copy the linklink copied!7. Left on your own? Social protection when labour markets are in flux

This chapter assesses challenges for social protection policies in a changing world of work and presents evidence of support gaps in public policies affecting different types of workers. Key policy challenges include a greater need for support resulting from greater employment instability or lower earnings among some groups; a reduced accessibility or adequacy of social protection measures that were designed around stable forms of dependent employment; and sustainability challenges, e.g. due to opportunities for avoiding participation in risk-sharing provisions. Accessing adequate support can be especially difficult for workers in less secure forms of employment. But support gaps are small in some countries that adopt fairly different social protection strategies, suggesting that accessible support can be achieved with different blends of social insurance and means-tested assistance. The chapter discusses alternative reform avenues and illustrates country approaches to prepare income support and reintegration measures for the future of work.


The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

copy the linklink copied!
In Brief
Key findings

Social protection systems play a key stabilising role, especially in the current context of heightened uncertainties about the pace and extent of labour market changes. The digital transformation will no doubt create many new opportunities, but will also render a growing number of current workers’ tasks redundant and will require substantial restructuring. There is evidence suggesting that these trends are already making job losses and employment changes more frequent for many workers (Chapters 2 and 3), increasing their needs for income and re-employment support. Effective social protection provides a buffer against the individual and social costs of these adjustments and can ensure that those losing their jobs have the time to find good job matches or undertake training if needed. In doing so, it can also counter calls for policy responses that stifle economic dynamism, such as creating barriers to trade or innovation. At the same time, the future world of work presents distinct and sizeable challenges that may undermine the prevention, protection or promotion capacities that have guided the development of present-day social protection systems. This chapter examines the nature and extent of these challenges and discusses reform avenues to tackle them.

The main findings are:

  • Future technological and labour market developments are inherently uncertain. But this can be no excuse for delaying reforms that are needed to make social protection future ready. Changing employment and income risks and elevated uncertainty, including about governments’ policy responses to labour-market changes, highlight the vital role of social protection in stabilising incomes and managing risks. But they also increase the individual and social costs of ineffective or inaccessible protection.

  • Some social protection systems are not well prepared for the faster pace of job reallocation (the destruction and creation of jobs in different firms and industries) that is likely to accompany the adoption of new production technologies. For instance, in a majority of OECD countries, fewer than one third of jobseekers receive unemployment benefits.

  • Technological advances make alternative work arrangements a viable option for a growing share of jobs and provide opportunities for organising work through contractual arrangements that may bypass traditional employer-employee relationships. Legal safeguards and social protection provisions that were designed around traditional forms of employment may no longer apply to workers with “non-standard” contracts, or not to the same extent. This not only creates inequitable, and possibly regressive, treatment of workers based on their employment status but also erodes the financial sustainability of social protection provisions.

  • Accessing social protection can be especially difficult for workers in less secure forms of employment even though their need for support is often particularly urgent. In some countries, workers engaged in independent work or short-duration or part-time employment are 40-50% less likely to receive any form of income support during an out-of-work spell than standard employees (e.g. Czech Republic, Estonia, Latvia, Portugal and Slovak Republic). Accessibility gaps can be especially large for the self-employed. For non-standard workers who do receive support, the level of benefits that are available during an out-of-work spell are often markedly lower than for standard employees (e.g. Greece, Italy, Slovenia and Spain). Unless access gaps are closed, further increases in non-standard employment will have negative consequences for inclusiveness and equity.

  • Pension coverage also tends to be less comprehensive for non-standard workers than for regular employees, exposing them to greater risks of low income and poverty in old age. In many countries, the self-employed can (partially or fully) opt out of pension schemes that are mandatory for dependent employees. In some countries, contributions are entirely voluntary for most self-employed. In others, mandatory contributions are lower for the self-employed than for employees (Austria and Portugal), or feature options that allow the self-employed to reduce their (mandatory) contributions (Poland and Spain). Each of these provisions leads to reduced future pension entitlements.

More volatile career patterns or a growing diversity of employment forms pose specific challenges for social protection provisions that link support entitlements or financing burdens to past or present employment. But future labour markets nonetheless leave room for a range of different social protection strategies.

  • Most countries’ social protection systems employ a mix of different design principles, such as targeting on current needs (means-tested assistance benefits), conditioning on past employment (earnings-related insurance benefits), or providing flat-rate entitlements (universal and unconditional support).

  • In the context of rapid job reallocation and a growth of alternative work arrangements, these provisions translate into specific policy challenges that differ between countries. Central social protection pillars, such as insurance or income-targeted assistance, will remain viable but they will need to adapt to new and changing risks.

  • Yet, given countries’ specific labour markets and institutions, a pursuit of generic policy prescriptions, such as a universal basic income or an exclusive reliance on last-resort safety nets, may be counter-productive as it can distract attention from positive reform steps that countries can take in the context of existing social protection strategies.

Many countries are actively assessing the challenges that automation and changing working arrangements pose for social support systems. Countries’ reform experiences provide valuable pointers to policy options and priorities. Some of the challenges concern “legacy” issues that are not new in the social protection debate. But the prospect of future labour market transformations often makes tackling them much more urgent.

  • Key priorities include: i) the correct classification of workers’ employment status (Chapter 4); ii) entitlement criteria that respond readily to changes in people’s need for support; and iii) making social protection rights portable between sectors or jobs.

  • Social protection provisions can themselves contribute to a growth of non-standard employment. For instance, in the Netherlands, the total employment cost for a dependent employee can be 60% higher than for an otherwise similar independent contractor. Different contribution burdens for different workers have existed for a long time in many countries and lower contributions can reflect specific risk patterns and fairness considerations. But when alternative forms of employment, such as contract gig work, micro jobs or casual work become more readily available, large disparities in labour costs between contractual arrangements are more likely to act as powerful drivers of employment and hiring decisions than in the past.

Novel forms of employment are also blurring the distinction between in-work and out-of-work categories. This raises new questions about the scope and ambition of employment-oriented social protection and activation (measures to strengthen people’s chances of re-employment and lessen any disincentives to work).

  • Comprehensive and tailored employment-support packages can be difficult to access for those in alternative work arrangements, reducing their chances of benefiting from the career opportunities that dynamic labour markets offer. Income support typically serves as the main gateway to labour market reintegration measures, and tackling gaps in income support provisions is therefore key in this context.

  • In addition, a careful review of the design and implementation of activation approaches should ensure that active labour market programmes remain well adapted to the needs and circumstances of jobseekers. A growing number of “part-time unemployed” (jobseekers with intermittent or part-time employment, including low-paid independent work) may call for a greater focus on facilitating advancement to good-quality employment, e.g. by shifting resources from work experience programmes or direct job creation towards job search assistance, tailored training and career counselling.

  • Policy makers should also review whether existing activation and reintegration measures yield an appropriate balance of demanding and supporting elements that is in line with policy objectives for job quantity and quality. For instance, enforcing job-search responsibilities and other requirements for the part-time unemployed may be a necessary counter-weight to extending benefit rights to these groups.

  • Likewise, with broadening options for when and how long to work, removing disincentives that discourage better-paid and more stable work becomes an increasingly pressing priority in the future world of work.

Adapting social protection to the future of work is likely to create additional financing strains at a time when social protection budgets are already under pressure in many countries.

  • Keeping funding levels in line with evolving needs for support requires a determined and co-ordinated approach and a policy debate on how new or expanded initiatives will be paid for, and who should pay, especially if evolving production technologies exacerbate a decline in the shares of national income that go to workers.

  • Key policy levers for tackling a shortfall of social protection resources include not only a suitable balance of revenues from labour and non-labour tax bases, but also the cost-effective delivery of support, as well as better revenue collection technologies and enforcement. Digital technologies are no panacea for increased efficiency but they can play an important role in these efforts.

  • Ensuring that social protection systems remain fiscally sustainable also calls for limiting opportunities for opting out of collective risk-sharing provisions, and for tackling unintended incentives that distort employment, hiring or layoff decisions.

copy the linklink copied!Introduction

The past 10-15 years have produced a remarkably broad global consensus that well-designed social protection (SP) is essential for inclusive labour markets and growth, and that it must be strengthened in order to meet global development goals  (OECD, 2018[1]; OECD, 2018[2]; ILO and World Bank, 2016[3]; United Nations, 2015[4]; European Commission, 2017[5]). But, in addition to longer-standing SP challenges linked to globalisation and population ageing, technological advances and associated changes in the world of work also give rise to concerns that existing SP strategies could be fundamentally compromised in some countries – e.g. World Bank (2018[6]), European Commission (2018[7]) and ILO (2018[8]). This chapter discusses the challenges that changing work patterns and an uncertain future of work pose for SP policies, and for those in need of support. It first presents new evidence on the readiness of existing SP systems to extend support to groups who face labour market risks as a result of digitalisation, automation and new forms of employment. The chapter provides concrete and “people-centred” indicators of the support that SP provides for individuals in specific labour market circumstances. The second part of the chapter reviews options to strengthen the prevention, protection and promotion functions of SP policies in different country contexts.

Automation, technology-assisted divisions of labour and algorithmic workforce management, e.g. through online platforms, have already brought fundamental change to production processes and to the world of work, and they will continue to do so. Eventually, as productivity gains and income growth feed through the economy, the resulting opportunities should be shared more widely. But productivity-enhancing technologies will not immediately and automatically lead to inclusive growth.1

The labour-market transformations that are underway add substantial uncertainty to workers’ careers and their incomes. Perceptions of income insecurity are driven by uncertainties on several levels, including the likely size of any gains and losses, their timing, and also the pace and direction of policy responses to those changes. Well-designed and accessible SP helps workers and their families succeed in a volatile world. Countries with more effective and employment-oriented SP systems are in a good position to cope with rapid change and to harness the resulting opportunities for shared and inclusive growth. But the labour market changes resulting from automation and new forms of employment also pose significant challenges if SP provisions fail to adapt to the evolving realities of the world of work.

This chapter explores gaps in existing social support provisions and those that may result from ongoing labour market transformations. Section 7.1 provides an overview of the main channels through which an acceleration in the job reallocation process and more varied and fragmented employment patterns may alter the functioning of SP and impact its effectiveness. Section 7.2 presents evidence on SP gaps between standard employees and workers in non-standard forms of employment. It summarises information on the statutory rules regarding SP for non-standard workers and discusses reasons why non-standard workers may not receive support even if statutory rules do not formally exclude them. The section then presents new empirical evidence on the level of support that standard and non-standard workers actually receive in practice. Section 7.3 considers options for how to address SP gaps, and for preventing SP from becoming less adequate in the future. The discussion is structured around illustrations of policy initiatives that countries have considered or implemented.

copy the linklink copied!7.1. Prevention, protection, promotion: Social protection and the future of work

SP helps individuals and families to manage risks and provides support to make economic or social disadvantage less concentrated on specific regions or groups, and less damaging for people’s longer-term prospects. It also seeks to maintain and improve living standards and to lower the costs associated with job reallocation by pooling labour market risks. From an economy-wide perspective, risk pooling, income-smoothing, redistribution and enabling support foster resilience against systemic uncertainties, including those related to the speed and magnitude of future labour market transformations.

SP can also act as a “backstop” to other policy levers, such as skills policy and labour market regulation (Chapters 4 and 6). While those policies seek to further workers’ employability and protect their interests, adapting them takes time. Necessary reforms may therefore lag behind rapidly evolving labour market realities. For instance, low-skilled workers are already the least likely to receive training and, where production processes become more fragmented and job tenure falls, firms’ and workers’ incentives to invest in firm-specific skills may weaken further. Accessible SP constitutes an additional layer of support when other policy channels are not (yet) effective.

Labour market transformations alter the functioning and the effectiveness of existing SP provisions through several channels. More volatile labour markets with growing under-employment (Chapter 3) and unequally distributed labour market risks produce growing demand for income protection and employment support. The financing of such support relies in large part on contributions or taxes levied on incomes from work, and this resource base can be at risk when labour’s share in national income declines or if individual financing burdens are lower for new or growing forms of non-standard work. Unequal financing burdens or SP entitlements, in turn, can shape labour market changes, e.g. by promoting certain types of dependent or independent employment while discouraging others.

In many countries, SP coverage is ill adapted to non-standard forms of employment, with unstable and overlapping jobs (see Chapter 4 for categories and definitions). For instance, own-account self-employed workers often have little or no access to key SP provisions, such as unemployment benefits and related job-search assistance. Those with on-call (including “zero-hours”) contracts may have access in principle but they might be effectively barred from out-of-work support due to legal ambiguities of what constitutes an “out-of-work” situation for somebody with no minimum working hours. Those with frequent employment gaps or transitions between activities may not meet relevant employment requirements or, if they do, the timing of support may not correspond to individual needs (Section 7.2).

Reduced de-facto coverage, in turn, compromises the sustainability of SP. Risk sharing through collective protection systems has unique strengths not only in terms of equity objectives but also on efficiency grounds (Chetty and Finkelstein, 2013[9]; Gruber, 1997[10]; Barr, 1989[11]). Key labour market risks, such as unemployment and low earnings, are uninsurable without government intervention, such as mandated membership, price setting or related regulation (Boeri and van Ours, 2013[12]). A shrinking group of SP members or contributors, e.g. as a result of readily available alternative work arrangements that allow bypassing SP provisions, can undermine the foundations of risk pooling. If mandates are partial or weakly enforced, those with comparatively lower risks (the “good risks” in insurance terminology) can minimise their contributions or choose to opt out entirely, implying greater financing burdens or weaker protection for the remaining higher-risk groups (Section 7.2.1). In the absence of regulation or public subsidies this produces further opt-out incentives and, ultimately, a cycle of escalating costs and declining reach of SP (Rothschild and Stiglitz, 1976[13]; Akerlof, 1970[14]).

Countries may operate different mandates in different parts of the SP system, e.g. if the unemployment scheme is optional for some workers while membership in the pension scheme is mandatory. Such differentiation can go against more encompassing concepts of social risk sharing across both individuals and risk types and, in doing so, it may exacerbate selective opt-in/opt-out behaviour. For instance, low-income/unemployment and longevity risks are typically negatively correlated (OECD, 2017[15]). When given the opportunity, higher-income or higher-skilled individuals may opt out of unemployment insurance but may wish to retain membership in a pension system that favours people with long life expectancy. Packaging unemployment, pensions and other SP elements through a unified set of mandates and regulations reduces the scope for “cherry picking”, and can help to maintain a more diversified risk pool, making SP attractive for broader sections of the population.

The extent of explicit or implicit redistribution between groups is another key determinant of the protection against earnings risks. Income assistance programmes that are financed through general tax revenues redistribute towards individuals in need without linking entitlements to own contributions. On the other end of the spectrum, individual savings, whether voluntary or mandated, also support consumption smoothing in the event of reduced earnings ability. But in the absence of any redistribution at all, vehicles such as individual savings accounts do not enable risk pooling across groups. They also provide little support to groups who are unwilling or unable to save enough (e.g. because they have care responsibilities or are born into poverty), and they cannot insure against catastrophic events that endanger livelihoods and overwhelm people’s own capacities to absorb or reverse a situation of economic need (e.g. a long-term or permanent loss of earnings ability). “Actuarially fair” insurance provisions also do not feature any explicit redistribution, as the present value of expected lifetime contributions equals the present value of expected lifetime benefits. But in practice, actuarially fair risk pooling nevertheless results in redistribution, in the sense that an insured person will typically get out either more or less than they paid in.2

The specific policy challenges that arise in evolving labour markets are shaped by the design principles governing existing SP systems (Sections 7.3.1 and 7.3.2). However, regardless of countries’ specific SP designs, the provision of adequate SP financing can be expected to move up the policy agenda, especially if existing coverage gaps are to be addressed (Section 7.3.5). Budgetary pressures and evolving perceptions of who benefits from SP, and who pays for it, may alter the political dynamics in the SP debate. In a rapidly evolving labour market, a major continuing challenge will be to ensure broad buy-in and a consensus among most people that they continue to be better off with adequate SP in place (Hills, 2017[16]).

copy the linklink copied!7.2. Social protection for alternative forms of employment: What are the gaps?

7.2.1. Statutory access

Statutory access varies by employment type and by SP branch. Temporary and part-time workers are in principle covered in the same way as permanent full-time employees in most countries and for most risks, as long as they satisfy minimum employment periods (Figure 7.1), earnings thresholds and other eligibility requirements, such as low family income. Some countries operate exemptions for specific non-standard contractual arrangements, such as casual employment, seasonal work or hybrid categories and some of these are noted below (see Chapter 4 for an overview of different non-standard forms of work). By contrast, statutory access to SP for self-employed workers is very frequently restricted in contributory SP systems (Figure 7.2). Indeed, SP provisions that were mostly set up with a steady employer-employee relationship in mind do not easily accommodate the self-employed:

  1. 1. Double contribution issue: Who should be liable for employer contributions in the absence of an employer? In practice, total formal contribution burdens are frequently lower for the self-employed than for dependent employees (Section 7.3.5). Requiring the self-employed to pay the equivalent of both employer and employee contributions brings formal burdens in line with dependent employees. But effective burdens may be higher for the self-employed, especially those with lower earnings, because minimum wages typically do not apply to them or because they may lack the bargaining power to shift any contribution-related costs onto their clients by charging higher prices.3

  2. 2. Fluctuating earnings and avoidance: The self-employed, along with some atypical employees such as on-call workers or those with zero-hours contracts, are often paid at irregular intervals, either because of time lags between work and payment, or because demand for their services is erratic (ISSA, 2012[17]). This complicates the calculation of contributions (as well as the assessment of entitlements). In particular, self-employed workers may be able to avoid or lower contributions by optimising their contribution base, e.g. through timing their work or earnings, see Section 7.3.3.

  3. 3. Moral hazard: Demand or price fluctuations affecting self-employed workers are difficult to distinguish from voluntary idleness and this complicates the provision of unemployment insurance in particular. For instance, there is no employer to confirm a layoff and efforts to re-establish a business operation are more difficult to monitor than the search for dependent employment.4 Where self-employed individuals can claim unemployment benefits, they typically need to meet relatively stringent requirements to demonstrate that their business is no longer operational.5

When self-employed workers do have access to social protection, it is frequently on a voluntary basis. This partly reflects specific risk patterns and fairness considerations, e.g. as entrepreneurs seek to make a profit in return for taking on business risks, may be less risk averse, and therefore may not require insurance to the same extent as employees. However, the same rationale for opt-outs could be invoked more broadly, e.g. for employees who face lower risks or are less risk averse than others. Ultimately, strong reliance on selective or voluntary SP membership widens the scope for gaming social risk-pooling systems, resulting in insurance becoming inefficiently narrow and unaffordable for those who need it. In particular, low-earning individuals may underinsure even when social insurance provisions offer attractive cost-to-risk ratios.6 Country experiences with voluntary schemes illustrate that selectivity typically leads to low coverage or a need for significant subsidies to keep risk sharing financially viable (Box 7.1).

copy the linklink copied!
Figure 7.1. Employment requirements for unemployment benefits range from 3 to 24 months
Number of months, 2018
Figure 7.1. Employment requirements for unemployment benefits range from 3 to 24 months

Note: (1) And at least 30 days of employment in the 12 months prior to the start of the unemployment spell; (2) Or 200 days in last two years; (3) Earnings condition also; (4) Assuming 40 hour work week; (5) 6 months in any one of the past two years plus minimum-earnings requirement. (6) Must also have been a member of the insurance fund for at least 12 months; (7) Or 26 weekly contributions in each of the previous two years. Must also have made 104 weekly contributions in whole career; (8) 28 weeks for repeated employment; (9) Must have 12 months of contributions since a previous unemployment spell; (10) Must also have been continuously in work for last 120 days.

Source: OECD tax-benefit model and policy database,


copy the linklink copied!
Box 7.1. Voluntary insurance: Country experiences

Membership in non-mandatory insurance schemes is often organised on an opt-in basis. Not all voluntary schemes are of opt-in type, however. For instance, voluntary pension insurance for marginal employees (mini jobs) in Germany are auto-enrolled since 2013. In all cases, voluntary membership risks adverse selection of members: where insurance premiums are uniform, those with the highest risk have the biggest incentive to join. If the scheme is entirely self-funded, this can lead to a vicious circle of contribution hikes and low-risk members leaving; if the scheme is publicly subsidised, costs can increase (see Section 7.1).

An example of this mechanism is the Canadian Special Benefits for Self-employed Workers (SBSE) scheme, which provides access to maternity and parental benefits, sickness benefits and care benefits for ill family members since 2010. Self-employed workers pay the same contribution as standard employees but they do not pay the part that employers would cover for dependent employees. In the first year of benefit pay-outs, over three quarters of claims were for maternity and parental benefits, two-thirds of opt-ins were women (who represent only 43% of all self-employed workers), and two-thirds were between the ages of 25 and 44 (compared to just one third of all self-employed). Opt-ins also had significantly lower incomes than other self-employed workers. As a result, premiums covered less than one-third of benefit payments (Employment and Social Development Canada, 2016[18]).

In 2007-8, reforms of the voluntary unemployment insurance in Sweden linked employee contributions to unemployment risk and raised average premiums by 300%. Membership in the Unemployment Insurance Funds dropped by around 10 percentage points in the following years. Groups that were particularly likely to exit the system included older workers over the age of 60, who have the lowest unemployment risk of all age-groups, and young workers under the age of 25, who typically have low earnings and short unemployment durations (Kolsrud, 2018[19]).

Self-employed workers in Austria can opt into a short-term sickness benefit programme, and about 8% of eligible self-employed workers do. In 2016, close to half of those who were covered received a benefit. The average benefit duration was 22 days, nearly twice the average sick-leave duration among dependent employees, who are subject to mandatory insurance, highlighting moral hazard risks. In response to the resulting deficits in the programme, the minimum benefit was cut significantly in 2017 (Fink and Nagl, 2018[20]).

Some schemes also offer choice in the level of contributions. For instance, self-employed workers in Latvia and Spain could choose the contribution base (and hence the level of contribution) in the unemployment and occupational injury insurance. Similar to selectivity mechanisms in insurance provisions with voluntary membership, higher-risk individuals can have an incentive to choose higher contributions in order to maximise their entitlements. Yet, if the system is explicitly redistributive, that is if it offers higher replacement rates for low incomes (contributions), there is a clear incentive to make the lowest possible contribution. In both Latvia and Spain, approximately nine out of ten self-employed workers chose to pay the minimum contribution (Arriba and Moreno-Fuentes, 2017[21]; Rajevska, 2017[22]).

Statutory access for non-standard workers varies by social protection branch

Unemployment benefits are the least accessible branch of SP for non-standard workers. Eleven of the 28 countries shown in Figure 7.2 (bottom left panel) do no not offer any kind of unemployment protection for self-employed workers. Access is also restricted for some forms of dependent non-standard work, e.g. casual workers in the United States, or para-subordinate workers in Italy (SSA and ISSA, 2017[23]; Raitano, 2018[24]). Extensions of unemployment-benefit coverage for self-employed workers have been legislated or discussed recently in a number of countries, including France and Ireland. Spain made a previously voluntary unemployment insurance scheme for self-employed workers compulsory from early 2019 (see also endnotes 9and 12).

The rules for accessing incapacity benefits – covering short-term sickness, work accidents and disability – vary across countries and types of non-standard work. In all three of these schemes, statutory access for non-standard dependent employees is mostly similar to standard employees. Exceptions include Australia, where casual workers are not entitled to cash sickness benefits (an employer-provided benefit), the United States, where casual workers do not have access to accidents-at-work insurance, and Italy, where some para-subordinate workers are not covered by short-term sickness insurance. However, only 14 of the 32 countries shown in Figure 7.2 offer similar access to benefits for self-employed workers (top right panel). Statutory access is weakest in the case of accidents at work. Many – although not all – genuine self-employed workers indeed have considerable control over their working environment and, as in the case of unemployment benefits, insurance against work accidents can therefore be prone to moral hazard (see also Chapter 4). But the exclusion of the self-employed does create important SP gaps for those with genuinely risky activities, including for workers who are wrongly classified as independent, or who are in the “grey zone” between self-employment and dependent employment, e.g. those with a de-facto employer.

When contingencies are independent of a specific job, protection for non-standard workers is more easily available. For instance, social assistance or minimum income schemes are typically financed through general tax revenue, and legal entitlement rules are based on need, regardless of past employment type, duration or stability. Family benefits, such as child allowances, are typically universal or means-tested, and statutory access to maternity benefits, which are often contributory, also tends to be similar for workers in standard and non-standard forms of dependent employment. An exception is Italy, where “workers on vouchers” and foreign seasonal workers do not have access to contributory family benefits (Jessoula, Pavolini and Strati, 2017[25]).7 Maternity benefits often have separate provisions for independent workers (Figure 7.2, top left panel). Yet, in all countries with compulsory maternity coverage for standard employees, self-employed workers can either opt into the main scheme voluntarily, or they have access to a separate benefit that is typically less generous than for dependent employees (lower benefit amounts and/or shorter duration).

Pension rules often differ between the self-employed and dependent employees (Figure 7.2, bottom right panel). In some countries, the self-employed can voluntarily join earnings-related schemes that are mandatory for employees (e.g. Germany and Australia).8 Chile has sought to incorporate the self-employed into the mandatory pension scheme through auto-enrolment since 2008. Opt-out provisions were retained, however, and the majority of the self-employed have continued to opt out so far. A few countries provide the self-employed with partial pension coverage, reducing both contributions and benefits in the mandatory scheme, or they subsidise pensions of the self-employed through a more favourable benefit formula. For instance, in Denmark, Japan, the Netherlands, and Switzerland contributions for the self-employed are mandatory for the basic pension pillar only. Box 7.2 summarises additional aspects of pension provisions for non-standard workers.

In countries where non-standard forms of work are frequent, statutory restrictions can inhibit effective insurance against major risks for a significant share of the workforce. For example, Greece and Italy, both countries with large shares of self-employed workers, show sizeable gaps in statutory access to sickness, invalidity and accident insurance as well as unemployment insurance for self-employed workers. Yet, statutory restrictions are also common among some countries where self-employment is less common, e.g. in the case of unemployment insurance in Canada, Japan, Norway, and United States. Figure 7.2 therefore does not display any obvious bi-variate correlation between SP provisions for self-employed workers and the incidence of self-employment. The lack of a bivariate link between statutory SP entitlements and the incidence of non-standard work is not surprising as the choice of employment form depends not only on expected benefits but also on many other factors, notably any taxes or mandatory contributions to the financing of SP (see Section 7.3.5).

7.2.2. How much protection is available in practice?

Comparisons of legal eligibility rules give an incomplete, and possibly misleading, picture of the support that is available in practice for different labour market groups. The limitations of “systems focussed” comparisons based on statutory access rules are twofold. First, beyond availability, the content and generosity of support will generally differ across types of workers and across countries. Second, statutory access is not the same as actual access:

  • Non-standard workers may have characteristics that make it difficult for them to meet entitlement criteria, even when they are the same as for standard workers.

  • The implementation of SP rules may differ between groups in practice and the implicit cost of claiming benefits may dissuade eligible people from applying.

  • Contributions to some SP elements may be voluntary for some categories of workers, who may opt out of or seek to bypass applicable rules if they perceive future benefits as small relative to the immediate individual cost.

copy the linklink copied!
Figure 7.2. Statutory access for independent workers is often limited
Statutory access to social protection for the self-employed compared to dependent employees (“employees”) by social protection branch and incidence of self-employment, 2017
Figure 7.2. Statutory access for independent workers is often limited

Notes: Gaps between standard dependent employees (full-time open-ended contract) and self-employed workers. “Partial access” to benefits can arise if a) eligibility conditions, benefit amounts or receipt durations are less advantageous for self-employed workers; b) insurance-based and non-contributory benefits co-exist and individuals can access only the latter (e.g. only basic pension and not earnings-related); or c) the self-employed can choose to declare a lower contribution base while dependent employees pay contributions on full earnings (possibly subject to a ceiling). “No access”: compulsory for dependent employees but self-employed are excluded. * Data on self-employment incidence is missing/incomplete for Estonia, Iceland and Luxembourg and refers to 2015 for the Slovak Republic and to 2014 for Latvia.

Maternity and family benefits: “no access”: access to neither maternity benefits nor family allowances, “voluntary/partial access”: voluntary or partial access to both schemes, or full access to one, but no access to the other. “Not applicable”: schemes do not exist.

Sickness, invalidity and work injury benefits: “no access”: access to none of these benefits, “larger gaps”: less than full access to all of the benefits and no access to at least one; “some gaps”: no access to one benefit and at least partial access to the others, or voluntary or partial access to all three benefits; “similar to employee”: at least partial or voluntary access to all benefits, and at least full access to one.

Old-age pensions: “voluntary at low earnings”: coverage is generally similar to employees but opt-out is possible if earnings are below a certain threshold. “Similar to employees”: mandatory coverage as for employees but contributions may still differ.

Source: Pensions: OECD (2017[26]), MISSOC (2018[27]), Social Security Administration (2018[28]). Working-age benefits: Australia: Whiteford and Haron (2018[29]), European countries: adapted from Spasova et al. (2017[30]), Canada, Japan, Korea: Information provided by country delegations to the OECD, USA, SSA and ISSA (2012[17]). Share of self-employment: OECD (2018), “Labour Force Statistics: Summary tables” and OECD Employment and Labour Market Statistics (database).


copy the linklink copied!
Box 7.2. Old-age pensions for workers with careers breaks or non-standard employment

Alleviating risks of old-age poverty can be particularly challenging for workers on non-standard contracts with low contribution levels and patchy contribution records. Pension systems in OECD countries have a number of features and characteristics that are relevant for non-standard employees and for the self-employed.

Contribution breaks and first-tier pensions

Where pension benefits are linked to contributions (and thus earnings), eligibility rules may require minimum contribution periods. Taking account of mandatory and quasi-mandatory schemes, a 10-year out-of-work spell combined with a late career start reduces pension entitlements by 20% on average across OECD countries (Figure 7.3, Panel A). In some countries with basic public pensions (Ireland, New Zealand, United Kingdom), mandatory pensions are not affected at all, while the penalty exceeds 30% in Mexico, Turkey and Chile.

Policy makers have a number of policy tools at hand to loosen the link between contribution histories and pension entitlements. Contribution periods can be credited for some out-of-work spells, such as unemployment or childcare-related leave, and first-tier pensions can be designed to be independent of work histories.

First-tier pensions are either contributory or non-contributory. Minimum pensions and contribution-based basic pensions are only granted to retirees with contributory records and their level may depend on the total contribution period. By contrast, residence-based basic pensions and guaranteed minimum income (social assistance) provisions for retirees are unrelated to contributions. Pension systems in most OECD countries include either a residence-based basic pension, a minimum pension, or a social assistance benefit alongside earnings-related pension provisions. Across the OECD, the level of income provided by non-contributory pensions is about 35% of median disposable income on average (Figure 7.3, Panel C). Access requirements vary considerably across countries with contributory first tier-pensions, however (Figure 7.3, Panel B). When minimum contribution requirements are relatively stringent, meeting them can be challenging for non-standard workers.

Old-age pensions for the self-employed

Figure 7.2 (right bottom panel) shows the extent to which pension scheme membership provisions differ between standard employees and the self-employed (see main text in Section 7.2.1). In addition, total old-age pension contribution rates are frequently lower for the self-employed than for employees, e.g. in Austria (18.5% versus 22.8%), Ireland (self-employed pay the employee part of 4% only), Norway (11.4% versus 22.3%) and, as from 2019, in Portugal (a reduction of roughly one fourth for self-employed). Some countries provide for lower contributions or exemptions for certain categories of self-employed workers, e.g. for some sectors or types of activity (Germany, Italy), during an initial period of business operation (Austria, Finland, France, Norway or Poland), for those combining independent and dependent employment (Belgium, Greece, Slovenia) or below an earnings threshold (Italy, Ireland, Luxembourg, the Slovak Republic and the United Kingdom). As a result, some self-employed might expect lower pension replacement rates than dependent workers with similar earnings, and they face higher risk of relying only on non-contributory benefits.

copy the linklink copied!
Figure 7.3. Pension provisions for workers with unstable careers vary widely across countries
Figure 7.3. Pension provisions for workers with unstable careers vary widely across countries

Note: Panel A: Parentheses show the additional years required for a full pension in countries where later retirement allows people with incomplete careers to access a full pension (i.e. without actuarial penalty). Panel B: Values for “Full” pension are not shown if they are the same as for the “Lowest” pension. In the Slovak Republic, there is no upper limit for the minimum pensions as it increases with each additional contribution year. Panel C: Non-contributory pensions include social assistance and residence-based basic pensions.

Sources and definitions: OECD (2017[15]; 2018[31]; 2017[26]).


An indirect way of lowering contribution burdens is to allow self-employed some flexibility in the contributions base that they declare (Finland, Latvia, Lithuania, Poland and Spain). When the contribution base can be chosen relatively freely, contribution floors are needed to counter avoidance strategies and the potential pension gaps that these create. But, depending on their design, contribution floors either lead to very high effective contribution rates for those with genuinely low earnings, or they exclude those with earnings below the minimum base from earnings-related pensions altogether. Most countries use taxable income as the contribution base for self-employed but some provide additional adjustments. In some countries, clients of the self-employed are required to pay part of the pension contributions that are due for some categories of self-employed workers (Austria, Germany, Italy and Portugal). This solution can be administratively demanding but can be an attractive method for broadening coverage while sharing SP financing costs.

  • Access to the overall support package is difficult to assess from the rules that govern separate individual SP elements. Depending on countries’ policy approaches, support for out-of-work or low-income groups is frequently spread across two or several SP branches. For instance, in-work support or guaranteed minimum income programmes can fill some of the gaps that first-tier out-of-work support leave for workers in independent, unstable, or part-time employment.

  • Finally, a focus on relative access gaps between standard and non-standard workers hides country differences in terms of the overall reach of support, and the coverage gaps that may exist for standard workers as well. For instance, on average across countries, two out of three jobseekers did not receive unemployment benefits in 2016, but coverage differed markedly between countries, ranging from under 10% in Italy, Slovak Republic, Poland, Greece and United States to more than 50% in Belgium and Finland (Figure 7.4).

From an inclusiveness perspective, a people-centred policy discussion requires information on the actual support that people receive in different labour market circumstances. This section presents new results on SP gaps that are observed in practice. The approach consists of estimating a statistical model of benefit entitlements controlling for the most important determinants of social benefits. As benefit access and amounts often depend on past events, the analysis relies on longitudinal household data that include information on current and past employment and earnings. The main variable of interest is the value of the total benefit package, rather than any individual category of social transfer, reflecting the fact that countries provide support through different channels and programmes. The sample comprises all working-age individuals aged 18-64 who are not retired and not in full-time education. The policy scope comprises the most important social transfers to working-age individuals and their families: unemployment and disability benefits as well as housing, family, in-work and guaranteed minimum income (GMI) transfers. Box 7.3 summarises the main steps of the empirical approach.

The focus on working-age cash benefits is a practical consequence of limitations in the panel data, which contain no systematic information on support that is provided in kind rather than in cash. Moreover, the panels are restricted to a time window of four years or less making the approach unfeasible for retirement benefits (which often depend on contribution histories over much longer periods). However, in the future-of-work debate, reasons other than data quality may motivate a specific interest in working-age support that is provided in cash. Indeed, for those experiencing technology-related unemployment or employment changes, income-support entitlements during working age are likely to be the most immediate concern. In addition, in-kind support, such as housing or active labour market programmes, is often tied to working-age benefits. While the rest of this section therefore focuses on working-age benefits, Box 7.2 above discusses implications of non-standard work patterns for retirement incomes.

copy the linklink copied!
Figure 7.4. Only a minority of jobseekers receive unemployment benefits
Coverage among unemployed and discouraged workers, 2016
Figure 7.4. Only a minority of jobseekers receive unemployment benefits

Note: In some countries, unemployment benefits can be received by people who have some work attachment (part-time unemployed) or who are not available for work (labour market inactive). The number of unemployment-benefit recipients can then be much higher than suggested by these coverage rates. See OECD (2018[32]) for detailed results by employment status. The OECD SOCR database provides benefit recipients totals by country ( “Discouraged” refers to people without employment who are available for work but are not actively looking for a job. “ILO unemployed” (based on the definition of the International Labour Organization) refers to out-of-work individuals who want a job, have actively sought work in the previous four weeks, and can start working within the next fortnight. Those who have made arrangements to take up paid employment or self-employment in the near future are also included in the definition of ILO unemployment. OECD-24 corresponds to the unweighted average of the countries shown. 2015 figures for Australia. LFS data for Sweden do not include benefits that are accessible to jobless individuals who: i) are not in receipt of core unemployment benefits; and ii) satisfy other conditions such as active participation in employment-support measures. ILO: International Labour Organization.

Source: OECD (2018[32]) using Household, Income and Labour Dynamics in Australia (HILDA) for Australia; European Union Labour Force Survey (EU-LFS) for European countries; and Current Population Survey (CPS) for the United States. Some European countries are excluded due to missing information in EU-LFS data.


Results are intended as shorthand summaries of benefit accessibility and generosity in comparative perspective and for both standard and non-standard workers (see Figure 7.5). The implied access gaps between standard and non-standard forms of work reflect policy provisions that were in force during the income reference period in the data (around 2014) and therefore do not account for policy reforms that were enacted since then, including in some of the countries where the estimated gaps are large.9 With that qualification in mind, estimated access gaps were largest in Estonia, where non-standard workers were only half as likely as standard workers to receive benefit support following job loss, and more than 60% of out-of-work individuals with past non-standard work are estimated to go without any benefit support during a 12-month out-of-work spell (Panel A). Gaps were also large in Czech Republic, Latvia, Portugal and Slovak Republic. A notable result is that several countries with comparatively good access to out-of-work support for standard employees also provide accessible support to those with past non-standard work (Austria, Belgium, France, Hungary and Luxembourg). Benefit accessibility for those with past continuous full-time work was also high in Iceland, Slovenia, Spain and the United Kingdom, but moderate gaps for non-standard workers existed in these countries (albeit statistically insignificant in the United Kingdom). In Greece and Italy, even standard workers only had a 50% chance of receiving benefits following a job loss.

copy the linklink copied!
Figure 7.5. Non-standard workers receive little support in some countries
Overall support package for working-age individuals, 2014-15
Figure 7.5. Non-standard workers receive little support in some countries

Note: Predicted benefit receipt during an entire year comparing: i) an able-bodied working-age adult who is out of work, had uninterrupted full-time dependent employment with median earnings in the preceding two years, and lives in a two-adult low-income household without children (“baseline: past-standard work”, triangle-shaped markers); and ii) an otherwise similar individual whose past work history is “non-standard”: mostly in part-time work, mostly self-employed, or interrupted work patterns during the two years preceding the reference year (“past non-standard”, light and dark diamond-shaped markers). Additional results for different categories of non-standard work are available for some countries (see main text).

Statistical significance refers to the gaps between baseline and comparator cases (90% confidence interval). Full-time students and retirees are excluded from the sample. Details on data and model specification are summarized in Box 7.3 and presented in further detail in Fernández, Immervoll and Pacifico (forthcoming[33]). The data source, the European Union Statistics on Income and Living Conditions (EU-SILC) covers additional countries but they are excluded here because effective sample sizes were small (e.g. Ireland, Lithuania), because the required micro-data were entirely unavailable (Germany), because key employment-status variables are recorded only for one individual per household (Denmark, Finland, the Netherlands and Sweden), or because of partial or partly conflicting information on income or benefit receipt (Norway).

Source: OECD calculations using EU-SILC panel data.


The predicted average size of the overall benefit package for recipients (Panel B) also varies enormously across countries, ranging from around 20% of national median incomes or less in parts of Central and Eastern Europe (Czech Republic, Slovak Republic, Poland, Hungary) to above 40% in France, Italy, Portugal, Luxembourg and Belgium. Generosity gaps for non-standard workers were largest in Southern Europe (Greece, Italy, Portugal, Spain), as well as in Estonia and Slovenia, exceeding 10% of median household income. As benefit amounts are measured over a year as a whole, these country patterns reflect differences in benefit levels in a given month, as well as in the duration of benefit payments.10 They also account for the full range of benefits that may be available to the different out-of-work individuals shown, i.e. not only unemployment benefits but also any cash housing or social assistance transfers.

Across countries, there is no obvious link between accessibility and generosity. For example, both the accessibility and generosity scores were high in Belgium and comparatively low in Poland. Benefit access in Italy was comparatively difficult for both standard and non-standard workers, but benefit levels for recipients were fairly high. Hungary shows the opposite pattern, with high implied coverage but low benefit levels. There is also no clear link between overall benefit generosity and the size of the gap between standard and non-standard workers. Generosity gaps were statistically insignificant in some countries with comparatively modest benefit amounts (e.g. close to 15% of median income in the Czech Republic and below 25% of median income in Hungary), but also in others where benefits exceeded 50% of the median (Belgium and Luxembourg).

In five countries both coverage and generosity gaps between standard and non-standard workers were statistically insignificant: Belgium, France, Hungary, Luxembourg and the United Kingdom. These results suggest that effective support for non-standard workers is achievable in the context of quite different social-protection systems and targeting strategies. For instance, out-of-work support in the United Kingdom is flat-rate, comprising an initial insurance benefit for dependent employees with the required contribution record, and means-tested support for active jobseekers with low family incomes and assets. Hungary and Belgium offer earnings-related unemployment protections to both standard and non-standard workers. In Hungary, non-standard workers, including the self-employed, can be entitled to unemployment benefits (Albert, Gáspár and Gal, 2017[34]). In Belgium, non-standard workers are also entitled to unemployment insurance benefits though benefit amounts are much more generous than in Hungary, and benefits for self-employed workers in Belgium account for household needs (De Wispelaere and Pacolet, 2017[35]). In both countries, means-tested support provides further layers of protection for those not entitled to insurance benefits. Recipient numbers of lower-tier social assistance are also substantial in Luxembourg but non-standard workers had (statutory) access to first-tier unemployment benefits as well (Pacolet and Op De Beeck, 2017[36]). In France, a key explanation for the insignificant coverage gaps is the very short qualification period for unemployment benefits (Figure 7.1), paired with a possibility to retain unused benefit entitlements for future out-of-work periods and to cumulate benefit rights across successive out-of-work spells for the (large and growing number) of workers with short-duration employment contracts. France also provides multi-layered income support that benefits workers in other types of non-standard employment (as well as others who may not qualify for first-tier insurance benefits). For instance, self-employed workers in France did not have access to unemployment benefits during the 2014-15 period shown in the results, but income-targeted social assistance and housing benefits do provide additional layers of income security for jobseekers with low household incomes.11,12

The empirical results highlight instances where SP gaps between standard and non-standard workers were large. But it is important to note that policy challenges can also arise in countries where the reported gaps are small or insignificant. First, the results refer to someone who remains out of work for an entire year, whose income is in the bottom 10%, and who has worked prior to the reference period. In other words, this is an individual who would qualify as deserving of income support by most standards and the provision of adequate support in such circumstances arguably represents a modest benchmark of effectiveness for SP systems. Support gaps may be significant and widespread for non-standard workers in other situations, for example: for those with some intermittent work during the year; for individuals with a working partner whose earnings lift household income beyond the poorest 10%; for those losing employment as a result of poor health; or for recent parents who take time off work to care for their children.

Second, gaps are calculated on the basis of micro-data that are necessarily backward looking. They therefore reflect recent labour-market realities and employment categories as recorded in these data. The present results therefore do not capture gaps that may exist for newly evolving alternative working arrangements, such as platform work. Updating the estimates at regular intervals, and as data with more granular information on employment categories become available, would allow monitoring SP gaps as the future of work takes shape. Such regular monitoring is desirable and should examine support gaps for a wider set of circumstances than is reported here.

Third, the gaps in Figure 7.5 are calculated taking an average across several types of past non-standard work, including part-time dependent employment, temporary work and self-employment. As shown in Figure 7.2 above, statutory access to many benefits is often especially limited for the self-employed, which suggests that obtaining support could be very difficult for some subcategories of independent workers, even in countries where gaps are insignificant between standard and non-standard workers more broadly. Caution is therefore required in drawing conclusions on specific policy challenges across countries solely based on Figure 7.5. Breakdowns of SP gaps for different types of non-standard work would further enrich policy reform discussions and preliminary results from such a more granular analysis are presented in Figure 7.6. These results disentangle part-time employment, unstable employment and self-employment and are available for six countries where data samples are sufficiently large for a disaggregated analysis: France, Greece, Hungary, Italy, Spain and the United Kingdom.

The estimates indicate that support gaps can indeed vary markedly across different forms of non-standard work but that this need not be the case.13 In Spain, part-time workers were somewhat less likely than standard employees to receive benefits during an out-of-work spell, but the difference was comparatively small (10 percentage points, Panel A). In the other five countries, benefit accessibility gaps between part-time workers and standard employees were statistically insignificant.

In Italy and Spain, those with interrupted work histories were less likely to receive out-of-work support than standard employees. But in four of the countries shown, benefit accessibility gaps for workers in unstable employment were statistically insignificant. In some of them, workers can qualify for unemployment insurance benefits after comparatively short periods in work (e.g. three months in France and six months in the United Kingdom, see Figure 7.1). It should be noted, however, that the gaps were calculated assuming the same earnings levels for the different worker categories (see figure notes). They therefore do not reflect accessibility issues that may exist for low-paid workers in particular. For instance, some countries require minimum earnings levels for employment to count towards unemployment insurance entitlements (e.g. approximately 16% of average full-time earnings in the United Kingdom). In these cases, accessing support can be more difficult for individuals with very short working hours or with extended out-of-work periods in-between employment spells.

Accessibility gaps for those with past self-employment were sizeable in four of the six countries, with implied coverage rates for out-of-work individuals as low as 10% in Italy and around 25% in Greece, Spain and the United Kingdom. Where unemployment insurance is not open to the self-employed (see Figure 7.2), they will need to rely on lower-tier income support, such as unemployment assistance or guaranteed minimum income benefits. These transfers typically feature strict access requirements, including income and asset tests, and are subject to significant non-takeup, which lowers their effective reach. Assistance benefits also tend to be less generous than insurance transfers. For those self-employed who do receive income support, predicted benefit amounts are therefore frequently lower than for standard employees. Exceptions are the United Kingdom (where flat-rate insurance and assistance benefits pay similar amounts), as well as France (where the generosity gap for the self-employed was bigger than in the United Kingdom, but statistically insignificant).

Benefit accessibility in France and Hungary did not vary substantially across different types of non-standard work: relative to standard employees, accessibility gaps were minor for all three categories (part-time work, self-employment and unstable work). Benefit levels were much lower in Hungary, however, while estimates for France suggest that income support may have been somewhat more generous for out-of-work individuals with a recent history of unstable or intermittent employment.

copy the linklink copied!
Figure 7.6. Support gaps can be sizeable for the self-employed
Overall support package for working-age individuals, 2014-15, selected countries
Figure 7.6. Support gaps can be sizeable for the self-employed

Note: Predicted benefit receipt during an entire year comparing: i) an able-bodied working-age adult who is out of work, had uninterrupted full-time dependent employment with median earnings in the preceding two years, and lives in a two-adult low-income household without children (“baseline: past-standard work”, triangle-shaped markers); and ii) an otherwise similar individual whose past work history is “non-standard”: mostly in part-time work, mostly self-employed, or interrupted work patterns during the two years preceding the reference year (“past non-standard”, light and dark diamond-shaped markers).

Statistical significance refers to the gaps between baseline and comparator cases (90% confidence interval). Full-time students and retirees are excluded from the sample. Details on data and model specification are summarized in Box 7.3 and presented in further detail in Fernández, Immervoll and Pacifico (forthcoming[33]). The data source the European Union Statistics on Income and Living Conditions (EU-SILC) covers additional countries but they are excluded here because effective sample sizes were small (e.g. Ireland, Lithuania), because the required micro-data were entirely unavailable (Germany), because key employment-status variables are recorded only for one individual per household (Denmark, Finland, the Netherlands and Sweden), or because of partial or partly conflicting information on income or benefit receipt (Norway).

Source: OECD calculations using EU-SILC panel data.


copy the linklink copied!7.3. Addressing social protection gaps: Key policy issues

The results presented in the previous section suggest that different SP strategies can be effective at limiting support gaps for workers in non-standard forms of employment. But where gaps do exist, they may become more widespread and reduce the effectiveness of existing SP provisions if ongoing labour market changes lead to a transformation of the standard employer-employee relationship, or to greater labour market churn and heightened employment instability.

copy the linklink copied!
Box 7.3. Benefit access and generosity: A statistical model accounting for key policy levers

The empirical work proceeds in two steps. A first step estimates the relationship between individual benefit receipt and a large number of key structural drivers of support, including previous and current work status and earnings, current household income, family composition, housing tenure and any health limitations.14 This type of information is typically available from household panel data. The current analysis uses three-year panels of the European Union Statistics of Income and Living Conditions (EI-SILC), and pools observations from two survey waves (2014 and 2015) to increase sample size.15 The effective sample sizes range between 3 500 observations in Iceland and 20 600 in Italy. The dependent variable is total social cash benefits received during an entire year. It therefore accounts for both the average generosity of monthly benefit payments and the durations of benefit entitlements (as well as any waiting periods or other possible gaps between benefit entitlement and pay-out). For benefits that are observed at the household rather than the individual level (family benefits, GMI benefits), amounts are allocated to all adult household members on a per-capita basis. Model specifications are consistent across countries and include the following independent variables, along with relevant interactions and higher-order terms: main employment status during the reference year; pre-transfer household income during the reference year; main employment status during the preceding two years; earnings during each of the preceding two years; family situation and number of children; health status; housing tenure and housing costs; education level; gender; and age.

Separate models are estimated for benefit receipt (yes/no indicator variable) and generosity (benefit amounts) using a generalized Hurdle approach, as the process that determines whether a person receive social benefits is not necessarily the same as the process that determines the amount received – see Wooldridge (2010[37]) and Cragg (1971[38]). The first model is a logistic regression for benefit receipt at the individual level. The second model is an exponential regression of benefit amounts (entitlements) estimated only on observations with positive benefits. The use of exponential regressions, rather than a standard log-linear model, sidesteps inference problems that arise with predicting levels for log-transformed dependent variables (Wooldridge, 2010[37]). A second step uses the estimated relationships for inference on the benefit gaps between standard and non-standard workers in concrete circumstances (“vignettes”) that are defined in a consistent way across countries. For ease of comparison, benefit amounts are shown as a share of median income in each country. The use of a vignette-based analysis facilitates the communication of complex statistical results in a comparative perspective, and the identification of relevant policy mechanisms underlying the observed gaps. A direct interpretation of the estimated coefficients is complicated by interaction effects, categorical variables and other nonlinear functional forms. Significant interpretation difficulties arise also in nonlinear models such as logistic regression as the raw coefficients are often not of immediate interest. In these cases, “marginal effects” (i.e. statistics computed from model predictions for different values of the control variables) allow summarising the entire vector of estimated parameters into a single value using the same metric as the dependent variable (here the probability of receipt and the benefit amount). Standard errors (computed by means of the Delta method), allow inference on the estimated gaps and their statistical significance.

Benefit “gaps” are calculated relative to a baseline “standard” worker who requires out-of-work support: an out-of-work 40-year old who was previously working full-time without interruptions and median earnings, and who lives in a low-income childless couple and in privately rented accommodation. Relevant characteristics for the comparator vignettes are as described in the notes to Figure 7.5 and Figure 7.6.

Source: Fernández, Immervoll and Pacifico (forthcoming[33]).

There are several reasons why gaps can arise, including:

  1. 1. Some economic activities do not give rise to SP entitlements (e.g. self-employment in many jurisdictions and casual or irregular work in some);

  2. 2. Past employment duration or social contributions are insufficient to qualify for benefits (e.g. if existing employment or contribution requirements are difficult to meet for those in unstable or marginal employment, see Figure 7.1).

  3. 3. Non-standard workers are treated differently during the claiming process (e.g. needs assessments for self-employed workers may be based on assumed rather than actual earnings, or an unemployed person searching for self-employment opportunities may not be considered an active job seeker); and

  4. 4. Already acquired entitlements are lost during a change in employment status or job (e.g. following a transition from dependent to self-employment when entitlements differ across employment statuses, or between jobs if they are tied to a specific employment relationship).

This section discusses policy options for addressing these and related gaps. It first discusses key challenges arising in the context of social-insurance systems (Section 7.3.1) and in universal and means-tested systems (Section 7.3.2). The remainder considers technical challenges related to volatile careers and earnings (Section 7.3.3), the role of activation policies in the future world of work (Section 7.3.4) and selected options for ensuring sufficient resources to support a growing role of SP in uncertain labour markets (Section 7.3.5).

7.3.1. Challenges in insurance-based social protection

Limited transferability of entitlements between jobs is a problem in earnings-related systems when entitlements differ across contractual arrangements or when they are tied to occupational schemes or to specific employers. Harmonising contributions and entitlements across the entire workforce, possibly accompanied by unified governance and administration structures, would make existing entitlements more portable.

Collecting and combining contributions from different schemes in one account, irrespective of the economic activity, is another – administratively simpler – way of ensuring that built-up entitlements are not lost in labour market transitions. For instance, Austria replaced its severance pay scheme with individual pension accounts in 2003. The previous severance-pay scheme only benefited employees upon lay-off, and therefore hindered labour mobility. In the new system, all dependent employees have pension accounts that receive regular employer contributions, but the account itself is independent of the employer and transferable across jobs. Transitions to self-employment suspend further employer contributions but they do not lead to a loss of entitlements. The measure increased job mobility for workers especially in distressed firms (Kettemann, Kramarz and Zweimüller, 2016[39]).

Entitlements derived from individual accounts can be subsidised explicitly (through credits) or implicitly (through the benefit formula). However, without substantial subsidies, individual accounts are a form of government-mandated saving whose scope and objectives differ from social insurance (see Section 7.1). In particular, “pure” individual accounts do not provide any risk sharing mechanisms and therefore cannot insure against catastrophic risks (such as long-term disability). Depending on how contributions are invested, they may also carry significant financial-market risks. As “pure” individual accounts do not incorporate redistribution, they are typically of less value to marginal or part-time workers.

Similar challenges limit the scope of publicly subsidised non-mandatory individual savings accounts that, in combination with means-tested safety-net benefits, have been suggested as an option for achieving adequate income protection for all, while keeping labour costs down (World Bank, 2018[6]). Disadvantaged and budget-constrained individuals can be unresponsive to savings incentives because of myopia, inertia, or information deficits. Subsidies then mainly divert savings toward subsidised products and crowd out unsubsidised savings. As a result, such programmes may mainly benefit individuals who already save even in the absence of subsidies or additional incentives (Chetty et al., 2014[40]).

One option to improve transferability of SP entitlements while maintaining acceptable replacement-rates for middle-income workers is to differentiate between earnings-related benefits (such as pensions, unemployment and short-term sickness benefits) and benefits that have no relation to earnings (health and long-term care) or where income smoothing is not a primary objective (longer-term unemployment or disability). Making the latter broadly accessible would help to safeguard basic entitlements irrespective of labour market transitions. As complements to these safety nets, earnings-related benefits could then be harmonised across contractual arrangements, to improve transferability (Cahuc, 2018[41]; Levy, 2008[42]).

7.3.2. Challenges in universal or means-tested social protection

Evolving labour markets arguably blur lines between traditional employment and different forms of independent work. Moreover, new types of atypical employment make it harder to reliably assess whether someone is working at all and how many hours they are putting into their job or jobs. As a result, tying social-protection entitlements and contributions to people’s employment status becomes more difficult. If existing strategies do not provide adequate coverage for all those in need, loosening the link between employment and entitlements could be one option for keeping social protection accessible, for resolving imbalances in SP entitlements across employment types, and for supporting labour market dynamism.

Moving towards greater universality through a form of basic income (BI) is an interesting proposal in this debate that has received considerable attention. No country has introduced a BI as a principal pillar of SP, however, and replacing large parts of existing support systems with a universal payment would be a major change. OECD simulations show that an unconditional payment to everyone at meaningful but fiscally realistic levels would require large tax rises as well as reductions in most current benefits, and would often not be an effective tool for reducing income poverty (OECD, 2017[43]; Browne and Immervoll, 2017[44]). Some disadvantaged groups would lose out when existing benefits are replaced by a BI, illustrating the downsides of SP without any form of targeting at all. In view of the immediate fiscal and distributional consequences of a fully comprehensive BI, reforms towards more universal income support would realistically need to be introduced gradually for specific groups (such as youth) or would need to be restricted in other ways.16 It would also require a parallel debate on how to finance a more equal sharing of the benefits of economic growth. From a broader economic-policy perspective, a downside of universal support is that, unlike out-of-work or needs-based benefits, it does not act as an automatic stabiliser: since it is paid regardless of income or employment status, spending levels do not go up during a downturn, and they do not fall during an upswing.

Another option is to strengthen needs-based support programmes by transforming some insurance programmes into means-tested assistance, or by expanding existing “safety-net” benefits for people with low incomes and no other resources. For example, Italy introduced a social assistance benefit, the Reddito di Inclusione in 2018. This will be replaced with the Citizen’s Income in 2019, which provides much higher benefit levels and aims to combine income support with activation measures (Bulman et al., 2019[45]). Such safety nets seek to ensure that those with high poverty risks receive basic support regardless of their past employment history or current work patterns. As tightly targeted schemes are financed from general revenues, their financing does not fall exclusively on workers. Non-wage labour costs can be reduced as a result, attenuating the incentives for automation (World Bank, 2018[6]) and encouraging formal employment (Levy, 2008[42]). Replacing contribution-based with general revenue-financed benefits can also help to align non-wage labour costs across employment forms, and thus ease distortions that arise from an unequal legal treatment of contractual arrangements (see Section 7.3.5).

Minimum-income safety nets are therefore an important element in countries’ strategies to alleviate poverty. However, targeted minimum-income schemes can be difficult to access in practice due to costly benefit-claiming processes, e.g. because of negative stigma, considerable information requirements and uncertainty regarding eligibility and entitlements, leading to substantial non-takeup (Bargain, Immervoll and Viitamäki, 2012[46]). They also do not offer significant income smoothing, except during spells with zero or very low (household) income. Unless complemented by benefits that provide significant replacement rates for middle-class workers, they do not provide insurance against major income losses for the majority with incomes well above the poverty line. In particular, out-of-work benefits paid in relation to household needs often do not provide any cash support to job losers in dual (or multiple-) earner households. Since re-employment support is typically tied to benefit receipt, the reach of active labour market programmes would then often exclude these groups as well (Section 7.3.4). A related challenge is that, while public support for safety nets can derive from an expectation that only the poorest require assistance, an exclusive reliance on targeting the poor minimises the majority’s stake in social protection and risks making funding less stable and potentially more susceptible to political cycles (Lindert, 2004[47]).

Needs-based provisions also risk crowding-out wages or support that is provided by employers as part of the pay package. In decoupling SP entitlements from jobs, a key policy challenge is therefore to ensure that government-provided transfers reach their intended recipients. The combination of an expanded in-work benefit and the introduction of a statutory minimum wage in the United Kingdom during the late 1990s provides a prominent example of such concerns, and of an approach for ensuring that income-tested in-work transfers benefit low-paid workers rather than their employers.17 However, an expansion of the number of low-paid independent workers who do not benefit from minimum-wage protection makes wage floors less effective in this respect.

7.3.3. Technical issues related to volatile careers and earnings

Despite these challenges, needs-based support will remain an essential complement to universal, insurance-based and employer-provided SP, and a necessary form of fall-back support while other SP pillars are adapted to a future world of work. For both needs-based and contributory systems, labour market changes present a number of very practical issues, however. In particular, the earnings of non-standard workers are typically subject to considerable fluctuations, notably in the case of self-employment. This complicates a reliable and timely assessment of household income and need, but also of the base for social contributions in insurance-based systems.18 Contribution payments that are not deducted at source but paid infrequently are also more visible and more likely to be perceived as taxes (Hershfield, Shu and Benartzi, 2018[48]).

Improved data transmission and real-time reporting of earnings tends to be costly (National Audit Office, 2018[49]) but can alleviate technical income measurement problems. Yet, a more fundamental question is what constitutes an appropriate payment or assessment frequency and a suitable reference period for assessing incomes and earnings. Payment modalities have implications for the responsiveness of SP systems, but also for the behaviour of claimants and contributors. Short reference periods make SP systems more responsive to people’s circumstances but can create incentives for gaming the system by timing incomes so as to maximise benefit entitlements or minimise contributions in a given period. For instance, earnings from self-employment can be shifted across periods, while contribution thresholds and ceilings depress individual contributions and total revenues when earnings are volatile. Longer reference periods sidestep such incentives especially in the case of contribution payments. But they can result in delayed and untimely support that does not match people’s current, and possible urgent, need for income or employment assistance. A related issue is that self-employed workers can struggle to fulfil minimum work or contribution requirements for benefit entitlement if there are significant time lags between performing work and receiving payment for it, e.g. if they only receive one payment for an extended period of work.

Volatile earnings may affect a growing number of workers in the future and accommodating earnings fluctuations in SP systems may therefore become more urgent. It is, however, not a new problem but rather a “legacy” issue that countries have been addressing in different ways, illustrating the trade-offs involved:

  • The Netherlands provides self-employed workers with interest-free loans to bridge temporary low-income periods and related liquidity problems (de Graaf-Zijl, Scheer and Bolhaar, 2018[50]). Building on such a system, one policy option could be a subsequent conversion of (some part) of such loans into benefits, e.g. once the low-income spell turns out to be persistent.

  • Denmark has harmonised unemployment-benefit entitlement rules for different types of dependent and independent workers by tying eligibility to taxable income over a three-year period, irrespective of the contractual arrangement. This also improves access for those who combine dependent and independent work. As an accompanying measure, self-employed workers who have ceased their business and receive benefits must not start a new one for six months. This “job search” period seeks to prevent self-employed from combining benefit receipt with continued independent work (OECD, 2018[51]).

  • Claimants of the Universal Credit, including the self-employed, in the United Kingdom are required to report earnings on a monthly basis. For longer-term benefit recipients, any earnings from self-employment are presumed to amount to at least the national minimum wage and reduce entitlements accordingly. This circumvents measurement problems for many self-employed claimants and is intended to avoid subsidising small businesses that are economically unviable, or independent workers who would be better off taking up employment in a regular minimum-wage job. However, while simple, this approach puts the self-employed with genuinely low earnings at a disadvantage (Citizens advice, 2018[52]; Low Incomes Tax Reform Group, 2017[53]). In particular, presuming minimum-wage earnings for a group that does not benefit from wage-floor regulations may exacerbate labour market disadvantages for workers who are faced with a choice between low-paid independent employment or not working at all, or who are actively encouraged by employment service providers to seek and accept independent work in the first place.19

  • New Zealand has instituted a welfare advisory group to address technical and broader issues related to labour market changes, including a growing availability of self-employment options, and their implications for SP capacity and responsiveness (Ministry of Social Development, 2018[54]).

7.3.4. Prevention and promotion: Activation policies in future labour markets

Large shares of workers transition in or out of jobs every year. Prior to the global financial and economic crisis, annual job separations and hires amounted to nearly 15% of total employment on average across the OECD. Job reallocation statistics were similar after the crisis, even if trends differed across countries (Falco, Green and MacDonald, forthcoming[55]) – see also Chapter 3. If accelerating technology adoption quickens the pace of job reallocation, the readiness and ability of workers to move from jobs in declining sectors or firms to expanding ones is likely to become an increasingly crucial determinant of future employment trends.

Changing work patterns and new forms of employment raise new questions about the scope and ambition of activation policies and employment-oriented SP. For instance, compared with out-of-work individuals with past standard employment, jobseekers with a history of self-employment can be less likely to rely on public employment services (PES) for their job search, e.g. if a lack of benefit entitlements means that there is no immediate financial incentive for regular interaction with the PES (Figure 7.7). The data in this figure do not provide information on people’s main channels for job search, or on the type of past self-employment (see figure notes), and this may explain why the differences between dependent employees and self-employed are in fact comparatively small in several countries, even in those where unemployment benefits are not typically available for self-employed. For some past self-employed, limited PES engagement may also reflect the availability of alternative job-search channels or a preference for a type of independent work that the PES is not perceived to facilitate. But limited engagement with the PES, for whatever reason, can become a growing concern when independent forms of work increasingly become substitutes for dependent employment.

copy the linklink copied!
Figure 7.7. In some countries, only a minority of jobseekers are in regular contact with the public employment service
By past employment status, 2017, in %
Figure 7.7. In some countries, only a minority of jobseekers are in regular contact with the public employment service

Note: Share of unemployed stating, in the context of questions about individuals’ job-search methods, that they “contacted the public employment office to find work during the previous four weeks”. Most respondents report more than one job-search method and data do not permit assessing the main one; this can be one reason why differences by previous employment status frequently appear small. Ages 25-64. 2013 data for Germany. Data for Ireland do not include information on past self-employment. Countries with fewer than 30 observations in any category are omitted.

Source: OECD calculations using EU Labour Force Survey.


Reintegrating displaced workers

“Displaced workers” – those facing involuntary job loss for technology-related or other economic reasons – find new jobs much more rapidly in some countries than in others. Whereas nearly 90% are re-employed within a year in Finland and Sweden, that share is only 30% in France and Portugal (OECD, 2018[56]). Providing timely and tailored employment support becomes more challenging when increasing numbers of workers are affected by job displacement, including through automation (see the discussion in Chapters 2 and 3). Some prevention and early intervention measures work best for workers who are laid-off after moderate or longer period of employment with the same employer. Indeed, one important difference between displaced workers and most other unemployed recipients of income and re-employment support is that there is in principle greater scope for proactive measures in the case of the former. Rapid response services, such as PES staff providing counselling or mediation events directly in a work place that will soon close, can jump-start the adjustment process by delivering re-employment services in a timely and targeted manner.

Such early intervention services can be quite effective but, except in the case of large-scale collective redundancies, are not widely used and require a form of social partnership that ensures that employers are actively involved in proactive re-employment measures, as for instance in Sweden (see also Chapter 5).20 Where labour markets are characterised by shortening job tenures or a diminished reach of traditional forms of worker representation, engaging employers in prevention and early intervention can become even more challenging. In a future world of work, these measures may become more difficult to access for growing shares of job losers as a result. Broadly available general activation and promotion programmes – for displaced workers and other jobseekers alike – are therefore likely to gain additional importance as central pillars of labour market reintegration strategies.

Activation of whom and towards what?

PES may need to adopt more effective outreach strategies to engage with workers who currently have work but may be at elevated risk of unemployment, such as those in unstable forms of work. Even though many countries support some groups of low-paid workers through in-work benefits or wage subsidies, the bulk of existing job-search assistance and related active labour market programmes exclude workers and focus on jobless people receiving out-of-work benefits. With the easier availability of task-based and independent forms of work, and a growing range of options for occasional short-duration or part-time employment, the traditional “in work” / “out of work” dichotomy appears increasingly anachronistic. In fact, in some countries, workers with intermittent employment and/or limited working hours already outnumber key categories of out-of-work individuals, such as those not working for family reasons, those unfit to work, or the longer-term unemployed (Figure 7.8). A strict focus on the jobless denies support from a large number of precarious workers who face barriers to higher-quality employment. It therefore becomes more and more unsatisfactory as basis for targeting activation and re-employment programmes.21

Extending access to active labour market support should move in tandem with tackling unintended benefit coverage gaps. For people with intermittent or precarious work this requires a careful review of key qualifying criteria for existing unemployment benefits, such as waiting periods, past employment requirements (see Figure 7.1) and rules regarding the extent and type of work that can be combined with benefit receipt (“part-time unemployment benefit”).22 In many OECD countries, some part-time work is compatible with the receipt of unemployment benefits and some countries offer partial benefits to workers whose working hours are cut.23 Yet, workers may not always be aware of this possibility (Stettner, Cassidy and Wentworth, 2016[57]). In contributory systems, effective out-of-work support for intermittent workers also requires provisions that allow “unused” entitlements (those that were not used in a given out-of-work spell) to be carried forward to future claim periods. An alternative (and partly functionally equivalent) strategy is the broader integration of out-of-work and in-work support programmes. The United Kingdom’s Universal Credit reform is a primary example of such a comprehensive strategy (Browne, Hood and Joyce, 2016[58]; Office for Budget Responsibility, 2018[59]).

The content and delivery of employment-oriented support also need to adapt. First, an evolving target group for active support requires sufficient and reliable resources for employment services and enough qualified frontline staff and caseworkers. User-friendly online services for jobseekers and automated systems that facilitate inter-agency access to relevant data (such as jobseekers’ labour market history and programme participation) can speed up initial jobseeker registration and other routine tasks.24 In general, technology can reduce pressures on operating budgets by creating more time for personal interaction between staff and jobseekers, and it enriches the toolkit for implementing broader activation strategies. For instance, IT-supported systems can support tailored and well-targeted services, e.g. through statistical profiling (OECD, 2018[60]). Yet, highly automated processes also create significant new risks, especially when introduced or used in a context of tight budget envelopes. An aggressive push towards replacing interpersonal contact with digital interfaces can compromise service accessibility and quality, notably for those with limited digital skills or those with complex needs where caseworker discretion is needed. Interventions and programme assignments using new generations of artificial-intelligence based profiling tools and decision support systems may also appear obscure or unfair and, as a result, might reduce their acceptance among some jobseekers, and their readiness for cooperating actively with PES.

copy the linklink copied!
Figure 7.8. Unstable work is common, but it may be outside the scope of activation measures
Employment Status over a 12-month period, average across seven countries
Figure 7.8. Unstable work is common, but it may be outside the scope of activation measures

Note: “Weak labour market attachment”: working for less than 45% of the potential full-time working hours in the calendar year (“unstable jobs”, “restricted hours”), or earning less than one third of the annual full-time statutory minimum wage (“near zero earnings”). Unweighted average across seven countries (Australia, Estonia, Ireland, Italy, Lithuania, Portugal and Spain), for 2014 and for ages 18-64 excluding students and military service.

Source: OECD Faces of Joblessness country studies ( using EU-SILC for European countries and the Household, Income and Labour Dynamics in Australia (HILDA) survey.

Second, new or growing forms of non-standard work require weighing the intended role and objectives of publicly provided labour market intermediation and employment services. A key issue is to what extent PES should actively connect people to very short-term work engagements (“gigs”), casual work (e.g. on-call employment or “zero hours” contracts) or independent forms of employment. Although the PES’ intermediation role can be partly redundant for platform work that is readily accessible online, some PES today already use web scraping technology to consolidate vacancies from a number (sometimes hundreds) of vacancy repositories. For instance, in the Netherlands, more than one third of vacancies that are listed on the PES job portal are sourced from other web resources using such technology. More fundamentally, however, a growing availability of alternative work arrangements raises the question whether employment services can and should prioritise different employment forms when recording openings in their vacancy databases or when offering vacancies to jobseekers. For instance, some Workforce Boards in the United States have held recruitment events for ridesharing services that are offered through online platforms (McKay, Pollack and Fitzpayne, 2018[61]) but such approaches frequently remain experimental and policies and practices differ both across and within countries. Even when active referral to non-traditional work is desired and compatible with PES procedures or guidelines, existing PES funding or reporting mechanisms may discourage it, e.g. if future earnings levels or stability are used for assessing the performance of reintegration measures but PES do not have ready access to information on earnings from self-employment. A growing availability of non-standard work contracts also highlights risks of creating “one-way streets” away from traditional forms of employment, e.g. if jobseekers are encouraged or assisted into self-employment but then have no clear route back to standard employment if they do not succeed in making a living from independent work.

More broadly, governments should review whether existing activation strategies for jobseekers, and those with unstable or low-paid work, strike a suitable balance between encouraging standard and non-standard employment, and are in line with policy objectives regarding job quantity and quality (OECD, 2018[2]). In the context of rights-and-responsibilities frameworks, such a review should for instance encompass the definition and application of “suitable-jobs” criteria, i.e. the type of jobs that benefit claimants must look for and accept in order to avoid benefit sanctions (Immervoll and Knotz, 2018[62]). In particular, integrating jobseekers into any kind of job may not be sufficient for individuals experiencing recurring cycles of precarious work and joblessness. Workers in sectors with significant exposure to automation or restructuring may also require access to employment services that re-assess their employment prospects, provide career guidance and develop coherent re-employment and, when needed, re-skilling strategies (see also Chapter 6).

Balancing incentives and support

Weak work incentives are rarely the only employment barrier for jobseekers, and often not the main one.25 But they are central to activation approaches which tie benefit receipt to active job-search and participation in labour market programmes. Importantly, a more fluid labour market with more options for when and how long to work creates more opportunities for acting on positive and negative incentives. This has significant implications for the scope of job-search and other behavioural requirements for benefit claimants, and for the design of tax-benefit systems more generally. First, incentives that favour specific employment forms can become more distortionary and economically damaging than they had been in the past (see Section 7.3.5).

Second, tax-benefit provisions, which create sudden income drops or gains as people vary earnings or working hours, are more likely to affect working time and earnings when choices are no longer constrained to, say, 40, 30 or 20 hours per week – see e.g. Saez (2010[63]). Avoiding excessive marginal effective tax rates (METRs), as caused by very high benefit withdrawal rates or by thresholds for tax liabilities or benefit entitlements, then becomes more desirable. Significant work disincentives, with METRs approaching or exceeding 100%, currently exist in a large number of OECD countries at lower earnings levels (Figure 7.9). With working hours and employment choices becoming more “elastic”, there may also be greater policy scope for setting explicit positive incentives towards socially desirable outcomes. This could for instance include tax reductions for inactive spouses taking up work as second earners (Immervoll et al., 2009[64]), or specific tax concessions or benefit bonuses for low-paid workers who move beyond marginal employment and work a certain minimum number of hours.26

Third, and perhaps most important, governments should review whether benefit reforms that are intended to tackle benefit coverage gaps create a need to rebalance the demanding and supporting elements of existing rights-and-responsibilities frameworks. Job-search and related activation requirements help in targeting support to genuine jobseekers and restrict opportunities for benefit receipt by others. An emergence of alternative working arrangements, with additional scope for arranging work or earnings patterns in a way that is compatible with benefit receipt, calls for additional efforts to formulate and enforce clear and reasonable responsibilities for benefit recipients. Likewise, extending the scope of job-search responsibilities and provisions for active participation in re-employment measures may be a necessary counter-weight to any extensions of benefit rights to new groups of jobseekers, such as part-time unemployed, those with intermittent employment, or those who entered unemployment after periods of self-employment.

copy the linklink copied!
Figure 7.9. Very high marginal effective tax rates can discourage people from working longer hours
Marginal effective tax rate (METR) for part-time workers, 2018
Figure 7.9. Very high marginal effective tax rates can discourage people from working longer hours

Note: Marginal effective tax rates (METRs) are the share of an earnings increase that is taxed away through higher taxes or reduced benefits. They are shown for an employee earning half the average hourly wage and working between 5 and 40 hours per week. The bars show the overall METR when moving from 5 hours per week to 40 hours per week. "Peak" is the maximum METR when increasing hours in 5-hour steps from 5 hours per week to 40 hours per week. Light-coloured bars indicate countries where peak METRs are at or above 100%, i.e. if there are situations when adding 5 working hours may not lead to higher net income. Calculations are based on four different family types (single, lone parent and dual-earner couple with and without children) and account for income taxes, employee-paid social contributions, and the following social benefits: minimum-income, cash housing, family, in-work benefits.

Source: OECD tax-benefit model TaxBEN (


7.3.5. Ensuring that social protection resources match evolving demands

Labour market transformations create new challenges for financing SP systems and can exacerbate existing, longer-standing ones. For a number of reasons, the resources that are needed for the provision of adequate working-age support can be expected to grow as a result. Accelerating labour market turnover in many countries (see Chapter 3), means that a higher share of workers would seek support both at any one point in time and throughout their careers. Closing SP gaps to include non-standard workers and supporting the transition and reintegration of displaced workers in declining industries will require corresponding SP budgets.

At the same time, the resource base for working-age support is being squeezed from several directions. The main funding base of SP systems is at risk of erosion if megatrends combine with rising market power of firms to perpetuate a decline of the labour share in national income – e.g. OECD (2018[32]), Sachs (2018[65]). On the spending side, population ageing puts growing pressures on pension, health and long-term care systems, which absorb increasing shares of available SP resources. Indeed, expenditures on old-age and survivor benefits over the past 25-30 years have grown substantially not only in total but, in spite of pension reforms, also on a per-capita basis. For instance, averaged across OECD countries with longer series of social spending, the spending on old-age and survivor benefits per individual aged 65 and older has grown from 22% of GDP per capita in 1990, to 32% in 2000 and 38% in 2013.27

Counteracting a decline in resources for working-age support requires a determined, coordinated and comprehensive approach that will need to include a policy debate on how new or expanded initiatives will be paid for, and who should pay. Key issues in this debate include ensuring that available resources are used efficiently. But it will also require debating options for broadening the tax base and strengthening revenue-collection technologies and enforcement (OECD, 2018[66]). In a context of a future of work with a growing choice of alternative working arrangements, a further challenge is that selective opting-in or opting-out raises sustainability risks in SP systems with voluntary membership (see Sections 7.1 and 7.2.1). This section focuses on two specific financing issues that are closely linked to the design of SP systems: incentives that SP financing provisions can create to encourage non-standard work, and achieving a suitable sharing of SP financing burdens across different employers.

Tackling financial incentives in favour of non-standard work

SP entitlements that are financed through an employment relationship represent non-wage labour costs. Where the cost of SP coverage is uneven across employment forms, employers and employees can lower costs by choosing between alternative working arrangements. Although these incentives are only one of many considerations that determine the choice of the most suitable employment form, differences in non-wage labour costs can be large. This is shown in Figure 7.10, which reports gaps in the “total payment wedge”, defined as the sum of personal income tax, social security contributions and other compulsory payments as a share of total labour costs, between dependent employees and different types of independent employment. For illustration purposes, results refer to someone with gross earnings equal to the average full-time wage in all cases.

Gaps are large where non-standard workers are excluded from certain parts of SP or if membership is voluntary. In both cases, legally mandated contributions can be significantly lower for non-standard workers as a result. In the Netherlands, the total payment wedge for an employee (51%) is more than twice as high as for an independent contractor (22%), meaning that total employment costs are 60% higher for an employee in this case.28 Most of the gap in the payment wedge (22 percentage points) is due to employer social security contributions, and a tax deduction for self-employed workers widens this gap further (OECD, 2018[67]).29 Of the seven OECD countries studied in Milanez and Bratta (2019[68]), the cost difference between hiring a dependent employee and the lowest-cost alternative working arrangement is smallest in Sweden, where independent contractors are included in most public SP programmes, with the exception of the bankruptcy fund (Kolsrud, 2018[19]). While Figure 7.10 focuses on statutory payments, further cost differences between employment forms can stem from employer contributions to SP measures that are part of collective agreements and provide top-up benefits for workers that fall within the scope of these agreements (see Section 7.2.1). De-facto gaps in contribution burdens can also be sizeable if participation in SP provisions is voluntary or allows some type of workers additional flexibility in the contributions they wish to make (Sections 7.1 and 7.2.1).

For workers, non-standard employment is especially attractive if contributions are lower than for standard dependent employment, while benefit entitlements are broadly similar (see Chapter 4). Where possible, both contributions and entitlements should be aligned across employment forms to prevent employers from engaging in regulatory arbitrage in the composition of their workforce. Several countries have taken steps to extend SP coverage to non-standard workers in order to curb a continued growth in their numbers. Austria gradually integrated independent contractors (freie Dienstnehmer), a hybrid between self- and dependent employment, into the social insurance system following concerns that employers used this contractual arrangement to avoid or evade social contributions. Since 2008, independent contractors are liable for the same (employer and employee) social security contributions as standard employees. While the number of independent contractors had grown steadily until early 2007, it began to fall after the reform was announced, reaching an all-time low in 2016 (Fink and Nagl, 2018[20]).30 In Italy, the gradual integration of para-subordinate workers into the general social insurance system since 2012 had similar effects (Raitano, 2018[24]).

copy the linklink copied!
Figure 7.10. Non-wage labour costs vary substantially across contractual arrangements
Total payment wedge by employment form, 2017, in percent
Figure 7.10. Non-wage labour costs vary substantially across contractual arrangements

Note: Relevant categories vary by country and country-specific employment forms are noted directly in the figure. The total payment wedge is the sum of income taxes, social contributions and other compulsory payments as a percentage of total labour costs.

Source: Adapted from Milanez and Bratta (2019[68]).


Balancing financing burdens across employers

A suitable balance of financing burdens between different groups of employers, including in sectors affected by automation-induced job displacement, may require new and innovative models of contribution financing. The financing of out-of-work support with conventional systems of employer contributions frequently does not link the employer costs to their layoff decisions and to the social costs that these layoffs entail. Without such a link, sectors or firms with more stable employment implicitly subsidise those where such layoffs are more frequent. Analogously to risk pooling between workers, this type of redistribution between firms can be desirable as an insurance against unexpected (e.g. cyclical) shocks. But it can lead to optimising behaviour when differences in layoff patterns are systematic or “structural”, e.g. if some firms reduce their labour costs by repeatedly hiring and laying off the same workers who can supplement their take-home pay with unemployment-insurance entitlements. In the context of automation, the absence of a link between financing and layoff behaviour creates distortions in favour of accelerated, and possibly excessive, substitution of workers by artificial intelligence or robots in some firms or sectors. Charging unemployment-insurance contributions as a function of layoffs (“experience rating” of unemployment insurance) could counter such distortions, place a greater share of the social cost of displacement at its source and contribute to a level playing field for labour-reliant and automation-intensive production technologies.

The United States is currently the only OECD country that applies experience rating of employer contributions in the context of unemployment insurance (OECD, 2018[32]), although experience rating is somewhat more common in other branches of SP, in particular invalidity pensions and work accidents (OECD, 2010[69]).31 For unemployment insurance, there is some evidence suggesting that experience rating reduces overall contribution burdens and increases employment.32 The specifics need to be designed with care, however, e.g. to avoid “cream skimming” (for instance, hiring only workers with low layoff risks to safeguard against rising insurance costs).

copy the linklink copied!7.4. Concluding remarks

This chapter has assessed the challenges for social protection (SP) in a future world of work and discussed reform avenues to tackle them. It has presented new estimates of gaps in existing SP provisions, including in employment forms that may become more prevalent as technological and workplace innovations broaden opportunities for alternative work arrangements or increase pressures on workers to accept them. Effective social risk sharing requires broad access to SP provisions and continued buy-in by workers. Unequal SP coverage or financing burdens could inhibit or undermine such a consensus, making protection inaccessible or too costly for those who need it most. Results suggest that reinforcing SP to tackle evolving labour-market risks is feasible in the context of different SP strategies, such as insurance or income-targeted assistance. But it requires determined efforts to adapt benefit provisions and employment support, and to secure an adequate resource base – sometimes by significantly upscaling it – for prevention, protection and promotion measures. One key challenge is to ensure a level playing field by tackling distortionary financial incentives to opt out of SP coverage, circumvent mandatory membership, or “game” the system by timing work or incomes to maximise entitlements or minimise contributions.

Preparing SP for the labour markets of the future requires a pro-active but iterative approach that addresses existing, and sometimes long-standing SP gaps, while adapting policy approaches as labour markets continue to evolve. This chapter presented “snapshots” of gaps in support that is available to broad groups of standard and non-standard workers, using individual-level data for the recent past. Future work should inform the policy reform debate by providing regular updates and by monitoring whether labour-market transformations lead to new gaps in SP provision. For instance, data permitting, such analysis should examine the accessibility and generosity of SP for specific circumstances of non-standard workers, such as employment through online platforms, as well as other forms of independent or dependent work that may emerge in the future.

copy the linklink copied!
Box 7.4. Policy directions

Governments should conduct a thorough review of their social protection (SP) systems to examine whether they provide reliable coverage against evolving labour-market and social risks. Most countries mix different SP design principles, such as means-testing or social insurance, and these provisions shape the ways in which rising non-standard work translates into specific SP access barriers. SP provisions themselves can drive trends towards non-standard employment. Where needed, SP provisions should be reinforced to ensure effective income and employment support for workers who are ill-equipped to benefit quickly from the opportunities of technological advances and dynamic labour markets.

Preparing SP for future labour markets requires a pro-active but iterative approach that addresses existing challenges while monitoring and adapting policy approaches as labour markets continue to evolve. Some challenges represent long-standing issues, but they can become more pressing as new technologies provide opportunities for alternative work arrangements. The correct classification of workers’ employment status is a pre-requisite for ensuring that they receive protection and support that is appropriate for their circumstances and risks (see Chapter 4).

However, even with well-defined legal categories and suitable enforcement in place, SP provisions can lead to significant support gaps for standard and, in particular, for non-standard workers. In order to ease access barriers to SP policy makers should consider:

  • Reviewing SP entitlement criteria, such as employment requirements, waiting periods and rules for combining or alternating benefit receipt with intermittent and other non-standard forms of work;

  • Enabling workers in independent forms of employment to build up rights to out-of-work support that is already available to standard employees;

  • Making SP provisions less rigid by ensuring that built-up entitlements are portable across jobs and forms of employment;

  • Maintaining or strengthening risk sharing across all labour market and income groups by tackling financial incentives that reduce the non-wage labour costs associated with non-standard work, such as reduced tax/contribution burdens or voluntary membership;

  • Making means tests more responsive to people’s needs by shortening the reference periods for needs assessments and by putting appropriate weight on recent or current incomes of all family members;

  • Subject to budgetary space, strengthening universal and unconditional forms of support, such as universal child benefits, as complements to existing targeted or insurance-based support measures.

Automation will lead to job displacement for many workers, while novel forms of employment are blurring the distinction between in-work and out-of-work categories. This raises new questions about the scope and ambition of activation and employment-oriented SP. Policy options and priorities include:

  • Tackling gaps in income support, which typically serves as the main gateway to labour market reintegration measures. This may require extending support for “part-time unemployed” and other jobseekers with intermittent, low-paid or independent employment;

  • Re-assessing the scope of claimants’ responsibilities, such as active job search, as a counter-weight to extending benefit rights. Such a review should ensure that the balance between supporting and demanding provisions remains in line with policy objectives regarding job quantity and quality. For instance, governments should consider if and when employment services should actively connect people to potentially precarious forms of work;

  • Ensuring that the content of active labour market programmes is well adapted to the needs and circumstances of an evolving client base. A growing share of part-time unemployed may call for shifting resources from work experience programmes or direct job creation towards tailored training and career counselling (see also Chapter 6).

Adapting SP to the future of work will create additional financing pressures at a time when SP budgets are already under pressure in many countries:

  • Keeping funding levels in line with evolving needs for support requires a determined and coordinated approach, including cost-effective SP delivery, better revenue-collection technologies and enforcement and a suitable balance of revenues from labour and non-labour tax bases;

  • Ensuring that SP systems remain fiscally sustainable also calls for tackling unintended incentives that distort employment or hiring decisions or encourage “gaming” of support systems by workers or employers;

  • The rationale for voluntary SP membership should be reassessed in light of labour-market developments. If new and emerging work patterns widen the scope for opting out of SP provisions, such opportunities could compromise the risk-sharing function of SP and erode its resource base;

  • Governments should also assess whether existing SP financing mechanisms achieve a fair balance of burdens between different employers, e.g. between those making little use of automation and those substituting large shares of their workforce with robots or artificial intelligence.


[80] Acemoglu, D. and P. Restrepo (2018), “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment”, American Economic Review, Vol. 108/6, pp. 1488-1542,

[81] Acemoglu, D. and P. Restrepo (2017), Robots and Jobs: Evidence from Us Labor Markets, (accessed on 19 November 2018).

[65] Agrawal, A., J. Gans and A. Goldfarb (eds.) (2018), R&D Structural Transformation, and the Distribution of Income, University of Chicago Press.

[14] Akerlof, G. (1970), “The Market for ’’Lemons’’: Quality Uncertainty and the Market Mechanism”, The Quarterly Journal of Economics, Vol. 84/3, p. 488,

[34] Albert, F., K. Gáspár and R. Gal (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts: Hungary”, in ESPN (ed.), ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts, European Commission.

[79] Allen, R. (2009), “Engels’ pause: Technical change, capital accumulation, and inequality in the british industrial revolution”, Explorations in Economic History, Vol. 46/4, pp. 418-435,

[78] Anderson, P. and B. Meyer (2000), “The effects of the unemployment insurance payroll tax on wages, employment, claims and denials”, Journal of Public Economics, Vol. 78/1-2, pp. 81-106.

[21] Arriba, R. and Moreno-Fuentes (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts Spain”, in ESPN (ed.), Access to social protection of people working as self-employed or on non-standard contracts, European Commission.

[77] Atkinson, A. (1996), “The Case for a Participation Income”, The Political Quarterly, Vol. 67/1, pp. 67-70,

[76] Autor, D. (2015), “Why Are There Still So Many Jobs? The History and Future of Workplace Automation”, Journal of Economic Perspectives, Vol. 29/3,

[46] Bargain, O., H. Immervoll and H. Viitamäki (2012), “No claim, no pain. Measuring the non-take-up of social assistance using register data”, Journal of Economic Inequality, Vol. 10/3,

[11] Barr, N. (1989), “Social Insurance as an Efficiency Device”, Journal of Public Policy, Vol. 9/01, p. 59,

[12] Boeri, T. and J. van Ours (2013), The Economics of Imperfect Labor Markets, Princeton University Press, Princeton, New Jersey.

[83] Bonoli, G. (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts Switzerland”, in ESPN (ed.), Access to social protection of people working as self-employed or on non-standard contracts, European Commission.

[58] Browne, J., A. Hood and R. Joyce (2016), “The (changing) effects of universal credit”, in Emmerson, C., P. Johnson and R. Joyce (eds.), IFS Green Budget, Institute for Fiscal Studies, London, (accessed on 8 January 2019).

[44] Browne, J. and H. Immervoll (2017), “Mechanics of replacing benefit systems with a basic income: comparative results from a microsimulation approach”, Journal of Economic Inequality,

[91] Browne, J. et al. (2018), “Faces of Joblessness in Ireland: Anatomy of Employment Barriers”, OECD Social, Employment and Migration Working Papers, No. 209, OECD Publishing, Paris,

[82] Brynjolfsson, E. and A. McAfee (2014), The second machine age : work, progress, and prosperity in a time of brilliant technologies, WW Norton & Company.

[45] Bulman, T. et al. (2019), “Tax-benefit reforms to support inclusiveness, productivity and growth in Italy”, No. forthcoming.

[41] Cahuc, P. (2018), “France: Social protection for the self-employed”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[74] Cahuc, P. (2018), “Wage Insurance, Part-Time Unemployment Insurance and Short-Time Work in the XXI Century”, IZA Discussion Papers, No. 12045, IZA, (accessed on 18 January 2019).

[75] Cahuc, P. and F. Malherbet (2004), “Unemployment compensation finance and labor market rigidity”, Journal of Public Economics, Vol. 88/3-4, pp. 481-501.

[73] Castell, L. et al. (2019), Take-up of Social Benefits: Evidence from a Nationwide Experiment in France, Paris School of Economics, Paris.

[9] Chetty, R. and A. Finkelstein (2013), “Social Insurance: Connecting Theory to Data”, in Handbook of Public Economics, Elsevier,

[40] Chetty, R. et al. (2014), “Active vs. Passive Decisions and Crowd-Out in Retirement Savings Accounts: Evidence from Denmark *”, The Quarterly Journal of Economics, Vol. 129/3, pp. 1141-1219,

[52] Citizens advice (2018), Universal Credit and Modern Employment: non-traditional work.

[72] Codagnone, C. et al. (2018), Behavioural Study on the Effects of an Extension of Access to Social Protection for People in All Forms of Employment, European Commission,

[38] Cragg, J. (1971), “Some Statistical Models for Limited Dependent Variables with Application to the Demand for Durable Goods”, Econometrica, Vol. 39/5, p. 829,

[50] de Graaf-Zijl, M., B. Scheer and J. Bolhaar (2018), “Netherlands: non-standard work and social protection”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[35] De Wispelaere, F. and J. Pacolet (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts: Belgium”, in ESPN (ed.), ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts, European Comission.

[90] Düll, N. et al. (2018), “Faces of Joblessness in Portugal: Anatomy of Employment Barriers”, OECD Social, Employment and Migration Working Papers, No. 210, OECD Publishing, Paris,

[18] Employment and Social Development Canada (2016), Evaluation of the Employment Insurance Special Benefits for Self-employed Workers,

[7] European Commission (2018), Access and sustainability of social protection in a changing world of work, Publications Office of the European Union,

[5] European Commission (2017), Commission Recommendation on the European Pillar of Social Rights,

[55] Falco, P., A. Green and D. MacDonald (forthcoming), “Are jobs becoming less stable?”, OECD Social, Employment and Migration Working Papers, OECD Publishing, Paris,

[89] Fernández, R. et al. (2018), “Faces of Joblessness in Spain: Anatomy of Employment Barriers”, OECD Social, Employment and Migration Working Papers, No. 207, OECD Publishing, Paris,

[33] Fernández, R., H. Immervoll and D. Pacifico (forthcoming), “Beyond repair? Anatomy of income support for standard and non-standard workers in OECD countries”, OECD Social, Employment and Migration Working Papers, OECD Publishing, Paris,

[20] Fink, M. and W. Nagl (2018), “Austria: How social protection rules affect self-employed and independent contractors”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[86] Ford, M. (2015), Rise of the Robots: Technology and the Threat of a Jobless Future, Basic Books, New York.

[71] Graetz, G. and G. Michaels (2015), Robots at Work, (accessed on 19 November 2018).

[10] Gruber, J. (1997), “The Consumption Smoothing Benefits of Unemployment Insurance”, American Economic Review, Vol. 87/1, pp. 192-205,

[48] Hershfield, H., S. Shu and S. Benartzi (2018), “Temporal Reframing and Savings: A Field Experiment”, SSRN Electronic Journal,

[16] Hills, J. (2017), Good Times, Bad Times (revised Edition): The Welfare Myth of Them and Us, Policy Press.

[92] Hofer, H., R. Hyee and G. Titelbach (forthcoming), “Where did all the independent contractors go?”, IHS Working Paper Series, IHS, Vienna.

[8] ILO (2018), Innovative approaches for ensuring universal social protection for the future of work,

[3] ILO and World Bank (2016), A shared mission for universal social protection. Concept note,

[64] Immervoll, H. et al. (2009), “An Evaluation of the Tax-Transfer Treatment of Married Couples in European Countries”, OECD Social, Employment and Migration Working Papers, No. 76, OECD Publishing, Paris,

[62] Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers”, OECD Social, Employment and Migration Working Papers, No. 215, OECD Publishing, Paris,

[17] ISSA (2012), Handbook on the extention of social security coverage to the self-employed.

[25] Jessoula, M., E. Pavolini and F. Strati (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts - Italy”, in ESPN (ed.), Access to social protection for people working on non-standard contracts and as self-employed in Europe, European Commission.

[39] Kettemann, A., F. Kramarz and J. Zweimüller (2016), “Beyond Severance Pay: Labor market responses to the introduction of occupational pensions in Austria”, Mimeo, University of Zurich,

[19] Kolsrud, J. (2018), “Sweden: Voluntary unemployment insurance”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[42] Levy, S. (2008), Good intentions, bad outcomes : social policy, informality, and economic growth in Mexico, Brookings Institution Press.

[47] Lindert, P. (2004), Growing Public, Cambridge University Press, Cambridge,

[53] Low Incomes Tax Reform Group (2017), Self-employed claimants of universal credit-lifting the burdens, Chartered Institute of Taxation, London, (accessed on 8 January 2019).

[84] Masso, M. and I. Kadarik (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts: Estonia”, in ESPN (ed.), ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts, European Commission.

[61] McKay, C., E. Pollack and A. Fitzpayne (2018), Modernizing unemployment insurance for the changing nature of work, The Aspen Institute.

[68] Milanez, A. and B. Bratta (2019), “Taxation and the future of work: How tax systems influence choice of employment form”, OECD Taxation Working Papers, No. 41, OECD Publishing, Paris,

[54] Ministry of Social Development, N. (2018), Establishing the welfare expert advisory group.

[27] MISSOC (2018), Mutual Information System on Social Protection,

[49] National Audit Office (2018), Rolling out Universal Credit, (accessed on 25 January 2019).

[56] OECD (2018), “Back to work: Lessons from nine country case studies of policies to assist displaced workers”, in OECD Employment Outlook 2018, OECD Publishing, Paris,

[2] OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD Publishing, Paris,

[67] OECD (2018), OECD Economic Surveys: Netherlands 2018, OECD Publishing, Paris,

[32] OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris,

[31] OECD (2018), OECD Reviews of Pension Systems: Latvia, OECD Reviews of Pension Systems, OECD Publishing, Paris,

[1] OECD (2018), Opportunities for All: A Framework for Policy Action on Inclusive Growth, OECD Publishing, Paris,

[60] OECD (2018), Profiling tools for early identification of jobseekers who need extra support, OECD, Paris, (accessed on 11 January 2019).

[66] OECD (2018), Tax Challenges Arising from Digitalisation – Interim Report 2018: Inclusive Framework on BEPS, OECD/G20 Base Erosion and Profit Shifting Project, OECD Publishing, Paris,

[51] OECD (2018), The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[43] OECD (2017), “Basic Income as a policy option: Can it add up?”, Policy Brief on the Future of Work, (accessed on 27 November 2017).

[26] OECD (2017), Pensions at a Glance 2017: OECD and G20 Indicators., OECD Publishing, Paris, http://doi:

[15] OECD (2017), Preventing Ageing Unequally, OECD Publishing, Paris,

[85] OECD (2014), Connecting People with Jobs : Activation Policies in the United Kingdom., OECD Publishing,

[69] OECD (2010), Sickness, Disability and Work: Breaking the Barriers: A Synthesis of Findings across OECD Countries, OECD Publishing, Paris,

[59] Office for Budget Responsibility (2018), Welfare trends report, (accessed on 8 January 2019).

[87] Pacifico, D. et al. (2018), “Faces of Joblessness in Italy: A People-centred perspective on employment barriers and policies”, OECD Social, Employment and Migration Working Papers, No. 208, OECD Publishing, Paris,

[88] Pacifico, D. et al. (2018), “Faces of Joblessness in Lithuania: A People-centred perspective on employment barriers and policies”, OECD Social, Employment and Migration Working Papers, No. 205, OECD Publishing, Paris,

[36] Pacolet, J. and L. Op De Beeck (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts: Luxembourg”, in ESPN (ed.), ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts, European Comission.

[24] Raitano, M. (2018), “Italy: Para-subordinate workers and their social protection”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[22] Rajevska, F. (2017), “ESPN Thematic Report on Access to social protection of people working as self-employed or on non-standard contracts Latvia”, in ESPN (ed.), Access to social protection of people working as self-employed or on non-standard contracts, European Commission.

[13] Rothschild, M. and J. Stiglitz (1976), “Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information”, The Quarterly Journal of Economics,

[63] Saez, E. (2010), “Do taxpayers bunch at kink points?”, American Economic Journal: Economic Policy,

[28] Social Security Administration (2018), Social Security Programs Throughout the World,

[30] Spasova, S. et al. (2017), Access to social protection for people working on non-standard contracts and as self-employed in Europe, European Commission, Brussels,

[23] SSA and ISSA (2017), “United States of America”, in Social Security Programs Throughout the World: The Americas, Social Security Administration, (accessed on 2 December 2018).

[57] Stettner, Cassidy and Wentworth (2016), A new safety net for an era of unstable earnings, The Century Foundation,

[70] Tamayo, H. and A. Tumino (2018), “Income protection of atypical workers in the event of unemployment in Europe”, EUROMOD Working Papers, No. EM18/18, ISER, University of Essex,

[4] United Nations (2015), Transforming Our World: The 2020 Agenda for Sustainable Development, (accessed on 4 December 2018).

[29] Whiteford, P. and A. Heron (2018), “Australia: Providing social protection to non-standard workers with tax financing”, in The Future of Social Protection: What Works for Non-standard Workers?, OECD Publishing, Paris,

[37] Wooldridge, J. (2010), Econometric analysis of cross section and panel data, MIT Press,

[6] World Bank (2018), World Development Report 2019: The Changing Nature of Work, The World Bank,


← 1. See Chapter 2 as well as Graetz and Michaels (2015[71]), Acemoglu and Restrepo (2017[81]; 2018[80]), Brynjolfsson and McAfee (2014[82]), Autor (2015[76]), Ford (2015[86]) and Allen (2009[79]).

← 2. Such ex-post redistribution in “actuarially fair” insurance systems is due to a number of reasons. First, risks may not be accurately known or knowable (e.g. if there is no functioning market for insuring certain types of risk, such as unemployment). Second, even if risks were known, they are an ex-ante concept. Those with unemployment spells that are longer or more frequent than suggested by their risk factors will get a better deal out of unemployment insurance, longer-lived individuals will see larger benefits from pensions, etc. Indeed, this type of redistribution from “lucky” people, towards those who are “unlucky” relative to their risk factors is the very point of risk pooling and distinguishes actuarially fair insurance from “pure” savings instruments.

← 3. Chapters 4 and 5, as well as OECD (2018[51]).

← 4. In addition, earnings levels may fall more readily for self-employed in response to market developments, e.g. because there are no minimum wages and downward wage rigidity does not apply to them. If entitled to unemployment benefits, those with poor earnings prospects may therefore have relatively strong financial incentives to become and remain unemployed.

← 5. For instance, claimants of unemployment benefits in Sweden are required to wind down or “freeze” their business and cannot claim benefits again for several years if they once again take up their previous self-employment activity after a benefit spell.

← 6. A recent survey of non-standard workers in Europe documented low willingness to pay for social protection (Codagnone et al., 2018[72]).

← 7. There is no federal maternity pay scheme in the USA.

← 8. Australia further provides tax concessions for small business owners who convert assets into retirement savings.

← 9. For instance, Spain took legislative measures to reduce the gaps in out-of-work support and contributions burdens between the self-employed and dependent employees (Royal Decree-Law 28/2018, of 28 December). France has also taken legislative measures to provide access to unemployment benefits for the self-employed (see endnote 12). Italy significantly expanded minimum-income provisions in 2018 and 2019 (see Section 7.3.2) and introduced a number of changes to the unemployment benefit system in 2015 (Pacifico et al., 2018[87]): Minimum contribution requirements are now shorter and maximum benefit durations have been extended. Certain categories of workers who were previously excluded from the unemployment insurance are now covered (e.g. seasonal workers) and some groups of long-term unemployed can rely on a means-tested unemployment assistance programme.

← 10. Detailed statutory information on benefit rates and durations is available from the OECD tax-benefit policy database through

← 11. Although non-takeup remains significant for means-tested benefits in France (Castell et al., 2019[73]), recipient numbers are comparatively high for these assistance benefits. See OECD SOCR database at

← 12. In summer 2018, the French parliament passed a law that provides for flat-rate unemployment benefits for jobseekers who become unemployed after a period of self-employment with earnings of at least 10 000 euro per year and subject to liquidation of the former business. The benefit, with a duration of 6 months, was scheduled to come into force in January 2019 but implementation has been delayed.

← 13. Fernández et al. (forthcoming[33]) extends this approach to other countries.

← 14. An alternative approach would be by means of deterministic tax-benefit microsimulation that uses sophisticated computer representations of theoretical entitlements in order to estimate benefit amounts at the individual level, see Tamayo and Tumino (2018[70]) and Browne and Immervoll (2017[44]) for recent multi-country applications. These models typically use cross-sectional data and cannot account for dynamic aspects, such as past work history and employment patterns, that are especially relevant for assessing entitlement gaps between standard and non-standard workers. They also focus on theoretical entitlements and cannot take full account of factors such as stigma, benefit sanctions or voluntary SP opt-ins or opt-outs. Together, these factors can account for differences between theoretical and actual benefit receipt.

← 15. The panel component of The European Union Statistics on Income and Living Conditions (EU-SILC) covers four years. But since it is a rotating panel, limiting the time-window of interest to only three years results in bigger samples.

← 16. E.g. by tying eligibility to socially useful activities, as in Atkinson’s (1996[77]) Participation Income.

← 17. For instance, Australia provides for income and asset-tested benefits, as well as SP entitlements, such as sick pay and carer’s leave, that are part of the remuneration package of dependent employees (Section 7.2.1). Casual workers, who are employed on an ‘as needed basis’ are typically entitled to a higher hourly pay rate than equivalent full-time or part-time employees. This usually incorporates a 25 per cent loading instead of entitlements to paid personal/carer’s leave, annual leave, notice of termination and redundancy pay under the National Employment Standards. The lack of workplace entitlements contributes to their more frequent receipt of general revenue-financed benefits (Whiteford and Heron, 2018[29]).

← 18. Different contribution bases are possible for the self-employed. Where profits are used as the contribution base, self-employed workers can seek to minimise contributions by inflating their costs. Basing contributions on revenues or turnover, by contrast, is technically more straightforward (e.g. for platform-based workers if platforms are obliged to reveal the relevant information). But it may result in excessive contributions for workers performing higher-cost activities. For instance, in France, micro-entrepreneurs pay social contributions based on their gross revenue. While contribution rates are lower for those workers in retail or hospitality industries than for workers with more limited costs such as the liberal professions (Cahuc, 2018[41]), they might be excessive if profit margins are low. More generally, self-employed workers draw income from both capital and labour. For those with significant capital income, this could substitute to some extent for social benefits and could reduce the need for publicly provided income smoothing. The distinction between capital and labour income is less meaningful for many self-employed whose capital investments are very limited, e.g. those performing specific tasks on internet platforms.

← 19. There have been concerns that the structure of performance-related payments to private providers may incentivise them to encourage or “push” clients towards self-employment (OECD, 2014[85]).

← 20. The job security councils operated by social partners in Sweden, demonstrate that early intervention measures can be offered to all displaced workers, including those affected by individual or small-scale layoffs, when employers and unions are constructively engaged.

← 21. Their barriers to higher-intensity employment are frequently similar to the employment obstacles faced by the jobless. See Faces of Joblessness country studies, available at

← 22. See Cahuc (2018[74]).

← 23. See OECD tax-benefit policy database,

← 24. Recent initiatives in this direction include the Korean WorkNet PES portal, the Digital First strategy in Flanders, Belgium, and the automatic collection and processing of earnings-related information in the context of the United Kingdom’s Universal Credit.

← 25. Fernández et al. (2018[89]), Browne et al. (2018[91]), Pacifico et al. (2018[87]), Pacifico et al. (Pacifico et al., 2018[88]), Düll et al (2018[90]), Fernández et al. (2018[89]).

← 26. The in-work tax credits (Working Credit and Working Family Tax Credit) that existed in the United Kingdom prior to Universal Credit are a primary example of such a policy configuration.

← 27. Increasing per-capita spending on pensions is in line with growing employment rates, especially among women, and the resulting growth in the number of people with significant pension entitlements. The average spending figures quoted in the text are for the following 18 countries: Austria, Australia, Chile, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, United Kingdom and United States.

← 28. Total employment costs (T), gross earnings (E) and payment wedge (PW) are related as follows: (1 + PW / (1-PW) ). Payment wedges of 51% and 22% translate into total employment costs of 204% and 128% of gross earnings, respectively.

← 29. The contributions gap is due to unemployment and disability insurance, as well as second-pillar pension contributions to which self-employed workers are not entitled (see Section 7.2.1).

← 30. The fall in the number of independent contractors was mainly due to a declining inflow, rather than by the dissolution of existing independent contracts. However, the likelihood of being in standard employment within one month of transitioning out of an independent contract increased from 11% before the reform’s announcement in 2006 to 13% after the reform’s implementation. In addition, the share of workers who transitioned out of the labour force upon exiting an independent contract dropped significantly, in line with the fact that prior to the reform, independent contractors were not covered by unemployment insurance (Hofer, Hyee and Titelbach, forthcoming[92]).

← 31. The US system is incomplete in the sense that employers are not charged the full cost of a layoff (Anderson and Meyer, 2000[78]).

← 32. For the United States, evaluation results in Anderson and Mayer (2000[78]) point to falling unemployment benefit claims and to positive overall employment effects. See also Cahuc and Malherbet (2004[75]).


This work is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and arguments employed herein do not necessarily reflect the official views of OECD member countries.

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Note by Turkey
The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union
The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Photo credits: Cover © Jacob Lund, Dragon Images, PaO_STUDIO, Daisy Daisy /

Corrigenda to publications may be found on line at:

© OECD 2019

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at