EUROPEAN COMMISSION

Strasbourg,13.3.2018

SWD(2018) 70 final

COMMISSION STAFF WORKING DOCUMENT

IMPACT ASSESSMENT

Accompanying the document

Proposal for a Council recommendation

on access to social protection for workers and the self-employed

{COM(2018) 132 final}

{SWD(2018) 71 final}

Table of contents

1.    Introduction: Political and legal context    

2.    Problem definition    

2.1.    A first look at how many people are potentially not covered by social security    

2.2.    Problem drivers    

2.3.    Consequences    

3.    Why should the EU act?    

3.1.    Legal basis    

3.2.    Results of the social partners consultations and of the public consultation    

3.3.    Subsidiarity: Necessity and value-added of EU action    

4.    Objectives: What is to be achieved?    

4.1.    General objectives    

4.2.    Specific objectives    

5.    What are the available policy Measures?    

5.1.    Baseline scenario    

5.2.    Measures discarded at an early stage    

5.3.    Description of the policy measures    

5.4.    Impacts of the policy measures    

6.    What are the Policy packages and their impacts?    

6.1.    Description of alternative policy packages    

6.2.    Impacts of the policy packages    

6.3.    Comparing the policy packages    

7.    Preferred Package    

7.1.    Preferred policy option    

7.2.    Overall impacts    

8.    The choice of legal instrument    

8.1.    Legal options    

8.2.    Preferred instrument    

9.    How will actual impacts be monitored and evaluated?    

9.1.    Monitoring    

9.2.    Evaluation    

Annex 1: Procedural information    

Annex 2: Glossary and Terminology    

Annex 3: Further information on the organisation of social protection systems    

Annex 4: Further information on non-standard employment and self-employment    

Annex 5: Visualisation of the Problem Tree    

Annex 6: Internal Drivers    

Annex 7: Overview of recent reforms in Member States    

Annex 8: The existing EU instruments    

Annex 9: Analytical methods    

Annex 10: Further information on impacts of policy measures and policy packages    

Annex 11: Stakeholder consultation    

Annex 12: Who is affected and how?    

Annex 14: Monitoring indicators    

Annex 15: Summary of case studies    

Annex 16: Behavioural study    

1.Introduction: Political and legal context

Globalisation, technological developments and demographic ageing have driven changes in the European labour markets for the last two decades and will continue to do so in the future. Employment will increasingly be more diverse and careers will be less and less linear. As labour markets evolve, social protection systems too need reforms in order to make our social model future-proof making the most of the future world of work.

Offering access to social protection is crucial for the economic and social safety of the workforce and the well-functioning labour markets that create jobs and growth. Yet there is a growing number of people who, due to their employment status, are left without sufficient access to social security.

The Commission Reflection Papers on harnessing globalisation 1 and on the social dimension 2 recall that national social protection systems in the EU differ due to different national traditions, political preferences and budgets but they face similar, transformative challenges. They underline the importance of social policies for a better distribution of the benefits of globalisation so people are confident that governments will protect and empower them in the adaptation to change. They also stress the importance of delivering new protections to help individuals use their full capacities in fast-changing labour markets.

During the public consultation 3  on the European Pillar of Social Rights many stakeholders expressed concerns regarding the obstacles faced by workers in non-standard forms of employment and by the self-employed in accessing social protection such as unemployment benefits, sickness benefits, disability benefits and pensions.

In his State of the Union speech in September 2017, President Juncker further stressed that a social fragmentation and social dumping are to be avoided in Europe. He underlined that, whilst national social systems will still remain diverse and separate for a long time, the EU should strive for a European Social standards Union and for a common understanding of social fairness in the single market.

Solemnly proclaimed in November 2017, the European Pillar of Social Rights 4 is the EU's political commitment to address these challenges and to uphold citizens' social rights in times of structural changes, including by stressing that "regardless of the type and duration of their employment relationship, workers, and, under comparable conditions, the self-employed have the right to adequate social protection" (principle 12). The Commission's initiative on "Access to social protection for workers and the self-employed" 5 is a key step to make principle 12 of the Pillar a reality on the ground. The initiative (2017/EMPL/003) is contained in the Commission Work Programme 2017 6 and 2018 7 and is a key component of the Social Fairness Package 8 to be adopted by the Commission on 13 March 2018.

The objective is to support people in non-standard forms of employment and self-employment who, due to their employment statuses, are not sufficiently covered by social security schemes and thus are exposed to higher economic uncertainty. By focusing on the level playing field for economic agents at national level, the initiative will also aim at avoiding short-term competitions in the internal single market, complementing the coordination of social security across borders 9 already set at EU level and ensuring long-term gains stemming from the economic benefits throughout the Union.

While acknowledging that many social protection schemes aim at protecting all citizens - whether or not active in the labour market - the initiative is expected to concentrate primarily on social security schemes which are governed, financed and designed in function of the labour market dynamics and leaves out of scope social assistance and other social protection schemes which are based on residence/citizenship status. The initiative focuses in particular on social security areas which are mostly impacted by the changing nature of work relationships, namely unemployment benefits; sickness benefits; maternity and equivalent paternity benefits; invalidity benefits; old-age benefits as well as benefits in respect of accidents at work and occupational diseases.

This Impact Assessment describes the current gaps in social security for people in non-standard forms of employment and self-employment across the Member States and discusses different policy options and impacts for EU action. It is important to note that the quantification of impacts described in this document does not capture the entire scope of the problem analysed. The limited nature of the quantification made stems from technical limitations of the statistical data and of the simulation models available. It also reflects the uncertainty related to the decisions on design and financing of social security systems. Since these decisions are to be left to the Member States, according to the principle of subsidiarity underlined in Article 153 (4) TFEU, the choices to be made are yet unknown and therefore a full measurement of their fiscal and economic impacts is not possible.

In light of the above, this document presents a proportionate assessment of impacts. The impacts are primarily described through a qualitative assessment and, where possible, a quantification of impacts is outlined.

Procedural information regarding the preparation of this Impact Assessment can be found in Annex 1. Annex 2 provides the glossary and definitions of key terms used in this document.

2.Problem definition

Social security serves to protect people against the financial implications of social risks, such as ill health, old age or job loss, and contributes to preventing and alleviating poverty and social exclusion. It also functions as a buffer during economic downturns by supporting domestic demand. Social security can take the form of in-cash benefits which replace or supplement work income (unemployment benefits, pensions, sickness benefits, child benefits) and/or in-kind benefits which provide support through the supply of goods and of services (healthcare, child care, training or job search assistance). Further information on the basic characteristics of social security systems in the EU can be found in Annex 3.

Well-designed social security systems are an important lever of competitiveness and sustainable growth as they facilitate labour market participation and support human capital investment in a world of rapid and growing labour market transformations. Adequate social security contributes to a better allocation of the labour force towards emerging and dynamic sectors of the economy by providing workers opportunities and incentives to participate in education and training and devote time to search for a job matching their skills. It helps people feel secure in order to fully embrace the opportunities created by new forms and patterns of work, new working relationships with employers, growing career transitions as well as self-employment. Social security systems thereby fulfil important functions for individuals, labour markets as well as at macroeconomic level, where they act as automatic stabilisers dampening welfare-reducing economic cycles.

This initiative addresses the problem that a significant and increasing share of people in the EU's labour market has insufficient social security coverage due to their employment status, in particular non-standard workers and the self-employed.

This development impedes social security systems from fully fulfilling their multiple functions outlined above. Such gaps increase risks to the welfare of the affected individuals and their families who endure greater economic uncertainty. The accumulated effects of such disparities in entitlements are likely to give rise to new inter- and intra-generational inequalities between those who have or manage to gain employment on contracts with full social rights and those who do not. Women may also be more disadvantaged by social security systems still geared towards linear full-time employment as they are more exposed to non-standard employment, particularly to part-time work, but also to temporary work and to fragmented careers due to interruptions related to children or caring for dependent relatives.

Negative consequences are also felt at the aggregate level, including in terms of higher labour market segmentation and unfair competition between firms employing staff on non-standard contracts exempted from social security contributions and those not doing so. Further, there can be effects on the financing base of social security systems, and labour market dynamism to the extent that mobility and transitions between different employment forms are hindered by differences in social security coverage. In the long run, it is the social and economic sustainability of national social security systems that is at stake. Gaps in access to social security for growing groups of the workforce will lead them to take recourse to safety nets of last resort in the case of social risk while the number of people contributing to social security systems will be proportionately smaller.

2.1.A first look at how many people are potentially not covered by social security

Social security systems were primarily developed for and remain geared towards workers in ‘standard employment relations’, implying a long-term, full-time work relationship 10 . Recently developments towards social security systems with more incentives to work have taken place, but the systems were not tailored to the specificities of people in different forms of employment, creating deficiencies in the way they are protected.

This potentially concerns a large share of people in the EU's labour market since the share of people in non-standard employment relationships and dependent self-employment has been increasing 11 . In 2016, 14% of the total 219 million people employed in the EU were self-employed, 8% were full-time temporary employees, 4% were part-time temporary employees, 13% were part-time permanent employees and 60% were employees with a full-time permanent contract (Figure 1). Altogether, people in all categories of non-standard employment and self-employment constitute around 39% of the employed population.

In addition to the increasing diversity in employment forms, careers are also becoming less linear. In the space of a generation, the average European worker has gone from having a job for life to having more than ten in a career 12 . People are frequently moving between different employment statuses and/or combining salaried employment and self-employment. Moreover, transitions from school to work are taking more time than before, and many young people accumulate several traineeships in the process, which in many Member States’ social security systems implies delaying the start of the payments of contribution and thus has an impact on the building up of entitlements. In general, evidence shows that the crisis has led to rising involuntary temporary and part-time work and to increased perceptions of job insecurity overall 13 , which may have been accentuated by the existing deficiencies in protection (further information on the levels and the observed trend increase in non-standard employment and on self-employment are presented in Annex 4).

Figure 1: Extent of different types of employment relationship in the EU28 in 2016

Source: Eurostat

Matsaganis et al. (2015) 14 estimated the risk for non-standard workers and the self-employed of not being covered 15 by three social benefit schemes: sickness, unemployment and maternity benefits. Their results show 16 a significant risk of not being covered by benefits for non-standard workers and the self-employed in relation to these three benefit schemes.

As shown in Figure 2, two million people employed on non-standard contracts in the EU could be at risk of not being covered by sickness benefits, with a significant variation across Member States. In particular, this risk concerns 5.1% of temporary full-time workers and 9.7% of temporary part-time workers in the EU. Concerning the risk not to be covered by unemployment benefits, eight million people in non-standard contracts are at risk: 31.9% of temporary full-time workers and 38.7% of temporary part-timers are at risk, against 0.1% of the standard workers (Figure 3). When it comes to maternity benefits, the proportion of non-standard workers at risk of not being covered is 1.8% for permanent part-timers, 8.5% for temporary full-time workers and 12.7% for temporary part-timers. As will be shown later, in general, the existing rules do allow non-standard employees to adhere to the corresponding social security schemes ("formal coverage"), although some specific categories may be excluded. The estimated figures thus also reflect the fact that despite being members, non-standard employees may often not be able to build up and access benefits ("effective coverage").

Figure 2: Distribution of people in employment aged 15-64 at risk of not being entitled to sickness benefits by type of employment, 2014 (% of the total in employment)

Source: Matsaganis et al. 2015

Figure 3: Distribution of people in employment aged 15-64 at risk of not being entitled to unemployment benefits by type of employment, 2014 (% of total in employment)

Source: Matsaganis et al. 2015

Among the self-employed in the EU, about 12 million are estimated not to be eligible for sickness benefits, again with a wide variation between countries (Figure 2). The risk of not being covered by unemployment benefits for the self-employed is as high as 54.5% (corresponding to ca. 17 million), again with significant cross-country differences (Figure 3). In both schemes the variation across Member States relates to the fact that insurance is mandatory in some Member States, voluntary in others and non-existent in another group of countries. The proportion of the self-employed who are estimated not to be covered by maternity or paternity benefits is much higher: 46.1%.

These figures illustrate the problem for three branches of social security. A qualitative mapping 17 by Member State, by social security branch and by employment category carried out for the purposes of this Impact Assessment (see Section 2.2.2 and Annex 6) shows that similar problems also exist for other branches of social security even if data limitations prevent the presentation of corresponding quantitative estimates.

Complementary evidence is available from a survey commissioned for this Impact Assessment. 18 Considering the self-reported rates of coverage by key social security branches, large gaps emerge between standard employees on the one hand and the self-employed and non-standard workers on the other (see Figure A.28 in Annex 16). These gaps are particularly pronounced when considering unemployment benefits and sickness benefits. Compared to standard employees, the incidence of self-reported benefit coverage in case of need is 50 percentage points (pps) lower for the self-employed and 29 pps lower for non-standard employees regarding unemployment benefits. For sickness benefits, self-reported benefit coverage is 35 pps for the self-employed and 33 pps for non-standard employees.

Differences are less pronounced in the case of old-age benefits between standard employees and the self-employed non-entrepreneurs (6 pps), while there is a difference of 22 pps between standard and non-standard employees. Similar gaps were registered in case of maternity, invalidity and accidents at works and occupational injuries.

It is important to notice that such gaps in self-reported coverage may reflect existing real limitations, but also reflect a lack of transparent information allowing people to be aware of benefits they are entitled to.

2.2.Problem drivers

The gaps of access in social security affecting non-standard workers and the self-employed are the result of external drivers to social security (like changes in labour markets, technological progress, globalisation and demographic change) as well as intrinsic ones. They are now reviewed according to the structure outlined in the problem tree shown in Annex 5.

2.2.1.External drivers

Several factors are shaping developments in the EU labour market(s) contributing to an increasing share of people being outside standard employment and thus potentially affected by gaps in social security coverage.

Intensifying globalisation has changed production processes and increased competition for firms 19 . Resulting needs to increase cost efficiency have, among others, been addressed by firms by lowering their labour costs – by offshoring and also by containing domestic labour costs and making it more flexible. As part of these efforts, employers may call in self-employed workers or hire temporary staff as well as people in new forms of work. Such flexible forms of labour can at the same time be more compatible with the above-mentioned requirements of production in a global economy. Moreover, there is an ongoing shift in the sectoral structure of the EU economy from manufacturing towards an increasing role of services. Since non-standard employment is more prevalent in the services sectors, this shift also implies an increase in this kind of employment 20 , including non-standard jobs in some high-skills service areas 21 .

The increasing recourse to existing forms of non-standard work as well as the emergence of new ones is also closely linked to recent technological developments transforming the world of work in unprecedented ways. New IT technologies, automation and digitalisation are changing the production and delivery of goods and services while technologies such as increased robotisation of standardisable tasks or the development of digital platforms are also having important impacts on the organisation of work 22 . These developments are giving rise to such phenomena as platform work, crowdsourcing and the collaborative economy. In conjunction with the above-mentioned changes related to globalisation, these technological transformations also resulted in higher use of traditional non-standard employment and self-employment and to the development of new atypical working arrangements 23 , such as workers with no employment contract (casual workers; "zero hours" contracts), workers who report working a very small number of hours (< 8 hours a week) or who hold a temporary contract for less than 6 months.

At the same time, an increase in non-standard forms of work may result from policies in response to population ageing, which is weighing on the size of labour forces and putting financial pressure on social security systems, both on the expenditure side (healthcare, pensions) and on the financing side (shrinking workforce) 24 . Offering flexible working practices 25 may play a key role in encouraging more people to enter the workforce (including women, low-qualified people and migrants), and may imply increasing the number of non-standard workers and expanding the diversity of new forms of employment.

The additional pressures of the recent economic crisis may have further contributed to the increasing use of new arrangements instead of traditional forms of work and contracts. While such new forms of work in relation to the developments outlined above respond to flexibility needs of (public and private) employers and workers, and have supported the integration of people in the labour market who otherwise risked to stay in unemployment, increases in labour flexibility may be associated with higher levels of job insecurity, as reflected in the emergence of unstable and informal jobs, with irregular salaries and reduced social security. Effective welfare policies are thus needed to accompany the changes to avoid larger inequalities within countries and heightened fragility of employment 26 .

Changes in labour markets institutions can mediate the effects of globalisation and technological progress on work and employment relations 27 . The ongoing rise of new working arrangements poses some challenges to the existing framework of employment legislation and, more broadly, of welfare policy – from minimum wages to benefits and pensions – which remains structured around the concept of standard employment contract. Furthermore, the employment/self-employment binary approach of employment legislation is no longer clear in labour markets. New forms of employment have blurred the boundaries between the two groups creating legal loopholes and leaving many people in the labour market unprotected.

The observed decrease in collective bargaining coverage 28 implies a loss of protection of working conditions for a growing number of workers 29 . Moreover, the instability of employment relationships, the fear of reprisals and the limited opportunity to join a union could sometimes hamper the "de facto" exercise of collective rights which are often being defined with reference to an open-ended contract model 30 , affecting especially non-standard employees. These developments thus imply an additional, complementary lack of protection.

2.2.2.Internal drivers

Alongside the external drivers, there are factors intrinsic to the social security systems which are root causes for the lack of access for non-standard workers and the self-employed. They regard the rules governing formal coverage, effective coverage, transferability and transparency. A summary overview of these drivers is presented in this section; Annex 6 provides more detail and background.

It should be noted that, although principle 12 of the European Pillar of Social Rights refers in general to social protection as an instrument to support fair and well-functioning labour markets and welfare systems, the internal drivers presented here relate only to a subset of social security branches.

A comprehensive mapping 31 of gaps in access for all branches – unemployment benefits, sickness benefits, old age benefits, benefits related to accidents and injuries at work, invalidity benefits, healthcare, family benefits, maternity/paternity benefits and long-term care - has been conducted. The mapping clearly showed that not all social protection areas are affected by labour market evolutions. 32  

The mapping showed that the areas where the type of work contract or employment status has a strong influence on the social protection coverage are unemployment benefits, sickness benefits, maternity and equivalent paternity benefits, invalidity benefits, old-age benefits, benefits in respect of accidents at work and occupational diseases; the analysis on these branches will thus be presented in more detail below.

The common element among these branches is that the benefits they provide have a function of income replacement in case the associated risk materialises and interrupts the person's work. For instance, sickness benefits provide income replacement in the case of prolonged illness. While healthcare also involves some cash benefits, these relate to the provision of services rather than to the replacement of income, and it has not been possible to identify a strong impact of a person's labour market status on the corresponding coverage. This branch is thus not included it in the scope of this initiative.

Another important point to note concerns the personal scope of the initiative. The initiative supports workers (as defined in Annex 2 33 ) which do not have a full-time open-ended contract, and the self-employed. Both groups show a high level of heterogeneity and the current forms of non-standard work and self-employment are described more in detail in Annex 2 and Annex 4. By virtue of the residual definition of non-standard workers given above, all current forms of non-standard employment, as well as of self-employment, are included in the scope of the initiative, without further sub-classifications. This also allows the initiative to be future-proof, since labour markets are evolving fast and technologies are giving rise to many different new forms of employment not yet known.

Formal coverage

Formal coverage means that a person in employment or self-employment is entitled to participate in social security systems. It is a prerequisite for benefitting from the protection granted by the scheme in case of need. Limited formal coverage means that the rules governing access to social security schemes do not allow individuals from specific groups to become members. Thus individuals are neither covered on a mandatory basis nor can they opt to join corresponding schemes on a voluntary basis.

Overall, people in non-standard employment usually have the same formal coverage by social benefit schemes as those in standard employment. However, in certain Member States, there are important exceptions relating to specific categories of non-standard workers, notably casual and seasonal workers, apprentices or trainees, on-demand workers, and those on temporary agency contracts as well as people working on certain contracts defined at national level, such as civil law contracts in Poland (more than half a million people) or mini-jobs in Germany (7.6 million) (see Table A.2 in Annex 6). Such coverage gaps are wide-spread geographically, they have been identified in fifteen Member States for unemployment benefits, twelve Member States for sickness benefits, twelve Member States for maternity benefits, seven Member States for accident at works and occupational injuries and thirteen Member States for old-age and survivors' benefits.

The situation is more critical for the self-employed. There are fundamental gaps in formal coverage regarding three core branches, where the self-employed as a group are excluded from membership. In particular, the self-employed, or important subgroups, do not have any access to unemployment benefits in eleven Member States (see Table A.3 in Annex 6). They cannot be insured against accidents at work and occupational injury through social security schemes in ten Member States and they are not covered by sickness benefits in three.

In other countries, the self-employed, or at least some categories, have formal coverage on a voluntary basis for unemployment insurance, sickness benefits and accidents at work in eleven, seventeen, nine countries, respectively. It is worth noting however that rates of enrolment in voluntary schemes are low (less than 1% to 20 %) in most Member States where this possibility exists. The low adherence of self-employed may be related to alternative individual coping strategies, affordability and financial disincentives, to unfavourable system design, lack of awareness or individual myopic behaviour. The latter may affect the individual decisions on intertemporal income distribution. This may be related to erroneous foresight, or to an unfounded fear that what has been promised for tomorrow might not materialise.

Effective coverage and transferability

Even when non-standard workers and the self-employed have formal coverage, they may experience gaps of effective coverage, that is, rules of social security schemes applicable to non-standard workers and the self-employed, compared to those applicable to standard employees, embed obstacles preventing them from accruing and taking up adequate entitlements in case of need. This may be because the rules are less favourable for the self-employed or non-standard workers or because identical rules for all three groups are more difficult to meet for people outside standard employment. For instance, given that temporary employees transition more frequently into unemployment than standard employees 34 , it can be relatively more difficult for them to accumulate a given qualifying period within the corresponding reference period to qualify for unemployment benefits.

According to the expert mapping 35 , the rules affecting effective coverage of the self-employed and of non-standard employees relate to time thresholds for contributions and entitlements (minimum qualifying period, minimum working period, waiting period or duration of benefits), to the way the income base is assessed for the self-employed or to the level of adequacy of the benefits (Annex 6). Overall, obstacles to effective coverage have been identified for at least nineteen Member States and in relation to six social security branches, notably unemployment, sickness, maternity, accidents at work and occupational injuries, invalidity and old age and survivors' benefits. Minimum qualifying periods and are particularly problematic for non-standard workers, while the self-employed mostly face problems related to longer waiting periods and shorter duration of benefits.

Regarding non-standard workers, minimum qualifying periods and minimum working periods in unemployment benefit schemes have been identified as problematic in six and four Member States respectively. Minimum qualifying periods are also challenging in the case of sickness benefits in five Member States. Waiting periods are unfavourable in six Member States in the case of maternity benefits (see Table A.6 in Annex 6).

Regarding the self-employed, they face longer qualifying periods for unemployment benefits in four Member States and are subject to longer waiting periods before receiving sickness benefits in eight Member States (see Table A.7 in Annex 6). The methods for calculating the income base or the reference income may also be unfavourable to them because they consider long previous periods of earnings or on future earnings instead of current earnings, not taking into account the fluctuations characterising incomes from self-employment. Moreover, in some Member States many self-employed may receive lower benefits compared to an employee with the same earnings because benefits are set at different levels according to the form of employment (e.g. pensions for the self-employed in Italy or sickness benefits for the self-employed in Belgium).

Transferability may refer to the possibility of transferring accumulated entitlements to another scheme following professional transitions between sectors or employment forms. 36 In this case a lack of transferability can hinder the building up of adequate entitlements over the course of a career characterised by frequent transitions as in the case of old age pensions.

Transferability can also refer to making qualifying periods in the previous employment status (or in concomitant employment statuses) count towards the qualifying periods in the new status. This is because the entitlement to benefits in the new scheme depends on contribution history but the entitlements in the old scheme cannot be transferred. In this second case, gaps in transferability may impede individuals for example of transitioning between employment and self-employment when these two statuses are covered by different schemes (see Table A.8 in Annex 6). In view of careers that are increasingly characterised by frequent changes between jobs and employment statuses, this regulatory failure is of rising importance. Obstacles to transferability can be due to lack of regulation in this area (four Member States), to extremely high administrative costs (one Member State) or to different rules governing different schemes (two Member States) which prevent people from transferring accumulated rights.

In the case of voluntary schemes, the transfer of entitlements may be influenced by behaviour of individuals. A number of factors can drive the decision to adhere or not to a voluntary scheme, including whether coverage is already available through the first job, the family context and private wealth accumulation which provide alternative means of coping with risks, as well as the individual's assessment of future risks facing them.

In some Member States the time profile of replacement rates in unemployment insurance depends on the contribution history and shorter contribution histories are likely to characterise workers on temporary contracts as compared to full-time workers. Here again behaviour of individuals matter. Particularly in the case of the self-employed, where the rules for declaration of income are less stringent, or where the individual has the option to (legally) choose what level of income to be assessed upon, they may report low incomes, leading to low entitlements. The low declaration of income due to myopic behaviour has been identified as an obstacle to the adequate protection of the self-employed in at least half a dozen of Member States 37 .

Transparency

Finally, there are gaps in the provision of transparent information to the general public, potential scheme members, schemes members and beneficiaries about the scheme's rules or about the individual obligations and entitlements. Transparency of information is ever more important due to the growing numbers of transitions between different employment status, contracts forms and social security schemes. A lack of simple and transparent information lowers the level of people's awareness on their rights and obligations towards social security and stops them from taking informed decisions. Indeed, a significant proportion of the population (see also Annex 6), especially among the self-employed and non-standard workers, is unaware of their social security coverage (up to 20% for some categories) 38 . If generic information about social security schemes is only absent in five Member States, personalised information, including an overview of rights and obligations and the possibility to obtain information through online simulation tools, is only available in about half of the Member States (see Table A.10 in Annex 6).

Despite the general acknowledgement of the challenges at hand, corroborated by the majority of stakeholders during the consultations organised in the framework of this Impact Assessment, national reform efforts are partial and uneven across the EU Member States. In recent years, some Member States started reforms to address some of the deficiencies related to social security coverage for non-standard worker and for the self-employed (see Annex 7 for a full overview of the national reforms). These reforms have mostly concerned the extension of social security coverage to people in non-standard employment and in self-employment, in particular concerning unemployment benefits coverage (e.g. France, Lithuania, Poland, Portugal and Sweden). Only a few reforms have addressed the problems of effective coverage (e.g. Spain), transferability (e.g. France and Ireland) and transparency. Moreover, although a number of countries have developed specific guidance and start-up programmes for the entrepreneurial self-employed, this has rarely also led to a redesign of social security taking into account their special needs.

The results of the implementation of already undertaken national reforms can only be evaluated in some years. However, the reforms seem partial in their scope (both regarding the branches of social security and groups of people in the labour market) and the efforts appear to be uneven across the Member States. The identified gaps in formal and effective coverage and the lack of transparency and transferability of accumulated rights and together with behavioural biases (and their influence on the voluntary take-up of social security schemes), therefore leave a large number of people in the labour market uncovered by social security.

2.3.Consequences

With an increasing number of people outside the scope of the social security systems, their ability to insure people against the financial implication of social risks, to prevent poverty, to support labour market participation and act as an automatic macroeconomic stabiliser is significantly reduced. Overall, a differential coverage of social security, based on contract and status of people in employment, has previously been criticised and documented 39 as both unfair and inefficient. It has consequences on those directly affected by the gaps but also on companies, on the economy and on society as a whole.

2.3.1.Consequences for non-standard workers and the self-employed

Insufficient coverage of social security significantly raises the risks to the welfare of the affected and their families. People who earn their living from non-standard employment or self-employment and do not have sufficient social security coverage are exposed to greater economic uncertainty due to the individualisation of risk. At the same time, some of these groups may be less able to deal with these risks on an individual basis. Together, this may lead to more precarious lives. Indeed non-standard workers and the self-employed are exposed to significantly higher poverty risk in most EU countries.

Non-standard employees' rate of being at-risk-of-poverty (AROP) is higher in all Member States (except in Malta) than the one for workers with a permanent and full-time job. On average in the EU28, around 15% of temporary and part-time workers are at risk of poverty, more than double the share observed for standard employees.

Figure 4: In-work at-risk-of-poverty rate employees according to their status, by countries, 2015, %

Source: Eurostat, Employees with a temporary job; Employees at part-time.    

Note: countries are ranked according to the growing poverty risk rate of workers with permanent jobs.

The significant variation that exists across Member States (see figure 5) reflects differences in both protection through the tax-benefit system and in income. Indeed, non-standard workers are estimated to have lower wages with respect to standard employees. For instance, it has been estimated that, controlling for workers characteristics, there remains a significant wage gap disadvantaging temporary employees relative to standard employees of 13% to 21% depending on the Member State 40 . The lack of social security coverage can thus compound disadvantages stemming from lower levels of income and unfavourable social security rules.

Similarly, the self-employed face a higher risk of poverty 41 than standard employees; their AROP exceeds 20% in the EU28, nearly three times the risk of standard employees. As reflected in Figure 5, observed ratios are much more pronounced in some Member States, with a maximum of 10 reached in Romania. Moreover, there is a negative correlation (- 0.6), between the at-risk-of-poverty of the self-employed and that of the employees 42 . This means that, in general, in Member States where employees are relatively well-protected against the risk of poverty, this is far less so for the self-employed. Where also the employees face a high risk, differences between both groups are less marked. When analysing the role of social security in alleviating poverty of both groups, it is noticeable that (see Figure 6 43 ) social security diminishes the poverty rate of both people in employment and self-employment in the majority of the Member States. However, across most Member States the reduction is visibly more pronounced for employees than for people in self-employment, which partly reflects the identified gaps in access for the latter group 44 .

The gaps in coverage of non-standard employment contribute to a lack of fairness within society, widening divides between generation, gender or social groups, since young people, the foreign-born and women are less frequently hired on full-time, open-ended contracts. The differences in coverage may also impede labour market mobility of these groups, which makes it even more difficult to close gaps between groups in the long-term.

2.3.2.Consequences for businesses and competitiveness

Differences in treatment related to social security create bias in the level playing field relating to companies’ hiring choices: firms may overuse employment contracts exempted from social security contribution. They transfer the risks and the costs of protection through social security systems to third parties (workers, dependant/bogus self-employed and/or public budget). Such practices contribute to maximise profits in the short-term, but they also have costs in the long run: they demotivate workers, lead to higher turnover, have reputational effects, and slowdown productivity gains and innovation at the company level. They also create unfair competition with companies that continue to contribute to social security systems and they threaten to erode the economic basis for standard employment as such jobs will be subject to competition from jobs exempted from the cost of social contributions.

During discussions on globalisation, business and political circles have expressed concerns that high employer contributions to social security make national economies less competitive. However, several studies have found a positive correlation between the openness and size of the welfare state, and social expenditure and competitiveness 45 . In particular they found that social spending increases competitiveness and show that reverse causality (i.e. that higher competitiveness leads to larger welfare states) is weak. They also find that inward FDI flows in OECD countries are positively affected by countries’ expenditure on social welfare.

2.3.3.Consequences for the labour market, public budgets, the economy and society

From an employment perspective, deficiencies in social security coverage increase labour market segmentation 46 , which can be associated with higher levels of unemployment (mainly for vulnerable groups) and lower quality of skills matching 47 . Moreover, the absence of social security make it less attractive for people to take non-standard jobs or move between and combine dependent employment and self-employment, preventing labour reallocation and hampering labour market's ability to respond to shocks as well as making segmented labour markets self-perpetuating, ultimately with negative implications on (total factor) productivity growth 48 .

Moreover, differences in social security coverage distort the playing field between people who pay no or lower social security contributions and those with full, obligatory contributions. It creates legal possibilities to avoid paying contributions and to underbid, leading to a race to the bottom and downward social convergence.

Insufficient protection through social security systems for certain groups may contribute to higher spending for public budgets in other areas. In the short term an increasing share of people in non-standard work or self-employment without social security coverage does not necessarily save public money and may even lead to rising costs for (generally tax-financed) safety nets of last resort that they may then rely upon. This is because, on the one hand, even if not paying social contributions and receiving social benefits, they are entitled to social assistance, minimum income or universal minimum healthcare benefit package. Increased reliance on the latter can have an additional efficiency cost including through delays in care seeking and overuse of costly emergency services instead of primary or integrated care services.

In the longer run, the absence of benefits to facilitate social and economic inclusion may make long-term unemployment and mental health problems more likely 49 . Therefore, if the longer-term welfare of people in non-standard work and self-employment is not secured, this will imply a social cost to be borne by these individuals and by the society, with higher costs to public budgets 50 .

Importantly, the main fiscal basis of social insurance remains, in many countries, standard work. An increasing share of people outside the systems may thus impair the fiscal sustainability of social security systems, an effect comparable in nature to the corresponding detrimental effects of population ageing and resulting increases in the dependency ratio 51 on their sustainability 52 . Annex 3 provides further information on the financing of social security systems and the challenges to their sustainability relating to the problem under review.

3.Why should the EU act?

3.1.Legal basis

The initiative "Access to social protection for workers and the self-employed" will support the Union's aims recognised in Art. 3 of the Treaty on the EU: to promote the well-being of its peoples, the sustainable development of Europe aiming at full employment and social progress, but also the aim to promote social justice and protection, equality between women and men and solidarity between generations.

Moreover, it will respond to Art. 9 TFEU, which states that "In defining and implementing its policies and activities, the Union shall take into account requirements linked to the promotion of a high level of employment, the guarantee of adequate social protection, the fight against social exclusion, and a high level of education, training and protection of human health". In addition, Art. 151(1) TFEU states that the Union and the Member States shall have as their objectives the promotion of "proper social protection".

Furthermore, the initiative will promote the aims of the Charter of the Fundamental Rights of the European Union as reflected in its dedicated article on Social security and social assistance (Art. 34).

Article 153(1)(c) and (2) TFEU enables the EU legislator to adopt measures in the field of social security and social protection of workers as long as it (i) avoids imposing administrative, financial and legal constraints in a way which would hold back the creation and development of small and medium sized undertakings, ii) does not affect the right of Member States to define fundamental principles of their social security systems; (iii) does not significantly affect the financial equilibrium of these systems. This legal basis can be used to address the challenges of access to social security and social protection for persons in non-standard employment. The Council is requested to act in unanimity.

Article 352 TFEU can provide the legal basis of an initiative related to the challenges of access to social security for people in different forms of self-employment. In this case the Council is also requested to act in unanimity.

The existing EU acquis related to the rights and obligations in relation to social protection is implemented through a variety of legislative and non-legislative provisions. However, none of these fully address the rights and obligations of people in non-standard employment and forms of self-employment. For an overview of available EU instruments, see Annex 8.

3.2.Results of the social partners consultations and of the public consultation

The European Pillar of Social Rights was jointly proclaimed by the European Parliament, the Council and the Commission on 17 November 2017. The consultations undertaken in the context of the Social Pillar had already seen many stakeholders express concerns regarding the obstacles faced by workers in non-standard forms of employment and by self-employed in accessing social protection such as unemployment benefits, sickness benefits, disability benefits, health insurance and pensions. Relevant contributions from other EU institutions, notably the European Economic and Social Committee, the Committee of Regions and European Parliament to the Consultation also highlighted issues in this area and, in the case of the European Parliament, called on the EU to act. 53  

Article 155 TFEU requires that labour and management are consulted on a range of issues concerning employment and social affairs set out in Article 153 of the Treaty, including the area of social security and social protection. Consultation of representative social partner organisations took place prior to the presentation of the proposal regarding access to social security of non-standard workers in accordance with Article 154 TFEU. The procedure of Article 154(2) TFEU is not as such applicable to Union action to address the challenges related to the self-employed on the basis of Article 352 TFEU. It was therefore on a voluntary basis that the Commission invited the social partners to share their views as regards the self-employed.

There was no agreement among social partners to enter into direct negotiations on any of the issues raised during the consultations While both sides broadly agreed with the challenges presented by the Commission during the consultations, they differ on the opportunity for EU action. On the side of labour, trade unions expressed support for a legislative action setting minimum standards at EU level. On the side of management, representatives of employers have underlined that the action should be left to Member States, given their responsibility on the design of social security systems. During the second phase of the consultation, all social partners agreed on higher transparency. Whereas trade unions call for full transferability, employers would like to limit transferability of minimum social protection rights and according to them there should not be an additional burden or cost of transferability.

All trade unions favour mandatory formal coverage, equalised to those of standard workers but national traditions should be taken into account. Employers referred to subsidiarity and expressed different views on the options. Some agree that social protection should be mandatory but employees should have the right to choose the form of coverage (public or private) and the freedom of choice for the self-employed is particularly highlighted. Other employers favour voluntary formal coverage in order to take into account the diversity of forms of employment of the heterogeneity within the self-employed.

The detailed outcomes of the consultation are presented in Annex 11.

In order to further support the preparation of the Commission's initiative, the Commission launched an open public consultation on 20 November until 15 January 2018, available in all languages. Targeted hearings with representatives of the self-employed, social insurance organisations and civil society and the Social Protection Committee (composed by representatives of Member States) were also organised in December 2017 (see also Annex 11 for further details).

3.3.Subsidiarity: Necessity and value-added of EU action

The problem of insufficient access to social security for a growing number of people in the labour market and its negative consequences on social cohesion, labour market dynamism and sustainable growth are wide-spread across Member States.

In principle Member States could take measures to improve the situation and several have indeed adopted reforms or are carrying national debates on the subject. In the Member States where reforms have been implemented, those reforms are steps in the right direction, but the evidence is that they are not comprehensive and systematic, i.e. they are not closing gaps in access for all groups and all branches of social security. Nor do current debates imply that all Member States necessarily plan to take action. Individual Member States may hesitate to regulate in this area on their own, as they may fear that such action could put their own companies at a disadvantage relative to companies from other Member States.

Action at EU level will support the implementation of principle 12 of the European Pillar of Social Rights so that all people in the labour market, regardless of their employment status, are protected against social risks such as unemployment or illness. Through a careful choice of the level of detail of the provisions, such EU action fully respects subsidiarity, irrespective of the specific legal instrument chosen. EU action can be a springboard for further national reforms and ensure that progress is not partial or uneven but is instead secured across groups, social security branches and Member States, creating a level playing field for economic agents in the internal market. The EU action can avoid short-term distortions of competition and ensure that all Member States move in the same direction at the same time promoting upward convergence, to the benefit the EU economy as a whole. In this respect, it is also worth noting that a special Eurobarometer survey on the Future of Europe shows strongly increasing support for more European level decision-making dealing with health and social security issues 54 .

EU action can also diminish the obstacles that the problem poses to the pursuit of the Union's objectives. Differences between employment statuses in access to social security can be detrimental the EU's fundamental objective of promoting full employment by hampering occupational labour mobility including towards higher productivity sectors. It can discourage workers from undertaking a certain type of employment or deter people from taking on the risks associated with self-employment, which thus hinder entrepreneurship and innovation. The problem can also be an obstacle to the achievement of the Union's objectives in relation to quality of jobs and poverty.

By enhancing convergence towards robust labour market institutions and social infrastructure, EU action can facilitate resilience, social cohesion and macro-economic adjustment within the euro area and beyond. As highlighted in the Five Presidents' Report on Completing Europe's Economic and Monetary Union 55 , in a single currency area, there is a need to build up the shock absorption capacity and labour market adjustment mechanisms of Member States.

4.Objectives: What is to be achieved?

4.1.General objectives

The general objective of the initiative "Access to social protection for workers and the self-employed" is to support all people who, due to their employment statuses, are not sufficiently covered by social security schemes and thus are exposed to higher economic uncertainty.

The initiative is a concrete way to fulfil the EU's objective to promote social justice and protection, as stated in Article 3 of the Treaty on the European Union. It further aims at implementing principle 12 of the European Pillar of Social Rights that states that 'regardless of the type and duration of their employment relationship, workers, and, under comparable conditions, the self-employed have the right to adequate social protection'. It also contributes to the implementation of other Pillar principles such as those relating to "secure and adaptable employment", "unemployment benefits" and "old-age income and pensions".

It will support Member States in adapting social security to present and future changes in labour markets. It will contribute to improving labour market dynamism by reducing labour market asymmetries in social security obligations and rights for different forms of employment. It will facilitate transitions between jobs, sectors and employment status by ensuring better preservation and transferability of acquired rights.

Its ambition is to foster sustainable growth as well as increase fairness, competitiveness and social justice by improving adequate access to social security for those without it, including many categories of people in non-standard employment and the self-employed, thereby addressing long-standing and new gaps and inequities in access. At the same time, ensuring that more people in the labour market contribute to the financing of social security systems would enhance their financial sustainability, contributing to intergenerational fairness and promoting social cohesion in the longer run.

4.2.Specific objectives

In order to reach the previous general objective, the initiative will have the following specific objectives:

-Closing formal coverage gaps: ensuring that legislation and collective agreement allow everyone in employment or self-employment to participate in the social security schemes.

-Addressing the problem of effective and adequate coverage and facilitating transferability of social security rights: ensuring that rules governing contributions and entitlements ensure the adequacy of protection and do not limit the accrual of entitlements and the access to benefits on the basis of employment status or type of employment relationship and do not distort the playing field between employment statuses or type of employment relationship. The initiative aims at preserving social security rights when workers or the self-employed change jobs, sectors of activity, type of employment relationship, or employment status.

-Increasing transparency on corresponding systems and rights: ensuring access to user-friendly information on rights and obligations to social security, irrespective of the employment relationship or employment status.

Formal coverage is a necessary, but not a sufficient condition to ensure adequate protection through social security systems. A person may be a member of a social security scheme without de facto being able to build and take up adequate entitlements to benefits and services, notably if the rules governing contributions and entitlements disadvantage some types of employment, employment statuses or career history. The rules should, on the contrary, promote social justice and a level playing field between employment statuses and types of employment relationship. Any difference between employment statuses or types of employment relationship regarding the rules governing entitlements should reflect the situation of beneficiaries and be justified by policy objectives. This objective concerns in particular the rules governing time thresholds, i.e. minimum qualifying period, minimum working period, waiting period and the duration of benefits. It also concerns the rules determining the assessment of the income assessment base.

If a person changes employment status, type of employment or economic sectors or if a person combines different jobs, the transferability and preservation of rights between the scheme available under the new status and the previous scheme covering the same risk should be facilitated. The preservation of rights should also be facilitated for persons combining different jobs and/or employment status at the same time. This is key in careers that are characterised by several job contracts and transitions. If rights are transferable, people will be more confident to make transitions, the labour market can be more dynamic and the allocation of resources more efficient.

Even when they have formal and de facto coverage, people may not benefit from social security if they do not have sufficient knowledge about their rights. If the rules governing social security schemes are transparent and if information about individual entitlements and obligations are easily accessible, individuals will be in a better positon to exercise their rights to social security.

Although these objectives are valid across all branches of social security, the issues have been found to be more acute in the branches that are more closely related to employment status, (unemployment benefits; sickness benefits; maternity and equivalent paternity benefits; invalidity benefits; old-age benefits; benefits in respect of accidents at work and occupational diseases).

5.What are the available policy Measures?

5.1.Baseline scenario

Under the baseline scenario the existing coverage gaps in social security would continue to exist across Member States and can be expected to lead to insufficient social security coverage for an increasing number of citizens.

On the one hand, external drivers like globalisation, transition towards a service economy, technological developments, demographic ageing outlined in section 2.2.1 are expected to persist. They will continue to rapidly change the European labour markets and make employment opportunities steadily more diverse. This implies that overall, employment on both old and new forms of non-standard contracts and in new forms of self-employment will likely constitute a steadily rising share of job-opportunities in the future. If observed recent trends continue, the share of traditional non-standard work in employment could rise from 25% in 2016 to 29% by 2030, while self-employed would remain broadly constant at 4 % (14% in 2016) 56 . For many individuals, most notably the young and the high-skilled, new forms of employment or own-account self-employment will represent an opportunity to be integrated in the labour market, expand choices and find a better work-life balance.

On the other hand, several Member States have taken steps in the recent past (see also Annex 7) to address some parts of the problem and change the internal drivers but overall progress is uneven across Member States, social security branches and targeted groups. Existing policies and planned reforms appear insufficient to prevent gaps in formal coverage, effective coverage, transferability and transparency in the absence of further EU action.

Over time an increasing share of the working population would thus be left without effective coverage under the baseline scenario.

There are a few relevant initiatives at EU level which address parts of the problem, yet a more comprehensive approach is lacking.

The existing EU guidance in the area of access to social protection 57 , particularly in the context of the European Semester and the Social Open Method of Coordination, will continue to provide recommendations for Member States' in this field. These efforts have yielded results in the past 58 by reinforcing cooperation between Member States and helping identify common objectives, but it is less likely that they could trigger the needed accelerated changes to social security systems through comprehensive action in the remaining areas, given their general nature and Member States' reservations that unilateral steps in some areas would entail risks for their competitiveness.

With the European Pillar of Social Rights, Member States and the EU institutions politically committed to strengthen the social acquis and deliver more effective rights to citizens. This political commitment should encourage Member States' actions and some other recent initiatives taken in the context of the Pillar can be expected to contribute only partially to solving the problem (see below e.g. for the Commission's proposal for a Directive on transparent and predictable working conditions in the European Union regarding transparency).

Regarding formal and effective coverage, the Directive 2010/41/EU 59 on equal treatment between men and women engaged in self-employed activity foresees the possibility to grant women (including spouses of the self-employed) the right to be entitled to maternity leave and benefits for at least 14 weeks. Since this Directive leaves to Member States decision if the right to such benefits should be granted on a mandatory or voluntary basis, levels of benefits and eligibility conditions vary widely among Member States. An independent expert evaluation concluded that the Directive was unsuccessful in delivering improved maternity and social benefits to the self-employed. 60  The Commission's proposal for a Directive on Work-Life Balance for Parents and Carers 61 provides for coverage by paternity and parental benefit and carers' benefit for all workers with an employment contract or relationship, including non-standard workers to whom the definition applies.

Regarding transferability, there is no requirement in EU law to ensure the transferability of occupational pension rights but the Directive 2014/50/EU 62 on the portability of supplementary pension rights encourages Member States to improve it and in fact, all Member States have opted to extend the same legal treatment to all workers (the Directive does not apply to self-employed persons) who leave an occupational pension scheme.

Moreover, several recent EU initiatives will contribute to addressing the problem of transparency. First, this concerns old-age benefits in the context of occupational pensions. This will be addressed by the Directive on minimum requirements for enhancing worker mobility between Member States by improving the acquisition and preservation of supplementary pension rights 63 (Portability directive) and by the recast Directive on the activities and supervision of institutions for occupational retirement provision (IORP directive) 64 . Both directives improve the provision of information on the pension schemes. The Portability directive also covers the protection of pension rights upon termination of employment 65 . The two directives should be transposed into national law by 21 May 2018 and 13 January 2019 respectively.

Second, the lack of transparency for other social security areas for workers will be partially addressed by the Commission's proposal for a Directive on transparent and predictable working conditions in the European Union 66 , but not for the self-employed. According to the proposal, workers will receive information on the social security institution(s) receiving the social contributions attached to the employment relationship and any protection relating to social security provided by the employer.

This legislative proposal also allows clarifying the boundaries between workers and self-employed bringing clarity on where the status of self-employed is unduly used (bogus self-employment). In relation to access to social security, such clarity helps ensure that those people erroneously considered self-employed can now have access to the available protection as workers (which is generally higher). However the personal and material scope of the two initiatives is different as the proposed directive does not cover the self-employed. Also, the proposed directive does not determine whether workers should have access to social security. It limits itself to secure that information on acquired rights is available to employees.

It could be argued that even without public intervention at EU or Member State level, the set-up of effective, efficient and accessible – in terms of prices/premiums - social security systems by social partners or professional organisation may solve the identified access problems. However, it is unlikely that such initiatives would lead to a comprehensive solution. Occupational schemes for pensions are indeed wide-spread in a number of Member States (including BE, DE, DK, FR, IE, NL, SE, SI and the UK), occupational funds may also play an important role in health insurance (e.g. IT) 67 and social partners may be involved in unemployment insurance schemes (e.g. DK, FR). However, there is no evidence of a trend increase in such schemes despite the existing problems. The challenges that have to be overcome are among others the need to mobilise a sufficient number of members to pool risks, while there has been a general decline in most Member States in the organisation of workers in trade unions and in trade union density.

Summing up, it is unlikely that the problem of insufficient access to social security will be comprehensively addressed by Member State action only with the support of the existing EU initiatives. As a consequence, it can be expected that the share of people in the EU who do not have access to social security will increase significantly under the baseline scenario.

5.2.Measures discarded at an early stage

When considering the policy measures, a number of approaches have been discarded:

Labour inspections: A non-legislative option to improve formal coverage that was discarded is increasing and improving labour inspections. Doing so could reduce the incentives of companies to replace employees with independent contractors not covered by social security who are in reality disguised employees to reduce the cost of labour 68 . However, that could potentially solve only a small part of the problem of lack of formal coverage and not solve the problem of lack of coverage for the majority of the self-employed 69 .

Universal basic income: In the face of rising non-standard employment and self-employment, one popular option could be delinking social insurance from employment by providing a universal social assistance benefit regardless of income, contribution history and employment status. Firstly, the focus of the current analysis is on the effects of employment status or contract type on the available protection through social security schemes. Coverage by social assistance and similar schemes is primarily based on criteria regarding residence and income, and thus they are out of the scope of the current initiative. Secondly, while this option may be feasible for countries that are in the process of setting up their social protection system, it seems to be more difficult to implement in countries that have already a well-established welfare system (as is the case for the EU Member States) 70 due primarily to strongly rooted traditions and preferences, the significant cost of switching from the old to the new model, and the extra cost associated with providing a comparably adequate protection for all 71 . Such an option would therefore very likely fail to garner the necessary political support for adoption and/or implementation. Therefore, the problem of insufficient access to social security is best addressed in the framework of the existing social security systems in the Member States.

Individual protection arrangements outside social security schemes: One option mentioned by stakeholders would be to restrict legislative action to demanding that all groups, notably including the self-employed, prove having some form of protection or insurance against the risks covered by social security systems, leaving to them the choice whether to arrange this via social security schemes or through purely individual private arrangements. First, private purely market-based arrangements are not available regarding all the risks under consideration (e.g. regarding unemployment), Second, where they exist, the solidarity manifested in collective schemes could be undermined, including by significant self-selection 72 where the corresponding potentially substantial costs could possibly weaken their financial sustainability. Possible measures and options therefore refer only to social protection concluded through an employment relationship which is coherent with the EU approach used in the Directive 2014/50/EU and Directive 2006/54/EC.

Benefits in kind: While benefits in kind account form an important part of the protection offered through some branches of social security, notably including health care, they are out of the scope of the current initiative. Regarding healthcare, an initial mapping did not provide sufficient evidence of significant differences in access related to labour market status. Similarly for other areas where benefits in kind are of key importance, the labour market situation is not expected to play a key role for access.

5.3.Description of the policy measures 

The objective to improve access to social security can be achieved by a combination of measures relating to three areas:

·Closing formal coverage gaps:

· Addressing the problem of effective coverage and facilitating transferability of social security rights

·Increasing transparency on corresponding systems and rights

The first dimension allows people to be member of relevant social security schemes. It is thus a prerequisite for the second dimension. We discuss measures for each of these areas. The policy measures retained are those that provide the alternatives that are most feasible (legally, technically and politically), coherent with other EU instruments, effective, relevant and proportional to tackle the problem drivers detected in section 2. In section 6.1, option packages will be built from the individual measures regarding the defined specific objectives.

As explained in Section 2.2.2, the personal scope (same or different measures to apply for workers and self-employed) considered across policy measures and options will be left general enough to allow the initiative to take into account new forms of work that could arise in the future. In terms of material scope, the measures would apply to the branches of social security which have been identified as most relevant of the Commission's initiative, namely unemployment benefits; sickness benefits; maternity and equivalent paternity benefits; invalidity benefits; old-age benefits as well as benefits in respect of accidents at work and occupational diseases.

The measures aim at eliminating or reducing obstacles which impair social security systems from providing people with adequate protection regardless of their employment status, while respecting Member States’ competences in design their social security systems.

Dimensions of the policy options left to Member States’ discretion in view of subsidiarity and proportionality and in respect of article 153(4) TFEU include i) the level of protection provided, ii) whether to extend coverage under existing schemes or by the creation of new schemes, iii) the way to finance the protection, which includes important design issues such as how to address the split of social contributions between employees and employers in the case of the self-employed and iv) combinations of schemes to be used (public, occupational or private). These issues are beyond the scope of the current initiative. Member States are best placed to decide about these dimensions in ensuring effective protection in the most efficient way given the diversity of existing social security systems across the Member.

5.3.1.Measures to extend formal coverage

Formal coverage can be addressed by granting access to social security schemes to previously excluded groups on a mandatory or voluntary basis. These alternatives may reflect different levels of ambition and may strongly influence the number of additional people that would be covered by social security as a result of this initiative.

i) mandatory coverage: extending coverage to groups of people previously excluded by virtue of their employment status or contract type 73 but also reviewing minimum income levels below which individuals in some Member States currently do not have access or only have access on a voluntary basis regarding certain branches of social security.

ii) voluntary coverage: previously uncovered groups would have the possibility to have join social security schemes, but the final decision on whether to use this option would be left to the individual.

The approach to coverage could be the same (either mandatory or voluntary) for all groups without access to social security (i.e. non-standard workers and self-employed) or could be differentiated between non-standard workers (mandatory coverage) and self-employed (voluntary) to reflect the differences of both groups regarding the allocation of risks, discretion over their work, decision-making power, financial autonomy and responsibility.

Combinations of the two alternatives could also be considered. Decisions on coverage could be linked to decisions on the level of protection (which is to be decided by Member States). In particular, a basic level of protection can be mandated and supplemented with additional coverage on a voluntary basis 74 . Such calibration of the mandatory element can also be motivated by the observation that uncovered individuals in most Member States do have access to non-contributory safety nets of last resort. Mandating a basic level of coverage close to the one provided by basic safety nets can thus also contribute to avoiding free riding behaviour of uncovered individuals and the corresponding costs to public budgets. Decisions on additional coverage, which are not associated with the same possible externality, are left to the individual 75 .

In their replies to the second phase social partners' consultation (see also Annex 11), all trade unions favour mandatory formal coverage, equalised to those of standard workers but national traditions should be taken into account. Employers expressed different views on these options. Some agree that social protection should be mandatory but employees should have the right to choose the form of coverage (public or private) and they highlighted in particular the freedom of choice for the self-employed. Others favour voluntary formal coverage in order to take into account the diversity of forms of employment of the heterogeneity within the self-employed. Many employers referred to subsidiarity that should prevail.

Regarding the open public consultation, around 2/3 of the respondents consider that social protection rights and obligations should be mandatory for every kind of job, irrespective of the type of contract. Only a small share of respondents think that social protection rights and obligations should be voluntary (5.9 % for non-standard work and 10.1 % for self-employed).

5.3.2.Measures to promote effective coverage and transferability 

Measures to address problems of effective coverage relate to the rules governing time thresholds (minimum qualifying periods, the minimum working periods, waiting periods 76 ) and to rules governing transferability. The latter concerns the two cases outlined earlier in Section 2.2.2: i) the build-up of entitlements over the course of a career characterised by transitions; ii) the preservation and transfer of entitlements from one scheme to a new one where there are minimum time thresholds to be respected.

During the second phase social partners consultation, views on effective coverage and transferability were mixed. Trade unions called for guaranteeing full portability by tying social protection entitlements to individuals through an equalised calculation and aggregation but without leading to an individualisation of social protection. On the other hand, some employers agree with the principle of transferability of rights but call for an impact assessment before creating legally binding realities or argue that subsidiarity prevails and that a mapping of existing transferable systems should be favoured. Others recommend the EU to limit itself to general provisions or would like to limit transferability to minimum social protection rights. The administrative cost of transferability should be kept at a minimum and no additional costs for the new employer to take on the provision of the transferred rights should occur.

Rules governing time thresholds

In schemes where minimum qualifying periods, minimum working periods, or waiting periods exist and are stricter for non-standard workers and the self-employed than for standard employees, or are the same as for standard employees, it can be more difficult for the two groups to fulfil them. Policy measures can adjust the minimum periods, so they are not longer than justified by policy objectives, including avoiding abuse. 77  

Time thresholds make the conditions for take-up of benefits more stringent and are used to minimise abuse, protecting the interest of the members of the scheme. The total elimination of time thresholds is therefore unrealistic because it would eliminate all disincentives to abuse embedded in the system. This hypothesis is therefore disregarded in our analysis below.

Modifications regarding qualifying periods, minimum working periods and waiting periods could be made for all insured people. Alternatively, they could be adjusted only for non-standard workers and/or the self-employed in the specific cases where the time thresholds have been identified as obstacles to reach the same social protection outcomes for everyone.

Rules governing transferability

i) Ensuring preservation, transferability and aggregation

Preservation of entitlements can support the build-up of entitlements over the course of a career where transferring is not possible (because for instance the person will not have coverage by another scheme after changing job or sector or employment status 78 ). The method of preservation may vary depending on the nature of the scheme. For instance, the value of pension rights may be adjusted in accordance with: inflation rate or salary levels (typically in a defined benefit scheme); return from investments derived by the scheme (typically in a defined contribution scheme). Preservation of entitlements can also mean that individuals also have the possibility of continue contributing to the former scheme 79 .

In cases where transferability is possible, then mandatory transferability ensures that all workers or self-employed leaving a scheme have the possibility to take their acquired rights with them. Conditions should also be set so that this transfer does not reduce substantially the entitlements (for instance due to unfavourable calculation of the transfer value or high administrative costs).

Aggregation of contributions period records within a given social security branch can support smooth transitions in the labour market. This implies that the past periods of employment or self-employment prior to changing employment status are taken into account as if the periods were completed in the current status 80 . Such aggregation should be made possible up to a limit justified by policy objectives, bearing in mind the need to preserve incentives for the functioning of the system.

ii) Personal social security accounts

This refers to disconnecting rights to protection through social security systems from professional status and connecting them to individuals, thereby providing better protection for people who change employers, jobs, sectors, employment status and even countries. Personal accounts can be set on the basis of existing schemes with a high coordination at national level among providers of all existing schemes or can be triggered through a fundamental revision of the social security system, implying a major reform effort to scale down the number of schemes related to specific employment statuses and occupational categories and to 'homogenize' their rules. Currently, only Ireland and Latvia have implemented personal accounts, while France is working to introduce them in its system 81 .

5.3.3.Increasing transparency on social security rights and obligations

During the second phase social partners' consultation, all social partners agreed with the need for higher transparency but they did not provide specific suggestions how this could be achieved.

Regarding measures to increase transparency, they could explicitly define what type of information would have to be provided, whether generic and/or personalised. In both cases they are expected to be free-of charge for the beneficiary.

i) Generic information on rights and obligations (broken down by employment status and contract type) could be provided, where not already done 82 , for instance via dedicated webpages relative to each relevant social security branch. The information provided should cover differences regarding the relevant groups, notably as regards standard employees vs. relevant categories of non-standard employees and the self-employed. The information should be easy to access and understand, and updated on a regular basis.

ii) Personalised information could be provided to individuals of their rights through dedicated webpages that could offer the possibility to calculate benefits (simulation tools) based on input on personal information such as income and employment situation. Other means to provide such personalised information would be via personal accounts where the personal information such as income and employment is already available and the calculation of benefits does not require any specific knowledge by the individual. Dedicated tools should be easy to find and use.

5.4.Impacts of the policy measures 

The assessment of impacts of policy measures 83 covers how the measures are likely to meet the defined objectives (effectiveness) and the key types of impacts, including economic and social impacts as well as those on national authorities. No environmental impact nor impact on fundamental rights has been identified.

The assessment of impacts combines qualitative arguments with quantitative results to the extent possible. A number of limitations need to be highlighted (see Annex 9 for other methodological considerations).

Firstly, the wide scope of the initiative addressing simultaneously different branches of social security in conjunction with the limitations in data available seriously constrains the possibility to provide a comprehensive picture. Available data only allow quantifications regarding some social security branches, and simulations are limited to the self-employed for the same reason. Furthermore there exists a certain asymmetry regarding the possibility to quantify costs versus – partially non-monetary – benefits. For instance, the positive impacts of reduced insecurity are hard to quantify, and neither do models allow combining micro-economic benefits of increased social security coverage with the more macroeconomic benefits of stronger economic stabilisation over the business cycle or potentially increased labour market transitions and dynamism.

Secondly, as explained earlier, subsidiarity implies that a number of key policy choices on how to improve access to social security are left to Member States, including decisions on the level of protection embodied in nations' social security systems, financing choices as well as the type of schemes through which coverage is provided. These choices fundamentally determine the qualitative and quantitative impacts that can be expected, impeding a more general or comprehensive illustration in this analysis.

Finally the evolving nature of the problem, as the rapid and fundamental changes in the labour market in the context of the new world of work described above, may also imply changes in the impacts. At the same time, it should also be noted that the choice of (non-binding vs. binding) legal instrument (see Chapter 8) may influence the degree of certainty with which the impacts materialise. The assessment of impacts is thus conditional on action being taken in response to the Initiative and could thus be regarded as an upper bound in this respect in case of a non-binding legal instrument.

5.4.1.Impacts of the measures extending formal coverage 

Effectiveness

Mandatory coverage has a very high effectiveness as, by definition, it implies coverage for all previously excluded. Mandatory coverage is a key means to implement the principle of solidarity, enabling also the most vulnerable categories to have access to social security schemes 84 . Moreover, mandatory coverage below a minimum level of benefit entitlement - corresponding to the one available from social safety nets - , can be justified by a need to limit that people would fall back non-contributory social safety nets.

Voluntary coverage can have a good effectiveness. Based on available evidence on enrolment in voluntary systems 85 it can be expected that a significant number of people would choose not to have coverage. At the same time, such an approach would respect the freedom of choice and reflects the heterogeneity of the self-employed, which were highlighted by key stakeholders during the consultation.

Choices on the financing of schemes - in the remit of Member States - can impact the degree of effectiveness, especially when access is granted on a voluntary basis. Indeed the way financing is set up can entail powerful (dis)incentives to participate. There are trade-offs between encouraging higher participation through public subsidies and public finance considerations. Financing also affects the degree of redistribution of income through the social security systems. For a given rate of contribution, incentives to enrol are higher for people who face the largest corresponding risks and are therefore most likely to benefit from the system, i.e. also generate a higher cost for it. Subsidies have successfully been used in some Member States to encourage broader membership and thereby reduce problems of adverse selection; however, such measures may also entail significant costs 86 . Alternatively, Member States could compel providers of social security schemes to offer their members identical obligatory benefits which do not depend on the amount of the contributions and by equalising risks and costs between different funds.

Where individuals are enrolled by default and need to take action to leave the system (through opt-out clauses), usually leads to higher coverage than opt-in measures (see Annex 10, Section 1).

Direct costs and other economic impacts

The direct cost of providing formal coverage depends on the number of people newly covered and the per-capita cost of coverage. The first and likely dominant component is clearly higher under mandatory coverage than under voluntary coverage, as not everyone will choose to be covered under the latter option. Differences in the per capita cost of coverage may arise due to problems of adverse selection, i.e. self-selection, under voluntary access. Resulting budgetary costs could be accompanied by savings from lower expenditure for social assistance and other safety nets that previously uncovered individuals would fall back on.

In addition to self-selection effects, moral hazard 87  could also influence the cost of extending coverage, notably regarding unemployment insurance for the self-employed, who have more control over the relevant risk than employees do in an employment relationship. While this is correct, the successful establishment of social security schemes protecting the self-employed in case of unemployment in the majority of Member States (the mapping in Chapter 2.2.2 only established corresponding formal coverage gaps in nine Member States) implies the potential for learning from peers about the necessary control mechanisms and rules which minimise this kind of behaviour. A recent study of the Danish unemployment insurance for the self-employed indeed finds a limited risk of moral hazard 88 .

Regarding formal coverage of the self-employed, simulations 89 (see also Annex 10 for all details) were undertaken for the extension of existing schemes of unemployment benefits for employees to the self-employed under the same rules in five Member States 90 . Results suggest that the cost of extending coverage in terms of the corresponding total expenditure for benefits could be in a range from below 0.1% to 0.4% of GDP annually, depending inter alia on the number of self-employed in the country, the unemployment risk they face, as well as on the generosity of the system in place. Comparing the simulated cost of a similarly generous unemployment insurance system for the self-employed across all Member States results in similar limited costs, albeit with higher cross-country variation.

Expressing the costs as a share of gross labour income of the self-employed, to be interpreted as contribution rates to be paid by them in a revenue-neutral scenario, results in contribution rates between less than 0.1% and up to 1.7% of gross labour income, in line or lower than contribution rates observed in many existing systems in Member States. Simulating costs for extending sickness benefits also results in limited costs in most Member States, with a significant degree of cross-country variation.

At the same time, depending on Member States' financing decisions, extending formal coverage can also avoid that the financing base of social security systems is skewed because an increasing group of people on the labour market is outside the system and does therefore not pay contributions.

Similarly, other choices about the nature of schemes may have implications for their cost. For instance, under voluntary coverage giving a choice between possibly more redistributive public and more individualised occupational schemes may be costly for the public scheme due to self-selection issues. The creation of new, separate schemes specifically designed for those not currently covered can weaken the solidarity ties between the insiders (in the old schemes) and outsiders (in the new schemes).

At the macro level, benefits (not simulated) from the extension of coverage can also be expected by strengthening the well-established key role of automatic stabilisers of public social insurance schemes', dampening business cycle fluctuations 91 . These help limit the loss of economic efficiency resulting from volatility in the economy, contributing to enhanced economic performance to the extent that large output fluctuations can, notably in the absence of wage flexibility, result in a trend increase in unemployment (hysteresis effects) and erode human capital.

A noticeable reduction in unfair competition between firms employing staff on non-standard contracts exempted from social security contributions and competitors not reducing costs in this manner could be expected from the extension of formal coverage. Such effects would be very pronounced under mandatory formal coverage for all groups while a voluntary approach would reduce but not eliminate possibilities for such arbitrage via contract forms.

A differentiated approach extending coverage on a mandatory basis to non-standard employees and on a voluntary basis to the self-employed could incentivise a movement towards the latter. In this respect, it is important to note that the enforcement would be facilitated by the recent proposal for a Directive on transparent and predictable working conditions, which allows to clarify the boundaries between workers and self-employed brings clarity on where the status of self-employed is unduly used (bogus self-employment). In relation to access to social security, such clarity will help ensure that those people erroneously considered self-employed can now have access to social security as workers.

Depending on financing decisions taken by Member States, an increase in labour cost associated with the extension of coverage could lead to a reduction in labour demand for the concerned groups, while work incentives for people paying contribution rates can also be reduced. In conjunction, both effects could imply a reduction in employment.

Finally, positive effects on labour market dynamism of removing the lack of social security coverage as an obstacle to transitions between self-employment, non-standard employment and standard employment could reasonably be expected. Labour market segmentation would be reduced.

Social impacts

Key expected impacts stemming from the improved protection would be an increase in the mutualisation of risk, increase of income security resulting in less precariousness and poverty. Given the higher level of effectiveness of mandatory coverage, this approach is also expected to bring higher social impacts.

The above-mentioned simulations regarding unemployment insurance show that there are noticeable reductions in the risk-of-poverty for the previously uncovered group of self-employed (for more details, see Annex 10). Coverage reduces the poverty risk among all self-employed noticeably (by up to 13 pps in the simulations) and softens especially the impact of job loss on poverty at household level (by up to 25 pps for household with former self-employed unemployed), even if the corresponding effects at the aggregate level are small.

It should be noted that to the extent that women currently are more disadvantaged by social security systems still geared towards linear full-time employment, the improved protection through the extension of mandatory formal coverage but also better effective coverage and transferability for non-standard employees would be especially beneficial for them.

5.4.2.Impacts of measures promoting effective coverage and transferability

Adjusting time thresholds

Effectiveness:

Time thresholds make the conditions for take-up of benefits more stringent thereby limiting inter alia fraud and budgetary costs. Nonetheless the time thresholds can be adjusted to a minimum justified by policy objectives while preserving features in the systems that avoid abuse 92 . This adjustment can significantly reduce the barriers for non-standard workers and the self-employed to the take up of benefits in case of need.

Direct cost and other economic impacts:

This measure would have a direct cost corresponding to the payment of benefits to persons who were previously unable to meet the time requirements. As this relates to persons who were already formally covered by a scheme and thus contributing to it, no increase in revenue would offset these benefit payments. At unchanged contribution rates, it can therefore result in higher costs for the provider of the scheme, notably including an impact on public budgets in the case of statutory schemes. Budgetary costs could be partly offset by the savings resulting from lower outlays on social assistance and from higher consumption-related tax revenue resulting from the improved effective coverage 93 .

As an example, the effects of relaxing qualifying periods regarding unemployment insurance (for all potential unemployed) have been quantified in dedicated microsimulations (see Annex 10, Section 3). In the simulations, the net costs for the unemployment benefit, going up by 0.13% of GDP. The initial budgetary costs are reduced due to increased tax revenues related to household consumption and positive effect on the welfare of individuals.

As regards other economic impacts, carefully choosing the extent to which time thresholds are adjusted should limit – but would not avoid - further indirect costs potentially resulting from behavioural changes in response to the adjusted requirements. Such costs could notably stem from reduced incentives to work associated with less stringent and more generous unemployment insurance.

Social impacts:

The measure is expected to have social impacts similar to those observed for formal coverage, resulting in a reduction in the individualisation of risk, income uncertainty, precariousness and notably a lower risk of poverty. The above mentioned simulations show a positive impact on disposable household incomes on the lowest quintiles of the income distribution and would significantly decrease the poverty risk for the group of the unemployed (see Annex 10 for details).

Ensuring preservation, transferability and aggregation of entitlements

Effectiveness

Preservation of rights can be expected to be moderately effective since it will at least allow individuals not to lose their accumulated entitlements and to preserve their value but the individual would have to rebuild his entitlements in the new schemes and re-start a contribution history. Here there is an interaction with the time thresholds discussed previously. Regardless of past contributions, the individual would have to wait before taking up benefits in the new scheme

The aggregation of past contribution records would have a good degree of effectiveness in ensuring that people are covered in the short term following transitions.

Mandatory transferability would have a high degree of effectiveness as it would allow individuals not to lose their previous accumulated entitlements when changing job or employment status in a way which entails the transition to a different social security scheme and at the same time enable them to continue contributing to social security. There is an interaction with the measures regarding effective coverage discussed above, to the extent that adjusting time thresholds (e.g. waiting or qualifying periods relating to unemployment or sickness insurance) reduces the need to transfer elements related to the past membership into another scheme. The lower the time thresholds are, the lower are the barriers to transferability across schemes.

The introduction of personal accounts, disconnecting rights to protection through social security systems from professional status and connecting them to individuals, would ensure coverage while reducing the need for transferability (and its related costs) between schemes that are tailored to specific employment statuses. In practice, entitlements would be attached to the individual regardless of the number of labour market transitions and coverage would not be impaired by the need to start a new contribution history at a change of a job for example.

Direct costs and other economic impacts

Preservation of entitlements is not expected to entail particular direct costs.

Aggregation of past contribution periods would impact benefit payouts between schemes in situations where people changing employment status are covered by different and financially independent social security schemes before and after the change. Setting up some protection of the financial interests of the "recipient" schemes would thus be necessary. This would relate to cases where most of the minimum period is accumulated – and the corresponding contributions paid- in the scheme under the old employment status, but the risk materialises soon under the new employment status, with the corresponding costs accruing to the associated "recipient" scheme. This would imply either requesting a certain minimum length of attachment to the new scheme i.e. setting a time threshold 94 , extending the coverage of the schemes by a few months after the person stops contributing or setting up a coordination mechanism for an allocation of contributions and costs across schemes. While the first alternative runs counter to the underlying objective of ensuring coverage also in the initial phase after a change of employment status, the second and third may entail larger costs associated to the administrative complexity.

Additional indirect costs related to the aggregation of past membership periods in other statuses could result from potential behavioural effects. As discussed earlier, minimum periods are in place to align people's behaviour with existing systems and avoid abuse. Policy objectives may thus set a limit to transferring past periods, like shortening minimum periods, to preserve the right incentives to the insured.

Indirect positive effects on the labour market can be expected from mandatory transferability. Transferability can be expected to facilitate labour mobility including between sectors as well as transitions between employment and self-employment. This would contribute to increased labour market dynamism and a better allocation of resources.

The introduction of personal accounts are expected to have the same positive effects on labour market dynamism as an approach securing accumulation transferability and aggregation but it would also entail increase the costs for public authorities and insurance providers as it would imply important changes in the overall coordination of existing social security systems. Estimating the costs of such a system change is not possible with available tools.

For all options measures above, it is important to consider the starting point of Member States. In countries with few or rather homogeneous schemes, the preservation aggregation and transferability of entitlements is likely to entail less administrative costs needed for the coordination between schemes than in countries with a significant number of diverse schemes and a rather complex administrative environment. The more homogeneous the conditions d the schemes are, the faster the transferability of entitlements is likely to be, re-assuring individuals and supporting their dynamism in the labour market.

Social impacts

Key social impacts should be qualitatively similar to the ones regarding measures on effective coverage, notably including reductions in precariousness and the risk of poverty since they will facilitate protection for those who alternate and/or combine different employment and self-employment statuses. Their importance can be expected to increase to the extent that careers are increasingly characterised by frequent transitions, with positive consequences on employment levels, competitiveness and productivity.

An important negative social impact of connecting the rights to individuals via the introduction of personal accounts could be a reduction in the solidarity and redistribution function of protection provided through social security systems and an associated increase in the individualisation of risks 95 .

5.4.3.Impacts of measures promoting transparency

Effectiveness

The effectiveness of measures would depend on the degree of specificity of the information provided, on the frequency at which it is available and on the fact whether individuals need to take action to obtain the information.

The provision of generic information alone would be moderately effective as it would not reflect the specificities of individuals' situation and action by individuals would be required to access the information. Personalised information is expected to be more effective. The frequency at which this information is provided would affect its effectiveness. If provided via online simulation tools or personal accounts, the information could be accessed at all times. The provision of personalised information could moreover have a strong impact on encouraging individuals to provide for their social protection, especially if the protection is not mandatory. It may thus help achieve the objective of coverage especially in voluntary schemes.

Direct costs and other economic impacts

The provision of generic information is likely to entail low direct costs.

The cost of providing personalised information through dedicated websites and tools would consist of a one-off cost and further costs on an ongoing basis for the necessary updating and maintenance. The costs of providing personalised information by default have been assessed to support the proposal on the activities and supervision of institutions for occupational retirement provision. The one-off cost is expected to amount, on average, to around 7€ per member 96 .

The costs of personalised information could be reduced by using digital solutions instead of paper. However, this should not be to the detriment of vulnerable groups that might lack access to such solutions and should take into account the needs of people with disabilities.

Importantly costs incurred in providing transparent information via new technologies are strongly influenced by the departing positions of Member States. In the maturity of IT environments and e-government solutions.

Personalised information is also expected to generate economic benefits as more people would be encouraged to save and the reliance on safety nets is expected to decrease.

Social impacts

An increase in transparency of social security rights and obligations is likely to entail positive social impacts. Awareness on the conditions to take up benefit is likely to increase. The perception of insecurity vis-a-vis the real protection provided by social security (see Annex 16 for more evidence on perceptions) is likely to decrease, with positive impacts on confidence indicators and higher perception of social fairness.

Increased transparency will also bring more accountability of the organisations administering social security schemes towards their contributors, here again having positive impacts on the general perception of social fairness.

6.What are the Policy packages and their impacts? 

6.1.Description of alternative policy packages

Policy measures described in section 5.3 are now grouped in alternative policy packages. To address all objectives of the initiative, each policy package must necessarily include measures relating to all the areas, notably formal coverage, effective coverage & transferability as well as transparency. Measures are grouped together in a way that allows them to be coherent and mutually reinforcing in achieving the objectives of the initiative.

Only working in combination can the three objectives promote adequate protection. As previously pointed out, formal coverage is only a prerequisite for benefitting from the protection granted by social security schemes in case of need. If not accompanied by measures ensuring the build up and take up of adequate entitlements, adequate and effective protection will not be achieved. Likewise, omitting measures on transferability would imply a lack of protection in the presence of transitions between employment statuses. The need for transparency is the highest when an individual has over time access to several schemes with differentiated rights, some mandatory others voluntary, due to a high number of employment transitions. Importantly, the lack of transparent information may hinder people's ability to be aware of their rights and thus benefit from the protection they may grant.

Package A is the baseline in the absence of new EU action.

Packages B and C differ with regard to the ambition relating to the formal coverage of the previously uncovered groups, where package C entails formal coverage for all and package B provides for a differentiated approach (mandatory coverage for non-standard employees and voluntary coverage for the self-employed), also in reaction to the emphasis stakeholders put on the specific characteristics and heterogeneity of the group of the self-employed.

It should be noted that establishing mandatory coverage for the self-employed is a realistic option. Based on the mapping presented in Chapter 2.2.2, there are 16 Member States where social security coverage is mandatory for the self-employed for at least some of the branches under consideration. For each one of these branches, there exist relevant examples of existing mandatory schemes for this group in at least a few Member States. The suggested option is thus in line with realities in the Member States.

The choice of measure for formal coverage has implications for the choices regarding the other measures in the packages. This is because – as mentioned above - formal coverage is a prerequisite for the other dimensions, notably for effective coverage and transferability. It is for this reason that only package C, the more ambitious package regarding formal coverage, maximising the number of people for whom effective coverage and transferability may become relevant issues in the first place, also includes the provision of personal accounts to facilitate transferability. Personal accounts tie rights to the individual rather than his employment status, requiring comprehensive changes to the set-up of current systems, and can thus be considered the more ambitious measure in this respect. Indeed individuals will only benefit from the introduction of personal accounts if they do have social security rights in the sense of being formally covered. Thus for package B which only extends formal coverage on a voluntary basis to the quantitatively important group of the self-employed, the introduction of personal accounts and the corresponding necessary changes to the set-up of social security systems would seem disproportionate. Package B thus aims to facilitate transferability in a more proportionate way, via changes to rules in the existing systems, notably establishing mandatory transferability of accumulated entitlements and allowing the aggregation of periods upon changes between schemes.

At the same time, it should be noted that regarding effective coverage, both packages propose the same measure – adjusting (to the extent compatible with policy objectives) time thresholds that have been identified as obstacles in this respect. Precise levels or thresholds for such adjustment cannot be fixed at EU level, limiting the scope for differentiation between the two packages. Firstly, such levels would depend on the precise set-up of the system and interactions with other parts of the tax-benefit system. Secondly and relatedly, this is due to subsidiarity regarding the system design.

Finally, it is logical that higher ambition regarding measures to improve formal and effective coverage as well as transferability reduces the needs for transparency – the less differentiated the corresponding rights are, the less need there is for personalised information. For this reason, package B (which entails a higher degree of differentiation in rights) involves requirements for the provision of generic and personalised information on social security rights and obligations. In package C, no measures to provide personalised information would need to be taken separately – but indeed such information will be readily available as a by-product of the introduction of personal accounts, without further efforts.

Table 1: Mapping of policy measures and policy packages

1 No dedicated measures are required for this, as personalised information would become available as a by-product of the introduction of personal accounts.

6.2.Impacts of the policy packages

The following assessment does not repeat the assessment of the individual measures undertaken in Section 5.4. Instead, it describes the joint impacts, notably highlighting interactions between the different measures grouped together both in terms of synergies and trade-offs.

Package A / Baseline Scenario: This option does not address the problems, but represents the baseline, as set out in section 2.4, with which other options can be compared.

Package B: The package would lead to high effectiveness regarding protection and thus pronounced positive social impacts while entailing corresponding costs for public authorities and for employers. The differentiated approach fully respects the freedom of choice of the self-employed, and thus fully takes into account the nature of self-employment while reflecting the significant heterogeneity among the concerned group. The provision of generic and personalised information improving transparency could stimulate higher voluntary enrolment, at a relatively small cost for public authorities. The package could make a noticeable contribution to supporting labour market transitions and dynamics as it helps ensure the build-up of entitlements along careers associated with changes between social security schemes while also providing for coverage in the short term following such changes. This may help improve the allocation of resources.

Package C:

This package has a very high effectiveness by ensuring all on the labour market have formal coverage. The introduction of personal accounts usefully complements this by also providing protection following professional transitions, again with very high expected effectiveness. In thus reducing the differences in protection between groups, the package limits the needs for the provision of information, although the introduction of personal accounts would also entail the provision of personalised information as a by-product. As regards effective coverage, it is identical to Package C.

Table A.14 in Annex 10 summarises the key impacts of the packages in a comparative manner.

6.3.Comparing the policy packages

The table below summarises the results of comparing the baseline scenario to the two alternative packages B and C. The comparison is based on elements discussed in more detail for each policy package in the sections above. A seven-stage qualitative grading system is proposed on the basis of the analysis of impacts in previous sections as follow:

Both packages B and C achieve an overall good level of effectiveness:

·Formal coverage: There will be a substantial increase in the number of previously uncovered people in the case of Package B. The extent of the increase in the number of self-employed covered by social security depends on how many people enrol in voluntary schemes, while non-standard workers will be mandatorily protected. The increase in the number of people protected is expected to be strong in the case of Package C, since everyone will be covered on a mandatory basis.

·Effective coverage and transferability: the two packages have the same requirements for what concerns the adjustment of the time thresholds which determine the qualification for benefits and are thus expected to have the same substantial level of effectiveness in this respect. The creation of personal accounts in Package C would effectively solve all problems of transferability by de-connecting rights from employment status and linking it to individuals. Establishing rules facilitating transferability in the existing system as per Package B is also expected to have a good effectiveness in this respect.

·Transparency: The provision of personalised information as per Package B would have a substantial effectiveness in increasing transparency. In Package C, the creation of personal accounts would entail the provision of personalised information available at all times as a by-product without further action, achieving a strong degree of effectiveness.

Social impacts are expected to be strongly positive in Package C. Mandatory coverage together with full transferability through personal accounts and aggregation of contribution periods strengthen the ability of people to effectively access social security benefits in case of need, contributing to the corresponding expected reduction of poverty among people in non-standard employment and among the self-employed. Positive impacts for the self-employed under Package B depend on enrolment rates; overall positive impacts could be substantially positive.

The direct costs of the packages comprise the cost of providing the benefits as well as costs related to improving transferability and transparency. As previously explained, no comprehensive quantification of the cost of the Packages is available. However, it can be expected that the corresponding cost is high in Package C entailing mandatory coverage for both groups and the creation of personal accounts, which would require important costs to be set up. For Package B, the cost associated to extending coverage is expected to be lower because through a voluntary approach for the self-employed, the number of individuals to be covered will be also lower.

Beyond the socio-economic criteria discussed above regarding the impacts of the packages, there are criteria of political nature against which the packages are assessed. Firstly, considerations on subsidiarity and proportionality are to be made. While both Packages respect subsidiarity since decisions of financing and design are left to the Member States, the mandatory approach of Package C for both groups leaves a low degree of flexibility in tailoring choices of formal coverage to the specificities of the self-employed. Given the special characteristics of and heterogeneity among the self-employed, a voluntary approach suggested in Package B appears more proportionate than Package C.

Secondly, the stakeholders' feedback gathered in the preparation of this impact assessment does not indicate a clear consensus among stakeholders, and in particular among representatives of the self-employed, for the strictest approach to formal coverage of the self-employed. Indeed views are mixed and several stakeholders recalled the importance to leave freedom of choice to the self-employed, given their heterogeneity and the current limited knowledge base capacity to fully assess how the problem is likely to evolve. Measures proposed in Package B are thus assessed as representing a good compromise between the different views expressed. As Package C does not take into consideration the arguments presented by some employers and by some of the self-employed representatives; it is assessed as satisfactory in this respect.

Finally, the analysis of coherence with other EU policies takes into account in particular the European Pillar of Social Rights. The two packages, which aim at a comprehensive improvement of coverage and also address comprehensively the issue of effective coverage and transferability correspond closely to the Pillar's ambition. Coherence with the EU2020 objectives related to poverty is very strong under mandatory coverage for all (Package C), where a high reduction in poverty can be expected and coherence is thus assessed as good regarding Package B.

7.Preferred Package

7.1.Preferred policy option

The preferred policy option is policy package B. This package is the one that best balances achieving the policy objectives with the related costs, with proportionality and with political considerations notably including stakeholder feedback and subsidiarity.

By striking a balance between views expressed in the formal social partner consultation and taking into consideration the specificities and heterogeneity among the self-employed, the preferred option represents a realistic and proportionate set of measures appropriate to contributing to a realisation of the ambitions of the European Pillar of Social Rights.

It is important to notice that voluntary schemes can be set up with different mechanisms to nudge individuals to adhere to the schemes. Opt-out clauses imply that the individual is by default included in the scheme but he/she can opt to move out of the scheme. Opt-in clauses imply that the individual is by default outside of the scheme but he/she can opt to adhere to the scheme. In both cases the individual has the freedom to choose whether or not he wants to be a member of a given scheme. Voluntary schemes with opt-out options are usually more effective in providing protection (see Section 1 in Annex 10).

Given the nature of self-employed work, a mixed approach may be considered. For those risks that are more difficult to evaluate due to asymmetry of information and in absence of a counterpart that verifies the information - as the employers for the dependent workers - opt-in clauses might be the most appropriate. These could apply to schemes such as unemployment and accidents at work and occupational injuries. On the contrary, an approach that nudges more effectively self-employed to be covered may be considered for schemes where the asymmetry of information is less problematic. Opt-out clauses may be considered for risks such as pensions, invalidity sickness and maternity.

7.2.Overall impacts

On workers and the self-employed

There will be an increase in the number of people protected, which will decrease the number of non-standard workers at risk of poverty and to some extent of the self-employed (depending on how many will enrol in voluntary schemes). There could also be a general moderate reduction in income inequality via impacts on the self-employed and the non-standard workers, although the improvement would depend on how many self-employed will enrol in voluntary schemes and on the hiring preferences of the employers.

Depending on Member States' financing choices, if part or all of the cost of extending coverage is left to the newly-covered groups, there would, as is usual in insurance schemes, be some income redistribution between those only bearing part of the cost and those for whom the risks and associated risks materialise. 97 At the same time, non-standard workers could experience a reduction in their disposable income if the number of the self-employed that enrols in voluntary scheme is low. A potential push towards (bogus) self-employment to avoid the payment of social security contributions would be limited by the clarity on the boundaries between workers and self-employed offered by the Commission's proposal for a Directive on transparent and predictable working conditions, which would help enforcement in this respect. Again depending on Member States' financing choices, there could be a limited loss in employment for affected groups in reaction to a potential increase in labour cost

Improved effective coverage and transferability will likewise result in a better protection of people in non-standard work and self-employment, including after transitions in the labour market which entail a change in social security scheme. Moreover, people in non-standard employment and self-employment will be able to maintain and use their accumulated entitlements.

Receiving personalised information will help them to know which rights they are entitled to. and will allow them to better prepare for the risk. Transparent information will also stimulate the participation of the self-employed in voluntary schemes or their choice to increase contribution rates in schemes allow to do it.

On companies

Depending on financing choices, companies are expected to experience an increase in the cost of managing fluctuations in production, in line with some labour cost increase related to non-standard employment and the self-employment. The increase in cost should be reduced in relation to non-standard employment because most categories of non-standard workers are already covered by social security schemes, so that the share of those to be mandatorily covered is small compared to the totality of the workforce. The cost increase due to hiring services of the self-employed will depend on how many self-employed will decide to enrol in voluntary schemes and on the extent to which the self-employed will pass on their cost through higher prices.

On the other hand, companies could benefit from an increase in productivity of individuals in the newly protected forms of employment, from reduced unfair competition and from the positive effects of social security systems on labour markets.

Modification of effective coverage, transferability and transparency do not impact on companies at large.

On Member States and insurance providers

The exact financial impact will depend on i) how many unprotected people are present in the Member States and ii) how the new schemes will be designed or the old schemes extended, combining public and non-public schemes, as well as on financing choices. For instance, if Member States will finance the schemes through taxes, public expenditure would increase, but this could be partially offset by reduced expenditure on last resort safety nets (e.g. social assistance) via a reduction in the number of people relying on them.

Adjusting thresholds to improve effective coverage could result in increased expenditure for (public or other) insurance providers, without corresponding changes on the revenue side. Establishing a set of rules for transferability and the protection of financial interests of schemes will also entail some costs and some redistribution between the schemes which are losing participants and those who are gaining. On the other hand, increasing the number of people protected and assuring that they are effectively protected in case of transitions can improve dynamism of national labour markets, the allocation of labour resources and the competitiveness of the economies. This can be expected to lead to lower recourse to benefits and thus have a positive impact on the financial sustainability of the insurance providers.

Concerning the provision of generic and personalised information, there would be an initial cost to set up the system but it would be followed by a limited cost for maintenance. These costs are expected to be offset by the positive impacts that transparency can bring on accountability and good governance of social security schemes. Furthermore, this could lead to improved social perceptions and sense of security, contributing to higher levels of consumer confidence, higher savings, more labour market dynamism and higher productivity.

8.The choice of legal instrument

8.1.Legal options

The legal instruments considered for the initiative are a Council Recommendation and a Directive; there could also be a combination of the two. The possibility of an increased focus on the issue of social protection in the framework of existing tools, notably the European Semester and the Open Method of Coordination, is also considered. A careful choice of the level of detail of the provisions would ensure full respect of subsidiarity, irrespective of the specific instrument chosen. Subsidiarity is thus not an argument for determining the choice between the different legal options since it is compatible with the considered legal instruments.

Article 153 (2) TFEU provides the possibility of adopting a Directive in the area of 'social security and social protection of workers' to set binding minimum requirements for gradual implementation in the Member States. Article 352 TFEU provides the legal basis of an initiative related to self-employment.

A directive could set binding minimum requirements, reflecting the ambition of the corresponding principle of the European Pillar of Social Rights and de facto establishing rights. A directive would bring a higher predictability of implementation while leaving some discretion to Member States to transpose according to their national specificities. Given its binding character, a directive would have a high degree of effectiveness in reaching the objectives of the initiative. Taking into account the national competences in the area of social security as foreseen in Article 153 (2) TFEU, a directive could establish clear policy goals but would not be able to provide detailed guidance on the specific design and financing of social security systems.

Alternatively, the Commission may adopt a proposal for a Council Recommendation providing guidance to Member States on how to achieve the policy goals set. Although such guidance would not be binding, it can nonetheless provide a basis towards a level playing field between Member States on the way they will address the gaps of access to social security. A Council Recommendation would be based on Article 292 TFEU in combination with Articles 153(1)(c) and 153(2), third subparagraph, as well as with Article 352 TFEU.

While a Council Recommendation may have a somewhat lower degree of effectiveness - as it would not give the EU capacity to legally sanction a Member State nor give citizens the possibility of appealing to court in the case of no implementation - the non-binding character of a Council Recommendation would also allow the EU to leave maximum flexibility to Member States on the way to achieve the goals. Reporting commitments can be included in the Recommendation itself and policy coordination instruments such as the European Semester can be used to promote progress.

An alternative to a stand-alone Recommendation or Directive would be an increased focus on access to social security in the context of the existing policy framework and instruments in the field of social protection, namely the European Semester, its country-specific recommendations (CSRs) and the Open Method of Coordination (OMC) on Social Protection and Social Inclusion. These instruments have already been used to stimulate dialogue at EU level on how to improve the access to social security for non-standard workers and the self-employed. However, the topic has not been covered in a systematic way in the European Semester – where it competes with a number of other key priorities - or on a regular basis in the OMC and guidance is usually limited and expressed in general terms. Consequently, only one country, the Netherlands, received a Country Specific Recommendation (CSR) on access to social protection for non-standard workers in 2017 98 and only few peer reviews were organised on this topic within the Social Protection Committee. This is partly due to the fact that the common objectives and indicators on access to social security gave little attention on how such access has evolved in function of changes in labour law and types of employment.

8.2.Preferred instrument

The preferred instrument is a Council Recommendation as it is regarded as the most effective and proportionate instrument at this point in time.

The consultations carried out in 2017, with social partners, relevant stakeholders and general public (see also Annex 11 for more details) indicated that there is no general consensus on the best instrument to be used. Some stakeholders underlined the stronger effectiveness of a directive; many others, including many Member States, showed little support for a legislative proposal, taking into account the different national points of departure, traditions, political preferences and budgets. Although a directive would likely to be more effective from a strictly economic point of view in reaching the objectives set out for this initiative, it is doubtful that it would gather support from all Member States, making it the less effective instrument overall.

Furthermore, the problem identified is of an evolving nature. The debate on access to social protection and the future of work is ongoing in several Member States and it will be influenced by ongoing reforms undertaken at national level. Certain aspects of it will also be influenced by the potential impact of some ongoing initiatives at EU level (including the Commission's proposal for a Directive of the European Parliament and of the Council on transparent and predictable working conditions in the European Union or the directive on pensions portability). The problem is however too important to postpone action waiting for the full materialisation of national reforms which, on their own and without general common guidance, may lead to uneven progress across groups (the self-employed vs. non-standard employees), social security branches and Member States.

Finally, the current statistical limitations make it difficult to fully monitor the evolution of the problem and such limitations – compounded with a number of choices left to the Member States such as the level and way of financing the action - have also strongly influenced the capacity to fully quantify the impacts of an EU action..

For the reasons explained above – the evolving nature of the problem, the ongoing reforms in some Member States and the strong statistical limitations – a Council Recommendation is a more proportionate approach than a Directive (or a Directive for certain elements of the proposal in combination with a Recommendation for the rest) which would impose binding outcomes.

The key added value of a Recommendation at this stage is to create momentum supporting and complementing national debates and reforms and guiding Member States' efforts in this. The Recommendation allows the EU level and the Member States to work further to understand the phenomena, its implications at national level and to stimulate and guide the reforms at national level. Importantly it will ensure that progress is not partial or uneven across target groups and social security branches. A focused approach is needed to give the issue the necessary political visibility, raise awareness and build momentum. This would help ensure that all Member States move in the same direction at the same time, thus promoting upward convergence to the benefit the EU economy as a whole. Indeed efforts are needed from all Member States in a coherent way and also in a somewhat simultaneous fashion to dispel doubts about possible first mover disadvantages.

A Council Recommendation can also contribute to create consensus on the best reform options, which is, for the time being, lacking at national and European debate, and is however necessary to design and implement reforms. The stakeholders' consultations (see below) reveal substantial divergence on the best instruments, as well as the direction of reforms.

Several examples from the recent past show how Council Recommendations can be effective instruments in achieving progress in the social field, notably the Youth Guarantee 99 and the Recommendation on Long Term Unemployment 100 . In particular, the Youth Guarantee, together with the Youth Employment Initiative, was instrumental in encouraging Member States' to step-up their efforts in this field. The Council recognised "the work done by Member States to mobilise additional financial resources, to strengthen and modernise their labour market institutions, and to support partnerships between the employment, education and youth sectors, as well as to engage with social partners, in line with national practices, and other stakeholders, in order to better deliver the Youth Guarantee." 101

The Recommendation could usefully be embedded in the existing processes such as the European Semester or the Social Open Method of Coordination, where Member States are already intensively involved. By stepping-up existing efforts to improve the knowledge-base and by providing specific guidance, the Recommendation would complement the analysis and guidance provided in the European Semester through country reports and country specific recommendations (CSRs). For countries with CSRs on this topic or related areas, the recommendation provides further guidance supporting their implementation effort. For countries without a CSR, it provides an EU level impetus for improvement of the coverage of the self-employed and non-standard workers. A Recommendation can also provide a coherent framework with a set of key principles, which could guide further exchanges, within the Social OMC, on how to improve the coverage and adequacy of protection through social security schemes, in practical terms.

9.How will actual impacts be monitored and evaluated?

9.1.Monitoring

Member States will be invited to collect and publish reliable statistics, building on existing statistics where possible but acknowledging the existing gaps, broken down by employment status, gender, age and citizenships and to report on the corresponding measures taken at national level on a regular basis in the framework of monitoring of the European Pillar of Social Rights, of the multilateral surveillance and, in particular, of the European Semester.

Moreover reinforced cooperation will be established with Eurostat to create indicators appropriate for this area. Although long-time series would not be available, it should be possible to track progress towards formal coverage, effective coverage and transparency since the publication of the Recommendation. Work on improving indicators has already started and could be available at the same time as the Member States start implementing the Recommendation.

At European level, monitoring of the implementation of the Council Recommendation in the Member States could be built on the reporting arrangements within the European Semester, in the same vein as done for other Council recommendations. In the current Semester, the Joint Employment Report already includes - for the first time - a short analysis of the gaps in protection through social security systems related to the employment status of people in the labour market. For the Member States where the gaps are more acute, dedicated country analysis will also be subject of multilateral and thematic reviews held by the Social Protection Committee.

On the basis of a commonly agreed list of indicators, the Commission will report to other relevant EU institutions, including the European Parliament, Council, and the European Economic and Social Committee on progress in the implementation of the Recommendation.

In line with the Five Presidents' report 102 , the Commission will also propose to further explore, in the context of the Social Protection Committee, the feasibility of establishing a benchmarking framework for access to social security (see annex 14), pursuing the ongoing effort to establish benchmarking frameworks in areas that affect the sustainability of EMU.

The benchmarking and monitoring framework will aim at evaluating progress made by Member States in decreasing the share of non-standard workers and the self-employed who have no formal coverage and no effective coverage, by social security branch. It will evaluate whether Member States have adjusted time thresholds to ensure better coverage. It will also check whether legal frameworks enable transferability of entitlements and aggregation of periods across schemes. It will examine whether generic and personalised information on social security entitlements is available. Outcome impacts such as the poverty reduction effects of social benefits for the self-employed and non-standard employees compared to standard employees will also be evaluated.

The table A.15 in Annex 14 presents the operational objectives of the preferred option and the corresponding indicators to measure progress with regard to these objectives.

9.2.Evaluation

The Commission will proceed to a review of the actions taken in response to this Recommendation and an evaluation of the impact of the Recommendation in consultation with the Member States and social partners at EU level in line with a review provision in the text of the Recommendation.

The evaluation will include an assessment of whether the operational objectives of the Recommendation have been reached. The benchmark against which progress will be measured is the baseline situation defined in this Impact Assessment.

A mapping of existing gaps of social security has been done for the purposes of this Impact Assessment and will be repeated in some years. The comparison of the two mappings – ex-ante and ex-post - will allow monitoring the evolution of the social security schemes, their coverage and their eligibility conditions., Targets can thus be established with reference to a reduction in the number of groups and issues identified as problematic across branches and Member States.

Annex 1: Procedural information

1. Lead DG, Decide Planning/CWP references

The lead DG is DG Employment, Social Affairs and Inclusion

Agenda planning/Work Programme reference 2017/EMPL/003.

Social Fairness Package: Initiative no. 8 in the Commission Work Programme 2018 103 .

2. Organisation and timing

The Impact Assessment was assessed by the Interservice Steering Group on 11/12/2017 (present DGs SG, ECFIN, GROW, TAXUD) and on 12/01/2018 (present DGs SG, SANTE, GROW).

The Analytical Document accompanying the second phase consultation of social partners on which the Impact Assessment is based, together with the second stage consultation document, was assessed by the Interservice Steering Group on 11/07/2017 (present DGs SG, GROW, HOME, JUST, SANTE, SJ, TAXUD) and on 13/10/2017 (present DGs SG, GROW, SANTE) and adopted following ISC (DGs consulted AGRI, BUDG, CNECT, COMM, COMP, EAC, ECFIN, ENER, EPSC, ESTAT, FISMA, GROW, HOME, JUST, MARE, MOVE, REGIO, RTD, SANTE, SG, SJ, TAXUD, TRADE).

The first stage consultation document, together with the TOR of the impact assessment study were assessed by the Interservice Steering Group on 27/03/2017 (present DGs SG, ESTAT, GROW, HOME, JUST, SANTE, SJ, TAXUD) and adopted following ISC (DGs consulted AGRI, BUDG, COMM, COMP, DEVCO, EAC, ECFIN, ECHO, ESTAT, HOME, HR, CNECT, JUST, MARE, MOVE, REGIO, RTD, SG, SJ, TAXUD, TRADE, GROW, SANTE, FISMA).

3. Consultation of the RSB    

The Impact Assessment report was reviewed by the RSB and discussed with the author DG in a meeting on 7 February 2018. On 9 February 2018 the RSB issued a positive opinion with reservations.

The revisions introduced in response to the RSB opinion are summarised in the table below:

4. Evidence, sources and quality

The following expert advice has fed into the Impact Assessment:

-From the Report of the public consultation "Accompanying the document establishing a European Pillar of Social Rights (SWD(2017)206)".

-From the commissioned study "Access to social protection for all forms of employment: Assessing the options for a possible EU initiative" by Fondazione Giacomo Brodolini.

-From the commissioned study "Behavioural study on the effects of an extension of the access to social protection for people in all forms of employment" by Open Evidence.

-From six commissioned case studies on special forms of employment (Mini-jobs in Germany, civil law contracts in Poland, self-employed without employees in the Netherlands, casual worker in Romania, economically dependent self-employed in Spain, project workers on continuous collaboration projects in Italy).

-From literature review as referred to in the footnotes.

Annex 2: Glossary and Terminology

1.Glossary

2.Definitions

1. Social security

Social security covers all measures providing benefits, whether in cash or in kind, to secure protection, inter alia:

·from lack of work-related income (or insufficient income) caused by sickness, disability, maternity, accidents at work or occupational diseases, unemployment, old age;

·lack of (affordable) access to health care;

·insufficient family support, particularly for children and adult dependants;

·general poverty and social exclusion 104 .

This initiative focuses on protection from lack of work-related income (or insufficient income) caused by sickness, disability, maternity or paternity, accidents at work or occupational diseases, unemployment or old age.

Formal coverage: a group can be identified as formally covered in a specific social security branch (e.g. old age, unemployment protection, maternity protection) if the existing legislation or collective agreement sets out that this group is entitled to participate in the scheme.

Effective coverage: a group can be identified as effectively covered in a specific social security branch if the individuals in this group have an opportunity to accrue benefits and can access to those benefits in case of need (currently or in the future).

A social security scheme is a distinct framework of rules to provide benefits to entitled beneficiaries. Such rules would specify the personal scope of the programme, entitlement conditions, the type of benefits, benefit amounts, periodicity and other benefit characteristics, as well as the financing (contributions, general taxation, other sources), governance and administration of the programme.

A benefit is a transfer in cash or in kind made by a public or private entity to someone entitled to receive it as part of a social security scheme.

Duration of benefits refers to the period of time during which an insured or the insured's dependents can avail of the benefits of the social security scheme.

Qualifying period means the period of membership of a scheme, required under national law or by the rules of a scheme, in order to trigger entitlements to the accumulated social security rights.

Minimum working period means that the person is required to have worked a minimum of hours/months/years before being eligible to receive the benefit in case of need.

Reference period is the period within which the qualifying period must be completed (for instance, 12 months within 24 months).

Waiting period: in some states benefits are not paid immediately upon the occurrence of a contingency, in other words the claimant must wait for a fixed period after the beginning of their unemployment, sickness, etc., before they are paid any social security benefit.

Transferability refers to the possibility i) to transfer accumulated entitlements to another scheme and ii) to make qualifying periods in a previous employment status (or in concomitant employment statuses) count towards the qualifying periods in the new status.

Transparency refers to the provision of information to the general public, potential scheme members, schemes members and beneficiaries about the scheme's rules or about the individual obligations and entitlements.

2. Employment statuses and types of employment relationship

A "worker" means a natural person who for a certain period of time performs services for and under the direction of another person in return for remuneration.

Self-employment refers to employment in which persons pursue a gainful activity for their own account 105 . 

Employment status refers to the classification of persons as either working in the framework of an employment relationship (worker) or working on their own account (self-employed).

Type of employment relationship refers to the various types of relationships that can differ regarding the duration of employment, the number of working hours or other terms of the employment relationship.

Non-standard employment is described by the ILO 106 as employment arrangements which deviate from the 'standard employment relationship' (full time, of indefinite length, part of a subordinate relationship between an employer and an employee). It includes temporary employment (fixed term, casual, seasonal work), part-time and on-call work, multi-party employment relationship (subcontracted labour, temporary agency work) and disguised employment (misclassified self-employment but also dependent self-employment).

New forms of employment 107 include employee sharing, job sharing, interim management, casual work (including intermittent work and on-call work), ICT-based mobile work, voucher-based work, portfolio work, crowd employment, collaborative employment. These forms include elements of non-conventional workplaces, support of ICT, different employment relationship organisation, different work patterns, networking. Eurofound underlines that there is currently no shared understanding of what constitutes ‘new forms of employment’.

Some specific forms of work are also referred to:

Casual work: according to Eurofound's definition 108 , casual work is a type of work where the employment is not stable and continuous, and the employer is not obliged to regularly provide the worker with work, but has the flexibility of calling them in on demand; it is work which is irregular or intermittent with no expectation of continuous employment. Workers’ prospects of getting such work depend on fluctuations in the employer's workload.

Eurofound divides casual work into two main categories:

-    On-demand work (including zero-hours contracts)

-    Intermittent work

The first, on-call work, involves a continuous employment relationship maintained between an employer and an employee, with the option for the employer to call the employee in as and when needed. In some countries employers are obliged to indicate a minimum number of hours, while in others they are not obliged to ever call the worker (i.e. zero-hours contracts). The second category of casual work, intermittent work, involves an employer approaching workers on a regular or irregular basis to conduct a specific task, often related to an individual project or seasonal work.

Crowd employment/platform work: “A European agenda for the collaborative economy” 109 defines the concept of collaborative economy, the presence of an online platform being a necessary element of the definition. It does not, however, refer to platform work, which is not formally defined at EU level. Platform work is carried out by service providers who can be professional or not; they can be self-employed persons or workers.

Eurofound (2015) defines crowd employment as an employment form that uses an online platform to enable organisations or individuals to access an indefinite and unknown group of other organisations or individuals to solve specific problems or to provide specific services or products in exchange for payment.

Crowd workers are very rarely, if ever, employed by the platform. Instead, they predominantly operate on a self-employed basis as independent contractors and outside the scope of employment legislation. Eurofound notes that crowd employment platforms have to follow general legal frameworks such as commercial codes, civil codes, consumer protection acts and data protection legislation, but there are no legal or collectively agreed frameworks specifically addressing crowd employment in Europe 110 . 

Domestic work: ILO defines “domestic work” in Article 1 of the Domestic Workers Convention, 2011 (No. 189):

(a) the term “domestic work” means work performed in or for a household or households;

(b) the term “domestic worker” means any person engaged in domestic work within an employment relationship;    

(c) a person who performs domestic work only occasionally or sporadically and not on an occupational basis is not a domestic worker.

'Temporary Agency Work' is a form of work where the worker has a contract of employment or an employment relationship with a temporary-work agency with a view to be assigned to a user undertaking to work temporarily under its supervision and direction 111 .  

Voucher-based work is 'a form of employment where an employer acquires a voucher from a third party (generally a governmental authority) to be used as payment for a service from a worker, rather than cash' 112 . This type of employment has been introduced in several Member States as a way to better regulate domestic work and improve protection of domestic workers, or tackle the problem of undeclared work in agriculture.

Traineeship is a limited period of work practice, whether paid or not, which includes learning and training components, undertaken in order to gain practical and professional experience with a view to improving employability and facilitating transition to regular employment 113 .

‘Bogus’ or ‘false’ self-employment refers to the phenomenon of workers who would usually meet the legal definition of a worker (see above) but instead are registered as self-employed. In this case, the subordinate employment relation is disguised as autonomous work, usually for fiscal reasons, or in order to avoid the payment of social security contributions and thereby reduce labour costs, or to circumvent labour legislation and protection, such as the provisions on dismissals. This can be characterised as an employer abuse of contractual relations, and these workers are also, compared to employed workers, at a higher risk of precariousness due to a lack of social security and pension coverage and a lack of access to some employment rights 114 .

Dependent self-employed are responsible for their own work and therefore formally do not work under the authority and responsibility of the main contractor even if de facto they are economically dependent from one employer/customer.

Economic sectors are sectors of the economy , grouped according to the kind of product s or service s they produced or according to ownership.

Annex 3: Further information on the organisation of social protection systems 

Social protection schemes may differ in a number of dimensions. There is no single classification available and recognised internationally. However, the European system of integrated social protection statistics, abbreviated as ESSPRO and developed by Eurostat defines a number of characteristics which can be freely combined to produce different groupings of schemes as required by analysis.

The categories are defined on the basis of (i) which body takes the essential decisions, (ii) the existence or absence of a legal obligation, (iii) the way entitlements are established, (iv) the scope of the scheme, (v) the level of protection provided. The classification is part of ESSPROS's regulation.

Decision-making

In total, Eurostat registered 1082 social protection schemes in Europe (including Switzerland, Iceland, Serbia and Turkey). Around 70 of them are not controlled by the central government or local authority. Among these, you can find many combinations: occupational or general schemes, compulsory or not compulsory, etc.

Typologies and financing of social protection schemes

Non-contributory schemes are social transfer schemes which normally require no direct contribution from beneficiaries or their employers as a condition of entitlement to receive relevant benefits. They include a broad range of schemes including universal schemes for all residents, categorical schemes covering specific categories of residents and means-tested social assistance schemes. They are usually financed through tax or other state revenues.

Universal schemes, in particular, are non-contributory transfer schemes, which cover all residents, and provide benefits for all, whether working or not and irrespective of income. Often the only condition attached to the receipt of the benefit is that the person must be a long-term resident or a citizen of the country.

Social insurance – like all forms of insurance – is a mechanism for (financial) risk pooling. It represents one of the most important instruments through which social security schemes operate, as it has done since the outset of national systems of social security, the earliest being the system introduced by Chancellor Bismarck in Germany in the 1880s. The insurance principle operates on the basis that a group of individuals (or perhaps their families), who are vulnerable to a specific risk (for example, disabling injury, or sickness) pay contributions into a pool or fund which is then used to pay benefits according to specified rules so as to indemnify those participants who suffer from the occurrence of the relevant risk.

Social insurance is financed by contributions, which are normally shared between employers and employees, with, perhaps, state participation in the form of a supplementary contribution or other subsidy from general revenue. Participation is compulsory, with few exceptions. Contributions are accumulated in special funds, out of which benefits are paid. A person's right to benefit is secured by his contribution record without any means-test. Within floor and caps contribution and benefit rates are often earnings-related 115 .

Occupational social security schemes provide workers, whether employees or self-employed, in an undertaking or group of undertakings, area of economic activity, occupational sector or group of sectors with benefits intended to supplement the benefits provided by statutory social security schemes or to replace them, whether membership of such schemes is compulsory or optional 116 .

Financing of social protection

There is a wide variety in financing structures of social protection systems across Europe. While on average, around 55% of the financing is relying on social contributions (employers and persons protected), the share of total social contributions in total resources actually varies from around 20% in DK to 80% in EE (Figure A.1). This diversity reflects the variety of the provision of social protection in Europe, such as typically whether social protection systems are more insurance based (and thus more financed by social contributions) or universal in nature (and then more financed by general taxation), but also other dimensions, such as the relative share in expenditure of the different main risks (for instance pensions are generally more financed based on social contributions and health expenditure more based on general taxation, see para below) or the relative importance of occupational versus publicly managed schemes.

Some estimates of the relative shares of social contributions and general taxes to the financing depends by type of social protection area considered can be derived 117 . In general pensions and family expenditure rely more heavily on social contributions (on average around 75% of total financing) as well as unemployment benefits (around 60% on average), while health care systems rely less on social contributions (on average 40%) and more on general taxation.

Trends in social security financing show that general taxes increasingly complement social contributions (Figure A.2a shows that this continued in terms of structure since the crisis hit in 2008, though to a lesser extent). The share of general taxation (and other sources) in total resources increased in the latest decades and is now close to 45% (against 35% in 1995, see Figure A.3b). It can be noticed that this increase in the share of general taxation corresponds mostly to a decline in the share of employers’ and employees’ contributions, while the share of social contributions from the self-employed or other persons covered remained broadly constant. This observed trend of decline in the shares of employers’ and employees’ contributions combined with a growing number of persons in self-employment poses a double challenge to the sustainability of social security systems, particularly Bismarckian ones 118 .

Figure A.1: Structure of social protection financing (2014)

(%GDP)                        (% total resources)  

Source: ESSPROS

Figure A.2: Trends in social security financing structures

a. b.

Source: ESSPROS

The structure of social benefits expenditure

At EU level, social expenditure on old age pensions and healthcare has grown over recent years as a percentage of GDP, while the expenditure devoted to family and unemployment benefits has tended to either remain static or decrease. In 2014, overall social protection expenditure (excluding administrative and other expenditures) accounted for 28% of GDP on average across the EU, ranging from around 14.5% in Latvia and Lithuania to over 30% in Denmark, France and Sweden (Figure A.3).

While there are clear variations across Member States in the actual composition of expenditure (Figure A.4), in all countries pensions (old age and survivors) and healthcare (sickness/healthcare and disability) account for by far the largest shares of social protection expenditure, with these averaging 44% and 35% respectively (or 13% and 10% of GDP) at EU level. In general, pensions account for the largest share of expenditure in almost all countries, with this being especially the case in Cyprus, Italy, Poland, and above all Greece. The other branches of social protection account for much more limited shares of social expenditure – at EU level family/child benefits account for 8% of total expenditure (or 2.4% of GDP), unemployment 5% (or 1.4% of GDP) and housing 2% (0.6% of GDP) although notable variations exist. For example, in Belgium, Finland, Ireland and Spain expenditure on unemployment benefits is well above the EU average, while in Denmark, Finland, Germany, Luxembourg and Sweden this is more the case for family/child benefits, and in the UK there is markedly more expenditure on housing benefits.

Generally speaking, the level of social contributions from the self-employed is significantly lower than that of employees, taking into account both employees and employers contributions. However, it is similar to the level of employees' contributions only (Figure A.5), which can typically relate to the design of social security provision (with less extensive or voluntary settings for the self-employed) 119 .

Figure A.3: Structure of social protection expenditure as % of GDP, 2014 (excluding administrative and other expenditures)

Source: ESSPROS

Figure A.4: Structure of social protection expenditure as % of total expenditure, 2014

Source: ESSPROS

The role of social dialogue in easing transitions

Social partners play a key role in the design of social protection schemes that derive from collective bargaining. These schemes often relate to a given economic sector, reflecting the structure of sectoral dialogue. In other instances, there are initiatives with a cross-sectoral dimension. In either case initiatives from social partners no longer consider solely the social protection benefits inherent to a given contract and take a wider perspective of securing transitions and whole professional careers.

For example, a variety of joint approaches have been used to address issues related to extending social protection to Media, Arts and Entertainment sector workers, mostly as an outcome of social dialogue. Some trade unions offer social protection coverage to freelance members. In some countries, specific schemes have been developed to provide coverage to Media, Arts and Entertainment workers – for example, the social security insurance fund for artists and writers in Germany 120 , which covers self-employed and freelance artists and writers, and the unemployment benefit for intermittent entertainment workers in France.

Figure A.5: Estimates of the relative apparent contributions of self-employed vs. employees in the EU (2014)

Source: ESSPROS

Note: Estimates of GDP per employment are derived from the national accounts for GDP levels and from LFS for employment levels; the number of self-employed is also taken from the LFS.

Since 1969 in France, performing artists working on a short-term basis (intermittent du spectacle) benefit from a rebuttable presumption of employment enshrined in the French Labour Code. This provision applies to performing artists and to technicians who regularly work on a short-term basis. The mechanism of the Guichet Unique du Spectacle Occasionnel (GUSO) is designed to facilitate the protection of short-term workers in France and their enjoyment of this status. This mechanism and the vocational training fund that it facilitates are an outcome of the strong social dialogue between employers and workers on behalf of the intermittent du spectacle. The Swedish 'Teateralliansen' (there is also a similar 'dansalliansen' which caters for dancers) is a structure specifically created to increase the social protection of freelance actors 121 .

Another case in point is the bipartite funds that offer extra social protection to agency workers. These funds are financed through employers' contributions and governed through a bipartite model. The funds are cross-sectoral and allow agency workers to benefit from their protection regardless of the sector in which they perform their assignments or between two assignments.

Annex 4: Further information on non-standard employment and self-employment

1.Self-employment: decline in traditional forms amid growth in new forms

1.1.The self-employed: group size, characteristics and heterogeneity

There were 32.7 million self-employed in the EU-28 in 2016, accounting for 15% of total employment. They were mainly concentrated in the following sectors: agriculture, forestry and fishing (14.1%), wholesale and retail trade (15.8%), construction (12.9%) and professional, scientific and technical activities (12%) 122 . Southern and Eastern European countries have the greatest percentage of self-employed people, with Greece topping the charts (30%). At the other end of the spectrum, two central and Northern European countries have the lowest proportion of self-employed, i.e. Denmark (8.3%) and Luxembourg (9%). The Netherlands and Ireland have self-employment rates above the European average (Figure A.6) 123 .

Figure A.6: Share of self-employed in total employment, by countries, 2016

Source: Eurostat, Employment by sex, age and professional status

Among the self-employed it is possible to distinguish between the self-employed with and without employees. In 2016, 71.5% of the self-employed were self-employed without employees 124 ; 28.5% of the self-employed without employees were registered as managers, professionals or associate professionals, while around 21.4% were agricultural or elementary workers, 16.2% were reported to be sales or service workers and 21% skilled or semi-skilled manual workers such as craft and related trade workers. The occupations of the self-employed vary considerably across countries. For example, in the Sweden, Belgium, Germany and Luxembourg, well over 50% of the self-employed without employees worked as managers, professionals or associate professionals in 2016, while in Romania they were just 3%.

The overall (marginal) decrease of the self-employed between 2008 and 2016 becomes an increase, if the primary sector is excluded from the analysis (Figure A.7).

Figure A.7: Share of self-employed in total employment in the EU28, 2008 and 2016

Source: Eurostat

Moreover, it is possible to note that there has been a change in the socio-economic structure of the self-employed population (Figure A.8). A high decrease can be noticed among the managers, and the occupations related to agriculture and manufacturing, while there has been an evident increase among the self-employed in services and sales and among the professionals.

Figure A.8: Share of self-employed in total employment by occupation in the EU28, 2008 and 2016

 Source: Eurostat

As also expressed by stakeholders during consultations, it is very important to go further in-depth in the analysis to identify and take into account the diversity among the group of the self-employed (going beyond the traditional split between the self-employed with/without employees) in order to assess to which extent granting them access to social security responds to an actual need 125 .

A study by Eurofound 126 sheds light on the heterogeneity of the self-employed using survey evidence exploited with the help of advanced econometric methods 127 .

In particular, the self-employed can be clustered into five different groups ('employers 128 ', 'stable own account workers' 129 , 'vulnerable' 130 , 'concealed' 131 , 'small traders and farmers' 132 - see also Figure A.9) in line with their characteristics along a pre-selected list of key dimensions relevant to their employment and work situation. The names of the groups were chosen to illustrate the most important aspect of the nature of each group. They are not commonly agreed or formal terms.

Of those five groups, two groups – the ‘employers’ and the ‘stable own-account workers’, representing about half of the self-employed or just over 16 million people - generally have more favourable characteristics. These groups are generally more independent with more discretion over their work, have bigger and economically more viable and sustainable activities and are more likely to be self-employed out of opportunity. For two other groups, ‘vulnerable’ and ‘concealed’, together a group of eight million people - or roughly the population size of Austria - the opposite is the case: they are generally more dependent and have less autonomy over their work. Also, these groups are in more precarious situations with lower levels of income and job security. They experience overall rather unfavourable working conditions. For the final group, ‘small traders and farmers’, which amount to also around eight million people, the picture is more mixed with a combination of both favourable and unfavourable conditions.

Figure A.9: Size of five groups of self-employed within the EU28 workforce (% of total employment)

Source: Eurofound (2017)

There is strong heterogeneity between the groups along a number of dimensions of high relevance for the present initiative. For example, nearly all the self-employed in the cluster ‘employers’ did not become self-employed out of necessity whereas only around 60% of the self-employed in the cluster ‘vulnerable’ are not self-employed out of necessity. The degree of being comfortable with the uncertainty and risks associated with the status of self-employment also differs markedly across the groups. While more than 90% of the ‘employers’ and close to 100% of the ‘stable own-account workers’ like being boss and taking decisions, the same holds true for less than three quarters of the ‘vulnerable’ and only about 55% of the ‘concealed’.

The clusters differ substantially with respect to their autonomy: Three quarters of ‘stable own-account workers’ have the authority to hire or dismiss employees while the shares for ‘employers’ and ‘small traders and farmers’ even exceed 90%. This only holds for below 20% of the ‘concealed’ and around 30% of the ‘vulnerable’ self-employed. Likewise, more than 90% of the ‘stable own-account workers’, ‘small traders and farmers’ and ‘employers’ have more than one client, compared to less than half of all ‘vulnerable’ self-employed.

There is also large diversity regarding the position in the income distribution and financial vulnerability. More than 60% of the ‘vulnerable’ are in the lowest income quintile, while this holds for only around 10% of the cluster ‘employers’ and 20% of the ‘small traders and farmers’. Close to 60% of ‘employers’ consider themselves financially secure in case of sickness. The other groups are more vulnerable, with only around 20% of ‘small traders and farmers’, 20%-30% of the clusters ‘concealed’ and ‘vulnerable’ and below 40% of ‘stable own-account workers’ being in the same situation. Looking at income levels alone does not give a sufficient picture of vulnerability in this respect. For instance among the ‘small traders and farmers’, almost two-thirds would be financially insecure in case of sickness, although incomes in this group tend to be at the higher end.

1.2.New forms of self-employment: dependent self-employment and bogus self-employment

In line with the European Court of Justice rulings 133 , the status of worker is determined by three elements: existence of remuneration in return for services; existence of work in a relationship of subordination where the employed determines the choice of activity, remuneration and working conditions; and the existence of effective and genuine activity of economic value. Out of these three elements, the relationship of subordination allows to draw a distinction between workers and self-employed persons and the most specific way of defining self-employment is the absence of subordination between the employer and the employees 134 .

Given the characteristics associated with entrepreneurship – decision-making power regarding the business, financial autonomy and responsibility – the ‘self-employed with employees’ are considered as being by nature self-employed. The category of ‘self-employed without employees’ is more diverse and, in some ways, ambiguous 135 . The group of economically dependent workers belongs to this category. To be categorised as an independent self-employed without employees, an individual should work as an autonomous subcontractor, who owns his/her means of production and works independently in order to accumulate capital and serve several clients 136 .

Thus, there are emerging work arrangements which are “in a grey area”, meaning that they do not easily fit into the traditional binary distinction between "employees" and the independent "self-employed”. In fact, there is an increase in formally self-employed work relationships that are marked by both economic and personal dependence. Given that these workers have only one (main) employer, they generate the whole, or at least a substantial part, of their income from this work relationship. Economic dependence may also mean that the worker takes (part of) the entrepreneurial risk without the trade-off in terms of employment security. Furthermore, due to the demands of their 'main' employer, the entrepreneurial possibilities associated with self-employment are also rather limited or absent in their case. These characteristics form the category of the dependent self-employed who work with only one client, under similar conditions as employees, but without the rights and protection that employees are entitled to under labour law, social protection rules and collective agreements. Dependent self-employment also means that both organisational boundaries and the boundaries set by labour and social security law between employment and self-employment become blurred 137 . Indeed, the distinction between the dependent self-employed and employees has important labour law, social protection and fiscal consequences:

·dependent self-employed are responsible for their own work and therefore formally do not work under the authority and responsibility of the main contractor even if de facto they are economically dependent on one employer/customer;

·access to social protection (e.g. in the case of temporary employment, occupational accidents, early retirement etc.) is more restricted for the dependent self-employed 138 .

·some working conditions (wages, working time, rest periods etc.) are governed by collective agreements or by specific legislative, administrative and regulatory provisions that are not applicable to the dependent self-employed 139 .

Therefore, the dependent self-employed are neither clearly separated from nor integrated with the employer they contract with. Moreover, despite the lack of data which would allow to easily distinguish between the traditional self-employed and the dependent self-employed, it is estimated that approximately 15% of all self-employed are economically dependent workers in the EU 140 . Moreover, there is evidence that the biggest difference between dependent self-employed and other groups of self-employed lies in their financial situation. Their households have a correspondingly higher level of difficulty making ends meet and they have the lowest level of job security, and the fewest opportunities for career advancement 141 . A Eurobarometer survey 142 shows that 49% of respondents who had started up a business or were currently taking steps to start one, stated that they were doing so because they saw an opportunity and 29% were doing so out of necessity, i.e. the possibilities for regular employment were absent 143 .

Moreover, among the dependent self-employed, another group with the only choice of working on their own account because of a lack of job opportunities 144 can be identified: that of the 'bogus self-employed' 145 .

‘Bogus’ or ‘false’ self-employment refers to the phenomenon of workers who would usually meet the legal definition of an employee but instead are registered as self-employed 146 . In this case, the subordinate employment relation is disguised as autonomous work, usually for fiscal reasons, or in order to avoid the payment of social security contributions and thereby reduce labour costs, or to circumvent labour legislation and protection, such as the provisions on dismissals 147 .

The OECD has described the bogus (or ‘false’) self-employment as consisting of ‘people whose conditions of employment are similar to those of employees, who have no employees themselves, and who declare themselves (or are declared) as self-employed simply to reduce tax liabilities, or employers’ responsibilities’. This category is particularly used in labour-intensive industries, such as construction, transport and the service sector 148 .

There are several factors leading to 'bogus' self-employment. One is that in almost all countries few formalities have to be fulfilled to set up as self-employed and the forms of bogus self-employment have gradually become more sophisticated 149 . Relatively restrictive national labour markets (i.e. highly regulated ones) and high non-wage labour costs are also factors creating 'bogus' self-employment 150 . From the point of view of the employers, it is a way to transfer costs, risks and responsibilities to the workers and to circumvent collective agreements, labour laws, payroll taxes and social security contributions 151 . On the other hand, while, in general, motives for people accepting a working engagement as dependent self-employed could include time flexibility and increased autonomy, together with further advantages in the form of tax benefits and lower social contributions, in the case of 'bogus' self-employment, reasons seem to be related only to their economic situation, since the characteristics of autonomy and flexibility are not maintained.

2.Non-standard work: temporary, part-time and new forms of employment 

Overall, the share of non-standard work in employment has increased over time. Using data for the EU28 and a longer available time series for the EU15, Figure A.10 shows that this has been the case especially for younger workers aged 25-39: in 1995, 23% of younger employees in the EU-15 had non-standard contracts. This proportion had increased to 32% for the same age group by 2016, implying that there has been a clear intergenerational component to this development.

Figure A.10: Employees by type of contract and age, 1995, 2005 and 2016

Source: European Commission, ESDE 2017, p. 77.

2.1.Temporary employment

In 2016, the proportion of employees aged 15–64 in the EU-28 with a fixed-term employment contract was 14.2% (Figure A.11). More than one in five employees in Poland (27.5%) and Spain (26.1%) had a temporary contract. Among the remaining Member States, the share of employees working on a contract of limited duration ranged from 22.3% in Portugal, down to 2% in Lithuania and 1.4% in Romania.

Figure A.11: Share of temporary workers in total employment (age 15-64), 2016

Sources: Eurostat

Both the number of temporary contracts and the temporary employment rate (number of temporary as percentage of all employees) show a strong cyclical (and seasonal) pattern. Both decline at the onset of the recession and tend to increase with recovery (and pick up the “double dip” around 2010-11), as many new hires are on temporary contracts 152 . Indeed, over the last three years (2014-2016), as countries emerged from the crisis, the growth in the number of permanent employees has exceeded the increase in the number of temporary employees. Thus, in the year preceding the third quarter of 2016, the number of employees with a permanent contract increased by 2.8 million (1.8%) in the EU-28, while the number of employees with a temporary contract increased by 420,000 (1.6%). Similarly, full-time employment has increased faster than part-time employment for nearly three years. In the years up to the third quarter of 2016, the number of full-time workers increased by 2.8 million (1.6%) while the number of part-time workers increased by 370,000 (0.9%).

Denmark, Netherlands, Slovenia, Malta, Croatia and Poland show some appreciable increase in the temporary rate since the onset of the recession in 2008. Spain, Germany, Latvia and Lithuania show a significant post-recession decrease. Given Spain’s large share of temporary contracts in Europe, the decline there has held down the European rate over this period, and vice-versa for Poland.

In a more long-term perspective, a clear increase in the share of workers on temporary contracts has been observed especially for younger workers (aged 25-39), see Figure A.12. At the same time, it is worth noting that women are more likely to work on temporary contracts than men, although the gap has been shrinking (Figure A.12).

Figure A.12: Share of temporary workers in total employment (by age and gender, 1995-2001 (EU-15) and 2002-2016 (EU-28)

Source: European Commission (2017 ESDE), p. 79.

Among temporary employment, agency work is experienced as one of the more insecure of contractual forms 153 . This is characterised by a particular “triangular” relationship, where workers are hired by an entity (the temporary work or employment agency) and then re-hired out or assigned to perform their work at (and under the supervision of) a user firm. In 1999, when this form of employment had just become legally permissible in most of Europe 154 , agency workers accounted for 1.2% of all employment in Europe. In 2015, it accounted for 1.9% 155 . The highest rate was in the UK (3.8%), followed by the Netherlands (3.0%) and France (2.1%). The lowest rate of 0.1% was in Greece. These figures are rather similar to those from 1999, with the exception of Germany where the agency worker rate increased from 0.7% to 2.4% of total employment.

The motives for employees to take on a temporary contract are primarily non-voluntary. Given the option of two identical jobs, one open-ended and the other with a temporary duration, the Labour Force Survey shows that only around 12% report that they did not want a permanent job 156 .

2.2.Part-time employment

Part-time work has grown for decades. This is due to the changing structure of the economy towards services and preferences for part-time work, not least related to the increase in female labour force participation. The proportion of the EU-28 workforce in the 15–64 age group reporting that their main job was part-time increased steadily: from 17.5% in 2007 to 19.5% by 2016, representing the type of non-standard employment that has undoubtedly grown since the onset of the recession. Over the period from 2007 and 2015, there was a widespread increase in the share of part-time workers, though it declined in Croatia, Lithuania, Romania and Poland 157 .

Figure A.13: Share of part-time workers in total employment (age 15-64), 2016

Source: Eurostat

As shown in Figure A.13, by far the highest proportion of part-time workers in 2016 was found in the Netherlands (almost 50%), followed by Austria, Germany, Denmark, the United Kingdom, Belgium, Sweden and Ireland. By contrast, part-time employment is relatively uncommon in Bulgaria (2%) as well as in the Czech Republic, Slovakia, Hungary and Croatia (between 4.8% and 5.8%). In these countries, non-standard employment is rare, and people may consider it to be an involuntary status.

The incidence of part-time work differs significantly between men and women. Almost one third (32%) of women aged 15–64 who were employed in the EU-28, worked on a part-time basis in 2015, a much higher proportion than the corresponding share for men (8.9%). 76.5% of women employed in the Netherlands worked on a part-time basis in 2016, by far the highest rate among EU Member States 158 .

In the most recent years, as the labour market improves, the part-time employment rate has declined only marginally, thus indicating that not all part time is voluntary. Self-reported involuntary part-time work (defined in relation to the response in the LFS survey “Person could not find a full-time job”) increased from 22.4% of all part-time work to 29.1% between 2007 and 2015 159 . There is a very strong association between national labour market performance and changes in the share of involuntary part-time work. Where the unemployment rate has increased most, there is a strong likelihood that the involuntary part-time share has also grown strongly. Greece, Spain and Cyprus are illustrative examples. The corollary is also true. The Member State with the most improved unemployment record over the period, Germany, is also the one with the sharpest decline in involuntary part-time share (Figure A.14) 160 .

Figure A.14: Changes in involuntary part-time employment share and in unemployment rate 2008-2015.

Source: LFS as calculated in Eurofound (2017)

Changes concern also the working hours. While women account for the majority of very low hours' part-time work, the male share has grown relatively faster since 1996 (from 1.5% to 2.8% of all male workers) compared to the female share (from 6.1% to 6.6% of all female workers). Highest shares were recorded amongst both young workers (less than 24 years of age) and workers above 64 years of age and it was in these groups, in particular the younger group, that the greatest growth in the share of very short hour part-time work was recorded.

2.3.New forms of work

As in the case of the self-employed the available data do not allow to capture the variety of non-standard forms of employment present in the labour market, and, in particular, the new forms of work.

Eurofound (2015) 161 distinguishes three categories of employer-employee relationship: relationships between employers and employees involving either multiple employers for each employee, one employer and multiple employees or even multiple employer-multiple employee relationships; provision of work on a discontinuous/intermittent basis or for very limited periods of time; and networking and cooperation agreements involving self-employed persons, especially freelancers.

On the basis of this broad categorization, the following new forms of non-standard employment were identified by Eurofound:

·employee sharing, where a single worker is jointly hired by a group of employers to meet the HR needs of various companies, resulting in permanent full-time employment for the worker.

·job sharing, where an employer hires two or more workers to jointly fill a specific job, combining two or more part-time jobs into a full-time position;

·interim management, in which highly skilled experts are hired temporarily for a specific project or to solve a specific problem, thereby integrating external management capacities in the work organisation;

·casual work, where an employer is not obliged to provide work regularly to the employee, but has the flexibility of calling them in on demand;

·ICT-based mobile work, where workers can do their job from any place at any time, supported by modern technologies;

·voucher-based work, where the employment relationship is based on payment for services with a voucher purchased from an authorised organisation that covers both pay and social security contributions;

·portfolio work, where a self-employed works for a large number of clients, doing small-scale jobs for each of them;

·crowd employment, where an online platform matches employers and workers, often with larger tasks being split up and divided among a ‘virtual cloud’ of workers;

·collaborative employment, where freelancers, the self-employed or micro enterprises cooperate in some way to overcome limitations of size and professional isolation.

In general, crowd work seems to be used as a supplement to total income. A survey conducted in the UK showed that in 2016 platform workers amounted to 4% of the employees, 25% of whom reported it as main job 162 . Another survey 163 implemented in the UK, Sweden, Germany, Austria and the Netherlands in the first two quarters of 2016 showed that between 5% and 9% of the online population were engaged in some type of crowd work 164 . It is usually a small supplement to total income 165 .

Furthermore, together with these new forms of employment, a variety of new forms of contracts has been recognised in Member States:

·'Civil law contracts for a specific task' there are different legal forms from a contract of employment and they are governed by the provisions of the Civil Code. They are particularly prevalent in Poland and commonly used when the employer does not plan to employ a worker under an employment contract, but requires the performance of certain activities or for some specified work to be completed.

·Mini-job is a concept coined in Germany, which refers to employment contracts for earnings of currently up to EUR 450 per month with reduced income tax and social security contribution rates and entitlements. 

·'On–demand work' involves a continuous employment relationship maintained between an employer and an employee where the employer has the option of calling in the employee as and when needed. There are employment contracts that indicate the minimum and maximum number of working hours, as well as so-called ‘zero-hours contracts’ that specify no minimum number of working hours, and the employer is not obliged to ever call in the worker.

Annex 5: Visualisation of the Problem Tree

Annex 6: Internal Drivers

This annex illustrates in detail the internal drivers described in section 2.2.2. The information in this section are based on ESPN synthesis and national reports on access to social protection for people in non-standard employment and in self-employment (2017), as well as on the report from Brodolini (2018, forthcoming) and on the report from Open evidence (2018, forthcoming). The information included reflects the information available at this stage. Further work with Member States will be undertaken to refine the picture. Member States where gaps were identified by experts are mentioned in the respective tables. However, there may be cases where gaps exist but have not been identified yet. Gaps in formal coverage and effective coverage are analysed for non-standard workers and for the self-employed. The annex indicates the branches in which major gaps of coverage by social security systems have been identified for non-standard workers and for the self-employed, and it shows as well the branches for which voluntary schemes are in place for the two groups.

The problems of transferability, transparency and behavioural biases are described for the two groups together. Regarding effective coverage, four criteria with a time dimension have been identified as obstacles on the basis of the information provided by the ESPN experts: long waiting periods, minimum qualifying periods, duration of benefits and minimum working period. Moreover, Member States in which transferability of accumulated social security rights across different social security schemes is not possible have been identified, as well as Member States in which there a lack of information/transparency regarding social security schemes.

Finally, problems of social security coverage have been also explored in relation to low enrolment rates observed in case of voluntary schemes.

1.Regulatory biases

This annex explored gaps in the following branches of social security:

(a) unemployment benefits;

(b) sickness benefits;

(c) maternity and equivalent paternity benefits;

(d) invalidity benefits;

(e) old-age benefits;

(f) benefits in respect of accidents at work and occupational diseases.

An initial, broader mapping 166 had shown that problems are concentrated in these areas.

1.1.Gaps in formal coverage

For the purpose of this Impact Assessment, gaps in formal coverage are identified in the cases in which people in non-standard employment or self-employment are not covered mandatorily by schemes in a given social security branch and do not even have the possibility to voluntarily adhere to them. Other than these regulatory deficiencies, in the case of voluntary coverage, behavioural choices may lead to people excluding themselves from formal coverage, which may be problematic. The sections below discuss how behavioural biases may play a role in such decisions and provide overview tables on the situation of formal and effective coverage across the Members States and across all branches of social security. Indeed the extent of gaps in formal coverage differs significantly between the various types of non-standard workers and the self-employed, as well as between branches of social security.

Non-standard workers are usually formally covered by social security in the same way as standard workers. However, in some Member States specific categories of non-standard employees are not formally covered for some or for all branches of social security. The main groups for whom this is the case in a significant number of countries are casual and seasonal workers, as well as trainees and apprentices. Furthermore, there are a number of national categories to which restrictions apply in the concerned Member State, such as 'marginal part-timers' and 'marginal freelancers' in Austria, 'mini-jobbers' in Germany and 'civil law contracts for a specific task' in Poland. Table A.2 gives a detailed overview of access gaps in Member States by social security branches and by category of non-standard work 167 .

Table A.2 - Lack of formal coverage to social security for people in non-standard employment

Note: a) Marginal part-timers; b) Agreement to perform a job; c) Mini-jobbers; d) civil law contracts; e) employees on "work agreement" with irregular income; f) domestic workers; g) on call jobs; h) temporary agency work.

Estimates for the size of the identified groups in the concerned Member States are not available in some cases. See Box A.2 for available figures. As regards the situation of the platform workers, recent research have tried to estimate the size of the platform economy and the extent to which it provides access to social security (see Box A.3).

The self-employed as a group lack formal coverage regarding key branches of social security, notably regarding unemployment benefits, accidents and occupational injuries as well as sickness benefit in a number of Member States. In some Member States, different rules apply for certain categories of self-employed (e.g. in Germany, Greece, Finland, France and Italy) Notably, the self-employed generally have no mandatory formal coverage and cannot opt-into voluntary schemes either for unemployment benefits in eleven Member States 171 , although some specific categories of self-employed can be covered on a voluntary basis in some of them 172 (Table A.3). The self-employed are also not formally covered by sickness benefits in three Member States 173 . Finally, the self-employed are also not covered in case of accidents and occupational injuries in ten Member States, although some specific categories of self-employed can be mandatorily covered 174 . It is also worth noting that in many branches and Member States, the self-employed have voluntary rather than mandatory formal coverage (see next section), which may lead to only a limited proportion of them being members of such schemes.

Table A.3 - Lack of formal social security coverage for the self-employed

Note: The table reports in which branches and in which Member States at least one sub-group of the self-employed is excluded from formal coverage in the sense that they have no mandatory coverage and cannot opt-into voluntary schemes either.

a) Only one or more sub-groups of the self-employed are not formally covered.

b) In these Member States only means-tested benefits are available to the self-employed while they are excluded from contributory schemes.

Voluntary coverage

In several Member States, adherence to certain social security schemes is not mandatory for the self-employed and some categories of non-standard workers, but is available on a voluntary basis (see Tables A.4 and A.5). Voluntary schemes can be supplementary to the existing mandatory schemes, or they can apply only to categories of workers which are not covered by the main schemes.

Table A.4: Voluntary social security schemes for people in non-standard employment

Note: a) supplementary schemes; b) if income below a certain threshold; c) For specific categories of non-standard employees.

Table A.5: Voluntary social security schemes for the self-employed

Note: a) additional schemes; b) if income below a certain threshold, c) for specific categories of the self-employed.

For Member States where voluntary schemes exist for all the self-employed or some groups, a low rate of enrolment can generally be observed. For instance, 5% of the self-employed are estimated to be covered by sickness benefits in Austria, 15.37% in Czech Republic, and almost no self-employed in Romania. As for old-age and survivors' pension schemes, 46% of self-employed are estimated to be covered in Belgium, only 650 self-employed in the Netherlands and 20% of the self-employed in Romania. Voluntary schemes for unemployment are also estimated to be under used: in Austria only 0.02% of the self-employed is covered; in Spain less than one quarter; in Finland 20% of the self-employed without employees and 10% of the self-employed with employees; in Ireland 10037 self-employed enrolled in a voluntary scheme 175 ; and in Romania only less than 1% is enrolled. Voluntary schemes for invalidity also result in low formal coverage: only 25% of self-employed were estimated to be covered in 2013 in the Netherlands (against 30% in 2006) and in Romania only 20% 176 .

1.2.Gaps in effective coverage and transferability

People in non-standard employment or in self-employment may be mandatorily formally covered or have the legal possibility to opt for access without, however, achieving effective coverage in the sense of de facto being able to build and take up adequate entitlements to benefits.

This may be due to rules governing the contributions and entitlements which are more difficult to meet for people in non-standard employment or self-employment. This may be the case because the requirements are stricter for these groups than for standard employees. Additionally, it may in some cases also be more difficult for them to fulfil identical requirements to those for standard employees. For instance, temporary employees are more frequently affected by unemployment, making it more difficult for them to accumulate necessary qualifying periods for unemployment insurance 177 .

Based on a more comprehensive mapping exercise, the criteria most frequently used by experts and thus used to identify gaps in effective coverage are:

-    Long waiting period, i.e. particularly long length of time between the occurrence of the risk and the receipt of the benefit.

-    Minimum qualifying period, i.e. the minimum period that people in non-standard employment and self-employed have to contribute to a scheme in order to start receiving the benefit.

-    Duration of benefits, i.e. the length of time during which the individual receive the benefit, in case this differ from that of standard workers.

-    Minimum working period, i.e. if the person is required to have worked a minimum of hours/months/years in order to be eligible to receive the benefit in case of need. This criterion is stricter than the minimum qualifying period, since a person can be, for instance, unemployed, but have the possibility to still contribute to the system, while when a minimum working period is required the person not employed nor in self-employment is not eligible for the benefit.

In addressing these issues, it is important to recognise that some of these criteria play a key role in the design of social security systems (see Box A.4 below).

Another problem affecting effective coverage relates to the income assessment base. This can be problematic in the case of high heterogeneity of rules among schemes which can determine a lower level of entitlements.

Finally, a lack of transferability when professional mobility (including between employment and self-employment) implies changing social security schemes can also determine gaps in effective coverage, as it does not allow the person to preserve the accumulated entitlements or make already fulfilled time thresholds count in the new scheme.

Rules related to time thresholds

As shown in table A.6, despite being formally covered, people in non-standard employment can lack effective coverage because the same rules and conditions as the standard employees apply also to people in non-standard employment, without taking into consideration the specificities of non-standard contracts and disadvantaging people working under this employment relationship. Long waiting periods to receive sickness benefits have been identified as an issue in two Member States and to maternity benefits in five Member States. A minimum qualifying period has been identified as an issue in case of unemployment in six Member States, in three countries concerning old-age, in six countries concerning sickness benefits, in one in case of maternity and in two concerning invalidity. Minimum working periods have been identified as an issue in four Member States for unemployment benefits, in two for old-age benefits, in one for sickness benefits, maternity benefits and invalidity. Finally, the duration of unemployment benefits is shorter than for standard employees in one Member State.

Table A.6: Time thresholds limiting effective coverage for the non-standard workers

Effective coverage can be problematic also for the self-employed despite being formally covered, because the eligibility conditions or rules are stricter than for standard workers. Waiting periods can be longer for the self-employed than for standard employees in case of unemployment in one Member State, and in eight Member States for what concerns sickness benefits. Or, the same waiting period can apply to self-employed and dependent employees. However, the latter are entitled to sick pay (i.e. they continue to receive their salary by the employers), which is not applicable to the self-employed for the nature of their working relationship, creating an effective difference in treatments. A minimum qualifying period in order to receive unemployment benefits has been identified as an issue in four Member States, in two for old-age and survivors' pensions and in one for invalidity benefits. Minimum working periods have been identified as an issue in three Member States concerning unemployment benefits, in two for old-age and survivors' pensions, in one for maternity and for invalidity and in one for accidents at work and occupational injuries. The duration of benefits is shorter for the self-employed with respect to standard employees in one Member State for unemployment, in two Member States for sickness benefits, in five Member States for maternity and in one for accidents at work and occupation injuries (see Table A.7).

Table A.7 - Time thresholds limiting effective coverage for the self-employed

1.3.Gaps in transferability

The risk of gaps in effective coverage for people in non-standard employment or self-employment does not only depend on the difficulties to fulfil eligibility criteria, but also on problems related to the transferability of social rights across schemes. People in non-standard employment and in self-employment can have problems in transferring their rights from a scheme to another, especially in Member States with highly fragmented social security schemes.

For our purpose, transferability relates solely to transfers across employment statuses and sectors, not across risk areas. This means that transferability is to be considered in situations where a person changes employment status (notably between dependent employment and self-employment) or economic sector and is covered by a social security scheme (scheme A) before and by a different one (scheme B) after the change, and where eligibility for a corresponding benefit (in scheme B) or the level thereof depends on past contribution history (length and/or contribution levels).

Lack of transferability is an issue for those branches of social security for which a minimum qualifying periods is required (this is especially the case of old-age pensions' schemes – see box A.5). In these cases, transferability can formally not be foreseen (i.e. there are no rules or laws which foresee this possibility), or there can be a problem related to effective transferability if the conditions are unfavourable (e.g. particularly high administration burden or costs). For example, in Italy, it is possible to accumulate the pension contributions made to different funds for receiving a single pension benefit. It is also possible to transfer the contributions made through the working life to a single fund, but this may be particularly expensive for the claimants of the benefit. In France, a note of the Pension Advisory Council 180 has shown that, even if in principle supplementary pension schemes are transferable, in practice, transferability only operates for contracts that obey to the same fiscal and social contributions treatment.

 Table A.8: Difficulties in transferability of social security rights

Note: the information have been collected by MISSOC experts for 21 out 28 EU Member States. Belgium, Bulgaria, Denmark, France, Romania, Sweden and Slovakia are not included.

1.4.Adequacy of the benefits

While lack of adequacy of benefits may be problematic for all forms of employment, it can be even more challenging for people in non-standard employment or in self-employment, which can "experience higher […] adequacy risk" 186 . This may importantly reflect income differences between groups, addressing which is beyond the scope of the current initiative, but can also be related to i) for the self-employed, the income assessment base taken into account for the calculation of benefits; ii) the length of contributory history; and iii) different rules for the calculation of benefits for different forms of employment (e.g. benefits can be defined as a lump-sum rather than in relation to earnings, as for salaried employment). Finally, differences in adequacy between groups could arise if at progressivity and reductions in social contributions for low-income groups do not apply equally to workers and the self-employed.

Income assessment base

Income assessment bases play a salient role regarding the accrual of entitlements to benefits for the self-employed. On the one hand, if the self-employed person pays social contributions on a very low income assessment base, s/he will receive a low level of benefits. On the other hand, if the reference base is too high, s/he may not be able to pay the required contributions. These issues may be related to the way of calculating the income base or reference income periods. Moreover, requirements for advanced payments may also hinder the self-employed person’s ability to build up entitlements. Income assessment bases calculated on long previous periods of earnings may not take into account the fluctuations characterising incomes from self-employment. This kind of estimation-based system may hinder the ability of the self-employed to pay social contributions if their current income decreases below the estimated one. At the same time, the freedom to choose (within certain limits) the level of contributions in some schemes can lead to low adequacy of benefits (see Box A.6 for an example).

Contributory history

Contribution histories can matter not only for effective coverage (via minimum requirements for their length, see the section above) but also for the adequacy of benefits in some social security branches. For instance, in some Member States the time profile of replacement rates in unemployment insurance depends on the contribution history. At the same time, the latter is very much linked to the form of employment of the individual (e.g. full-time vs part-time or temporary vs open-ended contracts).

Calculations of replacement rates for individuals with the OECD tax benefit model for unemployment benefits as well as housing and family benefits illustrate the major role of the contribution history in determining the amount of benefits and their evolution over time in some Member States (see Table A.9 below). For instance in Belgium an individual who contributed one year will have a total resulting replacement rate at the 18th month of unemployment which is 24.8 percentage points lower when compared to an individual who contributed five years. In some Member States the gap is large already after two months of receiving the benefits (i.e. Bulgaria, Lithuania, Hungary, Romania and Slovakia). Shorter contribution histories are likely to characterise workers on temporary contracts, to the extent that they also face more frequent interruptions by unemployment spells.

Table A.9: Replacement rate for unemployment, housing and social assistant benefits. Comparison 5 vs 1 year of contribution, 2016

Source: European Commission, elaborations on the basis of the OECD tax benefit model.

Different definitions of benefits for different forms of employment

In some Member States many self-employed may receive lower benefits compared to an employee with the same earnings because benefits are set at different levels according to the form of employment.

Cases identified by experts 196 include that in Belgium, the amounts of sickness benefits paid to the self-employed are equal to the minimum amounts paid to white-collar employees after six months of sickness. As a result, the adequacy of sickness benefits in cash is lower for the self-employed than for standard workers. Moreover, projections show a large decline in replacement rates for the self-employed by 2040. In 2010, a self-employed worker had a gross replacement rate of 77.2% and the rate for a salaried worker was 73.7%. In 2040, the figures would be respectively 47.3% for self-employed and 61.7% for employees 197 .

In Denmark, the self-employed receive lower unemployment benefit than people in salaried employment, with a shorter period of receipt of benefits. In Finland, on average, an employee earning €2,500 would get a monthly unemployment benefit of €1,500 (contribution-based), whereas the basic monthly allowance for the self-employed, with the same earnings, would be €697.

In Finland, compensation in the case of a work accident would be about €2,500 for the employee and €1,750 for the self-employed, or if the person is not eligible for income-related benefits, the universal minimum allowance would be approximately €622 a month.

1.5.Lack of transparency of entitlements and user-friendly information related to social security

A further aspect that should be taken into account when dealing with the problem of lack of social security coverage for people in non-standard employment and in self-employment concerns the issue of transparency and complexity of administrative procedures. The lack of easily accessible and clear information about rights and obligations relating to social security can hamper a clear understanding of working conditions and future risks. This may be of particular relevance for people in non-standard employment and self-employment who may face more frequent transitions and/or be covered by different set of rules. This situation is often produced by the complexity of social security schemes and the lack of easily accessible information systems, and is potentially compounded by the limited representation of the self-employed and non-standard workers in trade unions. A recent study identifies as causes of non-take-up of social benefits the inaccessibility and complexity of the procedures, the slow assessment, and the unawareness of people concerning benefits, entitlements and application procedure 198 .

Table A.10: Lack of information regarding social security schemes

Table A.10 above 199 provides a mapping about the availability of information regarding key social security systems in the Member States. Generic information about social security schemes is provided in most Member States, for instance via dedicated websites, although the information may not always be available broken down by employment status or contract. By contrast, the table reflects a wide-spread lack of transparency regarding personalised information, both regarding an overview of rights and obligations and the possibility to obtain information through online simulation tools. Such information is only available in about half of the Member States. Additionally, Box A.8 provides a more detailed overview regarding transparency in relation to pension rights.

A recent survey from Open Evidence shows the level of awareness of social security rights across people in different types of employment (Figure A.16) 206 . In general, the share of people unaware of their coverage is significantly higher among people which are not in standard employment. In general, it is 7% among full-time open ended employees, but 18% among own-account self-employed. Maternity benefits seems to be the ones for which people are more aware despite the status and the type of employment.

Figure A.16: Share of respondents unaware of their social security coverage, by type of employment

Source: Open Evidence

Invalidity and accidents at work and occupational injuries benefits seem to be the most problematic schemes in terms of transparency for non-standard workers, followed by old-age pensions' rights and unemployment benefits. 15% of temporary full-time workers are not aware of whether they have coverage regarding sickness benefits, while the percentage is lower in case of part-time (both open-ended and temporary). 16% of temporary part-time also seems unaware of whether they have coverage with respect to unemployment benefits (while the share both for open-ended part-time and temporary full-time workers is 13%).

Awareness concerning coverage regarding accidents at works and occupational injuries, and invalidity benefits is scarce also among the self-employed (21% for the self-employed without employees and 18% for the entrepreneurs are not aware of their entitlements). The share of those declaring to be unaware of whether they are entitled to old-age pensions is also high in the case of self-employed: 19% for the self-employed without employees and 17% for the entrepreneurs. Lack of awareness is also detected concerning unemployment benefits (again the percentage of self-employed that doesn't know if it is entitled to unemployment benefits is higher for the self-employed without employees, 17%, than for the entrepreneurs, 14%).

2.Behavioural biases

In the case of voluntary access to social security schemes, it is individuals' behaviour and choice that determine whether they have formal coverage or not. A number of factors can drive the decision not to adhere to systems, including whether coverage is already available through the first job in the case of voluntary access in the second job, the family context and private wealth accumulation which provide alternative means of coping with risks, as well as the individual's assessment of future risks facing them. This section explores how behavioural biases may contribute to low coverage rates observed in social security systems under voluntary access.

Aside possible financial disincentives influencing take up in voluntary schemes, individual preferences can also play a significant role in the way people adhere to existing social security schemes.

Individual's intertemporal distribution of consumption could be called myopic when the allocation of consumption to the present does not take sufficiently into account possible future risks. The tendency may be attributable to an erroneous foresight, or to a well-grounded optimism on the future, or to the fear that what has been promised for tomorrow might not materialise 207 . Researchers estimate that correctives for inadequate savings due to myopia could raise individuals’ lifetime welfare to an extent equivalent to almost an additional year’s worth of income 208 . Already in 1958, Samuelson found that individual myopia may produce suboptimal saving rates.

Myopic behaviour is increasingly identified by the academic literature to be a significant determinant of individual choices and is thus considered as a rationale for mandatory social security. 209 Especially concerning protection against longevity risk, workers are forced to save for retirement in order not to spend their accumulations without sufficient consideration for the future and thus end up in poverty.

Myopia can also limit the voluntary enrolment in social security schemes by people in non-standard employment and by the self-employed. Myopia can worsen individual judgment about future social risks and protection needs. Without comprehensive and clear information, individuals could prefer to forgo or reduce contributions when this possibility is allowed 210 .

Some results of the recent survey funded by the Commission can be interpreted as indication that a significant share of individuals can have this kind of myopic preferences, not giving enough importance to future risks (Figure A.17; for a full overview of the survey results see Annex 16).

Figure A.17: Importance of getting prepared for retirement now by current main employment status

Source: Open Evidence 2017

Individual preferences and behaviour also play a role in the assessment of the income base: the way in which the income assessment base is calculated can act as a disincentive to properly contribute or to legally declare the income. Where the rules for declaration of income are less stringent, or where the individual has the option to (legally) choose what level of income to be assessed upon, individuals may present the same myopic attitudes and erroneous foresight and lead them to under-report their present income. This under-reporting of income and the corresponding underpayment of contributions may prevent effective coverage or reduce adequacy (e.g. Bulgaria, Hungary, Estonia, Lithuania, Latvia, Romania, and Slovakia) 211 . For example, in Romania, due to both high contributory rates and the method of calculation of the assessment base only few self-employed contribute to social insurance. Indeed, the self-employed represent the 17.6% of employed people, but 90% of them are not covered for old-age pensions, invalidity, sickness or maternity benefits because of under-insurance 212 .

Annex 7: Overview of recent reforms in Member States

Annex 8: The existing EU instruments

The EU acquis related to the rights and obligations in relation to social security and protection is implemented through a variety of legislative and non-legislative provisions. However, none of these fully address the rights and obligations of people in non-standard employment and forms of self-employment.

The Open Method of Coordination (OMC), the European Semester and benchmarking

The Recommendation 92/442/EEC 227 on 'Convergence in social protection objectives and policies' identified commonly held objectives in the area of social protection and many initiatives in this field which followed can be seen as continuous process of implementation, monitoring and follow-up of this Recommendation, leading to the Social Protection Committee and the Social OMC. Together with the Recommendation on "common criteria concerning sufficient resources and social assistance in social protection systems" from the same year 228 , the 92 Recommendation had great impact on development and convergence of social protection systems in decades which followed.

The Recommendation was successful in establishing the firm bases for the Treaty of Amsterdam in 1999 where objectives started to be transformed into obligations 229 , led to the establishment of the High Level Working Party on Social Protection which was replaced by newly established Social Protection Committee in 2000, as the first body with the specific purpose of monitoring developments, promoting exchanges of information and good practice, and preparing an annual report on social protection. It offered also a basis for establishment of the Social OMC, particularly because of the resemblance of the methods of implementation.

One of the reasons why it was so influential is the success in setting up a strong dialogue between Member States and identifying common issues in social protection systems. In particular, the debate established a consensus among Member States and the EU institutions that social protection systems need to be modernised 230 . The reason why the Recommendation was so effective in raising the interest among Member States to exchange practice and actively participate in such debate might be its 'inviting', non-obligatory nature of the Recommendation and its aim of keeping the important balance between providing guiding principles in development of national systems and acknowledging that Member States remain free in determining how their systems should be organised and financed.

While the Recommendation gave a detailed rendition of many aspects of social protection and did so by making reference to ‘workers’, it did not mention explicitly people that fall outside the standard worker category. When it comes to the self-employed, the recommendation merely states that Member States should 'examine the possibility of introducing and/or developing appropriate social protection for self-employed persons.'

Later on, the common objectives and indicators on coverage of adequate and sustainable pensions, health care and long-term care set out in the context of the Social Open Method of Coordination 231 also gave little attention on how such coverage functioned in the context of labour law and types of employment.

Finally, 'Europe 2020' built on all of these steps when, in connection with the European Semester, it became the EU policy framework for policy coordination in the field of social protection with concrete and defined aims, implementation methods and monitoring systems.

EU Labour law: directives on non-standard forms of employment, on protection of employees in case of employers' insolvency and on working conditions

Over the years, an EU legal framework for protection of rights in some areas of non-standard employment has been developed through the Directive on Temporary Agency Work 232 and through negotiated agreements between social partners and implemented via the Directives on Part-Time Work and Fixed-Term Work 233 . But as the Directives apply to employment conditions, their impact on social protection is rather limited. While social protection closely linked to employment as part of pay, such as occupational pensions 234 , maternity pay and sickness pay is covered in these Directives the Court of Justice has made clear in its case-law that the Directives do not apply to statutory social security schemes 235 . In the preambles to the Agreements the social partners call on Member States to ensure that social protection arrangements are adapted to fit with evolving flexible forms of work. These preambles are not legally binding and, while Member States on several occasions have stated their commitment to make social protection more employment friendly, a recent review of the extent to which this has happened suggests that in several Member States there is ample room for improvement 236 .

The Directive 2008/94/EC on protection of employees in the event of employer insolvency covers protection of social security benefits under national statutory social security and supplementary occupational or inter-occupational pension schemes.

The Directives on work in specific sectors also provisions related to social protection. For instance, article 31 of Directive 2017/159 on the Work in Fishing Convention 237 establishes the right of fishermen to compensation in the event of injury due to occupational accident or disease. Similarly, the current maritime proposal includes a provision on social protection of seafarers on the compensation in case of death or long-term disability for occupational reasons.

Directives on equality between women and men 

The EU legal framework on equality between women and men guarantees a number of rights in the field of social protection. The Directive 1979/7/EEC on the progressive implementation of the principle of equal treatment for men and women in matters of social security covers statutory schemes which provide protection against the risks of sickness, invalidity, old age, accidents at work and occupational diseases and unemployment, as well as social assistance, in so far as it is intended to supplement or replace the statutory schemes. Occupational pensions are covered by the Directive 2006/54/EC on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation (recast). The EU legal framework also guarantees a number of rights in the field of maternity leave and parental leave.

Directive 2010/41/EU on equal treatment between men and women engaged in self-employed activity

At EU level, one step towards improving the protection of individuals engaged in self-employment and the coverage of social protection for spouses of the self-employed workers has already been made. The Directive 2010/41/EU 238 on equal treatment between men and women engaged in self-employed activity grants women (including spouses of the self-employed) the right to be entitled to maternity leave and benefits for at least 14 weeks; levels of benefits and eligibility conditions vary widely among Member States. Furthermore, where a system for social protection for self-employed workers exists in a Member State, it requires providing coverage also to spouses of self-employed workers engaged in their activities.

Social security coordination

Ever since 1959, the social security systems of the Member State have been coordinated so that persons moving within the Union could avoid losing their social security protection. However, Member States remain competent to determine the conditions for becoming affiliated to the social security system, provided that these conditions are not discriminatory.

The rules on social security coordination apply to national legislation on sickness, maternity and equivalent paternity benefits, old-age pensions, pre-retirement and invalidity benefits, survivors’ benefits and death grants, unemployment benefits, family benefits, benefits in respect of accidents at work and occupational diseases 239 .

Self-employed persons in the social acquis

The notion of "self-employed person" is not defined at EU level. Depending on the objectives pursued by different EU instruments, self-employed persons are covered or not.

Some directives concern only the employed person as they treat issues linked to employment contracts and are not relevant for self-employed persons (e.g. Directives 89/391/EEC, 91/533/EEC, 1999/70/EC, 96/71/EC). Other EU law provisions apply to both categories of persons, either by applying the same rules to self-employed as to employed persons (especially the current acquis implementing the principle of equal treatment, e.g. Directives 2000/78/EC, 2006/54/EC, 1979/7/EEC), or by setting different rules for self-employed persons (EU rules on social security coordination 240 ).

Directives related to supplementary and occupational pensions

The Directive 2014/50/EU on the acquisition and preservation of supplementary pension rights lays down rules aimed at facilitating the exercise of the right of workers to freedom of movement between Member States by reducing the obstacles created by certain rules concerning supplementary pension schemes linked to an employment relationship. The Directive (EU) 2016/2341 on the activities and supervision of institutions for occupational retirement provision requires that every institution provides both general and personalised information to the beneficiaries.

Annex 9: Analytical methods

The estimation of social and economic impacts include several sets of microeconomic simulations, to quantify the impact of the measures, as well as more qualitative analysis, allowing to better understand behavioural reactions and attitudes (through a survey) and country specific context (through case studies).

1. Evidence, sources and quality

The following expert advice has fed into the Impact Assessment:

-From the Report of the public consultation "Accompanying the document establishing a European Pillar of Social Rights (SWD(2017)206)".

-From the commissioned study "Access to social protection for all forms of employment: Assessing the options for a possible EU initiative" by Fondazione Giacomo Brodolini.

-From the commissioned study "Behavioural study on the effects of an extension of the access to social protection for people in all forms of employment" by Open Evidence.

-From six commissioned case studies on special forms of employment (Mini-jobs in Germany, civil law contracts in Poland, self-employed without employees in the Netherlands, casual worker in Romania, economically dependent self-employed in Spain, project workers on continuous collaboration projects in Italy)

-From literature review as referred to in footnotes.

2. Quantification of impacts - simulations on the basis of the EU-SILC database

The microeconomic simulations were carried out by using the latest available EU-SILC cross-section and panel data. Two complementary approaches were used: simulations in EUROMOD, the EU's tax-benefit microsimulation model, and ad-hoc simulations outside this modelling framework, respectively.

2.1 EUROMOD simulations

Simulations based on EUROMOD, the EU's tax-benefit microsimulation model, were commissioned by DG Employment. Further details can be found in the above-mentioned corresponding study, Fondazione Brodolini (forthcoming). EUROMOD was used to simulate the fiscal and distributional impacts of extending formal coverage by unemployment insurance to the self-employed in five Member States where they are currently excluded from the corresponding schemes, namely Greece, Croatia, Italy, Latvia and the Netherlands. Simulations were based on EU-SILC data for 2016. Available data did not allow performing similar simulations for non-standard workers since the data set does not distinguish between standard workers and non-standard workers.

The simulated unemployment shock was defined using historical data. Namely, the average transition rate from self-employment to unemployment or inactivity for 2007-2015 from EU-SILC data was considered, focussing on those self-employed workers (in year t-1) who become unemployed or inactive as a result of the unemployment shock (in year t). The self-employed are defined as individuals with positive self-employment income and no (dependent) employment income. From that sample, those subject to the unemployment shock are randomly selected. Table A.11 below reports on the unemployment risk underlying the simulated unemployment shocks.

Two scenarios were simulated:

- Scenario 1 ("earnings-related" scenario in the main text) assumes that recipients qualify for the unemployment benefit currently in place for employees if entitled, given their characteristics, and for as many months as dictated by the programme rules currently in place.

- Scenario 2 ("flat-rate" scenario in the main text) assumes that all formerly self-employed workers (in year t-1) who are either unemployed or inactive (in year t) qualify for a flat-rate unemployment benefit at 60% of median income, paid for at least as many months (in year t) as they had been self-employed (in year t-1).

In both scenarios, the rules of the systems in place are also extended on the financing side: self-employed workers pay the lowest unemployment insurance contribution rate currently applying in the Member State's existing system to employees and their employers. The corresponding information was retrieved from the MISSOC database. Table A.11 below presents the resulting contribution rates.

Reported results compare the outcomes in each scenario to the baseline. This is defined here as the counterfactual situation that might prevail after the unemployment shock and under the current state of affairs of no unemployment insurance coverage (and no unemployment benefit) for self-employed workers.

Table A.11: Unemployment risks and contribution rates 241

Distributional results relate to percentage point changes in the at-risk-of-poverty rate for several types of households, with respect to a poverty threshold that is kept constant at 60% of median equivalised household disposable income in 2016.

2.1 Ad hoc simulations for all Member States

The aim of the ad-hoc simulations is presenting some estimates about the budgetary costs that would be due to the provision of comparable unemployment benefits (henceforth UB) and sickness benefits for self-employed workers (henceforth SE) across all EU28 Member States. 242

The reported simulations are based on hypothetical scenarios where the same type of (earnings-related or flat-rate) UB scheme is introduced in all EU-28 countries, regardless of the design of existing schemes. Based on the estimated number of self-employed that would be eligible for benefits according to the evidence recorded in EU-SILC about the SE risks to become unemployed, corresponding benefit payments to them are simulated. Aggregating these payments provides an estimate of resources needed.

To get a magnitude of the estimated costs of these reforms, the total “gross” spending for the new simulated UB scheme for SE is expressed as a share of annual gross labour incomes of SE in the reported results, which can be interpreted as a contribution rate 243 .

This rate has to be interpreted as a gross rate in the sense that possible spending already existing for UB paid to SE is not taken into account and, thus, not subtracted from the simulated cost of the new UB scheme. To infer a “net” rate, the ratio between the unemployment benefit already paid to the self-employed and the corresponding wage bill should be subtracted from the “gross” rate.

It should be noted that the simulations reflect a “partial equilibrium” setting. This means that the model does not allow to capture behavioural reactions the introduction of benefits nor does it capture secondary effects due to the interplay between a possible change in contribution rates and received benefits or the other components of the public budget (in terms, for instance, of personal income tax, tax deductions or other means tested benefits).

Two types of benefits are simulated:

1.A replacement rate scenario: The unemployment benefit is paid according to the number of months spent in unemployment or inactivity for reasons different from retirement during 2014 for a maximum of 6 months. The amount of the benefit is equal to 50% of previous monthly gross labour income with a maximum monthly amount equal to 60% of national median gross labour income (from employment and self-employment).

2.A minimum level scenario: The unemployment benefit is paid according to the number of months spent in unemployment or inactivity for reasons different from retirement during 2014 for a maximum of 6 months. The amount of the benefit is equal to 30% of the national median gross labour income (from employment and self-employment), independently on previous individual monthly gross labour income as a self-employed.

Therefore, scenario a) is more consistent with the rules about unemployment benefits that apply to employees in most countries, while scenario b) is more consistent with the adoption of a minimum level of mandatory social security.

Sensitivity analysis was undertaken where the benefit was paid for 4 and 8 months instead of six or where the benefit level varied. Notably, the earnings-related benefit was set at 40% or 60% of previous monthly gross labour income instead of 50%. The flat-rate benefit was set at 20% and 40% of the national median gross labour income instead of 30%.

Estimates have been carried out by using a sample built merging 7 cross-sections of EU-SILC (2009-2015 cross-sections, i.e. covering the period 2008-2014), that, following harmonised definitions of the relevant variables, record workers’ self-reported statuses (i.e., employee, self-employed, unemployed or the various inactivity conditions) in each month of a certain year and also record annual gross labour incomes from employment and self-employment.

In the simulations, for each year of the period 2008-2014 recorded in EU-SILC 2009-2015 cross-sections, the focus is on those who were self-employed (both full-timers and part-timers) at January of each year and, being tracked their working status since February up until to December of that year in EU-SILC, measure the cost to cover these workers’ subgroups against unemployment risks is measured. To additionally consider the annual cost for paying unemployment benefits to those who have been self-employed in the previous and are still entitled to receive an unemployment benefit at the beginning of the next year, the results reported here add to the annual cost the spending for those who are still unemployed at the end of the year and have not exhausted the total period of receipt of UB, thus assuming they will not be again employed before the period of receipt of UB expires 244 .

The results reported consider as becoming unemployed those self-employed in the data who move into unemployment or inactivity for reasons other than retirement throughout the year. This broad definition is chosen in response to a possible underreporting of unemployment among the self-employed in EU-SILC data. It may allow to reduce a possible underreporting of unemployment risks for SE due to the fact that, being absent UB schemes for them in many countries, those who stop to work can answer in the SILC interviews to have become an inactive rather than an unemployed merely because they are discouraged, even if they would like to find a new job. Actually, for many not retired individuals, when they answer about their self-reported employment status the difference between the statuses of unemployed or inactive is rather slight and may depend on an individual feeling rather than on an objective status. And this holds true especially for SE when – being almost always absent UB schemes – a sort of official definition of unemployment, that makes the individual entitled to receive an UB, is not perceived by the them.

The report contains more details on the steps of the simulations, where each step has been run at the country level and pooling together the 7 cross-sectional waves, thus providing an average cost over a 7-year period.

Ad-hoc simulations on the extension of sickness benefits to the self-employed:

Differently from unemployment benefits, which can be simulated using information on periods spent in unemployment or inactivity, there is no information in SILC on which self-employed are more likely to receive a sickness benefit, if this benefit were provided. The contractor therefore simulated possible sickness benefit for the self-employed imputing benefits according to their characteristics, exploiting knowledge on employees who receive such benefits. More details can be found in the report, section 3.2.7.

3. Qualitative assessment

3.1. Comprehensive literature review and qualitative evaluation

The qualitative assessment of impacts is based on comprehensive literature surveys and a country expert questionnaire in 16 Member States (Denmark, the Netherlands, Ireland, Belgium, Germany, Latvia, Romania, Italy, France, Finland, Sweden, Spain, Greece, Poland and Slovakia). The questionnaire focuses on the most likely economic and social impacts of policy options through a qualitative evaluation made by national experts with reference to those social provisions where there is a lack of formal access for workers in non-standard forms of employment and self-employed or where the de facto coverage is highly limited.

More generally, as highlighted above, the use of quantitative tools to assess the impacts of the initiative implies a certain asymmetry given that available models capture more easily the costs than the benefits associated with the initiative. Special care will therefore be taken to highlight in a qualitative manner the benefits associated with the extension of social security to the previously unprotected groups in the labour market (in terms of changes in income distribution, labour market dynamism, sustainability of the welfare state). To give a complete picture, the qualitative analysis also points to additional negative effects not captured by the available quantitative models. For instance, there could be a reduction in labour demand, due to an increase in labour cost.

3.2 Behavioural survey

Furthermore, a behavioural survey based on an overall sample of 8000 people was conducted in 10 Member States (Germany, France, The Netherlands, Italy, Spain, Slovakia, Romania, Sweden, Poland and Belgium) to find out whether an extension of access to social protection would affect, and how, the behaviour of workers in non-standard forms of employment and self-employed. Respondents were recruited through an online panel so the sample is representative of the Internet population, but not representative of the population in general. The survey combines traditional online survey research and online experiment. It explores attitudes towards the existing national social protection systems (effectiveness in protecting people in different situations) and people's preferences (degree of risk aversion strength of preference for having social protection coverage as well as willingness to pay for it) as well as behavioural implications (whether availability of social protection would make self-employment more attractive, importance of transferability of rights). Socioeconomic information about the respondents was also collected, so as to control to specific characteristics of the individuals. Internet was used as a methodological tool for data gathering, using online panels in the selected countries. Annex 16 presents the main results of the survey.

3.3 Case studies

Finally, six case studies were conducted by external contractors to study relevant specific contract types in particular Member States:

·DE: Mini-jobs

·NL: self-employed without employees

·PL: civil law contracts

·RO: casual workers

·ES: economically dependent self-employed

·IT: workers on continuous projects (co.co.co.)

In the first place, the studies are used to describe the nature and the extent to which special forms of employment in the six concerned Member States are being used as a 'new form of non-standard work'. They also illustrate gaps in the de facto ability of people in these forms of employment to acquire adequate protection through social security systems.

In a second step, the case studies inform about the possible existence of derived rights (e.g. free health insurance for spouses, survivors pensions) social protection of people working in these special forms of employment and the take-up of social protection benefits at household level (e.g. widow pension). They also highlight the extent to which gaps in access to statutory social protection are compensated by other means (e.g. private health, pension or life insurance, position in the wealth distribution).

Finally, the case studies look at the costs of these forms of employment and their estimated potential benefits.

Annex 10: Further information on impacts of policy measures and policy packages

This Annex complements Chapters 5.4 and 6.2 on the impacts of the proposed policy measures and packages, respectively. It comprises four sections. Section 1 presents evidence on the impact of opt-out and opt-in clauses for enrolment in voluntary schemes. Section 2 provides details on the available quantification of the impact of measures to extend formal coverage and puts them in perspective. Finally, Section 3 presents an exemplary quantification of possible impacts of relaxing qualifying periods as a measure addressing effective coverage. Finally, Section 4 presents an overview table summarising main impacts of the policy packages.

1. Voluntary coverage under opt-in vs. opt-out clauses

Voluntary coverage by social security schemes can be implemented via opt-in clauses or opt-out clauses. Opt-in clauses require potential beneficiaries to take action to choose coverage whereas opt-out designs imply being covered is the default option, with action needed to avoid being covered. Experiments and observational studies show that making an option a default increases the likelihood that it is chosen (default effect). In the context of insurance, there is evidence that when consumers are faced with car insurance options that are very similar, except for the possibility to opt-in vs. opt-out of additional coverage, the former design option results in much lower levels of participation than opting-out (20% vs. 75% in a seminal study) 245 . In a similar vein, a large empirical literature shows the importance of default options in shaping decisions about retirement savings including savings plan participation and levels of contributions 246 . The following empirical evidence from Member States underlines the point regarding social security coverage:

In the UK pension system, flat-rate state pensions providing basic old-age income are complemented by voluntary contribution-based workplace pensions, set up or facilitated by the employer for their employees. To reverse the observed long-term trend decline in coverage, from 2012 the UK progressively switched to an opt-out system from the previous opt-in one. Workplace pension scheme membership subsequently increased from 50% in 2013 to 64% in 2015. The employee opt-out rate was 8% to 11% among larger sized employers and reached 17% in smaller firms.

In Germany, non-standard employees on a mini-job contract have voluntary access to the public pension system. In 2013, the design was changed from an opt-in to an opt-out option. The default conversion resulted in a tripling of the share of mini-jobbers covered by the pension system, albeit to a still low level (from below 6% to 18%) 247 .

As these two examples demonstrate, setting defaults can be an effective way of influencing behaviour also in a context of providing social security coverage.

2. Micro-simulations for the quantification of the impact of measures to extend formal coverage

Micro-simulations have been commissioned to Fondazione Brodolini 248 for the purposes of this Impact Assessment in order to estimate the economic costs and benefits of extending unemployment and sickness benefits to the self-employed. An introduction to these simulations is provided in Box A.9 (for further methodological details please see Annex 9) before their set-up and results are presented and discussed in more detail.

a) Extending unemployment insurance

Simulating the impact on public budgets of an extension of existing unemployment benefit systems

The simulations show the fiscal and distributional effects of extending unemployment insurance to the self-employed in a situation in which some reforms led to a homogenisation in a given Member State between employees and self-employed of rights to unemployment benefits and corresponding contributions. The exercise was carried out using the EU's tax-benefit model EUROMOD for five Member States (Croatia, Greece, Italy, Greece 250 , Latvia and the Netherlands) where the self-employed are currently generally excluded from coverage, since these Member States may be facing the most pronounced impacts.

The key scenario considered a simulation where coverage was extended to the self-employed under the same rules for earnings-based benefit entitlement in place for employees in each Member State - "earnings-related scenario". Rules of the existing systems regarding eligibility as well as the calculation of benefits are thus applied to the self-employed who become unemployed 251 . Results suggest that, comparing with a baseline situation where the same unemployment risk prevails but no unemployment insurance is in place, extending unemployment insurance could lead to additional total gross expenditure for benefits ranging from around 0.1% (NL, LV, FI) and 0.2% (GR) to 0.4% (HR, IT) of GDP annually, depending inter alia on the number of self-employed in the country, the unemployment risk they face 252 as well as on the generosity of the system in place (see Table A.12, column "Gross Spending / Earnings-related scenario").

As a first sensitivity analysis, to study the effects across these Member States of extending access to an insurance of similar generosity, a second scenario - "flat rate scenario” - is considered, where a flat-rate benefit is paid to the previously self-employed who have become unemployed, independently of previous labour income as a self-employed. The simulations considered a flat-rate benefit corresponding to 60% of national median disposable income 253 to the self-employed. The overall range of gross expenditure remains broadly unchanged at 0.1% - 0.4% of GDP, although for individual Member States (IT, LV), resulting expenditure was higher than in the scenario based on actual (less generous) systems in place 254 (see Table A.12, column "Gross Spending / Flat-rate scenario").

Table A.12: Simulated gross and net spending on unemployment benefits for the self-employed in % of GDP 255

These simulations for the two scenarios assume no behavioural reactions by those affected by the changes. The estimated impacts should therefore be considered as conservative; for instance increased underreporting or decreases in employment could lead to higher expenditure 256 . 

Depending on financing decisions by Member States, the simulated expenditure would only partially translate into costs for the public budget. For example, if the self-employed were requested to pay social contributions rates corresponding to the lowest level currently levied on employees, 257 parts of the public expenditure on benefits could be covered in this way. In this example, the estimated net public cost would not exceed 0.3% of GDP (HR), 0.2% (IT), 0.1% of GDP (LV, NL). No positive cost remains for Greece (see Table A.12, columns "Net Spending") 258 .

Simulating the cost of comparable unemployment benefits to the self-employed for all EU28 Member States

Complementary simulations were undertaken for all Member States for the introduction of a comparable system of unemployment insurance to the self-employed, regardless of the existing systems. So to facilitate the comparison of impacts previously self-employed unemployed are assumed to receive a benefit calculated according to the same rules across all Member States. Based on the estimated number of self-employed that would be eligible for benefits according to the evidence recorded in EU-SILC about the self-employed risk to become unemployed, corresponding benefit payments to them are simulated. Aggregating these payments provides an estimate of resources needed.

Notably, two sets of simulations were run involving two different options for coverage: in the first one, benefits amount to 50% of previous earnings ("replacement rate scenario"), in the second one there is a modest flat-rate benefit amounting to 30% of national gross median income ("minimum level scenario") 259 .

To get a magnitude of the estimated costs of these reforms, the total gross spending for the new simulated unemployment benefit scheme for the self-employed is expressed as a share of annual gross labour incomes of the self-employed, which can be interpreted as a contribution rate. 260 It is a gross rate and presents an upper bound in the sense that possible spending already existing for unemployment benefits paid to self-employed in Member States where these are covered is not netted out from the estimate 261 .

Figure A.18: Gross contribution rate to finance the unemployment benefit scheme for self-employed - Replacement rate scenario