¹ I employ the notions of social democratic, socialist and labour party in a generic and interchangeable way because the three labels do not indicate clear-cut programmatic or strategic differences among European moderate leftist parties.

² This article reports results from a broader project in which the non-class theory of social democratic party performance is laid out and empirically tested in detail. See Kitschelt, Herbert, The Transformation of Social Democracy (New York: Cambridge University Press, to appear).Google Scholar

³ Przeworski, Adam and Sprague, John, Paper Stones: A History of Electoral Socialism (Chicago: University of Chicago Press, 1986).Google Scholar

⁴ The critical importance of work generating a productive surplus for the concept of class has been reaffirmed by Poulantzas, Nicos, Classes in Advanced Capitalism (London: New Left Books, 1978), pp. 216 and 221Google Scholar, and Przeworski, Adam, Capitalism and Social Democracy (Cambridge: Cambridge University Press, 1985), pp. 84–5.CrossRefGoogle Scholar For a critique of Marxist conceptions of productive labour, see Giddens, Anthony, The Class Structure of Advanced Societies (London: Hutchinson, 1973), pp. 95–8.Google Scholar

⁵ There are some unproductive manual workers, e.g. those employed in the public sector. While the manufacturing and resource extraction sectors have the bulk of all productive wage earners, these sectors also include some unproductive white-collar employees, whereas the tertiary sectors also have a small proportion of productive workers. Nevertheless, manual labour and manufacturing sector employment are useful indirect tracers of the working class in the narrow sense of the term.

⁶ See Wright, Erik Olin, Classes (London: Verso, 1985)Google Scholar and, for critical discussion of his work, Marshall, Gordon, Rose, David, Newby, Howard and Vogler, Carolyn, Social Classes in Modern Britain (London: Unwin Hyman, 1989)Google Scholar, especially chaps 3 and 9, and Wright, Erik Olin, ed., The Debate on Classes (London: Verso, 1989).Google Scholar

⁷ Katznelson, Ira, ‘Working Class Formation: Constructing Cases and Comparisons’, in Katznelson, Ira and Zolberg, Aristide, eds, Working Class Formation (Princeton, NJ: Princeton University Press), pp. 3–41 at pp. 14–22.Google Scholar

⁸ See Lockwood, David, The Blackcoated Worker, A Study in Class Consciousness, 2nd edn (Oxford: Clarendon Press, 1989)Google Scholar, originally published in 1958.

⁹ See especially Beck, Ulrich, ‘Jenseits von Stand und Klasse? Soziale Ungleichheiten, gesellschaftliche Individualisierungsprozesse und die Entstehung neuer sozialer Formationen und Identitäten’, Soziale Welt, 34 (1983), 35–74Google Scholar; and Berger, Peter A., Entstrukturierte Klassengesellschaft? (Opladen: Westdeutscher Verlag, 1986).CrossRefGoogle Scholar

¹⁰ See especially Bourdieu, Pierre, Outline of a Theory of Practice (Cambridge: Cambridge University Press, 1977)CrossRefGoogle Scholar; and The Distinction: A Social Critique of the Judgment of Taste (Cambridge, Mass.: Harvard University Press, 1984).Google Scholar

¹¹ In this respect, the literature on ‘new’ social movements has made some theoretical and empirical advances. See especially Offe, Claus, ‘New Social Movements: Changing Boundaries of the Political’, Social Research, 52 (1985), 817–68Google Scholar; and Dalton, Russell J. and Kuechler, Manfred, eds, Challenging the Political Order: New Social and Political Movements in Western Democracies (Cambridge: Polity Press, 1990).Google Scholar

¹² I am here selectively summarizing the findings of Lane, Robert, The Market Experience (New York: Cambridge University Press, 1991).CrossRefGoogle Scholar

¹³ See, for example, Swenson, Peter, ‘Labor and the Limits of the Welfare State: The Politics of Intraclass Conflict and Cross-Class Alliance in Sweden and West Germany’, Comparative Politics, 23 (1991), 379–99.CrossRefGoogle Scholar

¹⁴ Throughout his study, The Market Experience, Lane makes the point that whatever beneficial or detrimental qualities are attributed to the market place, there is little link between market success and human happiness.

¹⁵ This has been demonstrated in a recent Swedish study of social stratification by Petersson, Olof, Westholm, Anders and Blomberg, Goran, Medborgarnas Makt (Stockholm: Carlssons, 1989).Google Scholar

¹⁶ As Marshall et al., Social Class in Modern Britain, p. 139Google Scholar and chap. 7, show, this problem also besets Wright's efforts to categorize authority relations by individuals' ‘control of organizational resources’.

¹⁷ It is therefore somewhat misleading to suppose that ‘affluence’ stimulates post-materialist, libertarian values. Much of the linkage between affluence and anti-authoritarian, participatory political demands is mediated via education and occupational task structure.

¹⁸ Berger, Peter L., The Capitalist Revolution (New York: Basic Books, 1991), pp. 57–68.Google Scholar

¹⁹ Kirkpatrick, Jeane, The New Presidential Elite (New York: Russell Sage, 1976).Google Scholar

²⁰ See Kern, Horst and Schumann, Michael, Das Ende der Arbeitsteilung? (Miinchen: Beck, 1984)Google Scholar and Piore, Michael J. and Sabel, Charles, The Second Industrial Divide (New York: Basic Books, 1984).Google Scholar

²¹ See Dunleavy, Patrick and Husbands, Christopher T., British Democracy at the Crossroads: Voting and Party Competition in the 1980s (London: George Allen & Unwin, 1985).Google Scholar

²² See Scarbrough, Elinor, ‘Review Article: the British Electorate Twenty Years On – Electoral Change and Election Surveys’, British Journal of Political Science, 17 (1987), 219–46, at pp. 229–30.CrossRefGoogle Scholar

²³ See Offe, Claus, Disorganized Capitalism (Cambridge, Mass.: MIT Press, 1985), chap. 5.Google Scholar

²⁴ Thus, empirical analysis shows, for example, that in Britain the distribution of voters' preferences is still more centred around economic left-right issues than in some other European countries.

²⁵ A refinement of this simple model with hypotheses predicting variance across countries can be found in Kitschelt, , The Transformation of European Social Democracy, chap. 4.Google Scholar

²⁶ Of course, many citizens still cast their vote based on party identification. All that matters for party competition, however, are voters who are available to competing appeals in the electoral market place.

²⁷ One may think of farmers or other rent-seeking economic status groups. Riker's voluntarist theory of ‘heresthetics’, according to which parties that consistently lose on a given dominant issue dimension will ‘create’ a new dimension, just as the Republicans in the middle of the nineteenth century ‘created’ the political issue of slavery to combat Democratic hegemony, ignores the requirement that there must be a structural propensity of the electorate to treat an issue as salient. See Riker, William H., The Art of Political Manipulation (New Haven, Conn.: Yale University Press, 1986).Google Scholar Politicians cannot create political dimensions out of thin air. Regardless of the form taken by the party system, studies estimating the dimensionality of European party appeals usually find only an economic left-right and a new politics dimension, the latter closely linked to my conception of libertarian versus authoritarian politics. See Budge, Ian, Robertson, David and Hearl, Derek, eds, Ideology, Strategy and Party Change: Spatial Analysis of Post-War Elections (Cambridge: Cambridge University Press, 1987).CrossRefGoogle Scholar

²⁸ For a detailed analysis of this proposition, see Cox, Gary W., ‘Centripetal and Centrifugal Incentives in Electoral Systems’, American Journal of Political Science, 34 (1990), 903–35CrossRefGoogle Scholar and Shepsle, Kenneth A., Models of Multiparty Electoral Competition (Chur: Harwood Academic Publishers, 1991).Google Scholar

²⁹ All these conjectures depend on certain further assumptions I cannot detail here. For example, voters are held to act sincerely rather than strategically.

³⁰ For an analysis of the interaction between electoral competition and government formation, see Laver, Michael, ‘Party Competition and Party System Change: The Interaction of Coalition Bargaining and Electoral Competition’, Journal of Theoretical Politics, 1 (1989), 301–24.CrossRefGoogle Scholar

³¹ This argument originates in the writings of Michels, Robert, Political Parties (New York: Free Press, 1962)Google Scholar, but then was popularized as the ‘law of curvilinear disparity’ by McKenzie, Robert, British Political Parties (London: Heinemann, 1955)Google Scholar and May, John D., ‘Opinion Structure and Political Parties: The Special Law of Curvilinear Disparity’, Political Studies, 21 (1973), 135–51.CrossRefGoogle Scholar For a critical discussion of this literature, see Kitschelt, Herbert, ‘The Internal Politics of Parties: The Law of Curvilinear Disparity Reconsidered’, Political Studies, 37 (1989), 400–21.CrossRefGoogle Scholar

³² Empirical support for this proposition is provided in Kitschelt, , The Transformation of European Social Democracy, chap. V.Google Scholar

³³ A detailed examination of this proposition can also be found in Kitschelt, , The Transformation of European Social Democracy, chap. VI.Google Scholar

³⁴ The difficulties of reconstructing class structures are described in Przeworski, and Sprague, , Paper Stones, pp. 194–200.Google Scholar

³⁵ See Flora, Peter, Kraus, Franz and Pfennig, Winfried, State, Economy, and Society in Western Europe, 1918–1975, Vol. 2 (Frankfurt, Main: Campus Verlag, 1987).Google Scholar

³⁶ Furthermore, I have checked data on the sectoral composition of each country's economy because the working-class contracts with an erosion of the industrial sector's share of total employment. Since the interpretation of my findings is similar to those derived from Table 1, I have not reproduced this analysis here.

³⁷ Definitions were changed in 1988. My comparison therefore includes only surveys taken between 1976 and 1987.

³⁸ The working class includes only respondents actually employed in blue-collar jobs, not individuals inactive in labour markets, even though they may be dependent on working-class bread-winners.

³⁹ See Rose, Richard and McAllister, Ian, Voters Begin to Choose: From Closed-Class to Open Election in Britain (Beverly Hills, Calif.: Sage 1986)Google Scholar; and Crewe, Ivor, ‘Labor Force Changes, Working-Class Decline and the Labour Vote: Social and Electoral Issues in Postwar Britain’, in Piven, Frances Fox, ed., Labour Parties in Post-Industrial Societies (New York: Oxford University Press, 1991), pp. 20–46.Google Scholar

⁴⁰ Holmberg, Sören and Gilljam, Mikael, Väljare och Val i Sverige (Stockholm: Bonnier Facta Bokförlag, 1987), p. 186.Google Scholar

⁴¹ The varying shift from class determination of the vote to issue voting and ideological dispositions is documented in the contributions to Franklin, Mark N., Mackie, Thomas T., Valen, Henry et al. , Electoral Change: Responses to Evolving Social and Attitudinal Structures in Western Countries (Cambridge: Cambridge University Press, 1992).Google Scholar

⁴² For Austria, see Plasser, Fritz, Ulram, Peter A. and Grausgruber, Alfred, ‘Vom Ende der Lagerparteien: Perspektivenwechsel in der osterreichischen Parteien- und Wahlforschung’, Österreichische Zeitschrift für Politikwissenschaft, 16 (1987), 241–58.Google Scholar For Britain, see Särlvik, Bo and Crewe, Ivor, Decade of Dealignment (Cambridge: Cambridge University Press, 1983), p. 103Google Scholar; and Crewe, , ‘Labor Force Changes, Working Class Decline, and the Labour Vote’.Google Scholar For Sweden, compare Holmberg, and Gilljam, , Väljare och Val i Sverige, pp. 223 and 292.Google Scholar

⁴³ See Lewis-Beck, Michael, Economics and Elections: The Major Western Democracies (Ann Arbor: University of Michigan Press, 1988).Google Scholar

⁴⁴ See Irwin, G. A. and van Holsteyn, J. J. M., ‘Towards a More Open Model of Competition’, West European Politics, 12 (1989), 112–38CrossRefGoogle Scholar; and Van Deth, Jan W. and Geurts, Peter A. T. M., ‘Value Orientation, Left-Right Placement and Voting’, European Journal of Political Research, 17 (1989), 17–34.CrossRefGoogle Scholar

⁴⁵ Lancaster, Thomas D. and Lewis-Beck, Michael S., ‘The Spanish Voter: Tradition, Economics, Ideology’, Journal of Politics, 48 (1986), 648–74.CrossRefGoogle Scholar

⁴⁶ Mackie, Tom and Franklin, Mark, ‘Electoral Change and Social Change’Google Scholar, in Franklin, , Mackie, , Valen, et al. , Electoral Change, pp. 33–57, at pp. 45–6.Google Scholar

⁴⁷ A more detailed review can be found in King, Desmond and Wickham-Jones, Mark, ‘Review Article: Social Democracy and Rational Workers’, British Journal of Political Science, 20 (1990), 387–413.CrossRefGoogle Scholar

⁴⁸ See especially the book's melancholic closing paragraph: ‘And they [socialist parties] do not have much of a choice: their organizational links, their ideological commitments, their daily habits, and their political projects tie them to their working-class roots. They are thus more likely to turn inward, to their working-class base, and suffer the electoral consequences… Ashamed of looking too far forward, mortally afraid of appearing irresponsible, left-wing political parties view socialism with embarrassment. Thus the era of electoral socialism may be over.’ (Przeworski, and Sprague, , Paper Stones, p. 185.)Google Scholar

⁴⁹ See Panitch, Leo, ‘Review of Paper Stones by Adam Przeworski and John Sprague’, American Journal of Sociology, 93 (1987), 490–3CrossRefGoogle Scholar; Burowoy, Michael, ‘Marxism without Micro-Foundations’, Socialist Review, 19 (1989), 57–86, at p. 64.Google Scholar

⁵⁰ Przeworski, and Sprague, , Paper Stones, p. 54.Google Scholar

⁵¹ Przeworski, , Capitalism and Social DemocracyGoogle Scholar, chaps 5 and 6. King, and Wickham-Jones, , ‘Review Article: Social Democracy and Rational Workers’, p. 394Google Scholar, have argued that Przeworski and Sprague's reconstruction of the socialist parties' strategic dilemma is inconsistent with Przeworski and Wallerstein's theory that the working class refrains from pursuing socialism. Whereas the former presupposes that workers want socialism, the latter shows that they do not. But Przeworski and Sprague never argue that workers want socialism. They claim only that workers are attracted to socialist parties by the prospect of particular ‘appeals’ and ‘visions’ that are never identified with socialism or anything else. Socialism presupposes a revolutionary party, abandoning the democratic majority principle, outside the electoral arena. For this reason, it is not an argument against Przeworski and Sprague to point out that workers abandon left-wing parties when their socialist appeals radicalize. What saves Przeworski and Sprague from King and Wickham-Jones' critique is their vagueness about left-wing parties' working-class ‘appeals’.

⁵² See Tingsten, Herbert, The Swedish Social Democrats: Their Ideological Development (Totowa: Bedminster Press, 1973)Google Scholar; Sainsbury, Diane, Swedish Social Democratic Ideology and Electoral Politics, 1944–1948 (Stockholm: Almqvist & Wiksell, 1980)Google Scholar; and Tilton, Timothy, The Political Theory of Swedish Social Democracy: Through the Welfare State to Socialism (Oxford: Clarendon Press, 1990).Google Scholar

⁵³ Przeworski, and Sprague, , Paper Stones, p. 72.Google Scholar

⁵⁴ Przeworski, and Sprague, , Paper Stones, p. 35.Google Scholar

⁵⁵ This hypothesis is consistent with the electoral composition of many radical revolutionary parties, such as those in Italy and France during the 1970s.

⁵⁶ Przeworski, and Sprague, , Paper Stones, p. 9.Google Scholar

⁵⁷ Przeworski, and Sprague, , Paper Stones, pp. 115–19Google Scholar, argue that the encapsulation of the working class prevented Scandinavian social democracy from pursuing an electorally optimal cross-class strategy mandated by a small trade-off between workers and allies. But if institutional constraints override opportunities and limits set on socialist party strategy, why then construct the conditions of socialist party success around trade-off parameters at all?

⁵⁸ See Lichbach, Mark, ‘Optimal Electoral Strategies for Socialist Parties: Does Social Class Matter to Party Fortunes?’ Comparative Political Studies, 16 (1984), 419–55.CrossRefGoogle Scholar

⁵⁹ Michael Wallerstein has pointed out to me in a personal communication that Przeworski, and Sprague, 's ‘trade-offs’Google Scholar then follow as a consequence of a simple ‘regression towards the mean’. ‘If social democratic election results are a stochastic process with v(t) = F(x(t)) + u(t) where F(x(t)) represents the predictable aspects of the vote (based on the class structure and past party strategies or whatever you like) and u(t) is a random variable with a mean of zero, then an unusually good election is likely to indicate a high value of u(t). The probability that u(t + 1) < u(t), and therefore that v(t + 1) < v(t) if x doesn't change very much, will be high. When u(t) is unusually low, then it is likely that u(t + 1) > u(t) and the social democrats will do better in the subsequent election.’

⁶⁰ See, for a thorough critique, Sainsbury, Diane, ‘Party Strategies and the Electoral Trade-Off of Class-Based Parties’, European Journal of Political Research, 18 (1990), 29–50.CrossRefGoogle Scholar

⁶¹ Data sources are as follows: Britain (1951–87): Särlvik, and Crewe, , Decade of Dealignment, p. 87Google Scholar and Crewe, Ivor, The Decline of Labour and the Decline of Labour: Social and Electoral Trends in Post-War Britain (paper prepared for delivery at the 1990 American Political Science Association Meeting in San Francisco), p. 7Google Scholar; Denmark (1964–84), Norway (1957–85), and Sweden (1956–85): Sainsbury, Diane, ‘Party Strategies and the Electoral Trade-Off of Class-Based Parties’Google Scholar; France: entire Left, legislative elections 1956–88 in Dalton, Russell J., Citizen Politics in Western Democracies (Chatham: Chatham House, 1988), p. 156Google Scholar; for socialists 1956–68, Przeworski, and Sprague, , Paper Stones, p. 154Google Scholar; for socialists and entire Left, 1973–1988, as well as communists 1973–88: Le Gall, Gérard, ‘Printemps 1988: Retour une gauche majoritaire’, Revue politique et parlementaire, 19 (1988) No. 942), 11–20Google Scholar; Germany (1957–87): Pappi, Franz Urban, ‘Parteiensystem und Sozialstruktur in der Bundesrepublik’, Politische Vierteljahresschriften, 14 (1973), 191–214, at p. 199Google Scholar; and Berger, Manfred, Gibowski, Wolfgang G., Roth, Dieter and Schulte, Wolfgang, ‘Die Konsolidierung der Wende: Eine Analyse der Bundestagswahl 1987’, Zeitschrift für Parlamentsfragen, 18 (1987), 253–84Google Scholar; Italy (1963, 1968, 1976): calculated from Farneti, Paolo, The Italian Party System (1945–1980) (New York: St. Martin's Press, 1985), p. 98.Google Scholar

⁶² The unstandardized regression coefficient b indicates the rise (positive sign) or decline (negative sign) of the proportion of workers' support from one election to the next as a function of the intercept and the percentage of allies voting socialist. For example, in Norway, the Left would receive 67.9 per cent more leftist support if no allies at all voted for the Left in the previous election. If 25 per cent of the allies voted Left in the previous election, socialist working-class support would increase by 67.9 − (2.14 × 25) = 14.4 per cent. On the other hand, if 35 per cent of the allies voted Left in the previous election, socialist working-class support would change by 67.9 − (2.14 × 35) = −7.0 per cent.

⁶³ Przeworski, and Sprague, , Paper Stones, p. 175Google Scholar, report much stronger correlations. In Norway the correlation is −0.65 and in France −0.88 for the entire Left and in the same range for the socialists. The authors, however, include only four to seven time-points, compared to the seven to thirteen time-points in my own data set.

⁶⁴ Data can be obtained from the author. The negative relationship is significant for Denmark, France (PCF) and Germany in data set I and for Belgium, France (PS and PCF), Germany and the Netherlands in data set II.

⁶⁵ Przeworski and Sprague do not include Britain, but the British Labour party does not face a communist competitor or a party with strong cross-class appeals and has close ties to the labour unions.

⁶⁶ Based on the national election studies from the 1950s to the 1980s, the average Alford Index for the Swedish Left has been 28.2, for Britain 37.5, for Norway 37.6, and for Denmark 39.7.

⁶⁷ National election studies show an average Alford Index of 1.8 for the French socialists, 6.0 for the Italian socialists, 12.0 for the French communists, 17.0 for the Italian communists and 17.8 for the German SPD. Eurobarometer Alford Indices for the 1976–88 period are 12.8 for the Belgian socialists, 17.3 for the Dutch Labour party, 1.1 for the French socialists, 0.5 for the Italian socialists, 10.4 for the French communists and 21.7 for the Italian communists.

⁶⁸ Przeworski, and Sprague, , Paper Stones, p. 162.Google Scholar

⁶⁹ I am presenting regression equations only for data set I because data set II includes less than two decades of elections. Findings for the latter have the correct signs, but are insignificant.