EUROPEAN COMMISSION

Brussels, 28.7.2023

SWD(2023) 265 final

COMMISSION STAFF WORKING DOCUMENT

IMPACT ASSESSMENT REPORT

Accompanying the document

Proposal for a Regulation of the European Parliament and of the Council

on European labour market statistics on businesses, repealing Council Regulation (EC) No 530/1999 and Regulations (EC) No 450/2003 and (EC) No 453/2008 of the European Parliament and of the Council

{COM(2023) 459 final} - {SEC(2023) 295 final} - {SWD(2023) 266 final}

Table of contents

1.Introduction: Political and legal context

2.Problem definition

3.Why should the EU act?

4.Objectives: What is to be achieved?

5.What are the available policy options?

6.What are the impacts of the policy options?

7.How do the options compare?

8.Preferred option

9.How will actual impacts be monitored and evaluated?

Annex 1: Procedural information

1.Lead DG, Decide Planning/CWP references

2.Organisation and timing

3.Consultation of the RSB

4.Evidence, sources and quality

Annex 2: Stakeholder consultation (Synopsis report)

1.Consultation strategy

2.Methodology and tools used to process the results

3.Stakeholder participation

4.Main results from the stakeholder consultation

5.Conclusions

Annex 3: Who is affected and how?

1.Practical implications of the initiative

2.Relevant sustainable development goals

Annex 4: Analytical methods

Annex 5: Assumptions used for the distribution of costs (one off and recurrent) to the BR typology

Glossary

1.Introduction: Political and legal context

Eurostat and the European Statistical System

Eurostat, the statistical office of the European Union (EU), ensures the production of high quality, comparable European statistics 1 according to the legal framework Regulation (EC) No 223/2009 2 (the “Statistical law”) and the statistical principles, notably those laid down in the European statistics Code of Practice 3 . These rules and principles aim to ensure, among others, the independence, impartiality, objectivity and reliability of European statistics, and through those objectives, public trust in the statistics. The main uses of European statistics are to serve EU policy design, implementation and monitoring, and their main users are EU Institutions.

The European Statistical System (ESS) is the partnership between Eurostat and the national statistical institutes (NSIs), as well as other national authorities responsible for the development, production and publication of European statistics in each Member State. This partnership also includes the European Economic Area (EEA) and European Free Trade Association (EFTA) countries, i.e. Iceland, Liechtenstein, Norway and Switzerland. Eurostat issues statistical regulations and methodological guides, organises expert groups, and assesses the quality of statistics and Member States’ legal compliance. In accordance with the EU principles of subsidiarity and proportionality, each ESS member develops a statistical system suitable to their individual institutional context, while still following the common rules.

Eurostat’s activities are further influenced by overarching policies such as the EU’s Better Regulation agenda 4 , which promotes open and transparent EU decision-making and evidence-based decisions, and European Commission President Ursula von der Leyen’s six political priorities, among them “An economy that works for people”. Implementing, monitoring and assessing these policies and priorities requires impartial and objective data – that is, official statistics.

Labour Market Statistics on Businesses (LMB)

Labour market statistics on businesses (LMB) are official statistics that describe businesses functioning in relation to labour markets. The areas covered by LMB mostly relate to level and composition of labour costs, distribution and structure of earnings, labour cost index, and job vacancy statistics. First statistics related to the level and structure of labour costs had already been collected since 1959 5 , with a periodicity of two to four years based on specific legislation for each data collection and covering different sectors of the economy (industry, wholesale and retail distribution, road transport, banking and insurance, services). Council Regulation (EC) No 530/1999 introduced systematic data collections on the level and composition of labour costs (labour cost survey, LCS) for the calendar year 2000 and at four-yearly intervals thereafter. The same act established the statistics on the structure and distribution of earnings (structure of earnings survey, SES) for the calendar year 2002 and for a representative month in that year, and at four-yearly intervals thereafter. Before Regulation (EC) No 450/2003 of the European Parliament and of the Council of 27 February 2003 on labour cost index (LCI) was adopted, the data had been collected on a voluntary basis since 1996. Similarly, Regulation (EC) No 453/2008 of the European Parliament established for the first time and regulated the data collection on job vacancies (JVS). Earlier on, the job vacancies data used to be collected on a voluntary basis, since 2001.

LMB statistics are relevant to policymaking in various areas including macroeconomic convergence, social cohesion, price stability and gender equality. They are the reference source for all key indicators on labour demand (job vacancy statistics), on labour costs structure and levels across countries and distribution of earnings. The labour cost index supports the assessment of the inflation/deflation risks within the euro area conducted by the ECB; job vacancy statistics show labour shortages, structure of earnings survey data offer evidence to analyse wage determinants. The latter also allows calculating/adjusting the gender pay gap which is an indicator for gender equality listed under the Sustainable Development Goals scoreboard (SDG 2030) and the European Pillar of Social Rights (EPSR). Both the labour cost index and the job vacancy are in the list of the Principal Economic Indicators for the Euro monitoring.

High quality and internationally comparable LMB statistics covering all Member States 6 are essential to support EU evidence-based policy making and a range of other initiatives by the EU institutions, Member States, and other statistical users (e.g., trade unions, non-governmental organisations, research institutes, the media, etc.). LMB statistics are used in many publications of Commission services, for instance the annual and quarterly Employment and Social Developments in Europe (ESDE) reports or the annual review of Labour Market and Wage Developments in Europe 7 . They are also very important for monetary policy and used by the ECB in the preparation of the Governing Council 8 meetings, many Economic Bulletins 9 and analyses published occasionally 10 . 

As LMB statistics play a pivotal role in policy-making, the fast-changing social and policy landscape is demanding further improvements of LMB statistics in terms of flexibility to address new needs of users and compilers, better data timeliness, comparability, consistency and quality reporting. Megatrends such as automation and shocks to the economy as those caused by the financial crisis of 2008 or the COVID-19 pandemic show that researchers and policy makers need accurate and timely information on the functioning of labour markets, in particular LMB statistics. The COVID-19 pandemic led to decreases in the number of hours worked and vacant posts and to changes in the labour cost structure due to labour subsidies. Some groups of workers have benefitted from improved employment conditions, including increased earnings, while others have seen their employment prospects and conditions decline. This also had an impact on labour shortages and the relative earnings levels of employees.

LMB indicators are used to monitor important European policies such as the EPSR 11 or the SDG 12 . In particular, the gender pay gap (GPG) is an SDG indicator under Goal 5 “gender equality” and is part of the EPSR dashboard, under the ‘Fair working conditions’ (Principle 6: “wages”).

Over time, the policy context for LMB has steadily evolved. In 2009, the Commission released a Communication to the European Parliament and the Council giving its vision for the production method of EU statistics for the decade (2010-2020) 13 . This document has explored the key changes in the business environment of the European Statistical System, and the implications for policymaking. It highlighted the importance of having an integrated system that allows countries to gather data from different sources, increasing the availability and reach of the analysis. It also highlighted the importance of increasing the quality of the data, considering that many external sources do not match the expected requirements for European statistics. In 2014, the Commission (Eurostat) began the process of modernising social statistics. This led to the adoption of the framework regulation for European statistics relating to persons and households, based on data at individual level collected from samples of persons and households 14 . In parallel, on 27 November 2019, the European Parliament and the Council adopted the Regulation on European Business Statistics (EBS), repealing 10 legal acts in this field. The LMB is at the intersection between both domains, belonging to social statistics in terms of topic covered and to business statistics according to the type of respondents (enterprises). This initiative on LMB should be seen as completing the modernisation round of social statistics.

Current issues with LMB statistics

As recognised by the evaluation, LMB statistics allow collecting high quality information that are widely used for the intended purposes. The strengths of LMB statistics lie in their coherence, efficiency, comparability and the fact that they are well established, reliable and widely used by organisations and policy makers at all levels. Moreover, no alternative data at the EU or wider levels exist.

However, some shortcomings that were already identified at the time of adoption of the legal acts were differently approached (by requiring feasibility studies that lead to amendment of the legislation, by making available financial support for creating the capabilities needed to give effect to the legal base), and in the case of some data collections not sufficiently resolved. They mainly concern the partial coverage of LMB data on earnings which limits their use for policy purposes.

For several EU Member States, structure of earnings survey and/or labour cost survey do not cover enterprises with one to nine employees or part of the public sector (NACE Rev. 2 section O: ‘Public administration, defence, compulsory social security’). This creates biases in key statistics such as the average earnings and labour costs and therefore, for the gender pay gap as well (see Table 1).

Table 0.Gender pay gap by size class of the enterprise, 2018

Source: Eurostat, calculation based on the SES 2018

Those limitations became more visible further to the recent policy developments in the areas of equality, skills, working conditions and monetary union monitoring to which the data at the EU level is essential and provides clear added value.

For instance, the monitoring of the new directive 15 on adequate minimum wages requires unbiased estimates of the median wages as a (country-specific) benchmark. This requires that Structure of Earnings data cover all segments of the economy, including enterprises with less than 9 employees and the public sector, whose importance varies across EU countries. 

In other respects, the lack of a legal obligation to provide information on gender pay gap (GPG) became more problematic as the current data collection is voluntary and does not cover all EU countries nor all GPG variables. This brings a risk factor for the monitoring of principle 2 (‘Gender equality’) of the European Pillar of Social Rights.

The frequency of SES and LCS is not sufficient anymore for a number of users and the timeliness of the LCI, SES and LCS need to be improved.

Moreover, some stakeholders consulted in the public consultation and interviews, expressed needs for more breakdowns (e.g., ISCO or NUTS) or additional variables, allowing a better analysis of wage determinants including for the gender pay gap. In the interviews, they shared the view that it would be beneficial to assess innovative data collection methods (e.g., web scraping for job vacancies), that have the potential to decrease the burden on data providers, provided they ensure sufficient quality.

The evaluation has shown that those shortcomings reduce the relevance, effectiveness, coherence and comparability of statistics across Member States, with negative impacts for decision-making based on them.

2.Problem definition

2.1.What is/are the problems?

The evaluation identified issues that could be structured around three main problems.

Problem 1: LMB statistics does not address new needs, sources and methods.

The current legal framework does not cover new and emerging users’ needs. For example, policy users expressed a growing interest in data on the skills of the employees (the occupation data collected at the detailed levels of the ISCO classification can be used as a proxy for skills) and their net earnings (e.g. to monitor the purchasing power of low wage and of minimum wage earners). LMB also miss a quarterly index of hours worked that would have allowed monitoring the impact of COVID-19 and other possible crises on the volume of work. Job vacancy statistics by region and occupation are not available either, except for a few EU countries. These were deemed crucial to build more comparable vacancy data, estimate needs and assess job shortages at the regional level.

At the moment, the data collection framework is not flexible enough to allow for adjusting new variables or the timeliness of statistics. There is also growing interest in non-discrimination and gender equality, but the collection of gender pay gap data is currently collected voluntarily and not delivered by all Member States. This results in data that are not as precise and comparable as possible or required by some data users. Moreover, data on the earnings of people with disabilities are not collected, even those based on a national (non-harmonised) concept of disability. This fails in supporting users in making evidence-based policies and to follow the requirements of Article 8 TFEU and the UN Convention on the Rights of Persons with Disabilities.

Users are also seeking timelier data, for some LMB statistics, namely more frequent data collections and releases (note that the frequency of collection and release currently varies by type of LMB statistics) and/or that statistics become available faster.

Another emerging trend not fully considered in the current LMB framework is the use of administrative data. Several Member States are making increasing uses of this type of source. This facilitates more frequent data collection (e.g., quarterly). The current legal framework is not, however, able to accommodate appropriate quality reporting for this type of data across Member States.

Therefore, some user needs are likely to remain unmet, representing a missed opportunity to maximise the impact of LMB statistics across stakeholders and especially policy makers and researchers making use of this data. The provision of data closer to users’ needs, coordinated across the Member States, would provide policymakers with the opportunity to make more rapid and reliable assessments and respond quicker to changing issues/needs.

Problem 2: LMB statistics do not capture important actors or phenomena of the EU economy

As the current data collection on the gender pay gap is voluntary, some Member States do not provide the data, or only partially so that estimates are not fully available for comparisons across countries and time. This hampers users to have a full overview of gender pay disparities across EU countries and act in order to promote equality between men and women as requested by article 10 TFEU, SDG 5 and the second principle of the European Pillar of Social Rights.

Interviewed stakeholders representing users stressed that extending the coverage of LMB statistics to micro enterprises would also be important, at least for the main variables of SES, and to the public sector as no coverage of these impacts the proportion of minimum wage earners and low-wage workers. This would also correct the existing biases in key figures such as the mean and median earnings which are used in many policy indicators such as the minimum wage to median earnings ratio (‘Kaitz index’). Covering only part of the economy also distorts derived indicators such as the gender pay gap. It prevents users from analysing the adequacy of wages in the public sector and assessing the corresponding wage policies of EU governments.

Similarly, adding more variables in structure of earnings survey would enrich the scope and depth of analysis and modelling based on this data source. This would allow refining further the existing decompositions of the unadjusted gender pay gap which face limitations in possibility for disaggregation (not accounting for wage penalties related to family situation, career breaks, etc.).

Problem 3: LMB statistics are not aligned with related statistical domains

The evaluation results show that the current legal framework can be further improved to align with other EU policies and legislations as those that are new or evolved since the LMB legislation adoption. For instance, more pressing needs have emerged with the adoption of the new directive on minimum wage adequacy 16 . It becomes even more important than before to collect and publish mean and median earnings for the whole economy, as a benchmark for minimum wage levels. It is also useful to develop concepts such as the net wages received by employees after deduction of all social contributions.

The same applies to recent developments on gender equality, for which the gender pay gap has become a prominent indicator and should be possibly adjusted for all social contributions and more precise estimates of the skills levels.

Due to the time elapsed since the adoption of the current LMB legislation, the latter is no longer aligned on the latest versions of the related legal texts. This concerns in particular the latest edition of the European System of Accounts (ESA 2010 17 ) which defines the main concepts (e.g. notions of employees, labour costs, wages) used in labour statistics. It is also relevant for European Business Statistics 18 whose general concepts have been updated with the adoption of the revised framework regulation and implementing acts.

A few stakeholders highlighted some issues of consistency in terminologies used across Member States. This was mentioned for example about labour costs, namely, the elements included capturing social security contribution. Similarly, other stakeholders mentioned some inconsistencies in collective agreement across different countries.

Coherence between different data collections is another dimension of quality as it is fundamental to address multi-dimensional issues. The fragmented legal framework poses an immediate risk of the lack of coherence and comparability between data collections.

Stakeholders affected

The persistence of the problems affects users of the LMB statistics who do not have the evidence that meets their demands and those of policy makers that miss it for an efficient decision-making process. Ultimately, EU citizens could be considered not to be able to fully capitalise on high quality research and efficient policies.

The next section explains more in detail these problems and assesses how significant they are in terms of scale and impact. It also examines the main drivers leading to the problems and reflects on the expected evolution of the problems in the absence of targeted action.

2.2.What are the problem drivers?

Each of the problems has specific causes, also known as drivers. The main causes of the problems identified so far are grouped under each problem. That is, each problem has specific drivers.

For each problem, the list of all relevant drivers is detailed in the next paragraphs.

Problem 1: LMB statistics do not address new needs, sources and methods

D1 – LMB statistics do not cover emerging needs

Over the time, policy data users alerted Eurostat that there are particular variables or breakdowns missing in the LMB that would allow them better analysis and policy monitoring. In that context, also during the public consultation several improvements needed were confirmed, for instance collecting and publishing the numerator (labour costs) and the denominator (index of hours worked) of the LCI ratio for several NACE aggregates was found crucial or highly important by 13 data users out of 35 replying to the public consultation, another 15 considered this improvement of medium importance and only 7 as low. For 5 out of 18 data producers this was unimportant, other 9 producers gave medium importance and 4 high or crucial importance. Similarly, quarterly data on the ‘total number of hours worked’ was needed by both types of stakeholders: data users (29 replies out of 36, of which 13 scored this improvement as crucial or highly important and 16 of medium importance) and producers (only 7 out of 18 gave it a low importance). Those needs were confirmed in the interviews with EU policy makers (e.g. DG EMPL; DG ECFIN). For lack of this information, it was not possible to monitor the impact of the COVID-19 crisis on the volume of work. This will also be relevant for future crises.

The results of interviews with main stakeholders (in particular European policy users) confirmed that there is a set of unmet needs related to missing variables that prevent from fully adequate analysis. Regarding SES, in the interviews, policy users, in particular DG JUST, DG EMPL, expressed interest in intersectional analyses that would build on sex, age, possibly disability, etc. as far as feasible in a harmonised fashion. This would allow a better analysis of wage determinants and the refining existing adjustment of the gender pay gap. Indeed, the gender pay gap is very difficult to interpret without adjustment for the structure of the male versus female labour force in terms of occupations, economic activities, seniority, family situation, etc. Refined adjustments would therefore contribute to a better benchmarking between Member States and to the monitoring of the Gender Equality Strategy, European Pillar of Social Rights, the Sustainable Development Goals (gender pay gap is an indicator for gender equality listed under the  EU SDG indicators set  Goal 5: ‘Gender Equality’). Collecting SES information on net earnings and detailed occupation would allow users to better monitor the promotion of equality enshrined in the Charter of Fundamental Rights. Getting data on vacancies and occupied posts (JVS) by occupation at the 2nd digit level of the ISCO classification 19  was considered crucial or highly important, in the public consultation, by 18 data users out of 31 replying the question, other 8 considered it as medium important and only 5 as unimportant. Out of 17 data producers, 7 put low importance to having JVS data by occupation. The JVS data by NUTS 2 region were found crucial or highly important by 14 out of 31 data users and 8 found this improvement as medium important. Among data producers replying to the question only 3 out of 16 found it unimportant. The JVS data by NUTS 2 region were found crucial in the interviews with DG REGIO. This data would allow identifying regional mismatches and measuring labour shortages for different skills. They would also be needed to check the quality of non-official vacancy data (online job vacancies).

Yet, the LMB statistics legal framework does not facilitate incorporating new elements on an ad-hoc approach. As underlined by the policy users at several occasions (e.g. meeting of the working groups, hearings), the new legal basis should as much as possible include the variables or breakdowns identified as ‘emerging needs’ and be open for future updates if the situation requires.

D2 – The quality framework of LMB statistics is not fit for the use of administrative data and innovative sources

In the past two decades, an increasing number of Member States has switched to administrative sources for the compilation of some LMB datasets, as least for some variables. In the case of the LCI, the use of administrative sources is already the standard in most countries; labour cost variables are generally taken from social security data whereas the number of hours worked is still collected via surveys. This is also the case for annual gender pay gap data which are compiled from the four-yearly SES data for the years when survey data are available and which are extrapolated with social security sources in some countries. In the latest structure of earnings survey (SES 2018), 10 out of 27 Member States reported using a combination of survey data and administrative sources. Administrative sources are less common for job vacancies domain as data from public employment services tend to overestimate the real levels, some filled vacancies being still recorded in the register of vacant post.

Without incorporating this type of data into the framework of LMB, there will be an untapped potential for adding variables and breakdowns as well as reducing the cost and burden. However, increasing the use of administrative data and potential innovative data sources requires an updated quality reporting framework for documenting those sources, assessing their quality and communicating to the users.

Administrative data have not been primarily collected for statistical purposes and they only partly match the concepts to be measured. In addition, non-statistical sources may have other impacts such as the lack of timeliness (due to the administrative process before their availability). Hence, using these sources, requires specific quality assessments (e.g. in quality reports) such as on the biases introduced by the original purposes for which these data are collected and measuring the extent and impact of the biases introduced in the quality of the statistics based on them. Specific quality reporting on administrative or hybrid data sources does not exist yet for LMB statistics. The current reporting is focused on the survey collection and the quality is mainly measured through the statistical variance and on possible deviations of the statistical sample from the target population of enterprises or employees.

Similarly, the experimental work on use of innovative data sources (for instance information on online job advertisements available on the internet) showed so far that the web-scrapped data (data extracted from the internet) provide results that differ in levels from the official statistical information collected by the members states directly from the enterprises but are following roughly the same time trends. Online job advertisements do not cover all job offers but they can complement the existing statistics with further comprehensive, detailed, and timely insights into labour market trends, allowing the early identification of new emerging jobs and skills. Being a potential important source for the future, the quality of these data collected from the web should be correctly assessed. Public consultation showed that for 18 users of the data adding a description of administrative sources/registers to the SES quality report was crucial or highly important, for 13 importance was medium and for 5 low. In total 36 users replied. Only 2 out of 20 data producers considered it unimportant.

Similarly for LCS, 13 data users out of 35 were stated that adding description of administrative sources was crucial or highly important, 16 that it was moderately important and only 6 that it was not important. Only 2 data producers out of 18 who replied considered it unimportant.

In case of LCI, only 3 out of 36 data users and 3 out of 18 data producers were not interested in information on administrative sources in the quality reporting.

For JVS, 9 data users (out of 31 who replied) and 1 data producers (out of 16) considered adding information on administrative sources in the quality reporting unimportant. For 11 data producers this was crucial of highly important and for other 11 moderately important.

D3 – Some LMB statistics are not timely or frequent enough

Data that are published with a long lag after the reference period lose their relevance and usability, in particular when the observed phenomenon quickly varies over time. Expectations are also higher and, although LMB data generally comply with the deadlines, they are not considered timely or frequent enough nowadays, at least by some users.

Stakeholders expressed a view that a higher frequency of data and quicker updates of some LMB datasets are needed. In the public consultation, the timeliness and the collection frequency were two of the less positively assessed elements of the statistical quality for SES: 30 % of respondents to the public consultation were not satisfied with the frequency of the survey (18 replies out of 59) whereas 51 % expressed satisfaction (30 out of 59) and 19 % being neutral (11 replies out of 59). For LCS dissatisfaction on the frequency reached 24 % (14 replies out of 58), with 48 % expressing satisfaction (28 replies out of 58), and 28 % being neutral (16 replies). Satisfaction with frequency was much higher for GPG, JVS and LCI with dissatisfaction only amounting to 2 % for each of these.

The results of the public consultation also showed that collecting SES data every 2 years instead of every 4 years, was considered crucial or highly important by 20 out of 36 data users replying this question against 7 out of 18 producers of LMB statistics. For 10 data users this improvement was of medium importance and for 10 data producers low.

At the same time, collecting LCS data every 2 years instead of every 4 years was considered crucial or highly important by 15 data users out of 33 who replied, for 10 of them it has medium and for 8 low importance. 12 out of 18 producers considered doubling frequency of LCS of low importance.

More frequent SES would provide more relevant figures for derived indicators such as the mean and median earnings. Some users (e.g. trade unions), during the stakeholders workshop, underlined this need that became crucial in time of quick changes (e.g., high inflation) to see if wages are adapted to the inflation. In addition increased SES frequency would allow more timely analysis of pay gap determinants.

Similarly to frequency, the timeliness of the SES and LCS data was also not satisfactory for 32 % of the respondents to the public consultation (19 replies), 39 % of them were satisfied (23 replies) and 29 % (17 replies) did not have an opinion.

Figure 0.Public consultation respondents’ satisfaction with data timeliness across datasets

Source: Public consultation

Among the data users, almost the same number was satisfied (14) as dissatisfied (13) with SES timeliness. For the LCS, 10 users were dissatisfied with timeliness against 12 satisfied.

Satisfaction with timeliness was much higher for JVS and LCI with dissatisfaction amounting to only 6 negative replies for JVS and 1 for LCI.

Improving timeliness was considered one of the three most supported priorities for the SES and LCS in the public consultation. Improving timeliness of SES (by 5 months) was considered crucial or highly important by 22 data users out of 36 who replied the question, 10 considered it medium important and only 4 as not important. Similarly, 15 data producers out of 18 who replied thought the timeliness improvement is either highly or medium important (9 and 6 replies respectively) and only 3 that it is low. When it comes to the LCS, shortening the time of data publication by 6 months was unimportant only to 4 out of 35 data users and 4 out of 18 data producers.

The interviews with EU policy makers (DG JUST, DG EMPL, DG ECFIN) confirmed those findings that were further strengthened by the employment organisations and trade unions during the stakeholders’ workshop. In particular, due to the COVID-19 pandemic and the current rise in inflation, the timeliness of data has become even more important to stakeholders. With the fast-paced changes in the labour market, stakeholders highlighted the need for more timely statistics to inform policy making. No alternative data sources exist at the EU level, therefore not addressing this issue will put in question correct and timely monitoring of EU policies LMB is used for.

Problem 2: LMB statistics do not capture important actors or phenomena of the EU economy

D4 – The public sector and micro enterprises are not captured by LMB statistics

The coverage of LMB statistics was a shortcoming identified already at the time of adoption of the legal acts. Those were differently approached (by requiring feasibility studies that lead to amendment of the legislation, by making available financial support for creating the capabilities needed to give effect to the legal base) and, in the case of some data collections, not sufficiently resolved. While there are only three Members States with difficulties in providing JVS data for the whole economy, more than half of them did not progress over time in gathering information on labour costs and structure of earnings for enterprises employing less than 10 persons or for NACE Rev. 2 section O (public administration, defence, compulsory social security), in the SES and LCS.

The economic activities required to be covered by the SES in the legislation are NACE sections B to S excluding O. Although most of the countries go beyond this requirement and provide information for section O, the lack of data from some of them prevents from calculating the EU aggregates. The above-mentioned shortcoming of the coverage of the SES influence also the comparisons of the earnings of men and women. Moreover, the section O covers enterprises that are female dominated.

Following the difference described above, the gender pay gap calculated for the whole economy (sections B to S) differs from the one in the aggregate not covering section O (sections B to S except O). The lack of the data for section O in some countries does not allow full comparisons across the EU.

Table 1.Gender pay gap by economic activity of enterprise, 2018

Source Eurostat, [earn_gr_gpgr2]

Extension of the coverage of the SES to micro firms was perceived as crucial or highly important by 22 out of 36 of data users replying the question, for 8 of them this improvement had low importance and for 6 medium. Out of 18 data producers, only one was considering this change unimportant.

Extension to NACE Rev. 2 section O was considered unimportant by only 5 data users and one data producer (out of 36 and 18 of them respectively). For 21 data users this change was crucial or highly important and for 10 the importance was medium. 10 data producers considered this change as crucial of highly important and 7 gave it a medium importance.

For the LCS, the level of satisfaction with the coverage was higher for size classes than for economic activities for both data users and providers.

Figure 1.Public consultation respondents’ satisfaction with size classes covered by LCS (all enterprises except those with less than 10 employees)

Source: Public consultation

Figure 2.Public consultation respondents’ satisfaction with economic activities covered by LCS

Source: Public consultation

Few countries provide LCS data for micro enterprises. Not covering all sizes of enterprises leads to biased results for the average labour costs which are generally lower in microenterprises compared with enterprises employing 10 persons and more, as illustrated in Figure 4 for the countries for which this information is available.

Figure 3.Total labour costs 20  in Euro, per employee in full-time equivalents, per hour, by size of enterprise, 2016

Source : Eurostat, [lc_ncost_r2]

Since the adoption of the LMB legislation and following the policy developments (e.g. introduction of the European Pillar of Social Rights), the data became more prominent and widely used, therefore requiring necessary improvements in quality.

For the structure of earnings survey reference year 2006, 11 Member States provided data on enterprises employing from 1 to 9 employees 21 . The situation did not improve significantly in the following waves of the survey (2010, 2014) and only 12 Member States (and 2 EFTA countries) transmitted those data in the latest data collection available for reference year 2018 22 . Similarly to the structure of earnings survey, only 11 Member States provided data for enterprises with 1 to 9 employees in the last labour costs survey (2016) 23 .  

Section O is covered on voluntarily basis in the structure of earning survey by 23 Members States, in labour cost survey by 22 and by 24 Members States in the JVS 24 .

This illustrates that voluntarily extension of the surveys coverage did not provide satisfactory outcome.

D5 – GPG data are provided on a voluntary basis

Data on the gender pay gap are provided by Member States on a voluntary basis leading to a partial coverage of countries (Ireland and Greece do not transmit) and statistical variables thus hampering cross-country comparisons. 

More than half of LMB data producers responding to the public consultation would find it crucial or of high importance to create a legal basis for the mandatory transmission of annual adjusted GPG data. Interestingly, only one data producer found this unimportant. Out of 37 data users replying the question, only 8 found that a legal basis on GPG is not needed. In the interviews, the data users (e.g. DG EMPL, DG JUST) offered detailed examples on how the data is used and stressed it´s importance as well as the problems caused by its voluntary status (e.g. incomplete coverage). Moreover, some miss certain determinants of the pay gap (e.g. number of children, work spell, total experience) to conduct research on the gender pay gap based on available statistics. Another problem they highlighted was that the gender pay gap statistics go through sizeable revisions, when SES becomes available. Finally, the proposal for directive on pay transparency 25  requires a sustainable production of annual GPG.

Figure 4.For you, how important would it be to add the following measures to the GPG data collection process: Creating a legal basis for the mandatory transmission of annual unadjusted GPG data?

Source: Public consultation

Problem 3: LMB statistics are not fully consistent with related statistical domains

D6 – Lack of alignments with definitions, concepts and approaches in related statistical domains

The evaluation found that the current legal framework for LMB statistics is internally coherent although this could be improved further by simplifying the legal architecture (replacing the three framework regulations currently in force by a consolidated text). This unified legal framework would benefit from systematic references to the corresponding concepts used in closely related domains such as the National Accounts and European Business Statistics. The legislation has been updated in both domains and their methodology revised (ESA 2010, EBS regulation) since the LMB legislation has been adopted. Aligning LMB is needed with a view to foster coherence across domains and clarity to the users of LMB legislation and the statistical offices of the Member States.

Moreover, the exclusion of “public administrations and defence, compulsory social security” (NACE Rev. 2 section O) from SES, leads also to a deviation from the scope used in national accounts. Civil servants and other government employees are covered in national accounts data but not always in LMB statistics.

The legal framework can also be further improved to align with other EU policies and legislations as those that are new or evolved since the LMB legislation adoption. This problem has rather limited impact on stakeholders and would be resolved by the above policy measures related to the extension of the scope of the LMB and updated referencing. Therefore, no specific policy measures addressing the alignment with other policies and legislations are needed.

2.3.Size of the problem(s)

Different problems or weaknesses identified in the evaluation of the legislation have different levels of severity that can be assessed based on the consequences of the identified shortcomings. Those mainly depend on the importance of the policy concerned and of the type of institutional user. The most important problems of the current legal framework is that it does not reflect the latest evolution of user needs and is not flexible enough to address those that will emerge in the future. This includes the fact that the gender pay gap data collection is provided on voluntary basis and therefore is not transmitted by all member states.

The timeliness of the data defined in the legislation currently in force is not sufficient anymore for the most important policy users (Problem 1). For instance, the labour cost index, could be provided to the Governing Council of the European Central Bank one month earlier than now, allowing quicker reactions to inflation/deflation risks stemming from by wages. This type of analysis and the timeliness of the LCI data have become even more crucial since inflation has regained momentum in EU economies. Timeliness and frequency of data are closely related. The data collected frequently to monitor phenomena changing quickly in time should be published almost immediately to be relevant. The data collected with lower frequency usually do not describe phenomena that are volatile but should be available with short delay to be useful for users who wait for it longer (in between the data collections). The policy users, in various occasions (expert groups meetings, hearings between the Commission services) often underlined the need for more timely and frequent LMB data. For instance, DG JUST, would need SES microdata more frequently than every four years in order to conduct detailed analyses of gender pay gap determinants.

Additional variables such as for instance career and family situation in the SES would also enhance the analysis of the gender pay gap. Similarly, information on net earnings would help good monitoring of the wages adequacy (problem 1). The indicators defined based on the LMB data are used to monitor policy initiatives that gained importance over time. For instance, the broader use of median wages data (coming from the structure of earning survey, SES) for policy monitoring, in the context of the directive on minimum wages, requires comparable data across EU countries whatever the share of SMEs in the economy or the relative size of the public versus private sectors (Problem 2). As illustrated in the evaluation report, the mean earning differs depending on the size of enterprises (Figures 1 and 9, evaluation SWD). Similarly, the current gender pay gap indicator is not provided for the whole economy as it is based on the same SES source (as illustrated in Table 1).

Moreover, the annual gender pay gap is provided on a voluntary basis only and in consequence not by all member states (Greece and Ireland do not provide data) and is on a potential risk to being discontinued.

Hence, the implementation and monitoring of the two new and very important directives on adequate minimum wages and (forthcoming) on pay transparency would not be accurate with the current LMB statistics.

2.4.How likely is the problem to persist?

LMB statistics are the only data source on the structure of earnings, labour costs and labour demand at EU level. No alternative data sources exist. Hence, without any changes, over a 10-year horizon, users’ needs are likely to remain increasingly unmet, missing an opportunity for evidence-based policy and good quality research. Considering the importance of LMB statistics for policies that generate impacts (e.g., inflation, gender pay gap, labour market performance, convergence), there is room for improving the LMB evidence provided for optimal decision-making process.

For instance, the partial coverage of EU economies leads to biases in key figures used for EU policies such as the job vacancy rate or the mean and median earnings. Without a revision of the current legislation, this problem will persist as observed in the past years despite Eurostat’s repetitive warnings (e.g. through the three-yearly report to EP/Council on job vacancy statistics).

Likewise, strategic policies related to equal rights will remain sub-optimal until the annual GPG data collection is made mandatory and SES data are complemented with additional wage determinants such as detailed occupations; new variables such as net earnings and (possibly) new information on discriminated and/or segregated sub-populations such as persons with disabilities.

In economic terms, users will continue to bear additional costs and burdens to search for data that are not available at the EU level.

The problem tree

Figure 6 below shows the links of the problems and their drivers as well as their main consequences.

Figure 5.Problem tree

3.Why should the EU act?

3.1.Legal basis

The legal basis for EU intervention in the area of European statistics is Article 338 of the Treaty on the Functioning of the European Union (TFEU), which confers on the European Parliament and the Council the power to adopt measures for the production of statistics where necessary for the performance of the activities of the Union, in accordance with the ordinary legislative procedure. Article 338(2) TFEU mentions six general principles that Union statistics must follow, namely impartiality, reliability, objectivity, scientific independence, cost-effectiveness, and statistical confidentiality while not entailing excessive burdens on economic operators. These principles are defined in Article 2(a)–(f) of Regulation (EC) No. 223/2009 on European Statistics and further detailed in the European statistics Code of Practice adopted according to Article 11 of the regulation. Article 338(2) TFEU also stipulates that Union statistics shall not entail excessive burdens on economic operators. This provision, together with that of Article 338(1) TFEU, according to which statistics shall only be produced where necessary, reflect an expression of the principle of proportionality.

Article 5 of the Treaty on European Union (TEU) enshrines three principles: conferral, proportionality and subsidiarity. In line with the principle of conferral, Article 338 TFEU empowers the Union to establish European statistics, so this is a clear specific objective to be pursued at EU level.

LMB statistics are part of the European statistics which are framed by Article 338 of the Treaty on the Functioning of the European Union (TFEU) as the legal basis. This is done following the ordinary legislative procedure in which case the European Parliament and the Council adopt measures to produce statistics for the Union to deliver on its role. Hence, according to Article 338, LMB statistics must conform to standards of impartiality, reliability, objectivity, scientific independence, cost-effectiveness and statistical confidentiality.

Finally, legislative action on LMB statistics falls under supporting competences, where the principle of subsidiarity authorises an intervention of the Union only if a specific issue cannot be addressed by the individual intervention of the Member States and provided the objective can be better achieved “by reason of the scale and effects of the proposed action” at EU level.

In this context, this section assesses whether a revision of the current legal framework governing European statistics on labour market on businesses is appropriate and justified, in view of its purpose to ensure high-quality data on the EU in line with the statistical principles and quality criteria applicable to European statistics.

3.2.Subsidiarity: Necessity of EU action

Currently, LMB data collections are undertaken under three separate Regulations that do not sufficiently address new needs; sources and methods neither cover important actors of the EU economy.

Voluntary data collections are appropriate instruments to pilot the production of new topics or characteristics, and to foster the incremental capability of national statistical systems to provide such new data. However, they tend to become inefficient over time as recurrent production costs eventually fail to generate substantial EU value added in terms of completeness and timeliness for all Member States.

The ESS provides an infrastructure for statistical information needed at the EU level. The system is designed to meet the needs of multiple users for decision-making in democratic societies. It complements the national needs assuring harmonised approaches and data comparability without duplicating the effort. The action does not go beyond the necessary activities of ESS partners while better ensuring the quality and comparability of statistics. Among the main criteria to be fulfilled by official statistics are consistency and comparability. Member States cannot achieve those criteria without a clear European framework, that is to say Union legislation laying down the common statistical concepts and transmission requirements.

In light of the variety of measures adopted at national level and given that providing a robust legal framework for the collection of LMB statistics is essential to maintain relevant and comparable statistics at EU level based on harmonised concepts and approaches to methodology, further action at EU level is justified. Moreover, while some problems are limited to some countries (e.g., coverage of JVS statistics incomplete in DK, FR, IT only), the general problems and requests from stakeholders apply to all Member States.

Moreover, the legislative action is necessary and proportionate as the problems identified can only be solved via the appropriate EU legislation and have a genuine EU-wide scope clearly linked to gaps of the current legislation (section 2). Without further EU legislative action, these problems will persist.

3.3.Subsidiarity: Added value of EU action

The added value of complete and comparable LMB statistics lies in their important input to feed into strategic EU policies (monetary and economic policies, Gender Equality Strategy, the European Pillar of Social Rights, Sustainable Development Goals). LMB statistics are also designed to meet the needs of multiple users, for the purpose of decision-making at all levels in the EU, as well as research and informing the general public.

For this purpose, statistics that are consistent and comparable across EU Member States are required. Coordinated action is essential to deliver coherent, reliable, timely and comparable data based on harmonised concepts. It is clear that without a European framework, Member States would not achieve the same results in terms of data, harmonised methodology and comparability of outputs and consistence, which can only be fully achieved by way of EU action.

Moreover, in order to ensure compliant provision of national data, a regulation is the most adapted type of EU action

Therefore, actions within this area will respect the principle of subsidiarity as framed by Article 5 of the TFEU.

4.Objectives: What is to be achieved?

Figure 7 sets out the main objective and specific objectives. There are no clear trade-offs between specific objectives. Each of them is quite distinct on a specific strand of actions to achieve the general objective.

Figure 0.Objective tree

Source: ICF

4.1.General objectives

The general objective of the action is to produce labour market and business statistics that are up-to-date, relevant, comprehensive in coverage of important economic actors, comparable across Member States and consistent with related statistical domains. It answers in the LMB domain the requirements of Article 338 TFEU for the production of statistics where necessary for the performance of the activities of the Union, in line with its principles of high quality and non-excessive burden.

4.2.Specific objectives

The general objective is made of 3 specific objectives:

·Adapting the regulatory framework to allow flexibility in meeting emerging needs, to release more timely statistics and to promote the use of innovative sources and  methods of duly assessed quality

·Improving the coverage of statistics to the whole economy and ensure exhaustive GPG data 

·Improving the consistency with related statistical domains 

Table 3 shows the connection between problem drivers and specific objectives. In other words, each driver could be seen as an obstacle to achieve the objectives. To the extent policy options would address drivers and problems, these would be helping to achieve the objectives.

Table 2.The problems and its drivers are obstacles to achieve the strategic and general objectives

Consistency with other EU policy objectives

As highlighted in the effectiveness part of the evaluation, the objectives of this intervention contribute to different European flagships:

·European Employment Strategy (2005/600/EC);

·Principle 6 of the European Pillar of Social Rights. 

·Directive on equal opportunities and equal treatment between men and women in employment and occupation (2006/54/EC)

· Directive on adequate minimum wages (EU) 2022/2041 ,

·Commission proposal for EP and Council Directive on pay transparency;

The policy options will also be assessed on these criteria.

5.What are the available policy options?

5.1.What is the baseline from which options are assessed?

In the baseline scenario, no action is taken. The current legal framework of LMB statistics divided into 3 different basic acts would continue to exist with the shortcomings described in the previous sections. In the baseline scenario there is no EU policy change and it does not require any particular measure or action. Hence, this is an extrapolation within the next 10 years of what would happen in a ‘no-change’ scenario. This scenario implies consequences for stakeholders as described in the problem definition, that is, economic, social and any other relevant impacts.

5.2.Description of the policy options

In order to set up policy options, an exhaustive list of all possible measures identified during preparatory work with NSIs, the Public Consultation and stakeholders’ consultation was set up for each data collection. In a second step, measures are bundled into regulatory frameworks or policy options and clustered by related policy drivers.

Policy measures

The policy measures are structured around the data collections to be regulated, amended or updated.

Job vacancies (JVS);

Labour cost index (LCI);

Labour cost survey (LCS);

Structure of earnings survey (SES); and

Gender pay gap (GPG).

The list of measures is presented in the table 4.

Table 3.List of measures by data collection

Policy options

We present in the following a list of options that correspond to different level of ambitions, but also costs and burden, for the LMB review subject to this impact assessment.

Options are defined as a set of individual measures, from those described in the previous part of the document, that would meet the user requirements identified in the evaluation and in the ‘problem definition’ part of this document.

In order to make grouping of the various measures into policy options, a first feasibility assessment of each individual topic and measure was carried out. This process involved assessing how the topics and measures relate to one another (mutually exclusive or complementary), and an initial screening of their likely effectiveness, as well as their technical, operational, economic and legal feasibility. Based on this assessment and in order to integrate the various policy measures in a meaningful set of policy options that would offer a choice to the policymaker, a realistic selection based on the hard or soft nature of the requirement was carried out. This builds on the distinction between required measures and voluntary measures on the one hand, and the extent to which measures were included at all in the option. Two categories of policy measures shall be tackled at the level of secondary legislation, namely those on improved quality reporting and on emerging needs for variables and breakdowns. Hence, in the description of policy options, they are grouped in only two sets of combined measures on quality and emerging needs respectively.

The policy options range from the least ambitious ('option 1’) – implementing most improvement measures but without legal enforcement (“soft” non-mandatory / voluntary measures) except for the annual GPG - to the most ambitious and costly (‘option 3’) - whereby all improvement measures identified in the evaluation would be implemented through legally binding provisions. Option 2 would enforce only a selected set of improvement measures. Options 1 to 3 are defined under one legislative act that will be an umbrella bringing together all current three LMB related primary legislations.

Finally, option 4 is the “conservative” option offering changes in every relevant framework regulation. It keeps the three exiting framework regulations and accommodates changes in their context when possible. It is further presented under part 5.2.4 but was discarded at an early stage as explained under part 5.3.

The baseline (option 0) and policy options 1 to 3 are detailed in Table 5 and in the following paragraphs. In this table the selection of measures is indicated according the following labels:

- Full regulatory approach with binding measures set in EU legislation are marked as “Full”

- Soft regulatory approach with non-binding measures set in EU legislation resulting with MS data transmitted on a voluntary basis is marked as “Soft”.

- Measures not included under this policy option are marked as “Not included”.

Table 4.Proposed policy options 0 to 3 clustered by drivers they address

5.2.1.Option 1: Reliance on the voluntary nature of contributions

In this option, the existing data collections and their legal obligations as foreseen in the three existing regulations are implemented under a new framework regulation. Moreover, the GPG, which is not covered by and does not fit in the scope of any of the 3 current framework regulations, is encompassed in this new framework regulation. The reason is the need of making GPG annual data collection compulsory (addressing driver five), aligning definitions, concepts and approaches as well as synchronising planning as set up by the sixth problem driver. Strong methodological links are necessary between SES and GPG (both dealing with earnings) and LCS and LCI (both dealing with labour costs).

However, selected new policy measures for improvement of the data collections requested by stakeholders will be implemented by voluntary data collections. This could be done using agreement between the NSIs or other national authorities and the Commission (Eurostat) as foreseen in the article 14 of the Regulation No 223/2009 28 by which members of the ESS agree to implementing new statistical developments / collections on a voluntary basis. It should be noted that increasing the frequency of the SES and LCS data collection on a voluntary basis was discarded as its high costs limit its implementation on a voluntary basis. Moreover, different periodicities of data collections (LCS and SES) among member states would lead to fragmentation of the statistical processing and dissemination. Only some policy measures (GPG 1 and All 1) would be enforced, in line with the approach of the new framework legal basis.

5.2.2.Option 2: High coverage, better timeliness and other measures

In this option all data collections and the annual GPG are implemented under a new framework regulation. The measures implemented will be compulsory. However, this option does not include the policy measures that have the highest impact in terms of burden and costs for businesses and NSIs. In particular the higher frequency for SES and LCS (every two-year instead of every four-year as currently) that would double the burden of these two surveys for businesses. Similarly, the coverage of micro businesses in LCS and some very ambitious improvements of timeliness are not implemented.

Namely, in this option the coverage will be extended to micro businesses for SES for the main information collected in the survey (SES 2) - while LCS will not be extended to these enterprises - measures associated to the fourth problem driver. Both, SES and LCS will be extended to cover NACE Rev. 2 section O (SES 1 and LCS 1) and JVS will cover full economy (including micro firms, public institutions and non-business economy, NACE Rev. 2 sections O to S)- JVS 1. Moreover, timeliness will be improved for the LCI for all countries (LCI 1) and larger countries will provide quick estimates allowing calculation the EU and euro area flash labour costs index twenty days earlier than the current date of T+70 days (LCI 2). For the SES, the timeliness will improve by two months (SES 3) and no improvement will take place for LCS (LCS 3 not implemented). Frequency of SES and LCS will not increase (SES 5 and LCS 4 not implemented) – measures associated to the third problem driver. All measures associated to the other policy drivers will be taken into account, including developing a legal basis for the GPG as under option 1 and covering emerging needs. Improved quality reporting will be assured for all data collections and better harmonisation and planning will be assured via one framework regulation. The objectives would be partially achieved: four out of six drivers would be fully answered and two only partially.

5.2.3.Option 3: Highest coverage and increased frequency

This option intends to implement, under a new framework regulation encompassing all data collections and the annual GPG, all and best policy measures listed above, including the high frequency measures for SES and LCS (every two-year data collection for each). This option would answer exhaustively the needs for improvement and harmonisation of all LMB statistics as expressed by stakeholders.

Namely, it includes all policy measures identified, for all actions answering all the 6 policy drivers, either in terms of coverage, timeliness and high frequency, quality reporting and use of administrative data and innovative sources, alignment of concepts across LMB domains, making GPG data collection compulsory and covering emerging needs. That means, the coverage of SES and LCS would be extended to micro firms and NACE Rev. 2 section O (SES 1, LCS 1, SES 4, LCS 2) and JVS to micro firms and NACE sections O to S (JVS 1). Timeliness of SES would improve to T+11 (SES 6) and LCS to T+15 (LCS 3). All measures associated to the other policy drivers will be taken into account, including developing a legal basis for the GPG as under option 1 and 2 and emerging needs. Improved quality reporting will be assured for all data collections. And better harmonisation and planning will be assured via one framework regulation.

Framing options 2 and 3 in the secondary legislation

The framework regulation will define domains and topics for the data collections. The technical elements including the list and description of variables and breakdowns will be specified in the secondary legislation.

The detailed variables (preliminarily identified in the evaluation) or breakdowns will be a subject for secondary legislation due to high level of details and granularity as well as their very technical character. Similarly, the framework regulation in options 2 and 3 will require Member States to provide quality reports on the sources and data collection methods, while their technical elements including the content of the quality report will be defined in secondary legislation. This approach allows for updating the contents of quality reports should new specific quality dimensions become relevant.

In addition, to assure flexibility in addressing future needs, some aspects (detailed topics, their corresponding periodicity and reference periods as well as transmission deadlines) will be set out in the annex of the framework regulation with delegated powers for the Commission to adapt them. This is the usual approach for statistical legislation, and it allows for updating technical elements to cater for future emerging needs.

5.2.4.Option 4: Implement actions in the existing legal frameworks

This option would implement policy measures by amending the existing three framework regulations. While it is possible to implement a large series of measures by amending Regulation (EC) 530/1999 for SES and LCS measures, Regulation (EC) 450/2003 for LCI and Regulation (EC) 453/2008 for JVS, there will be no legal framework for regulating the annual GPG nor better harmonisation and planning.

5.3.Options discarded at an early stage

The data collections covered in the LMB statistics are interlinked. They should use the same statistical definitions and be coordinated in timing to be fully effective. For example, indicators on labour cost levels are collected every four years (direct measurement) and updated every year by using the annual labour costs index as an auxiliary variable. As another case, the yearly gender pay gap is benchmarked every four years on the results of the structure of earnings survey. This requires definitions used for wages and other labour cost components to be identical in all these datasets in order to obtain unbiased yearly estimates. Updating each framework regulation separately would be a missed opportunity to best align the common concepts used in labour market statistics. Moreover, if the legal framework remains scattered, introducing any future new concepts into one existing regulation in future would require potential reopening of the other.

Moreover, the legal basis for the mandatory transmission of the annual unadjusted GPG data does not exist so far: this indicator and does not fit in the scope of any of the three existing legal frameworks. Hence, option 4 would not allow developing this legal basis that is an important request of stakeholders and is high in the policy agenda of the EU with the proposal of Directive on pay transparency, the EPSR and the SDG.

In addition, while a large set of policy measures could be implemented in each domain, e.g., the extension of coverage requested by stakeholders, high consistency and consolidation would not be ensured across the different domains and risks of divergences in national implementations would also increase. Indeed, the legal fragmentation of the 3 existing framework regulations makes it very difficult to achieve a coordinated and consistent approach when it comes to modifications.

In addition, there is a need for better coordination of the planning. While Regulation 530/1999 ensures an alternation of SES and LCS, and LCI and JVS regulations foresee quarterly transmissions to Eurostat, a full synchronised planning including GPG is needed. Links between LCI and LCS for providing annual estimates of labour costs based on indexes as well as between the LCS and SES (and SES and GPG) for coherence of statistical concepts and quality reports for all data collections allowing a smooth repartition of the burden for NSIs would not be ensured.

Finally, the undertaken initiative to modernise and streamline LMB statistics necessitates substantive amendments to the provisions of the legislative acts currently in force and granular approach dealing with every regulation separately will be administratively highly inefficient.

On the opposite, merging regulations into one integrated framework would foster clarity and consistency for all stakeholders, in particular the national statistical institutes, data providers, institutional and policy users, researchers, students and EU citizens in general. It is the best way to ensure that data producers use the same concepts in the same manner across all LMB datasets. It is also the simplest and most transparent way to align LMB concepts on the related definitions from European Business Statistics or National Accounts.

Moreover, the flexibility for introducing requirements that will be proved to be feasible (by for instance conducting pilot studies) will also be easier to manage if LMB statistics are all consolidated in one single basic act.

Finally, as LMB data collections are usually collected by the same bodies in the member states, the integrated framework would allow a better balancing and controlling of the burden on respondents over time, including for future new needs to be implemented in the context of the new flexibility.

Taking into account all these elements, option 4 has been discarded.

The features and differences of the options considered for the analysis of their impacts are summarised in Table 6.

Table 5.Options 0 (baseline) as well as 1, 2 and 3 under a single new framework regulation main features and differences

6.What are the impacts of the policy options?

The three policy options are assessed using a partial quantitative assessment of the efficiency of the policy options, with impacts monetised to the greatest extent possible, based on the available evidence.

Moreover, a qualitative assessment is carried out for each of the policy options covering the following criteria:

·Effectiveness;

·Coherence with the overarching objectives of EU policy and consistency with the proportionality principle;

·Efficiency;

·Impacts including on fundamental rights and wider social, environmental, digitalisation and on administrative burden of enterprises;

oBy its nature, statistics can only indirectly impact fundamental rights and contribute to social, economic developments by allowing monitoring them.

oAs the LMB statistics collects information on businesses it contributes to an administrative burden on them related to data reporting although limited.

oStatistical processing is a beneficiary and contributor to digital development.

For the qualitative assessment, a scoring approach is defined as indicated below.

Quantitative analysis of the options

Policy option 0 is the baseline scenario. The costs associated with this option are therefore an estimate of current costs, assuming these costs would remain constant over 10 years. The cost assessment of the baseline scenario has been carried out based on the evaluation results. It is important to note that the baseline costs estimated as part of this Impact Assessment are the costs associated with the current legal framework. The baseline costs estimated include all costs currently incurred to produce the various LMB datasets that could potentially be affected by the policy measures and options proposed. Datasets not regulated and voluntary are not included in the baseline costs.

The costs assessed for policy options 1, 2 and 3 are expressed in comparison with the baseline (policy option 0). These only estimate the additional costs associated with the implementation of each of the measures/actions required under each option, over and above the baseline situation, i.e. costs that would be incurred in any option. As such, total costs for each option are equal to the costs reported for policy option 0 plus those reported for each other relevant option.

The study supporting this impact assessment has been able to assess the minimum costs expected for a number of cost items, in particular:

ŸEC (Eurostat):

-one-off costs for amending the framework regulation and implementing regulations

-Recurring costs for monitoring compliance by MS.

ŸMSs/NSIs

-One-off costs for adapting the existing data collections to the new requirements under the revised LMB framework regulation

-Recurring costs for data collections

-One-off cost to communicate the change

-Recurring costs for reporting to the EU

There are other cost items and cost savings involved that the study could not ascertain, which include:

ŸThe costs of adapting IT systems to new regulations

ŸIndirect cost savings due to improved policymaking as a result of the changed framework regulation.

The assessment of the costs for businesses is based on a detailed study conducted by Eurostat in cooperation with the compilers (NSIs and other national authorities in charge of LMB).

Table 7 provides a high-level overview of the approximate costs. The assumptions are shown in Annex IV and V.

Table 6.Overview of costs, in Euro

Benefits such as potential cost savings or reductions in the administrative burden associated with each policy option could generally not be quantified. This depends in particular on future progress in IT and data collection techniques (such as ‘web-scraping’) which have not been implemented yet. These techniques are expected to decrease the LMB burden in the long run so that the estimates provided in this study should be taken as maximum amounts (caps). Estimates for the expected cost savings would be severely limited and hence add little value. In order to enable a choice to the policymaker between policy options 0 to 3, the available cost estimates and the qualitative assessments of the options are considered to be sufficient.

Many benefits on the availability, accessibility and use of LMB data sets have different effects on stakeholder groups (data users), making them challenging to define, quantify or monetise.

Analysis of the policy options

An overview of the qualitative assessment for the four options retained for the impact assessment has been carried out. It starts by an assessment of proportionality, followed by assessments of effectiveness, efficiency and coherence. A detailed scoring of the policy options is provided in the Section 2 of Annex 3.

6.1.Policy option 0 (baseline): No targeted EU action

The baseline passes the criterion on consistency with the proportionality principle, however on form of action (choice of instrument) it is uncertain whether this option can deliver satisfactory achievement of the objective and effective enforcement.

All effectiveness, efficiency and coherence criteria under policy option 0 score ‘0’ to enable meaningful comparisons with options 1, 2 and 3. For the baseline scenario there is no further EU action.

6.2.Policy option 1: Reliance on the voluntary nature of contributions

Figure 1.Intervention logic for option 1

Under policy option 1, there will be limited mandatory changes at the EU level and a larger number of voluntary changes. The measures, however, generally represent a continuation of the status quo under the baseline.

Policy option 1 passes the criterion on consistency with the proportionality principle, including choice of an instrument. This option scores “+” on effectiveness in all specific objectives. It scores “+” on impact on fundamental rights and social impact and is neutral “0” on impact on Sustainable Development Goals.

The results of the consultation on cost and burden with the NSIs and the further qualitative study showed that the introduction of a compulsory data collection on GPG would lead to an increase of costs of 34 % (+ 215 000 Euro) for this data collection. For the whole LMB, however, this means an increase of less than 1 % in total. The estimated increase in costs related to burden on enterprises would be negligible (less than 100 000 Euro). This is mainly linked to the fact that the GPG data collection is based on the existing sources and the burden is mainly on the statistical authorities to compile the data and prepare quality report.

This option would have limited effectiveness in addressing the problems identified with the current EU legal framework as, relying on national willingness, it would not ensure that improvements are implemented Union-wide. While it passes the proportionality principle simply because it is a simplification compared to the existing framework, its effects are modest and do not meet the expectations for a general review of LMB legislation. While there are some positive social impacts, the overall coherence with EU policies is not deemed well-served by this option, thus making it sub-optimal.

In summary, this option would be somewhat more effective in addressing the problems identified with the LMB framework compared to the baseline, although it still falls short from users expectations identified through the LMB evaluation exercise.

6.3.Policy option 2: High coverage, better timeliness and other measures

Figure 2.Intervention logic for option 2

As highlighted in the intervention logic, most of the drivers are addressed in option 2. It is therefore highly effective in achieving most of the policy objectives. In terms of coverage, option 2 does not foresee the extension of LCS to cover micro firms but extends SES to micro firms although only for main information (SES 2). Option 2 also offers better timeliness but the frequency remains unchanged – as with the baseline scenario (4-yearly). The option leaves limited gaps with the identified set of problems and drivers. While it is anticipated that the increase in timeliness will allow for better research and evidence from SES and LCI, the frequency of LMB statistics would not increase although this issue was often mentioned by stakeholders. Whereas cyclical indicators (JVS for labour shortages, LCI for hourly labour costs) are provided with sufficient frequency (quarterly), some users would appreciate a shorter periodicity for structural data such as the LCS and the SES. Nevertheless, option 2 would also cover important actors and segments of the population to allow for better research and policy analysis. This option also introduces new templates for the collection and quality reporting for the use of administrative data and innovative sources as well as more alignment in terms of concepts and definitions of relevant data domains including national accounts and the European business statistics.

The policy option 2 passes the criterion on consistency with the proportionality principle, including choice of an instrument. This option scores “++” on effectiveness related to specific objective 1.1 “Adapting the regulatory framework to allow flexibility in meeting emerging needs” and 1.3 “To promote the use of innovative sources and methods of duly assessed quality”. It scores however “+” on specific objective 1.2 “To release more timely statistics”. On specific objective 2 “Improving the coverage of statistics to the whole economy and ensure exhaustive GPG data” option 2 scores “+” and on objective 3 “Improving the consistency with related statistical domains” it is assessed to be“++”.

Impact on fundamental rights and on Sustainable Development Goals is positive “+” and social impact “++”. Indeed, in addition to the improvements related to the gender pay gap data, the extension of the SES coverage to the micro firms and NACE Rev. 2 section O would allow better assessment of the low–wage earners and contribute to the monitoring of the directive on minimum wages adequacy. Option 2 scores “+” on coherence as it would address most of the shortcomings of the baseline in terms of coverage of the economy and answering emerging needs and further harmonisation of standards (definitions, concepts, templates for quality reports). A clear effort on collecting these types of data would better ensure meeting overarching EU objectives.

The costs for compilers and time burden on businesses on statistical reporting related to data collections were assessed based on a quantitative survey of NSIs and Other National Agencies (ONAs) in charge of LMB statistics. This was complemented by a qualitative survey of the LAbour MArket Statistics (LAMAS) working group carried out in the July-August 2022 and the study conducted for the purpose of evaluation and impact assessment.

The Table 8 provides an overview of the increase in costs for NSIs across datasets for NSIs, for policy option 2

Table 7.Increase in costs of NSIs for policy option 2

Overall, the increase of costs for NSIs is approximately. 12 % (EUR 3.7 million) of which SES is the costliest with 1.5 million (20 %) followed by LCI with close to 0.9 million (9 %). Costs for mandatory data collection on GPG are marginal.

The table below provides an overview of the increase in costs for businesses across datasets.

Table 8.Increase in costs for businesses for policy option 2

Overall, the increase on businesses amounts to approximately EUR 4.7 million (11 %), of which EUR 724,000 (16 %) for JVS, EUR 12,000 (0.10 %) for LCI, EUR 457,000 (4 %) for LCS, EUR 3.43 million (21 %) for SES and EUR 92,000 for GPG.

To summarize, option 2 would create a limited increase in burden, mainly due to the extending SES to micro-firms. The costs implied (8.2 million) represent a negligible part (less than 1%) of the costs European Business Statistics (annual total costs of around 980 million Euro in 2017) 29 . The real amounts will be even smaller as ten NSIs have already extended SES to micro-firms. Therefore option 2 scores “++” on efficiency.

6.4.Policy option 3: Highest coverage and increased frequency

Figure 3.Intervention logic for policy option 3 

Option 3 fully captures all important business actors and allows for the monitoring of socio-economic phenomena across the whole EU economy. It leaves no meaningful gaps with the identified set of problems and drivers.

The extension of both the SES and LCS scope to micro-firms would also fill in the existing data gaps for this part of the economy. The addition of new data, improvements in timeliness, increased frequency of SES and LCS and alignment of definitions would result in more complete full filling users’ needs leading to better evidence-based policy making with issues such as discrimination and inequalities being better covered by European official statistics. All these improvements would significantly enhance the scope and depth of policy research on earnings, labour costs and labour shortages.

The policy option 3 passes the criterion on consistency with the proportionality principle, including choice of an instrument. However, the increased frequency and timeliness entail sustained further efforts that may, strictly speaking, not all be necessary to meet objectives of the initiative. Therefore the option is uncertain to pass on the criterion of costs being commensurate with the objective of the initiative.

This option scores “++” on effectiveness related to specific objective 1.1 “Adapting the regulatory framework to allow flexibility in meeting emerging needs”, objective 1.2 “To release more timely statistics” and 1.3 “To promote the use of innovative sources and methods of duly assessed quality”.

On specific objective 2 “Improving the coverage of statistics to whole economy and ensure exhaustive GPG data” option 3 scores “++” and on objective 3 “Improving the consistency with related statistical domains” it is assessed to be“++”.

Impact on fundamental rights, social impact and on Sustainable Development Goals is highly positive “++”.

Option 3 scores “+” on coherence as it would address most of the shortcomings of the baseline in terms of coverage of the economy and answering emerging needs and further harmonisation of standards (definitions, concepts, templates for quality reports). A clear effort at collecting these types of data would better ensure meeting overarching EU objectives.

The table below provides an overview of the increase in costs for NSI’s across datasets, for policy option 3

Table 9.Increase in costs for NSIs for policy option 3

The total cost of NSIs as reported in the Eurostat’s costs and burdens study suggests an overall increase of around + €19 million (+64%) of LMB statistics for policy option 3 as compared to the baseline. The higher increase would be for LCS 8.3 million (116 %), followed by SES (100%, +7.4. million).

The table below provides an overview of the increase in costs for businesses across datasets.

Table 10.Increase in costs for businesses for policy option 3

Similarly to NSIs, businesses would also face a significant increase in costs, In monetary terms, policy option 3 would increase costs for businesses of an estimated EUR 38 million, of which EUR 12,000 for LCI, EUR 2.7 million (61%) for JVS, EUR 15.2 million (142%) for LCS and EUR 20 million (125%) for SES. Overall this would mean 88% increase in costs for businesses compared to the baseline.

Option 3 would create much higher costs for NSIs and a significant increase in the burden for respondents. Therefore it scores “+” on efficiency.

Options 1, 2 and 3 are to the same extent likely to have a rather neutral indirect environmental impact when considering the European Green Deal and is not different from the baseline. There are no to negligible impacts on competitiveness, research and innovation and all three options are fully in line with the principle of “Do No Significant Harm”. All those options are also fully in line with the principle of digital-ready policymaking: digitalization is at the core of statistical processes within the European Statistical System (‘digital by default’). The specific measures will be implemented through digital techniques (web-forms, digital processes, IT programmes) with a view to minimize the costs for NSIs and burden on respondents for all options.

7.How do the options compare?

Table12 below summarises scoring for all options assessed.

Table 11.Summary assessment table

All options presented in the impact assessment are largely in line with the first judgement criteria on whether the option go beyond what is necessary to satisfactorily achieve the objectives because they are limited to what is necessary to satisfactorily achieve the objectives. They do not attempt to address any additional problems beyond those defined in the problem definition.

Moreover, policy options 1, 2 and 3 clearly pass the criterion the extent to which each policy option is limited to aspects that cannot be achieved satisfactorily by the Member States on their own. Ensuring consistency in the geographic, temporal and thematic breakdowns of LMB statistics at EU level requires EU intervention and cannot be left solely to the Member States (necessity and subsidiarity tests are passed).

The third criteria concerns the extent to which the costs (for the Union, national governments, regional or local authorities, economic operators or citizens) are commensurate with the objectives of the initiative. Based on the above assessment on cost and burden, all policy options are judged as having costs that are commensurate with the objectives of the initiative and passing the test. Even policy option 3, which would be the costliest option to implement, is judged as passing this test.

On the fourth criterion, as to whether the form of action (choice of instrument) is as simple as possible all options pass the test for the simple reason that a new framework regulation – required for all options – would be a clear simplification compared to the baseline. While the benefits for option 1 are cancelled out by its inherent inefficiencies, the form of action is still the simplest one possible.

Effectiveness

Option 1 would aim to increase effectiveness of the existing framework by considering elements that were not part of the existing regime. While these are voluntary in nature the mere consideration of these needs, and in responding to problems, this option compares favourably with the baseline. Consistency with related statistical domains would be improved somewhat, though the voluntary nature of the regime would greatly mitigate these potential benefits.

The impact on fundamental rights is not fully neutral. Given on the compulsory collection of the gender pay gap in particular the impact on fundamental rights of collecting these types of data will be partially served through this option.

Both option 2 and 3 would help to significantly improve the depth of statistics by further adapting the framework to evolving needs, sources and methods by considering numerous elements that are not part of the existing regime.

The coverage of statistics will be improved in particular for GPG data, extending SES to micro firms, better coverage of the economy through including NACE Rev. 2 section O “administration”, new elements of quality reporting allowing assessment of quality of innovative or administrative data sources, aligned definitions and concepts, and addressing emerging needs.

This together with improvements in LCI timeliness would further contribute to modernising social statistics.

In terms of social impacts there are positive effects in both option 2 and 3, with the mandatory collection of data on gender pay gap and on additional variables and breakdowns to be further detailed in the secondary legislation. The mandatory collection of information on employees (e.g. carer breaks) would expand the range of possible analyses on segregation, subject to legal constraints on privacy and data protection.

All three options have a rather neutral indirect environmental impact when considering the European Green Deal and not different from the baseline.

Overall option 3 compares favourably to option 1, and is more ambitious than option 2 when it comes to the frequency of SES and LCS. The timeliness and frequency improvements do match an expressed need of some users and thereby it makes the LMB more effective than under option 2.

Coherence

In terms of coherence both option 2 and 3 are expected to provide strong positive impact. On the other hand, option 1 would not address all the shortcomings of the baseline (option 0) in terms of the somewhat more complex nature of the various existing datasets, the divergence of outcomes due to the voluntary nature of many of the measures and thereby the rendering of coherence in EU policies somewhat less favourable than possible.

Both option 2 and 3 would address most of the shortcomings of the baseline in terms of coverage, answering emerging needs and further harmonisation of standards (definitions, concepts, templates for quality reporting). It is anticipated that these policy options would align LMB statistics with other EU flagship initiatives such as the EP and Council Directive on minimum wages, Commission proposal for a Directive on pay transparency and the principle 6 of the European Pillar of Social Rights.

Efficiency

All options would incur higher costs compared with the current LMB framework taken as the baseline. Costs in policy option 1 are mainly for NSIs and related to the GPG. Costs of option 2 (for Eurostat, NSIs and enterprises) are mainly driven by the costs of producing the statistics on wages and job vacancies on the extended population of enterprises, to a lesser extent on labour costs and the gender pay gap. In terms of impacts for respondents, option 2 would be far less burdensome (+ 11% increase in costs for businesses) than Option 3 (+88%).

Therefore, although option 3 presents the highest effectiveness, this comes at a high price and option 2 scores far better in terms of efficiency.

Potential obstacles that might be encountered for an effective implementation

The problem of the LMB statistics currently lies in non-legally binding collections (at EU level), particularly for micro firms. For instance, 12 member states (and 2 EFTA countries) transmitted those data in the latest SES 2018. Section O is covered in the SES by 24 member states and by 22 member states in LCS. Hence, the main impact in terms of extension of the coverage will be for the data collection of the SES on micro-firms, i.e., the effort will be for Belgium, Denmark Greece, France Croatia, Italy, Luxembourg, Hungary Malta, Austria, Poland, Portugal, Romania, Finland and Sweden. In job vacancy statistics, most member states already achieve a full coverage of economy. However, data for three of them (Denmark, France and Italy) are not fully comparable with those of the rest of the EU. In both France and Italy, public institutions are not covered within the sectors of public administration, education and human health (NACE Rev. 2, sections O, P and Q). Furthermore, France only includes units with 10 employees or more whereas Denmark only covers units within the business economy (NACE Rev. 2, sections B-N).

8.Preferred option

8.1.PREFERRED OPTION

Policy option 2 has been identified as the most effective option, which addresses the needs of data users and providers in the most cost-efficient way. In the online workshop with stakeholders this option was the preferred option. Although data users would like to see more ambitious proposals such as option 3, participants in the online workshop insisted that the selected option be realistic in terms of what can be achieved with Member States in the current context of cost and burden limitation and what could be possibly accepted for implementation by the EU statistical community. All options ensure proportionality. Although option 3 is the best performing option in terms of effectiveness, it is also the most expensive and causes the highest burden on enterprises. According to data providers, the implementation of option 3 would also raise serious difficulties on the field. The preferred option is therefore option 2.

8.2.REFIT (simplification and improved efficiency)

REFIT: Simplification and improved efficiency

Options 1, 2 and 3 would require a new framework regulation. This entails integrating the 3 existing framework regulations for LMB into one single basic act.

The favoured option (option 2), will lead to simplifications and improved efficiency through three main channels:

a) Merging the existing three framework regulations into one single legal act. A single framework regulation will clearly set the scope of LMB statistics and underline their common nature (collected from businesses) and purpose (to inform about developments in EU labour markets). Each concept will be defined only once and thus fostering consistency (also in future updates) and sparing users cumbersome cross-references. Moreover, at the member states level, a single and coherent framework regulation will facilitate use of the same tools for different purposes and limit consequently data collections and burden. Similarly, it will facilitate the use of administrative or innovative sources as the discussions with data owners will focus on the same concepts and finding the best source for them instead of having to repeat the same exercise with different concepts, building on different sources and delaying the decrease of the burden on businesses by using other / new sources.

b) Fostering the use of alternative administrative sources and modern digital techniques, including automatically transferring payroll data and web scrapping, which will play a role in mitigating the burden on enterprises in general and SMEs in particular. This may be achieved by extracting data on earnings and labour costs from payrolls and job advertisements from the internet. It is expected that new administrative or innovative data collections will also help covering emerging needs.

c) Limiting collection of variables on apprentices in the LCS only from those countries where they represent a sizeable share (more than 1%) of all employees;

The legislation will empower the Commission to adapt statistical requirements with implementing or delegated acts. Secondary legislation will cover technical details related to periodicities, reference periods, description of variables, modalities, statistical classifications and data breakdowns, precision targets, the metadata to be transmitted as well as on the content and the deadlines for transmission of the quality reports. These instruments are expected to be used for covering also future emerging needs.

Option 3 also entails the above simplifications but would extend the SES data collection to micro-enterprises for all variables. This is not the case for option 2 that limits this future data collection on micro-enterprises to a core set of key variables, most of which are available from administrative sources therefore simplifying as much as possible the extension of scope.

By contrast, option 1 may not lead to simplification and efficiency gains if different choices are made across LMB datasets, a clear risk due to the non-harmonised and uncoordinated nature of this option.

For the above reasons, only options 2 and 3 would pass the REFIT test as they would both improve the efficiency of evidence-based policymaking, including through enhanced possibilities for comparative research across member states. Nevertheless, option 2 would come first (++ against + for option 3) as it refrains from increasing the frequency of labour cost and structure of earnings surveys, thus bringing lower costs and larger efficiency gains.

At the same time, the additional measures added under the options, in particular options 2 and 3, can also require a slight additional need for follow-up, reporting (more frequent and timely data collection) and enforcement. Nevertheless, it is expected that some cost saving could be made by reducing the current administrative burdens borne by both the Commission (Eurostat) and NSIs.

The table below sets out a qualitative analysis of the possible simplifications brought about by this initiative. The stakeholder group benefiting from these savings is specified in the final column of the table.  

8.3.Application of the ‘one in, one out’ approach

No burdens on citizens will be caused by the initiative as it concern statistical collections on businesses.

For businesses as explained in the SME test, only a small subset of businesses are required to produce the datasets under LMB (sample surveys, i.e. only the very limited number of businesses selected in the sample), even under options 2 and 3.

While the specific information demands are higher for option 3, the total burden on enterprises remains negligible for option 1 and 2, higher for option 3.

8.4.SME test 

SMEs represent around 85% of new jobs in Europe and are hugely important to the EU economy.

The four steps of the SME Test are set out below, explaining the assessment carried out and estimated impacts on SMEs of the policy option 2.

SME Test: four steps

9.How will actual impacts be monitored and evaluated?

Monitoring and evaluation form an integral part of the policy cycle and must be undertaken through the implementation, application and review stages of the LMB initiative. The performance of the new LMB framework should be monitored and evaluated in relation to its operational objectives (shown in the table below). The next sections briefly describe the main considerations for monitoring and evaluation at each stage of the policy cycle.

Implementation stage

The choice of approach remains open, whether largely voluntary data collection under the LMB framework or different forms of mandatory changes to the LMB framework.

During the implementation phase of the new legal framework, the Commission (Eurostat) will continue organising regular expert group meetings with partner NSIs in the ESS to discuss and clarify any issues that may arise, continuing a long-standing history of good and close cooperation between Eurostat and its ESS partners on technical and statistical matters. This includes diligent joint preparation of key implementing acts regulating the detailed new statistical data and metadata requirement, which will be of central interest to both statistics users and producers. For instance, Eurostat can establish topical task forces to address specific technical challenges in achieving the objectives of the new legislation. The implementation stage is planned to conclude with a first evaluation focussing on the implementation, functioning and initial impacts of the new legal framework. To obtain sufficient information on the performance, this evaluation is planned within three to five years after the entry into force of the new legal framework, in line with better regulation advice that evaluations should have access to at least three full years of data.

A monitoring and evaluation clause is normally included in any new, or amended, legislative act to facilitate the collection of the necessary information.

Application stage

During the application stage, the Commission (Eurostat) plans to undertake an evaluation of the functioning and impact of the legislation every three to five years. The following table provides some suggestions on potential indicators for monitoring the effectiveness of the LMB initiative against its objectives. It also indicates some of the data sources that might be used.

It will be important to ensure that the indicators can be measured through methods and sources that are easily available and credible. Eurostat’s user satisfaction survey and more targeted surveys (e.g. among NSIs) could be relevant sources of evidence. Quality reports, annual compliance assessment exercises are also important tools for this purpose in the European Statistical System.

Possible key performance indicators, targets and data sources for operational objectives

Annex 1: Procedural information

1.Lead DG, Decide Planning/CWP references

2.Organisation and timing

After political validation of the LMB initiative in November 2021, an interservices steering group (ISG) chaired by Eurostat and composed of representatives of Commission DGs 31 was set up to supervise the progress on the combined evaluation and impact assessment including stakeholder consultations. The ISG met 5 times to discuss this evaluation SWD:

3.Consultation of the RSB

The RSB was consulted on this impact assessment on 18 January 2023.

The Board findings and recommendations were addressed in the impact assessment report as follows (text in italics):

Summary of findings

The Board notes the additional information provided and commitments to make changes to the draft report.

However, the report still contains significant shortcomings. The Board gives a positive opinion with reservations because it expects the DG to rectify the following aspects:

(1) The report does not clearly analyse the scale of the problem as well as its consequences for policy making. It does not adequately explain the rationale for intervention.

(2) The report does not sufficiently address the simplification potential of the preferred policy option. 2

What to improve

(1) The report should clarify the size of the problem. It should better explain the rationale for intervention, given that the evidence shows that most of the contacted stakeholders in the public consultation seem to be satisfied with the current situation. It should bring out more clearly any public interest case for new and more frequent data for EU policy areas.

   Specific section discussing on size of the problem (section 2.3) was added to the report.

The rationale for intervention was reinforced by refocusing the narrative from the analysis of satisfaction with the current LMB statistics to reflection of users related to the possible improvements. This is described mainly in the analysis of the problems drivers (section 2.2).

(2) The simplification potential of the preferred policy option should be better explained and elaborated. The report should explain how the replacement of the three existing Regulations with a single framework will foster simplification. It should identify and analyse any additional simplification aspects such as reducing the data requirements to the strict minimum and using data sources other than surveys.

Advantages of using a single legal act on LMB are discussed in sections 5.3 (in relation to the discarded option) and section 8.2 related to simplifications. Further simplifications are also described there. Details on using alternative data sources are provided in analysis of driver D2 in section 2.2 on problem drivers.

(3) The report should discuss potential problems with compliance and enforcement, where relevant. It should provide further information on the possible implementing or delegated acts of key importance for the impact or the success of the initiative.

Potential obstacles that might be encountered for an effective implementation are discussed in section 7 on comparison of the options.

Framing of options 2 and 3 in the secondary legislation is described in section 5.2 on description of policy options.

(4) The report should better explain the scale of provision of statistics on a voluntary basis by Member States and its consequences. In addition, it should provide further detail on how the sub-objectives on quality reporting, the use of administrative data and innovative sources by Member States and the response to emergent needs will be achieved.

The situation related to provision of data on voluntary basis is described in the description of current issues with LMB statistics. Biases in the data have been illustrated as some countries provide the data covering full economy voluntarily. More information about countries providing the data voluntarily was added in the analysis of the problem driver D4. The consequences and compilers opinion on voluntary character of GPG collection are described in the analysis of problem driver D5.

Use of administrative data and innovative sources as well as related quality reporting are described in the analysis of problem driver D2.

(5) The report should improve the presentation of stakeholder views. The consultation activities findings should be provided with the necessary granularity and should not be presented as aggregate percentages. The report should clarify to which consultation activities the results are referring to. The reader should also be able to understand what stakeholder category expressed a certain concern or opinion and to what extent this evidence is significant and robust.

Relevant graphical presentation and comments on stakeholder consultation was changed across the document, to cover more granular information on type of stakeholder and to add absolute figures.

(6) The report should better explain the problems with the evidence-base and data availability, its methodological choices as well as the limitations and the uncertainties of the consultation activities. It should explain how limited use of the public consultation results to assess possible impacts, in particular on respondents to the LMB statistics (enterprises), has been mitigated.

Issues related to data availability and mitigation of problems related to low coverage of enterprises in the results of consultation are further complemented in section 4 (Evidence, sources and quality) of Annex 1 on procedural information.

The Board notes the estimated costs and benefits of the preferred option in this initiative, as summarised in the attached quantification tables.

Some more technical comments have been sent directly to the author DG.

In order to limit the number of pages in the main text of impact assessment, the section on policy relevance of the LMB data was shortened; the detailed scoring of the policy options was moved to the Section 2 of the Annex 3. Scoring and narrative in the assessment of options were aligned.

4.Evidence, sources and quality

The LMB statistics are specific as for the type of information they provide and users they target, meaning a highly specialised domain with relatively scarce sources for evaluation. The stakeholders’ consultation (public consultation and targeted interviews as well as workshop with data users validating the policy options) was therefore extensively used to complement the first findings of the evaluation. To ensure the highest coverage of most important stakeholders, the public consultation was advertised through correspondence with most important data user (Including European Statistical Advisory Committee, Directors of Social Statistics, Business Statistics Directors Group, Labour Market Statistics Working Group, EUROFOUND, EIGE, ECB, DG EMPL CIVIC DIALOGUE, Employment Committee Indicators Group (EMCO IG) and Indicators' Sub-Group of the Social Protection Committee   (ISG SPC) of the Council, SME Envoy Network, LMB data collection correspondents, SES microdata users, European Economic and Social Committee labour market observatory, Business Europe, SMEunited, ETUC (European Trade Union Confederation), Platform of European Social NGOs and international organisations such as ILO, OECD and UNECE). In addition, different social media channels were used to catch general public. Unfortunately, very few replies from the general public were received. However, while not constituting a real survey outcome in the statistical sense (small number of replies, no probabilistic selection of the sample or respondents), the public consultation confirmed the needs accumulated over the years (for instance during the regular hearings between Eurostat and the Commission policy DGs). They were complemented by the targeted interviews and technical level correspondence between Eurostat and for instance the ECB. In addition, a careful analysis of Commission reports to the European Parliament and the Council on implementation of the labour costs index and job vacancy statistics, allowed to observe the developments in the implementation and compliance improvements of those data collections (evaluation SWD section 3.1). For its part, the systematic review of the quality reports that is a daily work in Eurostat allowed identification of main issues with the data. This way the information existing and/or collected in the consultation exercise together with the desk research analysing the legal LMB was triangulated to best evaluate the current initiative.

The call for evidence and the public consultation 34  were run via the Have Your Say website of the European Commission 35 . No feedback was received by the Commission (Eurostat) on the call for evidence. The public consultation was available in all official EU languages. These consultations aimed at gathering information and feedback from the general public and all relevant stakeholder groups as input into the evaluation and impact assessment as a step towards the preparation of a legislative proposal.

The public consultation was promoted via different channels: Have Your Say, Eurostat website a dedicated news item on 23 February 2022, social media campaign and Eurostat email banners, emails to stakeholders as well as nationally by several NSIs. Public Consultation (PC) questionnaire had a section assessing the LMB statistics in general and specific sections on each data collection that are a part of the LMB. The latter covered questions on satisfaction and attitude towards possible improvements or simplifications.

As NSI have been consulted by Eurostat several times throughout the review process it was agreed that their contributions would be most beneficial in an online survey format instead of semi-structure interviews. The questionnaire sent for the purpose of the targeted consultation with NSIs asked for more detailed opinions of statistics producers. The survey contained qualitative questions on costs and burden related to introduction of possible improvements per data collection. In addition, open questions on anticipated difficulties other than those related to cost or burden and identification of elements that were listed as potential improvement but are not needed in the opinion of the NSI.

Interviews were run, first for scoping purposes then in-depth. They had a semi-structured format and were addressing LMB users. No interviews with the NSIs were conducted due to the time contains. The most important element of them were integrated into the targeted on-line questionnaire.

One workshop was scheduled to present the draft findings of the impact assessment and to validate the results regarding the problem definition, feasibility, effectiveness, efficiency, and wider impacts of the policy options. Participants were given the opportunity to discuss their views and provide feedback on the research findings.

Consultation with SMEs was foreseen by the study accompanying supporting the impact assessment work to test directly the burden of the new proposals. Although several attempts were made to receive the information no replies were provided directly by the SMEs. Instead, information provided by the statistical institutes was used as a proxy assess the burden.

The responses to all consultation activities were carefully analysed. To analyse the public consultation results, all replies were first mapped onto the key stakeholder groups and then assessed by each group. Where relevant, the synopsis report presents the diverging views of various groups of stakeholders on the same issue.

3.Stakeholder participation 

The consultation activities mentioned above were implemented by Eurostat and/or the contractor supporting Eurostat during the evaluation and impact assessment - ICF S.A., Belgium. Table 2 presents the timeline of key consultation actions.

Table 2    Timeline of key consultation actions

In terms of level of participation, no replies were received during the 4-weeks consultation on call for evidence while the public consultation resulted in 79 replies from 21 Member States, Montenegro, and the United Kingdom. Most responses were received from Italy, followed by Latvia, Sweden and Belgium. Almost half of the respondent (46 %) represented public authorities, followed by academic and research institutions (19 %). Trade unions and companies represented each 4 % of the respondents. In terms of size, the majority of respondents came from large organisations with 250 or more employees (71 %), while only 6 % of respondents represented micro and small organisations.

In total 63 stakeholders were invited for interviews. Out of this number, 28 interviews were carried out with data users representing institutional policy users, trade unions, business associations, research organisations and international organisations.

A total of 27 responses were received during the on-line consultation with the NSI´s from the following countries: Austria, Belgium, Bulgaria, Czechia, Cyprus, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and Switzerland. Contributions were high quality and adequate and were used to feed into the impact assessment.

The workshop was attended by 9 participants.

No response from the media were received during the consultation. This was not considered to be problematic as the LMB statistics does not normally attract much of the media attention.

The consultation ensured appropriate stakeholder coverage for statistics producers and several types of statistics users. All NSIs of EU Member States and EFTA countries in charge of compiling LMB statistics and transmitting it to Eurostat were involved in the various consultation activities. Limited feedback was obtained from administrative data providers (three replies to the public consultation).

To ensure that the views of SMEs are capture and information is collected on the cost and burden on SMEs, particularly micro-enterprises for the different policy options, the study team has tried to collect information from different sources:

SME Europe was contacted and the organisation circulated request for data collection among their members

Local network contact points of Enterprise Europe were contacted in the following Member States: BE, BG, HR, IT, ES, NL, RO, FR, DE, SK, LV

The study team used a sample from the Orbis database to reach out to ten micro-enterprises from: CZ, EL, ES, IT, PL, SK

In addition during the workshop two organisations expressed interest to help facilitate data collection from micro-enterprises: Confederation of Danish Employer and the Austrian Federal Economic Chamber.

The Confederation of Danish Employer liaised with its members and provided four contact details to SMEs from which one organisation responded the questionnaire. All other sources resulted in no responses from SMEs.

Furthermore, LMB review was presented and discussed with Commission informal expert groups. These were mostly meetings of Eurostat-led expert groups with the participation of Member State experts, namely a LAMAS Working Group, the Directors of Social Statistics as well as the European Statistical System Committee (2 meetings in total).

4.Main results from the stakeholder consultation

4.1.General feedback on current statistics

Results of the public consultation show that overall stakeholders are satisfied with the quality of European labour market statistics on businesses. The main quality elements that were highly rated by stakeholders include:

Overall high quality

Relevance

Sufficiently reliable and accurate

Contribute to the public good and

Are impartial and objective.

Results of the semi-structured interviews confirm the findings of the public consultation and showed an overall positive outlook on the quality of LMB statistics while some areas for improvement were identified.

4.2.Areas for improvements

The elements of LMB that were found important to be further improved in the public consultation were:

extending the coverage of the survey to micro firms and for NACE Rev. 2 section O for the SES, LCS and for the missing countries in the JVS;

improving timeliness of the SES and LCS as well as LCI, in particular for the EU and euro area for the latter;

Collecting and publishing the numerator (labour costs) and the denominator (index of hours worked) of the LCI ratio for several NACE aggregates or quarterly data on the ‘total number of hours worked’ was found as needed by stakeholders.

Moreover, the consultation results showed that increasing the frequency to every 2 years for the SES was supported by 55 % of data users but only 39 % of producers of LMB statistics. The increase of frequency of the LCS was supported by 45 % of data users compared to 23 % producers.

During interviews, the users manifested a growing interest in new topics such as the impact of economic crises on the labour market, the inflation factor from the labour costs, mismatches of skills of the work force, types of pay schedules and possible discrimination in the labour market.

One of the most pressing problems highlighted by data users was in relation to data on gender pay gap, which is currently provided by Member States only on voluntary basis. For users this means that evidence is not comprehensive and suitable for cross-country comparison. In this regard, stakeholders offered detailed examples. Data users expressed also their need for detail on certain datasets including adding work history to study gender pay gap, data and availability of information by NUTS regions. Furthermore, with regards to the need of additional breakdown and variables, stakeholders expressed the need for additional detail, including more detailed level of disaggregation of information on occupation in the SES, adding work history to study GPG, (and data by NUTS region), which is not currently included in the SES. Moreover, further information on e.g. net earnings and annual earnings in kind would be appreciated by some GPG users as a way to analyse other aspects of gender segregation and/or discrimination in pay.

Stakeholders shared the view that it would be beneficial to include more innovative data collection method as it has the potential to decrease burden on data providers. The ESS is also working on new techniques such as web scraping. However, some stakeholders noted that one should ensure that data from innovative sources are of high quality and comparable with other data sources.

4.3.Qualitative assessment of areas needing improvements. 

The quantitative evaluation of costs and burden related to the improvements took place at the technical level prior the start of evaluation and impact assessment of the LMB review. Following problem definition of the LMB and identification of preliminary policy options, the NSI´s where addressed with the targeted consultation to assess the possible improvements. This assessment covered qualitative questions on improvements related to coverage, frequency and timeliness of the particular data collections constituting the LMB statistics and monetary and qualitative assessment addressing some punctual needs related to for instance missing variables in one data collection (SES).

Coverage of the LMB

Extending the coverage of the SES to micro firms (1-9 employees) for main information would cause significant increase in costs for 53% of the NSI´s (for 15% it will be one off cost for 38% regular every time the survey takes place). For 46% of NSI´s the increase will be no significant or zero. Extending the coverage of the SES to micro firms (1-9 employees) for whole set of variables would cause significant increase in costs for 62% of the NSI´s (for 4% it will be one off cost for 58% regular every time the survey takes place). For 38% of NSI´s the increase will be no significant or zero.

Extending the coverage of the LCS to micro firms (1-9 employees) would cause significant increase in costs regular every time the survey takes place for 62% of NSI´s. No increase or insignificant would be according to 38% of NSI´s.

Naturally, the extension of the coverage for JVS (both for the micro enterprises and for the NACE Rev. 2 section O) will affect only the 3 countries that do not do it yet.

For costs relating to extending the coverage to NACE Rev. 2 section O, over half of respondents (69%) noted no increase in costs for SES datasets. 19% noted an increase compared to the baseline every time the survey takes place. The remaining 12% noted costs when starting-up. For LCS, most respondents (81%) indicated no increases in cost. 15% of respondents indicated significant cost every time the survey takes place and the remaining 4% indicated a significant increase when starting up.

Frequency and timeliness

Transmitting all LCI results earlier, 65 days instead of 70 days after the end of the reference quarter would, according to NSI, not lead to significant increases in cost compared to the baseline (64%). Likewise, improving timeliness of the SES by two months is not likely to result in increases in cost compared to the baseline (58% of NSI´s).

Costs associated with collecting SES data every two years instead of every four years may be high as a large majority of respondents (84%) noted a significant increase compared to the baseline in cost every time the survey takes place. The remaining 16% noted no increase in costs.

For the LCS survey, most NSI´s (76%) reported an increase in cost every time the survey takes place every two years instead of four. The 20% of respondents indicated no increase in costs.

Additional variables

Replies to the NSI´s consultation showed that introduction of variables needed for deep policy analysis (e.g. disability status of employees according to the administrative concept in the SES, career brakes) would be problematic. The reason being the differences in the definitions at the national level, lack of this information to the employer, issues related to confidentiality and ethics as well as a high burden they cause on the employers.

At the same time other variables (e.g. occupation to be collected at more disaggregated level in the SES) seem to be relatively easier to be collected.

Other

Comments made by participants during the online showed that, while participants welcomed the additional measures and the full regulatory approach it was noted that there should be a balance between ambition and realism. It was emphasized that assessing the burden on enterprises for the proposed new measures would be crucial before making any final decisions.

5.Conclusions

The stakeholder consultation was implemented in line with Commission’s Better Regulation Guidelines. It reached successfully the intended stakeholder groups except for the administrative data providers and media organisations. Given the technical nature of the subject, the overall engagement of respondents was considered sufficient to support the back-to-back evaluation and impact assessment of the LMB.

The consultation supported the ongoing Commission initiative on LMB statistics and acknowledged importance of the data that constitute it. It also pointed out the existence of statistical gaps and the emergence of new statistical needs that cannot be satisfied within the current legal framework.

All stakeholders confirmed the need to plan statistical improvements although they did not always agree on the level of ambition of such improvements and statistics producers were somehow more prudent to changes than statistics users.

Statistical topics that gathered support across all stakeholder groups as priority topics were increase of the coverage in particular to the NACE Rev. 2 section O for all data collections and to microenterprises for the SES and better timeliness of the LMB data.

The main statistical topics where producer and users had different views concerned the increase of frequency of the SES or LCS. Producers and users also had different views regarding the addition of specific variables in SES.

Annex 3: Who is affected and how?

1.Practical implications of the initiative

The preferred option (option 2) would require the following changes in the LMB data collections.

Labour cost index

It is proposed to develop a flash LCI, to be transmitted by all countries representing more than 3% of the total number of employees at EU level (currently: Belgium, Germany, Spain, France, Italy, the Netherlands, Poland, Portugal and Romania). The flash LCI should be transmitted at t+45 days (in addition to the final data transmitted at t+70 days as is already the case) and published three working days later (i.e. at t+50 days at the latest).

Job vacancy statistics

For JVS, the main proposal consists in eliminating possible biases by covering all economic activities and enterprise sizes, which is not yet the case for three Member States. In France and Italy, public institutions are not covered within the sectors of public administration, education and human health (NACE Rev. 2 sections O, P and Q). In France, only units with 10 employees or more are included. Denmark only covers units within the business economy (NACE Rev. 2 sections B-N).

Gender pay gap

Annual (unadjusted) on gender pay gap data are currently collected under a gentlemen’s agreement. Under the LMB review, it is proposed to introduce a legal basis for the transmission of this important indicator. This will also serve the needs of the proposed directive on pay transparency whose monitoring will require annual on gender pay gap data.

Structure of earnings survey

It is proposed to extend the scope of the four-yearly Structure of Earnings Survey to micro firms (with 1 to 9 employees) and to public administration (NACE Rev.2 section O). The SES extension aims to correct the existing biases in the mean and median earnings which are used in a number of policy-relevant policies including the newly adopted directive on adequate minimum wages and for the on gender pay gap.

It is also proposed to shorten the deadline for transmitting SES data to Eurostat.

Labour cost survey

It is proposed to extend the scope of the four-yearly labour cost survey to public administration (NACE Rev.2 section O) and to shorten the deadline of transmission of data to Eurostat.

Moreover, the proposal will limit the data collection on apprentices to Member States where they represent more than 3% of the total number of employees at EU level (currently: Germany, Denmark, France, Italy and Austria).

General

To compensate for the resulting increase in the response burden, the LMB review will foster the use of administrative and innovative (e.g. web-scraping) sources for the compilation of LMB data. Quality reporting templates will be adapted to fit administrative data and multiple sources. In addition the emerging needs related to new variables in SES, indices for hours worked and labour costs in LCI and more detailed breakdowns of the JVS will be implemented These will be done in the secondary legislations.

The data are collected by the Members States based on the sample survey, based on the extraction of the information already available in the administrative registers, as a combination of the methods described above. National Statistical Institutes or Other National Authorities are free to choose the method that is the most efficient in terms of burden and the statistical quality.

2.Detailed scoring by policy option

Table 12.Policy Option 0: No targeted EU action (baseline)

2.1.Policy option 1: Reliance on the voluntary nature of contributions

Table 13.Policy Option 1: Reliance on the voluntary nature of contributions

2.2.Policy option 2: High coverage, better timeliness and other measures

Table 14.Policy Option 2: High coverage and timeliness; GPG regulated

2.3.Policy option 3: Highest coverage and increased frequency

Table 17.Policy Option 3: Highest coverage and increased frequency

(1) Estimates are gross values relative to the baseline for the preferred option as a whole (i.e. the impact of individual actions/obligations of the preferred option are aggregated together); (2) Please indicate which stakeholder group is the main recipient of the benefit in the comment section;(3) For reductions in regulatory costs, please describe details as to how the saving arises (e.g. reductions in adjustment costs, administrative costs, regulatory charges, enforcement costs, etc.;); (4) Cost savings related to the ’one in, one out’ approach are detailed in Tool #58 and #59 of the ‘better regulation’ toolbox. * if relevant

(1)Estimates (gross values) to be provided with respect to the baseline; (2) costs are provided for each identifiable action/obligation of the preferred option otherwise for all retained options when no preferred option is specified; (3) If relevant and available, please present information on costs according to the standard typology of costs (adjustment costs, administrative costs, regulatory charges, enforcement costs, indirect costs;). (4) Administrative costs for offsetting as explained in Tool #58 and #59 of the ‘better regulation’ toolbox. The total adjustment costs should equal the sum of the adjustment costs presented in the upper part of the table (whenever they are quantifiable and/or can be monetised). Measures taken with a view to compensate adjustment costs to the greatest extent possible are presented in the section of the impact assessment report presenting the preferred option.

3.Relevant sustainable development goals

Annex 4: Analytical methods

Detailed approach to the quantification of costs (efficiency)

The costs estimated as part of this evaluation were those associated with the current LMB framework specifically focusing on data collections defined by:

·Council Regulation (EC) No 530/1999 of 9 March 1999 concerning structural statistics on earnings and labour costs, (applicable for SES and LCS), 

·Regulation (EC) No 450/2003 of the European Parliament and of the Council of 27 February 2003 concerning the labour cost index (LCI)

·Regulation (EC) No 453/2008 of the European Parliament and of the Council of 23 April 2008 on quarterly statistics on Community job vacancies (JVS)

In addition the costs of voluntary gender pay gap data collection were assessed.

In order to get information on the current costs and the burden of the data collections, a survey was launched by Eurostat for National Statistical Authorities.

The questionnaire was structured along three parts, covering respectively:

(1) the costs for compilers (direct and indirect, both including FTE),

(2) the time per respondent (in most of the cases the respondent in LMB statistics is an enterprise, few countries however collect data for SES directly from the employees). For the sake of simplification, all costs for respondents are treated as costs of businesses.

(3) the time for all responding units. For the sake of simplification, all respondent related costs are treated as businesses costs.

Countries were asked to report the whole costs and burden for scenario of no changes in the legislation, and for the possible policy options, both the direct and indirect costs, after converting occupied posts in Full Time Equivalents (FTEs 37 ) Costs had to be reported in national currency and were converted into euros according to the latest exchange rates 38 available. The survey aimed to cover the whole production cycle - from the preparation of collecting the information via a survey or administrative sources, applying the corresponding legislation, the methodological work, the verification of the data in the country, all work related to dissemination - until the moment of sending the data to Eurostat.

On burden, countries were asked to report on time spent annually per respondent in minutes and of time spent annually by all respondents. The results were requested to be annual to account for differences in frequencies of LMB data collections (quarterly, yearly, 4 yearly). The burden was requested to be provided only in relation to scope required by the EU.

Eurostat has estimated the cost and burden for enterprises to provide data for LMB data collection.

For costs of compliers, (NSI), EU totals were derived as a simple sum. When a country did not report any costs or when the results had not been confirmed in the validation process, the values were imputed using the results from a country with a similar number of employees.

Given the low country coverage of the survey on burden by median time spend by all responding units, the values for missing countries were imputed by the EU median value calculated based on the available countries information. To monetise burden of respondents, time for all responding units in hours was multiplied by the mean hourly labour costs for the country concerned. It should also be noted that in some cases authorities extract data from administrative resources or data are derived from other surveys.

The data from compilers were complemented by results of study from the ICF as no clear distinction of one-off costs and recurrent costs were available from the compilers survey. The assumptions used for this assessment are provided in the Annex IV

Annex 5: Assumptions used for the distribution of costs to the BR typology

The contract support study on this impact assessment has assessed a total of three options:

Option 1: Reliance on the voluntary nature of contributions

Option 2: High coverage, better timeliness and other measure

Option 3: Highest coverage and increased frequency

Costs have been considered for three main stakeholder groups, namely 1) Member States and their National Statistical Institutes (NSIs), 2) the European Commission, including Eurostat specifically and 3) data producers/businesses. The cost of businesses participating in the data collection for all three policy options were assessed in the Eurostat study conducted with the NSIs (as described in the Annex IV). Similarly, the baseline costs of LMB were assessed for NSIs and businesses in the same study. The results collected as annual to account for differences in frequencies of LMB data collections (quarterly, yearly, 4 yearly). And in the consolidation of results, they were considered as recurrent and related to the data collections for the NSIs and costs for participating in surveys for the businesses.

Assumptions related to other types of costs (not related to data collection or participation of businesses in the LMB surveys) are summarised in the Table 1.

Further assumptions used (e.g. the hourly costs rates for the one-off costs in the NSIs) are annexed to the ICF study report.

Benefits were generally not appropriate for quantification. For example, the benefits to non-institutional data users from increased access to high quality labour market statistics on businesses would be challenging to quantify since this would depend on several additional factors, such as how this data would be used or the cost of accessing data through alternative sources. Therefore, estimates for benefits are not available.

1.Detailed overview of estimate costs

Mapping of the costs categories to typology used in the Annex II

(1)

Statistics in the ESS context are defined according to Article 3(1) of Regulation (EC) No 223/2009 (see footnote 2) as “quantitative and qualitative, aggregated and representative information characterising a collective phenomenon in a considered population.”

(2)

Regulation (EC) No 223/2009 of the European Parliament and of the Council of 11 March 2009 on European statistics and repealing Regulation (EC, Euratom) No 1101/2008 of the European Parliament and of the Council on the transmission of data subject to statistical confidentiality to the Statistical Office of the European Communities, Council Regulation (EC) No 322/97 on Community Statistics, and Council Decision 89/382/EEC, Euratom establishing a Committee on the Statistical Programmes of the European Communities ( OJ L 87, 31.3.2009, p. 164 ).

(3)

European statistics Code of Practice, revised version endorsed by the European Statistical System Committee on 16 November 2017 ( KS-02-18-142 ).

(4)

  https://ec.europa.eu/info/law/law-making-process/planning-and-proposing-law/better-regulation-why-and-how_en .

 

(5)

Council Regulations: No 10/1960, No 14/1961, No 28/1962, No 151/1962, No 101/66/EEC, (EEC) No 1899/68, (EEC) No 2259/7, (EEC) No 328/75, (EEC) No 494/78, Regulations (EEC) No 2053/69, (EEC) No 3192/73 and (EEC) No 494/78, (EEC) No 1596/81,  (EEC) No 3149/83, (EEC) No 1612/88, (EEC) No 3949/92, No 2744/95, (EC) No 23/97,

(6)

The study scope covers the 27 EU Member States plus the EEA Member States of Norway, Iceland and Lichtenstein.

(7)

  https://ec.europa.eu/social/main.jsp?catId=113#LMD  

(8)

  The Governing Council is the main decision-making body of the ECB. It consists of the six members of the Executive Board, plus the governors of the national central banks of the 19 euro area countries.

(9)

For instance: https://www.ecb.europa.eu/pub/economic-bulletin/html/eb201801.en.html#IDofArticle1 , https://www.ecb.europa.eu/pub/economic-bulletin/articles/2021/html/ecb.ebart202008_02~bc749d90e7.en.html#toc7

(10)

For instance: https://www.ecb.europa.eu/pub/pdf/scpops/ecb.op232~4b89088255.en.pdf?1ccf533dc92317c07a71721418088bd4

(11)

  https://ec.europa.eu/info/strategy/priorities-2019-2024/economy-works-people/jobs-growth-and-investment/european-pillar-social-rights_en  

(12)

  https://ec.europa.eu/eurostat/web/sdi/overview  

(13)

  https://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2009:0404:FIN:EN:PDF .

(14)

Regulation (EU) 2019/1700 of the European Parliament and of the Council of 10 October 2019 establishing a common framework for European statistics relating to persons and households, based on data at individual level collected from samples […] (Text with EEA relevance) (OJ L 261I, 14.10.2019, p. 1)

(15)

  https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022L2041  

(16)

  https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32022L2041  

(17)

  https://ec.europa.eu/eurostat/documents/3859598/5925693/KS-02-13-269-EN.PDF/44cd9d01-bc64-40e5-bd40-d17df0c69334  

(18)

  https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX%3A32019R2152

(19)

See also: https://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?TargetUrl=LST_NOM_DTL&StrNom=ISCO_88_CO&StrLanguageCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC

(20)

 Costs of apprentices not included 

(21)

Czechia, Denmark, Estonia, Spain, Cyprus, Latvia, Lithuania, Netherlands, Poland, Slovenia, Slovakia. Additionally, three MS (Ireland, Hungary and Finland) provided data for enterprises with 3 or 5 and more employees.

(22)

Bulgaria, Czechia, Germany, Estonia, Ireland, Spain, Cyprus, Latvia, Lithuania, Netherlands, Slovenia, Slovakia. Additionally, Hungary and Finland provided data for enterprises with 5 and more employees.

(23)

Bulgaria, Czechia, Estonia, Spain, Latvia, Lithuania, Netherlands, Portugal, Romania, Slovenia, Slovakia. Additionally, Hungary and Ireland provided data for enterprises with 3 or 5 and more employees.

(24)

In France and Italy, public institutions are not covered within the sectors of public administration, education and human health (NACE Rev. 2, sections O, P and Q) Denmark provides data for NACE sections B to N.

(25)

  https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52021PC0093&from=EN

(26)

Possible variables to consider at the level of secondary legislation: occupation at ISCO-3 digits, net earnings, gross earnings in kind, disability status of employees, career breaks of employees, number of employee’s dependent children, institutional sector of the enterprise.

(27)

In addition limiting the collection of labour cost survey data on apprentices to countries where they represent less than 1% of the total number of employees will serve as a simplification measure.

(28)

  https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32009R0223&from=EN  

(29)

  https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/1342-European-business-statistics-FRIBS-_en  

(30)

  https://intragate.ec.europa.eu/decide/sep/?view-dossier-details-id=DORSALE-DOSSIER-2021-24018

(31)

ECFIN, EMPL, JUST, JRC, RTD, SG and SJ

(32)

Working Group on labour market statistics (LAMAS WG).

(33)

ECB; Commission DGs: ECFIN, EMPL, GROW, JUST, REGIO, agencies: EIGE, EUROFOUND

(34)

See the Factual summary report on the public consultation of LMB: https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13266-European-labour-market-statistics-on-businesses-LMB-/public-consultation_en

(35)

https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13266-European-labour-market-statistics-on-businesses-LMB-_en

(36)

  https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/1342-European-business-statistics-FRIBS-_en  

(37)

FTE: Full Time Equivalents varies (in respect of working days and working hours per day) across countries, it is proposed to use the following standards: one year corresponds to 215 working days (used as well in Commission’s unit cost methodology), and one working day equals 8 hours worked. However, NSIs are free to decide to use their own (country wise) definition of ‘FTE’. In that case, the NSI was invited to communicate (for each survey) how many working days and hours one ‘FTE’ corresponded in the reporting country.

(38)

  https://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ert_bil_eur_m&lang=en

Euro/ECU exchange rates - monthly data [ert_bil_eur_m] extracted on 05 May 2021.