In a stimulating and important book, Politics against Markets: The Social Democratic Road to Power (reviewed in this journal 1986, No. 1,81-85), Gøsta Esping-Andersen outlines a 'theory of Social Democratic party formation and decomposition'. This theory is presented as the key explanation of the strength of Social Democratic parties as well as differing degrees of electoral success or failure of the three Scandinavian parties. The main question raised in this note concerns the adequacy and applicability of this theory in explaining the parties' electoral difficulties in the 1970's and early 1980's. The assessment of Esping- Andersen's argument, however, is prefaced by some comments on the nature of the difficulties of the Social Democratic parties in Denmark, Norway and Sweden, since these difficulties constitute the phenomena to be explained.

The Predicaments of the Parties

In analyzing the difficulties of the parties, Esping-Andersen examines a number of 'dimensions of party decomposition', such as aggregate electoral support, membership trends, the working-class vote, generational decomposition, etc. As in other diagnoses (Elvander 1979, 1980, Sainsbury 1983, 1984) the broad picture which emerges from his analysis is that the trend toward party decomposition appears most serious for the Danish Social Democrats (Socialdemokratiet, followed by the Norwegian party (Det norske arbeiderparti, DNA), and is least extensive for the Swedish party (Sveriges socialdemokratiska arbetareparti, SAP).

One problem with Esping-Andersen's analysis, however, is that it glosses over several worrisome trends for the Norwegian party and thus does not provide an adequate account of the DNA's difficulties. First, the extent of erosion in the party's share of the vote since the mid-1950's is actually very similar to that experienced by the SD, but the Norwegian party's polling strength remains superior because of its earlier stronger position. Second, in the early 1980's the drop in the working-class vote for the DNA began to approach that of the SD,¹ and young workers in particular have deserted the party. Third, the conclusion

that support for the DNA is actually growing among workers who are union members is open to question (Cf. Valen 1981, 113, Valen & Aardal 1983,88). Furthermore, the level of union membership among workers is lower in Norway compared to Sweden and Denmark, and in the 1981 election the proportion of non-unionized workers voting for the DNA fell considerably (Sainsbury 1985,6). Fourth, the capability of the DNA to mobilize young voters was only marginally better than that of the SD in the early 1980's (Cf. Glans 1983,127; Valen & Aardal 1983,57. Also Elvander 1980,311).

An additional weakness of the analysis is that the difficulties of the Norwegian Social Democrats appear to be largely a paler version of those of the Danish party when they are analyzed in terms of'dimensions of party decomposition'. A comparison these dimensions discloses similar difficulties, but it fails to identify clearly important differences between the parties' predicaments. This shortcoming reflects a major problem inherent in cross-national studies of a particular type of party - be it Social Democratic parties or Conservative parties. The problem consists of an emphasis on comparisons of the attributes of the parties under investigation the expense of comparisons of the contexts in which the parties operate. In order to understand the basic dissimilarities of the parties' dilemmas, it is necessary to look at the larger context: the voting trends for the other parties.

One of the most interesting dissimilarities in recent voting trends in Norway and Denmark is found in the electoral behavior of the youngest generation. Since the mid-1970's young voters in Norway have shifted to the right, and in Denmark they have moved to the left. From the 1973 election onwards younger Norwegian voters, as distinct from the previous generation, have shown a greater proclivity to favor the non-socialist parties, and a majority have voted non-socialist. Moreover, voting among young workers occurred on an unprecedented scale in both the 1977 and 1981 elections when around half of them voted for a non-socialist party. By contrast, a solid majority of younger Danish voters has voted socialist (i.e. for the Social Democrats or the parties to their left) in recent elections. Nor did the sharp drop in socialist voting among young workers in 1973 and 1975 continue in subsequent elections of the decade (Sainsbury 1985,8-9). Furthermore, in the early 1980's young Danish voters shunned the Progressives, the Conservatives, and Agrarian Liberals - the parties on the right of the political spectrum (Sauerberg 1982,52). Thus, contrary to Esping-Andersen's description, in the early 1980's it was young Norwegian voters and not their Danish counterparts displayed the sharpest polarization in party preferences.² Young Norwegian gravitated toward the Conservative and Progressive parties, at the same time as a much larger percentage of these voters supported the left socialists than did older voters.

A further dissimilarity between the two countries is that Norway experienced
a polarization in electoral choice in the late 1970's as the polling strength of the
centrist parties - especially the Liberals - waned. This 'polarization' was

accompanied by straight conversions of DNA voters to the Conservatives and a
substantial change in the pattern of second party preferences of DNA and
Conservative voters (Valen 1981, 215).

No similar phenomenon has emerged in recent Danish elections. Admittedly, the SD lost votes to Mogens Glistrup's Progress party in 1973, and the Progressives attracted working-class voters during the rest of the decade. In the 1980's, however, the SD's losses to the non-socialists have primarily gone to the centrist parties. For example, when the SD lost considerable ground in the 1981 election, its losses in terms of party switches were voters who changed to the small centrist parties and the left socialist parties (Borre 1982). As distinct from the Norwegian case, the recent and belated strong showing of the Danish Conservatives in the 1984 election occurred at the expense of the Progressives and the Conservatives' coalition partners - and not through inroads in the socialist bloc.

To sum up, the differences in the predicaments of the SD and the DNA are not merely a matter of degree, as a comparison of indices of party decomposition might lead one to believe. The Norwegian party's difficulties were associated with a general rightward shift in electoral choice, a polarization between the two largest parties, and straight conversions among DNA voters to the Conservatives. The Danish party's problems stem from a fragmentation rather than a decline in the socialist vote in the electorate. Instead the SD's share of the left vote has declined, and this decline has been accompanied by a drain in support to the right. Equally important, to the extent that the governmental status of the Social Democratic parties is dependent upon the strength of the socialist vote and not solely the party vote, divergent trends in socialist voting have far-reaching implications. short, larger patterns of party competition and not just erosion in the party strength of the Social Democrats have to be taken into account in an assessment of the parties' difficulties.

Explaining the Parties' Difficulties

Esping-Andersen's explanatory framework is rooted in the theoretical debate concerning the successes and failures of working-class politics and Social Democratic parties. In the opening, chapter he clarifies his position in the controversy and presents the 'theory of Social Democratic party formation and decomposition' based on three components: class structure, class formation, and class alliances. Class structure is viewed as setting certain limits and shaping options - but not as the determining factor of party strength. Instead state policy plays a crucial role in effecting class formation and coalitions.

Class formation, as distinct from class structure, involves the process by which a class becomes unified in the pursuit of collective political action. State policy which modifies the market (hence the title Politics against Markets) is vital to Social Democratic class formation. Two types of policy are especially important

in promoting class solidarity: 1) welfare reforms which are universalistic and provide generous benefits and 2) government controls over the business cycle securing a high level of employment. In fact, Esping-Andersen views these state policies as preconditions for the success and survival of Social Democratic parties, although his position on welfare reforms is actually quite equivocal (pp. 34-35). He further maintains that variations in Social Democratic reforms, specifically in the areas of welfare, housing and economic policies, explain why party decomposition varies so dramatically between the three Scandinavian parties.

With respect to class alliances, Esping-Andersen argues that the Social Democrats' commitment to democratic elections necessitates class coalitions in order to win a majority and gain power. In this way his analytical framework retains an emphasis on class structural change, but again state policy is stressed as the prime instrument in forging class alliances. His main point is that as the former alliance between the workers and the rural petite bourgeoisie dissolves, the survival of the Social Democratic parties depends upon their ability to achieve a new alliance with white-collar employees (Cf. Stephens, Chapter 2).

Perhaps the most novel aspect of this analytical framework is its emphasis on policy, although other writers have also focused on the importance of differing policy responses (Martin 1975a & b) and policy performance (Whiteley 1983). In any event, Esping-Andersen's major contribution, as he notes (p. 4), is a systematic attempt to explain trends in Social Democratic support as the outcomes of certain policies.

Despite the fact that a systematic examination of the impact of policies on party strength is a very fruitful enterprise, and Esping-Andersen's comparative policy analyses are one of the most valuable features of his book, it is justified to ask how adequate his thesis is in explaining party strength and party decomposition. doubts of a more general nature suggest themselves. The first is that his explanation rests essentially upon a single factor - the policy variable - even if it encompasses several policies. The second doubt stems from the fact that state policy is virtually a hydra, and there is a tendency to incorporate other explanations involving policies, which are incidental to the 'theory' outlined in the opening chapter.

Empirically, there are also problems which come to the fore in the discussion of party decomposition. His theory of party decomposition centers on the impact of policy on class politics in a very specific sense - the extent to which policies introduce new divisions in the class structure and in particular produce schisms within the working class, resulting in the erosion of party support. Three policies are of major importance in his explanation. He argues, firstly, that 'the advanced process of Social Democratic party decomposition, especially in Denmark, can be attributed to a pervasively liberalistic welfare state that enhances social stratification and cleavages cutting across class lines' (p. 149). Secondly, a

housing policy which creates divisions between renters and homeowners will
contribute to decomposition. Thirdly, decomposition will be positively related to
weak policy performance in controlling the business cycle (pp. 244-8).

Utilizing election survey data Esping-Andersen attempts to determine the degree to which opinions on these policies are a source of division and fragmentation in the electorate and among voters in the Social Democrats' traditional and potential constituencies. He also investigates the erosion in favorable attitudes towards Social Democratic policies. For the different policies he describes the patterns of voter opinion and cleavage in the three countries.

A major empirical problem here is that the cross-national variations in voters' attitudes presented in Chapter 8 and attributed to differences in Social Democratic policies are equally likely to be the product of dissimilar questions. This likelihood is especially pronounced with respect to welfare policy. The question used to measure Danish welfare state attitudes differs from the question used in the Norwegian and Swedish surveys, and it clearly inflates anti-welfare state sentiments by tapping attitudes toward abuse of welfare benefits.³ On the basis of this question he concludes that negative attitudes were extraordinarily high, and a 'special feature of Denmark is the especially low support for the welfare state among workers' (p.254). Other questions probing Danish voters' views on social reforms do not disclose the extreme anti-welfarism presented by Esping-Andersen, they do reveal an ebb in support of reforms in the mid-1970'5. Even then, however, Danish voters' priorities in public spending indicate overwhelming, approval of pensions, health services, and homes for the elderly (Esping-Andersen 413, Nielsen 1976, 144-5). Nor does an analysis using these other questions confirm Esping-Andersen's conclusion of much lower support among workers. Instead workers were generally more supportive of the welfare state than other occupational groups (Andersen 1982, 193, 202-3).

It also needs to be stressed that similarities in Scandinavian welfare reforms are far greater than differences, as Esping-Andersen's overview shows, and the divisive nature of Danish welfare policies is largely a matter of conjecture. Variations in welfare policy cannot account for the SD's debacle in the 1973 election.

A second major empirical problem is that, although his explanation appears to have some substantiation in the Danish case, it offers few clues in understanding trends toward Social Democratic decomposition in Norway. If Esping-Andersenportrays Danish welfare state as a cause of the SD's electoral problems and polarized protest, he describes the welfare state in Norway as promoting consensus and powerful solidaristic commitment (p. 252). Nor can Norwegian housing policy account for the trends toward party decomposition, at least in the way posited by his theory.⁴ As distinct from Danish housing policy, the Norwegianpolicy not produced a cleavage between renters and homeowners (p. 269). Finally,the DNA's record in regulating the business cycle and maintaining

full employment is impressive. Furthermore, Norwegian policy has been characterized by a larger measure of economic planning and controls on investmentsthan the other two countries. In Esping-Andersen's theory, governmentcontrol the economy is a major precondition for the long-range survival of Social Democratic parties, and control of the investment function is critical. It is, then, indeed ironic that Norway is the only country of the three in which an erosion in favorable attitudes toward state intervention occurred in the electorate including workers and DNA voters - from the late 1960's through the end ofthe 1970Y

The Norwegian case thus deviates sharply from the theory. Recognizing the lack of fit, Esping-Andersen argues that the DNA is not decomposing. Whether the DNA's problems qualify as decomposition is a moot question. As shown in the earlier discussion on the predicaments of the parties, the DNA faces serious difficulties, more serious than indicated by Esping-Andersen, and his theory fails to provide an explanation. Moreover, the empirical realities of the Norwegian case contradict his theoretical assumptions and predictions. In conclusion, Esping-Andersen's 'theory of Social Democratic party formation and decomposition' appears to be based on deductive reasoning and to have wide application. Upon closer scrutiny, the section of his theory dealing with party decomposition seems inspired by the Danish experience and has limited relevance when applied to Norway.

---

NOTES

1. Comparisons of the working-class vote for the three Social Democratic parties can be quite misleading because of different definitions of the working class. One problem in comparing the working-class vote of the SD and the DNA, and this difficulty is reflected in Esping-Andersen's comparison (Tables 4.3 and 4.4), is that the Danish figures frequently exclude pensioners with former working-class occupations, whereas they are included in the Norwegian percentages. The result of this inequivalency is that it deflates the working-class vote of the SD and boosts that of the DNA at least in the 1970's and early 1980's. For a more detailed discussion of these problems of comparability, see Sainsbury 1986.

2. Part of the difficulty lies in the fact that Esping-Andersen's discussion of the SD's difficulties bears strong traces of the analyses made in the mid-1970's during the party's darkest hour. This feature is most pronounced in his analysis of generational trends in the 1970s which relies exclusively on 1975 data in the Danish case.

3. The 1979 Danish election survey contains the question on abuse of welfare benefits used by Esping-Andersen as well as a question on support of welfare reforms (similar to the Norwegian and Swedish questions). The survey thus provides an excellent opportunity to compare responses to the two questions within the same sample. On the basis of the question on abuse of social benefits, 69 per cent of the Danes would be classified as displaying anti-welfare state attitudes and an unimpressive 20 per cent as supporting extension of the welfare state. (As Esping-Andersen has done, 'Don't know' responses have been excluded.) By contrast, the responses to the question on welfare reforms indicate that 31 per cent preferred cut-backs and 55 per cent favored existing reforms (Borre et al. 1983, 167, 176-7). Nor is this pattern of responses to the two questions peculiar to the Danes. It also exists for the Norwegians whose solidarity with the welfare state, according to Esping-Andersen, is overwhelming. In a 1973 Gallup poll Norwegians were asked about the extent of abuse of social benefits. Around 70 per cent believed that abuse was quite widespread (Kolberg & Viken 1978, 22), and thus using Esping-Andersen's procedure for the Danes, would qualify as anti-welfare state.

---

3. The 1979 Danish election survey contains the question on abuse of welfare benefits used by Esping-Andersen as well as a question on support of welfare reforms (similar to the Norwegian and Swedish questions). The survey thus provides an excellent opportunity to compare responses to the two questions within the same sample. On the basis of the question on abuse of social benefits, 69 per cent of the Danes would be classified as displaying anti-welfare state attitudes and an unimpressive 20 per cent as supporting extension of the welfare state. (As Esping-Andersen has done, 'Don't know' responses have been excluded.) By contrast, the responses to the question on welfare reforms indicate that 31 per cent preferred cut-backs and 55 per cent favored existing reforms (Borre et al. 1983, 167, 176-7). Nor is this pattern of responses to the two questions peculiar to the Danes. It also exists for the Norwegians whose solidarity with the welfare state, according to Esping-Andersen, is overwhelming. In a 1973 Gallup poll Norwegians were asked about the extent of abuse of social benefits. Around 70 per cent believed that abuse was quite widespread (Kolberg & Viken 1978, 22), and thus using Esping-Andersen's procedure for the Danes, would qualify as anti-welfare state.

4. Elvander mentions housing policy as an important issue contributing to the defeat of the DNA in the 1965 election, and he cites housing policy as a problem for the party in the Oslo area during the 1970's (p. 276, p. 286). In other words, policy may influence the party vote in additional ways than that hypothesized by Esping-Andersen.

REFERENCES

Andersen, J. G. 1982. 'Den folkelige tilslutning til socialpolitikken - en krise for velfærdsstaten?',
pp. 175-209 in D. Anckar, E. Damgaard & H. Valen, eds.,Partier, ideologier, väljare. Åbo: Abo
Akademi.

Borre, O. et al. 1983. Efter vælgerskredet. Aarhus: Politica.

Borre, O. 1982. 'The Danish Parliamentary Election of 1981', Electoral Studies 1, 250-55.

Elvander, N. 1979. Scandinavian Social Democracy: Its Strength and Weakness. Stockholm:
Almqvist and Wiksell International.

Esping-Andersen, G. 1985. Politics against Markets: The Social Democratic Road to Power. Princeton:
University Press.

Esping-Andersen, G. 1980. Social Class, Social Democracy and State Policy. Copenhagen: New
Social Science Monographs.

Glans. I. 1983. 'Valvindar välter teoribyggnader: Förändringar i klassers och kohorters partival i
Danmark och Sverige', Statsvetenskaplig tidskrift 86, 117-140.

Kolberg, J. E. & Viken, A. 1978. Trygdehetsens struktur', Tidsskrift for samfunnsforskning 19,17-41.

Martin, A. 1975a. 'Is Democratic Control of Capitalist Economies Possible?', pp. 13-56 in L.
Lindberg, et al., eds., Stress and Contradiction in Modern Capitalism. Lexington: Heath.

Martin, A. 1975b. 'Labor Movement Parties and Inflation: Contrasting Responses in Britain and
Sweden', Polity 7, 427-51.

Nielsen, H. J. 1976. 'Politiske meninger og vælgeradfærd', in O. Borre et al., Vælgere i 70'erne.
Copenhagen: Akademisk Forlag.

Sainsbury, D. 1986. 'Class Voting and Left Voting in Scandinavia: The Impact of Different Operationalizations
the Working Class', Paper, ECPR Workshop on Trends in Voting Behaviour.
Gothenburg.

Sainsbury, D. 1985. 'The Electoral Difficulties of the Scandinavian Social Democrats in the 1970'5:
The Social Bases of the Parties and Structural Explanations of Party Decline', Comparative
Politics 18, 1-19.

Sainsbury, D. 1984. 'Scandinavian Party Politics Re-examined: Social Democracy in Decline?'
West European Politics 7, 67-102.

Sainsbury, D. 1983. 'Functional Hypotheses of Party Decline: The Case of the Scandinavian Social
Democratic Parties', Scandinavian Political Studies 6 (New Series), 241-260.

Sauerberg, S. 1982. 'Den norske højrebølge - set fra Danmark', Økonomi og Politik 56, 40-54.

Valen, H. & Aardal, B. 1983. Et valg i perspektiv. Oslo: Central Bureau of Statistics of Norway.