Nordic Economic Policy Review 2025

The extension of collective agreements has generally modest positive effects on wages, particularly for lower-paid workers, but the impacts vary from sector to sector and may involve employment trade-offs. In Norway, most studies have found small positive wage effects in construction and shipbuilding, with mixed results in cleaning. Finland’s move to firm-level bargaining in forestry led to minimal overall changes, except for blue-collar workers in the paper industry. Notably, in this case, the abolishment of generally binding collective agreements did not lead to wage cuts. Evidence from Portugal and Italy shows a 1-2% increase in wages but similar employment declines due to extension. This highlights the wage-employment trade-off that policymakers must consider. Declining unionisation may pose a challenge to the legitimacy of extension, while posted workers and concerns about labour standards could support it. Overall, extension appears to be a useful but limited tool for supporting wages, with modest benefits to weigh against potential job losses.

Keywords: Collective bargaining, Wages, Posted workers

Acknowledgements: I would like to thank Torbjörn Hållö, Elin Svarstad and Roope Uusitalo for their comments.

Evidence from Nordic and other European nations suggests that the extension of collective bargaining agreements has generally modest positive effects on wages, particularly for lower-paid workers. However, the impact can vary from sector to sector and may result in employment trade-offs. In Norway, most studies have identified small positive wage effects from extensions in construction and shipbuilding, although the results were mixed for the cleaning sector. The Finnish move to firm-level bargaining in the forestry industry led to minimal changes in wages and wage dispersion, with only blue-collar workers in the paper industry receiving notable increases. In the Finnish case, the abolition of generally binding collective agreements did not lead to wage cuts. Evidence from other European countries, such as Portugal and Italy, indicates wage increases of 1-2% due to extension, but employment declines of a similar magnitude. This highlights the potential trade-off between higher wages and job losses that policymakers must consider. Looking ahead, declining unionisation rates in Finland may challenge the legitimacy of widespread extension. However, the increasing prevalence of posted workers and concerns about maintaining labour standards could support the use of extension mechanisms in other Nordic countries. Overall, the extension of collective agreements appears to be a useful but limited tool for supporting wage levels, with modest positive impacts that must be weighed against the potential negative effects on employment.

In principle, collective bargaining agreements bind employers who have signed the contract or belong to the employers’ association that signed it and employees who belong to the labour union that signed. In Nordic countries, however, collective agreements have what is known as erga omnes applicability, meaning that contracts apply to all workers in a firm, irrespective of whether they belong to a union. 

These contracts can be extended to non-signatory employers either voluntarily by the employer complying with the contract even though the company does not belong to the employers’ association (e.g. hängavtal in Sweden) or compulsorily by the government or courts extending the contract to all employers in an industry or occupation. Although in the academic literature the voluntary application of collective agreements by firms that do not belong to employers’ associations is seen as a particular type of extension mechanism, the actors in the labour markets do not necessarily share this view: they see it as another form of collective agreement.

See Scheller (2018) for more details about Hängavtal in Sweden.

The Nordic countries have different approaches to extension. Compulsory extension is possible in Finland, Iceland and, under certain circumstances, in Norway. Voluntary extension is common in Sweden and Denmark, but compulsory extension is not possible in these two countries. These differences reflect the differences between the national labour markets. For example, in Finland, the state has historically played a larger role in the labour market than in the other Nordic countries. In Denmark and especially in Sweden, the labour market parties have not favoured state intervention.

The extension of collective agreements is topical in the Nordic countries for two reasons. The first is the growing importance of what are known as ‘posted workers’, i.e. people working temporarily in Nordic countries while employed by a company operating from another EU member state. The employers of posted workers are not usually members of an employers’ association and, without extension, are not covered by collective agreements. The worry is that working conditions for posted workers are poorer than those stipulated in the collective agreement. The extension of collective agreements is one way to prevent such “social dumping.” The second reason concerns Finland, where the forestry industry abandoned sectoral bargaining, which also meant abolishing the generally binding collective agreement in the sector. This raised the question of how to guarantee minimum labour standards when there is no minimum wage and no generally binding collective agreement.

The remainder of this paper is organised as follows. I begin by discussing the arguments for and against the extension of collective agreements. This is followed by a brief description of compulsory and voluntary extension in Nordic countries. The main part of the article discusses Nordic evidence of the impact of extension on wages. I conclude by discussing the potential future for extensions in the Nordic countries.

The extension of collective agreements can have various impacts on the labour market. In general, collective agreements increase wages and reduce wage dispersion (See, e.g. Card et al., 2004). Fanfani (2023) and Card and Cardoso (2022) are recent European studies showing how collective agreements increase actual wages. In both studies, a 1% increase in the collectively bargained wage leads to about a 0.5% increase in actual wages.

The extension of collective agreements may reduce wage dispersion. Extensions bind all firms in a given industry, which should homogenise wages compared to a situation where some firms are not bound by any collective agreement or are bound by firm-level agreements. 

Extension is also often seen as “levelling the playing field” by making it impossible for workers or firms to compete by undercutting labour standards. This has been discussed in the case of posted workers in Nordic countries, where the worry has been that firms operating from other European countries may have an unfair competitive advantage if they are allowed to operate in Nordic countries while adhering to standards in their country of origin. This has been especially relevant in construction, where European firms have entered the Nordic market and often paid their workers less than their Nordic competitors have to pay under collective agreements. 

As a means of ensuring minimum standards in the labour market, the extension of collective agreements is a more flexible alternative to legislation. For example, a national minimum wage imposes a single minimum wage, whereas collective agreements stipulate different wage floors for different industries and occupations. Collective agreements are also easier to renegotiate than the political process of changing the minimum wage. However, in practice, it is difficult to quantify the difference between the two.

By imposing higher wages throughout an industry, an extension may have negative employment effects. Fanfani (2023) shows that collective bargaining agreements that led to wage increases in Italy also led to substantial job losses. Fanfani (2023) is relevant in this context because the minimum contractual wage levels are generally binding, although other provisions in the collective agreement are not. Martins (2021) studies the economic effects of the extension of collective agreements in Portugal and finds economically meaningful employment losses. Both studies show that the negative employment effects stem from reduced job creation. I discuss these studies in more detail later. Magruder (2012) shows that large firms in South Africa negotiate collective agreements, which lead to large job losses when extended to smaller firms.

Some see extensions as problematic from the perspective of basic rights. Firms are forced to comply with a contract negotiated by an employers’ association to which they do not belong. This is seen as breaching the negative freedom of association, that is, the right of firms to decide whether to belong to an employers’ association. However, the position of the International Labor Organization (ILO) and many European countries (for example, Germany) is that extensions of collective agreements do not violate negative freedom of association.

Extension may also affect the organisation of labour and firms. It increases the incentives for firms to join an employers’ association since the contract will bind them anyway, and as members they may be able to influence the stance of the association at the bargaining table. Extension reduces the incentives for workers to join trade unions since they receive the benefits of collective agreements anyway. However, in the Nordic context, this argument does not really apply, since the contracts always apply to all workers in the firm, irrespective of whether they are union members.

I now turn to a brief discussion of extension in the Nordic countries. A more elaborate discussion from a legal perspective is found in Bruun (2018).

In Finland, collective agreements are often extended to non-signatory parties. The decision is made by an independent committee under the auspices of the Ministry of Social and Health Affairs. Collective agreements must be submitted to this committee and it decides whether to extend the contract. The parties to the contract do not need to ask for it to be extended. The committee considers several criteria, but the contract’s coverage is the key determinant for extension. Other criteria include the stability of collective bargaining in the sector, the degree of organisation on both sides and the objective of using generally binding collective agreements as a safeguard for minimum working conditions for employees of companies not in employers’ associations. From the perspective of employers, almost the entire contract is extended. Some local bargaining options are not available to firms that are not members of employers’ associations.

This was changed in December 2024 when parliament approved changes to legislation that give all firms the same possibilities for local bargaining.

For employees, only the rights are extended, not the obligations. This means that industrial peace clauses in contracts cannot be extended. Only nationwide contracts can be extended, and thus, when the forest industry abandoned sectoral bargaining in 2022, it also meant that there was no longer a generally binding contract in that sector. This is the biggest change with respect to the extension of collective agreements in Finland. In a later section, I discuss the impact it had on wages.

In Norway, it is possible to extend contracts to foreign employers if it can be shown that they pay their foreign workers less than their Norwegian counterparts. This law was enacted in the 1990s when Norway joined the European Economic Area, but the first extensions were after the 2004 enlargement of the European Union. The decision is made by a committee that has one member from the employers’ side, one from the employees’ side, and three independent members. At least one party to the contract must ask for an extension. In principle, the whole contract can be extended, but in practice, trade unions have only asked for partial extensions, typically concerning wage levels.

The key extensions have been in the construction sector, where the contracts were first extended regionally in 2005–2006 and then nationally in 2007; shipbuilding, where the contract has been extended since 2008; and cleaning, where extension started in 2011. In 2015 and 2018, contracts were extended in several service sectors and the fishing industry.

See Flaarønning (Forthcoming), Table 1 for more details on the timing of extension of the contracts.

I will conduct an in-depth review of the Norwegian studies.

In Sweden, extension is not an option, but unions often force employers to enter into a collective agreement. The limits of the power of unions to force foreign firms to adhere to collective agreements have been widely discussed in Sweden. This started in 2007 when the European Court of Justice (ECJ) ruled against the construction workers’ union in a case where it had organised a ‘blockade’ of a Latvian construction firm to force it to accept the collective agreement in the sector (for more details about the ‘Laval case’, see, for example, Bruun (2018)). The Posted Workers Directive stipulates that posted workers cannot be paid below the legal minimum in the host country. However, as Sweden does not have a legal minimum wage, nor are the contracts extended to non-signatory parties, the company in question should have been allowed to adhere to the labour standards in the country of origin. This ruling led to the setting up of a commission to consider how to safeguard labour standards in Sweden within the EU’s regulatory framework. The commission saw that the extension of collective agreements, or a legal minimum wage, would not be consistent with the Swedish model of industrial relations. It recommended that if the collective agreement included a clearly defined minimum wage, then unions would be able to use a collective agreement to impose it on foreign firms if they did not already pay wages higher than the minimum. However, in practice, as a consequence of the ECJ ruling, the number of foreign companies covered by collective agreements has decreased dramatically (Arnholtz, 2023). In effect, trade unions’ rights to engage in collective action against foreign firms were restricted, which led to firms perceiving the threat of collective action against them as less credible (Arnholtz, 2023, p. 381).

In Denmark, extension is not an option, even for posted workers. The Danish response to the Laval ruling was to set up a tripartite commission that maintained the option of industrial action by unions against foreign firms to pressure them to sign collective agreements. However, this was under the condition that the collective agreement must be nationally representative and the wage demands transparent. The transparency requirement means that unions cannot demand local wage negotiations, which are an important part of the Danish wage-setting model. In practice, this means that unions can only demand the minimum wages stipulated in the collective agreement.

In Iceland, all collective agreements are binding (Ólafsdóttir, 2020).

Studying the causal effects of the extension of collective agreements on wages is difficult. The key problem is that while it is possible to observe what happens to the wages of individuals affected by an extension, it is impossible to know what would have happened to their wages if the agreement had not been extended. For example, if the data shows that wages increased by an average of 4% in an industry following an extension, one cannot conclude that the effect of extension is 4% since it is likely that the wages would have grown without the extension, let’s say by 2%.

In econometric parlance, the group that is subject to a policy is called the “treatment group” or the “treated group”. In this case, it is the individuals or sectors affected by the extension of collective agreements. A comparison group is used to measure what would have happened to the treatment group in the absence of the policy change. A good comparison group is one in which wages have developed similarly to the treatment group and have a similar reaction to economic shocks. Continuing with the example; if wages increased by 3% in industries that are similar to the industry with the extension, one can conclude that the causal effect of the extension was in the region of 1% (4%-3%). This method is called difference-in-differences. In the ideal case, there would be many industries where contracts would start to be extended at a single point in time and many comparison industries. It would then be possible to obtain statistically precise estimates of the causal effects. If the ‘treatments’ are at different points in time, this must be taken into account in the analysis; otherwise, the results may be biased (e.g. Goodman-Bacon, 2021, de Chaisemartin and d'Haultfoeuille, 2020, Callaway and Sant’Anna, 2021).

If there are only a few treated industries, or only one, it is more difficult to obtain precise estimates. In this case, the comparison group can be a weighted average of the industries in which contracts are not extended, and the weights are such that the wage trend in the comparison group is similar to that of the treated industry. This is called the synthetic control method (Abadie et al., 2010), and a slightly generalised version of it is known as synthetic difference-in-differences (Arkhangelsky et al., 2021). The following studies use either difference-in-differences, synthetic control, or synthetic difference-in-differences to study the effects of extensions.

An extension’s wage effects are likely to be context-specific, at least to some extent, since many institutional features affect how strongly they can affect wages. First, as we have seen, the nature of extension differs in different contexts. In some cases, almost the entire collective agreement is extended (for example, Finland), whereas in others, only the minimum wages are extended (for example, Norway). If the non-wage clauses of collective agreements have an impact on labour costs, the effects may be greater when the whole contract is extended. Second, the wage effects of an extension will probably be greater when the extended minimum wage is high relative to the average wage in the industry. For this reason, most studies on the effects of minimum wages use some kind of measure of the “bite” of the minimum wages, that is, some measure of how high the minimum wage is compared to the average (or median) wage (see e.g. Neumark and Wascher, 2008, Chapter 3). For these reasons, the estimated impacts of extension do not necessarily tell us anything informative about the impacts in other contexts.

As described above, in Norway, collective agreements have been extended in sectors that have seen an influx of foreign workers following EU enlargement in 2004. The effects of these extensions on wages have been widely studied. 

Eldring et al. (2011) provide a descriptive statistical analysis of wages and the extension of collective agreements in shipbuilding, construction and electrical installation in 2004–2009. Their results show that the proportion of employees receiving lower wages than the minimum stipulated by the collective agreement is smaller after extensions. They also show that wage distribution in the construction industry is more compressed after the extension.

Skjerpen et al. (2016) provide a more thorough statistical analysis of the wage impacts of the extension of collective agreements. They use random-effects panel data regressions to study how the extension of collective agreements affects wages in construction, shipbuilding and cleaning. Using data from 1997 to 2012, they find that extensions are associated with higher wages in construction and shipbuilding as well as for part-time workers in cleaning. Curiously, they find a negative association between wages and extension for full-time workers in cleaning.

Bratsberg and Holden (2015) focus on construction and study the impact of extension on wages in 2005–2007. To identify the impact of extension on the level of wages, they utilise the fact that extension was introduced at different times in different parts of the country. This allows them to conduct a difference-in-differences analysis. Their results show that extensions led to fewer employees receiving wages below the minimum stipulated in the collective agreement and that average hourly wages rose slightly. However, recent research has called into question the validity of the method used in this study. Several studies show that the fixed-effects methods used may lead to biased estimates in the case of staggered adoption of the treatment (e.g. Callaway and Sant’Anna, 2021, Goodman-Bacon, 2021, de Chaisemartin and d'Haultfoeuille, 2020). It would be interesting to see what the results look like using these more modern methods.

Knutsen (2022) studies the impact of extension on wages in shipbuilding, construction and cleaning. She studies the extensions in 2006/2007 (construction), 2009 (shipbuilding) and 2012 (cleaning). She uses fixed effects methods, as did Bratsberg and Holden (2015). Her results show that in cleaning, hourly and annual wages decreased, and hours worked increased. In construction, hourly wages increased, but hours worked decreased so much that annual wages decreased. The results are similar for shipbuilding, but the impact on annual wages is not statistically significant. She complements these analyses with industry-level analyses using the synthetic control method in order to form a comparison group of other industries that resembles, as far as possible, the ones in which agreements were extended (the treated industries). In practice, this means that the weights given to the different industries in the control group are chosen so that over time earnings in the synthetic control group develop similarly to the treated group. The results show that annual wages may have increased in shipbuilding following the extension and declined in cleaning and construction. Thus, the results are similar to the individual-level results for cleaning and construction but different for shipbuilding. However, the statistical significance of the results of this study is not clear.

Benedictow et al. (2021) study the same extensions as Knutsen (2022). They study hourly wages and find small positive impacts on wages in construction and shipbuilding. In cleaning, the estimate is also positive but not statistically significant. Strangely, they use different timing for the extension in shipbuilding and cleaning than Knutsen (2022). This probably reflects the fact that the extensions were not introduced at the beginning of the year, and the researchers have made different choices as to how they should construct the periods before and after.

It is a general problem with all studies of Norwegian extensions that they have only been introduced in a few sectors. The statistical theory behind the estimates assumes that they have been introduced in many industries. This casts doubt on the validity of the statistical analyses. Moreover, Knutsen (2022) is the only study that considers the fact that extension applies to all individuals in a given industry in statistical inference. This is the correct way to perform statistical inference, since the individual observations in a given industry are not independent of each other. Not taking this into account may lead to severely biased statistical inference, most probably overstating the statistical significance of the results. The synthetic control method addresses this shortcoming, but Knutsen (2022), for example, does not perform formal statistical inference.

Another problem highlighted by Knutsen (2022) is the difficulty of measuring hourly wages as they are not recorded in the registers used in the studies and have to be calculated by dividing either monthly or annual wages by working hours, and working hours are measured very imprecisely. The measurement error may lead to imprecise estimates if it is random or biased estimates if it is systematic.

Some studies consider aspects of extensions other than their impact on wages. Dapi (2016, Chapter 2) studies the impact of the extension of collective agreements in the cleaning industry on the number of temporary migrants employed in the industry. He uses the synthetic control method, and his results show that extension reduced the number of temporary migrants employed in this industry by about 20 % one year after the introduction of the minimum wage in the industry. Dapi (2016, Chapter 3) again studies the extension in the cleaning industry, this time the impact of it on the probability of job separation for workers who worked in the industry before it. He uses monthly data, and the impact estimate is the difference between the change in the average survival rate of a job from January–February in 2011 to September–December in 2011 and between the change in the average survival rate of a job from January–February in 2007 to 2010 to September–December in 2007–2010. His results show that low-paid workers are more likely to stay in their jobs after an extension.

Flaarønning (Forthcoming) studies the impact of Norwegian extensions on union density. His argument is that extension reduces the benefits of being a union member because all employees receive the benefits of collective agreements, which reduces the financial incentives to join a union. He uses difference-in-differences analysis that takes into account that extensions are adopted at different times in different industries. In his analyses, he uses only the industries in which the collective agreement is eventually extended. Thus, at any point in time, the treated group consists of industries in which the collective agreement has been extended, and the comparison group consists of industries in which it has not yet been extended. His results show that extensions lead to about a 3% decrease in union density. The analysis is otherwise solid, but the statistical inference does not consider the fact that individual observations within an industry are not independent, and thus, the results may seem more precise than they are in reality.

To summarise, several studies have found positive impacts on wages in the construction and shipbuilding sectors, with evidence of fewer employees being paid below the stipulated minimum and slight increases in average hourly wages. However, the effects in the cleaning sector are mixed, with some studies showing negative impacts on wages and increased working hours.

In October 2020, the Finnish Forest Industries Federation (FFIF) announced that it would no longer be a party to collective agreements. In practice, this meant that the forestry industry would move to company-specific agreements when the existing collective agreements expired. It would now also be possible for firms not to negotiate collective agreements at all. However, the coverage of collective agreements remains high, at 100% for blue-collar workers in the paper industries and about 93% of all employees in mechanical forest industries (Jutila Roon, 2024). 

This was a significant change, as it ended the long tradition of union agreements in a major industrial sector. It would also allow for significant changes to collective agreements, as there would no longer be a generally binding collective agreement in the forestry industry, and existing agreements would have no ‘ex-post effect’ (no ultra-activity). 

FFIF’s announcement sparked a heated debate. The labour unions in the sector condemned the move to company-specific agreements, and many commentators predicted a significant drop in wages because of company-specific agreements and the abolition of generally binding contracts.

Examples are https://www.hs.fi/politiikka/art-2000008361102.html, https://yle.fi/a/3-11574181, https://www.iltalehti.fi/politiikka/a/fa55e683-7cca-4c93-9085-7742f19426f4

In March 2021, Technology Industries of Finland announced that it would move to a hybrid model with company-specific agreements in addition to a collective sectoral agreement. In practice, this was only relevant for the software sector, where the coverage of a sectoral agreement would no longer be high enough for it to be declared generally binding. It was estimated that the collective agreement covered roughly 30% of employees in the sector, and based on this, the committee decided that the contract would no longer be generally binding.

https://finlex.fi/fi/viranomaiset/tyoehto/paatokset/2022/4602

These changes are part of a European trend towards the decentralisation of bargaining systems, that is, bargaining closer to the enterprise level. However, the shift from sectoral to company-level agreements in the forestry industry is a significant change, even from a European perspective.

Kauhanen (2024) studies what happens to the level and dispersion of wages when sectors move to firm-level bargaining and contracts cease to be generally binding. He uses administrative data on monthly wages and the synthetic difference-in-differences method. The method is suitable in this case, where there are only a few industries in which the collective bargaining system changes, many possible comparison industries, and a fairly long time period (January 2019–June 2023). This method generalises the synthetic control method, but the main idea is the same: the comparison group resembles the treated industry as far as possible.

His results show that moving to firm-level bargaining did not lead to major changes in the level of wages or wage dispersion. Most of the estimated effects are positive, but only a few are statistically significant. The results show that blue-collar employees in the paper industries were the only ones whose wages and within-firm wage dispersion increased after the change. For this group, the estimated effect on their average monthly wage is €296, which is about 6% of the average wage in the industry. Within-firm wage dispersion increased by 0.17 standard deviations, which is quite significant.

The findings presented here are limited to the short term. The long-term consequences may vary owing to several factors. First, company-specific agreements allow for more tailored pay systems, but these will require time to implement. Second, such agreements could potentially boost productivity by allowing for custom work schedules that better align with company needs. However, any gains in productivity and subsequent wage increases will only become apparent over an extended period.

Wages are also typically rigid downwards (Dickens et al., 2007), and firms are reluctant to cut wages even in institutional settings where it is possible (Du Caju et al., 2015, Galuscak et al., 2012). This means that the wage effects of extension and abolition of extension are probably not symmetric; extension may increase wages more than abolition would decrease them.

Kauhanen (2025) considers the same situation but focuses on the wage differences among groups of employees. He considers the differences between men and women, high and low educated and different age groups. The results show that the decentralisation of bargaining did not lead to increased wage differences between men and women or between high and low educated, but there is some evidence of increased wage differences between different age groups in IT services.

Overall, the move to firm-level bargaining and the abolition of generally binding collective agreements led to very small changes in wages. However, it did not lead to wage decreases, as might have been expected based on studies from other countries. One reason for this is that the coverage of collective agreements remained very high, although the contracts are now negotiated at the firm level. Importantly, it is still a nationwide union that negotiates contracts. This guarantees that there is some coordination among the contracts negotiated in different firms. In other circumstances, where the unions are weaker, the effects could have been different.

Skedinger (2010) contains an analysis of the impact of posted workers on the wages of Swedish workers in the construction sector. He studies this by comparing the average wages in 25 Swedish regions, where the number of posted workers varies both cross-sectionally and over time. He measures the importance of posted workers in a region using the number of collective agreements that Swedish unions have concluded with foreign firms. His results show that posted workers do not receive depressed wages in the Swedish construction industry. This might be because the influx of posted workers has been so small or because unions have been successful in forcing firms to accept the wage level stipulated in the collective agreement.

Martins (2021) studies the impact of extensions of collective agreements on wages, employment and the prevalence of contract work. Extensions are common in Portugal. He studies the time period 2007–2011, during which about 90% of the sectoral collective agreements were extended. The effects of the extension are identified by the timing of them. They were introduced in different months, and thus, for example, the employment change in the industries subject to extensions is compared with the employment change in industries not subject to extension, at least not yet.

The results show that continuing workers receive higher wages (1–2% depending on the specification), employment in the industry decreases by 2%, and the use of contractors (not subject to collective bargaining) increases by 1%.

Fanfani (2023) studies how wage increases stipulated in collective agreements affect actual wages and employment in Italy from 2006 to 2016. This study is relevant because minimum contractual wage levels are generally binding in Italy, as explained earlier. Similar to Martins (2021), he utilises the differential timing of the signing of collective agreements to identify the effects. The results show that a 1% increase in the base wage levels leads to a 0.45% change in the actual wage and about a 0.36% reduction in employment.

Developments in the Nordic labour markets affect the future of extensions to collective agreements. Nordic countries have largely relied on collective agreements to guarantee minimum standards in the labour market, but several trends threaten this reliance. In Finland, the decrease in unionisation threatens the legitimacy of extension. If only a minority of employees belong to unions, why should collective agreements determine the working conditions of all employees in an industry? There are, however, countries such as France where the coverage of collective agreements has remained very high even though union density is low. On the other hand, if the density of employers’ associations diminishes, the case for extension as a means of underpinning minimum standards in the labour market increases. The increasing role of posted workers may also support extension, but the labour market parties resist extension in Sweden and Denmark. Similarly, if the role of collective agreements as a guarantee of minimum standards in the labour market diminishes, the discussion of the need for a legal minimum wage will intensify. So far, labour unions in the Nordic countries have been sceptical of statutory minimum wages, although unions in other European countries have more positive attitudes toward them (Furåker, 2020).

Evidence from Nordic and other European nations suggests that the extension of collective agreements has generally had modest positive effects on wages, particularly for lower-paid workers. However, these impacts can vary across sectors and may result in employment trade-offs. In Norway, most studies found small positive wage effects from extensions in construction and shipbuilding, although the results were mixed for the cleaning sector. The Finnish move to firm-level bargaining in the forestry industry led to minimal changes in wages and wage dispersion, with only blue-collar workers in the paper industry receiving notable increases. Thus, in this case, the abolition of extension did not lead to lower wages, as one might have expected based on studies from other countries. However, the concurrent shift towards firm-level negotiations complicates the comparison. Evidence from other European countries, such as Portugal and Italy, indicates wage increases of 1–2% due to extensions but employment declines of a similar magnitude.

This highlights the potential trade-off between higher wages and job losses that policymakers must consider. Looking ahead, declining unionisation rates in Nordic countries may challenge the legitimacy of widespread extension, particularly in Finland, where contracts are often extended. However, the increasing prevalence of posted workers and concerns about maintaining labour standards could support the continued use of extension mechanisms. Overall, the extension of collective agreements appears to be a useful but limited tool for supporting wage levels, with modest positive impacts that must be weighed against the potential negative employment effects.

Abadie, A., Diamond, A., & Hainmueller, J. (2010). Synthetic control methods for comparative case studies: Estimating the effect of California’s tobacco control program. Journal of the American Statistical Association, 105, 493-505.

Arkhengelsky, D., Athey, S., Hirshberg, D. A., Imbens, G. W., & Wager, S. (2021). Synthetic Difference-in-Differences. American Economic Review, 111, 4088-4118.

Arnholtz, J. (2023). The embedded flexibility of Nordic labor market models under pressure from EU-induced dualization – The case of posted work in Denmark and Sweden. Regulation & Governance, 17, 372-388.

Benedictow, A., Dapi, B., Winger Eggen, F., Gjefsen, H. M., Kjelsrud, A., Måøy, J., Tofteng, M., & Svarstad, E. (2021). Evaluering av ordningen med allmenngjøring av tariffavtaler. Samfunnsøkonomisk analyse. Rapport 13-2021. Oslo.

Bratsberg, B. & Holden, M. B. (2015). Effekter av allmenngjøring i byggebransjen. Samfunnsøkonomen, 29, 68-80.

Bruun, N. (2018). Extension of collective agreements: The Nordic situation. In: Hayter, S. & Visser, J. (eds.) Collective Agreements: Extending Labour Protection. International Labour Organization.

Callaway, B. & Sant’Anna, P. H. C. (2021). Difference-in-Differences with multiple time periods. Journal of Econometrics, 225, 200-230.

Card, D. & Cardoso, A. R. (2022). Wage flexibility under sectoral bargaining. Journal of the European Economic Association, 20, 2013-2061.

Card, D., Lemieux, T., & Riddell, W. C. (2004). Unions and wage inequality. Journal of Labor Research, 25, 519-559.

Dapi, B. (2016). Essays on wage cyclicality and regulation, PhD thesis, University of Oslo.

De Chaisemartin, C. & D'Haultfoeuille, X. (2020). Two-way fixed effects estimators with heterogeneous treatment effects. American Economic Review, 110, 2964-2996.

Dickens, W. T., Goette, L., Groshen, E. L., Holden, S., Messina, J., Schweitzer, M., Turunen, J., & Ward, M. (2007). How wages change: Micro evidence from the International Wage Flexibility Project. Journal of Economic Perspectives, 21, 195-214.

Du Caju, P., Kosma, T., Lawless, M., Messina, J., & Rõõm, T. (2015). Why firms avoid cutting wages: Survey evidence from European firms. ILR Review, 68, 862-888.

Eldring, L., Ødegård, A. M., Andersen, R. K., Bråten, M., Nergaard, K., & Alsos, K. (2011). Evaluering av tiltak mot sosial dumping. FAFO-rapport.

Fanfani, B. (2023). The employment effects of collective wage bargaining. Journal of Public Economics, 227, 105006.

Flaarønning, T. (Forthcoming). Union decline through extension of collective agreements? British Journal of Industrial Relations, n/a.

Furåker, B. (2020). The issue of statutory minimum wages: Views among Nordic trade unions. Economic and Industrial Democracy, 41, 419-435.

Galuscak, K., Keeney, M., Nicolitsas, D., Smets, F., Strzelecki, P., & Vodopivec, M. (2012). The determination of wages of newly hired employees: Survey evidence on internal versus external factors. Labour Economics, 19, 802-812.

Goodman-Bacon, A. (2021). Difference-in-differences with variation in treatment timing. Journal of Econometrics, 225, 254-277.

Jutila Roon, M. (2024). Yrityskohtaiset työehtosopimukset. Selvitys neuvotteluista metsäteollisuudessa ja tietoalalla.Teollisuuden Palkansaajat analyysi 3/2024.

Kauhanen, A. (2024). The effects of the decentralization of collective bargaining on wages and wage dispersion: Evidence from the Finnish forest and IT industries. British Journal of Industrial Relations, 62, 319-334.

Kauhanen, A. (2025). The effects of the decentralization of collective bargaining on wage differences among employee groups. Applied Economics Letters, 1-4.

Knutsen, T. (2022). The effect of introducing industry-specific minimum wages. PhD thesis, University of Oslo.

Magruder, J. R. (2012). High unemployment yet few small firms: The role of centralized bargaining in South Africa. American Economic Journal: Applied Economics, 4, 138-166.

Martins, P. S. (2021). 30,000 minimum wages: The economic effects of collective bargaining extensions. British Journal of Industrial Relations, 59, 335-369.

Neumark, D. & Wascher, W. L. (2008). Minimum Wages. MIT Press.

Ólafsdóttir, K. (2020). The labor market in Iceland, 2000-2018. IZA World of Labor.

Scheller, E. (2018). Hängavtal–deras utbredning på svensk arbetsmarknad. Stockholm: Ratio. Arbetsmarknadsprogrammet. Rapportserie, del II.

Skedinger, P. (2010). The minimum wage revisited in the enlarged EU. In: Vaughan-Whitehead, D. (ed.) Chapter 12: Sweden: A Minimum Wage Model in Need of Modification. Edward Elgar Publishing.

Skjerpen, T., Kornstad, T., & Rybakla, M. (2016). Lønns-og sysselsettingsmessige virkninger av allmenngjøring innenfor bygge-, verfts-og renholdssektoren. Statistisk sentralbyrå, Notater 2015/2.