7. National statistics on pay and gender in different occupations – and their shortcomings

To understand the gender pay gap at an overall, national level, gender-disaggregated statistics are required that show how different factors relate to pay. It is not only single factors that are of interest but correlations between several different factors. These include, for example, occupational classifications in different levels of detail, gender distribution within occupations, labour market sector, full and part-time pay, age structure, migrant background and education.

The national statistical authorities of all the Nordic countries provide a statistical database for the general public, which enables users to create their own tables by combining different data. In the case of pay and salaries, for example, this might include choosing which occupations, pay measures, statistical measures, genders and years a table should cover. However, the choices are not unlimited, and different statistical authorities have prepared different bases from which to compile tables. The choices available at any given time are of course crucial with regard to the comparability of different occupations. In the following section, we provide our background thoughts for the analysis of the possibilities offered by the national databases.

The level of detail of the occupational classifications (from a general overview at the one-digit level to a more detailed breakdown four-digit level) associated with statistical data on pay can largely determine the possibility of analysing the pay situation of occupations of equal value. In the section Classification occupations, it is argued that the four-digit unit code, which offers the most detail, is the most relevant in the comparison of jobs of equal value, which is why there is particular focus on it in the review of the national statistical databases.

The gender distribution in an occupation is key for identifying gender pay gaps between occupations in which work of equal value is performed where the majority of those working in the occupations are women and men respectively. Comparisons between sectors are relevant, as pay levels may differ between sectors even if those working within a given occupation perform similar work.

In official reports on gender pay gaps, part-time pay is usually converted into full-time pay to allow for comparisons and calculate pay gaps. This practice obscures the role that part-time work plays in the pay gap, with part-time work being significantly more common among women than men. Part-time workers do not earn less simply because they work fewer hours; analysis of hourly pay reveals that part-time workers are also paid less per hour worked (Rubery & Grimshaw, 2015). Working part time can also affect pay in other ways: in addition to affecting hourly pay directly, it can affect opportunities for skills development, which in turn affects pay. The concept of part-time work is also not fixed: working almost full time can be seen as having a different impact on pay than working only a few hours a week.

Age also affects pay development, and occupations in which workers of one sex are generally younger than those of the other sex can be assumed to have a different pay structure than those in which the ages are more gender-mixed. In general, the gender pay gap increases with age; while it is smaller or non-existent among younger workers, men’s pay often increases more quickly, widening the gap among those of upper-middle age (Måwe, 2019). However, Wagner et al. (2020) find that age is correlated with education and increases the pay gap, particularly in occupations with low educational requirements.

There are relatively few studies on how migrant background affects pay in the Nordic countries (Måwe, 2019). No research has been conducted on how migrant background affects the pay gap across different occupations – for example, if one occupation has significantly more migrants than another. Such a study would also be difficult to conduct based on publicly available statistics, since pay statistics do not include country of birth.

Formal education is also an important factor, and the easiest to define, when valuing different occupations and offers an indication of how different occupations do, and should reasonably, compare, in terms of pay. The gender pay gap is often greater among high-paid workers than low-paid workers, even in percentage terms (Ekberg et al., 2023a; Måwe, 2019; Wagner & Teigen 2022). This is consistent with the findings of previous research: the pay gap is larger among the highly educated than the less-well educated (Wagner et al., 2020).

Given their importance in relation to the gender pay gap, the statistical variables that have been discussed here will be highlighted throughout this report. In the following section, we review some of the features of the statistical databases that are freely available in the different Nordic countries that make it difficult to compare men’s and women’s pay and gender pay gaps across occupations.

In the Finnish statistical database, the number of women and men employed (i.e. the basis for calculating gender distribution) in different occupations is provided at the four-digit level but in three different tables: for the private, central government and local government

As of the 2023 pay statistics, the municipal government sector is referred to as local government.

sectors. This makes it difficult to compare pay across sectors. All tables based on monthly pay generally include mean, median and percentile pay, as well as basic pay (without increments) and monthly pay (with fixed increments). The number of part-time employees, broken down by gender, is available in the pay tables for each sector, which present monthly pay at the four-digit level. However, part-time workers’ pay is not included in the mean or median pay data, i.e. this pay data is simply omitted and has not been converted into full-time pay.

Hourly pay is presented in separate tables for full-time and part-time employees, both overall and by sector, but only at the three-digit level for the overall statistics. However, the gender distribution of occupations is not available in the same table and has to be obtained elsewhere.

In addition to the national occupational classification, pay for women and men in the central and local government (but not private) sectors can also be searched by the occupational titles that appear in employment and other contracts, of which there are 674.

Only in the private sector is it possible to combine variables on occupation, age, pay and gender – but only at the two-digit level. Otherwise, there are no tables in which the variables of age and occupation can be combined, i.e. it is not possible to see whether the gender composition differs between age groups within an occupation. Gender is excluded in the table combining educational attainment and age with pay.

It is possible to combine education with gender and pay for each sector individually, but comparisons between sectors are difficult to make. The education codes used vary between sectors: for local and central government a level and field code are used (101 fields and 7 levels), for private sector employees who are paid monthly a uniform system is used that assigns 476 designations for different degrees, while private sector employees who are paid hourly are reported in the same way as employees in central and local government.

An archive is available that includes older tables, the earliest being from 2005, allowing statistical developments to be traced back. These are stored by year, meaning that comparisons of two or more points in time require data to be compiled from several different annual tables.

In the Norwegian statistical database available through Statistics Norway, a single table allows users to see data on occupational pay by sector, gender and working hours for the years 2015–2023. In this example table (11418), highlighted here, it is possible to combine a variety of information to make comparisons of the gender pay situation between occupations, such as information on the gender distribution within occupations and various statistical measures for comparisons.

In the Norwegian statistics, the sectors are divided into three: the private sector and publicly owned enterprises, local government and central government. Pay statistics on occupations in the sample table can be disaggregated from the one-digit to the four-digit level as per the Norwegian occupational classification STYRK-08, according to user preference. Part-time pay is converted to be equivalent to monthly full-time pay. In the table, part-time work is defined as less than 100% working time, and full-time as working 100% or more. In the table on gender and pay at the four-digit level, it is possible to select from the last ten reporting years.

The example table does not provide the age distribution (age structure) of occupations, but data on, for example, mean or median age for occupations at the four-digit level are available. Data on age distribution in the categories, under 40, 40–54, and 55 and over, for occupations at the four-digit level is available in another table. For those interested in pay distribution by level of education, statistics are available but not by gender or occupational classification.

Overall, in the Norwegian statistics, users can experiment with combining various statistical variables in the same table to compare women’s and men’s pay in occupations of equal value at the four-digit level, but multiple tables need to be combined to compare women’s and men’s pay in occupations in which work of equal value is performed in relation to age categories or education level.

Statistics Norway reports statistics and summaries relating to gender equality under the sub-area Population. The statistical database contains two tables on indicators for gender equality in municipalities in Norway. These tables do not include pay as an indicator of gender equality (but do include average gross income). These tables therefore do not provide data that can help to measure pay differentials between women and men in work of equal value.

The Swedish statistical database is based on five sectors: the central government, municipalities, regions, private sector salaried workers and private sector waged workers. One table combines four-digit occupations, the number of men and women employed and pay, as well as women’s pay as a percentage of men’s pay and regional pay, both within the labour market as a whole and by sector (either all five sectors or grouped into public and private sectors). Another table presents occupations at the four-digit level, the number of men and women employed, age in 10-year intervals and pay, and women’s pay as a percentage of men’s pay, both within the labour market as a whole and by sector. In a third table, the age variable is replaced by level of education, divided into seven levels. There is also a table that presents, in addition to the pure pay gap, the standard weighted gap (taking into account age, education, hours worked and sector) for each occupation individually.

The database includes the most recent table or tables compiled from previous years (from 2005 or 2014), allowing the user to see data from a specific year or all available data in a single table, for which the year can be selected. This means that historical developments are easy to trace.

Pay can only be linked to the variables of gender and part-time or full-time work for private sector salaried workers and only at the one-digit level; the number of hours worked is not specified for part-time workers.

There are often many empty cells in the Swedish tables, even when the data is apparently available in the database. For example, the number of employees is missing for many occupations in several pay tables (even when monthly pay is provided), despite this data being available in another table, at least for the year 2022. Thus, obtaining the gender distribution and pay for an occupation requires two tables.

The homepage for Statistics Sweden includes a link to ‘Pay statistics’, via which users can search for an occupational title and be directed to pay statistics with information on gender, sector, age and education. This entry point is useful for comparing pay for equal work, but not for comparing work of equal value.

Statistics Denmark (dst.dk) provides national statistics across nine different areas, with statistics on pay available under Labour and Income. On its website it provides 20 tables under the sub-heading Pay. For example, in one table (LONS20) users can obtain pay statistics by job function (at the one- to four-digit level based on the DISC-08 occupational classification), sector (overall, for the public sector overall or broken down by, for example, state or local government, enterprises and organisations/private sector), type of pay (hourly pay, fixed pay, total), employee group (total or divided into different categories), pay components (25 different variants based on pay increments or statistical measures), gender for full-time equivalent (total, women, men) and year (2013–2022). Three of the tables can be found under the heading Ligestillingsindikator, løngab (Gender equality indicator, pay gap) and allow the user to obtain data on annual national pay gaps from 2004 to 2022, pay gaps by occupational classification (1–4 digits according to DISC-08) and age in the years 2010–2022, and pay gaps by full- or part-time work and age in the years 2009–2022.

What makes it difficult to compare pay for work of equal value in the Danish context is the lack of data on gender distribution (number of women and men) in different occupations. This information is needed to know which occupations are male and female dominated so that pay gaps between these occupations can be analysed, rather than within the same occupation.

Gender equality is presented as a separate section on the Danish Statistical Authority’s website. Here it is easy to compare the size of the pay gap and its development between different groups, but data on pay itself is missing.

The link to the database can be found on the Statistics Iceland website, where gender-disaggregated pay statistics can be found under Society: Wages and income.

A large part of the labour market is not included in the Icelandic pay tables: the tables are based on about 100,000 employees, while the number of employees in Iceland in 2022 was about 200,000 (Statice.is). For example, the tables only include employers with more than 10 employees. The statistics are also somewhat skewed, as the tables from the database show that there are more women employed than men, while the proportion of women in the total labour force was slightly below 50%.

All the tables, except the one that presents occupation, gender and pay, identify occupations at the one-digit level only, making it difficult to identify occupations in which work of equal value is performed. At the one-digit level, gender-disaggregated tables on pay by sector, pay by industry, pay distribution by sector, pay distribution by industry and the gender pay gap are available.

According to Statistics Iceland, there is an Icelandic occupational classification, Ístarf21, that is based on ISCO-08. However, the pay tables in the statistical database still use the Ístarf95 classification. Ístarf21 breaks down occupations in much finer detail, but so far the advantages are not capitalised on. Ístarf95 includes 9 main groups based on ISCO-95, with 20 groups at the two-digit level and 52 groups at the three-digit level. However, several of the groups at the three-digit level include no data in the tables from the statistical database. This is also the case for the 156 four-digit codes. Some of these indicate the type of work and not the area of work as sub-codes within the three-digit code (general employees – skilled craft workers – skilled craft foremen), which should make it easier to assess whether occupations between two codes are comparable. Statistics Iceland also adds a fifth digit to some four-digit codes for the same purpose. In addition, Statistics Iceland has created some particular codes by, for example, combining codes that cover workers in the fishing industry (combining occupational with industry classifications) or combining supervisors from several different four-digit labour areas. In the latter case, several comparable occupational classifications have been combined, which makes the pay gap within the category (20%) interesting. Thus, the classification used includes a number of elements that facilitate pay comparisons between jobs in which work of equal value is performed.

Due to the small labour market in Iceland, all tables have several empty cells, i.e. some combinations of two or more variables (e.g. occupation, pay and part-time work) include no individuals, or so few that the table does not give the result.

The main table, which includes variables for gender (number of men and women) and different types of pay (basic pay, regular pay including fixed increments, all pay including bonuses) in mean, median and quartile terms for different occupations at the four-digit level, covers only full-time employees. No sectoral breakdown is provided.

The main table, which thus covers only about half of employees, includes more women than men and presents a different pattern than in the other countries: here the majority of occupations (at the three-digit level) are not male dominated, but gender equal (42%), and more occupations appear to be female dominated than male dominated (30% versus 27%). Moreover, in several of the occupations, especially in area 7, usually male-dominated craft occupations, there are so few women that the gender breakdown is not reported for men or women, and male dominance therefore cannot be easily verified in this particular material. However, actual pay between occupations can be compared. While it is not possible to report pay for either gender if the number of individuals is small, for understandable reasons, it should be possible to obtain the actual gender distribution.

In the sectoral table, occupations are only given in 11 categories. It also includes part-time workers and distinguishes between full-time and part-time workers – but the occupational classification is too coarse for the table to be used to compare men’s and women’s pay for work of equal value.

In another table, the rough occupational classification can be combined with industry. This can be used to identify differences in pay within the same occupational area depending on the industry, for example whether there are differences in pay within the group of clerks (group 4, clerical support workers) because their specific occupation is in a male- or female-dominated industry.

All tables cover the years 2014–2023, so comparisons over time are easy to make.

In the Gender Pay Gap section, there are a number of variables that show the development of pay and the pay gap: different sectors, full-time and part-time work, age groups, fields of activity and occupational classification across nine different occupational areas. However, the variables in the table cannot be combined and, as the occupational classification is very coarse, it is not really possible to compare occupations in which work of equal value is performed.

The table below summarises the main differences between the national databases. The information that can be obtained at the four-digit level can generally also be obtained at the more general levels, therefore additional notes are only provided at the three-, two- and one-digit levels if additional information is provided. The table is also otherwise highly general and serves mainly to illustrate how different data on pay and gender can be specified and presented with regard to occupation.

To summarise, there is varied scope for extracting variables that can be considered relevant in the comparison of the pay situations of occupations in which work of equal value is performed, and in most cases some information cannot be obtained. At present, the Norwegian database offers the greatest scope.

The different databases treat part-time work in different ways – the Finnish statistics omit part-time work entirely in some tables and in Denmark part-time work is converted into full-time equivalent; only the Norwegian statistics provide information on full- and part-time work as well as contracted weekly hours at the four-digit level. Age data in relation to gender and pay at this level is available in Norway, Sweden and Denmark, however in Denmark it is only provided in terms of the gender pay gap within the occupation, which is not useful when comparing different occupations. In Norway, the average age for different occupations is available at the four-digit level. Data on education levels and pay in relation to occupations is only available in Sweden, as rough categorisation. There is no indicator for migrant background as a selection option in any of the tables.

Much of what is missing from the publicly available databases can be customised by the Agency, subject to costs for compiling additional tables, meaning that the information exists, even if it is not publicly available.