Technical notes

Industrial relations indicators

Tables 1.1., 1.2. & 1.3. Trade union membership data and trade union density rates1

Sources

Various methods have been used to collect the data. An ILO questionnaire on union membership, union density and collective bargaining trends was sent to governments, employers and union representatives. Trade union profiles obtained through the ILO trade union network have been used as an additional source. For many countries in Africa, Latin America, and Central and Eastern Europe the data on trade union organization and employment were provided by the ILO Regional Offices. Furthermore, use has been made of data from special reports and research notes on Argentina, Brazil, India, Malaysia, Peru, Poland, the Russian Federation, South Africa and Uganda, prepared specifically for the Task Force. In total, the above ILO sources gave union membership data covering 64 countries. Absolute figures on union membership and employment are used in all cases in order to ensure that the density rates are calculated in the same way (see below).

A database established by J. Visser and published in OECD Employment Outlook (Paris, 1991) provided data covering the 1980s for 24 member States of the Organization of Economic Co-operation and Development. For the purpose of this report, the material has been updated to the mid-1990s, employing the same sources and methods as in the original report. An additional ten countries have been added on the basis of national sources, statistical yearbooks, government reports and published research, including case studies on countries in Africa, Asia, Latin America and Central and Eastern Europe.

Data were available for 1994 or in respect of 1995 nearly all countries. For more than 70 countries a comparison with an earlier period (mainly the mid-1980s) is made.

In some cases, difficulties have been encountered with regard to the quality and comparability of the data. However, comparability over time is not a major cause for concern, except in a few cases (i.e. transitional economies). Comparability across countries, especially between countries in different regions, is more problematic.

Unions and union membership

In order to calculate union membership, two preliminary questions are addressed: What is a trade union? and: Who do we count as a union member? The Australian Bureau of Statistics suggests: "an organization, consisting predominantly of employees, the principle activities of which include the negotiation of pay and conditions of employment for its members" 2 as a definition of a trade union. The definition includes workers' organizations and professional associations, even when collective bargaining is not their main activity. This might be explained either by the fact that some groups have other means of promoting and defending the employment and income interests of their members, or because employers do not negotiate with them. The first condition most often applies to professional groups and it is not always clear whether a particular professional association should have been included under the above definition of a union. The widest coverage, in this particular sense, is found in Europe. In Northern Europe, professional groups such as artists, doctors, architects, lawyers or even football players form their own employee associations. These associations may, or may not, have negotiated collectively on behalf of their members. They do, however, express the employment and income interests of those members. Such organizations are included in our statistics in accordance with national practice. However, self-employed members are excluded.

Recognition as a collective bargaining agent by employers and governments is a common problem for many unions. In Western and Northern Europe bargaining coverage is usually much wider than union coverage, whereas in most other countries bargaining coverage is lower than union coverage (when union coverage is calculated as a proportion of all wage and salary earners in employment). Organizations which are not recognized by the employer are not excluded from our statistics on union representation. Only if a workers' organization has no intention of representing the employment interests of its members and is not even minimally involved in the expression of workers' grievances is its inclusion in the statistics seriously questioned.

An advantage of this rather comprehensive definition of a trade union is that it recognizes that a union's central purpose is the representation of the employment and wage interests of workers, but that it is neutral with regard to the methods applied or the organizational form adopted by a union. Methods and forms of collective action vary widely across countries and over time. General, industrial, craft, occupational, enterprise unions or professional employee associations whose activities involve collective bargaining, pressure on legislators or public authorities, strikes, working-to-rule, petitioning, demonstrations or legal action are all included in the data base. Those employee associations "who are dependent on employers, who reject collective in favour of individual representation, do not seek a role in negotiations, or consist mainly of self-employed persons" 3 are excluded. Legally or voluntarily established works councils; 'in house' and 'company unions' which depend on the employer for their survival or finance; professional societies which mainly further the educational interests of their members; and unions of tenant farmers, shop keepers or mainly self-employed persons are not included.

The application of these criteria requires careful consideration in each case. Where information is available on all trade unions in a particular country, for instance through a survey or register of associations, it is possible to decide which organization is or is not a trade union for the purposes of the above definition. If the data are based on reports or estimates, it is assumed that all affiliates satisfy the criteria of a trade union.

Compiling membership statistics

There are two main methods of compiling membership statistics. The first is to carry out a household, enterprise or labour force survey. This method has clear advantages for calculating detailed union density rates by sex, employment status, industrial branch, enterprise size, educational attainment, level of earnings or other characteristics. The survey method gives better results when it is clear to respondents what is meant by a union and membership, and if problems of statistical sampling are solved. Data based on household, population or labour force surveys are now available on a fairly continuous, annual or biennial basis for the United States (from 1973 onwards), Canada (from 1984), Australia (from 1976), South Africa (from 1985), Great Britain (from 1989), Sweden (from 1987) and the Netherlands (from 1992). In Japan (from 1980), an enterprise survey exists; and in Great Britain (from 1980), a survey of establishments with more than 25 employees is available. This is our only source for the United States; for the other countries there is additional information.

The second method is a compilation of membership statistics from questionnaires completed by individual unions or trade union federations. In many countries this task is undertaken by an official registrar, a government office, a central statistical bureau, or one or more trade union federations. In some cases such data are compiled by independent researchers, either on the basis of unpublished registers, government surveys or even their own surveys. This kind of data offers advantages for a study of membership developments in relation to union type, membership concentration, inter-union competition, union politics and union ideology.

One of the main difficulties with the second method is statistical coverage or the identification of unions (i.e. not so much whether an organization should be considered as a 'trade or labour union, but simply locating and identifying the existence of small, new unions). This constitutes a problem in the case of unaffiliated unions as well as in countries where there is no obligation for a union to register. Official registration is not always useful either because it is sometimes used, with denial of recognition, as a means to hinder the formation of newly emerging unions (examples are found in Indonesia, South Korea, and Thailand as well as in Latin America before democratization: see Brazil) or because unions fail to return their membership or financial files to the registrar (e.g. India).

Under-reporting of non-affiliated and non-registered trade unions constitutes a problem in many countries but in the absence of surveys, elections, and independent press reports it is difficult to estimate its size. 4 It relates less to making comparisons within countries over time than to those drawn between countries. In general, statistical coverage has increased -- especially in western countries.

Many figures are based on self-reporting, i.e. membership claims made by the unions. These claims are checked with those of other unions, reports by researchers and experts who have greater insight into the industrial relations and union affairs in a particular country. Nevertheless, self-reporting of membership remains a starting point for many of the estimates.

Self-reporting of membership reflects different administrative and political practices and may yield incomparable and unreliable results. Unions may have reasons to overstate or understate their membership figures in reports to the press, public agencies, political parties, employers or competitors. They may apply different norms regarding who is to be considered as a 'member in good standing and may be slow to remove those who have left or no longer pay their contributions. Unions may include people who no longer consider themselves as members. Comparison with survey data suggests that some overstatement in reported membership is general but, in most cases, small.

Union density and labour force statistics: three denominators

As a measurement of relative rather than absolute size, union density rates are better suited to making comparisons, especially across countries, than absolute membership figures. Union density expresses union membership as a proportion of the eligible work force. Ideally, groups who are not legally permitted to join a union should be excluded from the calculation of union density statistics. In many countries, senior civil servants, the armed forces, police officers, security staff, teachers or domestic servants are not permitted to join a union. However, as Chang and Sorrentino (1991) 5 rightly observe, the eligibility to join a union shifts over time and across countries, and the strict application of such a criterion for calculating union density rates would make comparison across countries extremely difficult if not impossible.

Therefore, a common denominator is applied to permit comparison. Given the definition of a trade union, and the self-declared purpose and domain of most unions, this common denominator is defined as all people who earn their living on wages or salaries, including those who are employed in the public sector or work in government service. Normally, union density rates are standardized by calculating union membership as a proportion of the wage and salary earners in the same year (preferably on the basis of some annual average, or end-of-year data). Such data is directly comparable with household or labour force survey data. This method was used for calculating standardized annual union density rates between 1970 and 1989 for the OECD countries and is also one of the baselines in the ILO report. Although many unions, at least in Europe, retain membership of unemployed workers and those that have retired from the labour force (i.e. through reduced contribution rates or by offering special benefits), the calculation of standardized density rates requires their number to be subtracted from the 'active membership. This is possible, mostly with the help of surveys, financial data or on the basis of estimates from some unions in most OECD countries and in a small number of other countries. Where possible, the union membership data in this report exclude self-employed, unemployed and retired workers. Their number varies from 46% in Italy, 36% in Belgium, 19% in Sweden, 18% in the Netherlands, Austria or the Czech Republic to 6% in France and 5% in Swaziland. Generally, the share of retired and unemployed workers is higher if the union is 'older, if the union is organized in a declining sector (for example, manufacturing) and if the union is involved in the provision of benefits for such members. These conditions are more likely to apply in Europe (West and East) than, for instance, in Southeast Asia or Africa. Therefore, standardized or net rates have been calculated for European countries.

The major difficulty inherent in calculating union density rates on the basis of wage and salary earners in employment arises from the fact that, in many developing countries, the necessary employment data are missing. In many of these countries the line between employment, underemployment and unemployment, or between self-employment and employment for wages, is not easily drawn. In order to broaden the orbits of comparison, two more baselines for the calculation of union density rates are presented. One is connected to the concept of the non-agricultural labour force, the other to the formal wage sector.

The size of the non-agricultural labour force is estimated on the basis of ILO, OECD and World Bank data. The non-agricultural labour force is often regarded as the unions' main domain. As a rule, the number of union members in agriculture should be subtracted before union density rates are calculated. However, as it is rare for unions to organize any significant number of farm workers, this calculation has not been performed. Unionized farm workers are found only where there is a plantation system (i.e. in Malaysia) 6 or where unions have established a tradition of providing social security benefits. In the few countries where union density in agriculture is very high, for instance in the Scandinavian countries, the agricultural wage sector becomes so small (less than 5% of overall employment) that their number in the aggregate is ignored, without comparability being impaired.

The main advantage of using the non-agricultural labour force as the denominator for calculating union density rates is that data is available for almost all countries. The disadvantage is that many groups (such as the self-employed and unpaid family workers, especially in construction and in commercial, personal and household services) which do not belong to the target population of trade unions are also part of the non-agricultural labour force. The number of wage-earners is generally much smaller than the whole non-agricultural labour force -- also because the latter includes the unemployed in industry and services.

Finally, in some countries -- in particular Africa, South Asia and parts of Latin America -- employment data are only available for the formal wage sector. Thus, this sector constitutes the third common denominator the calculation of union density rates. Following the ILO (PREALC) approach in Latin America and the 1972 ILO Kenya Employment Mission Report, the informal sector is estimated by an addition of numbers from such occupational and status categories as employers, the self-employed, unpaid family workers, farm workers and domestic workers; with an estimation of the number of people earning less than the minimum wage; of the size of employment in small firms, usually with less than ten workers; and of the number of workers employed on irregular contracts or working only a few hours a week. The estimates take account of the fact that many of these categories overlap.

Table 2. Profile of national employers organizations

This table provides an overview of national employers organizations which are members of the International Organization of Employers.

Table 3.1. Collective bargaining structure in selected countries

This table includes three indicators of collective bargaining structures: level of bargaining, dominant level, and trend in bargaining levels. National level and sectoral level bargaining are combined in the table to emphasize the importance of collective bargaining outside the enterprise.

Table 3.2. Collective bargaining coverage rates in selected countries

The collective bargaining coverage rate shows the proportion of employees covered by collective agreements. The sources of the data differ and therefore caution is necessary when comparing data across countries. Collective bargaining coverage rates only concern the formal sector.

Tables 4.1., 4.2. & 4.3. Number of strikes and lockouts, workers involved and workdays not worked

Generally speaking, strikes and lockouts are temporary work stoppages wilfully effected by one or more groups of workers and temporary closures of one or more places of employment by one or more employers with a view to enforcing or resisting a demand or expressing a grievance. However, national definitions of strikes and lockouts may differ from country to country depending on the source of the statistics. The data presented in these tables may come from several sources, including strike notices, newspaper reports and direct inquiries addressed to employers or to workers organizations. The data cover strikes and lockouts together, as most countries do not distinguish between those two type of action in their statistics. Information on the differences in the scope, definition and methods used compiling the statistics presented in tables 4.1, 4.2 and 4.3 are given below (and indicated in the tables by the symbol +). Countries which do not have data are not included.

Table 5. Ratification of ILO Conventions on freedom of association and industrial relations

This table presents each countrys record of ratification of ILO Conventions on freedom of association and industrial relations. These conventions are international legal instruments that impose on ratifying States binding obligations, compliance with which is subjected to regular international supervision.

Socio-economic indicators

Table 6. Gross domestic product (GDP), population and poverty

The gross domestic product measures the total output of goods and services for final use produced by residents and non-residents, regardless of the allocation to domestic and foreign claims. The figures for 1994 are calculated according to the US dollar value in 1987.

The population data for the year 1996 and the annual growth of the population between 1980 and 1994 are United Nations estimates. The poverty rates are derived from a host of studies, mostly country studies that used a common methodology and similar poverty lines. They are, therefore, generally comparable over time, and between rural and urban areas, but possibly not between countries.

Table 7. Labour force structure and trends

The labour force and its annual growth rate are based on the ILOs most recent labour force estimates and projections. The economically active population is defined as: all persons of either sex who furnished the supply of labour for the production of goods and services during a specified time-reference period. National practice varies between countries as regards the treatment of groups such as the armed forces, members of religious groups, persons seeking their first job, seasonal workers or persons engaged in part-time economic activities. In certain countries all or some of these groups are included among the economically active, while in others they are considered to be inactive. However, the data on economically active populations do not include students, persons occupied solely in domestic duties in their own household, members of collective households, inmates of institutions, retired persons, persons living entirely on their own means, and persons wholly dependent upon others. The labour force participation rate is defined as the percentage of economically active children and adults in the 10 to 14 and 15 to 64 age groups respectively.

Table 8. Employment and unemployment

Statistics on unemployment refer, in principle to all persons of working age who during the reference period are without work, are available for work and are seeking work. These statistics are generally collected through labour force sample surveys. Where possible, the table shows the ILO estimates of unemployment rates as these are more comparable than the raw survey data.

Sources

Tables 1.1., 1.2. & 1.3. Trade union membership data and trade union density rates

Groups which are legally excluded from joining trade unions in selected countries:

Table 2. Profile of national employer's organizations

International Organization of Employers (IOE): Profile of IOE Members (Geneva), June 1996.

Table 3.1. Collective bargaining structure in selected countries

1. ILO: Questionnaire on trade union membership and collective bargaining coverage (Geneva, 1996).
2. L. Madhuku: "Worker participation in a developing country: The case of Zimbabwe", in Comparative Labour Law Journal (Philadelphia, University of Pennsylvania), No. 17, April 1996, pp. 603-612.
3. A. Goldín: Las relaciones industriales: el caso argentino, document prepared for the ILO (Geneva, 1996).
4. OECD: Employment Outlook (Paris, OECD, 1994), pp. 174-175.
5. A. Honeybone: Overview of the changes in the labour relations systems in New Zealand under the Employment Contracts Act, working paper (ILO, Geneva, 1991), pp. 11-15.
6. Estimate by C. S. Venkata Ratnam, International Management Institute, New Delhi 1996.
7. D. Ayadurai: "Malaysia", in S. Deery and R. Mitchell (eds.): Labour law and industrial relations in Asia: Eight country studies (Melbourne, Longman Cheshire, 1993), p. 83.
8. J. Visser: Trends and variations in European collective bargaining, research paper No. 96/2, Centre for Research on European Societies and Industrial Relations (CESAR), Amsterdam University, Feb. 1996, pp. 11-15.
9. P. Burgess (ed.): European management guides: industrial relations and collective bargaining (London, Incomes Data Services, Institute of Personnel and Development, 1996).
10. G. Fajertag (ed.): Collective Bargaining in Western Europe 1995-1996 (Brussels, European Trade Union Institute, 1996), pp.139-147.
11. P. Aro and P. Repo: Trade union experiences in collective bargaining in Central Europe. A report of an ILO survey in Bulgaria, Czech Republic, Hungary, Poland and Slovakia (ILO, Geneva, 1997).
12. K. H. Sivesind: "Bargaining and participation in Norway", in Workplace Europe -- New forms of bargaining and participation, 4th European Regional Congress of the International Industrial Relations Association (IIRA), Helsinki, 24-26 August 1994, pp. 153-174.
13. S. Clarke.: The development of industrial relations in Russia, document prepared for the ILO (Geneva, Dec. 1996), pp. 120-137.

Table 3.2. Collective bargaining coverage rates in selected countries

1. ILO estimate for 1995 (Harare, ILO/SAMAT), 1997.
2. ILO: Questionnaire on trade union membership and collective bargaining coverage (Geneva, 1996).
3. Estimates for 1995 by A. Goldín (University of Buenos Aires), 1996.
4. ILO estimates for 1993, 1994 and 1995 (Lima, Regional Office for Latin America and the Caribbean), 1996.
5. Industrial Relations Information Service (Ottawa), 1997.
6. Ministry of Labour, Chile, 1996.
7. Calculated from data in: R. Harbridge and A. Honeybone (eds.): Employment contracts: Bargaining trends and employment law update 1995-1996 (Wellington, Industrial Relations Centre, Victoria University of Wellington, 1996), p. 7.
8. Estimate by C. S. Venkata Ratnam for 1995, International Management Institute, New Delhi, 1996.
9. Yearbook of Labour Statistics 1992 (Ministry of Labour, Japan, 1993), p. 317, table 187.
10. Yearbook of Statistics 1996 (Department of Statistics, Malaysia, 1996), p. 220.
11. Collective bargaining: An update (Bureau of Labor and Employment Statistics, Philippines, 1996)
12. Calculated from data provided by Ministry of Labour, Singapore, 1997.
13. Y. C. Hwang and C.K. Chiao: "The transformation of industrial relations under democratization in Taiwan, 1987-1996", in Proceedings, Vol. 1, 3rd Asian Regional Congress of the International Industrial Relations Association (IIRA), Taiwan, China, 30 Sep.-4 Oct. 1996, p. 83.
14. "Denmark: Collective bargaining round begins", in European Industrial Relations Review (London, Eclipse Group), Feb. 1997, p. 277.
15. Ministry of Labour, Germany, 1996.
16. J. Visser: Trends and variations in European collective bargaining, op. cit., p. 32.
17. Ministry of Labour, Netherlands, 1996.
18. Ministry of Labour, Norway, 1996.
19. Ministry of Labour, Spain, 1996.
20. Mark Cully and Stephen Woodland, "Trade union membership and recognition" Labour Market Trends, June 1997, p. 238.

Table 4.1., 4.2 & 4.3. Number of strikes and lockouts, workers involved and workdays not worked

ILO: LABORSTA database (Geneva).

Table 5. Ratification of ILO Conventions on freedom of association and industrial relations

ILO: Lists of ratifications by Convention and by country, Report III (Part 2), International Labour Conference, 85th Session, Geneva, 1997.

Table 6. Gross domestic product (GDP), population and poverty

World Bank: World Data 1995 (Washington, DC).

United Nations: World population prospects 1950-2050: The 1996 revision (New York, 1997).

H. Tabatatai: Statistics on poverty and income distribution: An ILO compendium of data (ILO, Geneva, 1996).

Table 7. Labour force structure and trends

ILO: Economically active population 1950-2010, 4th edition (Geneva), Dec. 1996.

Table 8. Employment and unemployment

ILO: LABORSTA database (Geneva).

1. These technical notes are drawn from J. Visser: Global trends in unionization (ILO, Geneva, 1997).
2. J. Visser: "Trends in trade union membership", in Employment Outlook (OECD, Paris, 1991), p. 99.
3. idem.
4. Some of the main problems of statistical coverage, in this sense, are found in Cameroon, India, Indonesia, Mauritania, Namibia, Thailand and, in a minor sense, relating to small unaffiliated and mostly non-manual employee organizations or company staff unions, in Belgium and France.
5. C. Chang and C. Sorentino: "Union membership statistics in 12 countries", in Monthly Labour Review (Washington, Department of Labor), Dec. 1991, pp. 46-53.
6. W. Galenson: Trade union growth and decline: An international study (New York, Praeger, 1994).
7. G. Standing, J. Sender and J. Weeks: Restructuring the labour market: The South African challenge (ILO, Geneva, 1996).
8. J. Baskin: "The social partnership challenge: Union trends and the industrial relations challenge", in J. Baskin (ed.): Against the current: Labour and economic policy in South Africa (Johannesburg, Raven Press, 1996), p. 27.
9. F. Schiphorst: "The emergence of civil society: The new place of unions in Zimbabwe", in H. Thomas (ed.): Globalization and Third World trade unions: The challenge of rapid economic changes (London, Zed Books, 1995).
10. D. Galarneau: "Unionized workers", in Statistics Canada -- Perspectives, Spring 1996, pp. 43-52.