Collective bargaining coverage and extension procedures

  • Observatory: EurWORK
  • Topic:
  • Date of Publication: 17 December 2002

Franz Traxler/Martin Behrens

Collective bargaining coverage and mechanisms which extend the provisions of collective agreements beyond the members of the signatory organisations are important factors which strongly affect the procedures and practices through which wages, hours and working conditions are determined, and thus also affect economic development. This EIRO comparative study provides quantitative data on collective bargaining coverage levels and gives an overview of legally based extension mechanisms in 20 European countries - the 15 EU Member States, Hungary, Norway, Poland, Slovakia and Slovenia. The study examines trends in bargaining coverage and changes in the regulation and practice of extension since 1990, as well as exploring the views of the industrial relations actors on these issues.

This comparative study - based on the contributions of the European Industrial Relations Observatory (EIRO) national centres in the EU Member States, four candidate countries and Norway - analyses the coverage of collective bargaining and the existence and practice of procedures which extend the scope of collective agreements beyond the membership of the parties to these agreements.

Coverage and extension are key components of national industrial relations institutions in general and bargaining systems in particular. First, there is almost by definition a close association between coverage and extension. As comparative analysis has found (see National labour relations in internationalised markets. A comparative study of institutions, change and performance, F Traxler, S Blaschke and B Kittel, Oxford University Press, 2001), the extent to which extension mechanisms are used in a country is the most powerful single determinant of variations in the level of bargaining coverage across countries: coverage tends to increase significantly with the use of extension practices. Conversely, the applicability and actual use of extension mechanisms depends decisively on the nature of the bargaining system. For obvious reasons, extension rules can be implemented effectively only in connection with multi-employer collective agreements.

Given the positive impact of extension on bargaining coverage, coverage is significantly lower in countries with predominantly single-employer bargaining, as compared with countries where multi-employer bargaining prevails (see 'Bargaining (de)centralisation, macroeconomic performance and control over the employment relationship', F Traxler, forthcoming in British Journal of Industrial Relations, 2002). In countries where single-employer bargaining prevails, coverage correlates almost exactly with trade union density (see 'Collective bargaining and industrial change: a case of disorganisation? A comparative analysis of 18 OECD countries', F Traxler, European Sociological Review 12, 1996).

Furthermore, bargaining coverage and extension affect other areas of industrial relations. For example, higher levels of coverage are significantly associated with smaller wage differentials (see Wage inequality and varieties of capitalism, D Rueda and J Pontusson, Institute for European Studies Working Paper No. 97.6, Cornell University, 1997, and 'Wage-setting institutions and pay inequality in advanced industrial societies', M Wallerstein, American Journal of Political Science 43, 1999). Extension has a strong impact on the membership strength of employers' associations, as their density ratio (in terms of employees covered) significantly increases with the incidence of extension practices (see 'Collective bargaining in the OECD', F Traxler, European Journal of Industrial Relations 4, 1998, and Traxler et al 2001, cited above). This is because the practice of extending collective agreements to unaffiliated employers creates a very strong incentive for them to join associations. If extension provisions are established and used, employers cannot rule out the possibility of being obliged to become bound by a collective agreement, even when they are not members of the signatory employers' association. Hence, it is rational for them to join the relevant employers' association so as to be entitled to participate in the bargaining process, the outcomes of which may be binding on them in any case.

This EIRO comparative study, covering 20 countries - the 15 EU Member States, four candidate countries (Hungary, Poland, Slovakia and Slovenia) and Norway - presents:

  • quantitative data on collective bargaining coverage, pertaining to the countries' aggregate coverage, as well as coverage disaggregated by certain criteria;
  • an overview of legally established extension mechanisms, their practical relevance for collective bargaining, and mechanisms that work as functional equivalents to such legally established extension procedures;
  • information about the debates on coverage among the industrial relations actors (ie trade unions, employers' associations and governments) and about their views on the advantages and disadvantages of extension mechanisms; and
  • a summary of significant changes in coverage, as well as in legal provisions for and practices of extension, which have occurred since 1990.

Collective bargaining coverage

Data and definitions

The figures on bargaining coverage reported in this study refer to 2000 and 2001, or to the most recent data available. In principle, there are two possible data sources. Data on coverage may be provided by the bargaining parties. Alternatively, data may be available from household or labour force surveys which include questions as to whether the respondent’s job is covered by a collective agreement. However, any kind of systematic data collection is lacking in several countries. If possible, collective bargaining coverage has been estimated by the EIRO national centres in these cases. Reliable estimates could not be made for Finland, Greece, Ireland and Italy (though an estimated coverage rate of around 90% is often cited in the last-named case). As regards the UK, survey data are available only for Great Britain (ie excluding Northern Ireland).

Due to the manifold differences in the national collective bargaining systems across Europe, the raw data available directly from national sources are rarely comparable with one another. To achieve as high a level of comparability as possible, it is essential to proceed from a common operational definition of the rate of collective bargaining coverage.

Accordingly, the rate of collective bargaining coverage is an indicator of the extent to which the terms of employment are regulated by collective agreements. We define the coverage rate as the number of employees covered by a collective agreement as a proportion of the number of employees (ie wage- and salary-earners). Employees may not be covered by bargaining - where such bargaining occurs - for the following reasons. First, the parties to collective bargaining may be too weak to include all employees belonging to their formally demarcated scope of action. Second, the bargaining parties may explicitly exclude certain employee groups (eg part-time workers) from the purview of their collective agreements. In both cases, the employment terms of the employees concerned are regulated only by an individual employment contract. Third, certain categories of employees may be legally excluded from the right to collective bargaining. This may apply in particular to public sector employees or certain groups of them (eg the police and armed forces), whose employment terms are normally regulated unilaterally by the state under these circumstances.

There are two variants of the coverage rate, depending on whether or not one adjusts for the group of employees legally excluded from the right to bargain (see 'Collective bargaining: levels and coverage', F Traxler, OECD Employment Outlook, 1994, and Traxler 1996, cited above):

  • the unadjusted rate is defined as the number of employees covered by a collective agreement as a proportion of the total number of employees (regardless of whether certain groups are excluded from bargaining); and
  • the adjusted rate is defined as the number of employees covered by a collective agreement as a proportion of the number of employees equipped with the right to bargain (ie the total number of employees minus the number of employees excluded from the right to bargain).

The unadjusted rate measures the relative importance of the collective agreement as a regulatory mechanism in a certain country. The adjusted rate indicates the 'governance capacity' of a collective agreement within its own domain. To arrive at valid crossnational comparisons of bargaining coverage, it is important to differentiate between its two variants, since countries differ in the extent to which employees are excluded from the right to bargain.

Multi-level bargaining, as established in several of the countries considered here, creates another problem of comparability. Multi-level bargaining is likely to cause certain groups of employees to be covered by several collective agreements (eg an industry-level agreement and a company agreement) at the same time. Such 'multi-coverage' may in turn lead to 'multi-counting' of one and the same employee in statistics on coverage. This especially applies to statistics based on data provided by the bargaining parties. To avoid multi-counting, we refer in this study to the number of employees covered by any kind of collective agreement.

Another statistical problem may arise from the termination date of collective agreements. Figures may report only collective agreements in their first year of validity, thus excluding multi-year agreements which originate from an earlier year but are still valid. We refer in this study to all collective agreements which are valid during the year under consideration.

Finally, we refer only to the formal coverage of collective agreements, as demarcated by their scope. Hence we do not consider informal 'spillovers' of the terms of the collective agreements to firms not formally covered by them. However, the formal coverage of collective agreements does include those employees covered by extension procedures.

It was not possible for all countries to correct the raw data in line with the above definition for the purposes of this study. In these cases, the deviations from the definition are indicated. The most important deviations concern Spain and Norway, due to inflated figures caused by multi-counting as a consequence of multi-layered bargaining systems. In the case of Norway, the magnitude of error can be specified in that the available survey data are assumed to exaggerate coverage by 5%–10%. Conversely, the figures for Poland somewhat underestimate coverage. This is because a collective agreement, when concluded for an indefinite time, remains in force even after its termination until a new agreement is signed. There are no data on how many employees are covered by such agreements. Furthermore, it was impossible to calculate the adjusted rate for certain countries.

For all these reasons, the data reported here are not strictly comparable with previous data series from other studies covering the period from 1980 to 1996 (see Traxler et al 2001, cited above), which were able to control for inflated data for Spain and Norway and also document figures on unadjusted and adjusted coverage for a larger number of countries from the European and non-European Organisation for Economic Cooperation and Development (OECD) area.

Aggregate coverage

The aggregate (ie country-wide) rates of bargaining coverage for the 16 countries for which data are available are set out in table 1 below. Very rough estimates only are available for Belgium and Sweden, which refer to data from the first half of the 1990s and assume that no significant alteration of coverage has occurred since that time. In the case of Spain, the strong decline in coverage from 2000 to 2001 is a statistical 'artefact', since the figures on the number of employees covered are annual estimates by the bargaining parties which are rather insensitive to short-term fluctuations in employment. Hence, the estimated coverage rate shows a decline in times of expanding employment and and increase in economic recessions.

Table 1. Collective bargaining coverage, 2000-1
Country Coverage rate (%)
Unadjusted Adjusted
2000 2001 2000 2001
Austria 78 78 98 98
Belgium >90 >90 >90 >90
Denmark 83 83 83 83
France 90-95 90-95 90-95 90-95
Germany 67 nd nd nd
Hungary 33 31 36 34
Luxembourg 50 48 60 58
Netherlands nd 88 nd 88
Norway 70-77(a) 70-77(a) 70-77(a) 70-77(a)
Poland nd 1 nd nd
Portugal 87(b) nd 87(b) nd
Slovakia 51 48 51 48
Slovenia 100 100 100 100
Spain 75 68 91 81
Sweden >90 >90 >90 >90
Great Britain 36 36 nd nd

Notes: adjusted = coverage rate adjusted for employee groups excluded from the right to bargain; (a) = 1998; (b) = 1999.

Source: EIRO.

From a comparative perspective, the most notable finding is that 70% of employees or more are covered by collective agreements in the vast majority of countries examined. These high coverage rates (70% or above) apply in 10 of the 16 countries for which data are available. At 100%, coverage reaches its maximum level in Slovenia, as a consequence of the fact that businesses are obliged to be members of all-encompassing 'chambers' of commerce and industry which also act as employers' associations on behalf of their members in collective bargaining (SI0211102F). For the same reason, Austria records a coverage rate of almost 100%. In Great Britain, Hungary and Poland, which have the lowest coverage among the 16 countries, less than 40% of the total number of employees are covered. In Poland, which reports a coverage rate of a mere 1%, the proportion of employees covered is extremely low even given that the reported figure somewhat underestimates the actual level of coverage for the reasons mentioned above. This country thus deviates strongly from the European model of industrial relations which includes collective bargaining as an important mechanism for regulation of the employment relationship. Finally, there is a group of three countries (Germany, Luxembourg and Slovakia) with medium coverage rates (40%-70%).

Leaving aside the special situation of the candidate countries, where economic restructuring has resulted in rather low coverage in all cases but Slovenia, the very low coverage level in Great Britain, Hungary and Poland corroborates the abovementioned causal relationship between the bargaining system and bargaining coverage, in that single-employer bargaining prevails in all these countries.

In little more than half of those countries for which both unadjusted and adjusted coverage are available, these figures do not differ. This indicates that all groups of employees enjoy the right to bargain in these countries. In all the cases reporting differing levels of adjusted and unadjusted coverage, this difference ensues from the fact that certain parts of the public sector (civil servants and the armed forces) are formally excluded from the right to bargain. This difference is largest in Austria, as virtually all public sector employees lack the right to bargain (AT0005221F).

Coverage by employee categories

Table 2 below shows the level of bargaining coverage broken down by gender and employee status for the eight countries where such information is available. Gender-specific differences in coverage are rather small. Where such differences exist, the rate for women is higher than that for men in all countries but Luxembourg. Special research would be needed to explain the differences observed. However, the fact that top executives are sometimes excluded from the scope of collective agreements while women are under-represented in this group probably contributes to the generally higher rate for women.

Table 2. Collective bargaining coverage by employee categories, 2001
Country Coverage rate (unadjusted) (%)
Employee categories
Male Female Blue-collar White-collar
Hungary 28 35 30 33
Luxembourg 52 48 51 49
Netherlands 77(a) 80(a) Not applicable
Norway 69(b) 84(b) 78(b) 76(b)
Slovakia 46 50 51 46
Slovenia 100 100 100 100
Spain 68 68 92 38
Great Britain 35 37 34 36

Notes: (a) = private sector only; (b) = 1998.

Source: EIRO.

Similarly, coverage does not differ very much in terms of employee status. In countries where there are such differences, blue-collar workers record a higher rate than their white-collar counterparts. Again, this may have to do with non-coverage by bargaining of executives who are white-collar employees. In addition, this difference may reflect the fact that blue-collar workers are still the stronghold of unionisation.

Aside from top executives, bargaining systems as a whole do not exclude certain employee groups from their scope. Hence, employees in such 'non-standard' employment as part-time work is generally covered by collective bargaining. However, the reference is normally to the status of 'employee', implying that dependent self-employment is excluded. Exceptions to this rule are Greece, Italy, Sweden and Great Britain, where self-employment is covered to differing extents.

Coverage by sectors

The most fundamental divide in terms of sectors is between the private and public part of the economy - see table 3 below for data on 11 countries where such information is available. With the exception of Hungary, Portugal and Austria, public sector coverage is higher than coverage in the private sector, even though the right to bargain is not allotted to all groups of public sector employees in some countries. This suggests that public employers are more willing to negotiate than private employers are, if the right to bargain is established.

Table 3. Collective bargaining coverage in the private and public sector, 2001
Country Coverage rate (%)
Private sector Public sector
Austria 98 0
Denmark 71 100
Germany West 70(a) nd
Germany East 55(a) nd
Hungary 31 31
Netherlands 86 100
Norway 63(b) 100(b)
Portugal 89(c) 81
Slovakia 34 59
Slovenia 100 100
Sweden >90 100
Great Britain 22 73

Notes: (a) = 2000; (b) = 1998; (c) = 1999.

Source: EIRO.

Table 4 below provides data on coverage by sector of activity for 13 countries. Mining/quarrying, manufacturing, electricity/gas/water, construction and transport/storage/communication tend to show the highest coverage rates. These variations across sectors echo differences in the strength of trade union and employer organisation. In particular, the sectors recording high coverage rates in table 4 tend to be strongly unionised as well, as comparative research on union density has shown (see 'Trends in trade union membership', J Visser, OECD Employment Outlook, 1991, and Trade unions in western Europe since 1945, B Ebbinghaus and J Visser, Macmillan, 2000).

Table 4. Collective bargaining coverage by sector (%), 2000
Country Sector
1 2 3 4 5 6 7 8 9 10 11
Austria -- -- 100 100 -- 100 100 -- -- -- --
Denmark 54 50 76 80 75 57 50 78 -- 100 100
Germany West 51 96 -- -- 84 69 -- 73 -- 95 --
Germany East 33 93 -- -- 51 42 -- 58 -- 98 --
Hungary 19 41 40 97 9 16 23 64 27 10 40
Luxembourg 0 -- 75 100 100 50 0 95 75 100 100
Netherlands(a) 88 -- c.90 -- 98 81 99 90 49 100 96
Norway(b) 56 81 77 98 62 54 70 77 55 100 94
Poland(c) 12 2 2 3 1 0 1 1 2 0 0
Portugal(d) 88 48 94 98 91 86 89 88 69 -- 55
Slovakia 65 98-100 60 80 30 20 10 60 68 60 45
Slovenia 100 100 100 100 100 100 100 100 100 100 100
Spain -- 46 99 33 71 94 96 61 93 20 40
Great Britain (e) 15 34 31 64 24 18 9 48 -- 78 --

Industry classification follows NACE major divisions, as follows:

  1. agriculture, hunting and forestry, and fishing (NACE A B);
  2. mining and quarrying (NACE C);
  3. manufacturing (NACE D);
  4. electricity, gas and water supply (NACE E);
  5. construction (NACE F);
  6. wholesale and retail trade; repair of motor vehicles, motorcycles and personal household goods (NACE G);
  7. hotels and restaurants (NACE H);
  8. transport, storage and communication (NACE I);
  9. financial intermediation, real estate, renting and business activities (NACE J K);
  10. public administration and defence; compulsory social security (NACE L); and
  11. education, health and social work, other community, social and personal service activities (NACE M N O).

Notes: (a) = employees covered via extension are excluded and NACE classifications are not fully used in the Netherlands; (b) = 1998; (c) = 2001; (d) = 1999; (e) no aggregate data are available for divisions 9 and 11, though data are available for the individual sectors J, K, M, N and O.

Source: EIRO.

Coverage by bargaining level

Table 5 below provides figures on coverage by bargaining levels for 11 countries, grouped according to the two constituent levels - ie single- and multi-employer. In the vast majority of countries, multi-employer bargaining prevails in quantitative terms. Great Britain is the only EU Member State for which information is available where single-employer bargaining is prevalent. The reverse applies in the candidate countries, among which multi-employer bargaining predominates only in Slovenia. The EU thus differs significantly from the candidate countries in terms of the prevalent bargaining level (TN0207104F). In line with the association between coverage, extension practices and the bargaining level outlined above, aggregate coverage tends to be rather low in countries characterised by prevalent single-employer bargaining.

Table 5. Collective bargaining coverage by bargaining level, private sector, 2000
Country Bargaining level*
Multi-employer Single-employer
Austria 95 3
Denmark 45 25
Germany West(a) 63 7
Germany East(a) 46 10
Hungary 8 34
Luxembourg 60 40
Netherlands(b) 68 14
Portugal(c) 84 5
Slovakia 27 36
Slovenia 100 0
Spain(a) 66 9
Great Britain(d) 5 26

Notes: * each of the two categories includes the employees under both a single- and multi-employer agreement. In cases of multi-level bargaining, the sum of coverage by bargaining level does not equate to the corresponding aggregate coverage rate set out in in tables 1 and 3; (a) = private and public sector; (b) = 2001; (c) = 1999; (d) = workplaces with fewer than 10 employees excluded.

Source: EIRO.

Coverage by firm size

Data on bargaining coverage by firm size are sparse - table 6 below sets out information for six countries. However, leaving aside the very special case of Slovenia, they show a clear pattern which confirms earlier research (see Traxler 1994, cited above): coverage tends to increase with firm size. There are two main reasons for this. The propensity of large firms to join an employers' association is significantly higher than that of their smaller counterparts (see 'Two logics of collective action in industrial relations?', F Traxler, in Organised industrial relations in Europe: what future?, C Crouch and F Traxler, eds, Avebury, 1995). Likewise, unionisation tends to increase with firm size (see Visser 1991, cited above). This fact matters especially in countries with prevalent single-employer bargaining. Under these circumstances, extension mechanisms cannot mitigate the differences in strength of the unions and employers' associations across firm size, so that these differences directly translate into differences in coverage. Hence, differences in coverage across firm size are most pronounced when single-employer bargaining predominates – as in the cases of Great Britain and Slovakia (see table 6).

Table 6. Collective bargaining coverage by firm size (private sector; 2000)
Firm size (number of employees) Country
Netherlands(a) Slovenia
1-4 64 100
5-9 58 100
10-19 63 100
20-49 74 100
50-99 77 100
100-199 76 100
200-499 83 100
500 and over 89 100
. Norway(b)
1-5 36
6-9 56
10-19 60
20-49 66
50-100 72
100-199 74
200 and over 79
. Slovakia
1-19 0
20-49 5
50-249 40
250-499 84
500-999 95
1,000 and over 97
. Denmark
1-24 57
25-99 75
100 and over 79
. Great Britain(c)
1-24 10
25 and over 31

Notes: (a) = employees covered via extension are excluded; (b) = 1998; (c) = classification by workplaces.

Source: EIRO.

Collective bargaining coverage trends

Since 1990, the prevailing trend in collective bargaining coverage has been stability. As regards western Europe (the EU Member States and Norway), only three countries report considerable changes in coverage. While the coverage rate has increased in Denmark (DK0011104F), it has dwindled in Germany (DE0201299F) and Great Britain.

In Denmark, bargaining has expanded most strongly among white-collar employees, whose coverage rate has increased by 10 percentage points since 1994. Furthermore, coverage has increased by 7 percentage points in the private sector as a whole, with branches such as finance, transport and trade showing above-average growth. Both tendencies indicate that new groups have been integrated into the bargaining system.

One finds the opposite pattern in Germany and Great Britain. From the mid-1990s to 2000, the coverage rate declined from 72% to 63% in the western part of Germany and from 56% to 46% in the eastern part. A clear tendency towards a fall in coverage was also observable in Great Britain over much of the 1990s - the coverage rate fell from 49% in 1990 to 35% in 1998. Since then, the trend has been less pronounced with the rate falling from 36% in 1999 to 35.5% in 2001 (there was a break in the data series between 1998 and 1999). In both countries, collective bargaining has failed to integrate new workplaces, notably those created by the growing number of small and medium-sized enterprises (SMEs). Since new businesses, in particular SMEs, have emerged in all countries across Europe, the decline in coverage in Germany and Great Britain ensues from certain properties of the bargaining systems there. In Germany, a weakening of trade unions and employers' associations as well as a decrease in the use of extension practices are reported as the main factors in falling coverage. In Great Britain, the decay of multi-employer bargaining and the unions' failure to secure recognition for collective bargaining from employers in new workplaces has had a similar effect on coverage.

The context for bargaining trends in the candidate countries strongly differs from western Europe, as they have undergone fundamental economic and institutional transformations since the early 1990s. Under the old regime, bargaining coverage was 100%, although collective agreements often had little practical relevance and differed from the OECD countries in their focus on social welfare and health and safety at work (see Industrial relations between command and market. A comparative analysis of Eastern Europe and China, G Schienstock, P Thompson and F Traxler, eds, Nova Science Publishers, 1997). Hence, all countries but Slovenia saw a decline in bargaining coverage in the course of the transformation process. The companies in the emerging private sector have been neither keen on entering single-employer bargaining nor willing and able to set up strong employers' associations which might conduct multi-employer bargaining on their behalf (TN0207102F). The stability of coverage in Slovenia at an unrivalled level of 100% is mainly based on the obligatory membership of chambers of commerce and industry, which has survived the transformation process.

Debates on bargaining coverage

As far as debates on coverage are concerned, one can differentiate between issues directly addressing coverage and those implicitly affecting it. Overall, bargaining coverage is not a very controversial issue, since debates which directly refer to this issue are reported only from Spain, Poland, the Netherlands and Portugal:

  • in Spain, debate centres on the principle of 'ultra-activity' (ultraactividad) which means that, according to law, a collective agreement remains valid after its expiry, if it has not been renewed (ES0107150N). Both the government and employers want to abolish this provision for the sake of enhanced flexibility, something which meets strong trade union opposition;
  • in Poland, recent amendments to the Labour Code have introduced a far-reaching statutory 'hardship clause' (PL0209107F). Accordingly, the signatory parties can agree to suspend a collective agreement for up to three years, if a company faces financial problems;
  • in the Netherlands, liberal political parties in parliament have questioned the representativeness of trade unions in collective bargaining with regard to non-unionised employees, given that the bargaining coverage rate (at over 80%) far exceeds the rate of unionisation (around 25%) (NL0211104F); and
  • in Portugal the very strong regulatory role of the state in bargaining, in particular its pervasive extension practices, has been a matter of public debate (PT0011122F).

Issues indirectly affecting coverage concern the level of bargaining and the role of employers' associations and unions in the bargaining process. As outlined above, coverage tends to decrease with declining firm size. Hence, decentralisation to the company level threatens to undermine collective bargaining, as comparative analysis shows (see Traxler 2002, cited above). Examples of debates on these matters include the following:

  • in Italy, the General Confederation of Italian Workers (Confederazione Generali Italian del Lavoro, Cgil) - one of the three main trade union confederations - opposes further decentralisation of bargaining because of the low coverage rate of 'second-level' bargaining at company level. Cgil is also critical of a shift in bargaining from the sectoral to the territorial level, as favoured by Italian Confederation of Workers’ Unions (Confederazione Italiana Sindacati Lavoratori, Cisl) (IT9907250F);
  • in Hungary, there have been efforts to recentralise bargaining. In connection with attempts to establish a 'European-style' system of multi-employer bargaining, sectoral social dialogue bodies have been institutionalised, which may stimulate sectoral bargaining;
  • the Confederation of German Employers’ Associations (Bundesvereinigung der Deutschen Arbeitgeberverbände, BDA) blames the decline in bargaining coverage on the militancy of trade unions, claiming that their strikes and 'excessive' wage demands induce employers to leave their associations; and
  • in the UK, the trade unions' abovementioned problems in maintaining their status as a bargaining party led to a new statutory procedure for trade union recognition (UK0201171F) in 2000, which was the most controversial element of the Labour government's Employment Relations Act 1999 (UK0007183F).

Extension provisions

While in the previous section collective bargaining coverage was predominantly seen as a function of trade unions' and employers' ability to conclude collective agreements, in this section we focus on state-influenced provisions which may extend collective agreements to employers and employees not directly covered by these agreements. When referring to extension provisions, we do not include those provisions which affect the wages, hours, and working conditions of those groups of employees who are excluded from the right to collective bargaining. For the sake of analytical clarity, we also do not touch on cases where agreements or provisions thereof are extended by 'soft factors' such as informal agreement, habit, custom or other voluntary practices. Instead, we focus on three major types of extension, usually of multi-employer (typically sectoral) agreements, which are all based on some kind of legal act:

  • extension in the narrow sense (ie based on the 'erga omnes ' principle), which makes a collective agreement generally binding within its field of application by explicitly binding all those employees and employers which are not members of the parties to the agreement;
  • enlargement, which provides for a collective agreement concluded elsewhere to apply in sectors or areas where no union and/or employers' association capable of collective bargaining exists. Through enlargement, a certain collective agreement is made binding for a specific geographical or sectoral area outside this agreement's actual scope; and
  • functional equivalents, such as compulsory membership of the bargaining parties' organisations or legal provisions requiring government contractors to comply with the terms of a relevant collective agreement. Such functional equivalents are not based on formal extension mechanisms but do in effect extend the provisions of an agreement to a larger constituency.

As shown in table 7 below, extension provisions are quite common in the 20 countries considered. With the exception of Norway, Sweden and the UK, all countries have at least one extension mechanism such as extension in the 'narrow' sense, enlargement of collective agreements or functional equivalents. Various kinds of extension in the narrow sense apply in 15 of the 20 countries, while enlargement procedures apply in Austria, Portugal and Spain only. Functional equivalents can be found in six countries, though in three cases such provisions merely support other existing procedures. In Italy and Slovenia, extension mechanisms are exclusively based on functional equivalents and in Slovenia in particular, with companies' obligatory membership of chambers of commerce and industry, the impact on collective bargaining coverage is quite powerful.

Table 7. Varieties of extension provisions
Country Extension in the narrow sense Enlargement Functional equivalents
Belgium - -
Denmark - -
Finland -
France - -
Germany -
Greece - -
Hungary - -
Ireland -
Italy - -
Luxembourg - -
Netherlands - -
Norway - - -
Poland - -
Portugal -
Slovakia - -
Slovenia - -
Spain -
Sweden - - -
UK - - -

Notes: indicates existence of extension mechanism.

Source: EIRO

Types of 'erga omnes' extension

Provisions for the extension in the narrow sense (also known as 'erga omnes ' extension) represent the overwhelming majority of extension practices within the 20 countries under review. While concrete rules, requirements and procedures for extension differ widely, there are some common elements and practices which can be found in several countries. In four cases - see table 8 below - erga omnes extension applies quasi-automatically 'ex lege' (ie merely through the existence of a relevant law), but most commonly the initiative must come from at least one party to the collective agreement or one social partner. Public authorities, such as ministries of labour, play a crucial role in initiating the extension of an agreement in France, Greece, Portugal and, to some extent, also in Slovakia.

In the overwhelming majority of countries, the full range of provisions covered by a collective agreement is regularly extended while different provisions apply in Ireland, Belgium and Denmark. In Denmark, for example, extension provisions are used only to transpose the contents of EU Directives (DK0112158F), while in Ireland only provisions on minimum wages and working conditions may be extended. In Belgium only a limited range of issues is subject to extension. While limitations in these countries are part of the legal framework which governs extension mechanisms, other countries provide some more flexibility. In Hungary and Germany the law allows for the extension of a selected range of collectively agreed provisions, and particularly in Germany this leads to a rather piecemeal extension of selected provisions.

In the overwhelming majority of cases, extension is formally applied by some form of public act, usually a decision, decree or order issued by the government authority in charge of labour matters. Several countries have established minimum requirements for extension, most commonly minimum rates for coverage of the relevant agreement prior to extension. Such requirements for representativeness apply in Finland, Germany, Greece, Hungary, Ireland, the Netherlands and Spain. Other minimum requirements included that extension must be in the 'public interest' (Germany, Poland) or abolish disadvantages to employees or employers (Slovakia - SK0210102F). It should be noted, however, that although agreements are extended erga omnes, covering all employees and employers within their field of application, there are some differences - mostly depending on the different scope of collective bargaining in the particular country.

Table 8. Procedures for erga omnes extension of collective agreements
Country Initiative Extended to Minimum requirements for extension Special extension procedures
Austria Automatic (ex lege). Agreement automatically applies to all employees of employers affiliated to signatory organisations. - -
Belgium On request of signatory party but virtually automatic (ex lege). All employees and employers in a given sector. - Royal Decree issued by the competent Minister.
Denmark Request of social partners. All employees and employers in a given sector. - Law to be enacted by the government.
Finland Automatic (ex lege). All employees and employers in a given sector. National coverage of the agreement and 'representativeness' in a given sector. Special committee set up by Council of State needs to confirm representativeness of the agreement. Extension decision published in a public register.
France Almost automatic (ex lege) but formal request of Ministry of Labour or social partners required. All employees and employers in a given sector. . Executive order by the Minister of Labour after consultation with National Commission on Collective Bargaining (Commission nationale de la négociation collective).
Germany Request of at least one party to the collective agreement. All employees and employers in a given sector. Minimum of 50% of all employees in field of application must be covered prior to extension, extension must be in public interest, approval by a special collective bargaining committee (Tarifausschuss) required. Declaration by the Ministry of Labour and Social Affairs.
Greece Mostly on initiative of the Ministry of Labour but also on request of social partners. All employees and employers in a given sector. Employers already covered by agreement must employ at least 51% of employees in sector or occupation. Declaration by the Ministry of Labour after receiving opinion from tripartite Supreme Labour Council.
Hungary Application by contracting parties. All employees and employers in a given sector. Applicants need to present proof of their majority status. Decision by Minister of Employment and Labour after consultations with the Wage and Labour Committee of the tripartite National Interest Reconciliation Council ((Országos Érdekegyeztetó Tanács, OÉT).
Ireland Application by the parties to an agreement. All employees and employers in a given sector/region or company. Parties to the agreement must be 'substantially representative'. Decision of Labour Court to grant a Registered Employment Agreement (REA).
Luxembourg Automatic (ex lege). All employees and employers in a given sector. - Declaration issued by Grand-Ducal Regulation following a unanimous proposal of the parties represented on the Joint Conciliation Commission on advice of competent professional chambers.
Netherlands Request of one or more bargaining parties. All employees and employers in a given sector. Agreement must cover 'sufficient majority of relevant employees'. Decision by Ministry of Social Affairs and Employment.
Poland Motion to establish a multi-employer agreement filed by a bargaining party. All employees within a given multi-employer jurisdiction. Competing union must not have become involved in negotiations within 30 days after the motion was filed. Motion must be publicly announced.
Minister of Labour. All employees of employers which conducts the same or comparable activities. Extension must satisfy a 'vital social interest'. Decision by the Ministry of Labour.
Portugal Ministry of Labour. All employees and employers in a given sector - Extension order (Portaria de extensão) by the Minister of Labour. Employers, their associations, and unions have opportunity to object to the decision (but not to stop it).
Slovakia Request by social partners or on initiative of the Ministry of Labour, Social Affairs and Family. All employees of those employers not covered by the agreement which perform similar business activities and are subject to similar social conditions as the employers directly covered by the agreement. Extension should abolish disadvantageous situation of employees/employers that are not covered. Decree by the Ministry of Labour, Social Affairs and Family following positive decision by the joint Committee for the Extension of Sector Collective Agreements' Legal Obligation to Other Employers.
Spain Automatic (ex lege). All employees and employers in a given sector, subsector or geographical area. Representatives of 50% of relevant workers and employers must have signed the agreement. -

Source: EIRO.

To sum up, all these different provisions lead to significant differences in the use and range of erga omnes extension. In some 'low-threshold' cases, most notably Austria and Luxembourg, extension is granted almost automatically.

Enlargement procedures

Enlargement procedures, whereby a collective agreement is extended to a specific geographical or sectoral area outside its actual scope, are much less common in the 20 countries examined and exist only in Austria, Portugal and Spain:

  • in Austria, enlargement takes the form of an 'extension order', issued by the Federal Arbitration Board (Bundeseinigungsamt). The board acts upon written requests from employers' or employees' organisations. Collective agreements to be extended in this way must be of 'prevailing importance', the working conditions in the 'adopting' sector must be similar to those in the sector where the collective agreement originates and there must not be a competing agreement;
  • in Portugal, enlargement procedures apply in cases where there are no bargaining partners in a given sector or where the parties do not show any initiative to negotiate a collective agreement. Enlargement is granted in the form of an extension directive (Portaria de extensão) by the Minister of Labour. Although employers, their associations, and trade unions have the opportunity to object to this decision, they do not have the power to stop enlargement of an agreement; and
  • in Spain, enlargement is granted only where workers and employers are prejudiced by their inability to enter into a collective agreement in their field, due to the absence of legitimate bargaining parties. In addition, the law requires a positive decision by the joint National Consultative Committee on Collective Agreements (Comisión Consultiva Nacional de Convenios Colectivos) which also needs to be approved by the Ministry of Labour. Applications for enlargement are to be filed by one or both parties to a collective agreement.

The impact of extension on bargaining coverage

While, in general, extension mechanisms are a powerful variable which explains a great deal of variation in the level of collective bargaining coverage across countries, the intensity of this causal relationship depends to a large degree on the specific mechanisms and tools through which extension is available. As previous research has found, the existence of extension also benefits bargaining coverage by way of lowering the relative costs of becoming and remaining a member of an employers' association. It should be noted, however, that exceptions to this rationale apply. Two out of three countries which do not have any extension procedures (Norway and Sweden) have a fairly high collective bargaining density. With detailed data on the share of the workforce subject to the extension of collective agreements available for a small number of countries only - see table 9 below, which provides information on eight countries - the quantitative impact of extension provisions on bargaining coverage cannot be measured in this study.

Table 9. Employees covered by extension of collective agreements
Country Estimated % of employees covered by extension 2000 (2001) Other indicators of estimated impact of extension
Denmark 15%-17% (15%-17%)(a) nd
Germany nd Approx. 2.5% of all agreements are extended (2002)(b).
Hungary 0.34% (1.51%)(b) nd
Ireland nd An estimated 242,000 employees covered by REAs (see table 8 above) and JLCs (see table 10 below) (1998)(a). REA cases are rare, with five at the sector and one at the company level.
Netherlands 7% (7%) nd
Portugal nd 174 extension directives (1996), 43 extension directives (in first half of 2002).
Slovakia 8% (9%) nd
Spain 0%-1% (0%-1 %) nd

Notes: (a) = only a limited number of provisions are extended; (b) many agreements which are subject to extension do not include provisions on wages or wage levels.

Source: EIRO.

Functional equivalents

As shown in table 10 below, there are functional equivalents to extension provisions in six of the 20 countries analysed in the study. The contents and mechanisms vary widely and it is not always possible to estimate the specific impact of these equivalents on collective bargaining coverage. What is striking, however, is that provisions such as mandatory membership of employers' associations, as applies in Austria and Slovenia, seem to have a strong and positive impact on bargaining coverage. Thus, it is hardly surprising that in both countries coverage is almost complete. In the other cases, the direct impact on coverage is more opaque.

Table 10. Functional equivalents to extension of collective agreements
Country Mechanism
Austria Mandatory membership of the Chamber of the Economy (Wirtschaftskammer, WKÖ) for companies provides for almost complete collective bargaining coverage.
Finland Public officials are required to comply with the provisions laid down in the relevant collective agreements.
Germany Some federal states have passed laws on public procurement which require contractors in the construction and transport sectors to comply with the provisions defined by the relevant collective agreement (DE0201202F). Attempts have been made - so far unsuccessfully - to legislate nationally on this issue (DE0211205F).
Ireland Joint Labour Committees (JLCs) may be established by the Labour Court to define minimum rates of pay and employment conditions in particular sectors by Employment Regulation Order. JLCs are made up of representatives of employers and workers, and chaired by a representative of the Ministry for Enterprise, Trade and Employment.
Italy Courts use collectively agreed minimum wage levels as a point of reference when assessing whether wages and salaries conform with the constitutional requirement for fair pay. Furthermore, collectively agreed wage levels are taken into account in awarding public contracts, in order to identify 'abnormally low' bids.
Slovenia Almost total collective bargaining coverage results from obligatory membership for companies of the Chamber of Commerce and Industry of Slovenia (Gospodarska Zbornica Slovenije, GZE) or the Chamber of Crafts of Slovenia (Obrtna Zbornica Slovenije, OZE). Furthermore, the current law on labour relations (due to be replaced in January 2003) requires that certain issues must be covered in collective agreements.

Source: EIRO

Changes in the legal framework

Since the beginning of the 1990s, changes in the legal framework have occurred in many countries but by and large have provided only for some minor changes in procedures. In seven out of the 17 countries which have at least one of the three extension mechanisms ('erga omnes' extension, enlargement, or functional equivalents), the regulations on extension have not changed at all over this period. In another two cases (Denmark and Greece) the procedures were newly established during the 1990s, while changes in the regulations on extension have occurred in another eight countries:

  • in Finland, a 2001 reform of the Employment Contracts Act provided for a more precise definition of 'representativeness' and thus ended the rather inflexible rule which, in order for the extension of a collective agreement to occur, required that at least 50% of the employees in a given sector worked for an employer which was bound to the agreement (FI0107193F);
  • representativeness has also been an issue in the Netherlands, where new policy rules for the Ministry of Employment and Social Affairs, taking effect in 2003, will clarify procedures to determine the representativeness of parties to an agreement (NL0211104F);
  • in Germany, a 1998 reform empowered the Minister of Labour and Social Affairs to extend minimum wage agreements by decree, and thus made it possible to bypass any potential veto by organised employers within the relevant bipartite collective bargaining committee (DE9909117F);
  • in Ireland the introduction of a national minimum wage in 2000 (IE0107170F) made JLCs, which set minimum pay and conditions for certain sectors (see table 10 above), almost superfluous;
  • in Portugal (1992) and Slovakia (1996), amendments have set new time limits for the extension process, mostly in order to speed it up;
  • in Slovenia, a forthcoming reform of the Law on Collective Agreements will introduce a new extension provision, on top of existing functional equivalents; and
  • in Spain, a 1999 amendment to the Workers' Statute introduced clearer requirements for the process of enlargement. However, because the number of cases of enlargement has always been very low - only 92 cases were approved between 1986 and 1995 - even this change is unlikely to have a major impact.

To sum up, there is - perhaps surprisingly - a high level of stability in extension provisions and, despite occasional criticism and demands for deregulation (which will be discussed in more detail in the following section), the continuity is striking. One case, however, where more fundamental changes occurred stands out. In 1980, the then Conservative UK government repealed Schedule 11 of the 1975 Employment Protection Act and thus ended the British version of extension.

Debates on extension provisions

In the three countries without any extension provisions (Norway, Sweden and the UK), debates on this issue have arisen only in the UK, where trade unions have demanded an increase in the level of protection of the pay and conditions of public sector employees who are affected by the outsourcing of activities to the private sector (UK0111107F). These demands have not so far been met. In Italy, where a form of functional equivalent is in force, there were plans in the late 1990s to introduce a new form of erga omnes extension (IT9804226F). The draft bill (on which the social partners disagreed), never came to a final vote because differences between the political parties.

In six out of the remaining 16 countries (with no information available on Poland and Hungary) extension mechanisms are either widely accepted ( as in Austria, Belgium, Luxembourg, Portugal and Slovakia) or not a subject for public discussion, which suggest that neither the parties to collective agreements nor the public authorities feel any pressure to question or challenge extension practices (as reported from Greece).

In three cases, there are debates over adjusting extension mechanisms. The focus of these debates, however, is not on weakening the impact of extension on collective bargaining coverage but rather on addressing some formal deficits within the procedures or some unintended side-effects. In Ireland, for example, JLCs are considered to be too complex or ineffective and in Spain some initiatives are being taken to find joint rules for extension in small and dispersed workplaces. Some problems arise in the context of competing unions. In France (similar problems are reported from Portugal), where extension is usually supported by all parties, trade unions in particular are concerned about cases where agreements signed by 'minority unions' have been extended and granted the status of 'industry rules' (FR0108163F).

More mixed results can be observed in the Netherlands, Slovenia and Denmark. In Denmark, some uneasiness with extension is linked to the fact that it was introduced very recently as a consequence of a threat by the European Commission to take the government to the European Court of Justice (ECJ) over inadequate transposition of EU Directives. Despite some earlier opposition from the Dutch government, extension is still favoured by the social partners, though some have taken place on how to determine representativeness (NL0103127F). In Slovenia, the social partners have been involved in preparing a new law which is due to come into force in January 2003. The new law will include a new extension procedure which is, however, not to replace compulsory employers' membership of chambers, which (as seen above) already provides for 100% collective bargaining coverage.

Finally, there are two cases where extension mechanisms face some severe opposition, mostly from organised employers. In Finland, the Federation of Finnish Enterprises (Suomen Yrittäjät, SY) - which represents mainly SMEs - as well as the conservative National Coalition Party (Kansallinen Kokoomus) - one of the governing coalition parties - want to restrict the extension of collective agreements (FI9905105N). While some German sector-level employers' associations, most notably in the construction industry, still support extension, the BDA employers' confederation increasingly opposes it and uses its representation in collective bargaining committees to vote down an increasing number of applications for extension.


For the overwhelming majority of the countries surveyed in this comparative study, an analysis of recent trends in collective bargaining coverage reveals a high degree of stability. With information available for 16 of the 20 countries examined, it is striking that within the EU Member States and Norway bargaining coverage remains at a rather high level. Classifying the 16 countries for which data are available on a scale of high, medium and low coverage: 10 countries can be considered to be high density (over 70% collective bargaining coverage); three medium density (between 40% and 70%); and three low density (less than 40%).

High bargaining coverage is thus a common phenomenon, while there are also strong indications that only little significant change has occurred in coverage rates over the past decade or so in the current EU and Norway. In two countries (Germany and the UK) bargaining coverage has declined, while in one (Denmark) coverage has increased. In the run-up to EU enlargement, the Member States' systems of collective bargaining seem to be well established, and maintain to a large degree a great potential to regulate wages, hours and working conditions.

Similar developments can be observed in the case of provisions for the extension of collective agreements, which apply in the great majority of countries examined. While over the past decade several, mostly minor adjustments of the legal mechanisms for extension have occurred, the basic parameters and functions of extension provisions have remained unchanged. The notable earlier exception of the UK, where extension provisions were abolished in 1980, does not challenge this general observation. The impact of extension provisions on collective bargaining coverage varies between countries and cannot always be exactly specified. From the evidence available, however, it is fair to conclude that there is no substantial impact of extension provisions on either upward or downward change in collective bargaining coverage rates. Rather, extension provisions might have contributed to stabilising coverage.

This continuity is somewhat mirrored in the debate on both collective bargaining coverage and extension provisions. Some debates on coverage have arisen in Poland, Spain, and Portugal and issues related to extension have been subject to rather controversial debates in Finland and Germany. Overall, however, it can be said that despite a few demands for adjustment there are no proposals for a major change on the table.

When we include the four candidate countries in the analysis, the general picture of continuity somewhat changes. While in the area of extension provisions, developments in Hungary, Poland, Slovakia and Slovenia go along the same lines as the majority of western European countries in maintaining a stable legal framework, the situation is different in the case of bargaining coverage. With the exception of Slovenia, collective bargaining is in decline, reaching only medium coverage rates at best. As in the case of Hungary, where there are attempts to use special EU programmes to stabilise the collective bargaining system, it will be difficult to reverse the downward trend and to keep coverage rates from declining even further. This would lead us to assume that the forthcoming enlargement will bring increasing diversity of bargaining coverage rates within the larger EU and will contribute to strengthening the 'low coverage camp'. Given that coverage rates in the candidate countries are still not settled at a stable level, this will also reduce the general level of stability observed in this study. (Franz Traxler, University of Vienna and Martin Behrens, Institute for Economic and Social Research, WSI)

Useful? Interesting? Tell us what you think. Hide comments

Add new comment