%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
IntroductionFootnote 1
Unlike other branches of social insurance (industrial accident, pensions, health) the nature of the risk of unemployment hampered the earlier and more comprehensive introduction of public unemployment insurance. At the end of the nineteenth and long into the twentieth century, political elites and employers regarded more than a meagre means-tested public support for jobless workers as potentially raising reservation wages and thus distorting the supply of labour. Trade unions, for their part, offered voluntary unemployment cover and benefits to unemployed members as a way of resisting downward pressure on wages during times of recession. Maintaining unemployment funds allowed them some control over the level of wages, and the types of jobs unemployed benefit recipients were deemed to accept or permitted to turn down without loss of entitlement. Moreover, the prospect of unemployment benefit was a selective incentive for employees to join a trade union, and for trade unions to retain members during times of economic downturn (Rothstein, Reference Rothstein and Ashford1992; Scruggs and Lange, Reference Scruggs, Lange and Bormeo2001).
However, the disadvantages and limitations of purely trade union-run systems became increasingly visible in the 1920s. Membership was often affordable for better-paid skilled workers only, leaving a majority of unskilled labourers without unemployment cover. Mass and prolonged unemployment during the interwar years swelled the number of benefit claimants and quickly drained union funds. Hence, demands grew for a stronger state involvement in the provision, or at least the funding, of unemployment insurance. The models adopted during the first two decades of the last century consisted of either the so-called Ghent system (that is, public subsidies to voluntary trade union-run systems), named after the Belgian town in which it was first introduced in 1901, or compulsory state schemes. In the 1920s, the former was the more prevalent model in Europe, adopted in Belgium, the Netherlands, France, Switzerland and the Nordic countries. However, after World War II most of these countries replaced the Ghent system with compulsory unemployment insurance, which had already been introduced in Britain, Austria, Italy and Germany. In contemporary mature welfare states, membership in unemployment insurance is mandatory for the large majority of employees, and national schemes are run solely by the state, as in the UK, or in cooperation with social partners. Voluntary unemployment insurance administered by trade union-linked funds survived only in Denmark, Sweden and Finland.Footnote 2
This is not the place to discuss reasons for the shift from the Ghent model to mandatory unemployment insurance (on this see Viebrock, Reference Viebrock2004). Suffice to note that even in the three countries which retained the Ghent model, some characteristics of voluntary insurance have considerably changed. Most importantly, modern unemployment insurance funds are no longer run by trade unions but by agencies linked to but separate from trade unions, and funds are supervised by the state. Moreover, the requirement of a dual membership was abolished, enabling prospective members to join an unemployment insurance system without having to join a trade union at the same time.
In spite of initial anxieties on the part of national labour movements, this weakening of control over unemployment insurance has not led to a decline of trade union membership in the three Ghent countries. Instead, trade union density rates in Finland, Sweden and Denmark are distinctively high (see Table 1). Moreover, apart from a slight decline in Denmark in recent years, the Ghent countries have withstood the general trend of a diminishing relevance of trade unions elsewhere in Europe (see also Kjellberg, Reference Kjellberg2006). The figures for Norway in Table 1 suggest that high trade union membership is not simply a general trait of Nordic welfare states. Instead, the table underlines the apparently positive impact of voluntary unemployment insurance on the organisational strength of trade unions. It is thus unsurprising that social democratic parties in Nordic countries have supported this model, while centre-right and liberal governments have repeatedly tried to restructure or even replace it with compulsory public unemployment insurance.
Union density rates (adjusted data) for selected countries
Source: Visser (Reference Visser2006: 45).
In comparative social policy and industrial relations research, the link between unemployment insurance and cross-national differences in trade union density has been a topic of analysis for some time (for example, Neumann et al., Reference Neumann, Pedersen and Westergard-Nielsen1991; Rothstein, Reference Rothstein and Ashford1992; Western, Reference Western1993; Leonardi, Reference Leonardi2006). And yet, beyond general references, there is a lack of systematic investigation into the ways in which formally independent unemployment insurance systems might create recruitment effects for trade unions, and even fewer comparative analyses that suggest possible country-specific factors. Based on qualitative research and expert interviews, this article addresses this issue with reference to the operation of unemployment insurance in Denmark and Sweden. The following section covers relevant theoretical aspects and reviews the evidence of the link between voluntary insurance and trade union membership. Subsequently we contrast major institutional features (organisation, funding, benefit provision) in the two countries under investigation. We then identify four common and country-specific factors that make a dual membership (unemployment fund and trade union) appealing. The concluding part reflects on possible trajectories of voluntary unemployment insurance systems in the light of policy debates in Denmark and legislative changes introduced by the new centre-right government in Sweden in early 2007.
Voluntary unemployment insurance and trade union membership
In line with Olson's (Reference Olson1965) theory of selective incentives, it is no surprise that countries in which unemployment insurance is voluntary and run by trade unions have higher levels of trade union membership than countries with a compulsory public system (see Ebbinghaus and Visser, Reference Ebbinghaus and Visser1999, Reference Ebbinghaus and Visser2000). The absence of a compulsory state scheme and the availability of voluntary unemployment insurance organised by trade unions offers selective incentives, which help to overcome the ‘free-rider’ problem. Unlike wage bargaining, which works to the advantage of all workers within a particular industry or region, unemployment insurance restricts benefits to members only (Olson, Reference Olson1965). In addition, it helps to maintain links between trade unions and workers who are temporarily out of work. Of course, there are likely other reasons for the high and stable level of trade union representation in Nordic countries, such as union access rights in the workplace, connections with strong social democratic parties or labour market centralisation. However, the Ghent system of unemployment insurance is generally assumed to have played at least an important contributory role for the organisational strength of Danish and Swedish trade unions in the post-war period (see, for example, Rothstein, Reference Rothstein and Ashford1992).
Several quantitative comparative studies have provided empirical support for this assumption. Employing a range of statistical tests, Western (Reference Western1993, Reference Western1997) found that the existence of the Ghent system had a positive effect on trade union density during the period 1950 to 1985. The impact was particularly pronounced at times of rising unemployment. In countries with a compulsory public system, unemployed persons tended to terminate or suspend their trade union membership. In contrast, both unemployed and employed persons in Ghent countries remained members of, or increasingly joined, voluntary unemployment systems at times of economic downturn (see also Scruggs, Reference Scruggs2002). Job starters, too, were more willing to join a trade union in countries with voluntary unemployment insurance than in countries with a compulsory public system.
The case of Norway, as a country in which trade union membership is high by international standards but significantly below the rates found in the other three Nordic countries (see Table 1), might be regarded as an indication of the relative impact of the Ghent system (Albrechtsen, Reference Albrechtsen2004). Similarly, trade union density in Belgium, a country with compulsory unemployment insurance but a strong trade union role influence in this domain, is higher than in the Netherlands, even though the two countries are similar in many other relevant political and sociological aspects (Crouch, Reference Crouch, Manning and Shaw2000). Supported by recent time series analyses, Scruggs (Reference Scruggs2002) found a clear positive impact of the Ghent system on membership trends in the past three decades. Other authors, however, have raised doubts as to whether differences in trade union membership can solely be attributable to the presence or absence of the Ghent system (Shalev, Reference Shalev2007). Comparative investigations into historical developments of trade unionism in the post-World War II era point to high employment in Ghent countries as an important factor, as well as the strength of local trade union representation and their presence at the workplace (Hancké, Reference Hancké1993). Moreover, a theoretical puzzle remains. While it is plausible that during times of economic uncertainty, the membership of voluntary unemployment insurance funds increases, it is less clear why trade union membership rates rise at the same time. After all, unemployment insurance funds have been formally independent from trade unions for several decades. In addition, legal barriers as well as financial disincentives, which previously disadvantaged employees who joined unemployment insurance funds without becoming a trade union member, have long been removed or significantly weakened (see below).
The above studies offer some speculation on this point. Scruggs (Reference Scruggs2002) suggests that despite formal independence, trade unions might influence jobs offered by insurance funds to individual unemployed persons, or the type of jobs that are deemed suitable. Thus, unemployment insurance funds might discriminate in favour of union members (or might at least be perceived that they do so), not necessarily by altering rules of eligibility, but in the day-to-day administration of the programme. If there is a higher probability of being declared eligible for benefits, or of being offered more desirable employment, individuals might be inclined to join a trade union in order to be in the ‘insider’ job queue (Scruggs, Reference Scruggs2002: 292). A dual membership might thus be preferable to the option of only joining an unemployment insurance fund, especially at times of high unemployment. However, the mainly quantitative literature does not allow inferences about the validity of such assumptions, or about the mechanisms that might lead to actual or perceived preferential treatment in favour of job seekers who are also trade union members. The remainder of this article aims to identify such mechanisms.
Comparing voluntary unemployment insurance in Denmark and Sweden
One reason for modern welfare states to run mandatory rather than voluntary unemployment insurance is the problem of adverse selection. The extension of membership to employees who have a relatively low risk of unemployment is not only a question of social solidarity but financial viability. Voluntary unemployment insurance funds face the problem of ‘good risks’ declining to join, thus requiring membership fees to rise, which, in turn, decreases financial incentives to enter the scheme even more. These theoretical considerations suggest higher costs and/or lower levels of effectiveness of voluntary versus mandatory unemployment insurance. Empirically this is unconfirmed, as Table 2 demonstrates. In 2005, both the coverage ratio (share of the workforce included in unemployment insurance) as well as recipient ratio (share of unemployed in receipt of unemployment insurance benefits) were considerably higher in Denmark and Sweden than in Germany or the UK, two countries with compulsory unemployment insurance schemes. Differences in the latter are only partially explained with shorter maximum entitlement periods in the non-Ghent countries. If only claimants with a spell of unemployment below three months are considered, the ratio of insurance benefit recipients was higher in Germany (around 60 per cent) and in the UK (about 30 per cent), but still well below the rates in Denmark and Sweden. Other aspects, such as eligibility criteria, are further explanatory factors here. What is more, voluntary unemployment insurance can be at least as generous as compulsory systems and at the same time less costly as far as direct costs of individual insured persons are concerned. Because of the lack of earmarked fees and the existence of a general contribution to the National Insurance Fund, calculating such cost is difficult for the UK. However, direct unemployment insurance contributions in 2005 were certainly higher in Germany than average membership fees in either Danish or Swedish voluntary unemployment insurance funds, implying heavy tax subsidisation (see below).
Unemployment insurance systems compared, 2005
Notes: all data for 2005 unless otherwise stated; (1) for Denmark and Sweden: membership contribution plus average administration fee; Germany: 3.25% of an average gross wage paid in industrial employment, (Statistisches Bundesamt, 2005); UK: total National Insurance contribution for somebody earning an average weekly full-time salary in the private sector (figure in bracket represents contribution needed to fund expenditure on contributory JSA only); own calculations based on DWP (2006) and HM Treasury (2006); (2) for single person, no children, initial phase of unemployment, with previous wage of ‘average production worker’ in 2004 (OECD, 2006); (3) including ‘activation periods’ in Demark and Sweden; Germany: 18 months for claimants over 55; (4) portion of the labour force covered by unemployment insurance, 2002, Scruggs (Reference Scruggs, Clasen and Siegel2007); (5) recipients of unemployment insurance as share of all unemployed; for Germany: own calculation based on Statistisches Bundesamt (2005); for the UK: recipients of contributory JSA (with and without other benefits) as a share of all unemployed benefit claimants, February 2005 (DWP, 2006).
In sum, from a macro perspective, the Danish and Swedish examples illustrate that voluntary unemployment insurance can be more effective in social policy terms than mandatory systems. The problem of adverse risk selection is avoided by relatively generous benefits and low direct costs, both of which make membership in voluntary funds attractive. However, the above does not explain the individual preference for the, more expensive, option of entry in both the unemployment insurance fund and the (linked) trade union. Elucidating the reasons for this connection requires a closer look into major characteristics and the operation of contemporary Danish and Swedish systems of unemployment insurance. Referring to the situation in 2005, Table 3 provides a brief overview and orientation for the subsequent discussion. The table also highlights significant changes, which were introduced by the Swedish government with effect from January 2007.
Major characteristics of voluntary unemployment insurance systems in Denmark and Sweden, 2005 and relevant changes introduced in January 2007 in Sweden [in brackets and italics]
Administration and membership
First introduced in 1907, the Danish Ghent system of voluntary unemployment insurance is currently administered by 32 separate funds (a-kasser), all of which are governed by legislation, which stipulates common criteria for benefit eligibility and entitlement, as well as job search requirements, suitability of job offers and jobseekers' availability for employment. The control over a-kasser is exercised by the Arbejdsdirektoratet, a section within the Ministry of Labour. A major reform in 1969 circumscribed the remit of the sector-specific unemployment insurance funds, formally transferring their previous role in job placement activities to newly created public employment offices (Arbejdsformidlingen; AF) operating across all industries. Moreover, the option of charging up to 40 per cent higher membership fees from members who were not affiliated to the relevant trade union was abolished (Nørgaard, Reference Nørgaard1997: 333). The minimum number of members required to gain recognition by the state was increased (initially to 1,000, later to 5,000 and eventually to 10,000; Nørgaard, Reference Nørgaard1997: 340; Lind, Reference Lind, Waddington and Hoffmann2001: 117). In return, eligibility conditions were relaxed, benefit generosity improved, and the level of tax subsidies rose significantly. During the 1970s, the coverage of unemployment insurance was widened to part-time workers as well as self-employed people.
Apart from two funds for self-employed persons, the structure of unemployment insurance reflects the Danish trade union structure, which differentiates between blue- and white-collar employees, as well as educational and vocational grades achieved. A majority of the numerous trade unions are affiliated to one of the three umbrella organisations: the Landsorganisationen (LO) (for blue- and white-collar workers in the private and public sector), the Akademikernes Centralorganisation (AC) (professional association for employees with a university degree) or the Funktionærernes og Tjenestemændenes Fællesråd (FTF) (for white-collar workers in the public sector). Traditionally linked to the Danish social-democratic party, the LO is the largest organisation with more than half of all trade union members affiliated (Lind, Reference Lind, Waddington and Hoffmann2001). Employees with a particular occupational or educational background tend to join the unemployment insurance fund linked to the corresponding trade union. In principle, it has long been possible to join a fund that operates across sectors. However, exercising this choice has, until recently, been limited to the Christian unemployment insurance fund, which has traditionally recruited members who do not identify themselves with the trade union movement. Twenty years ago the membership of this fund consisted of merely 30,000 persons, but has since grown to almost 160,000, equivalent to 7 per cent of all insured persons, thereby making it the fourth largest single fund in Denmark (Arbejdsdirektoratet, 2005). After a legislative change in 2002, another three funds began to make use of the newly created option of recruiting members irrespective of their particular employment or educational background. In contrast to the majority of a-kasser, these three funds are actively engaged in publicising their services, aiming to recruit new members as well as tempting members from other funds with the prospect of lower membership fees (see below).
Since 2002, the membership of these competition oriented funds has increased. However, including the Christian fund, these four ‘open’ funds cover only about 15 per cent of the total membership in Denmark, despite membership fees that are about four times lower than those of the most expensive funds (Arbejdsdirektoratet, 2005). There is little indication that many more funds will decide to ‘open up’ and thereby breach the informal consensus among the large majority of funds that have ignored the option of recruiting members outside of their traditional industrial or occupational remit, thereby respecting and safeguarding traditional trade union demarcations. In short, the legislation aimed at stimulating more competition between funds has had only modest success.
In Sweden a major reform in 1974 supplemented the existing voluntary unemployment insurance system with a general tax-funded unemployment assistance (kontanta arbetsmarknadsstöd, KAS) for those who were either not members of one of the trade union-affiliated unemployment funds or did not fulfil the entitlement conditions. Benefits, however, were much lower than offered by unemployment insurance, and the entitlement period was relatively short. Usually it was groups at the margins of the labour market who only received this lower unemployment assistance: that is, those typically less unionised, such as women and young workers, than the male skilled workers (Björklund and Holmlund, Reference Björklund, Holmlund, Björklund, Haveman, Hollister and Holmlund1991: 123). As part of a far-reaching reform in 1997, this scheme was eventually abolished and replaced with a basic income security allowance, which covers all workers who have been in paid employment for a certain minimum period. The level of this basic allowance is considerably lower than unemployment insurance benefit.
The Swedish voluntary unemployment insurance system (officially termed earnings-replacement insurance) is run by 36 funds, which, similar to their Danish counterparts, have a private status. As in Denmark, legislation requires funds to have at least 10,000 members and the funds need to be officially registered with the National Labour Market Board (Arbetsmarknadsstyrelsen, AMS). The AMS stipulates uniform regulations (eligibility criteria, entitlement, benefit levels and so on) and, until the end of 2003, was responsible for supervising the unemployment funds, a task which has since been transferred to the IAF (Unemployment Insurance Board). While the funds establish whether claimants have fulfilled eligibility conditions and distribute insurance benefits, it is the IAF that determines the continuous benefit status on the part of jobseekers, such as regular renewal procedures.
Although formally independent, funds cooperate closely with particular trade unions and thus cover the same industrial sector as the relevant trade union. As in Denmark, trade unions are affiliated to one of the three confederations LO (blue-collar), TCO (white-collar) or SACO (academic). The LO has the closest links to the Social Democratic Party, whereas the other two confederations are politically more independent (Kjellberg, Reference Kjellberg2006). The occupational area or type of employment determines which fund can be joined. In the case of a job change, it is possible to change funds and existing entitlements are transferred. Membership in a fund is open to all employees, as well as to employers and, under certain conditions, until 2006 to students (see Table 3).
Although it has long been possible to be a member of a fund without joining the respective union, in practice almost all insured persons are also trade union members. However, in 1998 the formal separation between trade unions and unemployment insurance became more strongly accentuated with the introduction of the independent ‘Alfa’ fund (Alfa-Kassan). Unlike the trade union-affiliated unemployment insurance funds, its membership is open to the workforce across all sectors and occupations. However, its alternative status might be seen as being undermined by the fact that it is administered by the federation of the trade union-affiliated unemployment insurance funds. Although membership in the ‘Alfa’-fund has grown considerably in recent years (from 66,000 persons in December 2004 to 86,000 two years later: IAF, 2007), it has remained small in the context of a total of 3.8 million members of voluntary unemployment insurance funds in Sweden. What is more, its fee is somewhat higher than in most other funds, due to its higher administrative costs per unemployed person (IAF, 2005). In turn, this is linked to the fact that a large proportion of the members of the ‘Alfa’ fund are eligible only for the basic security benefit. In 2004, the ‘Alfa’ fund dealt with 76 per cent of the total of all entitlement days to this basic component within unemployment insurance.
Similar to the situation in Denmark, the competitive element of Swedish unemployment insurance is potentially to become stronger, as a result of significantly higher differences in membership fees introduced in 2007 (see below). Trade union members can remain in their old union but become members of cheaper alternative funds as long as they fulfil the occupational criteria for these funds. This is especially important for TCO-affiliated unions, which often have rivals in SACO-affiliated unions. We might thus expect changes both in trade union membership as a whole and in the relative strength of SACO and TCO in particular.
Funding and benefits
Unemployment insurance in both countries is funded by tax subsidies and individual membership fees, which can be set against personal tax. Disregarding a small annual duty to be paid for a public second pension (of DKK 144, or €19), in 2005 Danish membership in voluntary unemployment insurance comprised of a uniform annual insurance fee of DKK 3144 (€420) and an administrative fee, which ranged from DKK 624 (€86) in the least expensive to DKK 2748 (€368) in the most expensive fund. These differences in administrative fees reflect different costs, which are determined by the distribution of the risk of unemployment across funds and the structure of unemployment within the membership of a particular fund. Total membership is another factor since it influences the number of regional and local branches of a particular fund.
As a form of revenue, membership fees are less relevant than tax subsidies stemming from both general taxation and a hypothecated labour market tax of 8 per cent of gross earnings levied on all economically active persons. Moreover, while membership fees vary little from year to year, the share of expenditure covered by tax subsidies reflects changes in the level of unemployment since the latter determines total benefit expenditure. However, even at times of low unemployment, the bulk of the funding for unemployment insurance is tax financed and direct personal contributions are relatively low, which implies a strong individual incentive to join an unemployment insurance fund. Benefit generosity enhances this since transfers nominally replace 90 per cent of previous earnings. However, benefit income is taxed and a ceiling applies that means the actual replacement rate for all but those on previously very low earnings is well below 90 per cent (OECD, 2004; see also Table 2). Moreover, although nominally linked to developments in average earnings, for several years the benefit ceiling has not been fully uprated, thereby widening the gap between average earnings and the maximum benefit obtainable. While in 1994 the maximum unemployment benefit was equivalent to 66 per cent of average wages, by 2001 it had declined to 63 per cent (Hansen, Reference Hansen2003: 12).
However, there is a second type of income maintenance, which makes joining a Danish unemployment insurance fund highly appealing. In Denmark a-kasser administer not only unemployment insurance benefits but also a voluntary early retirement allowance (efterløn), which was originally introduced in 1979 in order to address mass unemployment with improved options for labour market exit. Unemployment insurance funds were made responsible for running efterløn, and eligibility became thus linked to membership in a fund. Moreover, initially there were no additional charges over and above the usual membership fee.
The importance of efterløn for unemployment insurance funds can be illustrated with a regulatory change introduced in 1992, which doubled the number of insurance years required for accessing the early retirement scheme. It also stipulated that, at the time of the prospective entry into early retirement, a minimum fund membership of 20 years (now 25 years) within the previous 25 (now 30) years would be required. In 1992, this rule applied particularly to employees below and around 45 years of age. Intent on safeguarding the chance of receiving the early retirement allowance, it was this age group that disproportionately joined unemployment insurance funds prior to the new law coming into effect (Parsons et al., Reference Parsons, Tranæs and Lilleør2003: 38).
Despite having made access somewhat more difficult in the early 1990s, efterløn maintained its role as a major route into retirement in Denmark. At any one point during the 1990s, about two-thirds of all members in unemployment insurance funds aged between 60 and 66 were in receipt of the allowance (Parsons et al., Reference Parsons, Tranæs and Lilleør2003), and today the fiscal scope of efterløn is larger than unemployment insurance. In 2004, insurance funds paid a total of about DKK 22.8 billion in unemployment benefit, but DKK 25 billion on efterløn (Politiken, 11 November 2005). This is remarkable since several changes introduced in 1999 were aimed at making the scheme less attractive. Since then, moreover, separate contributions have been payable, making it possible for members of unemployment insurance funds to choose only the less expensive option of unemployment insurance cover, rather than the package that includes efterløn. Recent data suggest that the reform has had some success, given that the ratio of younger members in a-kasser who select also to pay efterløn contributions has significantly declined (Bredgaard and Larsen, Reference Bredgaard, Larsen, Bredgaard and Larsen2005: 320).
There is no equivalent to unemployment funds running a similar system in Sweden. However, from the perspective of personal cost, the financial incentive to join a Swedish unemployment fund was, until recently, even greater than in Denmark, since membership fees were comprised of merely a small state levy and an administrative contribution which in 2005 was just €11 per month on average, and thus well below average contributions payable in Denmark (see Table 2). Moreover, until recently membership fees in Sweden varied much less across funds than in Denmark.Footnote 3 The low level of membership fees imply that, with 90–95 per cent of total expenditure, the Swedish system before 2007 was even more tax funded than the Danish system (Engblom, Reference Engblom2005; Lundgren, Reference Lundgren2006). From an individual perspective, a simple cost–benefit calculation made it thus highly rational for Swedish employees to join an unemployment insurance fund (Hatland, Reference Hatland1998). The financial savings from non-membership are small and it can be assumed that only individuals with a very low risk of unemployment, such as tenured civil servants, decide against joining an unemployment insurance fund. It is thus no surprise that the coverage is even slightly higher than in Denmark (see Table 2).
However, with the intention of weakening the attractiveness of voluntary trade union unemployment insurance, a number of significant changes have been introduced by the new centre-right coalition, which won the Swedish election in September 2006.Footnote 4 For example, conditions of eligibility have been tightened and benefit levels lowered (see Table 3). The most important change, though, is the steep increase in membership fees, shifting total funding heavily away from taxation. Whereas in 2004 contributions represented 9.4 per cent of unemployment benefit outlays (Lundgren, Reference Lundgren2006: 2), for 2007 membership fees are forecast, covering 46 per cent of total costs.Footnote 5 This has been achieved by raising the average monthly unemployment insurance contribution more than threefold from about SEK 100 (€11) to SEK 352 (€38). Another novelty is that unemployment insurance contributions varied more strongly in line with different unemployment rates of members in a particular fund. Whereas in 2005 insurance contributions ranged from SEK 84 (€9) in the cheapest fund to SEK 240 (€26) in the most expensive, in 2007 contributions varied between SEK 311 (€34) and SEK 656 (€71). This means that the new system as a whole, and some funds in particular, have become financially less attractive than before. One early consequence of this was a steep decline of about 6 per cent (210,000 persons) in total membership within unemployment insurance funds by April 2007 (EIROnline 2007).
In sum, due to the way the systems are designed, there are clear incentives for joining a voluntary unemployment insurance system in both Denmark and Sweden. In both countries, the level of individual fees (which can be set against personal income tax) is relatively low and the level of benefit support relatively high, especially for average and below-average wage earners. Legislative changes introduced recently in Sweden have certainly weakened the financial attraction of joining an unemployment fund compared to the situation before 2007, and there are indications of a significant drop in membership, which might cast some doubt on the sustainability of the system in the medium term (see below). However, before we discuss this issue, further aspects of the Danish and Swedish Ghent system have to be addressed that help to explain why the membership in an unemployment insurance fund is generally coupled with simultaneous membership in a trade union.
Voluntary unemployment insurance – still a recruitment tool for trade unions in Nordic countries
Supported by social surveys in both Sweden and Denmark, unemployment insurance cover is widely considered a major, if not the most important, factor for joining a trade union (for example, Lind, Reference Lind2004; Jørgensen, Reference Jørgensen2002; Timonen, Reference Timonen2003). However, the reasons which make the more expensive option of a dual membership more appealing than the cheaper option of merely joining an unemployment fund but not a trade union cannot be gauged from such surveys. In this section we identify two of those aspects which apply to both countries (identification with the trade union movement and institutional linkages), as well as two country-specific factors: active labour market support in Denmark and additional unemployment benefit cover in Sweden.
The traditionally strong and widespread identification with the trade union movement in both countries is certainly a factor that explains the popularity of a dual membership in both countries. Merely becoming a member of an unemployment insurance fund but not the respective trade union might be regarded as a sign of disloyalty towards fellow employees and their representation in the local workplace. However, a related and reinforcing second reason is a widely held perception that, despite a legal and formal separation between the two organisations, unemployment insurance cover is conditional upon trade union membership (see also Hatland, Reference Hatland1998). In recent years, Danish governments have aimed to create more transparency in this respect: that is, by encouraging members to switch between funds. However, as discussed earlier, anxious not to undermine traditional trade union demarcations, all but a few Danish funds have refrained from the option of recruiting members from other sectors, thereby perpetuating the perception that unemployment insurance funds and respective trade unions are closely linked. In addition, the formal separation between trade union and insurance fund at local level is undermined by the fact that, according to interviews in Denmark, it is not uncommon for individual personnel to work part-time in both organisations, which are often located within the same building. Dealing with staff who divide their working time between unemployment insurance fund and trade union can thus be regarded as another factor that contributes to a view of two closely linked organisations, despite a formal disconnection and separate membership fees. According to Kjellberg (Reference Kjellberg2006: 96), the formal separation between the two organisations in Sweden has not been widely understood either, and trade union membership is widely perceived as a prerequisite for access to an unemployment fund.
In addition to the absence of a clear institutional separation, there are country-specific reasons for dual membership, which are more directly associated with individual interests. In Denmark this applies to access to employment services; in Sweden to the prospect of better unemployment cover. As discussed above, a reform in 1969 created a structure in which Danish funds retained the responsibility for unemployment insurance, whereas the newly created public employment offices (AF) were put in charge of active labour market policy. However, this division remained somewhat insubstantial since unemployment insurance funds, and especially those linked to the LO-affiliated trade unions, maintained their active role as job brokers. Moreover, since their job placement rates were generally superior to those of the AF, unemployment insurance funds remained pivotal for job seekers. Eventually this situation led to the abolition of the AF's formal monopoly in job placement in the late 1980s and the de facto acknowledgement of the role of the a-kasser in the field of active labour market policy. As discussed, currently a-kasser offer their members a range of services aimed at job reintegration, such as information on vacancies and training options, counselling and general support with job search. It is noticeable that funds which closely collaborate with trade unions for semi-skilled and unskilled workers are particularly active in job-search services.Footnote 6 An important point here is the fact that some unemployment insurance funds offer these services in conjunction with the relevant trade union. Moreover, at local level there are often close contacts between a-kasser, the relevant trade union and employers in a particular sector of industry. The chance of a rapid reintegration into the labour market is thus higher, or perceived to be higher, for job seekers who are also trade union members.
In Sweden, active labour market services are not offered by unemployment insurance funds, which exclusively administer unemployment insurance benefits. Instead, public employment offices are in charge of job placement and job search assistance, while advice on benefit receipt or other relevant information for job seekers is offered by relevant trade unions directly. However, for many members of unemployment insurance funds, a simultaneous trade union affiliation has an important material impact in case of unemployment. As discussed, unemployment insurance benefits in Sweden are capped with a ceiling that is relatively low. Moreover, unlike regular indexation as in Denmark, this ceiling (as well as benefit floor) is adjusted by the government on a discretionary basis at irregular intervals. As a result, benefits for a growing share of unemployed people remain well below the nominal 80 per cent replacement rate. In 2003, the average replacement rate was 63 per cent of previous earnings (IAF, 2005), and around 45 per cent of those eligible for the income-related benefits did not receive the stipulated 80 per cent compensation level in 2005 (Lundgren, Reference Lundgren2006).
As a consequence, over time several schemes have developed and become increasingly relevant as a top-up to public unemployment insurance support (for an excellent overview, see Sjögren Lindquist and Wadensjö, Reference Sjögren Lindquist and Wadensjö2006). These include severance pay agreements, which mainly benefit people over the age of 40, as well as a multitude of ‘collective agreement-based insurance cover’, which top up state benefits generally to 80 per cent of previous individual earnings. The exact design of the latter varies across occupations and industries, but such arrangements are common within both the public and private sectors. A third scheme applies exclusively to trade union members. In 2005, several SACO unions and two TCO unions offered their members ‘collective insurance cover’ as part of their total trade union membership package. Providing income supplement in addition to public unemployment insurance benefits for between 100 and 120 days, this cover is normally funded out of slightly raised standard trade union fees. Finally, linking up with commercial insurance companies, some trade unions offer their members the option of purchasing supplementary individual insurance policies, providing benefit for a prolonged period of unemployment. Having become rather widespread in recent years (Engblom, Reference Engblom2005; Sjögren Lindquist and Wadensjö, Reference Sjögren Lindquist and Wadensjö2006), the point to note here is that collective additional cover provided by trade unions is conditional on simultaneous membership in the relevant unemployment insurance fund, a fact which certainly nurtures the link between the Ghent system and Swedish trade unions.
Conclusion
The discussion has shown that the voluntary Ghent system can be more effective in social policy terms (coverage and degree of social protection) than mandatory public unemployment protection. High levels of membership in unemployment insurance in Denmark and Sweden can be explained with reference to low direct personal costs and relatively high benefits in case of unemployment. Moreover, there are several factors that favour the simultaneous membership in unemployment insurance and trade union. The identification with the trade union movement suggests membership in the unemployment insurance fund close to the corresponding trade union. The post-tax financial costs of a dual membership are relatively low, but have recently risen in Sweden. Third, the formal separation between trade union and unemployment insurance does not always correspond with public perceptions of unemployment insurance as a component of the total benefit package gained by joining a trade union. Finally, many unionised job seekers have the advantage of potentially more effective job search support in Denmark and access to additional transfer income during periods of unemployment in Sweden.
Do these factors suggest that voluntary unemployment insurance is likely to remain a characteristic feature of Nordic welfare states in the medium term? Currently, the answer to this question seems to us, on balance, affirmative in the Danish case but more uncertain for Sweden in the light of recent changes. Certainly, political threats to the Ghent system can be identified in both countries. Arguably, in Denmark these include the decline in value of the benefit ceiling, the expansion of private unemployment insurance or the government's attempt to significantly lower the generosity of unemployment support in September 2003. However, with 3 per cent of all employees, the scope for additional private unemployment insurance has remained rather small. Moreover, the fact that it is the trade unions that are the providers of additional private unemployment cover – offered to their members as long as they belong to an unemployment fund – strengthens the system rather than erodes it. Finally, ever since the ill-fated attempt to reduce the level of benefits for the unemployed, which was uniformly criticised by trade unions and employer organisations alike, the conservative-liberal government has adopted a much more balanced position. While emphasising the need to increase competition between funds, the system of voluntary unemployment insurance is now portrayed as an important element within the Danish model of ‘flexicurity’ (Madsen, Reference Madsen, Campbell, Hall and Pedersen2006), which combines flexible labour markets, generous unemployment support and a strong emphasis on activation. At least for the time being, the Danish government has been anxious to promote this model as a way of maintaining economic competitiveness.
A prediction for Sweden seems rather more difficult. Certainly, the changes implemented in early 2007 have made the system less attractive. An even greater variation between membership fees across funds might foster a competitive situation, with employees breaking with the convention of joining the fund closest to their occupation or qualification, engaging in ‘shopping around’ for best deals. However, it is equally possible that developments will be similar to those in Denmark, where the system adapted to the option of becoming more market-based, with trade unions, by and large, finding ways of avoiding competition. Perhaps more seriously, the government plans to implement a compulsory unemployment insurance fund by 2008, which is strongly opposed by both social partners as well as the parliamentary left (EIROnline, 2007). If the government succeeds, traditional structures of the Swedish Ghent system will change dramatically, making the end of the system conceivable, which, in turn, would have serious repercussions for the strength of Swedish trade unions.
