10.7 Labour statistics and labour accounts#
10.7.1 Conceptual framework#
The labour statistics domain covers a wide range of labour-related topics. There is no single overarching internationally accepted framework for labour statistics or accounts like the SNA. However, there is a suite of internationally adopted statistical standards, guidelines and related indicator frameworks that can be referenced by those seeking to develop a comprehensive system of labour statistics. For those interested in compiling a labour account, there are national examples.
The International Labour Organization (ILO) is the UN international agency with the mandate to develop statistical standards in the domain of labour. It has fulfilled this purpose since its foundation in 1919 through the International Conference of Labour Statisticians (ICLS). The ICLS is a forum where countries discuss and adopt common definitions and agree on guidelines and standards (🔗) that subsequently act as a reference point for national and international labour statistics.
Reflecting the wide scope of labour statistics, many topics have been covered by resolutions and guidelines over time. Some of the main ones are:
employed persons and jobs;
unemployment and labour underutilization;
hours of work;
income from employment and labour costs;
organization of the labour market (i.e., statistics on collective labour agreements, industrial disputes and trade-union memberships);
health and safety at work;
informal employment and the informal sector;
child labour;
labour migration; and
key classifications including those relating to occupation and status in employment.
The ICLS has also endorsed the Decent Work Indicator Framework (🔗) as an overarching set of indicators covering ten substantive elements corresponding to the four strategic pillars of the ILO’s Decent Work Agenda (full and productive employment, rights at work, social protection and the promotion of social dialogue). Along with the range of resolutions and guidelines adopted by the ICLS, this indicator framework gives a sense of the wide scope of the labour statistics domain. Among other things, it enables a range of analytical approaches that can best be referred to as labour accounting, given that different countries use the term labour accounts to refer to different types of analysis.
10.7.2 Key resolutions and standards of the ICLS#
Among the many resolutions and standards which have been adopted through the ICLS, a number of resolutions are particularly key in setting a framework for labour statistics and relevant to labour accounts.
19th ICLS Resolution I: Resolution concerning statistics on work, employment and labour underutilization
At the 19th ICLS in 2013, a landmark set of standards concerning statistics on work, employment and labour underutilization was agreed (🔗). They are the basis for all statistics building on the concept of employment (which is the focus of labour accounting) although they extend beyond employment. The main features are:
The first statistical definition of the concept of work and a forms of work framework which identifies multiple forms of paid and unpaid work, distinguished by the beneficiary of the production, the receipt of payment (or not) and the purpose of the work (to gain workplace experience, to produce goods for household consumption etc.). Importantly, the framework recognises that people may engage in different working activities in the same period of time, enabling statistics to be produced on participation and time spent in different forms of work, as well as highlighting total work burden.
An updated definition of employment as work done in exchange for pay or profit. This is narrower than the widely applied definition adopted at the 13th ICLS that referenced all activities done to produce goods or services within the SNA production boundary. The 13th ICLS conceptually included some forms of unpaid work, such as subsistence farming within the scope of employment. The new framework identifies different forms of work for separate measurement and reporting, more closely aligning with policy needs and creating a more comprehensive framework.
Definitions of different components of labour underutilization to complement unemployment, namely time-related underemployment and the potential labour force (comprising people without employment that fulfil some but not all of the criteria of unemployment). Definitions are also provided for four different labour underutilization indicators, of which the unemployment rate is one.
The definitions from the 19th ICLS should be used as the reference point for analysis centred on employment or labour underutilization, such as labour accounting, as described below.
18th ICLS Resolution i: Resolution concerning the measurement of working time
These standards concerning the measurement of working time (🔗) define several different concepts of working time which can be used for different analytical purposes, such as the estimation of total hours worked or full-time equivalents (FTEs), which can be important in a labour accounting context. The working time concepts defined are:
Hours actually worked;
Hours paid for;
Normal hours of work;
Contractual hours of work;
Hours usually worked;
Overtime hours of work.
Among these concepts, hours actually worked, and hours usually worked are the ones most extensively used in measurement, and hours actually worked is the typical reference point for calculations of total hours worked or FTEs.
16th ICLS Resolution II: Resolution concerning the measurement of employment-related income
The most recent statistical standards covering employment-related income (🔗) were adopted at the 16th ICLS in 1998. The standards include a definition of employment-related income and different components of it. The standards distinguish between income related to paid employment and income related to self-employment recognising that the type of income earned can differ substantially between these two groups.
10.7.3 Other standards and guidelines#
A variety of other standards may be relevant to labour accounts depending on the analytical approach chosen. For example, a labour account may be aggregated using different classifications such as the international classification of status in employment (ICSE-93 or ICSE-18) or international standard classification of occupation (ISCO-08).
The latest international classification of status in employment (ICSE-18) was adopted at the 20th ICLS in 2018, replacing ICSE-93. The new classification is more detailed than ICSE-93 and provides improved analytical potential, building from 10 mutually exclusive categories at the lowest level of the hierarchy up to higher-level groupings. Another new feature of ICSE-18 is that it contains two sub-classifications built on the same ten detailed groups.
ICSE-18-R has a hierarchy that is aggregated based on the level of exposure to the economic risk of the worker in relation to their job/business, and at the top level distinguishes between workers in employment for pay and workers in employment for profit.
ICSE-18-A, by contrast, is aggregated based on the level of authority of the worker, distinguishing at the top level between independent workers and dependent workers.
In implementation, careful attention needs to be paid to the concepts, classifications and practices used to ensure the appropriate use of the data for labour accounts or any other analytical purpose. This is particularly important where, as is generally the case, multiple sources of data are used. Sources may differ in various ways, either conceptually or in measurement approach. For example, the reference point (employment concept) for a measure of working time used should align with the employment concept used in a measure of employment or any other measures used in the accounting framework. Likewise, attention should be paid to the classifications; for example, one source may use ICSE-93, while another uses ICSE-18, necessitating some adjustments during analysis. All such issues should be carefully considered in planning and undertaking the analysis.
10.7.4 Labour accounts#
The objective of labour accounts is to provide an integrated conceptual framework for producing a coherent and consistent set of aggregate labour market statistics. A labour account is designed to complement the existing suite of labour statistics to provide a logical framework for obtaining internally consistent estimates of key labour market variables. These key variables can assist users in making sense of seemingly inconsistent labour-related data, which are often based on different reference periods, populations, concepts, definitions and methodologies drawn from a variety of business and household surveys and other administrative sources.
A labour account provides a framework to bring together labour market data from multiple statistical sources into a coherent and consistent set of labour statistics. It is a complement to the existing suite of labour statistics. Users should continue to use the Labour Force Survey (LFS) for headline estimates of employment, unemployment and persons not in the labour force.
While there are currently no international standards for the production of a labour account, the ILO has documented two compilation approaches and a four-step process, which has been followed, to varying degrees, by the NSOs in Australia, Denmark, the Netherlands and Switzerland in compiling their own labour accounts.
A Step Forward to a Coherent and Timely Description of the Labour Market (🔗) describes the four-step process and the two compilation approaches, which are:
a cross-sectional approach involving confrontation and reconciliation of key labour market measures; and
a longitudinal approach, which incorporates changes to population and labour force via births, deaths, and net migration, and includes measures such as duration of employment.
A labour account provides a logical framework within which to bring together labour market data from a variety of business and household surveys and other administrative sources. The inputs from these sources and the derived statistics need to be harmonized, adjusted and integrated. Decisions that are taken in the production of a labour account within the national context need to be explicitly documented and published. The integration process can be viewed as a four-step procedure.
In the first step, the model and the identity equations are defined. The identity equations can be derived from the model. These identities may be seen as the most important part of the labour accounts, not only from the point of view of users of statistics, who will be presented with consistent data but even more from the producer’s point of view. These identities enable quality checks of sources. For example, in the context of the working time accounts covering registered paid employment, the document cites three main relational equations for employees:
jobs = employed persons - employed persons on leave + secondary jobs;
total hours worked = number of jobs * actual hours per job;
total compensation = number of jobs * compensation per job.
The second step involves the harmonization of definitions and classifications in source statistics and achievement of full coverage. The latter requires decisions to be made regarding which source is to be the primary source for each variable. Data are cut to size or expanded according to newly defined populations and definitions. After harmonization and achievement of full coverage, the aggregate values may still disagree in practice, but the findings from source statistics are aligned on shared parameters.
The third step is the minimization of measurement errors. It is highly likely that the harmonized data from the second step do not satisfy the definitional equations. By confrontation within identity relations, errors, irregularities and improbabilities are traced and corrected.
The fourth step involves balancing. Very small differences are neglected in the third step. These are smoothed out in the fourth and final step of the integration process. Here a mathematical procedure can be used by which the adjustments are minimized under the condition that the values attached meet the identity equations.
10.7.5 Australian Labour Account: Concepts, Sources and Methods, Oct 2019#
The Australian Labour Account (🔗) focuses on the cross-sectional approach and provides time-series data spanning 25 years. It has been designed to align with the 2008 SNA, as applied in the Australian System of National Accounts (ASNA), in particular, with the ASNA production and residency boundaries. This ensures direct compatibility with national accounts and productivity estimates, as well as providing a mechanism for bringing together conceptually related aggregate data from business, household and administrative sources. It is also consistent with the guidelines and standards established by the ILO.
Fig. 12 Australian Labour Account identity relationships – Jobs, Persons, Volume and Payments#
The scope of the Australian Labour Account is consistent with that of the national economy, as defined in the ASNA, which follows the 2008 SNA international standard. The framework incorporates four distinct quadrants: Jobs, Persons, Labour Volume and Labour Payments and covers all types of employment, including employees, self-employed and contributing family workers. The four quadrants are linked by a set of identity relationships, which the aggregate statistics must satisfy. These identities are shown in Figure 12. Some relationships are direct, such as the number of employed persons in the total economy is equal to the number of main jobs, while other relationships are considered indirect or derived, such that the relationship is based on an average or ratio measures such as average hours worked per job, or average labour income per employed person.
The Jobs Quadrant provides data on numbers of filled jobs derived separately from business and household sources, plus data on vacant jobs to provide a total number of jobs in the economy.
The Persons Quadrant includes data on numbers of employed persons, together with data on numbers of unemployed and underemployed persons (derived from household survey sources).
The Labour Volume Quadrant provides data on hours paid for (derived from business survey sources) and hours worked (from household survey sources), plus data on additional hours of work sought by unemployed and underemployed persons (from household survey sources).
The Labour Payments Quadrant provides data on labour income and employment costs (from business survey sources).
The Australian Labour Account combines data from the persons, jobs, labour volume and labour payments tables to calculate average hours worked, average remuneration (per person and per job), and average labour cost per hour worked. The data tables are compiled using interpolation, extrapolation, backcasting or benchmarking methods. The methods chosen are based on two main factors: the context in which the data were originally collected, and ability to fill data gaps between collection points or periods.
Various data sources (published and unpublished) are used in compiling the four quadrants. In general, the same data sources are used to compile both quarterly and annual labour account estimates. Quarterly survey estimates are benchmarked to annual survey estimates where possible. The method used to annualise data varies for each quadrant, depending on whether data are stock or flow estimates. The Australian Labour Account data at an industry level is derived where possible from data classified by industry reported in both business and household surveys. Where data at an industry level is not reported in surveys, the industry detail has been modelled using alternative sources.
Outputs from the Australian Labour Account consist of a number of spreadsheets and data sets, produced for both quarterly and annual data. Data are also produced for both balanced and unbalanced estimates. Quarterly data are produced at the industry division level, and the annual data are available at the industry subdivision level. Data are presented quarterly for the four quadrants for the balanced data for original (unadjusted), seasonally adjusted and trend estimates. For the unbalanced data outputs, data are presented quarterly for the four quadrants in original terms only.
The Australian Labour Account data provides an overarching picture of the Australian labour market over the past 25 years. This rich data source can be used for industry analysis of labour growth and performance in terms of people, jobs, hours, labour costs and income to better understand how the Australian labour market has evolved over time. In its Statement of Monetary Policy, February 2019 (🔗) the Reserve Bank of Australia made the following remarks regarding the Australian Labour Account:
“While the LFS will remain the best and most timely indicator of overall labour market developments (including employment growth and the unemployment rate), the ABS suggests that the Labour Account is likely to become the primary source of industry employment information.”
“The Labour Account’s industry employment numbers have a number of potential advantages over those in the LFS. The industry classification in the Labour Account aligns better with the measurement of industry value-added in the national accounts, and so the Labour Account should contribute to better estimates of industry productivity growth.”
“The Labour Account also takes a more comprehensive approach to estimate the amount of labour employed in each industry, for instance, by including the employment of non-resident visa holders.”
10.7.6 Further development#
There are no immediate plans to develop international standards which could be adopted at the ICLS and no standing international working group addressing the topic in a dedicated manner. This may change at some future date depending on the evolution of country demands.