International Comparisons of Annual Labor Force Statistics, Adjusted to U.S. Concepts, 10 Countries, 1970-2008

TECHNICAL NOTES

Introduction

This report presents selected labor force statistics adjusted to U.S. concepts for 1970 onward for the United States and nine developed foreign countries: Canada, Australia, Japan, France, Germany, Italy, the Netherlands, Sweden, and the United Kingdom.

Data in this report refer to the civilian working-age population and are based on figures mainly from national statistical agencies, but also from the Organization for Economic Cooperation and Development (OECD) and the Statistical Office of the European Communities (EUROSTAT). The U.S. data are from a labor force survey (also referred to as a household survey) called the Current Population Survey (CPS). Foreign-country data are based mainly on labor force surveys. All data come from secondary sources; that is, BLS does not conduct any surveys to collect data for foreign countries.

Foreign-country data are adjusted to U.S. concepts to the extent possible. Each section in this report provides an introduction to the U.S. concepts for the variable discussed but further details are available in the BLS Handbook of Methods, Chapter 1, "Labor force data derived from the Current Population Survey," at www.bls.gov/opub/hom.

The following are the exceptions to the adjustments made to foreign-country data:

 

• Lower age limits are adjusted to the age at which compulsory schooling ends rather than the U.S. lower age limit of 16 (for exceptions, see the Working-age population section).
• In the United States, persons on layoff are classified as unemployed because of weak job attachment; however,in some countries, persons on layoff are classified as employed because of strong job attachment. Strong job attachment is determined by national circumstances and is evidenced by, for example, payment of salary or the existence of a recall date.
• For some countries, no adjustment is made for deviations from U.S. concepts in the treatment of unpaid family workers and persons waiting to start a new job.
• While Australia and Japan exclude passive jobseekers from the unemployed, in accordance with the U.S. concept, Canada and the European countries do not. An adjustment is made to exclude them in Canada but not in the European countries where the phenomenon is less prevalent.
• Employment by sector data are not fully comparable with U.S. definitions for the United Kingdom prior to 1983 because data required to make adjustments are not available.

These "unadjusted" differences have a negligible effect on comparisons. Adjustments made for each country are discussed in the Country notes.

For further information on comparability issues, see Constance Sorrentino, "International unemployment rates: how comparable are they?" Monthly Labor Review, June 2000, pp. 3-20, at www.bls.gov/opub/mlr/2000/06/art1full.pdf.

Why are adjustments necessary?

Persons counted as employed, unemployed, or not in the labor force may differ across countries. Some types of workers that are categorized differently include new entrants to the workforce, persons on layoff or working part-time, students, and unpaid family workers. For example, whereas most foreign countries count all unpaid family workers as employed, the U.S. only includes them if they worked at least 15 hours per week. Also, foreign countries sometimes include the career military or national defense force in the labor force, whereas U.S. data are based on the civilian labor force. Another difference between U.S. and foreign-country definitions is with regards to age limits. The lower age limit of the working-age population according to U.S. concepts is 16 while most foreign countries collect data on the working-age population ages 15 and older. In addition, some countries may have an upper age limit.

To compare across countries, these definitional differences must be taken into account. Thus, the foreign-country data presented in this report are adjusted to a common framework — U.S. concepts. More details on adjustments made to foreign-country data for greater comparability are in the Country notes.

2

Country notes

Various methodological changes, such as changes in national data sources and BLS adjustment methods, have created breaks in the historical continuity of series for all countries except Japan. In the Tables, the letter (b) appears next to the first year that contains data affected by the change, indicating a break in series in that year. Some breaks have little or no effect on data and are therefore relatively inconsequential, while others affect comparability over time. This section provides background for each country’s breaks in series and measures the impact when overlapping data are available, as well as a description of the data sources and adjustments (except for age limits, which are indicated in the Working-age population section).

United States

Unadjusted and adjusted unemployment rates are from a monthly labor force survey called the Current Population Survey (CPS).

Adjustments. None.

Breaks. While current survey concepts and methods are very similar to those introduced at the inception of the CPS in 1940, various changes have been made over the years to improve the accuracy and usefulness of the data. These methodological changes and others, such as the introduction of new population controls from the decennial censuses, have caused many breaks in series, most of which had only a small impact on the comparability of data; however, the 1994 and 1990 breaks in series reflect major changes in methodology and are the first to be discussed below.

The 1994 break reflects a major redesign of the labor force survey questionnaire and its methodology. The redesign raised the unemployment rate by 0.1 percentage point in 1994. The 1990 break reflects the introduction of the 1990 census-based population controls, which raised the unemployment rate by 0.1 percentage point in 1990; female employment-population ratios and labor force participation rates are not affected by this break.

The 2004, 2003, 2000, 1999, 1998, 1997, 1986, and 1972 breaks also reflect revised population controls, although these breaks do not affect all series and their impacts are negligible. Additionally, there is a small break for employment and labor force levels in 1978, which reflects the introduction of an expansion in the sample and revisions in the estimation procedures. For further information on the breaks in series and their impacts, see “Historical comparability” in the “Household data” section of the BLS publication, Employment and Earnings, at www.bls.gov/cps/eetech_methods.pdf.

Sectoral employment is based on NAICS for 2000 onward and the Standard Industrial Classification (SIC) system for prior years. The change in classification system, which created a break in the sectoral employment series in 2000, reduced employment for 2000-02 in all sectors except services. The share of employment in agriculture decreased from 2.6 to 1.8 percent; in industry, from 22.2 to 22.0 percent; and in manufacturing (a subsector of industry), from 14.8 to 14.4 percent. Employment in services increased from 75.2 to 76.2 percent.

Canada

Unadjusted unemployment rates are from a monthly labor force survey.

Adjustments. The adjusted series are based on labor force survey statistics. For 1976 onward, BLS adjusts unemployment to:

• Include full-time students looking for full-time work.
• Exclude passive jobseekers.
• Exclude persons unavailable for work due to personal or family responsibilities or vacation.
• Exclude persons waiting to start a new job who did not seek work in the past four weeks (for 1994 onward only since these persons are similarly counted among the unemployed in the United States for prior years).

Statistics Canada provides BLS with the adjusted data for 1976 onward. The adjustments have had a consistently large impact over time, reducing the annual unemployment rate by 0.7-0.8 percentage point in the 2000s, 0.4-0.9 percentage point in the 1990s, and 0.2-0.4 percentage point in the 1980s. For further information on the adjustments and their impacts, see Constance Sorrentino, "International unemployment rates: how comparable are they?" Monthly Labor Review, June 2000, pp. 3-20, at www.bls.gov/opub/mlr/2000/06/art1full.pdf.
 

Breaks. For all series, there is a break in 1976, which reflects the incorporation of the 2001 Population Census results for 1976 onward. For the adjusted series, the 1976 break also reflects the implementation of the adjustments described above.

For the adjusted unemployment series, there is an additional break in 1994, which reflects a change in the definition of unemployment. Persons waiting to start a new job who did not actively seek work in the past four weeks are included among the unemployed prior to 1994; this group is excluded for 1994 onward.

Sectoral employment is shown on the NAICS basis for 1976 onward. These data are also benchmarked to the 2001 Population Census results. Prior to 1976, data are shown on the Canadian SIC basis. The change in classification system, which created a break in the sectoral employment series in 1976, reduced the share of employment in manufacturing from 15.7 to 15.0 percent in 1998.

Australia

Unadjusted and adjusted unemployment rates are from a labor force survey, which has been monthly since 1978 and was quarterly for prior years.

Adjustments. None. The unemployment rate is virtually unchanged when it is adjusted to U.S. concepts.

Breaks. For all series, there is a break in 1986. In April 2001, the Australian labor force survey questionnaire was redesigned to adhere more closely to ILO guidelines; data were revised for April 1986 onward, creating a break in 1986. The redesign reduced the unemployment rate by 0.2 percentage point in 1986.

For the unemployment series, there is a break in 2001, which reflects the reclassification of persons waiting to start a new job from not in the labor force to unemployed. The reclassification raised the adjusted unemployment rate by 0.1 percentage point in 2001.

Japan

Unadjusted unemployment rates are from a monthly labor force survey.

Adjustments. The adjusted series are based on labor force survey statistics. BLS adjusts employment and labor force to:

• Exclude unpaid family workers who worked less than 15 hours per week.
• Exclude military.

Further adjustments are made to unemployment rates by sex for 1985 onward. For details on the methodology, see Sara Elder and Constance Sorrentino, “Japan’s low unemployment: a BLS update and revision,” Monthly Labor Review, October 1993, pp. 56-63, at www.bls.gov/fls/mpres93.pdf, and Toshihiko Yamagami, “Utilization of labor resources in Japan and the United States,” Monthly Labor Review, April 2002, pp. 25-43, at www.bls.gov/opub/mlr/2002/04/art3full.pdf.

Breaks. None.

France

Unadjusted unemployment rates are based on a labor force survey, which has been continuous (i.e., covering all weeks of the year) since 2003 and was annual or semi-annual for prior years.

Adjustments. The adjusted series are based on labor force survey statistics.

Breaks. For all series except working-age population, there are breaks in 2003, 1990, 1982, and 1975. The 2003 break reflects revisions to labor force survey definitions to make them more consistent with ILO recommendations. The 1990 break reflects the incorporation of the 1999 census-based population controls for 1990 onward. The 1982 break reflects revisions to labor force survey definitions and updated population controls. The 1975 break reflects a change in BLS estimation methods and a revised definition of unemployment due to the introduction of a criterion on work availability.

Germany

Unadjusted unemployment rates are based on registered unemployment and wage and salary workers from an establishment survey.

Adjustments. The adjusted series are based on labor force survey statistics. BLS adjusts employment and labor force to:

• Exclude unpaid family workers who worked less than 15 hours per week prior to 1991.
• Exclude career military.

Breaks. For all series there is a break in 1991, which reflects the inclusion of former East Germany. Data for 1991 onward are for Germany (unified); prior to 1991, data are for West Germany. The unification raised the adjusted unemployment rate from 4.3 to 5.6 percent in 1991.

For the unadjusted unemployment series, there is a break in 2005, which reflects a change in national legislation that broadened coverage of the registered unemployed.

For the adjusted series, there are breaks in 2005, 1999, and 1984; working-age population is not affected by the breaks in 1999 and 1984. The 2005 break reflects a change in the periodicity of the German labor force survey data collection. For 2005 onward, data are collected continuously; prior to 2005, data were collected in April of each year. The 1999 break reflects the incorporation of an improved method of data calculation and a change in coverage to persons living in private households only. The 1984 break reflects a change in source: data are from the German Federal Statistical Office prior to 1984 and OECD for 1984 onward.

Italy

Unadjusted unemployment rates are from a labor force survey, which has been continuous since 2004 and was quarterly for prior years.

Adjustments. The adjusted series are based on labor force survey statistics. BLS adjusts employment and labor force to:

• Exclude unpaid family workers who worked less than 15 hours per week.
• Exclude career military.

Breaks. For all series, there are breaks in 1993, 1991, and 1986; working-age population is not affected by the 1986 break. Italy’s labor force survey became continuous in 2004 and Italy’s National Institute of Statistics (ISTAT) revised its definitions to more closely follow ILO and EUROSTAT recommendations; data were revised back to 1993, causing a break in that year. The revisions raised the unemployment rate by approximately 0.5 percentage point in 1993. The 1991 break reflects a revision in the method of weighting sample data, which raised the adjusted unemployment rate from 6.6 to 6.9 percent in 1991. The 1986 break reflects a revision in the survey questionnaire, resulting in a significant increase in the number of people reported as seeking work in the past 30 days. This revision raised the adjusted Italian unemployment rate from 6.3 to 7.5 percent in 1986.

Netherlands

Unadjusted unemployment rates are based on a combination of registered unemployment, wage and salary workers from an establishment survey, and a labor force survey, which has been continuous since 1986 and was biennial for 1973-85.

Adjustments. The adjusted series are based on labor force survey statistics. BLS adjusts employment and labor force to:

• Exclude unpaid family workers who worked less than 15 hours per week.
• Exclude military.

Breaks. For the unadjusted series, there is a break in 1996, which reflects updated population controls.

For the adjusted series, there are breaks in 2003, 2000, 1992, 1987, and 1983; working-age population is not affected by the 1992 break. The 2003 break reflects updated population controls. The 2000, 1987, and 1983 breaks reflect changes in sources. Prior to 1983, data are from Statistics Netherlands; for 1984-86, EUROSTAT; for 1987-2000, OECD; and for 2000 onward, EUROSTAT. The 1992 break reflects a change in labor force definitions.

Sweden

Unadjusted unemployment rates are based on a labor force survey, which has been continuous since 1993 and was monthly for prior years.

Adjustments. The adjusted series are based on labor force survey statistics. BLS adjusts employment and labor force to exclude the career military. In addition, BLS adjusts unemployment for 1987-2004 to include full-time students who are both seeking work and available for work; this group was classified as not in the labor force by Statistics Sweden for 1987-2004. Lastly, BLS adjusts all series so that there is no upper age limit.

Breaks. For all series except working-age population, there are breaks in 2005 and 1987. The 2005 break reflects the introduction of a new questionnaire. For the unadjusted series, the 2005 break also reflects two additional changes: (1) the unemployment definition was revised to include full-time students who are both seeking work and available for work and (2) population coverage changed from persons ages 16-64 to ages 15-74. The latter change had a minimal impact while the inclusion of students raised the not seasonally adjusted monthly unemployment rate by approximately 2 percentage points for April-June 2005.

There are several reasons for the 1987 break. First, a new questionnaire was introduced in that year. Questions regarding availability for work were added and the period of active work-seeking was reduced from 60 days to 4 weeks. These changes reduced the unemployment rate by 0.4 percentage point in 1987. Second, Statistics Sweden implemented new estimation procedures and a new reference week system in 1993. These revisions were implemented for 1987 onward, which raised the unemployment rate by 0.2-0.5 percentage point for 1987-93. Third, for the adjusted series, the break additionally reflects the classification of students seeking work as unemployed rather than not in the labor force for 1987-2004. This adjustment raised the unemployment rate by 0.1 percentage point in 1987 and by 1.0 to 2.2 percentage points for 1993-2004.

For the unadjusted series, there is an additional break in 1986, which reflects a change in population coverage from 16-74 years to 16-64 years.

United Kingdom

Unadjusted unemployment rates are based on estimates from the United Kingdom’s Office for National Statistics (ONS) as well as a labor force survey, which has been continuous since 1992, annual for 1984-91, and biennial for 1979-83.

Adjustments. The adjusted series for 1992 onward are based on labor force survey statistics. The 1971-91 series are based on administrative and labor force survey data and were created by ONS to be consistent with the series for 1992 onward. Data now comply with the current ILO definition of unemployment for 1971 onward. For the detailed methodology underlying the estimates, see Paul Doyle, “Consistent historical time series of labour market data,” at www.statistics.gov.uk/cci/article.asp?id=418. BLS adjusts employment and labor force for all years to exclude the career military.

Sectoral employment data are only partially adjusted to U.S. concepts prior to 1984. That is, the sum of sectoral employment data does not add up to employment shown in Table 2-1 for those years. Percent distributions by sector prior to 1984 are calculated using the sum of sectoral employment data rather than Table 2-1 employment in the denominator

Breaks. For the sectoral employment series, there is a break in 1984, which primarily reflects a change in concepts. Data are only partially adjusted to U.S. concepts prior to 1984 and are fully adjusted to U.S. concepts for 1984 onward. In addition, the source changed: data are from OECD prior to 1984 and ONS for 1984 onward.

 

Last Modified Date: October 2, 2009