Introduction

1Like the welfare state in many other countries, the Danish welfare state has faced “permanent austerity” (Pierson, 1998) during the last decades. High unemployment, public debt, an ageing population, etc. have put the welfare state under financial pressure. Even though there are conflicting evaluations of the development of the Danish welfare state in this period (e.g. Korpi 2002, Andersen 2000b and Green-Pedersen 2002a), most observers argue that the Danish welfare state has actually survived quite well. Some cuts have taken place, especially in unemployment benefits and early-retirement benefits, but other parts of the welfare state have actually been expanded under “permanent austerity”. However, it is important to be aware that the question of “what has happened to the (Danish) welfare state” can be asked from different perspectives which result in different answers (Green-Pedersen forthcoming). Most studies of the Danish welfare state focus on the generosity of the welfare state, e.g. the level and duration of benefits. However, it is also relevant to focus more on the institutional characteristics of the Danish welfare state. The general picture of continuity may hide important changes in the institutional characteristics which, on a long term, may have significant consequences.

2One of the most important institutional traits of the Danish welfare model is universalism, which in this paper is understood as a welfare model “providing adequate benefits for (almost) the entire population” (Rothstein, 1996). Even though the general picture of the Danish welfare state over the last decades is one of expansion as much as one of retrenchment, what has happened to the universalism of the Danish welfare state is an open question. This is the topic of this paper. Empirically, it focuses on the Danish pension system, a policy area that is frequently discussed in relation to universalism, but it also looks at a couple of other policy areas, especially child allowances.

3Theoretically, the paper will take its point of departure in the debate about the political underpinnings of the universal welfare state. The best-known theory of the universal welfare state is the power resource theory (e.g. Esping-Andersen, 1985), which argues that the universal welfare model is politically strong because it is based on a class compromise between the working class and the new middle class, serving the interest of both. Social Democratic parties are, therefore, considered the main advocates of universalism. However, Rothstein (1996) argues that the political strength of universal welfare is based on the moral logic involved in preventing a debate about how much society should give to those who cannot take of themselves.

4This paper offers a different argument about the politics of the universal welfare state. It argues that right-wing or bourgeois parties often push for more universal benefits because they are more market conforming than income – or means-tested benefits. Universal benefits do not damage market incentives to take a job or save for your own pension. On the other hand, left-wing or Social Democratic parties often argue for more income – or means-tested benefits because they are argued to be more redistributive.
The first section presents the theoretical debate about the universal welfare model. Then follows a section on the development of the Danish pension system in terms of universalism over the last decades. The next section looks briefly at universalism in other welfare schemes, showing that it has survived and, in some regards, even been expanded in Denmark in recent decades. Then follows a look at the current political debate in Denmark about universalism, which shows the weakness of the political arguments behind universalism. The paper ends with a conclusion.

The political underpinnings of the universal welfare state

5Before discussing the different theories of the political underpinnings of the universal welfare model, the concept of universalism itself deserves a few comments. The concept is often used in different ways, which, of course, is problematic as the different theories of the political underpinnings may refer to different things. A general question is whether the different definitions see the universal welfare model as the welfare model actually found in Scandinavia or as a theoretical model, which exists in various forms in Scandinavian and other countries (Andersen, 1999). This paper follows the latter solution and understands the universal model as suggested by Rothstein (1996) as one where the welfare schemes offer adequate benefits for (almost) the entire population. Based on this definition, attempts to limit the number of benefit recipients through, for instance, means testing are to make the welfare state less universal. Reducing the level of benefits may also be considered a move away from universalism in so far as they are reduced to a level where they can no longer be considered adequate for people with average or just above average income.

6The most famous theory of the political underpinnings of the universal welfare model is the “power resource” theory (O’Connor and Olsen, 1998). The core of the power resource theory is that the universal welfare model is established as a class compromise between the working class and farmers, and later, when there were fewer farmers, the new middle class of white-collar workers (cf. Esping-Andersen, 1985 and 1990). For the working class, it was crucial to secure a class alliance because it could never achieve a political majority on its own. Building a coalition with farmers, and later with the new middle class, secured a political majority. The price was that welfare state benefits had to be tailored to the interests of the coalition partners, which was also the reason for making benefits universal. Otherwise, farmers and the new middle class would turn towards the market. Benefits need to be for everyone and at a fairly high level to crowd out private alternatives.

7However, the argument that adequate benefits are necessary to achieve the crowding out is most relevant with regard to social risk where private alternatives such as pension systems exist. With regard to child allowances, for example, a private alternative is not really there and universalism is more a strategy to secure middle-class loyalty and willingness to pay taxes. The crowding out it effects is why the power resources model (e.g. Korpi and Palme, 1998), argues that Social Democratic parties should work for schemes such as the Swedish ATP, where people with higher income get higher benefits in order to avoid them turning towards the market.

8The paradox of the universal welfare state is thus that it serves the interests of the working class by benefiting the other classes as well. This is what Korpi and Palme (1998) label the paradox of redistribution. Giving benefits to large groups in society does not look like a strategy of redistribution, but, in fact, universal welfare states are the countries with the highest levels of redistribution. The reason is that even though every dollar spent on welfare is not highly distributive, there is still redistribution in the universal model, largely through the tax system, and as the number of dollars spent on welfare is high, the final level of redistribution is high. The universal welfare model is based on Titmuss’ old formulation (1976) that “welfare for the poor is poor welfare”.

9The political implication of the argument about the political underpinnings of the universal welfare model is that Social Democratic parties and other parties that fight for working-class interests should also promote and defend universal welfare benefits. In contrast, right-wing parties should work to avoid and reduce universal welfare benefits, in order to break the class alliance between the working class and other classes.

10Rothstein (1996 and 1998) has argued that the political underpinnings of the universal welfare model today are not so much Social Democratic parties as what he labels the “moral logic of the universal welfare state”. Once established, the universal welfare state affects the norms that govern the political debate of the welfare state. Universal welfare schemes become the starting point for political debate and the opponents of the universal welfare model must have strong arguments why citizens should suddenly be treated unequally. Furthermore, universal welfare states avoid problems with the implementation of welfare schemes, e. g. deciding whether people are actually needy enough to receive benefits and the concomitant debate that benefit recipients are not really needy. The whole distinction between the needy and not needy disappears in the universal welfare model. Once established, universal welfare benefits secure their own support. Rothstein’s argument thus resembles Pierson’s well-known argument (1996) that once established, welfare states create a new politics of the welfare state as they create their own political support and new interest groups.

11This paper agrees with Rothstein in as much as the strength of Social Democratic parties is not the main political underpinning of the universal welfare model today and that the very limited implementation problems involved in universal systems are part of their strength. Another strength of such schemes is their market conformity. Once income testing or means testing is introduced, they inevitably create a disincentive to behave according to the market, for instance, to look for a job or save for your pension. The reason is that you will lose benefits once you have an income from work or have your own pension.

12Such disincentives are likely to worry market-friendly, right-wing parties, which, therefore, are likely to prefer universal benefits to means – or income-tested benefits. This market friendliness of universal benefits is, of course, only there until universal welfare benefits crowd out private solution. However, as argued above, the “crowding out” problematic is most relevant with regard to pension schemes, where private alternatives exist. With regard to many other schemes, such as child allowance and student allowance, a private alternative is not really there.
Rothstein and the power resources theory also overlook a major weakness of the universal welfare model in political debates, namely its “counter-intuitiveness”. As proposed by, for instance, Goodin and Le Grand (1987), few would disagree with redistribution as at least one of the central goals of the welfare state. A majority of the Danish electorate also favours means testing or targeting at the general level as the welfare state should support those who need it most (Andersen, 2001, 17). As argued above, the universal welfare model is highly redistributive, but it is partly counter-intuitive: you have to benefit the rich to benefit the poor. The problem is that this is a social science finding which is considerably more complicated than the argument that if you help the less well-off, you should target benefits towards them and not give it to everybody. The counter-argument that “welfare for the poor is poor welfare” is difficult to advance very strongly in a public debate. If the debate is about where to focus necessary cutbacks, then it is very difficult to go against an argument saying that you should take benefits away from those who do not really need them and keep them for those in need. This “counter-intuitiveness” is the weakness of universalism in the public debate. Furthermore, the redistribution argument is most likely to be attractive to Social Democratic or other left-wing parties that focus on inequality. Therefore, they are likely to advance it as an argument against universalism.
The empirical sections below look at the development of the pension system and other welfare schemes. The focus is whether the welfare state has actually become more or less universal and the political forces mainly political parties that push it in one or the other direction.

More or less universal? The Danish pension system in times of economic crisis

13Before looking at the development of the Danish pension system over the last couple of decades, a few words on its historical development are necessary. The historical development of the Danish pension system differs from that of the other Scandinavian countries (Salminen, 1993), which is a significant reason why the Danish welfare state is considered the least universal of the Scandinavian welfare states from a power resources perspective (Esping-Andersen, 1985).

14During the late 1940s and 1950s, all three Scandinavian countries introduced national pension schemes covering the whole population and at fairly high levels. However, it was clear that for people with higher incomes, these schemes offered inadequate compensation, and this triggered political discussions about income-related pension schemes on top of the basic pension schemes. After intense political debate, Norway and Sweden introduced the ATP schemes offering earnings-related benefits on top of the basic pension. Universality was thus secured in these countries as everyone was covered by the same scheme, with adequate benefits also for people with high earnings, and private pensions were crowded out (Esping-Andersen, 1985, 145-178).

15In Denmark, the national pension was introduced in 1956 and expanded in 1964. At that time, the plan was a flat-rate pension for all Danish pensioners. During the 1960s, however, parliament introduced an income-tested element into the pension. The national pension thus came to consist of a universal, flat-rate, basic amount plus an income-tested pension supplement, equalling, on average, approximately 25 per cent of the basic amount. The scheme is a pay-as-you-go system financed though general taxation. An ATP scheme was actually introduced in Denmark in 1963, but contributions and benefits, which are tied to the number of hours worked – not earnings – were at a low level (Vesterø-Jensen, 1984). The fact that the need for a second-tier pension scheme in Denmark was not solved created an expanding market for private pensions (op. cit.).

16There is some disagreement as to why an extensive ATP scheme was not introduced in Denmark. Some link it to the political weakness of the Danish Social Democrats compared to their Scandinavian sister parties (Esping-Andersen, 1985). Others argue that it was more the result of disagreement between the Social Democrats and the trade unions, which were not really interested in such a scheme, and the fact that the Social Democrats linked the issue to their attempt to introduce economic democracy with central wage earner funds in Denmark (Pedersen, 2002; Green-Pedersen & Lindbom, 2003). Whatever the cause, part of the “pension question” was still unresolved in the early 1980s in the sense that blue-collar workers especially faced a huge drop in income on retirement.

17The national pension was improved in the late 1970s and was the most generous in Scandinavia (Øverbye, 1996). In the 1970s, around half of the pensioners received the full pension supplement, approximately 25 per cent a reduced part of it, and the remaining 25 per cent none of it (Vesterø-Jensen, 1985, 42-46). Still, even in combination with ATP, it was far from adequate as a pension solution for many blue-collar workers. The discussion of the pension issue in Denmark in the 1980s was thus centred on finding a pension solution for the groups that only had the basic pension and the limited ATP. The discussion of the degree of universality in the national pension was closely connected to the debate on finding a solution to this problem.

18For two reasons, it was very clear that the solution could not just be a general expansion of the national pension or the introduction of a Swedish-type ATP. First, a number of employees, especially in the public sector, had established funded occupational pensions in the 1960s as part of their collective wage agreements, and, in the early 1980s, around 40 per cent of Danish wage-earners had occupational pensions. If the national pension was expanded, these groups would end up with compensation rates above 100 per cent. The establishment of these pensions had thus laid down a path that precluded a general expansion of the national pension (Green-Pedersen and Lindbom, 2003). Second, due to Denmark’s economic situation, the tax increases that a huge expansion of the national pension would entail were impossible. Furthermore, the “greying” of the population became visible during the 1980s, and funded pension systems such as occupational pensions were preferable to the pay-as-you-go structure of the national pension. Therefore, all political actors advocated some form of occupational pensions (op. cit.).

19The political disagreements over the pension issue during the 1980s were first of all about the exact form of occupational funds. Should the system be introduced through legislation or through collective agreements? And should there be one scheme for all blue-collar workers or should they be decentralized to the level of individual trade unions (Green-Pedersen, 2003)? However, there was also political disagreement about the future of the national pension as a universal pension scheme once occupational pensions had been introduced. Should this element securing some degree of universalism be kept in the pensions system once occupational pensions had been introduced or should it be turned into a means-tested scheme?

20In 1985, the Danish Confederation of Trade Unions (LO) presented their vision of the pension system (LO, 1985). A suggestion to introduce one central occupational pensions system for all wage earners through legislation was followed by a suggestion to income test the entire national pension. The question caused some controversy between the Social Democratic parliamentary group and the trade unions as the former wanted short-term improvements of the national pension and was unhappy about income testing the basic amount, which had actually been part of the Social Democratic pension policy since the late 1970s (Petersen, 2002). The argument behind income testing the entire national pension was redistribution. A national pension targeted towards those with no other pensions could be more generous and thus better for pensioners with no or limited income besides the national pension (LO, 1985, 100). However, the suggestion to income test the basic amount was not part of the proposal that the Social Democratic parliamentary group put before parliament in 1986 (Smith, 1996).

21The right-wing governments in Denmark during this period did not support income testing of the entire national pension. Their pension policy centred on establishing decentralized occupational pension through collective agreements. Part of the strategy was to ease income testing of the pension supplement, which was done step-by-step, along with improvements of both the basic amount and the pension supplement. The number of people receiving the full pension supplement thus increased from approximately 50 per cent to approximately 66 per cent. The argument behind making the national pension more universal was to ensure that low-wage earners would actually benefit from establishing occupational pensions. If their income from occupational pensions would simply mean a smaller pension supplement, it was hard to ask them to introduce occupational pensions through their collective agreements. Thus, the decade of centre-right-wing governments in Denmark from 1982 to 1993 actually made the national pension more universal, mainly by making more groups eligible for achieve the pension supplement (Green-Pedersen, 1999).

22The only step away from universalism during the period of centre-rightwing governments was the introduction in 1984 of income testing of the basic amount based on considerable income from work for the pensioners aged 67 to 69, i.e. people who in fact continued to work. However, the change was marginal as it affected less than two per cent of the relevant pensioners. Furthermore, it is interesting that while the Social Democrats fiercely opposed all cutbacks made the by the centre-right government that took office in 1982, they actually supported this one (Green-Pedersen, 1999). They even teased the government that it had stolen an old Social Democratic proposal about income testing of the basic amount, a proposal which the right-wing parties had blocked while in opposition. The Social Democrats thus supported more income testing (Folketingets Forhandlinger, 1982-1983, 412-417).

23In 1993, a Social Democratic-led government took over. At that time, a solution to the pension question had been found. Starting in 1989 and continuing in 1991, occupational pensions were introduced for groups that did not already have them as part of collective agreements. In other words, the centre-right governments more or less achieved the pension solution they wanted. In the end, the trade unions found it more important to have additional pension coverage for their members than to wait for a possible solution through legislation. The new Social Democratic government responded to the final solution to the Danish pension problem by making the national pension more income-tested in the spring of 1993, right after taking office. The change, which was technical and hard to understand for the electorate, is a fine example of what Paul Pierson (1994, 19-22) calls a strategy of obfuscation.

24It was very difficult for pensioners to understand the effect of the changes made, which were the following: single Danish pensioners had always been entitled to an enhanced personal tax exemption, i.e. they paid no taxes on the national pension if they had no other income. One of the elements of the tax reform that the new Social Democratic government also introduced was to make all social benefits taxable and to compensate recipients by increasing benefits. This was also the case with the national pension, but the compensation was added to the income-tested pension supplement after a transition period. In addition, the basic amount for all pensioners was lowered slightly, and the pension supplement raised. In sum, the relative importance of the pension supplement in relation to the basic amount was increased. Today the two amounts are approximately equal for single pensioners, and the pension supplement equals approximately 50 per cent of the basic amount for pensioners living together as couple.

25For pensioners with no other income than the national pension, the relative importance of the pension supplement and the basic amount does not matter, but for the system in general the change made in 1993 was a step away from universalism and is a long-term saving device. As the occupational pension schemes mature, fewer pensioners will be entitled to the pensions supplement. The percentage of pensioners receiving the full pensions supplement declined to 62 per cent in 2002, after having increased to 69 per cent in 1993 (Statistiske Efterretninger, Social Sikring og Restvæsen, various years, and Statistiske Efterretninger, Social Statistik, 2002, 4). The change in the national pension scheme was made in connection with the tax reform containing tax reliefs, and the changes were calculated so that no pensioner would lose money after tax, including those hit by the increased income testing of the national pension (cf. Green-Pedersen, 2001b, 210-212).

26The major right-wing parties voted against the changes due to their resistance to the general tax reform. However, the debate in parliament revealed that they also had reservations concerning the increased income testing of the national pensions, which was the de facto effect of the changes. As usual, their argument was that the changes diminished people’s incentives to save for their own pension. The Social Democrats argued that if the effects of the tax reform were included, all pensioners would be better off, and that they had no problems with changes that would benefit needy pensioners most (Folketingets Forhandlinger, 1992-1993, 10057-10067). The Social Democratic-led government also extended the income-from-work testing of the basic amount to all pensioners, not just those aged 67 to 69. As with the change made for the 67-69-year-olds, the effects were marginal, but it was still a break with the universal national pension introduced by a Social Democratic-led government and opposed by the major right-wing parties.
Since then, the rules for the Danish national pension have remained largely unchanged. The only change worth mentioning was the introduction of a “senior cheque” at the end of 2002. This was a fairly limited amount given only to pensioners with no other income than the national pension. The cheque was the result of an agreement between the Liberal-Conservative government that came into office in 2001 and its supporting party, the Danish People’s Party. The latter demanded the cheque in return for supporting the 2002 budget. Initially, the cheque was given for one year, but the government and the Danish People’s Party have now agreed to make the cheque permanent. The argument behind the cheque was to help the most needy pensioners. The Social Democrats had no problems with this admittedly small step away from universalism and supported the cheque in the final vote, although they had no real influence.
Looking at the politics around the Danish national pension during the last two decades, it is hard to see the picture suggested by the power resources approach of the Social Democrats as the main political force behind universal benefits. If anything, the contrary seems to be the case. Under the centre-right-wing governments in the 1980s, the national pension moved more in the universal direction, whereas it moved in the opposite direction during the Social Democratic-led governments in the 1990s. The latter were fully in agreement with the line the Social Democrats had adopted since the late 1970s whereby once the pension problem had found a solution, the national pension should be more focused on the pensioners with no other income. The reason the non-socialist parties have generally pushed the national pension in the more universal direction is not the idea of universality itself, but their focus on giving people an incentive to save for their pensions. The idea that “saving should pay” works better with a universal than with an income-tested pension system, and this was the driving force behind the moves in the universal direction in the 1980s.
An interesting question is whether the universal elements of the national pension will survive in the future when the majority of pensioners receive their main pension income from occupational pensions. As many of these schemes were not established until the early 1990s, they are not that important for the majority of people retiring today, but they will become so in the decades to come. Today, 82 per cent of Danish wage earners are covered by occupational pensions, in most cases paying between 9 and 15 per cent of their income as pension contributions, so in the long run occupational pensions will be the most important part of the Danish pension system (Pensionsmarkedsrådet, 2001). Cox (1997) has argued that these changes could mean the end of solidarity in the pension system. When fewer people have the national pension as an important part of their pension income, their willingness to support it might decline. In addition, occupational pensions automatically increase inequality among pensioners, and this may create political pressure towards more benefits targeted at needy elderly like the senior cheque.
At present, opinion polls do not indicate a decline in support for a universal national pension, which enjoys very strong support among the electorate (Andersen, 2001, 18). The government could implement the same kind of “obfuscated” reform as it did in 1993, but there is little, if any, indication that the government would be willing to touch the national pension, so the alleged negative consequences of the changes in the pension system for a universal national pension have not emerged yet.

Universalism in other parts of the Danish welfare state

27The pension system is far from the only Danish welfare scheme with universal elements, and it is worth examining whether the same dynamics can be found in the other schemes.

28One type of scheme which is often universal is that of child allowances. In Denmark, they were part of the tax system until 1970, when universal child allowance was introduced. However, in 1977, the Social Democratic government made child allowance means tested, and the debate in parliament was interesting (Folketingets Forhandlinger, 1975-1976, 12012-12026). The Social Democratic speaker was not happy about making cutbacks, but when they had to be made, taking benefits away from the least needy was a fair way to do it. On the other hand, the Liberals had suggested that cutbacks should be made by cutting 10 per cent of everybody’s child allowance and thus preserving universality. The argument for this solution was that it would avoid the ensuing administrative costs of means testing, but also that it would, in fact, raise marginal tax rates and make work pay less. Once people increased their income, they would face higher tax brackets and a smaller child allowance.

29Means testing was introduced and tightened by another Social Democratic government in 1980. This, combined with the fact that the levels in the means testing were not fully indexed, cut the number of people receiving the full child allowance from 74 per cent in 1977, when means testing was introduced, to 41.8 per cent in 1985 (Knudsen, 1987). Limited indexation also decreased the real value of benefits significantly (op. cit.). As part of a tax reform in 1986, however, a broad majority in the Danish parliament reintroduced universal child allowance and a much higher level of benefit known as the child family allowance. This allowance has since been raised several times, especially for families with small children. Universal child allowance has thus survived in Denmark, but the period from 1977 to 1987, when means testing existed, shows that it was under pressure. As in the pension area, Social Democratic rule does not seem a good protection of universalism. In fact, the party has been willing to make cutbacks by limiting universality.

30Another policy area where there has been debate about universalism is home help for the elderly. Until 1989, affluent pensioners had to pay for home help, which was free for more needy pensioners. Already in 1986, a broad majority in parliament made home help free for everyone, which came into effect in 1989. In 1992, the same broad majority reintroduced fees for the affluent, but regretted the decision and made it free for everyone again later in 1992. The changes in 1986 were, as the changes to national pension in this period, driven by a wish to make saving for your pension pay. The reintroduction of fees in 1992 was a cost-saving device, but one that was difficult to administer, which caused the reversal later that year. Administrative costs of fees also seem to cause the stronger Social Democratic support for universalism in this areas than in others (cf. Folketingets Forhandlinger, 1992-1993, 3339-3354).

31Finally, the Danish student allowance has developed – after several reforms mainly in the mid – and late 1980s – from a scheme where benefits were heavily dependent on parents’ income into a universal and very generous scheme. For students under 20 in secondary education, the student allowance consists of a basic amount independent of parents’ income and a supplement for students with less affluent parents. These significant moves in the direction of a universal student allowance were made with broad political agreement about the principle of universalism, but with political disagreement about how generous a scheme was possible (cf. Green-Pedersen, 1999). In particular, the right-wing governments of the 1980s found the scheme too expensive and wanted universal benefits at a lower level.
Neither the pension system nor the examples from other schemes paint a picture of the Social Democrats as supporters of universalism and right-wing parties wanting means or income testing. In political disagreements about universalism, e. g. in connection with child allowance in the 1970s, the Social Democrats actually supported means testing on the grounds that it was the socially responsible solution, whereas the right-wing parties opposed it, arguing that it was a de facto disincentive to work.

The future of universalism in Denmark

32The empirical sections above indicate that the Danish universal welfare state has by and large survived. In some areas, e.g. home help for the elderly and the student allowance, the welfare state has actually become more universal. With regard to the pension system, the picture is more mixed, due to moves both towards and away from universalism. Furthermore, areas such as the child allowance and home help for the elderly have seen the reintroduction of means testing, so forces pushing away from universalism have also been at play during this period. Politically, it is very hard to find the picture suggested by the power resources theory of the Social Democrats wanting universalism against the wishes of right-wing parties. It was actually the centre-right-wing governments from 1982 to 1993 that made the moves in the universal direction, whereas most of the moves in the opposite direction happened under Social Democratic or Social Democratic-led governments before and after this period.

33The moves away from universalism were generally made with an argument that introducing means testing was a socially responsible way of making cutbacks as they would only hit the affluent. This kind of argumentation is clearly one of the challenges to the universal welfare model. The “counter-intuitiveness” of the universal welfare model makes it vulnerable to such arguments, not least if they are made by Social Democratic or other left-wing parties. The hot topic in the current welfare state debate in Denmark is how to handle the challenges from the “greying” population. This development will inevitably put pressure on social expenditure, which has caused a recurring debate about how to prioritize. In this debate, powerful actors such as the Danish Confederation of Trade Unions (LO) and the organization of the Danish municipalities (KL) have argued in favour of means testing the basic pension, child allowance and home help. They have often been supported by left-wing parties and individual Social Democratic politicians.

34The voters (Andersen, 2001) are to some extent in agreement with these organizations. A majority supports means testing of the child allowance, whereas a majority opposes means testing of the basic amount of the national pension. A recent poll, in which the formulation of the questions “forced” the voters to suggest cutbacks to meet the challenges from the “greying” of the population, also showed that means testing of the child allowance and fees for home help could find support (A4, April 14, 2003, www.lo.dk).

35The right-wing parties and the Social Democrats have replied that they are not planning cutbacks, so at present moves away from universalism are not about to happen. However, if future governments find cutbacks necessary, income testing, e.g. of the child allowance, seems likely. Retrenchment of the welfare state is politically very difficult and is the politics of blame-avoidance (Pierson, 1994), so if future Danish governments are forced to make cutbacks, they are quite likely to search intensively for areas where they can justify them in the eyes of the electorate (Green-Pedersen, 2002b). Income testing of schemes like child allowance seems among the few candidates.

36Another future threat to universalism has to do with securing adequate benefits and thus with the crowding out issue suggested by the power resources approach. The Danish health care system came under increasing attack during the 1990s due to waiting lists and alleged poor quality treatment and care. This criticism is not necessarily warranted, but if voters accept it, they show greater interest in private insurance on top of the national health insurance (Andersen, 2000a). Voters still strongly support a universal health care system (Andersen, 2001), but if they lose faith in it, they may undermine it by turning to private solutions. There are indications that such a process is slowly taking place (Andersen, 2000a), and, although private insurance has increased in recent years, it is far from inevitable that this will undermine the universal health care system.
The Social Democratic-led governments from 1993 to 2001 had limited success in restoring the confidence in the health care system, despite increasing budgets. One reason the right-wing parties won the 2001 election was exactly the problems in the health care system. The right-wing parties campaigned for more money for the health care system, and managed to convince the electorate that they would be better at improving the health care system than the Social Democrats (Andersen, 2003). The right-wing election promises, which they carried through after gaining government power, were more focused on re-establishing confidence in the public health care system than on making private solutions more attractive. Besides more money to do away with waiting lists, another initiative is an option to receive experimental cancer treatment abroad paid by the public health insurance. Thus, it does not seem that the right-wing parties have been interested in breaking one of the backbones of the universal welfare state, namely the universal health insurance scheme.

Conclusions

37The focus of this paper has been partly to present an overview of the development of universalism in Denmark during the decades of “permanent austerity”, especially in the pension system, and partly to discuss the Danish development in light of the theories about the political underpinnings of the universal welfare model. The two best-known theories are the power resources theory and Rothstein’s argument (1996) about the “moral logic of the universal welfare state”.

38In the theoretical section of this paper, the argument was that both theories underestimate the “counter-intuitiveness” of the universal model, namely that in order to benefit the poor you have to benefit the rich as well. There is strong support for this as a social science finding, but in the political debate, the argument that to benefit the poor you should target benefits at them is much more forceful and is a potential threat to the universal welfare state. Especially in times of economic crises, introducing means testing seems to be a “socially responsible way to save”. This was confirmed by the empirical section showing that moves away from universalism in Denmark during the economic crisis were justified by exactly such arguments, and they were typically advanced by the Social Democrats and left-wing parties. The strength of the universal welfare model was argued to be partly the limited implementation problems as argued by Rothstein, but also that it does not damage incentives to, for instance, save for your own pension or to take a job. Exactly such arguments were behind the moves in the universal direction in Denmark and these were mostly driven by right-wing parties focusing on making work and savings pay.
The universal Scandinavian welfare states are generally considered the achievement of the Scandinavian Social Democratic parties, implying that Social Democrats should work for universalism. However, this paper raises doubts as to whether this is true. The reason is that the universal welfare state intuitively conflicts with a social democratic idea that redistribution is a central welfare state goal. This was supported by the overview of the development of the Danish welfare state under the economic crisis. As argued by Pierson (1994) and Rothstein (1996), the very establishment of the welfare state alters the politics around it, and right-wing support for universalism may, therefore, be a result of the welfare state being a matter of fact and may thus be something new. Nevertheless, this paper supports more historical studies such as Baldwin (1991) and Nørgaard (2001), which also raise doubts as to whether universalism is really a Social Democratic goal and a Social Democratic achievement.