1Levels of both part-time and temporary employment (including fixed-term and temporary agency work) have been rising in recent decades not least due to a relaxation or lifting of rules on employment regulation governing the use of fixed-term and temporary agency work while keeping the regulation on permanent employment stable and the conscious promotion of part-time work by governments in the light of increasing female employment. Young people and women have been disproportionately affected by this trend. When the economy was hit by the recent crisis, temporary workers were often the first to lose their jobs and their average share thus decreased during the early phase of the crisis. Re-employment often takes place on the basis of temporary jobs, however, and the fact that temporary employment is on average already picking up again from 2010 illustrates this fact. It stands at 13.7 in 2012 for the EU-27 with slightly higher shares for women than men. 42.1% of young persons (16-24 years) in the EU-27 had a temporary employment contract in 2012 – as much as 36% of these stated that their main reasons for working on a temporary contract was that they could not find a permanent job. Part-time employment continued to grow during the economic crisis to a level of 19.2% in 2012 for the EU-27; the respective share for women and men are 32.1 and 8.4. This growth can be explained only in part by the use of worksharing measures. Particularly in countries that were most strongly affected by the economic crisis, the share of involuntary temporary and part-time employment increased markedly. These average movements notwithstanding, countries show very different patterns with regard to the extent of non-standard employment and developments during the crisis (Eurostat, online data base). 
2Previous research showed that non-standard employment does not always act as a stepping stone but is often permanent or recurring and more often associated with transitions out of employment to either unemployment or inactivity (Anxo and O’Reilly 2000; Erhel et al. 2010; European Commission 2009; Gash 2008; Leschke 2009; Muffels 2008). As regards the security side, the little research that is available shows that non-standard workers are disadvantaged in terms of access to unemployment benefits (to varying degrees depending on country and group of workers in question). Upon access, they may actually be in a position to receive proportionately higher benefit levels due to the progressive nature of some of the systems (low benefit ceilings, flat-rate schemes, etc.) (see Grimshaw and Rubery 1997; Talós 1999; Klammer and Tillmann 2001; Eurofound 2003; Leschke 2007, 2008; Schulze Buschoff and Protsch 2008).
3In the context of the flexicurity agenda, the deficient access to social security benefits and among them unemployment benefits, did gain some momentum in Europe during the second phase of the Lisbon Strategy.  This is neatly illustrated by the fact that the new integrated guidelines on economic and employment policies as part of the EU2020 strategy explicitly highlight the importance of ensuring adequate social security for workers on fixed-term contracts and the self-employed – interestingly parttime workers are not mentioned here (Council of the European Union 2010) . Similarly, the OECD Employment Outlook 2010 contains a study on the quality of part-time employment that, among other issues, looks at access of part-time workers to unemployment benefits. Both the European Commission and the OECD studies point to numerous forms of discrimination of non-standard workers with regard to access to unemployment benefits. Discrimination is more likely in primary than in secondary unemployment benefit schemes, but evident also in the latter.
4Lately, triggered by the economic crisis, the unemployment benefit coverage of non-standard workers has received considerable attention at European and national level. Fast and steep increases in unemployment, and the disproportionate affectedness of some labour market groups such as young and/or temporary workers – or more generally speaking workers with short contribution histories – emphasised the gaps in social security (see Council of the European Union 2011; European Commission 2010a and 2010b; ILO 2010; OECD 2009b; Immervoll 2009 for the OECD). These studies highlight that growing shares of workers may remain unprotected if temporary work and other non-standard work patterns become more common and they also point out that a lack in protection for these workers is of particular concern during an economic downturn as they are likely to experience a disproportionate share of job losses. There is some disagreement on how to best achieve a more comprehensive coverage of non-standard workers. While some proponents voice a preference for ensuring the availability and adequacy of lower-tier schemes such as social assistance rather than ad-hoc changes to unemployment insurance others remain more open in their policy recommendations pointing to the need for a debate on the relative roles of insurance and assistance benefits with regard to better coverage of non-standard workers. Recent volumes by Clasen and Clegg (2011) and Lefresne (2008) provide country case studies that address the extent to which benefit schemes have adapted – or have failed to adapt – over the last decades to the major changes that have affected labour markets including the recent economic crisis.
5Indeed, as a response to the recent economic crisis and the surging unemployment affecting specific labour market groups disproportionally, a substantial number of countries have improved the access of non-standard workers and/or young workers to unemployment benefits and some countries have also opened up their state-subsidised short-time working schemes to new labour market groups. 
6There are thus some seemingly contradictory developments with regard to non-standard workers in the crisis. On the one hand, these workers have in many cases been more affected by unemployment but, on the other hand, emphasis has in several countries been placed on improving their security in employment (short-term working schemes) or unemployment (unemployment benefits, active labour market policies). Part-time employment has continued to grow and new employment in the second phase of the crisis has often been created on a temporary basis. This paper thus sets out to ascertain whether the developments in the labour market and in the welfare system during the initial phase of the economic crisis can be seen as continuing the trend of labour market segmentation  or whether the crisis may indeed have contributed to containing some of the divisions created over the last decades.
7The next section focusses on the changes in the design of and access of various groups, and particularly youth and women, to unemployment benefits during the first years of the crisis (2008-2010). There are important limitations in the data with regard to unemployment benefit coverage. We do not have comparative figures on the coverage of non-standard workers. However, data, taken from the harmonised aggregate European Labour Force Survey, is available by gender and age which allows us to proxy access of part-time and temporary workers to a certain degree. Moreover, and importantly, labour force survey data on receipt of unemployment benefits has been found to be unreliable with regard to country differences. We therefore decided not to report absolute coverage levels but merely the relative differences between sub-groups and changes over time. Section 3 concludes discussing segmentation in both labour market and welfare state outcomes.
1 – Segmentation with regard to the welfare system? Access of non-standard workers to unemployment benefits
8As a result of the large output shocks experienced by the majority of European countries, EU unemployment increased markedly from 7.2% in 2007 to 9.7 % in 2010 and 10.6% in 2012. The EU average unemployment rate conceals large country differences in terms of extent and developments. Unemployment rates range from below 5.6% in Germany, the Netherlands and Luxembourg and as low as 4.4% in Austria to around 25% in Greece and Spain. Several countries (Estonia, Latvia, Portugal and Denmark) doubled and some (Ireland, Greece, Spain, Lithuania and Cyprus) tripled their unemployment rates between 2007 and 2012. Young and low-qualified workers were particularly affected by increasing unemployment. Both groups already had considerably higher-than-average levels of unemployment before the crisis. Youth unemployment is now close to 23%, more than 7 percentage points up from 2007.
9Substantial shares of unemployed persons do not have access to unemployment (insurance) benefits. In part, this is deliberate policy (i.e. a link between contributions and benefits is intended), while in other cases it is unintended, insofar as qualifying criteria for unemployment insurance are in many cases no longer in line with the changes in employment relationships (see e.g. OECD 2009a). As social security schemes are still largely based on “standard employment” (dependent, permanent and full-time) relationships, a number of basic principles underlying the design of unemployment insurance can lead to disadvantageous treatment of, among others, non-standard workers, in particular with regard to access to unemployment benefits. As has been highlighted above, women and youth are considerably more likely to have a part-time job, and, respectively, a temporary contract. Earnings or hours thresholds represent a direct criterion for excluding those working short hours. Moreover, the qualifying period (usually a minimum contribution period within a given reference period) can further restrict the access of persons whose contracts are of short duration. When the contribution period is measured not in months but in hours or days worked this can also represent a problem for workers with fewer than full-time hours per month. Moreover, in several countries the length of benefit payments depends on the previous period of contributions. Hours and earnings thresholds, as well as qualifying periods, are in most cases not of relevance for secondary benefit schemes. However, secondary schemes (unemployment assistance benefits) exist in only a limited number of European countries and new member states, in particular, do not, for the most part, have unemployment assistance benefits. Moreover, unemployment assistance benefits are not always universal (for example, they are in some cases restricted to unemployed persons with families), while benefits are usually very low and means-tested on the family or (more rarely) the individual and/or assets (compare OECD Benefits and Wages Indicators).
10This section seeks to highlight the disadvantages suffered by non-standard workers in access to unemployment benefits, with special reference to changes during the first period of the crisis (2008-2010). As unemployment benefit systems are designed to meet complex (and fast changing) conditions, there is no room here to refer to the detailed qualifying conditions and other design features. Regularly updated information on the design of unemployment benefits systems can be found in the MISSOC comparative tables (2012). Detailed information on unemployment benefit schemes, particularly with regard to part-time workers, is also available from a special OECD (2010b) survey.  These sources are used to give an aggregate view of possible disadvantages deriving from eligibility criteria such as hours’ and earnings thresholds and qualifying periods. The subsequent section will give an overview on changes in unemployment benefit schemes in the first part of the crisis, a period during which – as part of the broader stimulus measures – several countries have relaxed conditions and thus opened up their schemes to new groups of workers. Information for this section derives from comparative sources such as OECD, European Commission and Eurofound. Last but not least aggregate Labour Force Survey data on unemployment benefit coverage rates and recent changes will allow an idea to be gained of the “real” coverage of non-standard workers. As a breakdown by former contract form and/or working time is not available, we will use as proxies the comparison between men and women and youth and prime-age workers.
2.1 – Review of disadvantages with regard to access to and level of unemployment benefits
11OECD (2010a: table 4.2, p. 235) and OECD (2010b) provide information on the unemployment benefit systems of 21 of the EU27 countries.  This information allows us to gain an idea of the number of countries in which specific labour market groups, such as part-time or temporary workers, are faced with disadvantages in access to unemployment insurance benefits. As the unemployment benefit design is extremely complex and varied and subject to frequent change, the following is no more than an indication and it should be borne in mind that countries differ markedly in terms of specific design features and that these features are subject to interaction. This means that even if a country has several restrictive features (see footnotes 11-14), it might still turn out to have higher overall benefit coverage than another country with, at first sight, fewer restrictive features.
12According to the data collected in OECD (2010a and 2010b), in six out of 21 countries  the qualifying period is calculated in days or hours rather than in weeks or months, which will place at a disadvantage workers with fewer than full-time hours or fewer working days in a week. Moreover, seven countries  have a minimum hours or earnings threshold in addition to the usual eligibility conditions. Twelve countries  have a relatively short reference periods for contributions (the equivalent of one year of contributions in a two-year reference period or shorter) which can pose a problem, particularly for temporary (and other groups of) workers with short tenure and frequent spells of unemployment. The majority of countries (12 out of 21)  base the length of benefit payments on the contribution record which can create further disadvantages for non-standard workers. Again, it is important here to bear in mind that countries differ substantially with regard to the duration of unemployment insurance benefit payments. Countries such as the UK or the Czech Republic do not take into account contribution records for calculating the benefit period but grant uniformly short unemployment insurance benefits of only six months. This also has an implication for secondary assistance schemes (where they exist) as in these cases means-testing sets in at a very early stage.
13In most countries unemployment benefit levels are calculated as a share of previous earnings, placing part-time workers at a disadvantage with regard to absolute benefit levels.  Temporary employment often entails a wages penalty which is then also carried on into the benefit system. However, there are a number of mechanisms in unemployment benefit systems that can lead to redistribution between those with higher and those with lower earnings. The most obvious examples are countries such as the UK or Poland that apply universal flat-rate benefits. However, flat-rate benefits are in most cases very low and sometimes adjusted in case of part-time work (e.g. Ireland). Another example that leads to redistribution is the case of high universal minimum benefits (e.g. Sweden and Denmark); however, these are often adjusted in the case of part-time workers. Low benefit ceilings also contribute to increasing the net replacement rates of unemployed persons with lower former earnings as compared to those with higher earnings, a good example being Denmark, albeit with the part-time insured receiving only two thirds of the maximum benefit. Last but not least, some countries have introduced specific regulations to increase the unemployment benefits of low-wage-earners (e.g. Austria) (compare MISSOC 2010, OECD 2010b). Due to these redistribution mechanisms, the net replacement rates of part-time workers (insofar as they have access in the first place) can be higher than those of full-time workers (Leschke 2008).
2.2 – The first period of the crisis (2008-2010) - reforming unemployment benefits to better include non-standard workers
14In the past, improving the access of non-standard workers to unemployment benefits was seldom an explicit trigger for reform.  However, in some countries, as a side effect of other reforms, hours or earnings thresholds have been lowered or abolished and some schemes have been opened up to new groups of workers such as certain groups of self-employed. On the down side, some of the reforms aimed at improving the financial sustainability of unemployment benefit systems have further worsened the situation of non-standard workers. This is visible particularly in the trend to strengthen secondary systems by shortening the duration of non-means-tested benefits and tightening up qualifying criteria.
15The recent economic crisis, however, triggered a number of reforms in several countries that were explicitly aimed at better including specific non-covered or less well covered labour market groups, among them non-standard workers. This can also be seen as a reaction to the strong level of affectedness by (prolonged) unemployment of groups such as temporary workers, among them many young workers, which suddenly made the sometimes large gaps in unemployment benefit coverage very evident.
16In the following we will review the changes to unemployment benefit systems in the EU27  that have taken place in the first part of the crisis with regard to qualifying criteria (increasing eligibility), benefit level and duration (increasing generosity).  An important question is whether these changes represent long-term adaptations of the UI/UB systems to new trends in employment or whether they were shortterm and temporary reactions to the crisis. The fact that some of the measures were temporary in the first place and that the austerity packages of several countries placed a strong focus on cuts in the social and employment field (for a review of selected countries see Heise and Lierse 2011) hints at the latter. Active labour market policies – not a focus of this paper – and particularly job search support have also been strengthened in many countries.
17Qualifying criteria were relaxed in Finland, France, Portugal, Latvia and Slovenia; the latter three countries previously had very strict qualifying criteria. This is beneficial for employees with short tenure such as temporary workers. In Slovakia parental leave now counts as a contribution period for gaining access to benefits (for details see European Commission 2010b: 137; European Commission 2011: 76-78). Greece grants lump-sum payments to job losers and other vulnerable groups and, similarly, France accorded a one-off payment to jobseekers not eligible for unemployment benefits due to insufficient contributions.  Several temporary improvements have also been made to the Italian system for, among others, apprentices and project workers (one-off payments). The Spanish government approved a temporary flat-rate unemployment assistance benefit payable for six months to all unemployed persons whose unemployment insurance benefit had ceased since January 2009.  Meanwhile, from May 2009 Ireland actually reduced the benefit level for young claimants (18-24 years) (European Commission February 2011).
18Unemployment benefit levels have been increased, during the early period of the unemployment spell in Belgium, the Netherlands, Bulgaria, the Czech Republic and Poland, and regardless of the duration of unemployment in Latvia and Finland (European Commission 2011: 76-78).
19Benefit duration in unemployment insurance or assistance has been increased in Finland, Romania, Latvia and Lithuania, in the latter only in municipalities hit particularly hard by the crisis (European Commission 2011: 76-78). Inversely, Ireland, the Czech Republic, Poland, France and Denmark  have decreased the maximum duration of UI benefits. Second-tier systems such as social assistance have been improved in a number of countries, as evidenced by increases to, for example, housing support. Some countries (e.g. Estonia) had planned improvements to their unemployment benefits systems which, due to financial pressure on the systems, were not carried through.
20Table 1 gives an overview of the countries that modified their unemployment benefit schemes or their short-time working measures (for details on the latter see Leschke 2012) to improve coverage or increase benefit levels. It is interesting to note that a number of countries increased eligibility and or (initial) benefit levels while at the same time shortening the duration of benefits; France and Poland are obvious examples in this regard (strategy of exchanging or trading rights). However, there are also examples, such as Finland and Latvia, where all components of unemployment benefits were improved as a response to the crisis, and others, such as Ireland, where several components of unemployment benefits were reduced.
Countries with modifications in unemployment benefit systems or short-time working schemes during the first period of the crisis (2008-2010)*
Countries with modifications in unemployment benefit systems or short-time working schemes during the first period of the crisis (2008-2010)**depending on the country and the reform, the changes affected either all unemployment benefit recipients or certain categories of workers. For detailed information see text and footnotes.
2.3 – Changes in unemployment benefit coverage during the crisis
21Unemployment benefit systems differ markedly between countries not only in their overall coverage rate but also in their generosity with regard to benefit level and duration. One way of illustrating this is to look at country differences with regard to expenditure on unemployment benefits and on active labour market policies and (both measured as a percentage of GDP) in relation to the unemployment rate. As Figure 1 shows, there is no relationship between unemployment rates and either passive (PLMP) or active expenditure (ALMP), as illustrated for example by Spain and Latvia (Latvia, at comparable unemployment figures, spending less than one third of the amount that Spain is spending on unemployment benefits). Similarly, countries such as Denmark and Belgium with average unemployment rates, display the highest expenditure on active measures. Generally, the new member states, but also the UK and Greece, are low spenders. As a reaction to rising unemployment rates in the crisis, all countries increased expenditure on passive measures (in some cases the money was spent in a preventative way on short-time measures). The typical crowding out of active measures, insofar as as PLMPs and ALMPs are often paid from the same source, was evident in only a few countries, among them the UK and Italy. On the EU27 average, expenditure on active measures and on labour market services increased, albeit less steeply than expenditure on unemployment benefits.
Expenditure on unemployment benefits and active labour market policies as % of GDP and unemployment rates, 2010*
Expenditure on unemployment benefits and active labour market policies as % of GDP and unemployment rates, 2010*Data source: Eurostat, online database.
*Note: UK data for labour market services and active labour market policies refers to 2008.
22The remainder of the section looks at unemployment benefit coverage for different labour market groups and developments during the crisis. There are several problems with regard to the data. The aggregate European Labour Force Survey data provides information on persons registered at the public employment office and in receipt of unemployment benefit or assistance.  However, the information on persons in receipt of unemployment benefits contained in labour force survey data has been identified as unreliable due to, among other things, underreporting and misreporting (Immervoll et al. 2004: 58-67).  Indeed Immervoll et al. (2004) show for a number of countries that the figures on unemployment receipt rates from administrative data sources and those from labour force surveys differ substantially and that the differences are not always in the same direction. For this reason, we do not show absolute levels of benefit receipt but rather distributions and changes over time in relative terms. In the absence of information on the coverage of non-standard workers, the comparison between men and women (men being more likely to be permanently and, particularly, full-time employed) and between youth and primeage workers (prime-age workers being more likely full-time and, particularly, permanently employed) will allow us to obtain an approximation of the inclusiveness of the different countries’ benefit systems. Data is missing for several countries for reasons of confidentiality but EU27 aggregates are available. Moreover, the LFS data does not distinguish between unemployment insurance and assistance. The data allow assessment of benefit coverage at different durations of unemployment – due to fewer missing countries and less parsimonious data, we will focus for the most part on short-time unemployed (1-2 months).
23While countries can be seen to differ substantially in terms of unemployment benefit coverage, comparative information on this issue is sparse. Factors of relevance for overall coverage include different eligibility governing the schemes, varying incentives to register with the employment office, availability of assistance schemes and differences in the characteristics of unemployed workers.  Gender differences are evident but not extremely large on the EU27 average – at least for short-term unemployed (Figure 2). However, we know from other sources that men are more likely than women to have access to unemployment insurance benefits, whereas women often have to rely on unemployment assistance or other secondary benefits which are usually less generous and more likely to be means-tested (see e.g. Leschke 2008). Among the countries with available data coverage is higher for women than for men in only four countries – Romania, Bulgaria, Slovenia and, with only small gender differences, Denmark (probably due to the fact that, in this latter country, women are more likely than men to participate in the voluntary unemployment insurance). France shows approximately equal coverage. The other 14 countries with available data have higher coverage for men than for women, with particularly large differences in the UK, Slovakia and Italy. The large gender differences in the UK are likely to be related to strict qualifying conditions for contribution-based job-seekers’ allowance (unemployment insurance) and, at the same time, the strict means-testing applied to income-based job seeker’s allowance (assistance scheme) thereby affecting women more than men insofar as they are predominantly secondary earners in couple households.
Comparison of registered short-term unemployed in receipt of insurance or assistance by gender, 2010 (total=100)
Comparison of registered short-term unemployed in receipt of insurance or assistance by gender, 2010 (total=100)Data source: Eurostat, online database.
Note: Duration of unemployment 1-2 months. Age: 15-64 years. Countries are sorted by the difference between female and male coverage rates in ascending order.
24Examination of the same figure for youth (15-24 years) and adult workers (25-64 years) (Figure 3) reveals considerably larger differences with adults being, on average, approximately three times more likely to have access to unemployment benefits.  Complete data on this indicator was available for only 12 countries and differences between youth and adults were large in all countries except the UK; and most so in Sweden, Finland, Poland and Spain. These are all countries with above average shares in temporary employment. The examples of the UK and Germany show that with regard to youth coverage universal basic-benefit schemes work relatively well. However, the benefits payable under these schemes are usually low.
Comparison of registered short-term unemployed receiving insurance or assistance by age, 2010 (total=100)
Comparison of registered short-term unemployed receiving insurance or assistance by age, 2010 (total=100)Data source: Eurostat, online database
Note: Duration of unemployment 1-2 months. Countries are sorted by difference between youth and adult coverage rates in ascending order.
25Do we see any common trends as regards the developments of unemployment benefit coverage during the first period of the crisis? The modifications of the unemployment benefit systems in several countries should, in most countries, have positive implications for coverage. However, the characteristics of the newly unemployed  will also play a role and these are likely to have changed during the crisis. On the EU average, the first two crisis years, and particularly 2009, saw slight improvements in unemployment benefit coverage, whereas the level dropped back again in 2010 (Figure 4). The left-hand panel shows the countries with decreasing or stable benefit coverage between 2007 and 2010. Benefit coverage did not decline in linear fashion in these countries; in fact, in most cases, a higher than pre-crisis benefit coverage was recorded in 2008 and/or 2009 with coverage dropping again subsequently. The right-hand panel, shows the countries with increasing coverage comparing 2007 and 2010. In some of these countries initial drops in coverage particularly between 2007 and 2008 were observed. The countries with the strongest upward trend are Portugal, the UK, Spain and Italy. Not surprisingly, all countries that have relaxed the eligibility criteria to unemployment benefit schemes (Finland, France, Portugal, Slovenia, Slovakia) or introduced special payments to unemployed (e.g. Spain, Italy and Greece) show upward trends in coverage during the crisis. Upward trends in coverage of short-term unemployed in the remaining countries are likely to be induced by changing characteristics of the pool of unemployed, particularly increasing male unemployment.
Developments in short-term registered unemployed receiving benefit or assistance, total 2007=100
Developments in short-term registered unemployed receiving benefit or assistance, total 2007=100Data source: Eurostat, online database.
Note: Age 15-64 years; duration of unemployment 1-2 months. Countries are sorted by the direction and magnitude of change between 2007 and 2010. The UK figures refer to 2011/2007 as 2010 data is not available.
26The EU27 male average coverage rate increased slightly in the initial crisis phase but was back to its 2007 level in 2010 (Figure 5). Coverage for women has improved very slightly after an initial drop (Figure 6). Of the 19 countries with available data between 2007 and 2010, eleven countries displayed improving coverage for women and thirteen for men, with Belgium keeping coverage rates stable for men. Nine countries saw a drop in benefit coverage for women and five for men. The largest improvements for women were evident in Spain, the UK and Italy. Men saw the largest increases in Italy, followed by Spain and Portugal. Italy is well-known for its comparatively low unemployment benefit coverage (Jessoula and Vesan 2011). Greece and Denmark are interesting cases, insofar as men have seen their coverage substantially increase whereas coverage of women decreased during the crisis. Denmark is one of the few countries that had somewhat higher benefit coverage of women before the crisis. Hardly surprisingly, all countries (with available data) that improved the benefit system with regard to eligibility or duration saw positive outcomes for both men and women.
Developments in short-term registered unemployed men in receipt of benefit or assistance, 2007=100
Developments in short-term registered unemployed men in receipt of benefit or assistance, 2007=100Data source: Eurostat, online database.
Note: age 15-64 years; duration of unemployment 1-2 months. The UK figures refer to 2011/2007 as 2010 data is not available.
Developments in short-term registered unemployed women in receipt of benefit or assistance, 2007=100
Developments in short-term registered unemployed women in receipt of benefit or assistance, 2007=100Data source: Eurostat, online database.
Note: age 15-64 years; duration of unemployment 1-2 months. The UK figures refer to 2011/2007 as 2010 data is not available.
27Moving to the trends in unemployment benefit coverage for youth, on the EU27 average a substantial drop of close to 20 per cent from an already low level was recorded between 2007 and 2010, whereas the first crisis period had seen a short upsurge (Figure 7). This development has to been seen also in the context of youth being disproportionately affected by unemployment as well as by part-time and temporary employment during the crisis. Out of nine countries with available data, the Czech Republic saw the most drastic decline, followed by Poland and Austria. In France, on the other hand, the coverage of youth improved substantially, and Spain, Finland and the UK (the former two countries with large differences between coverage of youth and adults – Figure 3) also saw coverage increase markedly. All these three countries had deliberately improved benefit eligibility during the crisis. Crisis developments notwithstanding, countries still differ markedly in their ability to provide unemployment benefits to the general population and more specifically to youth. The case of Spain is particularly interesting. Here, both young and temporary workers were disproportionally affected by the crisis; at the same time benefit coverage – traditionally differing strongly between permanent and temporary workers (Leschke 2008) – increased. Overall unemployment benefit coverage of youth remains, however, comparatively low (Mato 2011).
Development in youth (15-24) registered as unemployed and in receipt of benefits/assistance, 2007=100
Development in youth (15-24) registered as unemployed and in receipt of benefits/assistance, 2007=100Data source: Eurostat, online database
Note: duration of unemployment 1-2 months.
28Overall, we can conclude that there are no clear country patterns with regard to improved access to unemployment benefits during the crisis. Improvements in coverage rates were not confined to the countries which deliberately relaxed qualifying conditions or increased the duration of benefits, as the changing characteristics of the unemployed (such as more male standard workers in the pool of unemployed) also impacted upon the coverage rate. Moreover, only a few countries with available data for all groups showed improvements for all the labour market groups examined here, clear examples being Finland, France, Spain and the UK – the latter one with some data limitations though. The Czech Republic, but also Germany and Austria showed decreasing coverage for all groups. When judging the results, it has to be kept in mind, however, that overall coverage rates vary strongly between countries, albeit with no good comparative data available. Table 2 summarises the findings of the above analysis.
Increased unemployment benefit coverage in 2010 as compared to 2007 – wrapping up the findings of the above analysis
Increased unemployment benefit coverage in 2010 as compared to 2007 – wrapping up the findings of the above analysisNote: data for the UK refer to 2007 and 2011, as 2009 and 2010 data are not available.
3 – Conclusions
29The aim of this paper was to assess whether the crisis period can be seen as perpetuating or even exacerbating the labour market and welfare segmentation, or whether it may actually have contributed to containing some of the duality that has developed in recent decades.
30Segmentation exists in both the labour market and the welfare system and the two spheres reinforce one another. Due to the fact that the design of unemployment benefit systems (and particularly unemployment insurance) is still, in most countries, strongly geared to “standard employment”, non-standard workers are disadvantaged in access to unemployment benefits. Comparative data on unemployment benefit coverage is not available for non-standard workers, we therefore had to use coverage rates for women and men and youth and adults to proxy access of non-standard workers. Both women and youth are disproportionally affected by part-time work and temporary employment, respectively, and it has been shown that, in the majority of countries (for which data is available), coverage of women is lower than of men, and that youth are at a particular disadvantage in the receipt of unemployment benefit.
31The crisis had a direct impact on temporary employment in that in most countries, during the first phase of the crisis, temporary workers were the first to lose their jobs, whereas in the second phase of the crisis new jobs were often created on a temporary basis. Part-time employment, on the other hand, continued to grow throughout the crisis. Individuals with low educational levels and young workers were particularly affected by unemployment, but also by subsequent increases in temporary employment and the still ongoing increase in part-time employment. Moreover, the countries disproportionately affected by the crisis not only saw unemployment rising but also, for the most part, saw further growth in part-time employment. Trends in temporary employment were more diversified in the countries particularly hardhit by the crisis, with some using it as external adaptation strategy (particularly Spain) and others (particularly Ireland) seeing their shares of temporary employment in total employment increase in spite of growing unemployment. Importantly, all countries disproportionately affected by unemployment saw involuntary forms of non-standard employment increasing – often by large margins, which points to the fact that in times of harsh labour market conditions the unemployed are forced to take up sub-standard employment. With regard to labour market developments, we have thus definitely witnessed further segmentation during the crisis. This is true both in relation to specific labour market groups, particularly youth and the low educated, and also with regard to specific countries, largely those disproportionately affected by unemployment.
32As regards the welfare system, varying trends in unemployment benefit coverage were evident during the crisis. In the light of the high degree to which specific labour market groups, such as temporary workers and among them youth, are affected by unemployment, and the fact that precisely these groups are usually less likely to qualify for unemployment benefits, a number of countries – such as Finland, France, Portugal, Latvia and Slovenia – deliberately relaxed (in many cases, but not all, permanently) the qualifying conditions leading to improvements in benefit coverage for these groups. Other countries saw unemployment benefit coverage decline during the crisis. As the available data can act only as a proxy (underreporting; no data on benefit coverage of part-time and temporary workers; no separate data on unemployment insurance and unemployment assistance) and information is either partly or completely missing for several European countries, it is difficult to judge whether the crisis has contributed to more or less segmentation in welfare coverage. There seems, in fact, to be a division between countries in this regard and the deliberate opening up of unemployment schemes in several countries to new groups of workers during the crisis can clearly be seen as a positive trend that contrasts with developments over the last decades. Also trends in coverage rates are often not linear within countries which emphasises the fact that the results are to some degree also driven by the changing pool of unemployed person. However, in the vast majority of countries, women and particularly youth remain in an inferior position with regard to benefit coverage and, on average, the coverage rate of youth has actually declined during the crisis. Last but not least, it is important to point out that, beyond the disadvantages in coverage for specific labour market groups, countries differ strongly in their overall coverage rates and generally also in terms of the generosity of unemployment benefits, as is evident from the expenditure figures. A considerable problem for cross-country research on this issue is that we lack reliable comparative data on absolute unemployment benefit coverage.
33In relation to the Europe 2020 strategy, it is interesting to note that social security coverage of nonstandard workers has now been explicitly placed on the agenda. Previously, during the Lisbon period, i.e. between 2000 and 2010, attention was focused not on social security but rather on the notion of employment security or employability. The economic crisis has further emphasised the coverage gaps and several institutions (including the OECD and the European Commission) have raised the question, particularly in relation to the crisis developments, of whether more encompassing unemployment benefit coverage should be achieved rather through first-tier schemes (with a stronger link to the labour market) or through second-tier schemes such as unemployment or social assistance. Generally, it remains to be seen whether the recent trends will have a longer-term influence on better inclusion of non-standard workers in unemployment benefit schemes. The recent austerity measures that in several countries placed a strong focus on cuts in the fields of employment and social policies, but also the fact that some countries improved access to unemployment benefits for specific categories of workers on a temporary basis only, raise doubts in this regard. In response to the labour market situation of non-standard workers, several European Commission documents have taken up the idea of a “single open-ended contract” (with no ex-ante time limit but providing a long entry phase with employment protection rights gradually increasing) that has been proposed for countries such as Spain, France and Italy in order to decrease labour market segmentation.
Associate professor at the Department of Business and Politics, Copenhagen Business School
For a detailed analysis of developments in (involuntary) part-time and temporary employment compare Leschke 2012.
Two studies on coverage were, for example, commissioned by the European Parliament dealing with how to create indicators on coverage of social security systems for people in flexible employment (Alphametrics 2005 and 2009).
See explanatory text under guideline 7 “Increasing labour market participation of women and men, reducing structural unemployment and promoting job quality”.
Short-time working schemes and the question in how far they are accessible to non-standard workers are not discussed in this article. For further information see Leschke 2012 and Hijzen, A. and D. Venn (2011).
Segmentation is used here to depict the separation of the labour market and welfare system in distinctive segments that are characterised by different degrees of e.g. job, earnings and income security. The term is used interchangeably with concepts such as dualisation or fragmentation.
An in-depth study on access of part-time and temporary workers to unemployment benefit schemes in Denmark, Germany, Spain and the UK using the European Community Household panel data (1993-2001) and institutional information up to 2006 is available in Leschke (2008).
There is no information on Romania, Bulgaria, Latvia, Lithuania, Cyprus and Malta as they are not members of the OECD. As the MISSOC comparative tables are not sufficiently detailed to complement the missing information for these countries, based on that source these countries will not be taken into account here.
Qualifying period in hours or days: Belgium, France, Greece, Hungary, Poland, Portugal.
Additional qualifying conditions based on previous hours or earnings: Austria, Belgium, Denmark, Germany, Finland, Poland, Sweden, UK.
Comparatively short qualifying period: Austria, Belgium, Germany, Italy, Luxembourg, Portugal, Sweden and – with particularly short qualifying periods – Ireland, (the Netherlands), Poland, Slovak Republic and Slovenia. The UK system combines the level of earnings and contribution periods, thereby placing at a disadvantage both individuals with low earnings and those with short tenure (for detailed information compare Leschke 2008).
Duration conditional on employment or contribution record: Austria, Estonia, France, Germany, Greece, Hungary, Ireland, Netherlands, Poland, Portugal, Slovenia, Spain.
Net replacement rates on unemployment for different wage levels and family types are available at the OECD Benefits and Wages homepage.
A notable exception is the reform of the Dutch unemployment benefit system, in the late 1980s and early 1990s, in reaction to increasing non-standard employment and particularly part-time work (Koopmans et al. 2005, 365-367).
OECD countries such as the US, Canada, Japan and Switzerland have also seen changes in their unemployment benefit systems (for more encompassing information see OECD 2010a).
Changes with regard to contributions that were evident in a number of countries are not reviewed here as they do not usually have a direct impact on the coverage of non-standard workers. They can, however, have an indirect impact if they set incentives to hire individuals on standard rather than non-standard contracts (see e.g. Spain where in the past the government has tried to set incentives to hire individuals on the basis of regular contracts by reducing contributions in such cases).
However, this measure was due to run for one year only, starting in February 2009 (OECD 2009c: 188).
Unemployment assistance in Spain is usually restricted to specific labour market groups such as unemployed persons with family responsibilities and older workers. The special benefit was removed in February 2011; it had covered about 700,000 unemployed people (Sanz de Miguel 2.3.2011).
Denmark formerly had a comparatively long universal duration of unemployment insurance benefits of four years; in 2010 this was reduced to two years.
The purpose of this information is to compare the ILO-based unemployment to registered unemployment.
Differences in the wording of survey questions across countries plays a crucial role here, for details see Immervoll et al. 2004: 59-60.
In principle, timing of assistance schemes and average duration of unemployment is also of relevance; however, these factors are left out here as we are looking at short-term unemployment of only one or two months.
OECD (2009c: 164) gives details on the countries where young jobseekers that have never worked can access to unemployment benefits. This is the case in Germany, Finland, Denmark Sweden, Ireland, UK, Belgium, Luxembourg and Greece (and Australia and New Zealand); however, there are in most cases strict conditions and mutual obligations.
The implications of the changing characteristics of the unemployed are not clear, as the effect could go in both directions. This can be illustrated by a simple example: more men than women have lost their jobs during the first years of the crisis and men are on average more likely to have access to benefits then women. On the other hand, temporary workers (with shorter tenure and therefore on average lower access than permanent workers) were also more likely to have lost their jobs.