Union Membership Statistics in 12 Countries
Chang, Clara, Sorrentino, Constance, Monthly Labor Review
Industrial relations practices differ widely among the developed countries, and union membership data serve as important background information for understanding how relationships between labor and management have evolved over time. Internationally comparable data would be helpful in assessing the relative roles of unions in different countries.
This report investigates the comparability of union membership statistics in the United States and 11 foreign countries and concludes that international comparisons should be made with caution. The figures published by each country are useful indicators of broad trends, but they should not be used to compare levels of unionization, commonly termed union density (union membership as a percent of paid employment).' In some cases, the unadjusted data also present a distorted indication of comparative trends.
Data adjusted for differences in coverage show that the gap between the United States and other countries in union density is not as wide as the unadjusted statistics would indicate. However, the United States remains a country of low union density in comparison with Canada, Australia, Japan, and most of Western Europe.
The United States is unique among the countries studied in that union density has fallen continuously since the mid-1950's. However, during the 1980's, it declined or stagnated in most of the countries examined. Sweden and Denmark had the highest levels and were also the only countries in which union density rose consistently until at least the mid-1980's. Unions in both Scandinavian countries have been highly successful in recruiting women and members of the growing service sector.
The material presented here is based on data and information compiled by the Bureau of Labor Statistics, as well as on several detailed studies by other researchers. George Bain and Robert Price's 1980 study yielded union membership data and presented a full discussion of problems associated with measurement of union-related data in the United States, Canada, Australia, and selected European countries.(2) Kenneth Walsh's and Jelle Visser's subsequent work examined methods of measurement of such data in many European countries.(3) Visser's 1991 update expanded coverage to several non-European countries.(1) The present report draws on Visser's work for adjustments of the European countries' data to a more internationally comparable basis.
The statistical offices in most developed countries have compiled data on union membership for many years. Table I presents these regularly published data for each country studied. However, differences in sources, reporting techniques, definitions, and coverage of the data render comparisons across countries difficult. These differences often reflect the widely varying institutional frameworks within which the unions operate. For instance, the fact that Swedish and Danish unions manage unemployment benefit funds means that unemployed union members remain on the membership rolls in these countries. This is somewhat less likely in other countries, where unemployment may lead to a lapse in union membership, especially if the duration is long.' In addition, retired and self-employed persons who belong to unions are included in the figures for some countries but excluded in others.
Sources. Union membership data are derived from two sources: household surveys and reports undertaken by the unions themselves. Currently, the United States is the only country that derives its time series data on union membership from a household survey. Other countries, including Australia and Canada, have also experimented with the survey technique, but data are available only for a few years. Australia has published results from supplements to its household surveys in August 1976, May 1982, and August 1986, 1988, and 1990. (See table 1.) In Canada, data exist from surveys on labor market activity carried out from 1986 to 1989. …