KILM 15. Average monthly wages

Introduction HIDE

KILM 15 presents trends in average monthly wages, both in nominal and real terms (i.e. adjusted for changes in consumer prices), for 115 economies.1 The inclusion of the series for the first time in the KILM 7th Edition was prompted by several considerations. Information on average wages represents one of the most important aspects of labour market information. Wages are a substantial form of income, accruing to a high proportion of the economically active population, namely persons in paid employment (employees). In most developed economies, more than 85 per cent of the employed population are paid employees, and the share of paid employees has been constantly rising in many of the newly industrializing countries (see KILM 3). For the first time in history, roughly half (49.9 per cent2) of the world’s workers were employees and thus wage recipients in 2009. Information on wage levels is essential to evaluate the living standards and conditions of work and life of this group of workers in both developed and developing economies. It helps to assess how far economic growth and rising labour productivity (KILM 17) translate into better living standards for ordinary workers and to the reduction of working poverty (KILM 18).3

There is also a particular need for information on average wages in planning economic and social development, establishing income and fiscal policies, fixing social security contributions and benefits, and in regulating minimum wages and for collective bargaining. Policy-makers, as well as employers and trade unions, pay close attention to wage trends. At the global level, the ILO’s biennial Global Wage Report analyses wage trends across different regions and discusses the role of wages policies (see box 15a).4 In addition to the relevance of wage data, international standards were long ago developed, adopted and implemented for the concepts, scope and methods of collection, as well as for the compilation and classification of wages statistics (see “Definitions and sources”). This should, in principle, facilitate international comparisons.

A final consideration relates to the presentation of the indicator both in nominal terms and in real terms. Series of wage statistics are generally available in nominal terms, expressed in absolute figures and in national currency. This reflects the way these data are collected, usually from those who pay wages (enterprises) or from those who receive them (paid employees). Wage statistics in nominal terms (and in national currency) are required by policy-makers, who set minimum wages in nominal terms, or by employers and trade unions, who bargain over nominal wage rates. Other data users also need nominal wage data, for example if they want to compare wage levels to other indicators that are available in nominal form (such as poverty thresholds or prices of goods), or if they want to convert them from one currency to another.

However, changes in nominal average wages are not necessarily very informative when it comes to assessing changes in the welfare of wage earners: They indicate only the earnings of an average employee in monetary terms, but not the amount of goods and services that can be purchased with wages. In other words, nominal wages do not provide information on the purchasing power of employees. This purchasing power is influenced by, among other factors, increases (or decreases) in prices of goods and services that employees acquire, use or pay for – i.e., by the inflation rate (see “Use of the indicator” for more information). Average monthly wages are therefore not only presented in nominal terms, but also in real terms by adjusting for changes in consumer prices. Note, however, that the consumer price index (CPI) reflects price changes as viewed from the perspective of the average consumer and that some wage earners might experience a different rate of price changes (for example, when they spend a higher proportion of their income on food items than the average consumer).5

Both the nominal and real average wage series are presented in national currency. This enables data users to calculate nominal and real wage growth rates without distortion caused by exchange rate fluctuations, and to link wage data to other data expressed in national currency. It also takes account of the fact that wage levels may not be strictly comparable across countries due to methodological differences, while growth rates are less likely to be affected by statistical effects (see “Limitations to comparability”). However, researchers who want to compare wages between countries can convert the nominal average wages into a common currency of their choice (which would not be possible in the case of an index).6

Table 15a shows average monthly wage series from the ILO’s Global Wage Database that was compiled for the latest edition of the Global Wage Report on the basis of official, national sources.7 Altogether, 115 economies are covered. The series referring to real average monthly wages was standardized to 2005 as a common base year.

Use of the indicator HIDE

The data on average monthly wages can be useful for a number of purposes, including the following:

(1) Real wages in an economic activity are a major indicator of employees’ purchasing power and a proxy for their level of income, independent of the actual work performed in that activity.8 Real wage trends are, therefore, useful indicators both within countries and across them. Significant differences in the purchasing power of wages, over time and between countries, reflect the modern world economy, and comparisons of the movement of real wages can provide a measure of the material progress (or regression) of the working population.

(2) Related to the point above, real average wages are an important indicator for monitoring changes in working conditions. They are one of the Decent Work Indicators in the substantive area of “Adequate earnings and productive work” and should be reviewed in conjunction with trends in working poverty (KILM 18) and the low pay incidence. Further, changes in the composition of employment – notably shifts between paid employment and self-employment – should be considered as contextual information (see status in employment, KILM 3).

(3) Trends in nominal wages can be used to inform adjustments in minimum wages, the lowest remuneration that employers may legally pay to workers under national law.9 While there is no single, recommended ratio between minimum wages and average wages, information on average wages can inform policy-makers when setting minimum wages and enable them to monitor whether those at the bottom of the distribution fall behind general wage increases.10 The Minimum Wage Fixing Convention, 1970 (No. 131), makes explicit reference to the general level of wages in a country as an element that shall be taken into account when determining the level of minimum wages.

(4) Likewise, the social partners – workers’ and employers’ organizations – rely on wage data for collective bargaining.11 A fundamental concern of employees and trade unions is to protect the purchasing power of wages, particularly in periods of high inflation, by raising nominal wages in line with changes in consumer prices. Real wage increases become feasible without putting the sustainability of enterprises into jeopardy when labour productivity is growing (see KILM 17). Since wage bargaining usually takes place at the sectoral or enterprise level, it can be useful to complement data on wages in the total economy with data disaggregated by sector, for example, the manufacturing sector.

Box 15a. The ILO’s Global Wage Report

The biennial Global Wage Report is the ILO’s flagship publication on wage trends and wage policies. It uses a number of indicators to analyze global wage developments, including the growth of average real wages, the low-pay incidence (defined as the share of wage workers with earnings below two-thirds of the median) and the wage share in national income. It is unique in its global scope and builds on data from some 115 countries and territories that between them account for 94 per cent of the world’s wage workers. Based on a standard methodology that corrects for the remaining response bias, the report documents wage growth for the world and in seven regions.

The report also provides practical illustrations of how collective bargaining, minimum wages and income policies can be building blocks of effective wage policies that contribute to equitable outcomes. It is inspired by the objective to promote “policies in regard to wages and earnings, hours and other conditions of work, designed to ensure a just share of the fruits of progress to all and a minimum living wage to all employed and in need of such protection”, one of the central elements of the Decent Work Agenda (see ILO Declaration on Social Justice for a Fair Globalization). The relevance of this approach has been underscored by the global economic crisis, during which many governments expanded income support policies and wage subsidies in order to stabilize domestic demand and to support recovery.

Further information is available from the Conditions of Work and Employment Programme (TRAVAIL), the ILO’s lead programme on wage data analysis and policy advice, or online at http://www.ilo.org/global/publications/ilo-bookstore/order-online/books/WCMS_145265/lang--en/index.htm.

(5) When used together with other economic variables such as employment, production, and income and consumption, trends in average real wages are valuable indicators for the analysis of overall macroeconomic trends, as well as in economic planning and forecasting. Importantly, they can indicate the extent to which economic growth and rising labour productivity translates into income gains for workers. These, in turn, influence aggregate demand, and countries with external surpluses can utilize wage policies to re-balance their economies by strengthening domestic consumption.

(6) At the international level, comparisons between countries of the movement of real wages over time can be relatively objective, since the growth rates are free from the influence of currencies and exchange rates. For the purposes of the Global Wage Report (see box 15a), the ILO has developed a methodology that uses national wage data to produce regional and global estimates on the growth of real average wages.12 These figures show gaps in performance between regions, and illustrate the differential impact of the global economic crisis on wage incomes of workers around the world (see “Trends”).

While the levels and trends of real wages of employees are an indication of the evolution of their purchasing power, it should, however, be noted that they are only an approximation to changes in the standard of living of the population, as the two concepts are not identical:

A common misuse of information on trends in real average wages is that of measuring changes in labour costs and/or unit labour costs in real terms. Firstly, wages represent only one part of total labour costs and the movements of average wages over time are not necessarily identical to those of total labour costs. Secondly, even if real wages rise and labour costs increase proportionately, this does not lead to an increase in real unit labour costs if wage increases reflect gains in labour productivity.

Definitions and sources HIDE

Statistics of real wages are not primary statistics. They result from the combination of two types of primary statistics – nominal wages and prices. The computation of real wage trends, therefore, requires some preliminary explanation of what is meant by real wages, and the method followed to compute them.

“Real wages” have been defined in the ILO resolution adopted by the Eighth International Conference of Labour Statisticians (ICLS) in 1954, as “the goods and services which can be purchased with wages or are provided as wages”.13 This definition establishes a useful basis for the computation of real wages and their comparison from one period of time to another, or between one country and another. The information required for the computation of real wages includes: (a) a wage measure expressed in monetary terms; (b) a series of prices of goods and services commonly purchased by employees; and (c) information on the consumption pattern of employees. Thus, in a given country, to provide an indication of changes in the purchasing power of wages resulting from changes in prices of consumer goods and services, the wage information – item (a), above – is combined with a consumer price index which, in principle, reflects items (b) and (c).

Different types of wage data correspond to different concepts of wages. The ILO resolution concerning an integrated system of wages statistics, adopted by the 12th ICLS (1973), contains – among other things – the definitions of “wage rates” and “earnings”.14 It also endorses the concept and definition of “labour costs” adopted by the 11th ICLS (1966).15 Another related, but conceptually different measure, “compensation of employees”, is used in connection with the national accounts. Guidelines were also adopted by the 16th ICLS (1998) for measuring the full income related to paid and self-employment.16

All these measures are designed to cover different aspects of wages and employment-related income. While wage rates are similar to price quotations for labour and measure the basic remuneration per time unit or unit of output, compensation of employees and labour cost correspond to a concept of cost to the employer and include components that do not actually represent current income to employees (see KILM 16).17 The wage measure that best corresponds to the concept of income to employees is that of earnings, as defined by the 12th ICLS (see box 15b). The 1954 resolution concerning the international comparison of real wages already indicated that “[a]s a point of departure for the purpose of computing ratios of real wages, wages should be average earnings ...”.

The price element is composed of two sets of data – a series of prices of goods and services commonly purchased by the reference population, and data on the consumption pattern of that population, that is, the quantities consumed which serve as an indication of the relative importance of the different goods and services at a given point in time. These two sets of data are present in a consumer price index (CPI), the definition of which is contained in the ICLS resolution concerning consumer price indices.18

Ideally, for the purposes of real wage computation, the two sets of data (wages and prices) should cover the same reference population (in the present case, the same employee group or category) and have similar geographic and industrial coverage and reference period.

Real average monthly wages presented in KILM 15 were calculated by the ILO based on nominal average monthly wages (also reproduced as part of the indicator) and annual average consumer price indices (CPI) from the International Monetary Fund.19 Some considerations that should be kept in mind with respect to both nominal average wages and CPI data will be discussed below.

Nominal average monthly wages are based on a variety of national sources, as published by national statistical agencies. In an ideal case, the indicator refers to monthly average wages in the sense of “earnings” (as defined by the 12th ICLS; see box 15b) for the entire economy and all employees in a given country. However, countries use different approaches when collecting wage data. Methodological differences relate to the type of source used, the coverage of the source, and how the data are aggregated to produce monthly average wages. When data for the target concept were not available, closely related wage series were used instead. Comparability issues that result from the variety of approaches to measuring wages are discussed in the next section.

The most common source for wage data – in particular in advanced economies, in Central and South-Eastern Europe and the CIS countries – are labour-related establishment surveys. They collect data at the source, namely from establishments that employ workers. Since establishments usually keep accurate records of all wages paid for their own book-keeping and for tax purposes, this approach has the advantage of producing reliable wage data without having to rely on the re-call of individual employees. However, in countries where enterprises routinely pay wages outside their normal book-keeping (so-called “envelope wages”) in order to avoid taxes and social security contributions, the establishment-based approach has limitations.

While most countries include firms regardless of size into establishment surveys, some countries exclude small firms with less than five or less than ten employees. Some countries also limit the coverage to the private sector (i.e. exclude the public sector) or to specific industries within the private sector (such as manufacturing). If small enterprises pay lower wages than large enterprises or wages differ between the public and the private sector, these exclusions will affect the level of the collected wage data – depending on how large differences are, and how many employees are excluded from the coverage. However, if wages in the excluded establishment move roughly in line with those enterprises for which data are available, these exclusions will only have a marginal effect on trends over time. Even data with less than full coverage can therefore be a useful proxy to analyze wage growth in an economy.

Establishment surveys usually draw their sample from an establishment register that is maintained either by the central statistical office or another institution, such as the Registrar of Companies. In developing countries with a large informal sector, this is a serious limitation since many small, unregistered establishments are missing from the sample frame. Also excluded are individual households employing paid domestic workers, which account for a significant proportion of total paid employment in some developing regions (see KILM 4).20 In some developing countries, establishment surveys therefore capture only a small proportion of all wage employees (those in the public sector and those in large, modern enterprises). Under these circumstances, collecting information from the recipients of wages can be the better alternative.

Household surveys, the second major source for wage data, have the advantage that they cover all employees regardless of where they work.21 Wage data from household surveys usually cover the public and private sector, formal and informal enterprises and all industrial sectors. There are, however, a number of subtle methodological differences that can affect comparability between countries of wage levels based on household surveys. For instance, some surveys collect data on the usual monthly wages while others ask for the actual wage received in the past month. At times it is also not clear whether respondents are asked to report their gross or net wages (i.e. before or after deduction of taxes and compulsory social security contributions). These differences can have a material effect on the reported level of wages, while they are less likely to have a major impact on trends over time as long as the survey instrument remains unchanged.

Finally, a few countries rely on administrative data sources such as social security records to compile wage data (Albania, Portugal and Singapore), or combine several different primary sources to produce a synthetic wage series (Austria, Bahrain, Canada and Latvia). In some countries the national accounts sections of central statistical offices produce the wage series that match the desired concept most closely (Algeria, Germany and Italy). However, national accounts are only a useful source for data on average wages when compensation of employees is disaggregated into its two major components – wages and salaries and employers’ social contributions – and when matching data on total wage employment exist.

While most countries report wages with a calendar month as a reference period, some report only daily, weekly or annual wages. In order to ensure comparability, these source data were standardized into the same monthly reference period, e.g. annual wages were divided by 12 months to produce average monthly wages.

The CPI series show measures of change over time in the general level of prices paid by consumers for goods and services. They combine information about consumption patterns at a given point in time and price changes. Where the original CPI series were not uniformly based on an index of 2005 = 100, they have been recalculated by dividing the index for each data set shown by the index for the base year and multiplying the quotient by 100. This operation is a simple re-basing of the index and does not involve any change in the weighting systems on which the original IMF data are based.

Consumer price indices are usually computed according to the following methodology: prices of selected goods and services are regularly collected from a sample of localities and outlets; special techniques for price collection are used in respect of such items as electricity, medical care, education, transport, communication, and so on. The source of the weights (the types and quantities of goods and services consumed) used for the index is generally a household expenditure survey (or a household budget survey or household income or expenditure survey).

As far as national resources permit, such surveys are representative in terms of household size, income level, regional location, socio-economic group and any other factors which have a bearing on household expenditure patterns. In a few cases, the weights are obtained from the national accounts or other sources, such as surveys of production, exports and imports, and retail trade. Administrative sources may also be used to obtain an estimated consumption pattern. In a large majority of countries, CPI series cover the whole territory and the entire population. In a few cases, the CPI is limited to urban areas (for example, in some Latin American countries).22

Limitations to comparability HIDE

As mentioned in the preceding section, country-specific practices differ with respect to the sources and methods used for wage data collection and compilation, which in turn have an influence on the results and comparability across countries. The main sources of information (establishment censuses and surveys, and household surveys) usually differ in terms of objectives, scope, collection and measurement methods, survey methodology and so on. The scope of the information may vary in terms of geographical coverage, workers’ coverage (for example, exclusion of part-time workers)23 and establishment and enterprise coverage (based on establishment size or sector covered). While household surveys encompass a greater range of jobs and workers than establishment surveys, they tend to experience problems associated with self-reporting of earnings.

Even when using the same concept of wages (for example, earnings), there are likely to be differences with regard to the inclusion or exclusion of various components (such as periodic bonuses and allowances, or payments in kind). Earnings statistics show fluctuations that reflect the influence of both changes in wage rates and supplementary payments. In addition, daily, weekly and monthly earnings are dependent on variations in hours of work (in particular, hours of paid overtime or short-time working), while hourly earnings are influenced by the concept of hours of work – hours actually worked, hours paid for, or normal hours of work – used in the computation (see KILM 7 for information on the various concepts pertaining to hours of work).

How much do these methodological differences matter? One way to assess the likely impact of differences in definition and measurement approach is to look at countries that produce more than one wage series. For illustrative purposes, figure 15a presents data from three countries that produce one wage series that closely matches the target concept and an alternative series where concepts and methodology differ substantially – for example, by limiting coverage only to full-time employees or excluding entire economic sectors. This provides for a ‘natural experiment’ to assess the likely impact of underlying differences.

In Sweden, the main series used in figure 15a refers to average monthly real wages for the entire economy and all employees (based on an establishment survey). By contrast, the secondary series (which is not included in KILM) refers to monthly average wages in September for non-manual workers in private sector enterprises with five or more employees, and it also excludes two industries.24 The latter series, therefore, substantially diverges from the target concept. Nonetheless, average wages in the secondary series are only approximately 15 per cent higher, and the trend over time is almost identical. A similar finding emerges from Australia that produces two alternative wage series; one from an establishment survey and the other from a household survey. Again, the level of measured average wages differs slightly, but both series show the same movement over time (the divergence in the last years can be attributed to a series break in the main series in 2007).25 A rare example for differences not only in level, but also in the trend over time is Germany. Here, the main series used in KILM 15 refers to real monthly average wages for the entire economy and all employees (derived from national accounts) and shows a decline between 1999 and 2009. When the coverage is restricted only to full-time employees (and excluding some industries), wages for this sub-group have grown over the same period. The divergence between the two time series is due to a composition effect, since the share of workers in atypical forms of employment (and with low wages) has grown substantially in the past decade.26

Figure 15a. Impact of methodological differences on wage statistics in Australia, Germany and Sweden, 1999-2009

Source: ILO, Global Wage Database, based on national sources.

The main insight from the analysis above is that a comparison of real wages over time within a given country raises fewer difficulties than cross-country comparisons of wage levels, which are more sensitive to methodological differences. This is reflected by publications such as the ILO’s Global Wage Report, which focuses primarily on growth rates of wages and compares them to trends in labour productivity and other indicators. However, one should be careful not to exaggerate the extent to which methodological differences compromise international comparability. When compared to a developing nation, even the secondary series for Australia, Germany or Sweden would correctly indicate that wages are much higher in industrialized countries.

The German data also point to a separate phenomenon, namely that trends over time are also influenced by changes in the employment structure – the relative shares of men and women, unskilled and skilled labour, full- and part-time workers, and so on – between various reference periods. It can therefore be useful to further analyse whether large compositional shifts have contributed to the observed changes in real average wages. In the German case, for example, the discrepancy between the two series is due to the substantial increase in the proportion of workers in atypical forms of employment (who receive lower wages than the average).27

When making comparisons of real wage trends between countries, one should keep in mind that this indicator is not only based on country-specific series of wages, but also that measures of real wages will be affected by the choice of the price deflator, that is, the CPI. The scope of CPIs can vary not only in terms of the types of household or population groups covered, but also in terms of the geographical coverage. Country-specific practices also differ regarding the treatment of certain issues relating to the computation of CPIs, including the treatment of seasonal items, new products and quality changes, durable goods and owner-occupied housing, the inclusion or exclusion of financial services and indirect taxes, and so on.

There are also differences in the methods used for collecting prices and compiling the indices. The price data for the different items are normally weighted in order to take into account the relative importance of each item with respect to total consumption expenditure. In most countries, the indices are computed in a derived form such as weighted arithmetic averages of prices for a selected number of representative items between the period under consideration and the base period, using a form of Laspeyres’ formula or Laspeyres’ chain index.

If workers and their families in all countries consumed the same goods and services in the same proportions and if this pattern of consumption did not change over time, determination of the purchasing power of wages would be relatively simple. But the main problem is that workers in different countries consume different goods and services, and consume them in proportions that do not follow a standard pattern. Differences in country-specific socio-economic structures lead to differences in the market basket of goods and services that people expect or hope to attain. Largely dissimilar consumption patterns are often a feature of countries with markedly different socio-economic structures. Such differences are encountered in comparisons between highly developed economies and developing economies, and between population groups with different social organization, different consumption habits and different philosophies of life.

Other factors may influence the comparability of real wage trends – and therefore purchasing power – across countries. One is the reference period of both wages and CPIs. Annual averages of hourly or monthly wages may be averages of information based on weekly, monthly or quarterly reference periods. In some cases, they are based on the whole calendar or financial year. On the other hand, the CPI data are annual averages of an index that is compiled, in most cases, monthly, or in a few cases quarterly or biannually. When nominal wages and CPI information do not refer to exactly the same period, this can give rise to problems for countries experiencing rapid inflation.

In spite of these comparability issues, which are inherent in the underlying statistical series, every effort has been made to choose information that is as close as possible to the target concept and thus comparable across countries. Apart from standardizing reference periods to monthly for those countries using a different reporting period, no further adjustments were made to the data in order to preserve the integrity of the data and ensure transparency. The indicators are presented with detailed methodological information so that KILM 15 should be a useful source for assessing trends in nominal and real average monthly wages in the countries concerned.

Trends HIDE

1. The number of economies with available data is 115, with 99 of these having data for more than one point in time.

2. ILO : Global Employment Trends 2011 : The challenge of a jobs recovery (Geneva, 2011); http://www.ilo.org/trends.

3. This was also the rationale for including average real wages into the ILO’s list of Decent Work Indicators; see ILO: Guide to the new Millennium Development Goals Employment Indicators and the full set of Decent Work Indicators (Geneva, ILO, 2009); http://www.ilo.org/global/publications/books/WCMS_145265/lang--en/index.htm.

4. ILO: Global Wage Report 2010/11: Wage policies in times of crisis (Geneva, 2010). Part of the analysis presented in this chapter draws on this publication.

5. For some purposes, other price measures such as the producer price index (PPI) or implicit GDP deflator might be more appropriate.

6. Historical exchange rates are available from a number of different sources, including from the IMF’s Exchange Rate Archive at http://www.imf.org/external/np/fin/data/param_rms_mth.aspx.

7. For a few countries, international databases such as the LABORSTA or EUROSTAT were used.

8. Real wages can also be viewed as remuneration for a given contribution in time, skills, and so on, to determine differences in wages received for the same or similar work. This approach requires the identification of similar jobs, occupations, or skill groups, within industries or between countries, and the availability of detailed data on occupational wages, prices of goods and services consumed by specific groups of workers, and so on. This analysis goes beyond the scope of the present indicator.

9. Note that minimum wages are set in nominal terms, so nominal average wages are the primary comparator. For a review of minimum wage legislation, see ILO, Working Conditions Laws Report 2010 (Geneva, 2010).

10. See Chapter 5.2 of the ILO: Global Wage Report 2010/11, op. cit.

11. See Chapter 5.1 of the ILO: Global Wage Report 2010/11, op. cit.

12. See Technical Appendix I in the ILO: Global Wage Report 2010/11, op. cit.

13. Resolution concerning the international comparison of real wages, adopted by the Eighth International Conference of Labour Statisticians, Geneva, 1954.

14. Resolution concerning an integrated system of wages statistics, adopted by the 12th International Conference of Labour Statisticians, Geneva, 1973; http://www.ilo.org/global/What_we_do/Statistics/standards/resolutions/lang--en/docName--WCMS_087496/index.htm.

15. Resolution concerning statistics of labour cost, adopted by the 11th International Conference of Labour Statisticians, Geneva, 1966; http://www.ilo.org/global/What_we_do/Statistics/standards/resolutions/lang--en/docName--WCMS_087500/index.htm (see box 16a in KILM 16).

16. Resolution concerning the measurement of employment-related income, adopted by the 16th International Conference of Labour Statisticians, Geneva, 1998; http://www.ilo.org/global/What_we_do/Statistics/standards/resolutions/lang--en/docName--WCMS_087490/index.htm.

17 Note, in particular, that under the 1993 System of National Accounts (SNA) the compensation of employees also includes employers’ social contributions (for example, to pension schemes or health insurance). These are not part of wages or earnings as discussed in this chapter.

18. Resolution concerning consumer price indices, adopted by the 17th International Conference of Labour Statisticians, Geneva, November-December 2003; http://www.ilo.org/global/What_we_do/Statistics/standards/resolutions/lang--en/docName--WCMS_087521/index.htm.

19. In the cases of Brazil and the United States, where national counterparts recommended the use of an alternative CPI that better mirrors the coverage of the wage data, we relied on national sources from the Instituto Brasiliero de Geografia e Estatistica (IBGE) and the Bureau of Labor Statistics (BLS), respectively. There are also cases where the IMF does not provide a CPI; in these cases a national CPI was used.

20. According to recent ILO estimates, the global share of domestic workers in paid employment is 3.6 per cent, but reaches 11.9 per cent in Latin America and the Caribbean and 8.0 per cent in the Middle East. See ILO: Global and regional estimates on domestic workers, Domestic Work Policy Brief No. 4 (Geneva, ILO, 2011).

21. Persons living in institutional households, such as military barracks, prisons or monasteries, are commonly excluded.

22. Detailed information regarding national practices in the compilation of consumer price indices is provided in ILO: Sources and Methods: Labour Statistics, Vol. 1: Consumer Price Indices. The Sources and Methods are available online at http://laborsta.ilo.org.

23. It should be noted here that wage series covering all persons employed should not be directly compared with series covering employees only, since a bias may be introduced with the inclusion of working proprietors and contributing family members.

24. Namely, it excludes ISIC Revision 3, tabulation categories A (Agriculture, hunting and forestry) and O (Other community, social and personal service activities).

25. For both series, monthly wages were obtained by multiplying weekly wages by 52 weeks and then dividing by 12 months.

26 See ILO, Global Wage Report 2010/11: Datenblatt Deutschland (Geneva and Berlin, ILO, 2010); available in German only at http://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/@publ/documents/publication/wcms_150027.pdf.

27. For an in-depth analysis of the German case, see Global Wage Report 2010/11: Datenblatt Deutschland, op. cit.

28. This section draws heavily on ILO: Global Wage Report 2010/11, op. cit. Apart from the data series for the countries reproduced in table 15, data for countries not covered by KILM were included, and, in some cases, index figures were used to calculate growth rates. To avoid response bias, which arises when non-responding countries have different characteristics from those of responding countries, the report uses a methodology to adjust for this bias. This standard methodology ensures that all regions are represented in the global wage trend in proportion to their size, and that the global wage trend is not distorted by differences in data availability between regions. See also the Technical appendix I of the Global Wage Report 2010/11.

29. According to official statistics deflated by the IMF’s consumer price index (CPI), China’s average wages grew by 13.1 per cent in 2007, 11.7 per cent in 2008 and 12.8 per cent in 2009. It should be noted that official statistics on wage growth published in the China Yearbook of Statistics refer only to “urban units”, which in practice cover mostly State-owned enterprises, collective-owned units and other types of companies linked to the State. An initial pilot survey of all enterprises conducted by China’s National Bureau of Statistics showed that average annual salaries in the private sector rose by only 6.6 per cent in 2009.

30. See for example, F. Peng, and W.S. Siebert: “Real wage cyclicality in Italy”, in Labour (2008), Vol. 22, No. 4, 2008, pp. 569–591; G. Solon, R. Barsky and J.A. Parker: “Measuring the cyclicality of real wages: How important is composition bias?”, in Quarterly Journal of Economics (1994), Vol. 109, No. 1, pp. 1–25; or P.J. Devereux and R.A. Hart: “Real wage cyclicality of job stayers, within-company job movers, and between-company job movers”, in Industrial and Labor Relations Review (2006), Vol. 60, No. 1, pp. 105–119.

© 1996-2011 International Labour Organization (ILO) | Copyright and permissions | Privacy policy | Disclaimer