related crime are gaining ground.

A growing dualisation of the labour force, clearly manifested in Sweden, happens when native Swedes have a low unemployment rate and high labour force participation at the same time as immigrants from non-EU/EEA countries have considerably higher rates of unemployment, low income, fixed-term work, precarious working conditions and lower union density (Bender 2023:204). The non-European proportion of the foreign-born workforce is higher in Sweden (57% of 2.2 million immigrants) than in other Nordic countries or the EU average (Konjunkturinstitutet 2024:73; SCB/Befolkning: Folkmängd efter födelseland 1900-2023.)

In Norway, there are serious concerns that a declining union density and coverage of collective agreements could threaten the front runner model. The tripartite Holden commissions III-IV addressed this (NOU 2023:30). The Norwegian employers’ associations have expressed a desire for increased union density, and it is also an issue of concern for the tripartite co-operation (see also Dølvik 2022).

To reverse the negative trend for union density, efforts to unionise immigrants and young workers should be given high priority. For the same reason, it is desirable to increase the proportion of the employees covered by ‘union clubs’/representatives. Young people and newly arrived immigrants should be given more information about trade unions and the Nordic labour market models. As a Norwegian study shows, tax-deductible union fees slow down or stop declines in union density, which is an argument in favour of the re-introduction of tax deductions for union dues in Sweden.

References

Ahtiainen, L. (2003, 2011, 2023). Palkansaajien järjestäytyminen Suomessa vuonna 2001, 2009, 2023. Helsinki: Työministeriö.

Andersen, S. K., Dølvik, J. E., & Ibsen, C. L. (2014). De nordiske aftalemodeller i åbne markeder – udfordringer og perspektiver. Oslo: Fafo-rapport 2014:08.

Barth, E., Bryson, A., & Dale-Olsen, H. (2025). Turning non-members into members: Do public subsidies to union membership matter? Journal of Economic Behavior and Organization. Retrieved from https://doi.org/10.1016/j.jebo.2024.106855

Bender, G. (2023). Swedish trade unions and migration. In L. Leopold et al. (Eds.), Migration and Integration in a Post-Pandemic World (pp. 199-231). Cham: Palgrave Macmillan. Retrieved from https://link.springer.com/chapter/10.1007/978-3-031-19153-4_15

Boeri, T., Brugiavini, A. & Calmfors, L. (2001). (Eds.). The role of unions in the Twenty-First Century. Oxford: Oxford University Press.

Böckerman, P. & Uusitalo, R. (2006). Erosion of the Ghent system and union membership decline: Lessons from Finland. British Journal of Industrial Relations, 44, 283-303.

Calmfors, L., Nederberg, J., & Persson, C. (2021a). Vad finns det för anledning att vara med i facket? Stockholm: Fores.

Calmfors, L., Nederberg, J., & Persson, C. (2021b). Vad finns det för motiv att vara med i facket? Ekonomisk Debatt, 49, 42-54.

Cools, S., Finseraas, H., & Rasmussen, M. B. (2020). The immigrant-native gap in union membership. Labour, 35, 24–51.

Dølvik, J. E. (2022). Strengthening the Nordic working life model. Fafo-paper 2022, 13.

Ebbinghaus, B. & Visser, J. (1999). When institutions matter: Union growth and decline in Western Europe, 1950-1995. European Sociological Review, 15, 135-158.

Ebbinghaus, B., Göbel, C., & Koos, S. (2011). Social capital, ‘Ghent’ and workplace contexts matter: Comparing union membership in Europe. European Journal of Industrial Relations, 17, 107-124.

Fazekas, Z. (2011). Institutional effects on the presence of trade unions at the workplace: Moderation in a multilevel setting. European Journal of Industrial Relations, 17, 153-169.

Frödin, O. & Kjellberg, A. (2018). Labour migration from third countries to Swedish low-wage jobs. Nordic Journal of Working Life Studies, 8, 1-20.

Høgedahl, L. & Møberg, R. J. (2022). Young workers in transition: Explaining the density gap by a life-course perspective. Nordic Journal of Working Life Studies, 12, 8.

Ibsen, C. L., Toubøl, J., & Jensen, D. S. (2017). Social customs and trade union membership. European Sociological Review, 33, 504–517.

Ibsen, F., Høgedahl, L., & Scheuer, S. (2013). Free riders: The rise of alternative unionism in Denmark. Industrial Relations Journal, 44, 5-6.

Ilsøe, A. (2013). Dealing with alternatively organized workers. Nordic Journal of Working Life Studies, 3, 85-107.

Kauhanen, A. (2025). How Extension of Collective Agreements Affects Wages. Nordic Economic Policy Review.

Kjellberg, A. (2001/2017). Fackliga organisationer och medlemmar i dagens Sverige. Lund: Arkiv förlag. Retrieved from https://lucris.lub.lu.se/ws/portalfiles/portal/34614668/Fackliga_organisationer_och_medlemmar_i_Sverige2017.pdf

Kjellberg, A. (2011). The Decline in Swedish Union Density since 2007. Nordic Journal of Working Life Studies, 1(1), 67–93. Retrieved from https://portal.research.lu.se/sv/publications/kollektivavtalens-t%C3%A4ckningsgrad-och-vita-fl%C3%A4ckar

Kjellberg, A. (2017). Self-regulation versus State Regulation in Swedish Industrial Relations. In Rönnmar, M. & Votinius, J. J. (eds.) Festskrift till Ann Numhauser-Henning. Lund: Juristförlaget i Lund. Retrieved from https://lup.lub.lu.se/search/publication/ac5caadc-5869-4753-ae32-9cae40377d4a

Kjellberg, A. (2019/2024). Kollektivavtalens täckningsgrad samt organisationsgraden hos arbetsgivarförbund och fackförbund. Lund: Department of Sociology. Retrieved from https://lup.lub.lu.se/search/publication/384bb031-c144-442b-a02b-44099819d605

Kjellberg, A. (2022). Facklig organisationsgrad ur ett nordiskt perspektiv. Lund: Department of Sociology. Retrieved from https://lup.lub.lu.se/search/publication/1b19000e-282c-45f3-9820-37cf4638210d

Kjellberg, A. (2023a). Kollektivavtalens täckningsgrad och vita fläckar. Stockholm: Arena Idé 2023. Retrieved from https://portal.research.lu.se/sv/publications/kollektivavtalens-t%C3%A4ckningsgrad-och-vita-fl%C3%A4ckar

Kjellberg, A. (2023b). Sweden: still high union density, but widening gaps by social category and national origin”. In Müller, T., Vandaele, K. & Waddington, J. (eds.). Trade Unions in the European Union. Brussels: Etui/Peter Lang.

Kjellberg, A. (2024). Den svenska partsmodellen ur ett nordiskt perspektiv: facklig anslutning och lönebildning. Stockholm: Arena Idé. Retrieved from https://portal.research.lu.se/sv/publications/den-svenska-partsmodellen-ur-ett-nordiskt-perspektiv-facklig-ansl

Kjellberg, A. & Ibsen, C. L. (2016). Attacks on union organizing: Reversible and irreversible changes to the Ghent-systems in Sweden and Denmark. In T. P. Larsen & A. Ilsøe (eds.). Den Danske Model set udefra, 279-302. Copenhagen: Jurist- og Økonomforbundets Forlag. Retrieved from https://portal.research.lu.se/sv/publications/attacks-on-union-organizing-reversible-and-irreversible-changes-t

Kjellberg, A. & Nergaard, K. (2022). Union Density in Norway and Sweden: Stability versus Decline, Nordic journal of working life studies 2022. Retrieved from https://portal.research.lu.se/sv/publications/union-density-in-norway-and-sweden-stability-versus-decline

Konjunkturinstitutet. (2024). Lönebildningsrapporten 2024. Stockholm: Konjunkturinstitutet

Lindellee, J. & Berglund, T. (2022). The Ghent system in transition: unions’ evolving role in Sweden’s multi-pillar unemployment benefit system. Transfer 28(2).

Nergaard, K. (2020a). Holdninger til fagorganisering. Oslo: Fafo-rapport 2020:33.

Nergaard, K. (2020b). Attitudes to unionisation. English summary of Fafo-rapport 2020:33.

Nergaard, K. (2024). Organisasjonsgrader, tariffavtaledekning og arbeidskonflikter 2022. Oslo: Fafo-notat 2024: 05.

Nergaard, K. & Stokke, T. A. (2010). Organisasjonsgrader og tariffavtaledekning i norsk arbeidsliv 2008. Oslo: Fafo.

Nergaard, K. & Svarstad, E. (2021). Fagorganisering blant unge nyansatte – hva betyr «et kollektivt trykk»?, Søkelys på arbeidslivet, 38(2), 81-98.

Nergaard, K. & Ødegård, A. M. (2022). Organisasjonsgrad blant innvandrere. Oslo: Fafo.

Nergaard, K. & Ødegård, A. M. (2024). Organisasjonsgrad blant innvandrere 2021. Oslo: Fafo.

NOU 2023:30. Utfordringer for lønnsdannelsen og norsk økonomi. Finansdepartementet 15. desember 2023. Oslo.

Rasmussen, M. B. & Pontusson, J. (2018). Working-Class Strength by Institutional Design? Comparative Political Studies, 51(6), 793-828.

Shin, Y-K. & Böckerman, P. (2019). ’Precarious workers' choices about unemployment insurance membership after the Ghent system reform: The Finnish experience'. Social Policy and Administration, 53(7), 921-938.

Stokke, T. A. (2000). Organisasjonsgrader i norsk arbeidsliv 1945–1998. Oslo: Fafo.

Svalund, J. & Berglund, T. (2018). Fixed-term employment in Norway and Sweden: A pathway to labour market marginalization?, European journal of industrial relations, 24(3), 261-277.

Toubøl, J. & Jensen, C. S. (2014). Why do people join trade unions? The impact of workplace union density on union recruitment. Transfer, 20(1), 135-154.

Vestin, E. & Vulkan, P. (2022). Cohort Differences in Swedish Union Membership 1956–2019 and the Role of Individualization. Nordic journal of working life studies.

Visser, J. (2002). Why fewer workers join unions in Europe: a social custom explanation of membership trends. British journal of industrial relations, 40(3), 403-430.