Resumé

From 1982 to 1993, a decade of tight economic constraints, the Danish welfare state was governed by bourgeois parties. The expectation of the power resources model, the most prevalent about the Scandinavian welfare states, was that the Danish welfare state would be turned further in a residual direction. Those expectations were, however, never fulfilled - on the contrary, the Danish welfare state was further expanded in a social democratic direction. A focus on how the bourgeois governments, often in a difficult parliamentary situation, tried to adapt the popular Danish welfare state to less favorable economic conditions accounts for important aspects of this development. The bourgeois governments relied on a combination of increased benefits to large groups of voters and well-hidden measures to strengthen public finances.

Introduction

For a long time, the factors determining welfare state development have been much discussed within political science. Numerous volumes have been published on the issue, coming up with nearly as many answers. Still, during the 1980s, it seemed that the so-called 'power resources model,' focusing on the power resources of the working class in particular, had gained a dominant position.¹ For instance, Esping-Andersen's (1990) identification of three welfare state regimes, which is based on the power resources model (cf. Kemeny 1995; Esping-Andersen 1985b), has been the point of departure of much welfare state research (e.g., Orloff 1993; Taylor-Gooby 1996; Bonoli 1996; and Kloostermann 1994), and it is cited in almost every scholarlywork welfare state development. Recently, the power resources model has, however, been challenged.² Most prominently, Pierson (1994; 1996) has questioned the relevance of the power resources model in explainingwelfare

---

* Christoffer Green-Pedersen, Department of Political Science, University of Aarhus. Universitetsparken, Aarhus C, Denmark. E-mail: cgp(a'ps.au.dk

plainingwelfarestate retrenchment - or the lack of it - in the USA and the UK, just as the relevance of the power resources model in explaining welfare state development in countries with strong Christian democratic parties has been challenged (Van Kersbergen 1995; Cox 1993). The aim of this article is to contribute to the investigation of welfare state dynamics by examining the development of the Danish welfare state from 1982 to 1993, when bourgeois governments ruled the country.

Apart from the work of Baldwin (1990), the explanatory power of the power resources model in relation to the Scandinavian³ welfare states has remained largely unchallenged. Considering that the Scandinavian countries the origin ground of the power resources model, it seems worthwhile examine these countries more closely. In the terminology of Kuhn (1970), the Scandinavian countries are 'paradigmatic cases' for the power resources model, and not being able to explain the development here would constitute a serious challenge to the model.

Denmark from 1982 to 1993 is a pertinent case in this regard, because the ten years of continuous bourgeois rule are unique in the Scandinavian context. Here, it has been almost impossible to examine what happens when bourgeois parties rule the welfare state for a longer period of time,⁴ simply because social democratic government has been the norm. As Shalev (1983, 329-30) has argued, it is problematic to test the explanatory power of the power resources model only on the basis of a prolonged reign of social democratic parties, since the effect of variation in the main explanatory variable cannot be investigated. The Danish welfare state from 1982 to 1993 thus makes a unique opportunity of testing the power resources model in a Scandinavian context.

This article argues that the power resources model is of limited relevance in explaining the development of the Danish welfare state from 1982 to 1993. According to the power resources model, ten years of bourgeois rule should turn the Danish welfare state further in a residual direction, i.e., give the market a more prominent role in the provision of welfare benefits. However,empirical shows that these expectations were never fulfilled. In fact, the Danish welfare state was further expanded in a social democraticdirection, in the direction of a generous welfare state relying on universal, flat rate, and tax-financed benefits. Alternative theories are not really available and, therefore, the article proceeds with a more open-ended discussion, which starts out by asking why the power resources model seems to have lost its explanatory power. In short, the problem seems to be that the power resources model cannot take into account what Pierson (1996) has called 'the new politics of the welfare state,' which points to the popular entrenchment of welfare states. Thus, the bourgeois governments from 1982 to 1993 faced the unpleasant task of adapting the very popular Danish welfare state to less favorable economic conditions. Generally, they tried

to do so by relying on a combination of increased benefits to large groups
of voters and well-hidden measures to improve public finances.

The Power Resources Model

The core idea of the power resources model is that the welfare state is part of the distributional struggle in capitalist societies between the working class and the bourgeoisie. A rapid transition from capitalism to socialism never occurred, and instead the welfare state was established. Its basic aim is to secure workers against the perils of the market, the most advanced welfare states being those where the market has the least influence on benefits. The establishment of the welfare state became possible with the introduction of universal suffrage, which strengthened the position of the working class relative to other classes (cf. Stephens 1979; Esping-Andersen 1985a; 1990; Korpi 1983; 1989; Esping-Andersen & van Kersbergen 1992). Welfare state variation has sometimes been explained simply by the power resources of the working class, i.e., its numerical strength, the strength of trade unions and the extent of its government participation (cf. Korpi 1989). Other studies have given more sophisticated explanations by including such factors as the possibilities of class alliances (Esping-Andersen 1985a; 1990) and the unity of the bourgeois opposition (Castles 1978). However, the bottom line remains that the welfare state is seen as an opportunity for the working class to limit the influence of the market on the distribution of wealth in society (cf. Esping-Andersen 1985b).⁵

When focusing on Scandinavia, the most celebrated explanation of the building of the welfare states is Esping-Andersen's (1985a; cf. also Esping- Andersen & Korpi 1986). Originally, the Scandinavian welfare states were established by red-green alliances between the working class and the farmers. the number of fanners declined, the working class was able to contract an alliance with the 'new middle class' by gradually replacing targeted benefits with universal ones, which were more attractive to the latter. The purpose of this new alliance, as seen from a social democratic perspective, was to avoid that the new middle class would turn to the market for welfare benefits, a tendency that would undermine political support for the welfare state (Esping-Andersen 1985a, 244-46).

According to Esping-Andersen, the comparatively weak Danish Social Democratic Party has only been partly successful in this strategy. The markethas to play a rather strong role in Denmark in relation to both pensions and housing (1985a, 155-90; Esping-Andersen & Korpi 1986; cf. also Vesterø-Jensen 1985). In combination with the inability of the Danish Social Democrats to secure full employment and to finance the welfare state after the two oil crises, this partial failure to keep the new middle class away

from the market has partly undermined the party's political basis. It has becomethe decomposed of the Scandinavian social democratic parties, a decomposition that has not only strengthened the bourgeois parties, but also created a strong left-wing opposition. Denmark is thus seen as the most obvious Scandinavian candidate for a middle class defection, a possibility which was already indicated by the 1973 tax revolt (Esping-Andersen 1985a, 281-85; Marklund 1988).⁶

The question then becomes what hypothesis concerning the Danish welfare under bourgeois reign to develop from this. If political parties, as argued by the power resources model (cf. Korpi 1989, 310-14), act as agents of class interests, the basic aim of bourgeois parties must be to increase the role of the market in welfare provision, as this is to the advantage of the class interests they represent. Still, according to the power resources model (Korpi 1989, 313; 1991; cf. also Esping-Andersen & Korpi 1986, 71-73), one cannot expect that bourgeois parties will always try to strengthen the role of the market in the short run. The reason is that strong labor movements be able to move the political center to the left, i.e., change 'the strategic context.' This may give right-wing parties a short-term interest in partly pursuing the policy agenda of the left.

However, in Denmark from 1982 to 1993, the bourgeois parties did seem to have the possibility of pursuing their basic class interests. First, as noticed above, analyses based on the power resources model show that the Danish bourgeois parties had already been somewhat successful in creating a role for the market in welfare provision.⁷ Second, the long period of bourgeois should give the bourgeois parties time to countervail 'the cumulative of social democratic rule' (Shalev 1983, 330). Thus, with the Danish economy on the brink of the abyss (Nannestad 1999), the bourgeois governments should have the possibility of introducing reforms, both in terms of cutting back on spending and, more importantly, in terms of changing the content of the welfare state. As Esping-Andersen & Korpi put it, '[i]n light of current welfare state cutbacks and legislated institutional changes, all indications are that Denmark is moving further away from the "Scandinavian model"' (Esping-Andersen & Korpi 1986, 55). How far in the residual direction the bourgeois parties would be able to push the Danish welfare state is difficult to say, and one would not expect a bourgeois revolution. Still, from the perspective of the power resources model, the Danish welfare state is to become more residual. However, this is not what happened.

The Danish Welfare State under Bourgeois Rule

A way of getting a first overview of the development of the Danish welfare
state from 1982 to 1993 is to look at the development of expenses on welfare

state purposes. As seen from Table I, these increased by about 20 percent in fixed prices from 1982 to 1992.⁸ This growth was enough to keep pace with the growth in GDP. The percentage of GDP spent on welfare state purposes thus remained constant around 37 percent, although it fluctuated somewhat over the decade. Judging from the development of public spendingon state purposes, there are thus few signs of cutbacks. In fixed prices, spending on welfare state purposes actually grew considerably.

These aggregate figures conceal both cuts and expansions. Most of the cutbacks took place in the beginning of the period. The most important ones were the changes in the indexation of cash benefits. Starting in 1982, the level of voluntary early retirement, unemployment and sickness benefits, was more or less frozen for three years. Considering that the level of inflation Denmark was around 10 percent in 1982 (Nannestad 1999), freezing benefits for three years could be expected to save a considerable amount of cash.⁹ A waiting day for sickness benefits was also introduced¹⁰ (Knudsen 1987).

Expansion took place all through the period, but, as indicated by Table 1, most distinctly after 1986. At the beginning of the period, maternity leave was improved and a new and more generous disability pension was introduced. 1986, both the national pension and child allowances were improved several times, just as social assistance became more generous. As part of the budget agreement for 1988, the benefits that had been partly frozen were also extraordinarily improved (Ploug & Kvist 1999; Plovsing 1994, 58-82).

As these examples indicate, it was mainly cash benefits that were improved.Thus, is worth noticing that, as seen in Table 2, expenses on cash benefits grew more than four times as much as expenses on services, which only grew slightly (cf. Green-Pedersen 1998).^J! Altogether, the Danish welfarestate not become any less generous under bourgeois reign. Cash

benefits were actually improved considerably. Still, there were important
cutbacks, especially during the first years of bourgeois rule.

Seen from the power resources model, spending measures are, however, only of limited interest (cf. Esping-Andersen 1990, 21). The more interesting question is whether the Danish welfare state moved away from its universal, flat rate and tax-financed structure.

Two surveys of social policy change in the period (Andersen 1997b, 8-20; Plovsing 1994, 58-82) both conclude that the Danish welfare state became more universal in character in the sense that benefits were made available to broader segments of the population, for instance by the abolishment of means testing. These changes also implied that benefits became more flat rate. To take a few examples, a universal, flat rate child allowance was introduced, as well as flat rate student grants for almost everyone above the age of 19. It is also worth noticing that the new disability pension act made benefits less dependent on medical criteria, just as the new social assistance act made benefits less dependent on the judgment of social workers (Plovsing 1994; Ploug & Kvist 1999). These benefits thus became more of a social right.

Looking at the way benefits are financed, the general picture is still one of tax financing. There is, however, some evidence of a shift away from tax financing in relation to unemployment and voluntary early retirement benefits, which are financed as one. Member and employer contributions to these schemes were raised a number of times, and employers were forced to pay for the first two days of unemployment (Nannestad 1999). In 1991, the years of membership of an unemployment fund required to be eligible for voluntary early retirement benefits were also raised from 10 to 20 years.^!2 These changes may be interpreted as an attempt to privatize the schemes (cf. Greve 1993, 71). However, the government still carries the marginal burden of the expenses and pays the major part of them,¹³ so the tendency towards a privatization of these schemes is only limited.

On the general level, there is thus little evidence that the Danish welfare
state lost its universal, flat rate and tax-financed character, just as it did not
become less generous. Actually, its seems that the social democratic structureof

ureofthe Danish welfare state was strengthened. Good examples of this are provided by a closer look at the two policy areas where, according to the power resources model, the struggle for new middle class support is most prominent, namely pensions and housing:

One researcher, arguing from the power resources model, has referred to the Danish pension system as a bifurcated system (Vesterø-Jensen 1985; cf. also Esping-Andersen 1985a,155-65; Esping-Andersen & Korpi 1986, 49-55). On the one hand, there is the national pension, which everyone is entitled to at the age of 67. On the other hand, an extensive market for individual pensions exists, together with various occupational pensions, especially for public employees. Thus, unlike in Sweden, an integrated superannuation scheme has never been established,¹⁴ neither in the private nor in the public sector (Overbye 1996). For a government wanting to increase role of the market, pensions seemed an obvious place to start, since private pensions were already widespread. The role of the market could be increased by weakening the national pension and /or by enhancing the attractiveness of private solutions, for instance through changes in the tax system. However, the bourgeois governments did just the opposite!

Right from the beginning, they declared that the national pensions would not be affected by the cuts they intended to make in cash benefits (Folketingets 1982-1983, 8) and later on, the level of benefits was raised several times. Thus, from 1982 to 1992, the real value of the national pension's basic amount increased by 17 percent for married couples and eight percent for singles, and the value of the pension supplement rose even more.¹⁵ Furthermore, the income testing of the supplement was eased in 1987,¹⁶ implying that the percentage of pensioners receiving the maximum supplement increased from 48 in 1987 to 67 in 1992 (Statistiske Efterretninger, Sikring og Retsvæsen, various years).

The only cutback in the national pension was the introduction, in 1984, of income testing of the basic amount. The income testing, however, only related to considerable earnings from work and only applied to pensioners aged 67-69. Thus, its impact was very limited: less than two percent of all pensioners were affected (ibid.). Furthermore, its is noteworthy that, unlike most other cuts in social expenditure introduced by the first bourgeois government, this one was supported by the Social Democrats. They even spoke of extending income testing to all pensioners, which they did after regaining power in 1993. The bourgeois government itself was less proud of the change in 1982, and the major bourgeois parties were against the extension the income test in 1993 (Folketingets Forhandlinger 1982-1983, 412-34;1993-1994, 2972-90).^I7

The bourgeois governments also diminished the attractiveness of private
pensions. First, and most importantly, a tax on the yield of private pensions
was introduced as of 1983, affecting all future pension savings. The revenue

from this tax was - and still is - quite considerable, and it has severe consequencesfor value of private pensions. Calculations (Henriksen et al. 1988, 202-3) show that a pension saver who was 30 years old in 1988, and whose savings would thus for the most part be affected by the tax, will end up with a compensation level more than 20 percent lower than that of a pension saver aged 55-59 in 1988. Second, the tax reform implemented in 1987 meant that marginal tax rates were lowered by two to seven percent (Matthiessen 1986). Thus, the value of the right to deduct contributions to pension schemes from taxable income was decreased. The maximum amount that could be paid into the popular capital pensions was also reduced(Hansen al. 1986). Altogether, the pension policy of the bourgeois governments may best be described as 'social democratic'¹⁸ in character. The national pension was strengthened and made more universal, while private pensions were made less attractive.

Through most of the post-war period, Danish housing policy has been characterized by an increasing role of the market. Most importantly, private home ownership has been stimulated through generous tax allowances (Esping-Andersen 1985a, 183-86; Esping-Andersen & Korpi 1986, 63-69). Again, one might expect that the bourgeois governments would try to increase the role of the market, or at least maintain the status quo. However, as was also the case with pension policy, the housing policy pursued by the bourgeois governments was 'social democratic' in character, although not as pronounced as in relation to the former.

The single most important event in Danish housing policy in this period was the limitation of the value of the tax allowance for home owners, i.e., the right to deduct interest on mortgages from taxable income (cf. Kjærgaard et al. 1986). This limitation was part of the 1987 tax reform and had severe consequences for Danish home owners. A limitation of this tax allowance was something the Social Democrats had long wanted, but not been able to carry through themselves (Esping-Andersen 1985a, 185-86).

As regards other aspects of Danish housing policy, such as rent allowances, control, and public housing, the general picture is one of continuity Harloe 1995, 484 98, 505-6), although certain changes did take place. Some cutbacks in rent allowances were carried through, but at the same time expenses grew considerably (Ministry of Finance 1992). At the beginning of the period, quotas for social housing were kept at a high level, but later cutbacks were implemented (Harloe 1995, 486-90). The rules concerning rent control remained largely unchanged. Altogether, changes in a residual direction seem to have been quite limited and, if the issue is the general Danish housing policy of that period, it must be characterized as social democratic, due to the 1987 tax reform.

Looking at the overall development of the Danish welfare state under
bourgeois reign, there is generally little evidence to support the hypothesis

that can be developed from the power resources model. It is true that the bourgeois governments did implement important cutbacks, especially at the beginning of the period, but benefits, especially cash benefits, were generally improved more significantly. As regards the content of the welfare state, the policy of the bourgeois governments was of a very social democratic character. This development is particularly pronounced in relation to the two areas highlighted by the power resources model, namely pensions and housing. Altogether, the Danish welfare state did not diverge any further from the Scandinavian model; actually it moved closer to it.

The Problems of the Power Resources Model

The logic thing to do next is to look for alternative theories which might explain the case. However, as such competing theories do not really exist, the remaining part of this article will be more of an open-ended discussion of what may explain the development of the Danish welfare state. One way to begin these deliberations is to ask why the power resources model seems incapable of accounting for the recent development of the Danish welfare state. An answer to this question would give some idea of where to look for alternative explanations.

Basically, the power resources model explains the development of welfare states in terms of conflicting class interests. According to the model, the welfare state is an instrument by which the working class may fight market influence on the distribution of wealth in society, a view that stems from the roots of the power resources model in Marxist class analysis. This view of the welfare state implies that, by definition, workers are pro-welfare state, the bourgeoisie is against it, while the new middle class is skeptical and will have to be tempted by luxury benefits. Thus, as Jackmann (1986) has pointed out, politics is class struggle by other means.

As argued by Pierson (1996, 150-51), this may long have been a fruitful way of analyzing welfare state development, but that no longer seems to be the case. According to Pierson, policy feedback from the welfare state has changed politics. Concurrently with its growth, the welfare state has attained a very strongly popular entrenchment and has created new supportiveinterest such as user groups and providers of public services. Pierson calls this 'the new politics of the welfare state' (1993; 1996,M5-51).Investigations the attitudes of the Danish electorate towards the welfare state support the idea of'popular entrenchment,' and draw a picture of welfare state attitudes somewhat different from the class perspective underlying the power resources model (Andersen 1993; 1997a; Nannestad & Paldam 1993). The Danish welfare state generally enjoys strong, but not uncritical, public support. Even among the voters of the bourgeois parties, a majority supports the welfare state (Nannestad & Paldam 1993,

305; Andersen 1988, 149), and the members of the new middle class are
those most in favor (Andersen 1993).

Feedback from the welfare state to politics is, of course, an important part of the power resources model. The working class must bribe the new middle class exactly because this strategy, if it succeeds, will create a feedback middle class support for the welfare state. The problem, however, is that such feedback mechanisms seem to be a lot stronger than expected by the power resources model and to work in ways not taken into account either. In Denmark, where the strategy of the working class succeeded only in part, the popular entrenchment of the welfare state is very strong, and, despite the prevalence of private pensions, the national pension is among the most popular welfare state schemes (Andersen 1991). In the UK and the USA, where the strategy of the working class failed, market solutions are prevalent. Therefore, these two countries, especially the USA, are seen by the power resources model as among the most obvious candidates for a backlash against the welfare state (Esping-Andersen 1990, 32-33). However, welfare state actually also proved to enjoy strong popular support in these two countries (Pierson 1994).

Alternative Explanations

Starting from the popular entrenchment of the welfare state and the idea, taken from rational choice theory (cf. Downs 1957), that governments vote maximize, or at least pay close attention to voters' reactions, we find that the situation of Danish governments, bourgeois or not, regarding the welfare can be described in the following way: Improvements of the welfare are attractive as a way of improving government popularity, while cutting back on the welfare state is risky business. During the years of bourgeois reign, the problems of Danish economy (Nannestad 1999) turned this situation into a real dilemma between economic constraints and popular On the one hand, the bourgeois governments needed a tight budgetary policy as part of the economic restoration, on the other hand, cutbacks in welfare state expenditures, which equal more than two thirds of overall public expenditure in Denmark, could easily imply severe electoral punishment. Thus, the bourgeois governments faced the unpleasant of having to adapt the very popular Danish welfare state to a less favorable economic situation.

One way out of this dilemma was to use what Pierson describes as strategiesof avoidance' (1994, 19-26, cf. also Weaver 1986), i.e., ways in which governments may limit the hazards of cutbacks. Especially the strategyof is interesting. Drawing on Arnold's (1990) idea of causal chains, the gist of this strategy is that governments, by the way they frame policy, make it difficult for voters to realize either that cutbacks are being

made or, if they do find out, that the government is the culprit. This idea is
similar to what Gray (1990) has called 'policy by stealth.'

The question then is if this line of thought makes some sense of the welfare of the bourgeois governments. It seems so. Even in an era of austerity, the Danish welfare state expanded further. Most of the expansion benefited broad groups of voters considered needy, such as pensioners and families with children. Furthermore, by making benefits more universal, the governments made sure that the improvements would benefit very broad segments of the electorate.

There are also at least three examples that the governments used the strategy of obfuscation or policy by stealth to improve public finances significantly. The first example is the introduction of the tax on the yield of private pensions, which was probably the single most significant contribution improving public finances. This is an example of such a strategy of obfuscation in two ways. First, this tax is a very complicated piece of legislation, to understand its exact consequences is difficult. Second, the negative consequences of the tax will not be felt by anyone until sometime in the distant future when someone gets a smaller pension. The benefits in terms of tax revenues, on the other hand, come to the government fairly quickly. The bourgeois parties might not have liked the tax and were also against it when in opposition (Vesterø-Jensen 1985, 252-58), but once in government and having to balance budgets, they saw the world in a different light. A tax that fairly quickly gave considerable revenues and whose negative were well hidden was simply too tempting.

The second example is how the bourgeois governments exported the blame for a tight fiscal policy to local governments (cf. Blom-Hansen & Pallesen 1998). Local governments in Denmark provide most welfare state services, equaling two thirds of the overall public consumption, whereas transfers are, by and large, in the hands of central government. Local governments enjoy some independence, notably due to their right to levy taxes, but central government has considerable influence on their economy, partly through its grants to local governments and partly through 'the budgetary cooperation': Each year, central and local government associations negotiate an agreement concerning the overall spending and tax levels of local governments. These agreements are not legally binding on the individual counties and municipalities, central government has some means of sanction if agreements are not kept (Albæk 1994; Blom-Hansen 1999). This institutional set-up gives central government an incentive to try to avoid the blame for a tight budgetary policy by keeping down the expenditure of local governments. In relation to local government expenditure, the central government can expect that at least part of the blame will be put on local politicians. Based on this, it is no surprise that Table 2 shows that expenses on transfers grew considerably, expenses on services grew only slightly.

The increases in member and employer contributions to unemployment and voluntary early retirement benefits provide the third example. It may be argued that they constitute a tendency towards a privatization of these benefits (Greve 1993, 71). The increases can, however, also be seen as hidden tax increases. The level of member and employer contributions is decided by parliament independently of the actual expenses on the two schemes, and the revenue accrues to the government like any other tax.¹⁹ Member contributions however, collected by the funds, and since membership is voluntary they do not look like a tax to the members. Employer contributions also affect very few voters personally.

Altogether, this line of thought, which is still far from an alternative theory to the power resources model, seems to account for important parts of the development of the Danish welfare state. It explains why the Danish welfare state grew and in a universal direction, just as it explains why transfers grew a lot more than services. In short, it seems that the bourgeois governments relied on a combination of improved benefits to large groups of voters and well-hidden measures to improve public finances. The best example of this is the 'social democratic' development of Danish pension policy, i.e., improvements of the national pension and the tax on the yield of private pensions.

However, there are also developments that run counter to this line of thought. The limitation of the tax allowance to home owners is difficult to understand in this way, since home owners, who constitute a large group of voters in Denmark, were affected very negatively and over a short span of time. The limitation was part of the 1987 tax reform which reduced tax rates but broadened the tax base (cf. Cox 1997), so perhaps the government hoped to be rewarded for the former.

During the first year of bourgeois reign, the government also implemented that were less well hidden to the voters, e.g., the lack of indexation and the introduction of a waiting day in sickness benefits. The reason why the government nevertheless implemented these cuts was probably Denmark was in dire economic straits, especially in the shape of a large budget deficit (Nannestad 1999). Furthermore, it seems that voter attitudes at that time gave the government some leeway in relation to the dilemma sketched above. A considerable 'crisis awareness' existed among the electorate, in the sense that it saw the need for measures to take the Danish economy away from the brink of the abyss (Petersen et al. 1996). Therefore, voters, even though they did not like the cutbacks, accepted them due to Denmark's economic troubles (Andersen 1988, 162-65).

To obtain a more comprehensive understanding of the development of the Danish welfare state from 1982 to 1993, the parliamentary situation of the bourgeois parties must also be included. All the bourgeois governments were minority and coalition governments and all of them, except the one

from 1984 to 1987, had to rely on the support of either all six bourgeois²⁰ parties or the Social Democrats to pass proposals in parliament. Furthermore,the governments accepted that in some situations, especiallyin to foreign policy, a majority outside the government passed legislation which the government then had to implement (Damgaard 1992).

The often weak parliamentary situation of the bourgeois governments influenced welfare state development in Denmark. An example is the introduction the universal and very generous student grant. This law was forced on the government by a majority consisting of the left wing parties and the Social Liberals, while the government itself voted against it. It was, however, also an exception in the sense that political compromising between the bourgeois governments and the opposition was generally the way the parliamentary situation influenced welfare state policy. One example is the 1988 budget agreement with the Social Democrats, which raised the level of most cash benefits considerably. There are also examples of the bourgeois governments suggesting cuts in social expenditure, in particular unemployment which they could not get through parliament. The Social Democrats vehemently opposed, especially during the first years of bourgeois almost all suggestions of cutbacks.

Based on this, one could argue that the Social Democrats, through political forced the governments to pursue a 'social democratic' policy, in direct opposition to their wishes. Such an argument would, however, the influence of the Social Democrats. Thus, it is worth noticing that the tax reform together with the universal child allowance was introduced between 1984—1987 when the bourgeois government was in a strong parliamentary position and did not need to compromise with the Social Democrats. Furthermore, the disagreements between government and opposition were generally about budgetary policy, not about the content the welfare state. For example, the government opposed the universal generous student grant, not because it was against universalism, but simply because it found the law too expensive (Folketingets Forhandlinger 1986-1987, 5803-23; 1987-1988, 8853-64). In short, it seems that the parliamentary of the bourgeois governments made it even more difficult for them to pursue a tight budget policy, but this did not force them to pursue a 'social democratic' welfare policy.

In summary, from the line of thought that any Danish government. bourgeois or not, has strong electoral incentives to work for a further expansion of the time-honored Danish welfare model, and if wishing to improve public finances, is well advised to do so by stealth, important aspects of the development of the Danish welfare state are easy to understand. factors, especially the parliamentary situation, have, however, also played a role.

Conclusion

When the first bourgeois government took office in 1982, one observer feared that this was the 'end of the idyll,' i.e., the Danish welfare state was in danger (Christiansen 1984). However, ten years of bourgeois rule expanded Danish welfare state further in a social democratic direction.²¹ Actually, one may argue that when the Social Democrats regained power in 1993, the Danish welfare state was in better shape than in 1982. It had been further expanded, its economic foundations were sounder, and public support intact.

This development is contrary to what could be expected based on the power resources model, according to which the Danish welfare state should have moved further in a residual direction. Furthermore, it seems that the incapacity of the power resources model in explaining the development of the Danish welfare state follows from its view of welfare state development as driven by conflicting class interests. This does not sufficiently take into account what Pierson (1996) has described as 'the new politics of the welfare state,' that is, the creation of the welfare state has changed politics as it has created politico-economic conditions favorable to its own survival and further expansion. An alternative comparative theory is not available, but it seems that one can understand important aspects of the development of the Danish welfare state from 1982 to 1993 by focusing on how the bourgeois tried to adapt the very popular Danish welfare state to a less favorable economic climate. They relied on a combination of improved benefits to large groups of voters and well-hidden measures to improve public finances.

Going back to the theoretical debate about the determinants of welfare state development, one of the central claims of the power resources model has been that politics, in terms of the party color and class base of the government, (Esping-Andersen & van Kersbergen 1992; Korpi 1989, 310-14; cf. also Castles & Pierson 1996). However, the development of the Danish welfare state during ten years of bourgeois reign gives evidence to the contrary. The Danish case seems a clear example of electorally driven policy convergence (cf. Jackman 1986). Thus, the Danish welfare state under reign, which may be considered a critical case for the power resources model, adds further evidence to Pierson's criticism of it. Furthermore, development of the Danish welfare state from 1982 to 1993 can, to some extent, be explained by a line of thought similar to Pierson's ideas of how governments act in light of the new politics of the welfare state. Therefore, this seems to be the direction where the study of welfare state development should be heading.

---

NOTES

1. The classics are Esping-Andersen (1985a; 1990): Stephens (1979); Korpi (1983). Cf. also O'Connor & Olsen (1998).

2. Criticism of the power resources model has, of course, always existed. For instance, the so-called 'state centered perspective' (e.g., Weir & Skocpol 1985) has challenged the explanatory power of the power resources model for a long time, but criticism of it has been rising through the 19905.

3. Sweden, Denmark and Norway.

4. All Scandinavian countries have had bourgeois governments for shorter periods of time, but since the power resources model emphasizes the cumulative hold of power by the social democrats (Shalev 1983, 330; Korpi 1991), a few years of bourgeois rule are not really suitable for testing the power resources model (cf. below).

5. The title of Esping-Andersen's book 'Politics against Markets' (1985a) is thus very apposite.

6. The introduction of economic democracy was suggested by Esping-Andersen as a way of revitalizing the Danish Social Democrats (1985a, 289-313).

7. Actually, Esping-Andersen suggests that the Danish Social Democrats have felt forced to implement bourgeois policy since the mid-1970s (1985b, 245).

8. The reason for not including figures for 1993 is that they are significantly affected by the policy of the Social Democratic-led government that took office in January 1993.

9. The level of inflation dropped in the following years, to 4.7 perscent in 1985 (Statistisk tiårsoversigt 1991), so the actual savings were less than what could have been expected.

10. The waiting dav was abolished again in 1986.

11. A caveat here is that the growth of the expenses on cash benefits was due to the rising number of unemployed and to demographic changes. This is partly true in the sense that if expenses on cash benefits are corrected for the rising number of unemployed and the rising number of elderly, they only rose by 22 percent in fixed prices. These corrections been made by calculating what expenses on unemployment insUrance, social assistance and the national pension would have been, had the number of unemployed elderly remained the same as in 1982. Even with these corrections, expenses on cash benefits, however, grew three times as much as expenses on services, so the main tendency is unchanged.

12. This gave the union controlled unemployment funds approximately 100,000 new members (calculations basedon Statistisk tiårsoversigt 1996).

13. As member and employer contributions are fixed, the exact percentage paid by central government varies with the level of unemployment. From 1982 to 1992, it varied between 61 percent (1987) and 85 percent (1982) (Statistiske Efterretninger: Nationalregnskab, finanser og betalingsbalance: Skatter og afgifter: Oversigt 1992; 1996). Employer contributions to social security schemes were abolished in 1988 and replaced by a 2.5 percent levy on the VAT base which is not included in the figures, so they should be interpreted cautiously.

14. In 1964, a contribution-financed scheme (ATP) was introduced. Benefits are, however. quite modest and depend on the number of hours worked, not on earnings, so this scheme is not really a superannuation scheme (cf. Overbye 1996).

15. All calculations are based on Jørgensen, various years; Statistisk tiårsoversigt. Statistiske efterretninger: Social sikring og retsvæsen.

16. Samspilsløsningen.

17. After regaining power, the Social Democrats also increased the relative importance of the income-tested supplement.

---

18. Despite the changes in the tax system, the prevalence of occupational pensions did increase considerably in the period, as the unions, led by the metal workers, negotiated pension schemes in the collective bargaining rounds of 1991. The unions had tried to achieve a pension reform up through the late 1980s but had failed, partly due to a lack of support from the government (Nielsen 1996).

19. On top of the contributions, members pay a small administration fee to their unemployment which varies from fund to fund.

20. The six bourgeois parties are the Liberals, the Conservatives, the Christians Peoples Party, the Center Democrats, the Social Liberals and the Progress Party. The second bourgeois government (1984-1987) could muster a majority with the support of only the Social Liberals and three members from Greenland and the Faeroe Islands (Damgaard 1992).

21. Some observers speak of cutbacks as a general description of the period (e.g., Andersen et al. 1996; Stephens 1995; Christiansen 1994).

ACKNOWLEDGMENTS

I'm grateful to Asbjørn Sonne Nørgaard, Peter Munk Christiansen, Robert Henry Cox, Kees
van Kersbergen, Christian Bertelsen and Peter Nannestad for many helpful comments to
earlier versions of this article.

REFERENCES

Albæk, E. 1994. 'Post 1970 'Budgetary Co-operation' between Central and Local Government
in Denmark,' Public Budgeting and Financial Management 6, 154-73.

Andersen, J. G. 1988. 'Vælgernes holdninger til den offentlige udgiftspolitik,' in Bentzon, K.
H.. ed., Fra vækst til omstilling. Copenhagen: Nyt Fra Samfundsvidenskaberne.

Andersen, J. G. 1991. 'Responsible Welfare Support in Denmark,' in Knudsen, T., ed., Welfare
in Denmark. Copenhagen: Ministry of Finance.

Andersen, J. G. 1993. 'Sources of Welfare States Support in Denmark: Self Interest or Way
of Life?,' in Hansen, E. J., et al., eds., Welfare Trends in the Scandinavian Countries. Armonk,
NY: M. E. Sharpe.

Andersen, J. G. 1997a. 'The Scandinavian Welfare Model in Crisis? Achievements and Problems
the Danish Welfare State in an Age of Unemployment and Low Growth,' Scandinavian
Studies 20, 1-31.

Andersen, J. G. 1997b. Beyond Retrenchment, Welfare Policies in Denmark in the 19905.
Department of Economics, Politics and Public Administration: University of Aalborg.

Andersen, K. V., et. al. 1996. 'Reorganizing the Danish Welfare State: 1982-1993. A Decade
of Conservative Rule,' Scandinavian Studies 68, 161-87.

Arnold, R. D. 1990. The Logic of Congressional Action. New Haven: Yale University Press.

Blom-Hansen, J. 1999. 'Avoiding the Joint-Decision Trap': Lessons from Intergovernmental
Relations in Scandinavia,' European Journal of Political Research, 35, 35-67.

Blom-Hansen, J. & Pallesen, T. 1998. The Fiscal Manipulation ofa Decentralised Public Sector:
Macroeconomic Policy in Denmark. Paper, Department of Political Science, University of
Aarhus.

Bonoli, G. 1997. 'Classifying Welfare States: a Two-dimension Approach,' Journal of Social
Policy 26, 351-72.

Castles, F. 1978. The Social Democratic Image ofSociety. London: Routledge & Kegan Paul.

Christiansen, N. F. 1984. "The End of the Idyll,' New Left Review 144, 5-32.

Christiansen, N. F. 1994. 'End of an Idyll,' in Anderson, P. & Camiller, P., eds., Mapping
the West European Left. London and New York: Verso.

Cox, R. H. 1993. The Development of the Dutch Welfare State. From Workers' Insurance to
Universal Entitlement. Pittsburgh, Pittsburgh University Press.

Cox, R. H. 1997. The Consequences of Welfare Retrenchment in Denmark,' Politics and
Society 25, 303-26.

Damgaard, E. 1992. 'Denmark, Experiments in Parliamentary Government,' in Damgaard,
E., ed., Parliamentary Change in the Nor die Countries. Oslo: Scandinavian University Press.

Downs, A. 1957. An Economic Theory of Democracy New York: Harper and Row.

Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity.

Esping-Andersen, G. & Korpi, W. 1986. 'From Poor Relief to Institutional Welfare States:
The Development of Scandinavian Social Policy,' in Eriksson, R. et al., eds., The Scandinavian
Model: Welfare Statesand Welfare Research. Armonk, N.Y.: M. E. Sharpe.

Esping-Andersen, G. & Van Kersbergen, K. 1992. 'Contemporary Research on Social
Democracy,' Annual Review of Sociology 18, 187-208.

Folketingets Forhandlinger, various years.

Gray, G. 1990. 'Social policy by Stealth,' Policy Options, 17-29.

Green-Pedersen, C. 1998. 'Udviklingen i udgifterne til overførsler og konsum i Danmark fra

Greve, B. 1993. 'Changes in Social Policy in Denmark During the Eighties,' in Greve, 8.,
ed., Future: Social Policy in Europe. Copenhagen: The Danish National Institute of Social
Research.

Hansen, H. S. et al. 1986. Skattereform og pensionsbeskatning. Copenhagen, A/S Skattekartoteket.

Harloe, M. 1995. The People's home. Social Renled Housing in Europe & America. Oxford:
Blackwell.

Henriksen, J. P., et al. 1988. Fordelingen af private pensioner. Copenhagen: Institutionen
Sociologi.

Jackman,R. 1986.'Elections and The DemocraticClassStruggle.' World Politics 39,123^16.

Kemeny, J. 1995. 'Theories of Power in The three Worlds of Welfare Capitalism: Journal of
European Social Policy 5, 87-96.

Kjærgaard, N., et al. 1986. Skattereformen og personbeskatning. Copenhagen: A/S Skattekartoteket.

Kloosterman, R. C. 1994. 'Three Worlds of Welfare Capitalism? The Welfare State and the
Post-Industrial Trajectory in the Netherlands after 1980,' West European Politics 17,
166-89.

Knudsen, R. 1987. Sociale kontantydelser 1976-1987. Copenhagen: The Danish National
Institute of Social Research.

Korpi, W. 1983. The Democratic Class Struggle. London: Routledge.

Korpi, W. 1989. 'Power, Politics, and State Autonomy in the Development of Social Citizenship,'
Sociological Review 54, 309-28.

Korpi, W. 1991. 'Political and Economic Explanations for Unemployment: A Cross National
and Long-term Analysis,' British Journal of Political Science 21,315-48.

Kuhn, T. 1970. The Structure of Scientific Revolutions, 2nd ed. Chicago: University of
Chicago Press.

Marklund, S. 1988. Paradise Lost? The Nordic Welfare State and the Recession 1975-1985.
Lund: Arkiv.

Matthiessen, L. 1986. Yderligere perspektiver på skattereformen. Working paper, Department
of Economics: University of Odense.

Ministry of Finance 1992. Udgiftsanalyser 1992. Copenhagen: Ministry of Finance.

Nannestad, P. 1999. 'Keep the Bumblebee Flying: Economic Policy in the Welfare State of
Denmark, 1973-1994,' in Albæk, E. et al., eds., Managing the Danish Welfare State under
Pressure: Towards a Theory ofthe Dilemmas of the Welfare State. Forthcoming.

Nannestad, P. & Paldam, M. 1993. 'The Demand for the Public Sector in the Rich Welfare
State of Denmark: Two Polis From 1990,' in Breton, A. et al., eds.. Preferences and Democracy.
Boston, and London: Kluwer Academic Publishers.

Nielsen, F. 1996. 'Danish Occupational Pensions in the 1980s: From Social Security to Political
in Shalev, M., ed., The Privatization of Social Policy: Occupational Welfare
and the Welfare State in America, Scandinavia, and Japan. London: Macmillan.

O'Connor, J. & Olsen, G. M. 1998. Power Resources Theory and the Welfare State. Toronto:
University of Toronto Press.

Orloff, A. 1993. 'Gender and the Social Rights of Citizenship,' American Sociological Review
58, 303-28.

Overbye, E. 1996. 'Pension Politics in the Nordic Countries: A Case Study,' International
Political Science Review 17, 57-90.

Petersen, E., et al. 1996. Danskernes trivsel, holdninger og selvansvarlighed under 'opsvinget.'
Department of Psychology: University of Aarhus.

Pierson, P. 1993. 'When effect becomes Cause. Policy Feedback and Political Change,' World
Politics 45, 595-628.

Pierson, P. 1994. Dismantling the Welfare State? Cambridge: Cambridge University Press.

Ploug, N. & Kvist, J. 1999. 'Continuity or Change in the Danish Social Security System,' in
Albæk, E. et al., eds., Managing the Danish Welfare State under Pressure: Towards a Theory
of the Dilemmas of the Welfare State. Forthcoming.

Plovsing, J. 1994. Socialpolitik. Copenhagen: Handelshøjskolens Forlag.

Shalev, M. 1983. 'The Social Democratic Model and Beyond. Two Generations of Comparative
State Research,' Comparative Social Research 6, 315-51.

Skatter og afgifter: Oversigt 1992 and 1996. Copenhagen: Danmarks Statistik.

Statistiske Efterretninger: Nationalregnskab, offentlige finanser og betalingsbalance, various
issues. Copenhagen: Danmarks Statistik.

Statistiske Efterretninger, Social Sikring og Retsvæsen, various issues. Copenhagen: Danmarks
Statistik.

Statistisk tiårsoversigt 1991 and 1996. Copenhagen: Danmarks Statistik.

Taylor-Gooby, P. 1996. 'Eurosclerosis in European Welfare States: Regime Theory and the
Dynamics of Change,' Policy and Politics 24, 109-23.

Van Kersbergen, K. 1995. Social Capitalism. A Study of Christian Democracy and the Welfare
State. London: Routledge.

Weaver, R. K. 1986. 'The Politics of Blame Avoidance,' Journal of Public Policy 6, 371-98.

Weir, M. & Skocpol, T. 1985. 'State Structures and the Possibilities for "Keynsian"
Responses to the Great Depression in Sweden, Britain and the United States,' in Evans, P.
8., ed., Bringing the State Back In. Cambridge: Cambridge University Press.