The hours of work relate to any period of time spent by persons in the performance of activities which contribute to the production of goods and services within the general production boundary as defined by the United Nations System of National Accounts.

The various types of statistics on hours of work presented in Chapter 4 are indicated in the tables by the following codes¹:

• (a) Hours actually worked
• (b) Hours paid for
• (c) Hours usually worked
• (d) Normal hours of work

The series generally relate to employees of both sexes, irrespective of age. Data by sex are presented in around half of the series.

Since 1962, there are international statistical standards only for the hours actually worked and the normal hours of work concepts, adopted by the Tenth International Conference of Labour Statisticians². These are confined to persons in paid employment and describe work situations that are more typical for production workers in manufacturing establishments. While they have served as a guide to countries when producing national estimates, many national definitions have now become more comprehensive. These international standards are currently being revised to improve the coverage of persons and work activities, and to better reflect diverse working time arrangements. New international standards are expected to be adopted by the forthcoming 18th International Conference of Labour Statisticians in 2008, that will include other concepts such as hours usually worked and hours paid for. The following paragraphs provide definitional elements for these concepts as they are being measured by many countries.

The concept of hours actually worked relates to the time that persons in employment spend on work activities during a specified reference period, comprising:

• (a) time spent directly on production (producing goods and services, including paid and unpaid overtime);
• (b) time spent to facilitate production, necessary to work activities or to enhance the performance of persons (design, prepare, maintain the workplace, procedures, tools, including receipts, time sheets, reports), changing time (donning necessary work clothing), transporting activities (door to door, bringing agricultural produce to market) and work-related train-ing;
• (c) time spent in-between main activities (awaiting customers, stand-by for reasons such as lack of supply of work or power, machinery breakdown, accident), travel time to meetings or work assignments, active oncall duty (as for health and technical service professional);
• (d) resting time (short rest or refreshment breaks including tea, coffee or prayer breaks).

Hours actually worked excludes time not worked, whether paid or unpaid, such as:

• (a) annual leave, public holidays, sick leave, parental leave, etc.;
• (b) meal breaks;
• (c) time spent on travel from home to work and vice versa, also known as commuter travel, that is not actually time spent working.

Hours actually worked covers all types of workers, whether in self-employment jobs or in paid employment jobs; it may be paid or unpaid and carried out in any location, including the street, field, home, etc.

The hours paid for concept covers workers in paid employment and comprises all hours, whether worked or not, that have been paid by the employer. When compared to the hours actually worked of employees, it covers all paid hours, even those not worked, such as paid annual vacation, public holidays, paid sick leave, other paid leave, and excludes all unpaid time worked (for example, unpaid overtime).

The concept of normal hours of work covers a subset of persons in paid employment (e.g. those covered by labour laws and regulations) and relates to the hours in a day, week or year that they are expected to be at the disposal of their employer according to legislation, collective agreements or arbitral awards. When compared to the hours actually worked of employees, normal hours of work excludes all overtime (paid and unpaid) but includes all time which is paid at normal rates even if not worked.

The hours usually worked relates to the hours actually most commonly worked per week by persons in paid and self-employment during a long reference period such as a month, season, or other long period, i.e., the typical value of the weekly hours worked. As compared to the normal hours of work, all usual overtime is included in hours usually worked. All time worked not on a usual basis is therefore excluded.

In general, hours of work data are obtained from two main sources, namely establishment surveys or censuses and household sample surveys. Two other sources are official national estimates or administrative records of social insurance schemes. These sources are identified in the tables as a code in parentheses to the right of the country name. The source codes are explained in page XVI.

The first source relates to payroll data derived from establishment sample surveys or censuses, that often furnish at the same time statistics on wages and on employment. These statistics often relate to hours paid for and, to a lesser extent, to hours actually worked. Statistics derived from establishment-based surveys tend to have limited worker coverage, as they cover employees or a subset of them (e.g., wage earners or salaried employees) or who work in establishments above a certain size or in certain industries only. The coverage of the statistics may therefore be importantly limited in countries where most workers are engaged, for example, in small sized establishments or in self-employment jobs. These types of limitations are indicated in footnotes.

The second main source relates to labour force sample surveys or other household based surveys. Statistics derived from this source often relate to hours actually worked, and sometimes to the hours usually worked. In theory they cover the employed population as a whole, including the self-employed. Some statistics from labour force surveys relate to the hours of work in the main job only. This and other types of limitation are indicated in footnotes.

The statistics presented in Tables 4A and 4B relate to employees, as defined by the International Standard Classification of Status in Employment, ICSE-93. When statistics are derived from labour force surveys, they can refer to total employment. When they are derived from establishment surveys, they can refer to a subgroup of employees, namely to wage earners or to salaried employees. Some countries further limit worker coverage to “adults”, “skilled” or “unskilled” employees, in which case it is specified as a footnote. There are no international definitions for wage earners or for salaried employees. The former are generally equated with “manual”, “production” or “blue collar” workers, and the latter with “white collar workers”.

Worker coverage is indicated at the centre of the page for each series.

The series are presented on the basis of the average number of hours of work per week; in a few cases hours per day or per month are shown. These exceptions are indicated in footnotes.

For statistics derived from establishment surveys, average hours actually worked or average hours paid for per week or per month are normally compiled by dividing the total number of hours actually worked or paid for during a week or a month by the average number of workers on the payrolls during the same period. Average hours actually worked or paid for per day are generally compiled by dividing the total number of hours actually worked or paid for during a week, fortnight or month by the total number of days actually worked or paid for during the same period.

In labour force surveys, average hours actually worked are normally derived by dividing the hours actually worked by all persons in employment (or all employees) by the number of persons in employment (or employees). In these calculations, persons absent from work during the whole survey reference week should be included in the calculation. In some series, however, they are excluded to better reflect a “typical” work week, in which case this is indicated in a footnote.

In deriving averages, statistics may be based on one single point in time (i.e., when the survey is carried out once in the year), on a set of points in time (i.e., monthly or quarterly surveys) or on continuous observations (i.e., surveys that cover all weeks in a year). The more points in time are involved in the calculation, the less the estimate will be affected by seasonal fluctuations, annual and sick leave actually taken during the year as well as by variations in strike activity.

In making comparisons of statistics on hours of work, it should be borne in mind that the data are influenced by the averaging method used, the number of data points used, and practices regarding the number of days normally worked per week, regulations and customs regarding weekend and overtime work, the extent of absence from work, etc.. They will also be affected by industrial and job coverage. Statistics that exclude workers in agriculture, where hours worked follow different patterns than in other industries, will not be comparable with those that include them. Similarly, statistics that relate to the main job only will be lower than statistics that cover multiple jobs³.

The data shown in this table refer to, in principle, the average hours of work according to the broadest economic activity distinctions of all major divisions or categories of the International Standard Industrial Classification of all economic activities, ISIC. Two versions of ISIC are used to present statistics, ISIC rev.2 and ISIC rev.3⁴. Major Divisions (for ISIC rev. 2) or Tabulation Categories (for ISIC rev. 3) are shown as codes with an indication of the classification which is used, and the name corresponding to each code is given in the Appendix. Where the national divisions or categories differ from the international groups, this is indicated in footnotes.

This table shows average hours of work in manufacturing by major group (in ISIC rev. 2) or division (in ISIC rev.3) in the manufacturing industry. Where the national major groups or divisions differ from the international groups, this is indicated in footnotes.

^1. For the full set of statistical time series on hours of work held in the ILO data base, see the ILO Department of Statistics statistical dissemination website: (http://laborsta.ilo.org).

^2. For the full text of the resolution, see ILO: Current international recommendations on labour statistics (Geneva, 2000) or the ILO Department of Statistics’ website (www.ilo.org/public/english/bureau/stat).

^3. For information on the differences in scope, definitions and methods of calculation, etc., used for the various national series, see ILO: Sources and Methods: Labour Statistics (formerly Statistical Sources and Methods), Vol. 2 : ~ Employment, wages, hours of work and labour cost (establishment surveys)~ , second edition (Geneva, 1995); Vol. 3 : ~ Economically active population, employment, unemployment and hours of work (household surveys)~ , second edition (Geneva, 1990); Vol. 4 : ~ Employment, unemployment, wages and hours of work (administrative records and related sources)~ (Geneva, 1989). All these volumes can be consulted online at our statistical website http://laborsta.ilo.org.

^4. The detailed classifications can be consulted on the UN Statistical Division website at: http://unstats.un.org/unsd/class/default.htm.